Eco Case Study

giggles.desmoin

IMFFullReport2017.pdf

Home >Business & Finance homework help >Economics homework help >Eco Case Study

Global Financial Stability Report, October 2017

G lobal Financial Stability R

eport Global Financial Stability Report

Wo r l d E c o n o m i c a n d F i n a n c i a l S u r v e y s

I N T E R N A T I O N A L M O N E T A R Y F U N D

17OCT

IM F

O C T

Is G row

th at R isk?

Is Growth at Risk?

Blank

Global Financial Stability Report October 2017

Is Growth at Risk?

Wo r l d E c o n o m i c a n d F i n a n c i a l S u r v e y s

I N T E R N A T I O N A L M O N E T A R Y F U N D

Cover and Design: Luisa Menjivar and Jorge Salazar Composition: AGS, An RR Donnelley Company

Cataloging-in-Publication Data

Joint Bank-Fund Library

Names: International Monetary Fund. Title: Global financial stability report. Other titles: GFSR | World economic and financial surveys, 0258-7440 Description: Washington, DC : International Monetary Fund, 2002- | Semiannual | Some issues also have thematic

titles. | Began with issue for March 2002. Subjects: LCSH: Capital market—Statistics—Periodicals. | International finance—Forecasting—Periodicals. |

Economic stabilization—Periodicals. Classification: LCC HG4523.G557

ISBN 978-1-48430-839-4 (Paper) 978-1-48432-056-3 (ePub) 978-1-48432-057-0 (Mobipocket) 978-1-48432-059-4 (PDF)

Disclaimer: The Global Financial Stability Report (GFSR) is a survey by the IMF staff published twice a year, in the spring and fall. The report draws out the financial ramifications of economic issues high- lighted in the IMF’s World Economic Outlook (WEO). The report was prepared by IMF staff and has benefited from comments and suggestions from Executive Directors following their discussion of the report on September 21, 2017. The views expressed in this publication are those of the IMF staff and do not necessarily represent the views of the IMF’s Executive Directors or their national authorities.

Recommended citation: International Monetary Fund. 2017. Global Financial Stability Report: Is Growth at Risk? Washington, DC, October.

Please send orders to: International Monetary Fund, Publications Services P.O. Box 92780, Washington, DC 20090, U.S.A.

Tel.: (202) 623-7430 Fax: (202) 623-7201 E-mail: [email protected]

www.bookstore.imf.org www.elibrary.imf.org

International Monetary Fund | October 2017 iii

CONTENTS

Assumptions and Conventions vi

Further Information and Data vii

Preface viii

Foreword ix

Executive Summary x

IMF Executive Board Discussion Summary xiv

Chapter 1 Is Growth at Risk? 1 Financial Stability Overview 1 Large Systemic Banks and Insurers: Adapting to the New Environment 2 Monetary Policy Normalization: A Two-Sided Risk 17 Has the Search for Yield Gone Too Far? 23 The Rise in Leverage 32 Could Rising Medium-Term Vulnerabilities Derail the Global Recovery? 42 Box 1.1. A Widening Divergence between Financial and Economic Cycles 47 Box 1.2. Cyberthreats as a Financial Stability Risk 49 References 51

Chapter 2 Household Debt and Financial Stability 53 Summary 53 Introduction 54 How Does Household Debt Affect Macroeconomic and Financial Stability? 56 Developments in Household Debt around the World 58 Financial Stability Risks of Household Debt: Empirical Analysis 62 Conclusions and Policy Implications 70 Box 2.1. Long-Term Growth and Household Debt 72 Box 2.2. Distributional Aspects of Household Debt in China 73 Box 2.3. A Comparison of US and Canadian Household Debt 75 Box 2.4. The Nexus between Household Debt, House Prices, and Output 77 Box 2.5. The Impact of Macroprudential Policies on Household Credit 79 Annex 2.1. Data Sources 81 Annex 2.2. Methodology 84 References 87

Chapter 3 Financial Conditions and Growth at Risk 91 Summary 91 Introduction 92 Financial Conditions and Risks to Growth: Conceptual Issues 93 How Do Changes in Financial Conditions Indicate Risks to Growth? 96 How Well Do Changes in Financial Conditions Forecast Downside Risks to Growth? 100 Policy Implications 108 Annex 3.1. Financial Vulnerabilities and Growth Hysteresis in Structural Models 109 Annex 3.2. Estimating Financial Conditions Indices 113

iv International Monetary Fund | October 2017

G LO B A L F I N A N C I A L S TA B I L I T Y R E P O R T: I S G R O W T H AT R I S K ?

Annex 3.3. The Conditional Density of Future GDP Growth 115 References 116

Tables

Table 1.1. Sovereign and Nonfinancial Private Sector Debt-to-GDP Ratios 34 Annex Table 2.1.1. Countries Included in the Sample for Household Debt and Data Sources 81 Annex Table 2.1.2. Household Survey Data Sources 82 Annex Table 2.1.3. Description of Explanatory Variables Used in the Chapter 83 Annex Table 2.2.1. Logit Analysis: Probability of Systemic Banking Crisis 84 Annex Table 2.2.2. Panel Regression Estimates for Three-Year-Ahead Growth Regression on

Household Debt and Policy Interaction Variables 86 Table 3.1. Forecast of GDP Growth Distribution for the Global Financial Crisis with and

without Financial Conditions Indices 104 Table 3.2. Market Consensus Forecasts for the Global Financial Crisis Were Considerably

More Optimistic Than Forecasts Based on Financial Conditions 104 Table 3.3. Forecast of GDP Growth Distribution for the Global Financial Crisis: Comparing

Partitioned and Univariate Financial Conditions Indices with Autoregressions 105 Annex Table 3.2.1. Country Coverage 113 Annex Table 3.2.2. Data Sources 114 Annex Table 3.2.3. Partitioning of Financial Indicators into Groups 115

Figures Figure 1.1. Global Financial Stability Map: Risks and Conditions 2 Figure 1.2. Global Financial Stability Map: Assessment of Risks and Conditions 3 Figure 1.3. Search for Yield, Asset Valuations, and Volatility 4 Figure 1.4. Global Systemically Important Banks: Significance and Business Model Snapshot 6 Figure 1.5. Global Systemically Important Banks: Capital, Liquidity, and Legacy Challenges 7 Figure 1.6. Global Systemically Important Banks: Market Activity 9 Figure 1.7. Global Systemically Important Banks’ International Activity 10 Figure 1.8. Global Systemically Important Banks: Financial Performance Gaps 12 Figure 1.9. Life Insurance Companies’ Profitability and Capital 13 Figure 1.10. Changes in Life Insurance Companies’ Business Models 14 Figure 1.11. Life Insurers’ Market Valuations and Risk Outlook 16 Figure 1.12. Simulated Mark-to-Market Shocks to Assets and Liabilities 17 Figure 1.13. Central Bank Balance Sheets and the Sovereign Sector 19 Figure 1.14. Policy Rates, 10-Year Government Bond Yields, and Term Premiums 20 Figure 1.15. Emerging Market Economy Capital Flows 22 Figure 1.16. Global Fixed Income Markets and US Corporate Credit Investor Base 24 Figure 1.17. Emerging Market Economies: Debt Issuance, Portfolio Flows, and Asset Prices 25 Figure 1.18. Low-Income Country External Borrowing and Vulnerabilities 26 Figure 1.19. US and Emerging Market Corporate Bond Spread Decomposition and Leverage 27 Figure 1.20. Long-Term Drivers of the Low-Volatility Regime 29 Figure 1.21. Leveraged and Volatility-Targeting Strategies 30 Figure 1.22. Vulnerability of the US Corporate Credit Investor Base to Shocks 31 Figure 1.23. Group of Twenty Nonfinancial Sector Credit Trends 33 Figure 1.24. Group of Twenty Nonfinancial Private Sector Borrowing 35 Figure 1.25. Group of Twenty Nonfinancial Private Sector Credit and Debt Service Ratios 36 Figure 1.26. Chinese Banking System Developments 38 Figure 1.27. China: Regulatory Tightening Has Helped Contain Financial Sector Risks 39 Figure 1.28. Chinese Banks: Financial Policy Tightening and Credit Growth Capacity 40 Figure 1.29. Bank Profitability and Liquidity Indicators 41 Figure 1.30. Global Financial Dislocation Scenario 43 Figure 1.31. Emerging Market Economy External Vulnerabilities and Corporate Leverage 45 Figure 1.1.1. Financial and Economic Cycles 48

C o n t e n t s

International Monetary Fund | October 2017 v

Figure 2.1. Household Debt-to-GDP Ratio in Advanced and Emerging Market Economies 54 Figure 2.2. First- and Second-Round Effects of the Buildup of Household Debt on Financial Stability 57 Figure 2.3. Growth and Composition of Household Debt by Region 60 Figure 2.4. Household Debt: Evidence from Cross-Country Panel Data 61 Figure 2.5. Effects of Household Debt on GDP Growth and Consumption 64 Figure 2.6. Effects of Household Debt on GDP Growth: Robustness Tests 65 Figure 2.7. Micro-Level Evidence Corroborating the Macro Impact 66 Figure 2.8. Banking Crises and the Role of Household Debt 67 Figure 2.9. Bank Equity Returns and Household Debt 68 Figure 2.10. The Impact of Household Debt by Country and Institutional Factors 69 Figure 2.1.1. Long-Term per Capita GDP Growth and Household Debt 72 Figure 2.2.1. Characteristics of China’s Household Debt 73 Figure 2.3.1. US and Canadian Household Debt Developments and Characteristics 75 Figure 2.4.1. Panel Vector Autoregression Dynamic Analysis 77 Figure 2.4.2. Consumption Response to House Prices 78 Figure 2.5.1. Macroprudential Policy Tools and Household Credit Growth 79 Annex Figure 2.1.1. Loan Characteristics, Rules, and Regulations 82 Figure 3.1. Tighter Financial Conditions Forecast Greater Downside Tail Risk to Global Growth 97 Figure 3.2. Risk of Severe Recessions Is Especially Sensitive to a Tightening of Financial Conditions

in Major Advanced and Emerging Market Economies 98 Figure 3.3. In Emerging Market Economies, Changes in Financial Conditions Also Affect Upside Risks 99 Figure 3.4. Higher Price of Risk Is a Significant Predictor of Downside Growth Risks within One Year 101 Figure 3.5. Rising Leverage Signals Higher Downside Growth Risks at Longer Time Horizons 102 Figure 3.6. Waning Global Risk Appetite Signals Imminent Downside Risks to Growth 102 Figure 3.7. Probability Densities of GDP Growth for the Depths of the Global Financial Crisis 103 Figure 3.8. In-Sample and Recursive Out-of-Sample Quantile Forecasts: One Quarter Ahead 106 Figure 3.9. In-Sample and Recursive Out-of-Sample Quantile Forecasts: Four Quarters Ahead 107 Annex Figure 3.1.1. Conditional Densities of Growth with High and Low Asset

Prices—One-Period-Ahead Forecasts 110 Annex Figure 3.1.2. One-Period-Ahead GDP and Financial Conditions 111 Annex Figure 3.1.3. Asset Prices and Credit Aggregates before and after a Financial Crisis 111 Annex Figure 3.1.4. Simple Debt Tax Ameliorates Risk of Leverage-Induced Recessions 112

vi International Monetary Fund | October 2017

ASSUMPTIONS AND CONVENTIONS

The following conventions are used throughout the Global Financial Stability Report (GFSR):

. . . to indicate that data are not available or not applicable;

— to indicate that the figure is zero or less than half the final digit shown or that the item does not exist;

– between years or months (for example, 2016–17 or January–June) to indicate the years or months covered, including the beginning and ending years or months;

/ between years or months (for example, 2016/17) to indicate a fiscal or financial year.

“Billion” means a thousand million.

“Trillion” means a thousand billion.

“Basis points” refers to hundredths of 1 percentage point (for example, 25 basis points are equivalent to ¼ of 1 percentage point).

If no source is listed on tables and figures, data are based on IMF staff estimates or calculations.

Minor discrepancies between sums of constituent figures and totals shown reflect rounding.

As used in this report, the terms “country” and “economy” do not in all cases refer to a territorial entity that is a state as understood by international law and practice. As used here, the term also covers some territorial entities that are not states but for which statistical data are maintained on a separate and independent basis.

The boundaries, colors, denominations, and any other information shown on the maps do not imply, on the part of the International Monetary Fund, any judgment on the legal status of any territory or any endorsement or acceptance of such boundaries.

1CHAPTE R

International Monetary Fund | October 2017 vii

This version of the Global Financial Stability Report (GFSR) is available in full through the IMF eLibrary (www. elibrary.imf.org) and the IMF website (www.imf.org).

The data and analysis appearing in the GFSR are compiled by the IMF staff at the time of publication. Every effort is made to ensure, but not guarantee, their timeliness, accuracy, and completeness. When errors are discovered, there is a concerted effort to correct them as appropriate and feasible. Corrections and revisions made after publica- tion are incorporated into the electronic editions available from the IMF eLibrary (www.elibrary.imf.org) and on the IMF website (www.imf.org). All substantive changes are listed in detail in the online tables of contents.

For details on the terms and conditions for usage of the contents of this publication, please refer to the IMF Copyright and Usage website, www.imf.org/external/terms.htm.

FURTHER INFORMATION AND DATA

viii International Monetary Fund | October 2017

The Global Financial Stability Report (GFSR) assesses key risks facing the global financial system. In normal times, the report seeks to play a role in preventing crises by highlighting policies that may mitigate systemic risks, thereby contributing to global financial stability and the sustained economic growth of the IMF’s member countries.

The analysis in this report has been coordinated by the Monetary and Capital Markets (MCM) Department under the general direction of Tobias Adrian, Director. The project has been directed by Peter Dattels and Dong He, both Deputy Directors, as well as by Claudio Raddatz and Matthew Jones, both Division Chiefs. It has ben- efited from comments and suggestions from the senior staff in the MCM Department.

Individual contributors to the report are Ali Al-Eyd, Zohair Alam, Adrian Alter, Sergei Antoshin, Magally Bernal, André Leitão Botelho, Luis Brandão-Marques, Jeroen Brinkhoff, John Caparusso, Sally Chen, Shiyuan Chen, Yingyuan Chen, Charles Cohen, Claudia Cohen, Fabio Cortes, Dimitris Drakopoulos, Kelly Eckhold, Martin Edmonds, Jesse Eiseman, Jennifer Elliott, Aquiles Farias, Alan Xiaochen Feng, Caio Ferreira, Tamas Gaidosch, Rohit Goel, Hideo Hashimoto, Sanjay Hazarika, Dong He, Geoffrey Heenan, Dyna Heng, Paul Hiebert, Henry Hoyle, Nigel Jenkinson, David Jones, Mitsuru Katagiri, Will Kerry, Jad Khallouf, Robin Koepke, Romain Lafarguette, Tak Yan Daniel Law, Feng Li, Yang Li, Peter Lindner, Xiaomeng Lu, Sheheryar Malik, Rebecca McCaughrin, Kei Moriya, Aditya Narain, Machiko Narita, Vladimir Pillonca, Thomas Piontek, Breanne Rajkumar, Mamoon Saeed, Luca Sanfilippo, Jochen Schmittmann, Yves Schüler, Dulani Seneviratne, Juan Solé, Ilan Solot, Yasushi Sugayama, Jay Surti, Narayan Suryakumar, Nico Valckx, Francis Vitek, Changchun Wang, Jeffrey Williams, Christopher Wilson, and Xinze Yao. Magally Bernal, Breanne Rajkumar, and Claudia Cohen were responsible for word processing.

Gemma Diaz from the Communications Department led the editorial team and managed the report’s produc- tion with support from Linda Kean and editorial assistance from Sherrie Brown, Lorraine Coffey, Susan Graham, Lucy Scott Morales, Nancy Morrison, Katy Whipple, AGS, and Vector.

This particular issue of the GFSR draws in part on a series of discussions with banks, securities firms, asset management companies, hedge funds, standard setters, financial consultants, pension funds, central banks, national treasuries, and academic researchers.

This GFSR reflects information available as of September 22, 2017. The report benefited from comments and suggestions from staff in other IMF departments, as well as from Executive Directors following their discussion of the GFSR on September 21, 2017. However, the analysis and policy considerations are those of the IMF staff and should not be attributed to Executive Directors or their national authorities.

PREFACE

T wice a year, the Global Financial Stability Report (GFSR) assesses the degree to which developments in the financial sector may affect future economic conditions by

analyzing macro-financial linkages and then identifies policies to mitigate risks to growth from the financial sector. At the current juncture, investor risk appetite is buoyant globally: since the last report in April, funding conditions have continued to improve, asset return volatility has receded to multiyear lows across markets, and global capital flows have surged. This easing of financial conditions has supported global growth and financial inclusion, with credit being allocated to benefit a broad range of borrowers. These favorable conditions create a window of opportunity to strengthen the financial system that should be seized, since experience has taught us that it is during times of easy financial conditions that vulnerabilities build.

Chapter 1 of this GFSR documents how the con- tinuation of monetary accommodation in advanced economies—necessary to support activity and boost inflation—is associated with rising asset valuations and higher leverage, and how this environment makes the system more vulnerable to future shocks. Chapter 2 focuses on household leverage, showing that ample credit growth portends benign conditions in the near term but larger downside risks in the medium term—and thus creates an intertemporal tradeoff. Chapter 3 takes this logic a step further and directly links the easing of financial conditions to downside risks to GDP growth. Easy financial conditions fuel growth in the shorter term, but when those condi- tions are coupled with a buildup in leverage, risks to growth rise in the medium term. In fact, we propose to measure financial stability by a measure of Growth

at Risk, defined as the value at risk of future GDP growth as a function of financial vulnerability.

The analysis in all three chapters underscores that some of the factors that have contributed to recent gains in financial stability could put growth at risk in the medium term in the absence of appropriate policies to address rising financial vulnerabilities. Macropru- dential policies, such as those that address underwriting standards, are the primary tool for guarding against future risks to growth from the global financial system. Now is the time to further strengthen that system, particularly by focusing on nonbank institutions, whose vulnerabilities are rising. Macroprudential policies that mitigate the buildup of medium-term risks can also help to better balance monetary policy tradeoffs.

Whereas vulnerabilities are rising in the nonbank financial system, the safety of the global systemically important banks (GSIBs) has improved significantly. Those banks have more capital and more liquidity and are subject to tighter supervision, thanks to the pivotal reforms undertaken after the 2008 global financial crisis. Yet some GSIBs still struggle to adapt their business models to ensure their continued health and profitability, which is critical if they are to fulfill their primary mandate: lending to the real economy. A review of the unintended consequences of the postcri- sis regulatory reforms will likely lead to some stream- lining in the implementation of banking regulations, but it is essential that the overall high level of capital and liquidity be preserved, regulatory uncertainty be avoided, and the global financial regulatory reform agenda be completed. Equally essential is continuing international regulatory cooperation.

Tobias Adrian Financial Counsellor

International Monetary Fund | October 2017 ix

FOREWORD

x International Monetary Fund | October 2017

Near-Term Risks Are Lower The global financial system continues to strengthen

in response to extraordinary policy support, regulatory enhancements, and the cyclical upturn in growth. The health of banks in many advanced economies contin- ues to improve, as progress has been made in resolv- ing some weaker banks, while a majority of systemic institutions are adjusting business models and restoring profitability. The upswing in global economic activity, discussed in the October 2017 World Economic Outlook (WEO), has boosted market confidence while reducing near-term threats to financial stability.

But beyond these recent improvements, the environ- ment of continuing monetary accommodation—neces- sary to support activity and boost inflation—is also leading to rising asset valuations and higher leverage. Financial stability risks are shifting from the bank- ing system toward nonbank and market sectors of the financial system. These developments and risks call for delicately balancing the eventual normalization of monetary policies, while avoiding a further buildup of financial risks outside the banking sector and address- ing remaining legacy problems.

The Two Sides of Monetary Policy Normalization

The baseline path for the global economy, envisaged by central banks and financial markets, foresees contin- ued support from accommodative monetary policies, as inflation rates are expected to recover only slowly. Thus, the gradual process of normalizing monetary policies is likely to take several years. Too fast a pace of nor- malization would remove needed support for sustained recovery and desired increases in core inflation across major economies. Unconventional monetary policies and quantitative easing have forced substantial portfolio adjustments in the private sector and across borders, making the adjustment of financial markets much less predictable than in previous cycles. Abrupt or ill-timed shifts could cause unwanted turbulence in financial markets and reverberate across borders and markets. Yet the prolonged monetary support envisaged for the major

economies may lead to the buildup of further financial excesses. As the search for yield intensifies, vulnerabilities are shifting to the nonbank sector, and market risks are rising. There is too much money chasing too few yield- ing assets: less than 5 percent ($1.8 trillion) of the cur- rent stock of global investment-grade fixed-income assets yields over 4 percent, compared with 80 percent ($15.8 trillion) before the crisis. Asset valuations are becoming stretched in some markets as investors are pushed out of their natural risk habitats, and accept higher credit and liquidity risk to boost returns.

At the same time, indebtedness among the major global economies is increasing. Leverage in the non- financial sector is now higher than before the global financial crisis in the Group of Twenty economies as a whole. While this has helped facilitate the economic recovery, it has left the nonfinancial sector more vul- nerable to changes in interest rates. The rise in leverage has led to a rise in private sector debt service ratios in several of the major economies, despite the low level of interest rates. This is stretching the debt servicing capacity of weaker borrowers in some countries and sectors. Debt servicing pressures and debt levels in the private nonfinancial sector are already high in several major economies (Australia, Canada, China, Korea), increasing their sensitivity to tighter financial condi- tions and weaker economic activity.

The key challenge confronting policymakers is to ensure that the buildup of financial vulnerabilities is contained while monetary policy remains supportive of the global recovery. Otherwise, rising debt loads and overstretched asset valuations could undermine market confidence in the future, with repercussions that could put global growth at risk. This report exam- ines such a downside scenario, in which a repricing of risks leads to sharp increases in credit costs, falling asset prices, and a pullback from emerging markets. The economic impact of this tightening of global financial conditions would be significant (about one- third as severe as the global financial crisis) and more broad-based (global output would fall 1.7 percent relative to the WEO baseline with varying cross-coun- try effects). Monetary normalization would go into

EXECUTIVE SUMMARY

e x e C u t I v e s u M M a r y

International Monetary Fund | October 2017 xi

reverse in the United States and would stall elsewhere. Emerging market economies would be disproportion- ately affected, resulting in an estimated $100 billion reduction in portfolio flows over four quarters. Bank capital would take the biggest hit where leverage is highest and where banks are most exposed to the housing and corporate sectors.

Deleveraging in China: Challenges Ahead Steady growth in China and financial policy tighten-

ing in recent quarters have eased concerns about a near-term slowdown and negative spillovers to the global economy. However, the size, complexity, and pace of growth in China’s financial system point to elevated financial stability risks. Banking sector assets, at 310 percent of GDP, have risen from 240 percent of GDP at the end of 2012. Furthermore, the grow- ing use of short-term wholesale funding and “shadow credit” to firms has increased vulnerabilities at banks. Authorities face a delicate balance between tightening financial sector policies and slowing economic growth. Reducing the growth of shadow credit even modestly would weigh on the profitability and broader provision of credit by small and medium-sized banks.

Global Banks’ Health Is Improving The health of global systemically important banks

(GSIBs) continues to improve. Balance sheets are stron- ger because of improved capital and liquidity buffers, amid tighter regulation and heightened market scrutiny. Considerable progress has been made in addressing legacy issues and restructuring challenges. At the same time, while many banks have strengthened their profit- ability by reorienting business models, several continue to grapple with legacy issues and business model chal- lenges. Banks representing about $17 trillion in assets, or about one-third of the GSIB total, may continue to generate unsustainable returns, even in 2019. As problems in even a single GSIB could generate systemic stress, supervisory actions should remain focused on business model risks and sustainable profitability. Life insurers have also been adapting their business strate- gies in the low-yield environment following the global financial crisis. They have done this by reducing legacy exposures, steering the product mix away from high guaranteed returns, and seeking higher yields in invest- ment portfolios. Meanwhile, supervisors need to moni- tor rising exposure to market and credit risks.

Policymakers Must Take Proactive Measures Policymakers must take advantage of the improving

global outlook and avoid complacency by addressing rising medium-term vulnerabilities. • Policymakers and regulators should fully address

crisis legacy problems and require banks and insur- ance companies to strengthen their balance sheets in advanced economies. This includes putting a resolution framework for international banks into operation, focusing on risks from weak bank busi- ness models to ensure sustainable profitability, and finalizing Basel III. Regulatory frameworks for life insurers should be enhanced to increase reporting transparency and incentives to build resilience. A global and coordinated policy response is needed for resilience to cyberattacks (see Box 1.2).

• Major central banks should ensure a smooth normalization of monetary policy through well- communicated plans on unwinding their holdings of securities and guidance on prospective changes to policy frameworks. Providing clear paths for policy changes will help anchor market expectations and ward off undue market dislocations or volatility.

• Financial authorities should deploy macroprudential measures, and consider extending the boundary of such tools, to curb rising leverage and contain grow- ing risks to stability. For instance, borrower-based measures should be introduced and/or tightened to slow fast-growing overvalued segments, and bank stress tests must assume more stressed asset valua- tions. Capital requirements should be increased for banks that are more exposed to vulnerable borrowers to act as a cushion for already accumulated expo- sures and incentivize banks to grant new loans to less risky sectors.

• Regulation of the nonbank financial sector should be strengthened to limit risk migration and excessive capital market financing. Transition to risk-based supervision should be accelerated, and harmonized regulation of insurance companies—with emphasis on capital—should be introduced. Tighter micro- prudential requirements should be implemented in highly leveraged segments.

• Debt overhangs—especially among the largest borrowers as potential originators of shocks—must be addressed. Discouraging further debt buildup through measures that encourage business invest- ment and discourage debt financing will help curb financial risk taking.

xii International Monetary Fund | October 2017

G LO B A L F I N A N C I A L S TA B I L I T Y R E P O R T: I S G R O W T H AT R I S K ?

• Emerging market economies should continue to take advantage of supportive external conditions to enhance their resilience, including by continu- ing to strengthen external positions where needed, and reduce corporate leverage where it is high. This would put these economies in a better position to withstand a reduction in capital inflows as a result of monetary normalization in advanced economies or waning global risk appetite. Similarly, frontier market and low-income-country borrowers should develop the institutional capacity to deal with risks from the issuance of marketable securities, including formulating comprehensive medium-term debt man- agement strategies. This will enable them to take advantage of broader financial market development and access, while containing the associated risks.

• In China, the authorities have taken welcome steps to address risks in the financial system, but there is still work to do. Vulnerabilities will be difficult to address without slower credit growth. Recent policies to improve the risk management and transparency of the banking system and reduce the buildup of maturity and liquidity transformation risks in banks’ shadow credit activities are essential and must continue. How- ever, policies should also target balance sheet vulnera- bilities at weak banks. The government’s commitment to reducing corporate leverage is welcome and should remain a priority as part of a broader effort to insulate the economy against slower credit growth.

• Although significant progress has been made in developing the postcrisis policy response, progress remains uneven across the various sectors, with several design and implementation issues remain- ing outstanding. Ensuring that the reform mea- sures are completed and implemented is essential to minimize the likelihood of another disruptive crisis. Completing the reform agenda will also allow policymakers to conduct a comprehensive evalua- tion of the impact of the reforms and fine-tune the agreed measures. This will allow them to address any material unintended effects their cumulative implementation might have on the provision of key financial services. This is critical to provide contin- ued assurance that reforms have delivered on their objectives and to stave off emerging pressures to roll back these measures, which would only make the financial system more vulnerable.

• Finally, implementation of structural reforms and supportive fiscal policies (as examined in Scenario

Box 1 of the October 2017 World Economic Outlook) would lift global growth and generate positive eco- nomic spillovers, reinforcing financial policy efforts.

Household Debt and Economic Growth Chapter 2 examines the short- and medium-term

implications for economic growth and financial stability of the past decades’ rise in household debt. The chapter documents large differences in household debt-to-GDP ratios across countries but a common increasing trajec- tory that was moderated but not reversed by the global financial crisis. In advanced economies, with notable exceptions, household debt to GDP increased gradu- ally, from 35 percent in 1980 to about 65 percent in 2016, and has kept growing since the global financial crisis, albeit more slowly. In emerging market econo- mies, the same ratio is still much lower, but increased relatively faster over a shorter period, from 5 percent in 1995 to about 20 percent in 2016. Moreover, the rise has been largely unabated in recent years. The chapter finds a trade-off between a short-term boost to growth from higher household debt and a medium-term risk to macroeconomic and financial stability that may result in lower growth, consumption, and employment and a greater risk of banking crises. This trade-off is stronger when household debt is higher and can be attenuated by a combination of good policies, institutions, and regulations. These include appropriate macroprudential and financial sector policies, better financial supervision, less dependence on external financing, flexible exchange rates, and lower income inequality.

Financial Conditions Can Predict Growth The global financial crisis showed policymakers

that financial conditions offer valuable information about risks to future growth and provide a basis for targeted preemptive action. Chapter 3 develops a new macroeconomic measure of financial stability by link- ing financial conditions to the probability distribution of future GDP growth and applies it to a set of 21 major advanced and emerging market economies. The chapter shows that changes in financial conditions shift the whole distribution of future GDP growth. Wider risk spreads, rising asset price volatility, and waning global risk appetite are significant predictors of increased downside risks to growth in the near term, and higher leverage and credit growth provide

e x e C u t I v e s u M M a r y

International Monetary Fund | October 2017 xiii

relevant signals of such risks in the medium term. Today’s prevailing low funding costs and financial market volatility support a sanguine view of risks to the global economy in the near term. But increasing leverage signals potential risks down the road, and a scenario of a rapid decompression in spreads and volatility could significantly worsen the risk outlook for global growth. A retrospective real-time analysis

of the global financial crisis shows that forecasting models augmented with financial conditions would have assigned a considerably higher likelihood to the economic contraction that followed than those based on recent growth alone. This confirms that the analytical approach developed in the chapter can be a significant addition to policymakers’ macro-financial surveillance toolkit.

xiv International Monetary Fund | October 2017

E xecutive Directors broadly shared the assess- ment of global economic prospects and risks. They observed that global activity has strengthened further and is expected to rise

steadily into next year. The pickup is broad based across countries, driven by investment and trade. Nev- ertheless, the recovery is not complete, with medium- term global growth remaining modest, especially in advanced economies and fuel exporters. In most advanced economies, inflation remains subdued amid weak wage growth, while slow productivity growth and worsening demographic profiles weigh on medium- term prospects. Meanwhile, several emerging markets and developing economies continue to adjust to a range of factors, including lower commodity revenues.

Directors noted that, while risks are broadly bal- anced in the near term, medium-term risks remain skewed to the downside, with rising financial vulnera- bilities. These include the possibility of a sudden tight- ening of global financial conditions, a rapid increase in private sector debt in key emerging market economies, low bank profitability and pockets of still-elevated non- performing loan ratios, and policy uncertainty about financial deregulation. Directors also pointed to risks associated with inward-looking policies, rising geopo- litical tensions, and weather-related factors.

Given this landscape, Directors underscored the continued importance of employing a range of policy tools, in a comprehensive, consistent, and well- communicated manner, to secure the recovery and improve medium-term prospects. They recognized that major central banks have made every effort to commu- nicate their monetary normalization policies to markets. The cyclical upturn in economic activity provides a window of opportunity to accelerate critical structural reforms, increase resilience, and promote inclusiveness.

Directors stressed that a cooperative multilateral framework remains vital for amplifying the mutual benefits of national policies and minimizing any

cross-border spillovers. Common challenges include maintaining the rules-based, open trading system; preserving the resilience of the global financial system; avoiding competitive races to the bottom in taxation and financial regulation; and further strengthening the global financial safety net. Multilateral cooperation is also essential to tackle various noneconomic challenges, among which are refugee flows, cyberthreats and, as most Directors highlighted, mitigating and adapting to climate change. Concerted effort is also needed to reduce excess global imbalances, through a recalibra- tion of policies with a view to achieving their domestic objectives as well as strengthening prospects for strong, sustainable, and balanced global growth. In this con- text, as a few Directors emphasized, the IMF also has a role to play by continuing to strengthen its multilateral analysis of external imbalances and exchange rates.

Directors agreed that continued accommodative monetary policy is still needed in countries with low core inflation, consistent with central banks’ mandates. Fiscal policy should gear toward long-term sustain- ability, avoid procyclicality, and promote inclusive growth. At the same time, fiscal policy should be as growth friendly as possible, using space, where avail- able, to support productivity and growth-enhancing structural reforms. In many cases, policymakers should prioritize rebuilding buffers, improving medium-term debt dynamics, and enhancing resilience. Efforts to raise potential output should be prioritized based on country-specific circumstances, including increasing the supply of labor, upgrading skills and human capi- tal, investing in infrastructure, and lowering product and labor market distortions. Social safety nets remain important to protect those adversely affected by tech- nological progress and other structural transformation.

Directors noted that income disparities among countries have narrowed, but inequality has increased in some economies. They saw a role that well-designed fiscal policies can play in achieving redistributive

IMF EXECUTIVE BOARD DISCUSSION SUMMARY

The following remarks were made by the Chair at the conclusion of the Executive Board’s discussion of the Fiscal Monitor, Global Financial Stability Report, and World Economic Outlook on September 21, 2017.

I M F e x e C u t I v e B o a r d d I s C u s s I o n s u M M a r y

International Monetary Fund | October 2017 xv

objectives without necessarily undermining growth and incentives to work. Directors generally concurred that there may be scope for strengthening means-testing of transfers in many countries and for increasing the progressivity of taxation in some others. Most Direc- tors noted that any consideration of a universal basic income would have to be weighed carefully against a host of country-specific factors—including existing social safety schemes, financing modalities, fiscal cost, and social preferences, as well as its impact on incen- tives to work—which, in the view of many Directors, raised questions about its attractiveness and practical- ity. Directors emphasized that improving education and health care is key to reducing inequality and enhancing social mobility over time.

Directors underlined the continued need for emerg- ing market and developing economies to bolster economic and financial resilience to external shocks, including through enhanced macroprudential policy frameworks and exchange rate flexibility. They noted that a common challenge across these economies is how to speed up their convergence toward living standards in advanced economies. While priorities differ across coun- tries, many need to improve governance, infrastructure, education, and access to health care. In several countries, policies should also facilitate greater labor force partici- pation, reduce barriers to entry into product markets, and enhance the efficiency of credit allocation.

Directors observed that the global financial system continues to strengthen, and market confidence has improved generally. They recognized the substan- tial progress made in resolving weak banks in many advanced economies, while a majority of systemic institutions are adjusting business models and restoring profitability. However, a prolonged period of monetary accommodation could lead to further increases in asset valuations and a buildup of leverage in the nonfi- nancial sector that could signal higher risks to finan- cial stability. These developments call for continued vigilance about household debt ratios and investors’ exposure to market and credit risks. In this context, Directors stressed the need to calibrate the path of nor- malization of monetary policies carefully, implement macro- and microprudential measures as needed, and address remaining legacy problems.

Directors noted a generally subdued outlook for commodity prices. They encouraged low-income developing countries that are commodity export- ers to continue improving revenue mobilization and strengthening debt management, while safeguarding social outlays and capital expenditures. Countries with more diversified export bases should further strengthen fiscal positions and foreign exchange buffers. Across all low-income developing countries, an overarching chal- lenge is to maintain progress toward their Sustainable Development Goals.

Blank

Financial Stability Overview Near-term financial stability risks have declined with the strengthening global recovery, but medium-term vulnera- bilities are building as the search for yield intensifies. Risks are rotating from banks to financial markets as spreads and volatility compress while private sector indebtedness rises.

The Global Recovery Is Improving the Near-Term Outlook for Financial Stability

Near-term risks to financial stability continue to decline. Macroeconomic risks are lower (Figures 1.1 and 1.2) amid the global upswing in economic activ- ity, discussed in the October 2017 World Economic Outlook (WEO). Emerging market risks have also declined, underpinned by the pickup in global activity and benign external conditions. This environment of benign macroeconomic conditions and continued easy monetary and financial conditions—but still sluggish inflation—is fueling a marked increase in risk appetite, broadening investors’ search for yield.

Systemically Important Banks and Insurers Continue to Enhance Resilience

Global systemically important banks (GSIBs) and insurers have strengthened their balance sheets by raising capital and liquidity but are still grappling with remaining legacy issues and business model challenges.

Prepared by staff from the Monetary and Capital Markets Department (in consultation with other departments): Peter Dattels (Deputy Director), Matthew Jones (Division Chief ), Paul Hiebert (Advisor), Ali Al-Eyd (Deputy Division Chief ), Will Kerry (Deputy Division Chief ), Zohair Alam, Sergei Antoshin, Magally Bernal, Luis Brandão-Marques, Jeroen Brinkhoff, John Caparusso, Sally Chen, Shiyuan Chen, Yingyuan Chen, Charles Cohen, Fabio Cortes, Dimitris Drakopoulos, Kelly Eckhold, Martin Edmonds, Jesse Eiseman, Jennifer Elliott, Caio Ferreira, Tamas Gaidosch, Rohit Goel, Hideo Hashimoto, Sanjay Hazarika, Geoffrey Heenan, Dyna Heng, Henry Hoyle, Nigel Jenkinson, David Jones, Jad Khallouf, Robin Koepke, Tak Yan Daniel Law, Yang Li, Peter Lindner, Rebecca McCaughrin, Aditya Narain, Machiko Narita, Vladimir Pillonca, Thomas Piontek, Mamoon Saeed, Luca Sanfilippo, Jochen Schmittmann, Juan Solé, Ilan Solot, Yasushi Sugayama, Narayan Suryakumar, Francis Vitek, Jeffrey Williams, and Christopher Wilson.

After a painful period of restructuring and absorption of elevated charges for past misconduct in the form of fines and private litigation, the outlook for sustainable profitability is improving, but strategic reorientation remains incomplete. The next section assesses risks from large global banks and life insurance companies.

Medium-Term Vulnerabilities Are Rising and Rotating to Nonbanks

Many asset valuations have continued to rise in response to the improved economic outlook and the search for yield (Figure 1.3, panel 1), driving down a broad range of risk premiums (Figure 1.3, panel 2). While increased risk appetite and the search for yield are a welcome and intended consequence of unconven- tional monetary policy measures, helping to support the economic recovery, there are risks if these trends extend too far. Compensation for inflation risks (term premiums) and credit risks (for example, spreads on corporate bonds) are close to historic lows, while volatility across asset markets is now highly compressed (Figure 1.3, panel 3). Some measures of equity valuation are elevated, but relative to yields on safe assets (that is, the equity risk premium) they do not appear overly stretched. This prolonged search for yield has raised the sensitivity of the financial system to market and liquidity risks, keeping those risks elevated. The widening diver- gence between economic and financial cycles within and across the major economies is discussed in Box 1.1.

A key stability challenge is the rebalancing of central bank and private sector portfolios against a backdrop of monetary policy cycles that are not synchronized across countries. Too quick an adjustment in monetary policies could cause unwanted turbulence in financial markets and set back progress toward inflation targets. Too long a period of low interest rates could foster a further buildup of market and credit risks and increase medium-term vulnerabilities.

Credit risks are already elevated, given the deteri- oration in underlying leverage in the nonfinancial sector—households and firms—of many Group of Twenty (G20) economies. Despite low interest rates,

IS GROWTH AT RISK?1CHAPTER

1International Monetary Fund | October 2017

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

private sector debt service ratios in many major econ- omies have increased to high levels because of rising debt. Weaker households and companies in several countries have become more sensitive to financial and economic conditions as a result.

The Global Recovery Could Be Derailed

Prolonged low volatility, further compression of spreads, and rising asset prices could facilitate addi- tional risk taking and raise vulnerabilities further. Investors’ concern about debt sustainability could eventually materialize and prompt a reappraisal of risks. In such a downside scenario, a shock to individ- ual credit and financial markets well within historical norms could decompress risk premiums and reverber- ate worldwide, as explored later in this chapter. This could stall and reverse the normalization of monetary policies and put growth at risk.

Large Systemic Banks and Insurers: Adapting to the New Environment The large internationally active banks at the core of the financial system—so-called global systemically important banks (GSIBs)—have become more resilient since the crisis, with stronger capital and liquidity. Banks have made sub- stantial progress in addressing legacy issues and restructuring challenges—while adapting their business models to the

new regulatory and market landscape. Strategic reorien- tation has led to a pullback from market-related business. Banks have, however, retained a presence in international business and cross-border loans. These strategic realignments have come amid changing group structures, as activity is increasingly channeled through subsidiaries. Despite ongoing improvement, progress is uneven and adaptation remains incomplete. About a third of banks by assets may struggle to achieve sustainable profitability, underscoring ongoing challenges and medium-term vulnerabilities.

Life insurers were hit by the global financial crisis, but have since rebuilt their capital buffers. However, they are now facing the challenge of a low-interest-rate environment. In response, insurers have adapted their business models by changing their product mix and asset allocations. But in doing so, they have been increasingly forced out of their natural risk habitat in a search for yield, making them more vulnerable to market and credit risks. Investors still worry about the viability of some insurers’ business models and find it difficult to assess risks, resulting in weak equity market valuations. Policymakers should seek to strengthen regulatory frameworks and increase reporting transparency.

Global Systemically Important Banks

Global banks remain critical pillars of international financial intermediation. These GSIBs provide a wide range of financial services for companies, institutions,

Global financial crisis

Source: IMF staff estimates. Note: The shaded region shows the global financial crisis as reflected in the stability map of the April 2009 Global Financial Stability Report (GFSR).

Away from center signifies higher risks, easier monetary and financial conditions, or higher risk appetite.

Emerging market risks Credit risks

Market and liquidity risks

Risk appetiteMonetary and financial

Macroeconomic risks

Risks

Conditions

Figure 1.1. Global Financial Stability Map: Risks and Conditions

April 2017 GFSR October 2017 GFSR

Risk appetite has grown markedly as near-term stability risks have declined.

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

–1

–2

–1

Higher risk

–2

–1

Easier

Tighter

Higher risk appetite

Lower risk

Lower risk Lower risk

Higher risk

Overall (8)

Overall (10)

Domestic fundamentals

(4)

Volatility (2)

Corporate financing

(2)

External financing

(2)

Macroeconomic conditions

(4)

Uncertainty (2)

Inflation or deflation risks

(1)

Sovereign risks (1)

Overall (3)

Asset allocation preferences

(1)

Excess returns (1)

Flows to risky assets

(1)

Overall (11)

Banking sector

(4)

Household sector

(3)

Corporate sector

(4)

Overall (12)

Liquidity and funding

(5)

Volatility (2)

Valuations (3)

Position and correlation risks

(2)

Figure 1.2. Global Financial Stability Map: Assessment of Risks and Conditions (Notch changes since the April 2017 Global Financial Stability Report)

1. Macroeconomic risks have fallen, and macroeconomic conditions have improved.

2. Emerging market risks are lower, driven by improved fundamentals and external financing conditions.

3. Credit risks are unchanged, with improvements in the banking sector contrasting with increasing corporate and household sector risks.

4. Monetary and financial conditions remain accommodative, as slightly higher real rates are offset by easier lending conditions and financial conditions.

5. Risk appetite continues to increase, as reflected in robust capital flows to emerging markets and increased performance and allocations to risk assets.

6. Market and liquidity risks are unchanged, as compressed risk premiums and low volatility offset less-extended market positioning and improved trading liquidity conditions.

Source: IMF staff estimates. Note: Changes in risks and conditions are based on a range of indicators, complemented by IMF staff judgment. See Annex 1.1 in the April 2010 Global Financial Stability Report and Dattels and others 2010 for a description of the methodology underlying the global financial stability map. Overall notch changes are the simple average of notch changes in individual indicators. The number in parentheses next to each category on the x-axis indicates the number of individual indicators within each subcategory of risks and conditions. For lending conditions, positive values represent a slower pace of tightening or faster easing.

Overall (6)

Monetary policy

conditions (3)

Financial conditions

index (1)

Lending conditions

(1)

Central bank balance sheet

(1)

Higher risk

Unchanged

Macroeconomic risks

Unchanged

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

1. Search for Yield (Percentile rank)

The global search for yield has compressed risk premiums across some assets ...

2. Cross-Asset Valuations (Percentile rank)

3. Realized Volatility (Percentile rank)

… while volatility remains near precrisis lows.

Sources: Bank of America Merrill Lynch; Bloomberg Finance L.P.; Dealogic; Haver Analytics; Organisation for Economic Co-operation and Development; Thomson Reuters; and IMF staff estimates. Note: The color shading is based on valuation quartiles. Red (dark green) denotes low (high) premiums, spreads, volatility, and issuance quality, as well as high (low) issuance and house price to income. In panel 1, quality of issuance shows spreads per turn of leverage. Quantity of issuance is 12-month trailing gross issuance as percent of the outstanding amount. In panel 2, CAPE is the trailing 12-month price-to-earnings ratio adjusted for inflation and the 10-year earnings cycle. Forward P/E is the 12-month forward price-to-earnings ratio. Equity risk premiums are estimated using a three-stage dividend discount model on major stock indices. Term premium estimates follow the methodology in Wright 2011. Corporate spreads are proxied using spreads per turn of leverage. For house-price-to-income ratio, income is proxied using nominal GDP per capita. The percentile is calculated from 1990 for CAPE, forward P/E, equity risk premiums and term premiums, from 1999 for EM term premiums, from 2000 for house-price-to-income ratio, and from 2007 for corporate spreads. In panel 3, the heatmap shows the percentile of three-month realized volatility since 2003 at a monthly frequency. CAPE = cyclically adjusted price-to-earnings ratio; DM = developed market; EM = emerging market; FX = foreign exchange; Govt = government; P/E = price to earnings.

Figure 1.3. Search for Yield, Asset Valuations, and Volatility

United States 83 79 85 7 6 74

Germany 62 33 86 9 14 39

Japan 28 17 87 5 65 8

United Kingdom 85 60 96 8 8 92

Emerging Markets 25 58 84 19 5 44

CAPE Equity Risk Premiums

Term Premiums (10-year)

Corporate Spreads

House Prices to Income

Forward P/E

High yield

Spreads

Quantity of

Issuance

Quality of

Issuance

12 13 1714 15 162006 07 08 09 10 11

Global financial

crisis

European debt crisis

Precrisis buildup of

risks

Latest0605 1211 1716151413042003 10090807

Oil sell-off, China growth

worries, Brexit, US election

Equity DM 10%

Equity EM 6%

Govt. Bond DM 14%

Credit DM 14%

Credit EM 5%

FX DM 22%

FX EM 35%

Commodities 16%

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

and individuals across many countries.1 Together, these 30 banks hold more than $47 trillion in assets and more than one-third of the total assets and loans of thousands of banks globally. They have an even greater role in certain key global financial functions: collec- tively they comprise 70 percent or more of certain international credit markets (for example, syndicated trade finance), market services, and the international financial infrastructure. GSIBs are central to the inter- national financial system (Figure 1.4, panel 1).

All GSIBs share systemic importance. At the same time, they are a diverse group, with differences in business mix and geographic positions. The 30 GSIBs encompass business models ranging from those that are market focused to those that are consumer focused and from highly specific transaction banking models to all-embracing universal banks (Figure 1.4, panels 3 and 4). About half of GSIBs, by assets, are universal banks, offering a mix of services. Unsurprisingly, most operate on more than one continent. But almost a third of these banks, by assets, are largely domestic businesses (mostly in China and the United States).

GSIBs Are Undergoing Business Model Transitions

In the aftermath of the crisis. GSIBs have been reorienting their business models in three overlapping phases (Figure 1.4, panel 2). First, a process of legacy cleanup has been ongoing for most banks. As these legacy challenges recede, banks have entered a phase of strategic reorientation, which continues to affect both their lines of business and geographic scope. As banks have progressed in these first two phases, the focus is shifting to resolution regimes and the associated need to reconfigure interna- tional group structures for some banks. These multiyear adjustments—still ongoing—have been necessary to

1Global systemically important banks (GSIBs) are identified based on size, interconnectedness, cross-jurisdictional activity, impact on financial institution infrastructure (for example, the payments system), and complexity (Basel Committee on Banking Supervision 2014). GSIBs included in the analysis are based on the list published in November 2016, the latest available at the time of this report, and include the following: China (4)—Agricultural Bank of China (ABC), Bank of China (BOC), China Construction Bank (CCB), Industrial and Com- mercial Bank of China (ICBC); Japan (3)—Mitsubishi UFJ Financial Group (MUFG), Mizuho Financial Group (MFG), Sumitomo Mitsui Financial Group (SMFG); Continental Europe (11)—Banco Santander (SAN), BNP Paribas (BNP), Crédit Agricole (CA), Credit Suisse (CS), Deutsche Bank (DB), Groupe BPCE (BPCE), ING Groep (ING), Nordea Bank (NDA), Société Générale (SG), UBS Group (UBS), Uni- credit Group (UCG); United Kingdom (4)—Barclays (BARC), HSBC Holdings (HSBC), Royal Bank of Scotland (RBS), Standard Chartered (STAN); United States (8)—Bank of America (BOA), Bank of New York Mellon (BNY), Citigroup (C), Goldman Sachs (GS), JP Morgan Chase (JPM), Morgan Stanley (MS), State Street (STT), Wells Fargo (WFC).

support resilience and achieve more sustainable profitabil- ity in the new environment. Progress on these fronts has been positive, but uneven, and challenges remain.

Global Banks Have Fortified Balance Sheets and Continue to Address Crisis Legacies

The resilience of GSIBs has improved over the past decade as they have adapted to enhanced prudential standards. They have significantly strengthened their balance sheets with an additional $1 trillion in capital since 2009 while reducing assets. Adjusted capital ratios (incorporating reserves against expected losses) have in aggregate risen steadily since the undercapitalized precrisis period (Figure 1.5, panel 1). GSIB liquidity has also improved: loan-to-deposit ratios are down from the elevated levels a decade ago, and reliance on short-term wholesale funding has fallen (Figure 1.5, panel 2).

In tandem with higher capital and more liquidity, GSIBs have also made significant progress in dealing with legacy challenges from the 2008–09 financial crisis and its aftermath. • Banks have made progress in cleaning up legacy

assets, facilitated by carving out noncore portfo- lios (mainly legacy impaired loans and bonds) for aggressive disposal and runoff (Figure 1.5, panel 3). About two-thirds of GSIB noncore assets have been disposed of; US GSIBs are the most advanced in this process. In contrast, several European banks continue to take high charges to provide for and write off legacy bad debts.

• Second, charges for past misconduct in the form of fines and private litigation have eased from a high level. These charges totaled an estimated $220 billion between 2011 and 2016, equivalent to 27 percent of underlying net income for European banks over the period and 19 percent for US banks. Although some of these charges were the result of misbehavior in personal financial services (insurance products in the United Kingdom, consumer protection in the United States, private banking tax evasion at the global level), most stemmed from market businesses (US residen- tial mortgage-backed securities, fixing of the London interbank offered rate) and international transactions (anti–money laundering measures) in which GSIBs dominate. From a financial stability point of view, the litigation charges should strengthen incentives for more prudent future business practices.

Despite progress in disposing of legacy assets and dealing with past misconduct, GSIBs continue to cope

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

10 20 30 40 50 60 70 80 90

100

0 25 50 75 100

Bank loans Bank assets

Total exposures Level 3 assets

International loans EM US$ project finance

EM US$ syndicated loans Payments

Underwriting revenues Derivatives

Equity revenue FICC revenue

Wealth managers

Consumer banks

Transaction banks

Investment banks

Corporate banks

Universal banks

Business models

$110 trillion $79 trillion $0.6 trillion $26 trillion $50 billion

$87 billion $60 billion

$600 trillion $48 billion

$2,500 trillion $300 billion

$54 trillion

Global market

Legacy Strategy Structure

Ø NPL cleanup

Ø Portfolio runoff

Ø Conduct charges

Ø Restructuring costs

Ø Line of business adjustments

Ø Geographic scope

Ø Efficiency and capabilities

Ø Subsidiarization

Ø Cross-border funding

Receding Continuing Emerging

Sources: Bank financial statements; Bank for International Settlements; Basel Committee on Banking Supervision; Bloomberg Finance L.P.; Dealogic; Haver Analytics; Office of Financial Research; S&P Capital IQ; SNL Financial; and IMF staff estimates. Note: In panel 1, global market size for total exposures, level 3 assets, payments, and over-the-counter derivatives are calculated using the GSIB indicator metrics. “Total exposure” is a proxy for banks’ total asset exposures, which includes total consolidated assets, derivatives exposures, and certain off-balance-sheet exposures. This is the same as the denominator used for the Basel III ratio. EM US$ project finance includes syndicated loans only. GSIBs’ apparently low share of international loans reflects the nearly pure domestic focus of the local category banks as shown in panel 3. In panel 1, global banking loans and assets are calculated using a sample of 3,500+ banks. See footnote 1 in the text for an explanation of the abbreviations in panels 3 and 4. EM = emerging market; FICC = fixed income, currencies, and commodities; GSIB = global systemically important bank; NPL = nonperforming loan.

Figure 1.4. Global Systemically Important Banks: Significance and Business Model Snapshot

1. GSIBs’ Global Market Share by Asset or Activity, 2016 (or latest) (Percent; US dollars)

2. Bank Business Model Challenges

3. GSIB Business Models and Geographic Strategies

4. GSIBs: Revenue Mix by Line of Business, 2016 (Percent of revenue)

Global Regional Local

Universal Bank Balance of household and business services C, JPM, HSBC, DB, STAN, BNP, MUFG

CA BOA, ABC, CCB, ICBC

Corporate Bank Lending to businesses BARC, SMFG UCG, MFG

Investment Bank

Capital markets services, advisory, mergers, and secondary market sales and trading

GS, CS 3

Transaction Bank

Corporate transaction services (including payments) and institutional services (settlement, clearing, custody)

BNY, STT 1

Consumer Bank

Retail banking including lending (mortgages, credit cards, other unsecured credit), savings products, and retail payment services

ING, SAN, SG NDA, BOC, RBS

BPCE, WFC 23

Wealth Manager

Asset management, private banking, and insurance MS, UBS 4

52 18 31 100

Percent of GSIB Assets

Business Model

Description Geographic Reach

International credit Market services

Market infrastructure Overall balance sheet exposure

M S

U B

B O

S G

W FC

N D

R B

B PC

S A

IN G

S TT

B N

S M

M FG

B A

R C

U C

G S

C S

IC B

A B

M U

FGD B

C C

B O

S TA

JP M

B N

H S

B C

C A

Markets Corporate and investment banking Wealth management Consumer Transaction Commercial

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

100

120

140

160

20 08 09 10 11 12 13 14 15 1

Europe excluding United Kingdom

United States

China Japan Total United Kingdom

Europe excluding United Kingdom

United States

China Japan Total United Kingdom

2005 06 07 08 09 10 11 12 13 14 15 16 Adjusted capital (trillions of US dollars, right scale) Total assets (trillions of US dollars, left scale) Adjusted capital to total assets (percent, right scale)

100

2005 06 07 08 09 10 11 12 13 14 15 16

Long-term securities and others (left scale) Short-term borrowing and repos (left scale) Deposits (left scale) Loans to deposits (right scale)

100

150

200

300

100

150

200

250

20 08 09 10 11 12 13 14 15 16

500

1,000

1,500

2,000

20 09 10 11 12 13 14 15 16

Percent of assets (total) Percent of assets (Europe) Percent of assets (United States)

Ba si

s po

in ts

(o f e

qu ity

)

Ba si

s po

in ts

(o f e

qu ity

)

Cumulative conduct charges (right scale) Cumulative restructuring charges (right scale)

Bi lli

on s

of U

S do

lla rs

Bi lli

on s

of U

S do

lla rs

Bi lli

on s

of U

S do

lla rs

Noncore assets (left scale) Noncore Assets Litigation Expenses Restructuring Costs

Sources: Bank financial statements; Bloomberg Finance L.P.; Dealogic; S&P Capital IQ; SNL Financial; and IMF staff estimates and analysis. Note: Adjusted Tier 1 capital equals shareholders’ equity, minus 45 percent (an estimate of average gross loss given default) of reported nonperforming loans, plus loan-loss reserves. In panel 1, total assets are adjusted for the netted derivatives. In panel 3, conduct and restructuring charges (in basis points of equity) are on an estimated posttax basis, assuming charges adjusted by effective tax rates.

Figure 1.5. Global Systemically Important Banks: Capital, Liquidity, and Legacy Challenges

1. Capitalization

Global banks are better capitalized ...

2. Liquidity (Percent)

... and hold higher liquidity ...

3. Legacy Challenges: Noncore Assets, Litigation Expenses, and Restructuring Costs

... and have made good progress in addressing legacy challenges.

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

with restructuring charges. Most of these are severance and other charges stemming from branch and staff reductions motivated by banks’ efforts to reduce their operating cost structures. Continental European and UK banks are most affected; their restructuring charges in 2016 amounted to $13 billion, equivalent to 25 percent of their underlying net income. Although some GSIBs have made substantial progress in reduc- ing staff, others (particularly some European GSIBs) still report large restructuring charges.

Global Banks Have Reduced Market-Related Business

Strategically, GSIBs have reduced their market-related functions—investment banks have made some of the biggest cutbacks (Figure 1.6, panel 1). This move came as earlier overexpansion and excess capacity collided with regulatory changes that increased risk-asset weight- ing and capital charges and drove a sharp decline in profitability of banks’ other lines of business (Figure 1.6, panel 2). Fixed income, currency, and commodity (FICC) businesses, in particular, have become less attrac- tive to all but a few high-volume or high-margin players, which have taken a greater share of a shrinking revenue pie (Figure 1.6, panels 2 and 3). In this environment, US banks have gained market share, and activity is now concentrated in fewer players.

While GSIBs’ declining exposure to financial mar- kets will reduce their risk, there may be associated costs to market liquidity. Evidence that this change affects market liquidity in normal times is mixed, and greater participation by nonbank market intermediaries could help address the fragmentation of market liquidity. What is less clear is whether global banks’ reduced capacity to intermediate in financial markets could affect the resilience of liquidity in periods of stress. Similarly, the supply of risk management services that require GSIB balance sheet space and capital could be reduced or provided to fewer clients. The balance between reduced GSIB riskiness and potential costs to liquidity during stress is an issue deserving of careful ongoing consideration.2

2Work is underway at the Financial Stability Board, in collabo- ration with standard-setting bodies, to evaluate the impact of the regulatory reform agenda. But it will likely take some time to realize the full impact of changes in bank business models on financial activity. Adrian and others (2017) also document the stagnation of broker-dealer balance sheets associated with deleveraging.

Global Banks Overall Continue to Operate Internationally

In contrast to declining market intensity, GSIBs overall have remained central to the provision of interna- tional credit and services (including total loans and spe- cific product markets, such as syndicated lending, trade finance, and project finance). International balance sheet commitments and revenue mix have remained quite sta- ble across almost all GSIBs (Figure 1.7, panel 1). Even as non-GSIB banks shrank international loans aggressively during 2009–13 (owing to balance sheet pressures), GSIBs as a group maintained their international lending volume (Figure 1.7, panel 2).

Those GSIBs less impacted by the financial crisis have maintained or expanded their international role. This may in part be motivated by the relative profitability of international operations. Across a sample of 724 banking subsidiaries, foreign banking operations have been more profitable than domestic business for Japanese and continental European and UK GSIBs (Figure 1.7, panel 3). Japanese banks, whose international loans have con- tributed to raising profitability, have continued to pivot aggressively toward international markets—maintaining their reliance on potentially volatile wholesale foreign currency funding—accompanied by a general expansion of corporate loans and foreign securities investments. Shifts in international exposures of continental Euro- pean and UK banks reflects three main crosscurrents. A few—mainly UK banks—have emphatically cut exposures in an international arena where they suffered large losses. Some (mainly French) banks were forced by balance sheet constraints to retrench. For many others, international lending remains an attractive business to which they have demonstrated commitment within the constraints of their balance sheet capacity and expo- sure limits.3 In contrast, US GSIBs, whose domestic operations are highly profitable, have maintained or slightly pulled back the international proportion of their loan portfolios.

Subsidiarization Presents a Structural Challenge for Some Banks

Largely in response to national regulatory pressures, several GSIBs more reliant on branching have begun gradually shifting their international lending from a direct cross-border model to one based on lending via

3This could suggest that reduced international exposure may be more a cyclical than a structural phenomenon for GSIBs, as sug- gested for the broader banking sector by McCauley and others 2017. See also Caruana 2017.

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

JPM

BOA

CITI HSBC

BARC STAN

United States (47%) Continental Europe (27%)

RBS

BNP

CS SG UBS

MUFG MFG Others

$132 bn

JPM

BOA

CITIHSBC BARC

STAN RBS

BNP CS

SG UBS

MUFG MFG Others

$87 bn

United Kingdom (23%) Others (3%)

United States (52%) Continental Europe (23%)

United Kingdom (17%) Others (8%)

3. FICC Trading Revenues, 2010 and 2016

FICC revenue pool has shrunk with a shift in market share toward US banks.

0 20 40 60 80 100

China

2010 2016

Japan Continental Europe

United Kingdom

United States

Sources: Bank financial statements; Basel Committee for Banking Supervision; Bloomberg Finance L.P.; equity research reports; European Central Bank; Federal Reserve Board; S&P Capital IQ; SNL Financial; and IMF staff estimates and analysis. Note: In panel 1, market intensity is an index scaled (1 to 100) of relative exposures across the 30 GSIBs over 2010 to 2016. Each exposure is based on an average of (1) market-risk-weighted assets divided by total risk-weighted assets; (2) Level 3 assets divided by total assets; (3) notional derivatives relative to total assets; and (4) average value at risk relative to risk-weighted assets. In panel 2, business type is identified for each subsidiary entity based on a sample of 934 foreign and domestic subsidiaries of the 30 GSIBs. Banking (724 subsidiaries) includes corporate, commercial, and consumer banking, and the advisory part of investment banking. Markets (156 subsidiaries) include underwriting, secondary market trading in securities, currencies and commodities, and dealings in derivative contracts. Wealth management (46 subsidiaries) includes asset management, private banking, and insurance. See footnote 1 in the text for an explanation of the abbreviations in panels 1 and 3. FICC = fixed income, currencies, and commodities.

Figure 1.6. Global Systemically Important Banks: Market Activity

1. Market Intensity, 2010 and 2016 (Index, maximum intensity = 100)

Market intensity has declined sharply ...

2. GSIBs by Home Region: Average Return on Assets, by Business Type, 2014–16 Average (Percent)

... as banks avoid relatively unprofitable markets businesses.

–0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8

Markets Banking Wealth management 9.5

CCB

ABC

BOC

ICBC

SMFG

MUFG

MFG

HSBC

STAN

BARC RBS

ING

UCG

BPCE

SAN

BNP

NDA

UBS

BNY

WFC

STT

BOA

CITI

JPM

2016 2010

United States

Continental Europe

United Kingdom

Japan

China

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

0.0

0.3

0.6

0.9

1.2

1.5

0 20 40 60 80 100

CCB

ABC

BOC

ICBC

SMFG

MUFG

MFG

HSBC

STAN

BARC

RBS

ING

UCG

BPCE

SAN

BNP

NDA

UBS

BNY

WFC

STT

BOA

CITI

JPM

2010 2016

2006 07 08 09 10 11 12 13 14 15 16

China Japan Continental Europe

United Kingdom

United States

China Japan Continental Europe

United Kingdom

United States

From GSIBs From non-GSIBs

2011–13 averages 2014–16 averages

Domestic Foreign

Sources: Bank financial statements; Basel Committee for Banking Supervision; Bloomberg Finance L.P.; European Central Bank; Federal Reserve Board; S&P Capital IQ; SNL Financial; and IMF staff estimates and analysis. Note: Degree of internationality is an index scaled (1 to 100) of relative exposures across the 30 GSIBs over 2010 to 2016. Each exposure is based on an average of (1) percent of revenue from nonhome regions; (2) international loans divided by total loans (or international assets divided by total assets); and (3) foreign deposits divided by total deposits. For panel 2, see notes in Figure 1.6 for sample descriptions. In panel 3, subsidiary return on assets are based on reported earnings. The reported earnings of subsidiaries in the United Kingdom and the United States may be understated due to the booking of conduct charges in those jurisdictions. See footnote 1 in the text for an explanation of the abbreviations in panel 1. GSIBs = global systemically important banks.

Figure 1.7. Global Systemically Important Banks’ International Activity

1. Degree of Internationality, 2010 and 2016 (Index, maxiumum degree = 100)

GSIBs’ international activity has remained stable overall.

2. International Loans (Trillions of US dollars)

GSIBs are increasing their share in international lending despite an overall reduction.

3. GSIBs by Home Region: Average Return on Assets, Domestic and Foreign Banking Subsidiaries, 2014–16 Average (Percent)

Foreign banking operations are more profitable than domestic entities for many banks.

4. GSIBs by Home Region: Overseas Subsidiaries’ Deposits as Percent of Total Liabilities, 2011–13 and 2014–16 Averages (Percent)

Subsidiaries of European and US GSIBs have increased their funding through local deposits.

United States

Continental Europe

United Kingdom

Japan

China

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

foreign subsidiaries (“subsidiarization”). The aggre- gate share of GSIB lending extended through foreign subsidiaries has risen from 40 percent to 60 percent of international lending since 2009 and may continue to increase gradually as banks respond to regulatory pressure to house their activities in each international jurisdiction within local legal entities with adequate local capital and liquidity. This has motivated banks to shift funding from cross-border (interbank and intragroup) funding toward local deposits (Figure 1.7, panel 4).

These structural adjustments have helped improve the resolvability and funding resilience of large, highly interconnected global banks, which strengthens financial stability. Healthy subsidiaries may also be better able to withstand pressure on their parents or other affiliates, which may have a positive effect on the stability of host countries. These considerable benefits come with some possible unintended costs. Keeping individual pools of capital in subsidiaries across a group may lower returns on equity as banks maintain higher levels of capital than before subsidiarization. Lower mobility of capital and liquidity might also compromise GSIBs’ capacity to respond to solvency or liquidity shocks.4 This may be more significant for banks that have a globally inte- grated capital and liquidity model (most investment banks) than for consumer banks. Moreover, regulatory impediments to the flow of liquidity, risk management, and funds deployment within the euro area contribute to higher costs and reduced activity, adding to business model and economic challenges. Again, officials will need to consider the balance of costs and benefits of these structural adjustments.

Progress toward Sustainable Profitability Is Uneven

Uneven progress in tackling legacy charges, business model adaptations, and group structure has led to varied profitability, as well as a mixed outlook across GSIBs (Figure 1.8, panel 1). In part, this owes to the vigor and timeliness in addressing legacy and capital chal- lenges from the global financial crisis. Responding early has paid off. US bank profitability, for example, has reached levels in line with or exceeding 8 percent cost of equity, a conservative estimate of investors’ required returns, and approach management-stated targets for their returns. European banks’ 2016 profitability, in contrast, was more mixed, with several banks generating

4Chapter 2 of the April 2015 Global Financial Stability Report (GFSR) discusses these issues further; see also Cetorelli and Goldberg 2012; Reinhardt and Riddiough 2015; and Fiechter and others 2011.

low returns, in part because of their slower progress in addressing legacy issues. Overall, about half of GSIBs by asset size remain below an 8 percent return on equity.

The outlook for sustained profitability is becom- ing more favorable as legacy issues are more fully addressed, business model improvements are imple- mented, and the global recovery strengthens.5

Following a period of strong cyclical and structural profitability headwinds over the past five years, prof- itability drivers are turning up (Figure 1.8, panel 2). After restructuring, weak and challenged banks’ assets are set to increase again. This is expected to arrest their revenue declines and to improve their reported cost-ratio dynamics. Along with an expected cyclical improvement in net interest margins, these develop- ments should help increase return on assets.

However, even with these improvements and better outlook, analysts expect one-third of the GSIB assets (about $17 trillion) to generate below-sustainable returns in 2019 (Figure 1.8, panel 3). For these banks, profitabil- ity has been restrained by structural forces such as high operating costs, low operating efficiency, and highly com- petitive home markets, exacerbated in several cases by weak information technology systems. Banks that exhibit both thin capital buffers relative to future regulatory requirements and relatively weak profitability to build those buffers over the next few years warrant heightened attention (see Figure 1.8, panel 4). Some banks continue to grapple with legacy issues, while others, particularly European investment banks, still face the fundamental problem of defining and executing profitable business models. An environment of low domestic interest rates also affects the profitability of Japanese GSIBs. These banks seek continued international expansion to offset compressed domestic profitability, and supervisors must bear in mind that such expansion increases currency and maturity mismatch risks (see IMF 2017d). Problems in even a single GSIB could generate systemic stress, so supervisory action clearly needs to remain focused on business model risks and sustainable profitability.

5This report defines banks as “weak” if they are expected to gen- erate return on equity below 8 percent in 2019, “challenged” if the expectation is between 8 and 10 percent, and “healthy” if more than 10 percent is expected. Investor surveys, cited in the October 2016 GFSR, suggest that the cost of equity is at least 8 percent. The current cost of equity—inferred from current market prices using a Gordon Growth model—is almost 11 percent for GSIBs as a whole; individual bank estimates for the cost of equity range from 8 to 15 percent. Bank management medium-term profitability targets are consistent with this view: the target for 11 out of 21 GSIBs is a return on equity above 10 percent; for the remaining 10 banks, it is between 8 and 10 percent.

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

0.0 0.2 0.4–0.3 –0.2 –0.2–0.1 0.0

ST T

W FC

JP M G S

M S

B N

B O

A C

N D

U B

IN G

B N

C S

SA N C A

B PC

U C

D B

H SB

R B

B AR

ST AN

SM FG

M U

M FG

C C

AB C

IC B

B O

–10 –5 0 –3 2 7–0.4 –0.2 0.0 0.2 0.4–7 0 7

Healthy

Challenged

Weak

100

Asset growth

Healthy (ROE ≥ 10) Challenged (8 ≤ ROE < 10) Weak (ROE < 8)

2011–16 2016–19

Revenue/assets Cost/assets Return on assets Leverage Return on equity

4 5 6 7 8 9

10 11 12

10 12 14 16 18 20

Pr ofi

ta bi

lit y

(R O

CET 1 ratio

United States Continental Europe United Kingdom Japan China

Sources: Bank financial filings; Bloomberg Finance L.P.; SNL Financial; and IMF staff analysis. Note: Underlying profit is reported net income excluding conduct and litigation charges, restructuring costs, and noncash valuation adjustments. In panel 1, CS has an ROE of –0.3 percent in 2016. Management’s ROE targets, where not available directly, are estimated from their stated return on tangible equity targets, assuming a constant ratio of current tangible equity to total equity. In panel 2, asset growth is on an annualized basis. In panels 2 and 3, future asset forecasts are estimated using consensus RWA forecasts and assuming constant RWA density. In panel 3, a balanced sample of the current 30 GSIBs are considered for the entire duration. In all panels, 2016 numbers are used for BPCE due to lack of analyst forecasts. Forward-looking analyst forecasts consensus is gathered from Bloomberg. In panel 4, the colors correspond to those in panel 1. See footnote 1 in the text for an explanation of the abbreviations in panels 1 and 4. CET 1 = common equity Tier 1 capital; GSIB = global systemically important bank; ROE = return on equity; RWA = risk-weighted asset.

Figure 1.8. Global Systemically Important Banks: Financial Performance Gaps

1. GSIB Return on Equity: 2016 Underlying, 2019 Consensus Forecasts, and Management Medium-Term Target (Percent)

Most US GSIBs should reach profitability targets, but European and Japanese GSIBs face significant gaps.

2. GSIBs: Annualized Asset Growth in Percent and Changes in Profitability Drivers and Metrics (Percentage points)

Balance sheet reflation and cost improvement are expected to help profitability ...

3. Percent of GSIB Assets by Return-on-Equity Thresholds, 2019 Consensus Forecasts

... whereas global banks, representing about one-third of GSIB assets, are still expected to have weak profits.

4. GSIBs: Profitability and Capital Position, 2019 Consensus Forecasts (Percent)

Some banks have thin capital buffers and weaker profitability prospects.

2000–03 2004–07 2008–11 2012–16 2019E Target

494844 39

90 72

37 23

16 10

15 29

39 51

7 19

16 20162016201620162016 Medium-term target 2019 forecast 8 percent cost of equity

WFC

STT

BNYMSGS

BOA

JPM

STAN

RBSBARC

HSBC CS

UBS NDA

SAN UCG

BPCESG

BNP ING

SMFGMFG MUFG

ABC BOC

CCB

ICBC

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

Further Policies Are Needed

Regulation and supervision of global systemically important banks have been considerably tightened in recent years, with detailed frameworks governing capital and liquidity and much more vigorous and regular monitoring. There has been less progress in making a resolution framework for international banks operational. Challenges include the need for further strengthening national resolution regimes, the development of cross- border resolution plans with adequate loss-absorbing capacity to make them effective, and close coordination between home and host-country regulators and resolution authorities, providing sufficient comfort for host coun- tries that a centralized resolution strategy would protect their interests. Only with such a framework in place will it be possible to avoid the potential negative consequences that can flow from the imposition of capital and liquidity requirements for GSIBs on a market-by-market basis.

In addition, regulators should have a strong focus on risks from weak business models to ensure that weaker banks are able to achieve sustainable profitability. As dis- cussed in previous GFSR reports, this applies beyond the global banks that are the focus here. In particular, although euro area banks have made further progress in cleaning up their balance sheets, nonperforming loan ratios remain high in some countries, and profitability is still a challenge. Without a more concerted effort to reduce nonperforming assets and improve business models, financial stability con- cerns could be reignited in the euro area. More generally, continued progress toward completing banking union remains essential to strengthening the financial stability foundations of the euro area banking sector.

Finally, it will be important to finalize Basel III to further strengthen the financial sector and create a more level international playing field. At a minimum, any proposals by national regulators to substantially ease capital, liquidity, or prudential standards should be considered carefully in light of their potential to damage the agenda of global regulatory harmonization.

Insurers Life Insurers Have Rebuilt Capital Buffers since the Crisis

Life insurers were hit hard by the global financial crisis. Profits tumbled, particularly in the United States (Figure 1.9, panel 1), and capital buffers fell.6

6This analysis is based on a sample of more than 80 life insurers from Belgium, France, Germany, Italy, Japan, the Netherlands, Nor- way, Spain, Sweden, the United Kingdom, and the United States. The sample covers almost two-thirds of total assets of life insurers in Europe, Japan, and the United States.

2005 06 07 08 09 10 11 12 13 14 15 16

–4

–3

–2

–1

Continental Europe Japan United States

2005–07 2008 2009–16

Global financial crisis

Sources: Bloomberg Finance L.P.; and IMF staff estimates. Note: In panel 1, return on assets is calculated by dividing net income by total tangible assets minus separate accounts. In panel 2, the shareholder equity ratio is calculated by dividing the sum of common equity plus retained earnings by tangible assets minus separate accounts. In both panels, for Japan, separate accounts were not deducted in the denominator due to lack of data.

Figure 1.9. Life Insurance Companies’ Profitability and Capital

Amid falling yields and bullish asset markets, life insurers have managed to restore profits ...

1. Life Insurers: Return on Assets (Asset-weighted indices, period averages)

... allowing them to retain earnings and lift capital buffers.

2. Life Insurers: Shareholder Equity Ratio (Percent)

Continental Europe

JapanUnited States

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

But insurers have been able to build capital since then (Figure 1.9, panel 2). Bullish equity and bond markets have raised the value of the portion of insurers’ assets that are marked to market, helping boost earnings, dividend payouts, and capital.

Life Insurers Have Been Adapting Their Business Models to Cope with Historically Low Returns

While building capital levels, life insurance compa- nies have also been adapting their business models in

response to the low-yield environment. Several changes have been made in the face of lower investment spreads. First, insurers have reduced the guaranteed returns on new policies (Figure 1.10, panel 1). Second, they have adjusted their product mix (Figure 1.10, panel 2). Euro- pean insurers have gradually sold more unit-linked poli- cies. These policies sell units similar to those in a mutual fund and shift market risk to policyholders. US insurers have moved from variable to fixed annuities, which are easier to hedge. Japanese insurers have favored the sale of

100

2008 16

54 33

24 18

67 75

100

2008 16

811 42

23 4

65 83

AAA/AA/A BBB Not IG NR and other

2006 07 08 09 10 11 12 13 14 15 16

2009 15

100 bps

50 bps

25 bps 2009 16 0

100

2007 15

Europe: gross written premiums

United States: sales

Japan: gross written premiums

Unit linked Nonlinked

Fixed annuities Variable annuities

Life insurance AnnuitiesGuaranteed returns

Japan

Investment returns Japan

Guaranteed returns United States

Investment returns United States

Guaranteed returns Germany

Investment returns Germany

100

2004 16

Cash, loans, real estate, and other

Equities Foreign bonds Domestic bonds, maturity > 10 years

Domestic bonds, maturity < 10 years

United States Europe

Figure 1.10. Changes in Life Insurance Companies’ Business Models

1. Average Investment Returns and Guaranteed Returns (Percent, on existing portfolios)

Facing investment spread compression, life insurers in Germany, Japan, and the United States have reduced guaranteed returns ...

2. Changes in Insurance Product Mix (Percent)

... and have been gradually changing their product mix.

3. European and US Life Insurers: Bond Asset Allocation (Percent)

Searching for yield, US and European life insurers have invested more in lower-rated bonds ...

4. Japanese Life Insurers’ Investment Portfolio (Percent)

... and Japanese life insurers have increased duration and holdings of foreign bonds.

Sources: Bundesbank; NLI Research Institute; and Office of Financial Research. Note: bps = basis points.

Sources: European Insurance and Occupational Pension Authority; Life Insurance Association of Japan; and Life Insurance and Market Research Association.

Sources: SNL Financial; and IMF staff estimates. Note: Not IG = noninvestment grade: bonds with ratings lower than BBB–; NR = not rated. NR and other may include some loans.

Source: Bank of Japan.

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

insurance products over saving products. However, these changes have been slow to affect balance sheets given the large amount of legacy policies that remain.

In addition, insurers have been adjusting their asset mix to higher-yielding and less liquid assets, moving out of their natural investment habitat in search of yield. • Insurers have taken on more credit risk. Despite

risk-sensitive capital requirements, at least one-third of US and European insurers’ bond portfolios now have a BBB rating or lower (Figure 1.10, panel 3).7 Additional risk taking has also been taking place in the United States—for example, using unregulated subsidiaries, which do not face the same capital requirements as insurers.

• Insurers have taken on more market risk. Japanese and US insurers have extended the maturity of domestic bond holdings to better match the dura- tion of their liabilities and enhance yields. Over the past five years, portfolio durations in the United States have increased from about five to eight years overall. Japanese life insurers have also invested in higher-yielding foreign bonds, partly exposing them to currency risk (Figure 1.10, panel 4).

• Insurers have taken on more liquidity risk. Exam- ples include commercial property, infrastructure financing, private placements, structured securities, and mortgage loans. In the United Kingdom, about 25 percent of annuities are currently backed by illiq- uid investments, and insurers have plans to increase that proportion to 40 percent by 2020.8

Market Concerns about Insurers Persist

Despite these changes, insurers continue to face profitability pressure (Figure 1.11, panel 1), and investors remain concerned about life insurers’ business models, as reflected in market valuations. Half of the US and European insurers in the sample, by assets, now have a price-to-book ratio both below precrisis levels and below one (Figure 1.11, panel 2), reflecting concerns over future profitability in a low-rate environ- ment, as well as difficulties in assessing risks. • Profitability: Despite efforts to change business

models, insurers in a significant group of countries continue to face both high guaranteed returns and

7Part of this change can be attributed to downgrades of bonds that were already in the bond portfolios of insurers.

8See Bank of England 2017.

high duration mismatches (Figure 1.11, panel 3).9 If low interest rates persist, investment returns could continue to decrease for the next decade, a situation that would leave life insurers in the Netherlands, Germany, Sweden, and Norway facing negative spreads within a few years. Even if interest rates were to increase by 100 basis points, many insurers would still face this risk (Figure 1.11, panel 4).

• Risk assessment: Investors continue to have difficul- ties adequately assessing risk in the sector because regulatory regimes are evolving and disclosure is inadequate. For example, discount rates used to value future liabilities differ between insurers and are often higher than market risk-free rates, result- ing in an underestimation of liabilities. Regulatory gaps (discussed later in this chapter) make it hard to compare risks in insurers across countries. Options embedded in some insurance contracts are also hard to value, making it difficult to assess balance sheet risks.

Life Insurers Are More Vulnerable to Market and Credit Risks

Business model adjustments on the asset side have made insurers more vulnerable to a decompression of risk premiums and falls in asset prices. A sharp decline in equity and real estate markets, combined with an increase in credit spreads and a flight to high-quality sovereign bonds, would amount to a double hit on insurers’ balance sheets in this scenario. Asset values would fall, while liabilities would increase as risk-free rates used to discount future liabilities decline. Fig- ure 1.12 shows a simulation of such a scenario, in which assets and liabilities are fully marked to market. However, current accounting and regulatory rules exempt insurers from marking all their liabilities to market and allow them to dampen market shocks through adjustments to liabilities. In the simulation, life insurers in Italy, Spain, and the United States would be affected by their lower-rated sovereign and corporate bond holdings. Insurers in Germany, the Netherlands, Norway, and Sweden would be affected by the relatively long duration of their liabilities.

If such a shock were to occur, it could mean that life insurers would be unable to fulfill their role as financial intermediaries, precisely when other parts of the finan-

9See Chapter 2 of the April 2017 GFSR.

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

100

2005–07 2014–16

ROE < 8% 8% ≤ ROE ≤ 10% ROE > 10%

0 1 2 3 4 5

Europe United States

Half of European and US sample

–6 –4 –2

0 2 4 6 8

10 12 14 16

–3.5 –3.0 –2.5 –2.0 –1.5 –1.0 –0.5 0.0 0.5 1.0 1.5

IRL ITA

GBR

JPN

NOR SWEDEU

NLD

KOR CHE

CAN USA

FRA BEL

Long-term yield on sovereign bonds minus average guaranteed returns (percent)

D ur

at io

n of

li ab

ili tie

s m

in us

du ra

tio n

of a

ss et

s (y

ea rs

)

20 17

(y ea

r- to

-d at

e av

er ag

Precrisis (2005–06 average)

Bank averages

2005–07 2014–16 United States Europe Japan

2005–07 2014–16

1. Life Insurers: Return on Equity (Period average, percent of sector assets per category)

Legacy liabilities are a drag on their profitability ...

2. Life Insurers’ Price-to-Book Ratios

... such that half of European and US insurers are valued below their book values and below precrisis levels.

3. Duration Mismatch and Guaranteed Return Spreads

Guarantees and duration mismatches remain high for a large part of the sector.

4. Projected Number of Years until Bond Yields Fall below Guaranteed Returns

Some insurers may soon face negative investment spreads.

Sources: Annual reports; Autorité de Contrôle Prudentiel et de Résolution; Bloomberg Finance L.P.; Bundesbank; De Nederlandsche Bank; European Insurance and Occupational Pensions Authority; Moody’s Investors Service; National Association of Insurance Commissioners; Nationale Bank van België; NLI Research Institute; Office of Financial Research; Organisation for Economic Co-operation and Development; SNL Financial; and IMF staff estimates. Note: In panel 1, the implied cost of capital was about 10 percent before and after the global financial crisis. In panel 3, the size of the bubble relates to the share of liabilities with guaranteed returns to total life insurance liabilities. Green = countries with insurance sectors that have low guaranteed returns and low or negative duration mismatch. Yellow = countries with insurance sectors that have either high guaranteed returns or a high duration mismatch. Red = countries with insurance sectors that have both high guaranteed returns and high duration mismatch. In both cases in panel 4, guaranteed returns continue to decline. In the case of a 100 basis point increase in bond yields, Belgian, Japanese, and US investment yields are not expected to fall below guaranteed returns. Data labels in the figure use International Organization for Standardization (ISO) country codes. ROE = return on equity.

Figure 1.11. Life Insurers’ Market Valuations and Risk Outlook

USA BEL JPN NLD DEU SWE NOR DEU NLD SWE NOR

Current interest rate environment

100 basis point increase in sovereign and corporate bond yields

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

cial system are also failing to do so.10 This highlights the importance of guarding against complacency and the need for additional policy focus on nonbank financial institutions and financing markets and the extension of macroprudential tools.

Policies Are Needed to Ensure Greater Insurer Resilience

Life insurers face growing vulnerabilities in the continued low-interest-rate environment. Policymakers should ensure that as insurers adapt to this environ- ment they do not take excessive risks. Risk assessment in the insurance sector suffers from opaque and het- erogeneous financial disclosure and deficiencies in the accounting and regulatory regimes. Policymakers must continue to strengthen regulatory frameworks and increase reporting transparency.

Greater public disclosure of timely information on key metrics to assess interest rate risk (namely, guaran- teed returns and duration mismatches) would motivate insurers to further adapt their business models and build additional capital buffers. Liabilities are often not valued using current market prices (Japan, United States) or are understated by country- and firm-specific adjustments (Europe), hampering comparability. In the United States, there is no consolidated capital require- ment, and sector-wide stress tests are not regularly undertaken, which leaves the potential for firms to mask risks. In Europe, the lack of loss-absorbing capac- ity in some instruments eligible as regulatory capital harms the credibility of reported solvency positions. Regulators are encouraged to close these regulatory gaps. In particular, the International Association of Insurance Supervisors should accelerate its efforts to establish a global insurance capital standard that ade- quately addresses these underlying vulnerabilities.

Monetary Policy Normalization: A Two-Sided Risk Central bank balance sheets have grown considerably due to large-scale asset purchase programs. This has forced substantial portfolio adjustments in the private sector and across borders, reducing government bond yields, term premiums, and credit spreads while boosting equity valuations. As the global recovery progresses, a key stability challenge is to gradually rebalance central bank and private sector portfolios against the backdrop of monetary policy cycles that are not synchronized across countries.

10See also Chapter 3 of the April 2016 GFSR.

Too quick an adjustment could cause unwanted turbu- lence in financial markets and international spillovers. However, the expected process of normalization is likely to be gradual, with continued easy monetary conditions and low volatility that could foster a further buildup of financial excesses and medium-term vulnerabilities.

Managing the gradual normalization of monetary pol- icies presents a delicate balancing act. The pace of nor- malization cannot be too fast or it will remove needed support for sustained recovery and desired increases in core inflation across major economies. The substantial rebalancing of private portfolios that has occurred also makes the adjustment of financial market prices much less predictable than in previous cycles. On the other hand, the likely prolonged period of low interest rates could further deepen financial stability risks as investors take on more risk in their search for yield.

–10

–8

–6

–4

–2

AUT BEL DEU ESP FRA GBR ITA JPN NLD NOR POR SWE USA

Changes in liabilities Changes in assets

Sources: Bank of Japan; European Insurance and Occupational Pensions Authority; Life Insurance Association of Japan; Moody’s Investors Service; National Association of Insurance Commissioners; and IMF staff estimates. Note: Cash flows are fixed. Derivative positions and loss absorption by policyhold- ers and by taxes and regulatory adjustments are not taken into account. This implies that results should be considered an upper-bound impact. Shocks are applied to aggregate sector balance sheets of solo life insurers as of 2016:Q3 (Europe), 2016:Q1 (Japan), and 2015:Q4 (United States). The following shocks are applied: equity (–10 percent); real estate (–6 percent); sovereign debt yield AAA–A (–50 bps), BBB (+100 bps), < BBB (+100 bps); corporate bond yields AAA–A (+50 bps), BBB (+150 bps), < BBB (+200 bps); risk-free rates (–50 bps). Data labels in the figure use International Organization for Standardization (ISO) country codes. bps = basis points.

Figure 1.12. Simulated Mark-to-Market Shocks to Assets and Liabilities (Percent)

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

Uncertainty around Central Bank Balance Sheet Adjustments

Large-scale asset purchase programs by the major central banks have led to a considerable shift in port- folios by domestic and foreign investors (Figure 1.13, panels 1 and 2). Central banks in Japan, the United Kingdom, the United States, and the euro area have increased their holdings of outstanding government securities to 37 percent of GDP, up from 10 percent before the global financial crisis. These purchases have produced marked shifts in asset allocations across major advanced economies during their respective periods of quantitative easing (QE). • The Bank of Japan’s QE program, the most aggres-

sive of those of major advanced economy central banks, led domestic banks and pension funds to reduce their Japanese government bond holdings. The European Central Bank’s QE program also had a large impact in altering the composition of port- folios: foreigners significantly reduced their holdings of government debt, followed by domestic banks and pension funds. In the United States, the Federal Reserve’s QE programs led to a more muted shift: foreigners reduced their holdings of Treasuries as the accumulation of foreign exchange reserves slowed, as did insurance companies and pension funds, but other investors increased their holdings, including banks (to satisfy liquidity requirements), households, and mutual funds. The extent of the QE programs across central banks largely reflected the severity of the deflationary pressures experienced since the crisis began.

• Some 100 percent or more of the supply of gov- ernment bonds has been absorbed by central bank purchases in the euro area and Japan. Official demand for Japanese government bonds exceeded net issuance in early 2013, while official purchases of euro area government debt eclipsed net issuance in 2016 as the growth in government deficits slowed (Figure 1.13, panel 3). But even though the Federal Reserve’s QE programs were large in absolute terms, they were more modest relative to net issuance, which explains their more muted impact on investor portfolio rebalancing.11

11Federal Reserve asset purchases accounted for a lower share of net issuance of US Treasuries, but a much greater share of quasi-agency mortgage-backed securities (net issuance in excess of 100 percent).

• By reducing the stock of fixed income instruments available to the private sector, central banks crowded out traditional investors, such as banks, insurance companies, and asset managers, to differing degrees (Figure 1.13, panel 4). This prompted some private investors to reach for duration, credit, and liquidity risk to increase returns—an intended and beneficial consequence of asset purchase programs.

Going forward, portfolio rebalancing will have an impact on term premiums and broader risk premiums through two main channels. First, by releasing partic- ular assets, central bank balance sheet normalization will increase their net supply to the public and may increase their term and risk premiums (the portfo- lio balance channel) (Figure 1.13, panel 4). Second, normalization will be associated and consistent with higher future short rates (the signaling channel).

There is significant uncertainty as to the magni- tude of the adjustment in term premiums, given the unique set of conditions—large central bank bal- ance sheets, a prolonged period of accommodation, diverging monetary policy cycles, and uncertain effects of postcrisis reforms and portfolio substitution. The magnitude holds great import: sovereign bond yields are the benchmark rate for a wide range of other assets, and term premiums are an input for broader risk premiums.

Historically, policy rates and term premiums have not always moved in unison; indeed, they diverge quite often (Figure 1.14, panel 1). Once the central bank starts increasing policy rates, it also provides forward guidance, reducing uncertainty (over interest rates and inflation). Consequently, bond risk and term premi- ums decline. Indeed, term premiums actually declined during the two most recent US tightening cycles; even previous monetary tightening cycles draw at best a mixed picture.12

But historical precedent may not be a helpful guide, given the large size of central bank balance sheets and compressed term premiums (Figure 1.14, panel 2). In the case of the United States, the Federal Reserve esti- mates that market expectations of a gradual unwinding and fall in the maturity of its securities holdings would increase the term premium by about 15 basis points by the end of 2017, at which point QE would still be holding down term premiums by a total of about

12Adrian, Crump, and Moench 2013.

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

–2

2010 11 12 13 14 15 16 17 18 19 20 21 22

100

2004 05 06 07 08 09 10 11 12 13 14 15 16

–500 0

500 1,000 1,500 2,000

2000 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17

Euro area—private2

Japan—private United Kingdom—private United States—private G4—official3

–1,000

–500

500

1,000

1,500

2,000

2,500

3,000

Jan. 17Jan. 15Jan. 13Jan. 11Jan. 09Jan. 07Jan. 2005

Fed ECB BOJ BOE Total Stock

Aggregate stock as a share of GDP

Domestic central bank Foreign official sector Domestic bank Domestic nonbank Foreign bank Foreign nonbank

Pr iv

at e

in ve

st or

–500 0

500 1,000 1,500

2000 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17

–200 0

200 400 600 800

2000 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17

United States

Euro Area

Japan

Net issuance Sovereign bond purchases

Forecast

1. Change in Central Bank Balance Sheet Assets (Billions of US dollars, 12-month rolling sum, left scale; percent of GDP, right scale)

Central bank balance sheets have expanded because of large-scale asset purchases ...

2. Advanced Economy Sovereign Bond Holdings by Investor Type (Percent)

... leading domestic and foreign central banks to capture a sizable share of sovereign debt.

3. Government Bond Issuance and Official Demand (Billions of US dollars, 12-month moving sum)

Large official purchases have outstripped net issuance in the euro area and Japan ...

4. Change in Stock of Advanced Economy Sovereign Debt, by Region of Issuance and Holder1 (Trillions of US dollars, cumulative change since beginning of 2010)

... but going forward, the private sector will need to absorb additional supply.

Sources: Bank of England; Bank of Japan; European Central Bank; Federal Reserve; government sources; Morgan Stanley; World Bank; Arslanalp and Tsuda 2012, updated; and IMF staff estimates. Note: Panels 2–4 exclude agency debt securities. In panel 4, debt stocks are converted to US dollars using end of quarter exchange rates; ECB net purchases are assumed to decline to a reduced pace and the asset purchase program extended to June 2018; Fed net purchases are assumed to follow the path outlined by the Fed starting in 2017:Q4; BOJ net purchases are assumed to equal forecast net supply; BOE net purchases are assumed to equal zero from 2017:Q1 onward. BOE = Bank of England; BOJ = Bank of Japan; ECB = European Central Bank; Fed = US Federal Reserve; G4 = euro area, Japan, United Kingdom, United States; QE = quantitative easing. 1Forecasts use forecasted central government net lending/borrowing. 2The following member countries of the euro area are included: Austria, Belgium, Finland, France, Germany, Greece, Ireland, Italy, Netherlands, Portugal, and Spain. 3Until end-2016, debt absorbed by central banks and foreign and supranational institutions; from 2017 onward, aggregated central bank purchases.

Figure 1.13. Central Bank Balance Sheets and the Sovereign Sector

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

85 basis points, with the portfolio balance channels accounting for two-thirds of the impact.13 An inflation surprise on the upside could also lead to a sharp jump in term premiums.

Potential International Spillovers Pose Additional Challenges and Risks

Because of the different starting points and time paths for both economic recovery and the state of financial repair, the international aspects of balance

13Bonis, Ihrig, and Wei 2017.

sheet normalization and spillovers are significant for two reasons: • The domestic effects of balance sheet normalization

may be transmitted to other economies because global financial markets are highly integrated. Balance sheet normalization in major advanced economies could tighten financial conditions in other countries, raising long-term rates and induc- ing capital outflows from those countries. This is because term premiums exhibit a high degree of comovement, particularly if they originate from shocks from the largest global bond markets,

–200 –100

0 100 200 300

400 500 600 700

JPN GBR CAN DEU USA

Current Maximum Mean Minimum

t – 0 End-2017 End-2018 End-2019 End-2020 –1

2007 08 09 10 11 12 13 14 15 16 17 YTD

Japan United Kingdom United States Euro area

Average Maximum Minimum

Date of minimum

Jun. 16 Sep. 16Sep. 16 Jun. 16 Jun. 16

1994 cycle 1999 cycle 2004 cycle Current cycle Quarters after first rate hike

–100 –50

0 50

100 150 200 250 300 350 400

0 1 2 3 4 5 6 7 8 9 10

0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8

Cumulative change in term premiums Cumulative change in policy rate Treasury 10-year yield (right scale)

Ba si

s po

in ts

Pe rc

en t

1. Federal Funds Rate and Term Premiums during Previous Monetary Policy Cycles

Policy rates and term premiums have diverged during recent monetary policy tightening cycles ...

2. Term Premiums in Advanced Economies (Basis points, 1990–2017)

... but term premiums are near historical lows in several major economies.

3. Market-Implied Cumulative Change in Policy Rates (Basis points)

Monetary policy cycles are diverging ...

4. Overnight Indexed Swap Forward Rate Curves for Advanced Economies (Percent)

... and markets expect a slow pace of tightening.

Sources: Bloomberg Finance L.P.; and IMF staff estimates based on Wright 2011. Notes: Panel 4 shows annual average three-month overnight indexed swap (OIS) rates on forward contracts for tenors from six months to five years. The OIS forward curves are constructed from the US dollar, euro, Japanese yen, and British pound, and the average, maximum, and minimum are computed for each tenor across the four jurisdictions. Data labels in the figure use International Organization for Standardization (ISO) country codes. YTD = year to date.

Figure 1.14. Policy Rates, 10-Year Government Bond Yields, and Term Premiums

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

such as the United States, Germany, and Japan (see the October 2016 GFSR). These heightened cross-border dynamics could potentially trigger a large simultaneous increase in global rates. This poses challenges because of diverging monetary pol- icies (Figure 1.14, panel 3) and paths for normaliza- tion (Figure 1.14, panel 4).

• Differences in balance sheet repair across countries could create additional sources of financial stress as monetary policy normalizes. For example, euro area sovereign term spreads could increase further as the prospect of reduced monetary accommoda- tion moves closer. Although this could partly reflect rising inflation expectations, it could also signal increased credit risks in countries with high debt burdens given the prospect of further reductions in European Central Bank (ECB) net asset purchases.

How Will Emerging Market Economies Fare amid Reduced Central Bank Support?

Large-scale monetary accommodation has under- pinned a significant portion of portfolio flows to emerging market economies. Model estimates indicate that about $260 billion in portfolio inflows since 2010 can be attributed to the push of unconventional poli- cies by the Federal Reserve (Figure 1.15, panel 1).14

These estimates suggest that the expected steady pace of Federal Reserve policy normalization over the next two years (as described in the baseline of the October 2017 WEO) could reduce portfolio flows by about $35 billion a year (Figure 1.15, panel 2). Countries that benefited the most during the boom period could see the largest moderation in inflows. If so, Chile, Mexico, and South Africa would be expected to experience the greatest decline in inflows

14Estimates for portfolio flows are obtained using a model adapted from Koepke 2014. The model estimates the impact of external “push” and domestic “pull” variables on portfolio flows to emerging markets, consistent with the capital flows literature. The dependent variable is monthly data from the Institute of International Finance on nonresident portfolio flows to emerging market economies (that is, foreign purchases of emerging market stocks and bonds). Inde- pendent variables include push factors, pull factors, and a constant term. Push variables include a proxy for global risk aversion (the US corporate BBB spread over Treasuries), three-year-ahead expectations for the federal funds effective rate, and the change in assets on the Federal Reserve’s balance sheet. Pull variables include an emerging market economic surprise index compiled by Citigroup and the Morgan Stanley Capital International Emerging Markets Index. The (positive) constant term captures the sizable passive component of portfolio flows, which is due to portfolio growth and passive reallo- cation (and thus unrelated to push or pull factors).

relative to the size of their economies, estimated at a cumulative 1.0 to 1.5 percent of annual GDP over the next two years (Figure 1.15, panel 3). It is worth noting, however, that emerging market economies with previously large inflows are generally those with deeper and more liquid markets that are able to with- stand outflows better. Countries that have benefited the most from inflows owe some of this benefit to strong domestic factors, such as improving growth and external positions and declining corporate vulner- ability. To the extent that such favorable conditions are maintained, the impact of a less favorable external environment would be mitigated, including via other types of foreign capital inflows, such as foreign domestic investment.

Emerging market economies should be able to handle this reduction in inflows in a relatively smooth manner, given their enhanced resilience and stronger growth outlook. However, a rapid increase in inves- tor risk aversion would have a more severe impact on portfolio inflows and prove more challenging, particularly for countries with greater dependence on external financing. For example, Malaysia, Poland, South Africa, and Turkey are projected to have sizable external financing needs through 2020 (Figure 1.15, panel 4). However, pressures from external shocks can be mitigated by large external asset holdings of domes- tic investors and banks.

Monetary Policy Changes Should Be Well Communicated to Prevent Excessive Market Volatility

The baseline path for the global economy foresees continued support from accommodative monetary policies, as inflation rates are expected to recover only slowly. Too quick an adjustment could cause unwanted turbulence in financial markets while removing needed support for the recovery. To ensure a smooth nor- malization of monetary policy, monetary authorities should provide and follow well-communicated plans on unwinding their holdings of securities and, if needed, provide guidance on prospective changes to the framework. At the same time, authorities need to be mindful of potential global spillovers as normaliza- tion proceeds. These efforts will help anchor market expectations and avoid undue market dislocations or excessive volatility.

Central banks with still-expanding balance sheets will need to take appropriate measures to alleviate col-

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

lateral scarcity pressures in order to support liquidity resilience and efficient market functioning. • For the European Central Bank, subdued inflation

points to the need for monetary policy to remain accommodative for an extended period.15 To this end, the ECB has committed to keeping policy rates at their current levels until well past the horizon of net asset purchases. It will be important to adhere to this commitment, thus ensuring the credibility of forward guidance and maintaining accommodation even if supply constraints neces-

15See IMF 2017c.

sitate scaling back net asset purchases next year. Moreover, reinvesting the proceeds from maturing assets would keep the central bank balance sheet from shrinking.

• For the Bank of Japan, stubbornly low inflation underscores the importance of maintaining sus- tained accommodation through its “quantitative and qualitative easing with yield curve control” framework.16 The Bank of Japan should carefully calibrate its yield curve policy in the event of downside risks, including by considering lowering

16See IMF 2017d and IMF 2017e.

–1.0

–0.5

0.0

Oct. 2017 Apr. 18 Oct. 18 Apr. 19 Oct. 19 –10

MYS TUR POL ZAF CHL MEX IND COL BRA CHN IDN RUS

–100

100

200

300

400

2010 12 14 16 –80

–60

–40

–20

–1,350

–1,200

–1,050

–900

–750

–600

–450

–300

–150

Oct. 2017 Apr. 18 Oct. 18 Apr. 19 Oct. 19

Flows impact of Fed balance sheet reduction Flows impact of Fed policy expectations Fed balance sheet reduction (right scale)

–1.5

Portfolio balance (Fed QE) Fed policy expectations Global risk appetite EM domestic factors

China India Brazil Indonesia Turkey Poland Colombia Malaysia Chile Mexico South Africa

Short-term debt on remaining maturity basis (2018–20 average) Current account deficit (2018–20 average) External financing requirement (2018–20 average) 15 percent external financing requirement threshold

1. Model Estimates: Cumulative Contributions to Emerging Market Portfolio Flows (Billions of US dollars)

A large portion of portfolio flows has been driven by US monetary policy accommodation.

2. Estimated Cumulative Monthly Contributions to Emerging Market Portfolio Flows, 2017–19 (Billions of US dollars)

Estimates point to a substantial reduction in portfolio flows due to US monetary policy normalization ...

3. Estimated Cumulative Impact of External Factors on Portfolio Flows (Percent of GDP)

... with some countries likely to experience reduced inflows of 1–1.5 percent of annual GDP over the next two years.

4. External Financing Requirements (Percent of GDP)

This could prove challenging for those with large external financing needs.

Sources: Federal Reserve; and IMF staff estimates. Note: Data labels in the figure use International Organization for Standardization (ISO) country codes. EM = emerging market; Fed = Federal Reserve; QE = quantitative easing.

Figure 1.15. Emerging Market Economy Capital Flows

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

the yield curve—in coordination with appropriate fiscal support and with consideration to the profit- ability of financial institutions and the functioning of the Japanese government bond market—should deflation pressure persist. Moreover, it is important for the Bank of Japan to continue to monitor the market liquidity and functioning of the Japanese government bond market and to consider appro- priate measures to alleviate shortages in the event of liquidity stress.

Has the Search for Yield Gone Too Far? The low-interest-rate environment has stimulated a search for yield in markets, pushing investors beyond their traditional risk mandates. This has compressed spreads, reduced the compensation for credit and mar- ket risk in bond markets, contributed to low volatility, and facilitated the use of financial leverage. While these supportive financial conditions have helped boost growth, as intended, they have also raised the sensitiv- ity of the financial system to market risks. Prolonged normalization of monetary policy could extend these trends. Unless well managed, these rising medium-term vulnerabilities could lead to significant market disrup- tions if risk premiums and volatility decompress rapidly.

Too Much Money Chasing Too Few Yielding Assets Has Created a Search for Yield

After nearly 10 years of extraordinary monetary accommodation, as well as changing structural factors such as demographics and slower growth, the universe of global fixed income looks very different than before the global financial crisis. While the size of the fixed income market has exploded—one of the major investment- grade benchmark indices has increased from about $19.5 trillion in 2007 to $45.7 trillion in 2017—the portion of bonds with yields that meet investor targets has shrunk dramatically. In 2007, about 80 percent of the fixed income index ($15.8 trillion) yielded over 4 percent—the approximate required return for many absolute return investors such as pension funds and insurance companies (Figure 1.16, panel 1).17 But

17For example, the required return on investment for insurance companies = the guaranteed returns promised to policyholders + the cost of their equity * leverage. These numbers differ between markets. For the United States, this is 3.6 percent + 10 percent * 0.10 = 4.6 percent. For Europe, this is 2.3 percent + 10 percent * 0.07 = 3.0 percent. This assumes no additional sources of profit, such as underwriting margins, so the required return should be seen

this proportion has now shrunk to less than 5 percent ($1.8 trillion) (Figure 1.16, panel 2).18

In the United States, this dearth of higher-yielding securities combined with the portfolio rebalancing effects of QE has resulted in a search for yield. There has been a marked shift of foreign investors out of their traditional positions in US Treasury bonds and agency securities and into higher-yielding US corporate bonds (Figure 1.16, panels 3 and 4). Non-US investors now rank among the largest holders of US corporate bonds, at nearly 30 percent of outstanding debt, up from 12 percent in 1990 and one quarter before the start of quantitative easing policies. Marginal demand has been especially pronounced among Asian investors, with flows from insurance and pension funds from Japan and Taiwan Province of China accounting for almost two-thirds of all foreign institutional flows into US investment-grade credit over the past three years.

The Search for Yield Has Also Led to Greater Capital Flows and More Borrowing by Low-Income Countries

In emerging market economies, the search for yield—combined with stronger growth and lower corporate vulnerabilities—has supported a notable rebound in portfolio inflows. Nonresident inflows of portfolio capital reached an estimated $205 billion in the year through August and are on track to reach $300 billion for 2017, more than twice the total observed during 2015–16 and on par with the strong pace of inflows from 2010–14 (Figure 1.17, panel 1). The primary beneficiaries of portfolio inflows have been large emerging market economies, including Colombia, Mexico, South Africa, and Turkey. Some have used this period to enhance policy buffers in the form of higher international reserves (Figure 1.17, panel 2). This has helped compress yields and spreads for sovereigns and firms, lifting asset valuations and external bond issuance (Figure 1.17, panels 3 and 4).

Low-income countries have also benefited from the search for yield by expanding their access to interna- tional bond markets. Bond issuance has risen sharply since the start of 2017, with the total volume $7.4 bil- lion close to the record level in 2014 (Figure 1.18, panel 1). Despite strong global demand for yield,

as an upper bound. Nevertheless, absolute return investors require historically high real rates. For pension funds, the required return is the discount rate applied to liabilities.

18Bank of America Global Broad Market Index.

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

low-income countries face less favorable borrowing conditions, reflecting less liquid markets, weaker credit profiles, and the lack of an issuance track record (Fig- ure 1.18, panel 2). Borrowing has generally been used to fund infrastructure projects, refinance debt, repay arrears, and increase budgetary flexibility.19 However, this borrowing has been accompanied by an underlying deterioration in debt burdens (Figure 1.18, panel 3).

19See IMF 2017a.

In low-income countries, greater reliance on foreign borrowing leaves them vulnerable to a decompression of global risk premiums. This vulnerability reflects several factors, including higher total debt stocks and greater debt servicing needs and high exposure to flight-prone foreign asset managers and hedge funds. Low-income countries would be most at risk if adverse external conditions coincided with spikes in their external refinancing needs. Although near-term debt rollover needs are small, many low-income-country

–1–0 0–1 1–2 2–3 3–4 4–5 5–6 6–7

$15.8 trillion

7–8 >8

Corporate Quasi and foreign government Securitized/collateralized Sovereign

Asia 4.2%

USA yield 4.6%

Other Europe 4.3%

CAN 5.3%

GBR 4.7%

Latin America 7.3% FRA 3.6%

DEU 2.1%

JPN 3.3% EMEA 5.6%

ESP 2.2%

Global total outstanding $1.5 trillion

Yield (percent) –1–0 0–1 1–2 2–3 3–4 4–5 5–6 6–7 7–8 >8

Yield (percent)

1. Global Investment-Grade Fixed Income Instruments, 2007 (Trillions of US dollars)

In 2007, a variety of asset classes generated returns in excess of 4 percent.

2. Global Investment-Grade Fixed Income Instruments, 2017 (Trillions of US dollars)

In 2017, corporate debt is the only significant asset class that provides a comparable return.

3. Yields of US Dollar Corporate Bonds Outstanding

US corporate bonds make up the majority of the US dollar corporate bond universe ...

4. Holdings of US Corporate Bonds and Loans, by Investor Type (Percent)

... drawing foreign investors beyond their traditional risk habitats.

Sources: Bank of America Merrill Lynch; Bloomberg Finance L.P.; Federal Reserve; Haver Analytics; and IMF staff estimates. Note: Panels 1 and 2 are based on the Bank of America Global Bond Market Index. Data labels in the figure use International Organization for Standardization (ISO) country codes. EMEA = Europe, Middle East, and Africa.

Figure 1.16. Global Fixed Income Markets and US Corporate Credit Investor Base

$1.8 trillion

2008 09 10 11 12 13 14 15 16 17

Foreign sector Insurance companies Banks and broker dealers

Mutual funds Households/hedge funds

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

issuers face a significant repayment hump after 2021 (Figure 1.18, panel 4). Indeed, annual principal and interest repayments (as a percent of GDP or inter- national reserves) have risen above levels observed in regular emerging market economy borrowers.

Credit and Market Risks Are Increasingly Being Mispriced

Low yields, compressed spreads, abundant financ- ing, and the relatively high cost of equity capital

have encouraged a buildup of financial balance sheet leverage as corporations have bought back their equity and raised debt levels (as discussed in the April 2017 GFSR). This means that the share of lower-rated com- panies in major US, European, and global bond indi- ces has increased (Figure 1.19, panel 1). This trend of worsening credit quality also means that the estimated default risk for high-yield and emerging market bonds has remained elevated (Figure 1.19, panels 4 and 5).

Despite declining credit quality, the compensation for credit risk in key corporate bond markets has

–100 –75 –50 –25

0 25 50 75

100 125 150 175

2011 12 13 14 15 16 17F

–200

–100

100

200

300

400

2000 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 –5

0 5

10 15 20 25 30 35 40 45 50

So ut

h Af

ric a

Tu rk

Co lo

m bi

M ex

ic o

Ch ile

In di

Po la

Br az

In do

ne si

Ch in

Ru ss

Portfolio debt Portfolio equity Reserves Other investments

EM debt inflows EM equity inflows

–450 –350 –250 –150

–50 50

150 250 350 450 550

2011 12 13 14 15 16 17F

Amortizations/buybacks Gross issuance Coupons Net financing

1. Nonresident Portfolio Flows to Emerging Markets (Billions of US dollars, four-quarter rolling sum)

Portfolio flows to emerging markets have rebounded in recent quarters.

2. Cumulative Nonresident Capital Inflows and Change in Gross Reserves, 2010:Q1–17:Q1 (Percent of GDP)

Some emerging markets have used foreign inflows to build reserve buffers.

3. Hard Currency Sovereign Issuance (Billions of US dollars)

Emerging market sovereign gross and net issuance is at record levels.

4. Hard Currency Corporate Issuance (Billions of US dollars)

Corporate gross issuance is back to 2013–14 levels, but net issuance remains subdued.

Sources: Haver Analytics; Institute of International Finance; JPMorgan Chase & Co.; and IMF staff estimates. Note: Panel 2 uses four-quarter sum of GDP to 2017:Q1. Panels 3 and 4 are JP Morgan estimates. Panel 4 omits direct investment and financial derivative liabilities. EM = emerging market; F = forecast.

Figure 1.17. Emerging Market Economies: Debt Issuance, Portfolio Flows, and Asset Prices

205 225 186

12 63 110

3 20

0 13

57 58

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

actually fallen. One way to gauge this is to measure the amount of spread per unit of corporate leverage paid to investors. For every increase in the lever- age multiple (measured by debt to earnings before interest, taxes, depreciation, and amortization), the spread received has declined sharply for both US dollar–denominated and emerging market bonds (Figure 1.19, panel 2). A decomposition of bond yields suggests that the amount of spread left for mar- ket risk has fallen, particularly for high-yield bonds

(Figure 1.19, panels 3–5). Similarly, other estimates of market risk premiums in bond markets suggest that compensation has declined steadily over time (Figure 1.19, panel 6). To reach the average levels from 2000 to 2004, market risk and term premi- ums would need to rise about 200 basis points for investment-grade bonds and about 450 basis points for high-yield bonds. Market risk and term premiums would need to rise about 375 basis points for emerg- ing market bonds.

0 1 2 3 4 5 6 7 8 9 10 11 12

2009 10 11 12 13 14 15 16 17YTD

Africa Asia Latin America Number of issuers (right scale)

2018 19 20 21 22 23 24 0

100

0 5 10 15 20 25 30 35

Pu bl

ic g

ro ss

d eb

t ( pe

rc en

t o f G

D P)

Public interest expenses to revenues (percent)

Cameroon Rwanda Honduras

Vietnam Nigeria Mozambique

Côte d’Ivoire Ethiopia Ghana

Zambia Kenya Tanzania

Sources: Bloomberg Finance L.P.; Bond Radar; and IMF staff estimates. Note: Sample includes 74 low-income countries that were both International Development Association and IMF Poverty Reduction and Growth Trust (PRGT) eligible as of end-2014. Four countries (Bolivia, Mongolia, Nigeria, Vietnam) have graduated from the list of PRGT-eligible countries. Data labels use International Organization for Standardization (ISO) country codes. EM = emerging market; YTD = year to date.

Figure 1.18. Low-Income Country External Borrowing and Vulnerabilities

1. International Sovereign Issuance of Low-Income Countries by Region (Billions of US dollars)

Low-income sovereign bond issuance has risen sharply in 2017, nearing previous peaks.

2. Low-Income Country Coupons at Issuance and Secondary Emerging Market Yields (Percent)

Market access conditions improved recently, but remain less favorable compared with other issuers.

3. Interest to Revenues and Public Debt, 2012–18

Debt burden indicators have deteriorated.

4. Sovereign International Bond Servicing Needs (Billions of US dollars)

Tighter external financial conditions would affect those with large rollover needs.

Oct. 2011 Aug. 12 Jun. 13 Apr. 14 Feb. 15 Dec. 15 Oct. 16 Aug. 17

EM B-rated sovereigns Individual country issuances

Rising issuance Stressed issuance

Recovering issuance

EM BB-rated sovereigns

Principal Interest

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

–2.0 0 2 4 6 8 10 12 14 16 18

Ja n.

2 00

Au g.

0 7

M ar

. 0 8

O ct

. 0 8

M ay

0 9

D ec

. 0 9

Ju l.

Fe b.

1 1

Se p.

1 1

Ap r.

N ov

. 1 2

Ju n.

1 3

Ja n.

1 4

Au g.

1 4

M ar

. 1 5

O ct

. 1 5

M ay

1 6

D ec

. 1 6

Ju l.

–2 0 2 4 6 8

10 12 14 16 18 20 22

Ja n.

2 00

Ja n.

0 3

Ja n.

0 4

Ja n.

0 5

Ja n.

0 6

Ja n.

0 7

Ja n.

0 8

Ja n.

0 9

Ja n.

1 0

Ja n.

1 1

Ja n.

1 2

Ja n.

1 3

Ja n.

1 4

Ja n.

1 5

Ja n.

1 6

Ja n.

1 7

1999 2001 03 05 07 09 11 13 15 17

Market risk premiums Term premium Default risk compensation Risk-neutral Treasury yield Total

–2 0 2 4 6 8 10 12 14 16 18 20 22

Ja n.

2 00

Ja n.

0 1

Ja n.

0 2

Ja n.

0 3

Ja n.

0 4

Ja n.

0 5

Ja n.

0 6

Ja n.

0 7

Ja n.

0 8

Ja n.

0 9

Ja n.

1 0

Ja n.

1 1

Ja n.

1 2

Ja n.

1 3

Ja n.

1 4

Ja n.

1 5

Ja n.

1 6

Ja n.

1 7–1

0 1 2 3 4 5 6 7 8 9

10 11 12

Ja n.

2 00

Ja n.

0 1

Ja n.

0 2

Ja n.

0 3

Ja n.

0 4

Ja n.

0 5

Ja n.

0 6

Ja n.

0 7

Ja n.

0 8

Ja n.

0 9

Ja n.

1 0

Ja n.

1 1

Ja n.

1 2

Ja n.

1 3

Ja n.

1 4

Ja n.

1 5

Ja n.

1 6

Ja n.

1 7

100

200

300

400

500

600

700

800

2008 09 10 11 12 13 14 15 16 17

Global emerging markets United States

Market risk premiums Term premium Default risk compensation

Risk-neutral Treasury yield Total

Market risk premiums Term premium Default risk compensation

Risk-neutral Treasury yield Total

Developed market US dollar high yield Global US dollar investment grade, excluding emerging markets Emerging markets

United States Euro area Global

1. Quality Breakdown of the Investment-Grade Index (Percent of sample with BBB rating)

A high proportion of ratings are clustered at the bottom end of the investment-grade rating range.

2. Emerging Market and US Dollar Bond Spreads per Turn of Leverage (Basis points per turn of leverage)

Risk-adjusted spreads have compressed to postcrisis lows.

3. US Dollar Global Investment-Grade Bond (Excluding Emerging Markets) Yield Decomposition (Percent)

Risk premiums grind tighter for investment ...

4. US Dollar Developed Market High-Yield Bond Yield Decomposition (Percent)

... and high-yield risk premiums fall to near new tights after an energy-related pop in 2016.

5. US Dollar Emerging Market Bond Yield Decomposition (Percent)

Emerging market bond risk premiums are also grinding lower ...

6. Markets Plus Term Premiums for Emerging Market and Developed Market Investment-Grade and High-Yield Bonds (Percent)

... driven by declines in term and market risk premiums.

Sources: Bank of America Merrill Lynch; JPMorgan Chase & Co; Standard & Poor’s; and IMF staff calculations. Note: Market risk premium is the difference between the observed monthly bond spread and the estimated default risk compensation. Default risk compensation is estimated monthly by breaking down each index’s holdings into Standard & Poor’s (S&P) ratings buckets. Then, based on each bucket’s rating and average duration, an average cumulative default probability is derived by referencing S&P’s ratings transition tables. These results are weighted by the duration and ratings distribution of the corresponding index. Investment-grade spread, duration, and weightings are derived from the JPMorgan JULI ALL ex-EM index. High-yield data are derived from the JPMorgan Developed Market High Yield index. Emerging market data are derived from the JPMorgan EMBI Global index. Loss given default is always assumed to remain constant at 60 percent. Panel 5 includes both investment-grade and high-yield bonds.

Figure 1.19. US and Emerging Market Corporate Bond Spread Decomposition and Leverage

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

Volatility Is Compressed

The bountiful liquidity provided by major cen- tral banks through their QE programs, as well as the expectation that central banks will react swiftly to market stress, has further strengthened the link between low risk premiums and low volatility. The impact of economic and financial conditions on US equity volatility is examined through an explanatory

model, which offers three main findings (Figure 1.20, panel 1).20 • First, stable macroeconomic fundamentals have

reduced volatility, as captured by the volatility of

20The analysis is centered on the United States as the most repre- sentative measure of global market volatility, given that the United States accounts for over one-third of the global equity market and dominates trading of implied volatility futures.

–1.0

–0.8

–0.6

–0.4

–0.2

0.0

0.2

0.4

0.6

0.8

2004–07 average

10 11 12 13 14 15 16 17

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

2010 11 12 13 14 15 16 17

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

2010 11 12 13 14 15 16 17

Top 50 Bottom 300

Macroeconomic fundamentals Corporate performance Funding and liquidity conditions

External spillovers VIX index Model-fitted VIX index

1. Drivers of Declining Equity Volatility (Z-score, number of standard deviations)

Equity volatility touched record lows in 2017.

2. Realized Volatility of Individual Stocks (US/S&P 500 stocks, 90-day historical volatility)

S&P 500 index volatility is suppressed by large firms ...

3. Net Income (Percent of assets, four-quarter moving averages)

... whose earnings are stronger and more stable ...

4. Dividends and Stock Repurchases (Percent of assets, four-quarter moving averages)

... and whose payouts are more generous.

Sources: Bloomberg Finance L.P.; and IMF staff calculations. Note: The Chicago Board Options Exchange Volatility Index (VIX) model is an ordinary least squares regression using quarterly data since 2004:Q1. Macroeconomic fundamentals include US GDP growth and the rolling 12-month standard deviation of the Citi US Economic Surprise Index. Corporate performance includes net income to assets and payouts to assets for Standard & Poor’s (S&P) 500 firms. Funding and liquidity conditions include the TED spread (the difference between the interest rates on interbank loans and on short-term US government debt, “T-bills”); average euro, Japanese yen, and British pound one-year cross-currency basis swap rate; and supply of US Treasuries net of Federal Reserve purchases. External spillovers include the average of 10-year Greek, Italian, Portuguese, and Spanish yield spreads to the German 10-year yield. The VIX is used as the dependent variable in the volatility model.

Figure 1.20. Long-Term Drivers of the Low-Volatility Regime

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

economic surprises and the strength of underlying growth. Accommodative monetary policy has helped support this economic environment.

• Second, the accommodative funding and liquidity conditions provided by monetary policy have left volatility lower than in previous cycles.

• Third, corporate performance has remained stable and contributed to steady investor earnings expectations and reduced volatility.

This steady corporate performance—and associated low realized volatility measures—has been driven in part by large-cap companies (Figure 1.20, panel 2). The market performance of large-cap companies has been underpinned by stronger and more resilient earn- ings (Figure 1.20, panel 3). At the same time, how- ever, cash-rich US corporations have used payouts via dividends and stock repurchases to smooth equity valu- ations and compress volatility (Figure 1.20, panel 4). With payouts rising to a high percentage of assets, this tool may be less available to smooth earnings. Finally, increased dispersion of returns across sectors, which may reflect potential policy shifts in the United States and abroad, has also contributed to reduced volatility of the overall index.

Low Volatility, Financial Leverage, and Liquidity Mismatches Could Amplify a Market Shock

Low volatility can increase the sensitivity of the financial system to market risk. First, in standard portfolio risk models, low volatility enables investors to increase their exposure to financial assets and so their sensitivity to market risk. Second, low volatility can create incentives for investors to increase financial leverage, which collectively can amplify market shocks. An example of this effect is the increased popularity of so-called volatility-targeting investment strategies (Figure 1.21, panel 1). These strategies seek to keep expected portfolio volatility to a specific targeted level. Lower market volatility (in both global equity and bond markets) then means that greater financial lever- age is needed to meet volatility targets (Figure 1.21, panel 2).21

However, during volatility spikes, these strategies can lead to significant asset sales to pare back leverage.

21Derivatives such as equity index futures are commonly used to achieve greater financial leverage by volatility-targeting invest- ment strategies.

Such an episode took place in August 2015,22 when a representative volatility-targeting investment strategy cut its global equity exposure drastically (Figure 1.21, panel 3).23 The size of US equity holdings held by volatility-targeting investment strategies may be larger than $0.5 trillion today.24 Although this is less than 2.5 percent of the market capitalization of all US publicly traded equities, the trading volume related to deleveraging from these trading strategies could be much larger, particularly at times of equity mar- ket stress.25

The low-interest-rate environment has also raised bond market risk. Low interest rates have reduced coupons of newly issued bonds. While this has been a boon for issuers, helping to reduce debt servicing costs, it has come at the price of higher market risk for investors. The prices of those bonds are more sensitive to changes in interest rates (increasing their duration). This market risk is illustrated in Figure 1.22, panel 1, which simulates the impact of an immediate 100 basis point shock on long-term interest rates. The analysis shows that this impact has increased over time as dura- tion has increased. Losses in bond funds might lead to outflows from asset managers. Indeed, the sensitivity of outflows appears to have increased in relation to periods of large negative returns in US high-yield bond funds (Figure 1.22, panel 2). A significant outflow might trigger sales of riskier and less liquid assets held by open-end mutual funds, which could lead to sub- stantial changes in the price of these instruments and

22The Chicago Board Options Exchange Volatility Index (VIX) increased sharply to 40.7 percent on August 24, 2015, its highest level since September 2011, from 13.0 a week earlier. While rising concerns about a hard landing in China amid a significant decline in oil prices were major drivers of the increase in market volatility, market participants’ concern about a perceived end to the Federal Open Market Committee quantitative easing policy may have also played a major role in the equity market sell-off.

23The Standard & Poor’s (S&P’s) 500 index exposure for a repre- sentative volatility-targeting investment strategy uses the AQR Risk Parity Fund mutual fund as its proxy portfolio.

24This estimate assumes that the universe of volatility-targeting investment strategies holds on average a portfolio in which global US equities account for 60 percent of the exposure and bonds account for 40 percent. The result is also adjusted by an estimated leverage number based on the volatility targets of different volatility-targeting investors. US equity exposure is assumed to be about half of the exposure to global equities. This is similar to the average geographic breakdown of equity investments in the AQR Risk Parity Fund over the past two years.

25Chandumont 2016 estimates that selling from volatility- targeting funds accounted for between 9 and 16 percent of all trading volume in S&P 500 futures during August 24–26, 2015.

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

affect the value of these assets held by other investors. Figure 1.22, panel 3 shows that mutual funds hold a greater share of the high-yield bond market than in the past.

Prolonged normalization of monetary policy could mean continued low volatility and a further buildup of exposures, duration, and financial leverage. This would make the financial system even more sensitive to mar- ket risk, storing up medium-term vulnerability.

Efforts Are Needed to Help Lessen Stability Risks Regulators should be attentive to the potential for

a substantial increase in asset market volatility to con- tribute to destabilizing feedback effects such as asset fire sales and adverse liquidity and leverage spirals. To lessen these risks, financial regulators should continue working to ensure that financial institutions maintain robust risk management standards at all points in the credit, business, and interest rate cycles. In addition,

20 15

:Q 1

Q 2

Q 3

Q 4

16 :Q

Q 2

Q 3

Q 4

17 :Q

Q 2

Investment Strategy Volatility Target (percent)

Flexibility to Deviate from Volatility Target

AUM Mid-2017 Growth in AUM Past Three Years (percent)

Variable Annuities 8–12 Low $440 billion 69

CTA/Systematic Trading 15 Medium $220 billion 19

Risk Parity Funds 10–15 Medium–high $150–$175 billion ...

0.5

1.0

1.5

2.0

2.5

2012 13 14 15 16 17

Global equity index volatility (left scale) Global bond index volatility (left scale) Leverage of 60/40 portfolio with a 12 percent volatility target (right scale)

100

120

Exposure to global non-US equities (right scale) Exposure to US equities (right scale) VIX index (maximum quarterly level, left scale)

Sources: Annuity Insights; Barclays Capital; BarclayHedge; and IMF staff calculations.

Figure 1.21. Leveraged and Volatility-Targeting Strategies

1. The Growth of Volatility-Targeting Investors

2. Leverage for a Theoretical Volatility-Targeting Investment Portfolio1

(Sixty-day moving average)

Lower volatility drives investors to increase financial leverage to meet their return and volatility targets ...

3. Global Equity Exposure for a Representative Volatility- Targeting Investment Portfolio2

(Percent/net asset value)

... leading to rising equity exposures that are prone to sell-offs during volatility spikes.

Sharp reduction in equity exposures as volatility spiked in August 2015

Sources: Bloomberg Finance L.P.; Federal Reserve; Investment Company Institute; and IMF staff calculations. Note: AUM = assets under management; CTA = Commodity Trading Advisor; VIX = Chicago Board Options Exchange Volatility Index. 1The leverage calculation for a theoretical volatility-targeting investment strategy assumes a theoretical investment portfolio consisting of 60 percent global equities/40 percent bonds and an annual return volatility target of 12 percent. Leverage is defined as total investment exposure divided by the net asset value of the portfolio. The calculation uses a 60-day realized volatility moving window on the returns of equity and bond investments. The MSCI World Index is used as the proxy for equity investments; the Bloomberg Barclays Global Aggregate Total Return Value Unhedged index is used as the proxy for bond investments. 2The S&P 500 index exposure for a representative volatility-targeting investment strategy uses the AQR Risk Parity mutual fund as its proxy portfolio. The exposure data are obtained using Bloomberg’s port function and reflect the percentage exposure of the fund’s portfolio to equity index futures as a percentage of market value.

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

supervisors, regulators, and firm management should closely monitor and assess financial institutions’ exposure to asset classes where there are indications that the search for yield has contributed to valua- tion pressure.

There is also a need for regulators to endorse a clear and common definition of financial leverage in invest- ment funds and to improve data transparency, partic- ularly with respect to derivatives. Lack of progress on regulation on the use of derivatives is a concern given that the use of financial leverage through derivatives

appears to be on the rise as fund managers seek to enhance low yields, particularly in strategies that target a specified level of price volatility.

Policymakers should continue to strengthen supervi- sory frameworks relating to liquidity risk management. This could be done by building on recent initiatives and recommendations to include greater flexibility in redemption and dealing frequency,26 marking illiquid

26See US SEC (October 2016), FSB (January 2017), IOSCO (July 2017), and UK FCA (February 2017).

Projected UK losses (left scale) Projected euro area losses (left scale) Projected US losses (left scale) Percent loss (right scale)

United States (left scale) Europe (right scale)

1. Estimated Loss to Fixed-Income Mutual Funds Following a 100 Basis Point Shock to Interest Rates

Higher duration leaves investors more vulnerable to interest rate risk ...

100

150

Bi lli

on s

of U

S do

lla rs

Pe rc

en t

250

200

300

350

1994–95 99–2000

(Taper tantrum)

2004–06 Jun. 13 Mar. 17

Sources: Bloomberg Finance L.P.; EPFR Global; Federal Reserve; Investment Company Institute; and IMF staff estimates. Note: In panel 1, data are based on prior periods of US monetary policy tightening starting in February 1994, July 1999, July 2004, and December 2015 and periods of large interest rate moves since the global financial crisis. The Barclays Capital Global Aggregate index is used as a proxy for duration of an average fixed-income portfolio. Total fixed-income mutual fund assets are used to calculate the dollar losses from a parallel 100 basis point increase in interest rates. Panel 2 shows periods when cumulative losses have exceeded 5 percent. There have been only four periods over the past decade when cumulative monthly losses on US high-yield bond benchmarks have exceeded 5 percent—a typical threshold used by investors when implementing stop-loss strategies. These risk management strategies are commonly used by investors to reduce their holdings in risky assets if prices breach certain prespecified loss limits. By closing out the position, the investor is hoping to avoid further losses.

Figure 1.22. Vulnerability of the US Corporate Credit Investor Base to Shocks

2. Flows and Performance of US High-Yield Bond Mutual Funds (Periods when cumulative losses exceeded 5 percent)

... at a time when there is greater sensitivity of investor outlows.

Ra tio

: o ut

flo w

s to

p er

fo rm

an ce

0.0

0.6

0.4 0.5

0.2 0.1

0.3

0.7 0.8 0.9 1.0

–35 –30 –25 –20 –15 –10 –5

Cumulative returns (percent)

3. Mutual Funds Holdings as Share of Total High-Yield Bond Market (Percent)

Liquidity mismatch risk is also a concern amid the rise in illiquid assets held by mutual funds globally.

2008 09 10 11 12 13 14 15 16

Sep–Nov 2008

Jun–Sep 2015

Aug–Sep 2011

Nov 15–Jan 2016

Rising sensitivity of investor outflows to periods of large losses

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

assets to market, and the treatment of institutional investors, as well as through better guidance on the use of particular risk management tools and enhanced disclosure requirements.

For borrowers in frontier markets and low-income countries, authorities should develop institutional capac- ity to deal with the risk that accompanies increased issu- ance of marketable debt securities. Authorities should formulate a comprehensive debt management strategy that incorporates exchange rate, interest rate, and liquid- ity risks associated with the issuance of external debt and explore liability management operations to mitigate refinancing risk.27 Authorities should ensure efficient use of the borrowed funds by strengthening public invest- ment management. They should also enhance investor relations programs to better understand and inform the international investment community regarding their debt issuance strategy.

The Rise in Leverage Leverage in the nonfinancial sector has increased since 2006 in many G20 economies amid easy financing conditions. While this has helped facilitate the recovery in aggregate demand, it has also made the nonfinancial sector more sensitive to changes in interest rates. Private sector debt service burdens have increased in several major economies as leverage has risen, despite declining borrowing costs. Debt servicing pressure could mount further if leverage continues to grow and could lead to greater credit risk in the financial system. China has seen a rapid buildup in leverage, so the recent derisking measures are a welcome step. Yet continued rapid credit growth and accumulated vulnerabilities at smaller banks make it challenging to fully address systemic risks.

Group of Twenty Nonfinancial Sector Leverage

Aggregate G20 Debt-to-GDP Ratios Are Higher than before the Global Financial Crisis

Among G20 economies, total nonfinancial sector debt—borrowing by governments, nonfinancial companies, and households from both banks and bond markets—has risen to more than $135 trillion, or about 235 percent of aggregate GDP (Figure 1.23, panel 1).28 This partly reflects economic develop-

27See IMF 2017b. 28G20 aggregates are based on the 19 individual economies in the

group (the 20th member is the European Union).

ments since the global financial crisis. The rise in sov- ereign debt is largely due to the downturn in GDP, but is also due in part to the necessary actions taken by governments to stabilize economies and financial sectors. Private sector credit growth has helped facil- itate the subsequent recovery in aggregate demand, and so has cushioned economic growth against further downside risks. But higher debt has made the nonfinancial sector more sensitive to changes in interest rates.

In G20 advanced economies, the debt-to-GDP ratio has grown steadily over the past decade and now amounts to more than 260 percent of GDP. In G20 emerging market economies, leverage growth has accelerated in recent years. This was driven largely by a huge increase in Chinese debt since 2007, though debt-to-GDP levels also increased modestly in other G20 emerging market economies (Fig- ure 1.23, panel 2).

Overall, about 80 percent of the $60 trillion increase in G20 nonfinancial sector debt since 2006 has been in the sovereign and nonfinancial corporate sectors (Figure 1.23, panel 3). Much of this increase has been in China (largely in nonfinancial companies) and the United States (mostly from the rise in general government debt). Each country accounts for about one-third of the G20’s increase. Average debt-to-GDP ratios across G20 economies have increased in all three parts of the nonfinancial sector (Figure 1.23, panel 4).

There has also been a broad increase in nonfinancial debt-to-GDP ratios across individual G20 econo- mies since 2006; only Argentina and Germany have experienced a decline in total nonfinancial sector debt to GDP (Table 1.1). In some economies, individual sectors have deleveraged. For example, household debt to GDP fell in Germany and the United States, in particular. Nonfinancial corporate leverage declined the most in Argentina, Japan, and the United Kingdom. But in the majority of cases in the G20, nonfinancial debt-to-GDP ratios have risen.

While gross liabilities have risen, the development of net debt—gross debt minus financial assets—has varied across the nonfinancial sector in G20 advanced economies (Figure 1.23, panel 5). General government net debt rose along with gross debt over the decade since 2006. Nonfinancial private sector net debt, how- ever, fell as savings and higher asset prices helped build up financial assets more quickly than liabilities. This, in turn, has helped support the recovery in spending

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

G20 emerging market economies, excluding China

G20 advanced economies China

Nonfinancial companies Households General government Nominal GDP

Debt (left scale) Cash (right scale)

1. Gross Debt and GDP (Trillions of US dollars)

Debt has been rising more quickly than GDP ...

3. Change in Gross Debt, 2006–16 (Trillions of US dollars)

... and in sovereigns and firms ...

100

120

140

Sources: Bank for International Settlements; Bloomberg Finance L.P.; Haver Analytics; IMF, World Economic Outlook database; and IMF staff calculations. Note: Data are adjusted for foreign exchange movements by converting to US dollars at the end-2016 exchange rate. Advanced economy nonfinancial corporate debt is shown net of estimated intercompany loans. In panel 3, OTH = other Group of Twenty (G20) economies. Panel 4 shows the average debt-to-GDP ratio across the G20 economies, by sector. Panel 5 shows debt minus financial assets as a percent of GDP. Panel 6 is based on a sample of more than 2,600 nonfinancial companies in continental Europe, Japan, the United Kingdom, and the United States. Each dot shows average debt and cash to assets for the same 50 firms. Data labels in the figure use International Organization for Standardization (ISO) country codes.

Figure 1.23. Group of Twenty Nonfinancial Sector Credit Trends

2. Gross Debt-to-GDP Ratios by Region (Percent)

... largely in advanced economies and China ...

100

500 1,000 1,500 2,000

4. Average Gross Debt-to-GDP Ratios by Sector (Percent)

... with debt-to-GDP ratios above precrisis levels.

92 94 96 98 2000 02 04 06 08 10 12 141990 16

92 94 96 98 2000 02 04 06 08 10 12 141990 16 92 94 96 98 2000 02 04 06 08 10 12 141990 16

5. Advanced Economy Net Debt-to-GDP Ratios by Sector (Percent)

Private sector financial assets have risen ...

6. Advanced Economy Nonfinancial Corporate Debt and Cash (Percent of assets)

... but cash is unevenly distributed among firms.

0 2,500

General government Nonfinancial companies

USA 11.1

OTH 6.8

OTH 3.0

OTH 4.7

AUS 0.7

USA 1.5

IND 0.8

USA 4.5

CAN 0.7

JPN 2.7

GBR 1.4

CHN 4.4

CHN 14.4

Households

CHN 3.9

20 06 08 10 12 14 16

General government

–60

–50

–40

–30

–20

20 06 08 10 12 14 16

–280

–260

–240

–220

–200

–180

–160

20 06 08 10 12 14 16

Nonfinancial companies

Households Firms with the highest debt have the lowest cash

Sample (ordered by debt-to-assets ratio)

100

150

200

250

300

General government

Households Nonfinancial companies

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

and GDP. But it is important not to draw too much comfort from this development. While debt accumula- tion is not necessarily a problem, one lesson from the global financial crisis is that excessive debt that creates debt servicing problems can lead to financial strains. Another lesson is that gross liabilities matter. First, in a period of stress, it is unlikely that the whole stock of financial assets can be sold at current market values— and some assets may be unsellable in illiquid condi- tions. Second, the aggregate data used here do not account for differences in the distribution of assets and liabilities. For example, the younger population might have a greater proportion of debt in the household sector, while the older population might have a greater proportion of financial assets.

A similar argument can be made about cash holdings in nonfinancial companies. Although cash holdings may be netted from gross debt at an individ- ual company—because that firm has the option to pay back debt from its stock of cash—it could be mislead- ing to do so in the aggregate data generally used in this section. This is because the distribution of debt and cash holdings differs between companies. Figure 1.23, panel 6, which is based on debt and cash stocks held by a sample of more than 2,600 European, Japanese, and US companies, shows that those with higher debt also tend to have lower cash holdings and vice versa.

Although G20 gross private nonfinancial debt has increased in the aggregate, the reasons for higher leverage differ across sectors. For example, changes in household leverage appear to be broadly associated with lower borrowing costs and house price move-

ments (Figure 1.24, panel 1). Higher house prices, driven up by buoyant market conditions and risk appetite, mean that not only is more borrowing needed to purchase properties but also that more collateral is available to support the increased borrowing. Lower interest rates make new borrowing more attractive for households. Chapter 2 examines household indebted- ness in more detail. It finds that household debt has continued to grow over the past decade across a broad set of countries. It also concludes that high growth in household debt in the medium term is associated with a greater probability of a banking crisis.

The increase in corporate debt has taken place during loose financing conditions, just as during the period before the global financial crisis (Figure 1.24, panel 2). Low interest rates probably stimulated greater demand for credit from companies as larger debt became more affordable, leading to changes in capital structures. Easy financing conditions—a combination of low interest rates, buoyant market valuations, and low volatility—have reduced the probability of default as measured by credit models, which is likely to have increased the willingness of lenders to supply credit to companies.29

However, this contemporaneous default proba- bility is based on current market conditions, which might not last. If there are adverse shocks, a feedback

29Growth in private sector debt in some emerging market econ- omies may also be linked to improvements in credit infrastructure (such as increased use of credit registries and improvements in credit risk evaluation) as well as policies to foster lending to small and medium enterprises and financial inclusion.

Table 1.1. Sovereign and Nonfinancial Private Sector Debt-to-GDP Ratios (Percent)

Advanced Economies Emerging Market Economies JPN CAN USA GBR ITA AUS KOR FRA DEU CHN BRA IND ZAF TUR MEX RUS SAU ARG IDN

General Government

2006 184 70 64 41 103 10 29 64 66 25 66 77 31 45 38 10 26 70 36

2016 239 92 107 89 133 41 38 96 68 44 78 70 52 28 58 16 13 54 28

Households 2006 59 74 96 90 36 105 70 44 65 11 14 10 39 9 12 8 12 4 11

2016 57 101 79 88 42 123 93 57 53 44 23 10 35 18 16 16 15 6 17

Nonfinancial Corporations

2006 100 76 65 79 67 73 83 56 49 105 39 38 33 27 14 32 28 20 14

2016 92 102 72 73 71 79 100 72 46 165 44 45 37 67 28 52 50 12 23

Total 2006 343 221 225 210 205 187 183 164 180 142 118 125 104 81 64 49 66 93 61

2016 388 295 259 250 246 243 232 226 168 254 145 125 124 113 103 84 78 73 68

Sources: Bank for International Settlements; Haver Analytics; IMF, World Economic Outlook database; and IMF staff calculations. Note: Dark shading denotes a higher debt-to-GDP ratio in 2016 than in 2006. The table shows debt at market values. Advanced economy nonfinancial corporate debt is shown net of estimated intercompany loans where data are available. Data labels in the table use International Standardization Organization (ISO) codes.

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

loop could develop, which would tighten financial conditions and increase the probability of default, as happened during the global financial crisis. Thus the low contemporaneous default probability could mask risks associated with the buildup of corporate lever- age, a phenomenon that has been called the “volatility paradox.”30

Higher Private Sector Debt Has Raised Servicing Costs and Could Increase Vulnerabilities

While debt has generally increased relative to GDP, it happened in a period of falling and low interest rates. So what happened to debt affordability over this period? This question is important because measures of debt affordability tend to be good vulnerabil- ity signals, particularly when debt levels are high.31 Although lower interest rates have helped lower sover- eign borrowing costs, in most of the G20 economies where companies and households increased leverage, nonfinancial private sector debt service ratios— defined as annualized interest payments plus income amortization—also increased (Figure 1.25, panel 1).

Moreover, there are now several economies where debt service ratios for the private nonfinancial sectors are higher than average and where debt levels are also high. Figure 1.25, panel 2, shows that this is partic- ularly the case for the nonfinancial private sector in Australia, Canada, and China, and for the household sector in Korea (debt service ratios for households and nonfinancial companies are available only for G20 advanced economies).

The distribution of debt within an economy’s corpo- rate and household sectors is also important in assessing payment pressures. While the aggregate data on debt service ratios used here do not allow an examination of the distribution, other work might shed some light on this question. The April 2017 GFSR found (for com- panies in the United States) a deterioration in interest coverage ratios for those most indebted, particularly in the energy sector. In emerging market economies, however, commodity companies and industrials made up a significant proportion of firms with weak interest

30See Adrian and Shin 2013 and Geanakoplos 2010 for a discussion of the leverage cycle, and Brunnermeier and Sannikov 2014 and Adrian and Brunnermeier 2016 for a discussion of the volatility paradox.

31Chapter 2 discusses household debt service capacity as a vulnerability indicator. See also work at the Bank for International Settlements on this issue, including Drehman, Juselius, and Korinek 2017; BIS 2017; and BIS 2012.

Economies with nominal house price growth greater than 25 percent (2006–16)

Economies with nominal house price growth less than 25 percent (2006–16)

–4

–3

–1

–2.0

–1.5

–1.0

–0.5

0.0

0.5

1.0

1.5

2.0

2000 02 04 06 08 10 12 14 16

St an

da rd

d ev

ia tio

ns fr

om m

ea n

Cr ed

it ga

p (p

er ce

nt ag

e po

in ts

) –20

–10

2. Average Nonfinancial Corporate Credit, Financing Conditions, and Default Probability (Four-quarter moving average)

Figure 1.24. Group of Twenty Nonfinancial Private Sector Borrowing

1. Change in Household Gross Debt to GDP, 2006–16 (Percentage points)

Household debt has risen broadly with house prices.

Corporate debt has built up with easy financing conditions.

Financing conditions (right scale)

Model-based probability of default (right scale)

Credit (left scale)

CH N

CA N

KO R

AU S

TU R

ID N

M EX ZA

D EU FR A

IT A

JP N

G BR US

IN D

BR A

RU S

Sources: Bank for International Settlements; Bloomberg Finance L.P.; Haver Analytics; Moody’s CreditEdge; Organisation for Economic Co-operation and Development; and IMF staff calculations. Note: In panel 1, house price growth is from 2008 in Brazil; from 2010 in China, India, and Turkey; and is not available for Argentina and Saudi Arabia. Panel 2 shows the average Group of Twenty: corporate debt-to-GDP gap, financing conditions (average of corporate borrowing rates, book-to-market ratios, and implied volatility), and probability of default from the Moody’s KMV model (based on a sample of more than 41,000 companies). Data labels in the figure use International Organization for Standardization (ISO) country codes.

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

Figure 1.25. Group of Twenty Nonfinancial Private Sector Credit and Debt Service Ratios

1 2 3 4 5 6 7 8 9 10 11 12 Years since start of boom

–10

1 2 3 4 5 6 7 8 9 10

1. Change in Private Nonfinancial Sector Debt and Debt Service Ratios, 2006–16

2. Debt Service Ratios and Debt, 2016 (Percent)

3. Change in Credit-to-GDP Ratio (Percentage points)

4. Cumulative Real House Price Growth (Percent)

–6

–4

–2

Ch an

ge in

d eb

t s er

vi ce

r at

io (p

er ce

nt ag

e po

in ts

)

Change in debt to GDP (percentage points)

200 300 400 500 600 700

Nonfinancial companies

D ev

ia tio

n fr

om m

ea n,

p er

ce nt

ag e

po in

Gross debt to income (percent) 40 60 80 100 120 140 160 180

Households

D ev

ia tio

n fr

om m

ea n,

p er

ce nt

ag e

po in

Gross debt to income (percent)

Greater debt payment pressure

Years since start of boom

0 50 100 150 200 250

Nonfinancial private sector

D ev

ia tio

n fr

om m

ea n,

p er

ce nt

ag e

po in

Gross debt to GDP (percent)

–40 –20 0 20 40 60 80 100 120

Debt service ratios have increased with higher leverage, despite low interest rates.

Debt service ratios in some countries are now at high levels ...

... in economies with credit booms ... ... and house price growth.

Sources: Bank for International Settlements; Bloomberg Finance L.P.; national statistical offices; Organisation for Economic Co-operation and Development; and IMF staff calculations. Note: Debt service ratios are defined as annualized interest payments plus amortizations as a percentage of income, as calculated by the Bank for International Settlements. In panel 1, the size of the circles is proportional to debt to GDP in 2016. In panel 2, income is gross disposable income plus interest payments (plus dividends paid for firms). Panel 3 shows Group of Twenty economies with higher demeaned nonfinancial private sector debt service ratios and debt levels against past booms. Past booms are for a sample of 43 advanced and emerging market economies where the credit-to-GDP gap rose above 10 percent. The start and end dates of the booms are defined as periods when the credit gap was above 6 percent. Financial crisis dates were taken from Laeven and Valencia 2012. Data labels in the figure use International Organization for Standardization (ISO) country codes.

Canada

Average of credit booms that did not lead to financial crises

–10

–5

USA

GBR

KORJPN

ITA

DEU

FRA

CAN AUS

–3

–2

–1

6 TUR

ZAF

RUS

MEX

IDN

IND

CHN

BRA

USA

GBR ITA

KOR

JPNDEU

FRA

CAN

AUS

TUR

ZAF

RUS

MEX IDN

IND

CHN

BRA

USA

GBR

KOR

JPN ITA DEU

FRA

CAN

AUS

–2.0

–1.5

–1.0

–0.5

0.0

0.5

1.0

1.5

2.0

USA

GBR

KOR

JPNITA

DEU

FRA CAN

AUS

Australia

China

Korea

Average of credit booms that led to financial crises

China

Australia

Canada

Korea

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

coverage ratios. Similarly, ECB 2017 shows that the dis- tribution of household debt service ratios reveals greater vulnerability among those that had more recently taken out a mortgage to finance a house purchase than was evident from the aggregate figure.

Although not all credit booms lead to recessions, it is interesting to compare the credit booms in econo- mies most likely to face payment pressures with past experience. While the boom in Australia is similar to the average of past credit booms that did not lead to a financial crisis, the boom in Canada has been longer than the average of these benign booms, and the boom in China has been steeper than the average of past credit booms that did coincide with a financial crisis (Figure 1.25, panel 3). In addition, in three of the economies with the highest debt service ratios, there has been a steep increase in real house price valuations (Figure 1.25, panel 4).

Experience has shown that a buildup in leverage associated with a run-up in house price valuations can develop to a point that they create strains in the non- financial sector that, in the event of a sharp fall in asset prices, can spill over to the economy. For example, Chapter 2 finds that the relationship between future GDP growth and household debt is driven mostly by mortgage debt. This could be because of the procycli- cality of home equity lines of credit, or more generally because of wealth effects that lead households to cut consumption when the value of their housing assets declines.32

Overall, there are now several major economies where debt servicing pressure in the private nonfinan- cial sector is already high. Weaker households and companies in these countries could have trouble repay- ing their debt if interest rates rise or if incomes fall.

Policies Are Needed to Reduce Vulnerabilities in the Private Nonfinancial Sector

Policymakers should address the risks from contin- ued increases in debt and leverage across sectors by drawing on, and enhancing where needed, an appro- priate mix of macroprudential and microprudential policies, preemptive regulatory measures, and close monitoring of balance sheets.

Higher household debt burdens should be reduced where debt servicing pressures are already high and should not grow further where debt servicing is

32See also Mian and Sufi 2011 and Schularick and Taylor 2012.

currently manageable but debt levels are elevated. This can be achieved through a combination of measures, including limits on debt-service-to-income and loan-to-value ratios, and measures to restrict loan contracts. Some countries have undertaken measures to address high house price valuations and deter further buildup of household debt. Policy measures, however, must carefully balance minimizing the medium-term risks to financial stability while not harming the potential long-term benefits of financial inclusion and development.

Policymakers should vigilantly monitor nonfinancial corporate leverage. Macroprudential measures extended through banks (such as sectoral capital requirements or risk weights on foreign currency credit) could also be considered to reduce or prevent a further buildup in cor- porate debt. In addition, tax reforms that reduce incen- tives for debt financing could help attenuate the risk of a further buildup in leverage and may even encourage firms to lower existing tax-advantaged leverage. More broadly, measures to foster smooth corporate delever- aging should be deployed where needed, including by strengthening corporate restructuring mechanisms.

China: From Derisking to Deleveraging— Challenges Ahead

The rapid rise in nonfinancial sector leverage in China in recent years, along with the size, complex- ity, and pace of growth of its financial system, point to continued financial stability risks. Banking sector assets are now 310 percent of GDP, nearly three times the emerging market average and up from 240 percent at the end of 2012. Rapid increases in intrafinancial-system credit have been an important factor in this growth (see Figure 1.26, panel 1). This reflects both the growing use of short-term wholesale funding to boost leverage and profits (Figure 1.26, panel 2) and shadow credit to firms and other non- financial borrowers (Figure 1.26, panels 3 and 4), particularly by small and medium-sized banks.33 This

33Shadow credit refers to banks’ nonloan, nonbond credit to nonfinancial borrowers. This includes assets that are on balance sheet (trust beneficiary rights, specialized asset management plans, and other structured assets) and off balance sheet (bank-sponsored wealth management plans). Estimates of off-balance-sheet bank credit are calculated as 65 percent of outstanding wealth management plans, which deduct the portion of underlying plan assets that are claims on financial or public sector counterparties, as reported in China Bank Wealth Management Market Annual Report 2016.

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

has increased the opacity of intermediation, increased the use of unstable short-term funding, and raised sensitivity to liquidity stress.

China recently introduced a range of prudential and administrative measures to contain these vulner- abilities. Efforts to derisk the financial system using better-designed regulatory tools (such as the Macro- prudential Assessment, or MPA) aim to slow growth in banks’ supply of shadow credit, reduce dependence on interbank funding, and contain regulatory arbitrage.34

34Among examples of such measures are the People’s Bank of China’s inclusion of wealth management products in its MPA frame- work, counting negotiable certificates of deposit toward the pru- dential limit on interbank liabilities, and tightening corporate bond collateral requirements for exchange-traded repurchase agreements.

On-balance-sheet shadow credit products at small and medium-sized banks declined sharply in late 2016 and early 2017. Growth in off-balance-sheet shadow credit, in the form of wealth management products, has also recently reversed by the largest amount in the post- crisis period (Figure 1.27). This coincided with rising interbank and bond market interest rates and stalling corporate bond issuance.

Authorities Face a Delicate Balance between Tightening Financial Sector Policies and Slowing Credit Growth

Curbing shadow credit could have an out- size impact on banks’ capacity to increase credit. Bank-level data show that roughly half of lenders’

Figure 1.26. Chinese Banking System Developments

Big 5 Small and medium-sized banks Total

Other assets Claims on financial sector Claims on nonfinancial sector

Shadow credit: off balance sheet Shadow credit: on balance sheet New loans

Big 5 Small and medium-sized banks

1. Contribution to Bank Asset Growth (Percent, year over year)

3. Net Increase in Private Nonfinancial Credit (Trillions of renminbi)

4. Bank Shadow Credit: Net Increase and Ratio to Deposits (Trillions of renminbi and percent)

2. Nondeposit Funding (Maturing < 1 year, percent of total assets)

Intrafinancial system credit has driven bank growth ...

... but also reflecting significant shadow credit ... ... particularly from small and medium-sized banks.

... increasing reliance on risky funding ...

2014 15 16 2014 15 16

–2 0 2 4 6 8

10 12 14 16 18

2014 15 16 17 5

2011 12 13 14 15 16

2014 15 16

Shadow credit as percentage of total new credit

41% 53%51%

Intrafinancial system credit

Sources: Haver Analytics; People’s Bank of China; SNL Financial; and IMF staff calculations. Note: Shadow credit refers to banks’ nonloan, nonbond credit to nonfinancial private borrowers, both on and off balance sheet. For a complete definition, please see footnote 33. Panels 2, 3, and 4 are based on publicly available financial statement data for 32 of China’s largest banking groups.

1313 9

Net increase (trillions of renminbi)

Shadow credit to deposits (percent)

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

estimated credit in recent years was extended via such products.35 As shadow credit typically requires less capital and provisioning than regular loans, reduc- ing its growth would free up only enough capital to support a smaller increase in lending, leading to a net slowdown in the flow of total credit. For instance, if banks expanded shadow credit by 27 percent—the pace in 2016—their projected retained earnings would support total credit growth (loans and shadow credit) of 17 percent year over year, just above the actual growth rate in 2016. If banks instead kept shadow credit constant, increasing only loans, the same amount of retained earnings would support credit growth of 11 percent, in line with nominal GDP growth in the second quarter of 2017 (Fig- ure 1.28, panel 1).

Banks face a trade-off between using retained earnings to address vulnerabilities or support credit growth.36 If some retained earnings are used to increase the pace of loss recognition, or increase capital and provisions against a modest portion of existing shadow products, credit capacity would decline further (Figure 1.28, panel 2). Balance sheet vulnerabilities from shadow credit would also recede only gradually at smaller banks, remaining elevated relative to the biggest banks (Figure 1.28, panel 3).

Derisking Will Weigh on Some Banks’ Profitability and Business Models

Shifting away from shadow credit products and interbank funding will improve bank balance sheets over time, but in the short term could also decrease bank profitability, weakening buffers at already vul- nerable banks and reducing capacity to expand credit. Bank earnings in China have fallen in recent years, driven by an uptick in provision expenses and lower net interest margins (Figure 1.29, panel 1). Small and medium-sized banks have sustained profitability in

35Based on publicly reported data for a sample of 32 of China’s largest banking groups. This calculation excludes corporate bonds held in banks’ securities portfolios. The total credit provision from these banks depicted is equivalent to roughly 90 percent of the total increase in nonfinancial credit in 2015 and 2016 (as measured by Total Social Financing flows).

36Banks can avoid this trade-off through recapitalization. Chinese banks have announced planned increases of RMB 66 billion in new common equity for 2017, or about 2 percent of end-2016 common equity at small and medium-sized banks. Raising capital in public markets is complicated, however, by rules against raising capital when price-to-book ratios are below 1.

On-balance-sheet shadow credit proxy Unsecured interbank liabilities

Net issuance

20 12

:Q 1

12 :Q

2 12

:Q 3

12 :Q

4 13

:Q 1

13 :Q

2 13

:Q 3

13 :Q

4 14

:Q 1

14 :Q

2 14

:Q 3

14 :Q

4 15

:Q 1

15 :Q

2 15

:Q 3

15 :Q

4 16

:Q 1

16 :Q

2 16

:Q 3

16 :Q

4 17

:Q 1

17 :Q

–1,500

–1,000

–500

500

1,000

1,500

2,000

2,500

3,000

–400

–200

200

400

600

800

1,000

1,200

2015 16 17

... and so did off-balance-sheet shadow credit.

2. Bank-Sponsored Wealth Management Product Net Issuance (Billions of renminbi)

Interbank lending and shadow credit dipped sharply in 2017 ...

Figure 1.27. China: Regulatory Tightening Has Helped Contain Financial Sector Risks

1. Small and Medium-Sized Banks: Monthly Change in Selected Balance Sheet Categories (Billions of renminbi, three-month average)

Sources: CEIC Data Co. Ltd.; China Banking Regulatory Commission; Haver Analytics; media reports; People’s Bank of China; Wind data; and IMF staff calculations.

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

part by shifting their business model toward shadow credit activities, which account for a growing share of revenue (Figure 1.29, panel 2) and balance sheets, with shadow products surpassing loan growth over the past three years by a wide margin.

A return to traditional lending would strain profits at smaller banks via several channels. Net interest income from loans and deposits fell from 1.7 percent of assets in 2011 to just 1.0 percent in 2016, reflect- ing the changing asset mix but also the higher (and relatively liberalized) interest rates in the shadow credit market (Figure 1.29, panel 2).37 Profitability could suffer if more credit flows through the formal loan market, which is subject to more conservative provisioning rules and macroprudential controls on sector allocation. Any tightening in shadow credit

37The deterioration in net interest margins is mostly attributable to the traditional lending and deposit-taking business, whereas shadow investment and funding activities have had a neutral or posi- tive contribution on a net basis, particularly at smaller lenders.

activities would likely crimp net fees and commis- sions, which have doubled since 2011 at smaller banks on the back of higher off-balance-sheet income related to shadow products.

Reducing wholesale funding will also weigh on credit growth, particularly at small and medium lenders. These banks have funded much of their growth via nondeposit-funding sources with shorter maturities. Nondeposit funding maturing in less than one year has risen to about 34 percent of assets, from 22 percent in 2011, with over half maturing in less than three months (Figure 1.29, panel 3). The result has been a sharp increase in short-term borrowing to finance long-maturity assets, with short-term nondeposit funding exceeding similar-maturity nonloan assets by about 6 percent of assets, or RMB 2.8 trillion (see Figure 1.29, panel 4). Any meaningful reduction in short-term mar- ket funding would require liquidating longer-term assets.

To be successful, regulatory tightening on lend- ers must be accompanied by reforms that reduce

0 5 10 15

2016 (actual)

... or increased capital against 10 percent of shadow credit1

... with a 25 bps fall in ROA ...

No shadow credit growth ...

Figure 1.28. Chinese Banks: Financial Policy Tightening and Credit Growth Capacity

1. Realized and Projected Credit Growth Capacity1 (Percent)

Curbing shadow credit slows overall credit ... ... and would reduce vulnerabilities only gradually.

... especially if other weaknesses are also addressed ...

3. Shadow Credit to Capital (Percent)

2. Net New Bank Credit, Realized and Projected Capacity (Trillions of renminbi; percent)

2017 total credit growth capacity

11%

2014 15 16 172 173

Shadow credit growth assumption

27%

Actual credit growth

Projected credit growth capacity 27%

100

200

300

400

500

600

700

800

2013 14 15 16 17E

Small and medium-sized

banks Shadow credit

growth assumption

Big 5 banks 27% 0%

Sources: Company annual reports; SNL Financial; and IMF staff calculations. Note: Shadow credit refers to nonbond, nonloan credit to nonfinancial private borrowers, both on and off balance sheet. For a complete definition, please see footnote 33. bps = basis points; E = estimated; ROA = return on assets. 1Credit growth capacity is calculated at the bank level for 32 firms as the maximum net new credit possible given assumptions for growth in shadow credit (on and off balance sheet) and common equity Tier 1 (CET1) capital. Changes in shadow credit affects the CET1 available to support credit growth. New shadow credit is assumed to carry regulatory capital risk weightings of 25 percent, whereas off-balance-sheet shadow credit carries a risk weighting of zero. Assumes firm-level profitability, dividend payout ratio, CET1 ratio, and loan mix from 2016 stay constant. 2Projected credit growth capacity assuming shadow credit growth of 27 percent year over year. 3Projected credit growth capacity assuming shadow credit growth of 0 percent year over year.

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

the economy’s vulnerability to slower credit growth. Authorities’ recent efforts to improve banks’ risk man- agement and reduce maturity and liquidity transfor- mation risks in shadow credit activities are necessary and must be deepened. Stability risks will nonetheless remain elevated, however, if banks support continued rapid credit growth: they will have fewer buffers to recognize losses, profitability could compress further at weaker lenders, and incentives for regulatory arbi- trage will remain strong. Raising new equity would allow banks to raise provisions and capital without slowing credit growth, but must be accompanied

by reforms to strengthen bank risk management and governance.

A broader reform package could help mitigate the economic impact of slower credit growth and tighter regulations while addressing vulnerabilities. On the borrower side, authorities must build on their commitment to reduce corporate leverage, resolve nonviable firms, and improve credit efficiency.38 With lenders, regulation to reduce shadow credit risks and

38IMF 2016b, 2016c, and 2017f discuss progress and recommen- dations on these topics in more detail.

Net income (percent of average assets, left scale)

Less than three months Three to 12 months Over 12 months

Net interest margin (percent, left scale) Provisions (percent of average assets, right scale)

Net fee and commission income NII: other interest income minus other interest expense NII: loans and deposits Provision expense

Big 5 Small and medium-sized banks Total

1. Selected Profitability Indicators (Percent of average assets, percent)

Bank earnings are lower due to narrower margins and rising provisions ...

0.9

1.4

Sources: SNL Financial; and IMF staff calculations. Note: Shadow credit refers to banks’ nonloan, nonbond credit to nonfinancial private borrowers, both on and off balance sheet. For a complete definition, see footnote 33. NII = net interest income. 1Assets and liabilities available on demand or maturing in three months or less.

Figure 1.29. Bank Profitability and Liquidity Indicators

2. Small and Medium-Sized Banks: Selected Revenues and Expenses (Percent of assets)

... but would be worse without shadow-related income.

–0.75

3.00

0.00

0.75

1.50

2.25

0.9

1.2

1.5

1.8

2.0

2.3

2.6

0.1

0.6

0.2

0.3

0.4

0.5

2011 12 13 14 15 16

1.0

1.1

1.2

1.3

2011 12 13 14 15 16

3. Small and Medium-Sized Banks: Nondeposit Funding by Maturity (Percent of assets)

Growing use of risky short-term funding ...

2011 12 13 14 15 16

–0.74

Shadow- credit- related income

0.63

0.82

1.05

0.61

0.98

–0.69

1.17

0.53

0.98

–0.50

1.32

0.44

0.87

–0.31

1.42

0.34

0.76

–0.28

1.58

0.31

0.67

–0.30

1.73

2011 12 13 14 15 16

4. Banks: Short-Term Nondeposit Funding Minus Short-Term Nonloan Assets1 (Percent of assets)

... has led to worsening maturity mismatches.

–8

–6

–4

–2

2011 12 13 14 15 16

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

regulatory arbitrage should be further strengthened. Policies should target reducing balance sheet vulner- abilities at weak banks, including through restricting dividend payouts. Restructuring or resolving nonvia- ble financial institutions would also support corporate debt restructuring and strengthen risk management and governance incentives. The forthcoming IMF– World Bank Financial Sector Assessment Program report on China will discuss financial sector stability issues in China in more detail and provide specific recommendations.

Could Rising Medium-Term Vulnerabilities Derail the Global Recovery? Concerns about a continuing buildup in debt loads and overstretched asset valuations could have global economic repercussions. This section uses a scenario analysis to illustrate how a repricing of risks could lead to a rise in credit spreads and a fall in capital market and housing prices, derailing the economic recovery and undermining financial stability.

This section illustrates how shocks to individual credit and financial markets well within historical norms can propagate and lead to larger global impacts because of knock-on effects, a dearth of policy buf- fers, and extreme starting points in debt levels and asset valuations. A sudden uncoiling of compressed risk premiums, declines in asset prices, and rises in volatility would lead to a global financial downturn. With monetary policy in several advanced economies at or close to the effective lower bound, the economic consequences would be magnified by the limited scope for monetary stimulus. Indeed, monetary policy nor- malization would be stalled in its tracks and reversed in some cases.

The Global Macrofinancial Model documented in Vitek 2017 is used to assess the consequences of a continued buildup in debt and an extended rise in risky asset prices, from already elevated levels in some cases. This dynamic stochastic general equilib- rium model covers 40 economies and features exten- sive macro-financial linkages—with both bank- and capital-market-based financial intermediation—as well as diverse spillover channels.

This scenario has two phases. The first phase features a continuation of low volatility and com- pressed spreads. Equity and housing prices continue

to climb in overheated markets. As collateral values rise, bank lending conditions adjust to maintain steady loan-to-value ratios, facilitating favorable bank lending rates and more credit growth. As discussed, leverage in the nonfinancial private sector has already increased over the past decade across major advanced and emerging market economies. In the scenario, a further loosening in lending conditions, com- bined with low default rates and low volatility, leads investors to drift beyond their traditional risk limits as the search for yield intensifies despite increases in policy rates.

As presented earlier, market and credit risk premi- ums are close to decade-low levels—leaving markets exposed to a decompression of risk premiums. Thus, the second phase begins with a rapid decompression of credit spreads and declines of up to 15 and 9 per- cent in equity and house prices, respectively, starting at the beginning of 2020. This shift reflects debt lev- els breaching critical thresholds, prompting markets to grow concerned about debt sustainability, while risk premiums jump, aggravating deleveraging pres- sures. As risk premiums rise, debt servicing pressures are revealed as high debt-to-income ratios make bor- rowers more vulnerable to shocks. The asset repricing is moderate in magnitude, but is broad-based across jurisdictions and leads to a tightening of financial conditions. Flight to quality flows reduce long-term bond yields in safe havens and raise them in the rest of the world. Segments with higher leverage and extended valuations are hit particularly hard, leading to higher funding costs and debt servicing strains.

Underlying vulnerabilities are exposed, and the global recovery is interrupted. Figure 1.30 summarizes the main impacts and spillovers: • The global economic impact of this scenario is

broad-based and significant, about one-third as severe as the global financial crisis.39 The level of global output falls by 1.7 percent by 2022 relative to the WEO baseline, with varying cross-country impacts.

• The severity of the economic impact on the United States is cushioned by stronger bank buffers, milder house price declines, and more monetary policy

39The results are broadly consistent with Chapter 2, which finds that increases in household debt from already elevated levels signal high economic risks, and with Chapter 3, which concludes that rising private sector leverage signals higher downside risks to growth over the medium term.

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

Euro area United States ChinaOther advanced economies Other emerging market economies

Mortgage debt: median country Corporate debt: median country

High Medium Low High Medium Low

1. Price of Equity

–15

–10

Pe rc

en t

Pe rc

en ta

ge p

oi nt

Pe rc

en t

Pe rc

en t

–5

2017 18 19 20 21 22

3. Nominal Policy Interest Rate

–2

–1

5. Output Losses 6. Reductions in Bank Capital Ratios

4. Mortgage and Corporate Debt

2. Price of Housing

–6

–4

–2

–2.0

–0.5

–1.0

–1.5

0.0

0.5

1.0

1.5

Figure 1.30. Global Financial Dislocation Scenario

Financial stability risks build up for two more years, as equity and house prices continue to rise amid low volatility and narrow spreads, followed by an eventual sharp repricing.

Monetary policy responses are limited by policy space in some countries. A decompression of risk premiums leads to an abrupt deleveraging.

Output losses are broad-based. Rising defaults reduce capital at banks.

Source: IMF staff estimates. Note: The variables in all panels are expressed as deviations from baseline. In panel 5, countries are shaded according to the following magnitudes of output losses: (1) smaller than 1.8 percent of GDP (“low impact”), (2) between 1.8 percent and 2.3 percent of GDP (“medium impact”), and (3) greater than 2.3 percent of GDP (“high impact”). In panel 6, the thresholds for reductions in bank capital ratios are (1) smaller than 0.625 percentage points (“low impact”), (2) between 0.625 and 0.675 percentage points (“medium impact”), and (3) greater than 0.675 percentage points (“high impact”).

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

space compared with other advanced economies, despite relatively high equity valuations. The Federal Reserve reverses interest rate hikes during the second phase of the scenario, cutting the policy rate by 150 basis points to 1.75 percent by 2022.

• The euro area suffers a larger output loss because the policy rate is at the effective lower bound and—as a result of renewed financial fragmentation—term premiums rise in high-spread euro area economies. Government debt ratios climb because nominal output is lower and debt service costs are higher for these economies.

• Emerging market economies are disproportionately affected by the correction in global risk assets. The flight to quality prompts outflows from their equity and bond markets, putting pressure on curren- cies and challenging countries with large external financing needs.

• Corporate and household defaults rise on the back of higher interest costs, lower earnings, and weaker growth. Default rates do not breach global financial crisis levels but return to levels consistent with prior cyclical peaks. Firms in some euro area countries and China with excessive debt overhangs are more sensitive to the increase in credit costs. Household leverage and high house prices in Australia and Canada make these economies more susceptible to risk premium shocks.

• Higher credit and trading losses, in turn, reduce bank capital ratios to varying degrees worldwide. Banking systems in advanced economies are health- ier compared with the precrisis period, while lever- age is less of a potential amplifier. Chinese banks suffer outsize declines in capital, but strong policy buffers could be used to mitigate the financial and economic impacts.

Emerging Markets Would Suffer a Retrenchment in Foreign Capital Inflows

Drawing on the above scenario, the potential for emerging market stress due to pressures on portfo- lio inflows is examined in more detail, including by taking into account the likely reduction in these flows from Federal Reserve balance sheet normalization (as discussed earlier). • During the first phase of the scenario, portfolio

flows to emerging market economies are supported by rising investor risk appetite. This partially offsets the drag on portfolio inflows from US monetary

policy normalization observed during 2017–19. As a result, there is a (net) reduction in portfolio inflows to emerging market economies of about $25 billion a year, compared with $35 billion under the baseline (Figure 1.31, panel 1).

• During the second phase of the scenario, the asset market correction triggers a more rapid retrench- ment in capital inflows to emerging market econ- omies of about $65 billion over the first four quarters, in addition to the projected reduction of $35 billion in inflows associated with continued Federal Reserve balance sheet normalization. The combined effect results in a reduction of portfolio inflows of some $100 billion during the first four quarters of the correction (and about $65 billion during the subsequent four quarters).

• At the country level, the associated portfolio inflow reduction during the first two years of the shock to global risk premiums ranges from 1.6 to 2.3 percent of GDP for the most affected countries (Fig- ure 1.31, panel 2). Such a reduction is likely to lead to an outright reversal of portfolio flows, at least during some quarters, considering that the decom- pression of risk premiums is likely to be more rapid in some periods than in others (rather than unfold- ing at a steady pace as depicted in this exercise).

The buildup in external financing pressures could be particularly challenging for countries with large and rising projected current account deficits. For example, Colombia, South Africa, and Turkey have projected current account deficits in the range of 3 to 4½ per- cent of GDP in 2019 (Figure 1.31, panel 3). More- over, emerging market currencies would come under pressure, limiting space for monetary policy to ease. In turn, higher domestic interest rates would affect firms’ debt servicing capacity, hitting those with still high lev- els of corporate leverage and increasing risks to weaker banking systems (as explored in the April 2017 GFSR) (Figure 1.31, panel 4).

Emerging Market Policies

In emerging market economies, policymakers should take advantage of current favorable external conditions to further enhance their resilience, includ- ing by continuing to strengthen external positions where needed and reduce corporate leverage where it is high. Deploying policy buffers and exchange rate flexibility would help buffer external shocks, while

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

improving corporate debt-restructuring mechanisms and monitoring firms’ foreign exchange exposures would lower corporate vulnerabilities. Advances in these areas would leave these economies better placed to cushion any reduction in capital inflows that may occur from monetary policy normalization in advanced economies.

However, capital outflow pressures could become more significant if there is a severe retrenchment in global risk appetite, as in the scenario described earlier.

Such pressures should usually be handled primarily with macroeconomic, structural, and financial policies, although the appropriate response will differ across countries depending on available policy space (see IMF 2012, 2015, 2016a). Where appropriate, exchange rate flexibility should be a key shock absorber, but in countries with sufficient international reserves, foreign exchange intervention can be useful to prevent disor- derly market conditions. In periods of stress, liquidity provision may also be needed to support the orderly

25th percentile 50th percentile 75th percentile Peak Current: below peak

2013 2017 2019

Baseline ScenarioAdditional impact under scenario

Impact under baseline

Net impact under scenario

1. Estimated Cumulative Reduction in Emerging Market Portfolio Flows (Billions of US dollars)

US monetary normalization and a global asset market correction would increase capital outflow pressures.

Figure 1.31. Emerging Market Economy External Vulnerabilities and Corporate Leverage

2. Estimated Peak Reduction in Emerging Market Portfolio Flows (Percent of GDP, reduction over four quarters)

Countries that previously received large inflows may see sizable outflows.

1.0

4.0

3.0 3.5

2.0 1.5

2.5

4.5 5.0 5.5 6.0

3. Current Account Balances (Percent of GDP)

Outflows could be challenging for countries with large current account deficits ...

4. Corporate Leverage (Total debt to EBITDA, multiple)

... and put pressure on those with high levels of corporate leverage.

ARG

IND

MEX

ZAF

TUR

ARG

BRA

CHN

IDN RUS

BRA CHN

IDN

RUS

–7 –6 –5 –4 –3 –2 –1

0 1 2 3 4

Tu rk

So ut

h Af

ric a

Ch ile

Co lo

m bi

In do

ne si

Br az

M ex

ic o

In di

Po la

Ru ss

Ch in

M al

ay si

–300

–200

–100

100

2017 18 19 20 21 22 –2.5

–2.0

–1.5

–1.0

–0.5

0.0

Ch in

In di

Br az

In do

ne si

Tu rk

Po la

Co lo

m bi

M al

ay si

Ch ile

M ex

ic o

So ut

h Af

ric a

Sources: Bank of America Merrill Lynch; Bloomberg Finance L.P.; Capital IQ; Haver Analytics; and IMF staff calculations. Note: Data labels in the figure use International Organization for Standardization (ISO) country codes. EBITDA = earnings before interest, taxes, depreciation, and amortization.

2003 04 05 06 07 08 09 10 11 12 13 14 15 16

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

functioning of financial markets. Capital flow manage- ment measures should be implemented only in crisis situations, or when a crisis is considered imminent, and should not substitute for any needed macroeco-

nomic adjustment. When circumstances warrant the use of such measures on outflows, they should be transparent, temporary, and nondiscriminatory and should be lifted once crisis conditions abate.

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

Prolonged monetary accommodation—and a continuing need to sustain economic momentum— has contributed to a widening divergence between financial and economic cycles. Rapid inflation of asset prices has ensued as large output gaps necessitate an unusually protracted period of low interest rates. This asset price growth has been accompanied by gather- ing strength in credit growth and rising leverage, the combination of which has facilitated strong financial expansion across several economies. Such financial expansions have generally been accompanied by less remarkable economic recoveries, leading to only slowly dissipating negative output gaps. This divergence creates a challenge for monetary and financial policies to support economic recovery while ensuring that medium-term risks do not build. • In the United States, a maturing financial cycle

expansion has combined with a slowly closing output gap. The combined growth of asset prices (equity, bond, property) since the recent recession has seen one of the longest and largest cyclical expansions since 1970, albeit from a relatively weak starting point (Figure 1.1.1, panel 1). This growth across asset markets has only moderated a little from its peaks, while credit growth has been gathering momentum.

This box was prepared by Paul Hiebert, Yingyuan Chen, and Yves Schüler (Deutsche Bundesbank).

At the same time, an unusually large negative output gap has been slow to close, suggesting a need for complementary macroeconomic and financial sector policies to support the economic recovery while attenuating the financial cycle upswing as needed.

• In the euro area, the divergence between financial and economic cycles is also growing. A strong asset price boom is only slightly off recent peaks, while credit growth is slowly recovering (Figure 1.1.1, panel 2). This contrasts with a persistently large negative output gap—also suggesting a need for continued accommodative macroeconomic policies and tighter financial sector policies, as warranted in particular euro area member countries.

• The financial cycle in Japan, in contrast, has been more muted in tandem with a weak economic recovery, while asset price inflation has been volatile and oscillating around long-term trends in recent years. Recently, however, stronger credit growth has emerged along with a narrowing of the negative output gap.

• In other economies where debt service ratios for the private nonfinancial sectors have risen to high levels—such as Australia, Brazil, Canada, China, and Korea—there is a particularly strong need for financial sector policy vigilance to guard against any further buildup of imbalances.

Box 1.1. A Widening Divergence between Financial and Economic Cycles

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

–8

–6

–4

–2

–8

–6

–4

–2

0 4 8 12 16 20 24 28 32 36 40 –0.04

–0.03

–0.02

–0.01

0.00

0.01

0.02

0.03

0.04

0 4 8 12 16 20 24 28 32 36 40

2009 11 13 15 17:Q12009 11 13 15 17:Q1

–0.10 –0.08 –0.06 –0.04 –0.02

0.00 0.02 0.04 0.06 0.08 0.10

–0.04

–0.02

0.00

0.02

0.04

0.06

0.08

0.10

–0.04

–0.02

0.00

0.02

0.04

0.06

0.08

0.10

0 4 8 12 16 20 24 28 32 36 40

US Cycles: Current versus Historical since 1970

Asset and Credit Cycles and Output Gap

United States Euro area Japan

2009 11 13 15 17:Q1

Maximum of past cycles Minimum of past cycles Current cycle

Quarters from start of cycle

1. Asset Cycle (Quarterly index, deviation of filtered real growth from its historical average)

2. Credit Cycle (Quarterly index, deviation of filtered real growth from its historical average)

3. Output Gap (Percentage points)

4. Asset Cycle (Quarterly index, deviation of filtered real growth from its historical average)

5. Credit Cycle (Quarterly index, deviation of filtered real growth from its historical average)

6. Output Gap (Percentage points)

The US financial expansion and output gap are noteworthy by historical standards ...

... as a cumulative gap grows between financial and economic cycles across major advanced economies.

Sources: Bank for International Settlements; IMF, World Economic Outlook database; national sources; and IMF staff estimates. Note: Cycles are dated using National Bureau of Economic Research recession dates. Cycles capture low-frequency movements around long-term rates. Real asset price cycles combine momentum common to equity, corporate bond, and house price indices—deflated using national consumer price indices. The credit cycle is real total nonfinancial sector credit. For more information on the underlying methodology, see Schüler, Hiebert, and Peltonen 2017.

Figure 1.1.1. Financial and Economic Cycles

Box 1.1 (continued)

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

Cyberthreats to financial institutions are growing, and events in 2016 and 2017 have altered the threat landscape substantially. There has been a sizable increase in the impact and sophistication of financially motivated cyberattacks on financial institutions.1 Cyberthreats can be related to financial gain— including malware attacks—or can aim to destroy information technology systems. Some estimates place the economic losses of a hypothetical major global cyberattack as high as $53 billion (Lloyds 2017). While the magnitude and frequency of attacks have grown, their nature has evolved as perpetrators have adopted operational models that replicate legitimate businesses, such as the use of vertically integrated software packages and cloud-based operations. This evolution renders the technology both more potent and easier to access. Moreover, because cyberthreats are international and can become systemic, private sector institutions are not well positioned to respond effec- tively on their own. A coordinated regulatory approach is needed, which would result in a consistent risk mitigation framework to support financial stability.

The systemic risk ramifications of a cyberattack could be substantial. There are several channels through which cybersecurity events could threaten financial stability: (1) data breach, (2) disruption of business, (3) integrity attack (modifications to internal data), and (4) malicious activities (financial gain). Greater reliance on technology, combined with the interconnection of the global financial system, means that many, if not all, participants in the system are at risk. Banks and financial market infrastructures, in particular, harbor the potential for contagious cyberrisk, given their interconnection—so that attacks on individual financial institutions can quickly fan out across national financial systems and beyond. A recent example concerns the June 2017 “NotPetya” attack, disguised as ransomware, which among others severely hit bank operations in Ukraine. Information technol- ogy systems in the country, including automatic teller machines, were rendered unusable. Problems spilled across borders2 at a total global cost of some $850 mil- lion. Other interconnected financial institutions, such as financial infrastructures (for example, payment,

This box was prepared by Tamas Gaidosch and Chris Wilson. 1For example, the number of stolen identities rose 95 percent

year over year in 2016, according to Symantec. 2For example, two multinational companies estimated losses

from NotPetya exceeding $130 million each.

clearing, and settlement systems), are also at risk. Insurance companies are less exposed through connect- edness; however, their indirect exposure through their cyberinsurance risk underwriting can be significant and is not fully understood.3

A global and coordinated policy response is needed to ensure resilience to cyberattacks and combat cybercrime. Regulators have begun introducing cybersecurity regulations. Among recent initiatives, the European Parliament—following up on the EU-wide Cybersecurity Strategy—adopted the directive on secu- rity of network and information systems; the European Banking Authority issued guidelines on information and communications technology risk assessment; the Bank of England launched a vulnerability testing framework and set out a supervisory statement on cyberinsurance underwriting risk; the Board of Governors of the Federal Reserve System, the Office of the Comptroller of the Currency, and the Federal Deposit Insurance Corporation jointly published a notice of proposed rulemaking regarding enhanced cyberrisk management standards; the Committee on Payments and Market Infrastructures and the Board of the International Organization of Securities Com- missions issued cyberguidance for financial market infrastructures; and the New York State Department of Financial Services issued Cybersecurity Requirements for Financial Services Companies. The EU-wide Gen- eral Data Protection Regulation, effective May 2018, although not specific to the financial sector, will never- theless have a significant global impact on the system, given its extraterritorial applicability and potentially drastic fines for data breaches.4 While regulations converge on common themes, their sectoral applica- bility and level of detail vary, which presents compli- ance difficulties for international operations. Tackling cybercrime effectively means attacking its business model. The risks of being engaged in cybercrime must be raised significantly, underpinned by stronger inter- national coordination.

Beyond ensuring resilience, regulation has increas- ingly focused on prevention. Frameworks are being designed for the identification and prevention of cyberincidents, as well as for timely recovery and information sharing. Ongoing initiatives by financial

3As evidenced by the recent supervisory statement of the Bank of England on cyberinsurance underwriting risk.

4Fines can be up to 4 percent of yearly turnover or €20 mil- lion, whichever is greater.

Box 1.2. Cyberthreats as a Financial Stability Risk

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

regulators typically include practical countermeasures such as requirements on penetration and resilience tests (for example, testing how far into an organiza- tion’s system hackers can go and how well the system defends itself and recovers). As these regulations take hold, harmonization of minimum standards is needed to help smooth implementation, especially for institu-

tions operating across borders and sectors. More inter- national coordination would be helpful to share good practice, identify emerging risks, and raise standards across the entire global system—including, as needed, broader cross-border cooperation and information sharing with intelligence and other agencies outside the financial sector, among others.

Box 1.2 (continued)

C H A P T E R 1 I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

References Adrian, Tobias, and Markus Brunnermeier. 2016. “CoVaR.”

American Economic Review 106 (7): 1705–41. Adrian, Tobias, Richard Crump, and Emanuel Moench. 2013.

“Pricing the Term Structure with Linear Regressions.” Staff Report 340 (revised), Federal Reserve Bank of New York.

Adrian, Tobias, Michael Fleming, Shachar Or, and Erik Vogt. 2017. “Market Liquidity after the Financial Cri- sis.” Staff Report 796 (revised), Federal Reserve Bank of New York, June.

Adrian, Tobias, and Hyun Song Shin. 2013. “Procyclical Lever- age and Value-at-Risk.” Staff Report 338 (revised), Federal Reserve Bank of New York.

Arslanalp, Serkan and Takahiro Tsuda. 2012. “Tracking Global Demand for Advanced Economy Sovereign Debt”, IMF Working Paper 12/284, Updated, International Monetary Fund, Washington, DC.

Bank for International Settlements (BIS). 2012. “Do Debt Service Costs Affect Macroeconomic and Financial Stability?” BIS Quarterly Review (September): 21–35.

———. 2017. Annual Report: The Global Economy: Maturing Recoveries, Turning Financial Cycles? Basel.

Bank of England. 2017. “Changing Risks and the Search for Yield on Solvency II Capital.” Speech by David Rule, July 6.

Bank Wealth Management Registration and Trusteeship Center. 2017. China Bank Wealth Management Market Annual Report 2016 (in Chinese). http:// www .chinawealth .com .cn/ resource/ 830/ 846/ 863/ 51198/ 52005/ 961636/ 1495184467803885769503 .pdf.

Basel Committee on Banking Supervision. 2014. “The G-SIB Assessment Methodology—Score Calculation.” Bank for International Settlements, Basel.

Bonis, Brian, Jane Ihrig, and Min Wei. 2017. “The Effect of the Federal Reserve’s Securities Holdings on Longer-Term Interest Rates.” FEDS Notes, Board of Governors of the Federal Reserve System. https:// www .federalreserve .gov/ econres/ notes / feds -notes/ effect -of -the -federal -reserves -securities -holdings -on -longer -term -interest -rates -20170420 .htm.

Brunnermeier, Markus, and Yuliy Sannikov. 2014. “A Macro- economic Model with a Financial Sector.” American Economic Review 104 (2): 379–421.

Caruana, Jaime. 2017. “Have We Passed ‘Peak Finance’?” Inter- national Center for Monetary and Banking Studies, Bank for International Settlements, Basel.

Cetorelli, Nicola, and Linda Goldberg. 2012. “Banking Global- ization and Monetary Transmission.” Journal of Finance 67 (5): 1811–43.

Chandumont, M. L. 2016. “The Volatility Regime.” Actuary 13 (1): 36–43.

Dattels, Peter, Rebecca McCaughrin, Ken Miyajima, and Jaume Puig. 2010. “Can You Map Global Financial Stability?” IMF Working Paper 10/145, International Monetary Fund, Washington, DC.

Drehman, M., M. Juselius, and A. Korinek. 2017. “Account- ing for Debt Service: The Painful Legacy of Credit Booms.” BIS Working Paper 645, Bank for International Settle- ments, Basel.

European Central Bank (ECB). 2017. “Box 2: Financial Vulnerabilities of Euro Area Households.” Financial Stability Review (May).

Fiechter, Jonathan, Inci Otker-Robe, Anna Ilyina, Michael Hsu, Andre Santos, and Jay Surti. 2011. “Subsidiaries or Branches: Does One Size Fit All?” IMF Staff Discussion Note 11/04, International Monetary Fund, Washington, DC.

Financial Stability Board (FSB). 2017. “Policy Recommenda- tions to Address Structural Vulnerabilities from Asset Man- agement Activities.” January 12.

Geanakoplos, John. 2010. “The Leverage Cycle.” NBER Macro- economics Annual (24).

Guscina, A., G. Pedras, and G. Presciuttini. 2014. “First-Time International Bond Issuance—New Opportunities and Emerging Risks.” IMF Working Paper 14/127, International Monetary Fund, Washington, DC.

International Monetary Fund (IMF). 2012. “The Liberalization and Management of Capital Flows: An Institutional View.” Washington, DC.

———. 2015. “Managing Capital Outflows: Further Opera- tional Considerations.” IMF Policy Paper, Washington, DC.

———. 2016a. “Capital Flows—Review of Experience with the Institutional View.” IMF Policy Paper, Washington, DC.

———. 2016b. “China—2016 Article IV Consultation.” IMF Country Report 16/270, Washington, DC.

———. 2016c. “China—Selected Issues.” IMF Country Report 17/271, Washington, DC.

———. 2017a. “Macroeconomic Developments and Pros- pects in Low-Income Developing Countries.” Policy Paper, Washington, DC.

———. 2017b. “The Medium-Term Debt Management Strategy: An Assessment of Recent Capacity Building.” Policy Paper 072817, Washington, DC.

———. 2017c. “Euro Area Policies—2017 Article IV Consulta- tion.” IMF Country Report 17/235, Washington, DC.

———. 2017d. “Japan—2017 Article IV Consultation.” IMF Country Report 17/242, Washington, DC.

———. 2017e. “Japan—Financial System Stability Assessment.” IMF Country Report 17/244, Washington, DC.

———. 2017f. “China—2017 Article IV Consultation.” IMF Country Report 17/247, Washington, DC.

International Organization of Securities Commissions (IOSCO). 2017. “Consultation on CIS Liquidity Risk Management Recommendations.” July.

Koepke, Robin. 2014. “Fed Policy Expectations and Portfolio Flows to Emerging Markets.” Working Paper, Institute of International Finance, Washington, DC.

Laeven, Luc, and Fabian Valencia. 2012. “Systemic Banking Crises Database: An Update.” IMF Working Paper 12/163, International Monetary Fund, Washington, DC.

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

Lloyds. 2017. “Counting the Cost: Cyber Exposure Decoded.” Emerging Risks Report 2017. https:// www .lloyds .com /~/ media/ files/ news -and -insight/ risk -insight/ 2017/ cyence / emerging -risk -report -2017 ---counting -the -cost .pdf.

McCauley, Robert Neil, Agustín Bénétrix, Patrick McGuire, and Goetz von Peter. 2017. “Financial Deglobalisation in Banking?” Working Paper 650, Bank for International Settlements, Basel.

Mian, Atif, and Amir Sufi. 2011. “House Prices, Home Equity-Based Borrowing and the US Household Leverage Crisis.” American Economic Review 101: 2132–56.

Reinhardt, Dennis, and Steven Riddiough. 2015. “The Two Faces of Cross-Border Banking Flows.” IMF Economic Review 63 (4): 751–91.

Schularick, Moritz, and Alan Taylor. 2012. “Credit Booms Gone Bust: Monetary Policy, Leverage Cycles, and Financial Crises, 1870–2008.” American Economic Review 102 (2): 1029–61.

Schüler, Yves S., Paul P. Hiebert, and Tuomas Peltonen. 2017. “Coherent Financial Cycles for G7 Countries: Why Extend- ing Credit Can Be an Asset.” Working Paper 43, European Systemic Risk Board, Frankfurt am Main.

UK Financial Conduct Authority (FCA). 2017. “Illiquid Assets and Open-Ended Investment Funds.” Discussion P DP 17/1. February.

US Securities and Exchange Commission (SEC). 2016. “Investment Company Risk Management Program Rules.” October 13.

Vitek, Francis. 2017. “Policy, Risk and Spillover Analysis in the World Economy: A Panel Dynamic Stochastic General Equi- librium Approach.” IMF Working Paper 17/89, International Monetary Fund, Washington, DC.

Wright, Jonathan. 2011. “Term Premia and Inflation Uncer- tainty: Empirical Evidence from an International Panel Dataset.” American Economic Review 101 (4): 1514–34.

Summary

A lthough finance is generally believed to contribute to long-term economic growth, recent studies have shown that the growth benefits start declining when aggregate leverage is high. At business cycle frequen- cies, new empirical studies—as well as the recent experience from the global financial crisis—have shown that increases in private sector credit, including household debt, may raise the likelihood of a financial

crisis and could lead to lower growth. Globally, household debt has continued to grow in the past decade. This chapter takes a comprehensive look

at the relationship between household debt, growth, and financial stability across a sample of 80 advanced and emerging market economies. Besides aggregate macro-level analysis, the chapter also delves into micro-level data on individual household borrowing to shed additional light on how household indebtedness affects growth and stability at the aggregate level.

The chapter finds that there is a trade-off between the short-term benefits of rising household debt to growth and its medium-term costs to macroeconomic and financial stability. In the short term, an increase in the house- hold debt-to-GDP ratio is typically associated with higher economic growth and lower unemployment, but the effects are reversed in three to five years. Moreover, higher growth in household debt is associated with a greater probability of banking crises. These adverse effects are stronger when household debt is higher and are therefore more pronounced for advanced than for emerging market economies, where household debt and credit market participation are lower.

However, country characteristics and institutions can mitigate the risks associated with rising household debt. Even in countries where household debt is high, the growth-stability trade-off can be significantly mitigated through a combination of sound institutions, regulations, and policies. For example, better financial regulation and supervision, less dependence on external financing, flexible exchange rates, and lower income inequality would attenuate the impact of rising household debt on risks to growth.

Overall, policymakers should carefully balance the benefits and risks of household debt over various time hori- zons while harnessing the benefits of financial inclusion and development.

HOUSEHOLD DEBT AND FINANCIAL STABILIT Y2CHAPTER

53International Monetary Fund | October 2017

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

Introduction Considerable attention has been paid to household

debt since the global financial crisis as it has continued to grow in a wide range of countries (Figure 2.1). The median household debt-to-GDP ratio among emerging market economies increased from 15 percent in 2008 to 21 percent in 2016, and among advanced economies it increased from 52 percent to 63 percent over the same period. At the same time, in the highest quartile, the household debt-to-GDP ratio fell only slightly from 88 percent to 86 percent in advanced economies and continued to rise from 28 percent to 32 percent in emerging market economies. While this increase reflects to some extent the intended effects of expansionary monetary policy, central banks in various advanced and emerging market economies have recently warned against the financial stability risks of high household debt and high debt-to-income ratios when inflation and wage growth are low (see, for example, Reserve Bank of Australia 2017, Bank of Canada 2017, Bank of England 2017, South African Reserve Bank 2017, and Banco Central de Chile 2017).

Household debt and access to credit can help boost demand and build personal wealth, but high indebt- edness can also be a source of financial vulnerability. According to the permanent income hypothesis, higher debt indicates higher expected income. It also allows households to make large investments in housing and education and helps smooth consumption over time. In other words, debt allows households to acquire goods and services now and repay gradually, through higher (anticipated) income. In the long term, higher private sector credit supports economic growth (Beck, Levine, and Loayza 2000) although the precise link between growth and household debt is more elusive (Beck and others 2012). Nonetheless, even if positive in the long term, high household indebtedness can cause significant debt overhang problems when a country unexpectedly faces extreme negative shocks. The experience of the global financial crisis suggests that high household debt can be a source of financial vulnerability and lead to prolonged recessions (Mian and Sufi 2011). Broader cross-country studies also indicate that increases in

The authors of this chapter are Nico Valckx (team leader), Adrian Alter, Alan Xiaochen Feng, and Xinze Yao, with contribu- tions from Machiko Narita, Feng Li, and Xiaomeng Lu, under the general guidance of Claudio Raddatz and Dong He. Atif Mian was a consultant for this chapter. Claudia Cohen and Breanne Rajkumar provided editorial assistance.

household debt may predict lower future income growth and financial crises in the medium term (Mian, Sufi, and Verner, forthcoming; Jordà, Schularick, and Taylor 2016). As household borrowing increases the economy grows quickly in the short term but becomes highly leveraged. In this situation, a macroeconomic shock may increase unemployment and reduce output in the medium term because of financial disruptions or nominal rigidities (for example, downward wage rigidity, a zero lower bound on interest rates, or fixed exchange rates) that may prevent full adjustment to the shock.

The macroeconomic and financial risks arising from increasing household debt may not be equally important across countries at different stages of development and with different financial and institutional characteristics. Emerging market economies may be less prepared to deal with the consequences of a household deleveraging pro- cess because of limited institutional capacity. For exam-

100

125

1980 85 90 95 2000 05 10 15

10th–90th percentile 25th–75th percentile Median

1995 99 2003 07 11 15

Source: IMF staff calculations. Note: Panels show the cross-country dispersion of household debt-to-GDP ratios. See Annex 2.1 for sample coverage.

Figure 2.1. Household Debt-to-GDP Ratio in Advanced and Emerging Market Economies (Percent)

1. Advanced Economies

2. Emerging Market Economies

C H A P T E R 2 h O u S E h O L d d E B T A N d F I N A N C I A L S T A B I L I T Y

International Monetary Fund | October 2017

ple, lack of effective personal bankruptcy regimes may prevent households and lenders from efficiently dealing with debt overhang. On the other hand, household debt is lower in emerging market economies than in advanced economies reflecting a higher prevalence of financial fric- tions that reduce households’ access to debt. The balance between more financially and institutionally developed economies’ ability to deal with the consequences of higher household debt and the higher debt resulting from those very characteristics will likely determine the effect of household debt on economic growth and financial stability immediately and over the medium term.

This chapter takes a comprehensive look at the relationship between household debt, macroeconomic performance, and financial stability across a broad sam- ple of countries. It largely abstracts from the long-term considerations related to financial inclusion and financial access and focuses instead on the short- to medium-term consequences of household debt increases. It does so using a larger sample of advanced and emerging market economies than hitherto investigated to shed new light on the conditions under which household debt increases are more likely to predict subpar macroeconomic perfor- mance, large economic downturns, and financial crises.1 Furthermore, it also explores micro-level data based on national surveys for selected countries to document a series of stylized facts and the underlying mechanisms behind the aggregate results. Specifically, the chapter aims to answer the following questions: • How strongly is household debt aligned with future

GDP growth and consumption? Does the pattern differ between advanced and emerging market economies? Does the relationship depend on the institutional context, such as the terms of household debt contracts and various institutional factors?

• At the individual household level, what role do income differences play in household borrowing and consump- tion decisions? Is the household debt-to-income ratio very different across income groups and countries?

• How strongly is an increase in household debt asso- ciated with the probability of financial crises? Does household debt represent a neglected crash risk?

• What are the implications for macroprudential and other policies?

1See Chapter 3 of the April 2012 World Economic Outlook for an earlier analysis of household debt, Chapter 3 of the April 2011 Global Financial Stability Report for an analysis of housing finance and financial stability, and the October 2016 Fiscal Monitor for an analysis of private versus public sector debt.

The main findings are as follows: • On average, an increase in household debt boosts

growth in the short term but may give rise to macro- economic and financial stability risks in the medium term. Real GDP initially reacts positively to increases in household debt, as do consumption, employ- ment, and house and bank equity prices. However, after one or two years, the dynamic relationship between debt, GDP, consumption, employment, housing, and bank equity prices turns negative. Higher household debt is associated with a greater probability of a banking crisis, especially when debt is already high, and with greater risk of declines in bank equity prices.

• But the negative medium-term consequences of increases in household debt are more pronounced for advanced than for emerging market economies. In the latter, the short-term positive relationships between household debt and GDP growth, consumption, and employ- ment are stronger and the negative medium-term association with these variables is weaker. These rela- tionships are explained by the lower average household debt and credit market participation in emerging mar- kets, which may mean narrower and less costly delever- aging from a macro perspective. Or it may imply less room for overborrowing at the aggregate level in coun- tries where other financial frictions constrain access to debt for a larger share of the population.

• Country characteristics and the institutional setting play an important role. These negative medium-term effects are reinforced when household debt is high in countries with more open capital accounts and fixed exchange rates, whose financial systems are less devel- oped, and where transparency and consumer financial protection regulation is absent, quality of supervision is lower, and income inequality is larger. While these characteristics are more prevalent in emerging market economies, the lower initial levels of household debt in this group compensate for their amplifying effect for the average emerging market economy in the sample. Nonetheless, these results show that the overall consequences of household debt increases may vary importantly across countries and can be benefi- cial, even at high levels of debt, when the right mix of policies and institutions is in place.

• Lower-income groups tend to be more vulnerable. Household surveys confirm that, within countries, the share of lower-income households in total debt has grown. These households typically have higher

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

debt-to-income, higher debt-service-to-income, and higher debt-to-assets ratios, which makes them more vulnerable to adverse shocks than higher-income households.

• Macroprudential tools are useful. Macroprudential tools that target credit demand, such as restrictions on loan-to-value and debt-to-income ratios, seem to help constrain the growth in household credit.

The remainder of the chapter is organized as follows: The chapter first lays out a conceptual framework for household debt and macro-financial stability. It then describes some general developments in household debt, both from a macro and a micro (disaggregated) perspective. Next, it turns to empirical analysis of financial stability risks posed by household debt and the comovement between household debt, income, and consumption for both advanced and emerging market economies. The findings of the chapter lead to ques- tions about the regulatory framework that influences household debt decisions and risk taking, which are addressed subsequently. The last section concludes and presents relevant policy implications.

How Does Household Debt Affect Macroeconomic and Financial Stability? This section discusses some of the key models and mecha- nisms through which changes in household debt affect the macroeconomy and financial stability. First, it reviews some long-term relationships between household debt and growth. Next, it discusses the permanent income theory and some alternative models that yield different effects.

Higher financial inclusion and financial development can have positive effects on long-term growth, but the relationship between household debt and long-term growth is more elusive. Extensive literature has docu- mented that financial development and the corresponding increase in private credit by both firms and households lead to higher growth (Levine 1998; Beck and Levine 2004, among others). However, the link between house- hold debt and long-term growth has been more elusive, with earlier papers arguing that the growth consequences of household debt depend on the use of borrowed resources, and more recent evidence finding a weak relationship between household debt and GDP growth.2

2For the earlier papers on the conditional relationship between some proxies of household debt and growth, see Jappelli and Pagano 1994 and De Gregorio 1996. For recent analyses that directly

More recently, Arcand, Berkes, and Panizza (2015) and Sahay and others (2015b) find that when private sector debt reaches a certain level, the positive effects on per capita growth start to decline, which they relate to the diversion of resources from productive sectors and to rising financial stability risks when the economy becomes highly leveraged (see Box 2.1 for further discussion and a direct analysis of the long-term relationship between household debt and growth).

At the business cycle frequency, the permanent income theory argues that household debt has benefi- cial effects on the macroeconomy and on financial sta- bility. Households that anticipate an increase in future income will increase their debt to smooth their con- sumption or make large investments in nonfinancial assets or education (Friedman 1957; Hall 1978).3 A smoother intertemporal consumption pattern improves household welfare and contributes to macroeconomic stability, while credit and asset markets accommo- date the financing needs of households (Uribe and Schmitt-Grohé 2017). As such, household debt also enhances financial stability.

But newer theories and empirical evidence show that the relationship between household debt and macro-financial stability can also be negative. More recent consumption and debt theories relax some of the assumptions of the permanent income model and consider the consequences of borrowing constraints, negative externalities, and behavioral biases.4 These

consider measures of household debt finding statistically insignifi- cant relationships to long-term growth, see Beck and others 2012; Angeles 2015; and Sahay and others 2015a.

3In this context, demographics and the distribution of income and debt matter. Younger households that anticipate future income growth would borrow more against their future income (Blundell, Browning, and Meghir 1994). Rajan (2010) and Kumhof, Rancière, and Winant (2015) have argued that increased income and wealth inequality led to the rapid growth of household debt in the United States and eventually to the financial crisis in 2008. Coibion and others (2017) find that, over the period 2001–12, income inequality may have indirectly operated as a screening device for banks, given that they lend less to low-income households in high-inequality regions in the United States.

4Market incompleteness may also play a role in households’ borrowing and saving decisions. Sheedy (2014) argues that financial contracts are typically not contingent on all possible future events. Because households do not have access to insurance against future risks that could affect their ability to repay debt, the bundling together of borrowing and a transfer of risk are inefficient. In the same vein, Deaton (1991), Carroll (1992), and Aiyagari (1994) argue that households may maintain a “buffer stock” of precaution- ary savings to smooth out future consumption. This suggests that debt may have a more limited role for macro-financial stability.

C H A P T E R 2 h O u S E h O L d d E B T A N d F I N A N C I A L S T A B I L I T Y

International Monetary Fund | October 2017

market imperfections may result in household debt becoming a source of vulnerability, with consequent risks for macro-financial stability. Some of the effects are illustrated in Figure 2.2. More specifically: • Borrowing constraints, leverage, and aggregate demand:

If aggregate demand determines the level of output, a contraction in demand by highly indebted households will not always be compensated for by an increase in demand by those that are less indebted, which may lead to a recession (Eggertsson and Krugman 2012; Korinek and Simsek 2016). In this type of model, adverse shocks to highly indebted households, such as a reduction in the value of collateral, trigger borrow- ing constraints that lead to a deleveraging process that may further reduce the value of collateral. The pres- ence of nominal rigidities, such as a zero lower bound for nominal interest rates or nominal wages that can-

not adjust downward, amplifies the consequences of these shocks.5 For instance, adverse shocks to house prices (or stock prices) reduce homeowners’ equity in their housing assets (or households’ net wealth, respectively). If sufficiently large, this reduction could trigger large debt defaults and impose further downward pressure on house prices (or stock prices, respectively), leading to a debt deflation spiral (Fisher 1933), as illustrated in Figure 2.2.6 This sequence

5A broad set of macroeconomic models with financial frictions predict that high leverage reduces borrowing capacity and amplifies the impact of negative macroeconomic shocks (Kiyotaki and Moore 1997; Bernanke, Gertler, and Gilchrist 1999; Brunnermeier and San- nikov 2014, among others). Although these models focus on firms instead of household debt, the mechanism applies more broadly and is incorporated into newer studies described in this section.

6Note, however, that household debt defaults can also facilitate adjustment to lower debt levels, because it increases the resources

Debt default/

bankruptcy

Housing Financial assets Other assets

Human capital

Debt Mortgages Consumer credit

Other liabilities

Financial sector

Housing/ securities market

Assets Liabilities

High debt level

Fisher’s debt- deflation: declines in asset prices

Bank capitalization is impaired, banks reduce lending

Downward price spirals due to collateral constraints

Worsened household balance sheets lead to more defaults, bankruptcies

Household Sector

Debt overhang

Real economy

Corporate investment and

employment

Household Sector

Initial effect after a negative shock hits highly indebted households (for example, income shock, credit tightening) Second-round effects

Income Expense

Labor income Capital income

Consumption Debt service Other expenses

Deleveraging reduces aggregate demand

Declines in corporate investment and private employment

Declines in household income

Households cut back consumption further due to lower income

Source: IMF staff. Note: This figure depicts the interactions between household debt, the financial sector, and the real economy. The balance sheet view (panel 1) shows assets and liabilities (debt) at the household level, whereas the cash flow view (panel 2) shows household income and expenses in the form of consumption and debt service. The two main channels through which household debt and consumption interact are deleveraging and debt overhang. Debt overhang may adversely affect aggregate demand through deleveraging or a crowding out of consumption by the debt service burden. Deleveraging can occur through forced or accelerated repayment of debt, reduction in new credit, and increased defaults or personal bankruptcies. From a legal standpoint, default follows from a situation in which assets and income are insufficient to cover debt-servicing costs, and bankruptcy from lack of sufficient assets and income to repay the debt. There may be second-round effects, such as Fisher-type debt-deflation dynamics, that may be caused by downward asset price spirals.

Figure 2.2. First- and Second-Round Effects of the Buildup of Household Debt on Financial Stability

1. Balance Sheet View 2. Cash Flow View

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

generates negative spillovers. It can cause stress to bank capital and balance sheets and thereby harm the rest of the economy and compromise financial stability. Since, when taking on debt, households do not internalize the potential impact of their decisions on aggregate demand and other households, they borrow too much from a social perspective. Hence, better outcomes could be achieved by ex ante policies that reduce the debt level, or constrain its increases (Korinek and Simsek 2016).

• Behavioral biases: Short-sighted households may strongly prefer current consumption over future consumption, or neglect crash risk. Households that value too much current consumption (hyperbolic discounting) tend to postpone saving decisions indefinitely and to contract an excessive amount of revolving debt (Laibson 1997). Overoptimism may also lead households to borrow too much, resulting, for instance, in higher credit card debt (Meier and Sprenger 2010). Consistent with the idea of overoptimism, not only among households but also among market participants, recent evidence shows that credit expansions forecast equity crashes (Baron and Xiong 2017). Households that base their expectations solely on extrapolations from past events, when house prices have been growing, may increase their borrowing during housing booms because they expect their home equity to continue growing (Fuster, Laibson, and Mendel 2010; Shiller 2005).7 Alternatively, households may neglect cer- tain low-probability risks, such as potentially large defaults on mortgages affecting AAA-rated securities exposed to these defaults (Gennaioli, Shleifer, and Vishny 2012). Or they may vary in their optimism about returns on risky assets (Geanakoplos 2010), with optimistic agents borrowing from pessimistic ones to purchase assets that serve as collateral. This process may amplify asset prices and leverage cycles and impair financial stability. Finally, tax treat- ment (interest deductibility) may also play a role in explaining a bias toward debt financing for house- holds, much as it does for firms (IMF 2016b).

households have at their disposal to cover non-debt-related expenses and maintain their consumption levels (Elul 2008). Such a financial decelerator mechanism may explain why debt overhang is more costly (as measured by consumption loss) in countries where the cost of debt default is very high.

7Cheng, Raina, and Xiong (2014) find that even real estate professionals (midlevel managers in securitized finance) had overly optimistic beliefs about house prices.

To summarize, the exact nature of the relationship between household debt and future growth and financial stability may depend on several factors. The relation- ship may be positive if agents behave in a rational, forward-looking manner and contract debt solely with an eye on future income growth and returns to capital in the absence of financial frictions and binding bor- rowing constraints. However, the relationship between household debt and macro-financial stability may turn negative for the reasons described above. The negative relationship may be more likely when households borrow primarily for nonproductive purposes or experience inad- equate returns on their investment. High debt may bring about sharp adjustments in their consumption pattern— through deleveraging—and affect other parts of the economy. Depending on how well a country can absorb macro-financial stress or on the policies and institutions in place—such as the monetary stance, fiscal space, qual- ity of regulation and supervision, capital account open- ness, and the degree of foreign-currency-denominated loans—some episodes of debt overhang and deleveraging may be absorbed more easily than others, in response to exogenous shocks affecting households.

Developments in Household Debt around the World This section shows that household debt levels are higher in advanced economies than in emerging market economies and mainly comprise mortgage debt, while household debt has grown substantially in emerging market economies. Micro-level evidence indicates that lower-income households are less likely to borrow, but those that do tend to have riskier borrowing profiles.

Household debt to GDP is higher in advanced economies than in emerging market economies, but there is considerable heterogeneity within each group. On average, in 2016, the household debt-to-GDP ratio reached 63 percent in advanced economies and 21 percent in emerging market economies, reflecting differences in financial depth and inclusion across these groups of countries.8 But even in advanced economies, it ranges from about 30 percent of GDP in Latvia to more than 100 percent of GDP in Australia, Cyprus, Denmark, Switzerland, and the Netherlands (Figure 2.3, panel 1). In some emerging market economies, house-

8In this chapter, household debt comprises loans by households from banks and other financial institutions. In some countries, this also includes nonprofit institutions serving households.

C H A P T E R 2 h O u S E h O L d d E B T A N d F I N A N C I A L S T A B I L I T Y

International Monetary Fund | October 2017

hold debt remained very low, at less than 10 percent of GDP in 2016, in Argentina, Bangladesh, Egypt, Ghana, Pakistan, the Philippines, and Ukraine, while in others, such as Malaysia, South Africa, and Thai- land, it exceeded 50 percent of GDP. More broadly, the cross-country distribution of the household debt- to-GDP ratio is positively correlated with differences in financial development (Figure 2.3, panel 2).

Mortgage debt makes up the bulk of household debt in advanced economies, but less so in emerging market economies. It accounts for more than 50 percent of total household debt in most advanced economies, whereas among emerging market economies it captures one-third or less of total household debt (Figure 2.3, panel 3). Indeed, differences in mortgage debt explain a large fraction of the difference in household debt between emerging market and advanced economies. Although the characteristics of mortgages vary widely across countries and jurisdictions, a survey of IMF country desks finds that most mortgages are recourse loans: after a default the lender can try to seize additional household assets to cover the debt if the market value of the house is insufficient (see Annex Figure 2.1.1). Other debt consists primarily of consumer credit, which is typically used to smooth out short-term fluctuations in consumption and income but can also be used to finance microenterprises.9

Household debt has grown substantially in many countries over the past decade and has kept growing in recent years, especially among emerging market economies. Household debt-to-GDP levels fell in the United States and the United Kingdom after the global financial crisis of 2007–08 and in various European countries—most notably, Iceland, Ireland, Portugal, Spain, and the Baltics—in the wake of the European sovereign debt crisis (Figure 2.3, panel 1). In Germany, household debt has fallen as a percentage of GDP since 2000. Notwithstanding these recent declines, the level of household debt to GDP remains high by historical standards in most of these countries and has kept grow- ing in other advanced economies, such as Australia and Canada (Figure 2.3, panel 5). In a number of emerg- ing market economies—most notably Chile, China, Malaysia, Thailand, Paraguay, Poland, and some central and southeastern European countries, household debt to GDP expanded rapidly over a short time, from as low

9For instance, urban Indian households report about one-fifth of their debt to be for business-related purposes. In addition, rural households use two-fifths of their debt for productive purposes, with the highest share among the wealthier households (see Badarinza, Balasubramaniam, and Ramadorai 2016).

as 10 percent of GDP in 2005 to more than 60 per- cent of GDP in some cases. This is also reflected in the rapid rise of median household debt-to-GDP ratios in emerging market regions: from between 5 percent and 10 percent in 2000 to between 17 percent and 22 per- cent in 2016 (Figure 2.3, panels 5 and 6).

Changes in household debt ratios are driven mainly by debt increases rather than low or negative income growth. In theory, the household debt-to-GDP ratio may go up if debt increases more, or declines less, than GDP does. The rapid rise in the household debt-to-GDP ratio from 1990 to 2007 is due mainly to rapid increases in inflation-adjusted household debt, in both advanced and emerging market economies, amounting to 6.7 percent and 13.4 percent a year, respectively—far exceeding the growth of real GDP and real disposable income (Figure 2.3, panel 4). This rise was facilitated by the sharp decline in interest rates and easier and more widespread access to credit. Hence, debt servicing may not have risen that much. During this period, net wealth also rose on account of strong real house price increases. After 2008, the growth in house- hold debt slowed to 2 percent a year in advanced econ- omies, reflecting a retrenchment of households in the wake of the global financial crisis, and to 6.6 percent a year in emerging market economies. In both cases, debt continued to exceed the rate of GDP growth, leading to increases in the ratio of household debt to GDP.

The overall trend in household debt to GDP is very similar to that of the debt-to-assets ratio. For a subsam- ple of 18 Organisation for Economic Co-operation and Development countries, increases in household debt to assets are highly correlated with household debt-to-GDP ratios (Figure 2.4, panel 6). Thus, increases in debt are usually accompanied by rising leverage, meaning that a focus on net wealth may mask underlying vulnerabilities that arise from procyclical asset values. The trend is most notable for mortgage debt—which constitutes the bulk of household debt in many countries—for which there is large comovement with the housing market cycle. As a result, households are less able to tap into their housing wealth to smooth consumption after a shock. Therefore, following the recent empirical literature and without losing much generality, the rest of the empirical analysis focuses on the debt-to-GDP ratio.10

10In the ensuing analysis, using the debt-to-assets ratio instead of the debt-to-GDP ratio for a subset of 26 Organisation for Economic Co-operation and Development countries for which such data are available yields qualitatively the same results (see Figure 2.6, panel 2).

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

100

150

200

0 20 40 60 80 100 120 140

20 16

2007

EMEs AEs

120

150

H ou

se ho

ld d

eb t-

to -G

D P

ra tio

H ou

se ho

ld d

eb t-

to -G

D P

ra tio

Financial development (index)

100

150

200

Share of mortgage debt

EMEs AEs

15 G

D P

In co

m e

RH H

H H

D /G

D P

G D

In co

m e

RH H

H H

D /G

D P

1990–2007 2008–16

EMEs AEs

100

1980 85 90 95 2000 05 10 15

Euro area Other AEs CEEC

1995 2000 05 10 15

Asia Africa Middle East Latin America

Sources: Bank for International Settlements; CEIC Data Co. Ltd.; Economic Cycle Research Institute; Haver Analytics; IMF, International Financial Statistics, Monetary and Financial Statistics, and World Economic Outlook databases; Jordà-Schularick-Taylor Macrohistory Database; Svirydzenka 2016; Thomson Reuters Datastream; and IMF staff calculations. Note: For countries included in regional breakdowns, see Annex 2.1. In panel 2, financial development is the index taken from Svirydzenka 2016. Panel 4 reports median annual growth rates for each country group and period for real GDP, real disposable household income, real household debt (RHHD), and household debt-to-GDP ratio (HHD/GDP). Dashed line in panel 1 denotes the 45-degree line. AEs = advanced economies; CEEC = Central and Eastern European countries; EMEs = emerging market economies; income = real disposable household income.

Figure 2.3. Growth and Composition of Household Debt by Region (Percent)

1. Household Debt-to-GDP Ratio, 2007 and 2016 2. Household Debt-to-GDP Ratio and Financial Development, 2013

3. Household Debt-to-GDP Ratio and Mortgage Share of Debt, 2016

4. Decomposition of Annual Changes in Household Debt Ratio

5. Advanced Economies and Central and Eastern European Countries: Median Household Debt-to-GDP Ratio

6. Emerging Market Economies in Asia, Africa, the Middle East, and Latin America: Median Household Debt-to-GDP Ratio

0 0.2 0.4 0.6 0.8 1.0

0 20 40 60 80 100

C H A P T E R 2 h O u S E h O L d d E B T A N d F I N A N C I A L S T A B I L I T Y

International Monetary Fund | October 2017

Q1 Q2 Q3 Income quintile Income quintile

Q4 Q5 0

400

800

1,200

Q1 Q2 Q3 Q4 Q5

7 8 9 10 11 12

Pa rt

ic ip

at io

n ra

log (Real GDP per capita)

0 10 20 30 40 50

M or

tg ag

e pa

rt ic

ip at

io n

ra te

Overall participation rate

Q1 Q5

100

200

300

400

0 50 100 150

M ed

ia n

de bt

-t o-

in co

m e

ra tio

Household debt-to-GDP ratio

AEs EMEs

1995 2000 05 10 15

Household debt-to-GDP ratio (left scale)

Household debt-to-assets ratio (right scale)

Sources: Bank for International Settlements; country panel surveys; Euro Area Housing Finance Network; Luxembourg Wealth Study; Organisation for Economic Co-operation and Development (OECD); US Survey of Consumer Finance; and IMF staff calculations. Note: Panels 1 and 2 show the cross-country dispersion across income quintiles, evaluated at the median for mortgage borrowers (quintile 1 to quintile 5, from lowest to highest income). Dashed lines in panels 4 and 5 denote the 45-degree line. For country coverage, see Annex 2.1. Panel 6 shows debt, asset, and wealth ratios for a subsample of 18 OECD countries for which such data are available since 1995. AEs = advanced economies; EMEs = emerging market economies.

Figure 2.4. Household Debt: Evidence from Cross-Country Panel Data (Percent, unless noted otherwise)

1. Loan Participation Rate, 2010 2. Debt-to-Income Ratio, 2010

3. Loan Participation versus per Capita GDP, 2013 (X axis = US dollars purchasing power parity)

4. Mortgage Participation Rate and Overall Participation Rate, 2013

5. Median Debt-to-Income Ratio and Household Debt-to-GDP Ratio, 2013

6. Household Debt-to-GDP Ratio and Debt-to-Assets Ratio

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

Lower-income groups typically participate less in credit markets, and their credit profiles are weaker. Household survey data from 25 countries show that households in the lowest income quintiles participate much less in mortgage (and overall) credit markets (Figure 2.4, panel 1). Those that do, however, have, on average, higher risk profiles, with higher debt-to-assets and debt-to-income ratios as well as higher debt service ratios (defined as total debt repayment as a percentage of total income) (Figure 2.4, panel 2). This suggests that lower-income households are most vulnerable to cyclical fluctuations in income and are less likely to benefit from positive wealth effects, given their rela- tively low net asset holdings. From a bank’s perspective, these customers generally represent a higher credit risk, which, in turn, may explain the relatively low participa- tion rate, indicating the presence of credit constraints.

Differences in participation across countries explain part of the differences in debt ratios between advanced and emerging market economies. As with other measures of financial inclusion, household credit participation increases with economic development, as measured by real GDP per capita (Figure 2.4, panel 3).11 As credit participation increases, it initially covers mainly high-income families and then moves more aggressively toward easing access for lower-income families, as reflected by the curvature of the respective income groups’ lines (Figure 2.4, panel 4). Thus, high credit participation by low-income families is mainly an advanced economy phenomenon; lower-income countries grant access to credit mainly to higher-income households. Since not all households have debt and since debt-to-income ratios vary significantly across households, macro-level measures of household debt (such as debt-to-GDP and debt-to-net-wealth ratios) underestimate the true burden of indebted households (Figure 2.4, panel 5).12 This underestimation could be especially relevant for emerging market economies where participation rates are low and where low macro-level indebtedness may coexist with significant micro-level household indebtedness (see Box 2.2 for an analysis of Chinese households).

11See also Demirgüç-Kunt and Klapper (2012), who find that account penetration is higher in economies with higher national income, as measured by GDP per capita.

12The aggregate measures of household indebtedness correspond to an income-weighted average of individual household debt ratios. Households with no debt but positive income, as well as differences in indebtedness across households, lead to differences between aggre- gate and micro-level measures.

The dynamics of household debt are linked to the evolution of house prices. For example, household debt in Canada and the United States evolved very similarly until the global financial crisis (Box 2.3). After the crisis, household debt continued to rise in Canada but fell in the United States as house prices followed different paths: declining in the United States while continuing to appreciate in Canada. As a result, US households’ lever- age for mortgage holders, reflected in the debt-to-income ratio, remained broadly constant, while Canadian mortgage borrowers’ debt to income increased across all income groups and is now much higher than for US households. These patterns suggest that household debt and housing prices have common dynamics (Box 2.4). Similarly, in China, where house prices rose by 16 per- cent in real terms, the debt-to-income ratio increased across most income groups between 2011 and 2015, and especially for lower-income households (Box 2.2).

Financial Stability Risks of Household Debt: Empirical Analysis Increases in household debt have a positive short-term but a negative medium-term relationship to macroeconomic aggregates such as GDP growth, consumption, and employ- ment. They also predict downside risks to GDP growth and a higher probability of a banking crisis. However, the strength of the negative association depends on the level of household debt to GDP, getting stronger when this level exceeds certain thresholds. The short-term positive effects are generally stronger and the medium-term negative effects are consistently weaker for emerging market economies.

Household Debt and Growth, Consumption, and Employment

When household debt increases, future GDP growth and consumption decline and unemployment rises relative to their average values. Changes in household debt have a positive contemporaneous relationship to real GDP growth and a negative association with future real GDP growth, in line with various recent empirical studies.13 Specifically, a 5 percent increase in household debt to GDP over a three-year period forecasts a 1¼ percent decline in real GDP growth three years ahead (Figure 2.5, panel 1).14 These results do not seem to be

13See, for instance, Mian, Sufi, and Verner, forthcoming; Jordà, Schularick, and Taylor 2016; and Lombardi, Mohanty, and Shim 2017.

14The empirical model includes country fixed effects, so that all variables can be interpreted as deviations from their sample averages.

C H A P T E R 2 h O u S E h O L d d E B T A N d F I N A N C I A L S T A B I L I T Y

International Monetary Fund | October 2017

driven by potential endogeneity concerns.15 A further breakdown shows that household debt is correlated with future declines in private consumption (Fig- ure 2.5, panel 2) but less so with government consump- tion and investment. It is also negatively correlated with the current account deficit. These findings suggests that household debt booms finance consumption expan- sions, often through current account deficits that revert later when consumption and GDP growth also decline. Increases in household debt are also associated with significantly higher unemployment up to four years in the future (Figure 2.5, panel 3).

The short-term positive association between changes in household debt and GDP growth is stronger and the medium-term negative relationship weaker for emerg- ing market economies than for advanced economies (Figure 2.5, panel 1). On the other hand, consumption expands less in the short term and declines less in the medium term after household debt increases in emerg- ing market economies (Figure 2.5, panel 2), while the results for unemployment follow a similar pattern as those for GDP (Figure 2.5, panel 3). This suggests that the trade-off between the benefits of increased household participation in credit markets and the risks to macro- economic stability is less striking for these countries, most likely because of lower average household debt, although institutions and policies may also play an important role, as discussed later. Moreover, the evidence on long-term growth reviewed in Box 2.1 suggests that, in the long term, increases in household debt appear positively related to growth up to a certain level.16

Increases in household debt are associated with height- ened downside risks to future GDP growth for all coun- tries, but in emerging market economies they also predict

15Results obtained using instrumental variables yield qualitatively similar and quantitatively larger estimates than those obtained through ordinary least squares. In these estimations, changes in household and firm debt-to-GDP ratios were instrumented by the interaction between a country’s degree of capital account openness and US financial conditions and global liquidity (broad money). Micro-level regressions discussed below—which are much less likely to be affected by potential endogeneity—provide additional support for the causal interpretation of these results.

16The cumulative effect of an increase in household debt on growth, consumption, and employment, inferred from Figure 2.5, is negative in advanced economies and neutral to marginally negative in emerging market economies. However, such an exercise implicitly relates changes in household debt to longer-term growth outcomes, which is more adequately addressed in the framework reviewed in Box 2.1. According to those results, an increase in the household debt-to-GDP ratio raises long-term growth as long as the final ratio is below a threshold between 36 and 70 percent of GDP (corre- sponding to a 90 percent confidence interval).

higher upside risks. Quantile regression results show that changes in household debt have important implications for movements in the distribution of future GDP growth (Figure 2.5, panel 4). Initially, household debt is associ- ated with strong positive output growth (the right tail of the distribution), especially among emerging market economies. But three to five years ahead, increases in household debt seem to have a clearer association with below-average movements of future growth (the left tail of the distribution of future real GDP growth).17 This pattern is consistent with the deleveraging and aggregate demand externalities that arise after a period of rapid growth in household debt, resulting in a volume of borrowing above the socially optimal level that leads to important corrections after a shock. It is interesting to note that, among emerging market economies, increases in household debt are associated with worse negative and stronger positive future growth outcomes compared with advanced economies. This finding may reflect the more extreme historical experiences in this group of coun- tries; they benefit more from financial development and improved access to finance but also suffer more strongly during episodes of debt overhang and financial crises.

Supply-driven increases in household debt are more damaging to future growth. Using changes in financial conditions to identify supply- and demand-driven increases in household debt, similar to Mian, Sufi, and Verner, forthcoming, shows that the supply-driven component of household debt has a stronger impact on future GDP growth than the demand component (Figure 2.5, panel 5). Similarly, a monetary policy loosening (negative Taylor rule residuals) reinforces the negative relationship between household debt and future economic activity.

The negative medium-term association between GDP growth and growing household debt is largely absent at low levels of debt to GDP. At very low levels of house- hold debt to GDP, below 10 percent, the association between increases in debt and future real GDP growth is positive; it turns negative when household indebted- ness exceeds 30 percent of GDP (Figure 2.5, panel 6). Beyond that point, the correlation declines slightly, but it maintains its negative sign. The presence of this nonlin- earity is consistent with recent findings of a bell-shaped

17In advanced economies, an increase in household debt is neg- ative for medium-term GDP growth across the entire distribution of future GDP growth (all quantiles), whereas in emerging market economies, the impact of household debt on future GDP growth is negative only in the left tail of the distribution (when future growth is below average).

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

–0.3

–0.2

–0.1

0.1

0.0

0.2

t t + 1 t + 2 t + 3 t + 4 t + 5 t + 6

All AEs EMEs

–0.3

–0.2

–0.1

0.1

0.2

0.3

t t + 1 t + 2 t + 3 t + 4 t + 5 t + 6

All AEs EMEs

–0.1

0.1

0.2

t t + 1 t + 2 t + 3 t + 4 t + 5 t + 6

All AEs EMEs

–0.2

–0.1

0.1

0.0

0.2

0.3

0. 15

0. 50

0. 85

0. 15

0. 50

0. 85

0. 15

0. 50

0. 85

0. 15

0. 50

0. 85

0. 15

0. 50

0. 85

0. 15

0. 50

0. 85

0. 15

0. 50

0. 85

0. 15

0. 50

0. 85

0. 15

0. 50

0. 85

All AEs EMEs All AEs EMEs All AEs EMEs t t + 2

Percent

t + 4

–0.5

–0.4

–0.3

–0.2

–0.1

Joint Supply Demand

–0.5

–0.3

–0.1

0.1 0.0

0.3

0.5

<10 10 20 30 40 50 60 70 80 90 100

Sources: Bank for International Settlements; CEIC Data Co. Ltd.; Economic Cycle Research Institute; Haver Analytics; IMF, World Economic Outlook database; Jordà-Schularick-Taylor Macrohistory Database; Penn World Table; and IMF staff calculations. Note: Panels 1, 2, and 3 are from panel regressions of rolling three-year real GDP growth (consumption and unemployment, respectively) up to six years ahead, on lagged changes in household and corporate debt-to-GDP ratios (over a three-year period), controlling for lags of the dependent variable, and country and time fixed effects. Panel 4 shows quantile regression coefficient estimates for changes in the household debt ratio, using the same specification as the panel regression model. Panel 5 breaks down changes in household debt-to-GDP ratios into supply and demand factors, where local financial conditions are assumed to signal supply-side factors, and the residual to reflect other (demand) factors. Panel 6 shows coefficient estimates from a panel regression estimation, conditioning the effect on changes in household debt, and interacted with various debt thresholds. Colored bars indicate that the effects are statistically significant at the 10 percent level or higher. See Annex 2.2 for details of the estimation methodology. AEs = advanced economies; EMEs = emerging market economies.

Figure 2.5. Effects of Household Debt on GDP Growth and Consumption

1. Impact on Real GDP Growth (Regression coefficients)

2. Impact on Real Consumption Growth (Regression coefficients)

3. Impact on Unemployment (Regression coefficients)

4. Quantile Regression of Real GDP Growth (Regression coefficients, 15th, 50th, and 85th quantiles)

5. Demand and Supply Effects (Regression coefficients)

6. Real GDP Growth Threshold Effects (Regression coefficients at various household debt-to-GDP levels)

C H A P T E R 2 h O u S E h O L d d E B T A N d F I N A N C I A L S T A B I L I T Y

International Monetary Fund | October 2017

relationship between financial deepening and long-term growth (Sahay and others 2015b) and studies relating this to increased financial risks (see also Box 2.1). While the threshold above which increases in household debt more strongly signal risks to real activity is low, it is gen- erally above the levels reached by emerging markets in this sample. This finding may partly explain the milder association estimated for this group of countries.

The relationship between future GDP growth and household debt is driven mostly by mortgage debt. The finding that the mortgage debt component is statistically significant and the nonmortgage component is not (Fig- ure 2.6, panel 1) goes somewhat against the argument that increases in debt accompanied by a simultaneous accumulation of assets are less risky, because households may be able to tap into these assets when facing shocks. This could be due to the procyclicality of home equity lines or—more generally—to wealth effects that lead households to cut consumption when the value of their housing assets decline.18 Further evidence confirms that the accumulation of assets does not dampen the consequences of increased indebtedness. Changes in the household debt-to-total-assets ratio are associated with growth declines only at horizons beyond five years ahead, with increases in household debt to GDP remaining significant at shorter horizons (Figure 2.6, panel 2). These results suggest that, at business cycle fre- quencies, it is primarily households’ debt service capac- ity, approximated by a higher debt-to-GDP ratio, that signals vulnerabilities rather than their solvency position.

Similar results are found in micro-level data: high debt-to-income ratios make households more vulnerable to income shocks. Micro longitudinal data for five euro area countries show that high household indebtedness in 2010, right before the European sovereign debt crisis, caused a significant reduction in consumption between 2010 and 2014 (Figure 2.7, panel 1).19 Furthermore, consumption declined more for the most indebted

18Boom-bust cycles in housing prices that accompany increases in household debt could be driving the results reported above, but further analysis shows that lagged house price growth is not very sig- nificant in growth forecasting regressions. Additional evidence from dynamic panel vector autoregression techniques shows that house price shocks are associated with a gradual rise in household debt, whereas household debt shocks lead to significant increases in house prices in the short term, up to two to three years, but are followed by a fall in house prices afterward (Box 2.5).

19The macroeconomic and unexpected nature of the shock makes it unlikely that the results are driven by the reverse causality argu- ment that individual households borrowed preemptively to hoard liquidity and smooth consumption.

–0.3

–0.2

–0.1

0.1

0.0

Mortgage Nonmortgage

–1.0

–0.8

–1.3

–0.5

–0.3

0.3

0.5 Household debt-to-GDP ratio Debt-to-assets ratio

Sources: Bank for International Settlements; CEIC Data Co. Ltd.; Economic Cycle Research Institute; Haver Analytics; IMF, World Economic Outlook database; Jordà-Schularick-Taylor Macrohistory Database; Penn World Table; and IMF staff calculations. Note: This figure shows coefficients of household debt variables in panel regressions of real GDP growth, one to six years ahead, on lagged changes in household and corporate debt-to-GDP ratios (over a three-year period), controlling for lags of the dependent variable, and country and time fixed effects. Panel 1 splits household debt into mortgage and nonmortgage debt-to-GDP ratio. Panel 2 includes changes in the household debt-to-GDP ratio and changes in the household debt-to-assets ratio in the panel regression. Estimations are performed over subsamples for which data are available compared with analysis in Figure 2.5. Colored bars indicate that the effects are statistically significant at the 10 percent level or higher.

Figure 2.6. Effects of Household Debt on GDP Growth: Robustness Tests (Regression coefficients)

t t + 1 t + 2 t + 3 t + 4 t + 5 t + 6

1. Mortgage and Nonmortgage Debt

2. Debt-to-Assets and Household Debt-to-GDP Ratios

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

households (Figure 2.7, panel 2), which also perceived themselves to be the most financially constrained (Fig- ure 2.7, panel 3). The larger reduction in consumption by highly indebted households at the micro level and the corresponding decline in aggregate consumption observed in macro data are consistent with the effects of aggregate demand externalities arising from delever- aging. Evidence for China also shows that consumption of households with high debt-to-income ratios responds more strongly to income shocks (Figure 2.7, panel 4 and Box 2.2). Hence, highly indebted households’ higher marginal propensity to consume may amplify the effect of negative income or credit shocks on China’s econ- omy, in line with evidence in advanced economies (for

example, Mian, Rao, and Sufi 2013). Similar results are found for advanced economies, such as Australia, although they are less pronounced.

Financial Stability Risks and Neglected Crash Risk

Increases in household debt are also good early warning indicators for banking crises.20 A simple look at the data shows that increases in household debt peak about three years before the onset of a banking crisis (Figure 2.8, panel 1). Formal evidence from a logit

20Previous research documenting similar findings includes Gourin- chas and Obstfeld 2012; Drehmann and Tsatsaronis 2014; and Jordà, Schularick, and Taylor 2016.

–100

–50

100

0 500 1,000 1,500 2,000 2,500

Ch an

ge in

c on

su m

pt io

n (p

er ce

nt o

f i nc

om e)

Ch an

ge in

c on

su m

pt io

n (p

er ce

nt o

f i nc

om e)

Past debt-to-income ratio –4

–3

–2

–1

0 <1 1–3 3–5 >5

Past indebtedness (debt-to-income ratio)

0–1 1–2 2–3 3–4 4–5 5–6 >6

Debt-to-income ratio (mortgage debt)

DTI = 1 DTI = 5 DTI = 1 DTI = 5

China Australia

Sources: European Central Bank Household Finance and Consumption Survey; Household, Income and Labour Dynamics in Australia Survey; China Household Finance Survey; and IMF staff calculations. Note: Panels 1–3 present data from euro area countries with a panel dimension (Belgium, Cyprus, Germany, Malta, Netherlands). The change in consumption-to- income ratio is computed over 2010–14. For panel 4, see Boxes 2.2 and 2.4 for additional information. DTI = debt-to-income ratio. 1In panel 4, results are based on data for households tracked between 2013 and 2015 for China, and between 2006 and 2015 for Australia.

Figure 2.7. Micro-Level Evidence Corroborating the Macro Impact

1. Euro Area: Initial Debt-to-Income Ratio and Changes in Consumption, 2010–14

2. Euro Area: Drop in Consumption among Indebted Households, 2010–14 (Percent of income)

3. Homeowners Not Applying for Loans Due to Perceived Credit Constraint, 2014 (Percent)

4. China and Australia: Response of Consumption to Income Shocks1

(Percent)

C H A P T E R 2 h O u S E h O L d d E B T A N d F I N A N C I A L S T A B I L I T Y

International Monetary Fund | October 2017

panel data model shows that a rise in the household debt-to-GDP ratio contributes to a greater probability of banking crises three years ahead (Figure 2.8, panel 2). The marginal effect, at about 1 percent, is economically significant, since the unconditional crisis probability is about 3.5 percent for the countries under examination. The relationship between increasing household debt and financial crises is more pronounced when house- hold debt is high (65 percent of GDP). This is broadly consistent with the nonlinear effects found for the relationship between household debt and GDP growth, with the higher threshold resulting from the extreme nature of crises as compared with episodes of growth declines. The existence of nonlinear effects suggests that debt increases in already highly indebted households may be hard to sustain when facing a negative income shock, leading them to drastically reduce consumption and default on their debts.

Increases in the household debt ratio predict negative equity excess returns (over the risk-free rate), especially for the banking sector. Such predictability is present for both the banking sector and the overall stock market index (Figure 2.9, panel 1). This negative correlation may reflect investor overoptimism and a systematic neglect of the risk of equity crashes (so-called neglected crash risk) during periods of high growth in household debt (Figure 2.9, panel 2). Further analysis with quantile regressions shows that the negative association between increases in household debt and future equity returns is stronger in the lower tail of the return distribution than in the upper tail, confirming that investors appear to systematically neglect the risk of equity crashes. Although the neglected crash risk affects all sectors, predictability is stronger for bank stock returns, suggest- ing that rising household debt is often associated with neglected banking sector vulnerabilities.21 As discussed later in the chapter and shown earlier, these vulner- abilities may arise both from the ensuing decline in growth associated with the deleveraging process or from higher debt defaults from overindebted households. The predicted decline in overall stock market returns suggests that growth contractions explain part of these results. But consistent with a simultaneous role for

21Risk-adjusted abnormal returns of the banking sector are com- puted to measure the performance of bank stocks relative to market returns. Abnormal returns are defined as the capital asset pricing model regression residuals with quarterly data. For each country, the coefficient on market excess return, that is, the market beta, is esti- mated in each year based on past return data to avoid using future information that is unknown in that year.

–5

–4 –3 –2 –1 0 +1 +2 +3 +4 Average

Household debt Corporate debt

0.5

1.0

1.5

2.0

Change in corporate debt

Change in household debt

Change in household debt × high household

debt level

1. Increase in Household and Corporate Debt Ratios around Banking Crises (Percent)

2. Probability of a Banking Crisis: Marginal Effects (Percentage points)

Sources: Bank for International Settlements; CEIC Data Co. Ltd.; Economic Cycle Research Institute; Haver Analytics; IMF, International Financial Statistics, and Monetary and Financial Statistics databases; Jordà-Schularick-Taylor Macrohistory Database; Laeven and Valencia 2013; Thomson Reuters Datastream; and IMF staff calculations. Note: Panel 1 shows the average growth in ratios of household and nonfinancial corporate debt to GDP before and after a banking crisis, as well as the uncondi- tional average growth rate. Panel 2 shows the marginal effects of a panel logit model for banking crises for 34 countries, with country fixed effects, levels, and changes in ratios of household and nonfinancial corporate debt to GDP. It also shows the interaction effect with a high household debt dummy variable, set at 65 percent of GDP, representing the top quintile of the distribution. The effects are significant at the 10 percent confidence level. Banking crises are taken from the updated database by Laeven and Valencia (2013).

Figure 2.8. Banking Crises and the Role of Household Debt

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

rising defaults, increases in the household debt ratio are often associated with higher growth of nonperforming loans in the country’s banking sector three years later, confirming that rapid growth in household debt is asso- ciated with greater banking stress in the future.

When Is Household Debt More Likely to Predict Low GDP Growth?

The consequences of an increase in household debt for future growth differ substantially across countries. The estimated debt-to-GDP-growth relationship exhibits substantial heterogeneity within both advanced and

emerging market economies (Figure 2.10, panel 1). The median coefficient for the three-year-ahead impact of an increase in debt on GDP growth is –0.5 for advanced economies and –0.13 for emerging market economies. Within each group of countries, the dispersion of the estimated coefficients is large, although more so for emerging market economies, which also have a larger share of positive country-level coefficients. This dis- persion suggests that, in addition to the initial level of household debt documented earlier, country-specific and institutional factors may play a role in mediating the relationship between rising household debt and future economic activity. To investigate the role of various leading factors, separate panel regressions add interactions between household debt and a number of institutional and country-specific characteristics to the panel regression between changes in household debt and three-year-ahead GDP growth (Figure 2.10, panel 2).22

Having an open capital account and a fixed exchange rate regime increases the risks associated with rising household debt. An open capital account has multiple benefits for financial integration and access to foreign capital (Mussa and others 1998; Stulz 1999), but it also exposes countries experiencing large capital inflows to sudden stops (Calvo and Reinhart 2000). In this sample, a more open capital account results in a stronger negative association between increases in household debt and future GDP growth.23 This result might arise from the accumulation of foreign-currency-denominated debt, similar to findings by Mian, Sufi, and Verner (forthcom- ing). As noted in the literature, capital flows that sustain episodes of foreign debt accumulation are frequently followed by sudden stops that force strong corrections in consumption, particularly in emerging markets. This pattern is consistent with a larger differential effect of capital account openness in this group of economies. Along similar lines, having a fixed exchange rate regime reduces an economy’s flexibility to accommodate exter- nal shocks, resulting in a larger contraction in aggregate demand, especially in the presence of nominal wage rigidities (Schmitt-Grohé and Uribe 2016). Interestingly,

22Additional analysis also attempted to relate the effect of house- hold debt on banking crises documented earlier to institutional and country-specific variables, but no significant interaction effects were detected, probably because of the relatively smaller coverage, over time, and number of countries and crises observations, relative to the panel data growth regression analysis.

23In this analysis, capital account openness is measured as de jure openness. The results do not change when using de facto measures such as capital flows as a percentage of GDP.

–8

–7

–6

–5

–4

–3

–2

–1

One year ahead

Two years ahead

Three years ahead

Four years ahead

Five years ahead

One year ahead

Two years ahead

Three years ahead

Four years ahead

Five years ahead

1. Banking Sector Abnormal Returns (Regression coefficients)

2. Bank Equity Crash Risk (Marginal effects)

Source: IMF staff calculations. Note: Panel 1 shows coefficients from regressions of future bank equity risk-adjusted abnormal returns, one to five years ahead, using past three-year changes in the household debt-to-GDP ratio as independent variables. Panel 2 shows the marginal effect of the change in the household debt ratio (normalized by the standard deviation) on the probability of equity crashes in the next one to five years. Bank equity crashes are defined as annual bank equity returns lower than one standard deviation below the mean, as in Cheng, Raina, and Xiong 2014; and Baron and Xiong 2017. Solid bars mean that the response is statistically significant using 95 percent confidence intervals.

Figure 2.9. Bank Equity Returns and Household Debt

C H A P T E R 2 h O u S E h O L d d E B T A N d F I N A N C I A L S T A B I L I T Y

International Monetary Fund | October 2017

this analysis shows that it is the combination of a fixed exchange rate regime and capital account openness that magnifies the risks associated with increasing household debt. This finding is consistent with the limitations that such a regime poses for accommodating the conse- quences of large changes in capital inflows (IMF 2016a).

Financial development and the quality of bank supervision seem to mitigate the medium-term negative relationship between increases in household debt and GDP growth. Credit expansion in a more financially developed environment entails lower risks because the financial system is better able to assess credit risk and allocate credit and is better prepared to deal with their consequences. Moreover, countries where banking supervision is more stringent and capital requirements are stricter appear able to reduce the negative effect of household debt on GDP growth. The same effect is found for banking systems that have higher capital ratios or a larger distance to default. All these measures directly or indirectly reflect the quality and conservatism of the banking supervision—supervisors may stop banks from paying out high dividends to shareholders and instead require them to retain higher capital buffers, thereby limiting, to some extent, the bank lending channel.

Among institutional variables, the existence of credit registries significantly reduces the risks signaled by rising household debt. Having access to broad information on individuals’ levels of debt and payment histories (both positive and negative) reduces the possibility of overbor- rowing, improves origination standards, and reduces borrowing costs for good creditors. In addition, char- acteristics of the debt frameworks—such as protection against predatory lending—temper the negative asso- ciation with future GDP growth, but are not robustly significant. Other aspects of the institutional framework, such as various characteristics of the household credit market obtained through a survey of country desks, do not appear to have a significant effect in reducing the risks signaled by household credit expansion.24

The effect of household debt on GDP is somewhat larger in more unequal societies. The role of inequal- ity is not obvious because of two countervailing forces (Figure 2.10). On one hand, richer households tend to have lower debt-to-income (DTI) ratios and higher participation (Figure 2.4). A higher level of inequality

24For the list of housing market characteristics see Annex Figure 2.1.1. The lack of significance for several of these and other institutional measures may result from the reduced samples for which they are available or the limited time variation of the data (some being available for a single year).

(.. .,–

(– 2,

–1 .5

]

(– 1.

5, –1

]

(– 1,

–0 .5

]

(– 0.

5, 0]

(0 ,0

.5 ]

(0 .5

,1 ]

(1 ,1

.5 ]

(1 .5

,2 ]

[2 ,..

Advanced economies Emerging market economies

–1.5

–1.0

–0.5

KA openness Float or fix

Income inequality Transparency

Financial dev

Supervisory strict

1. Distribution of Country-Specific Coefficients (Relative frequency, percent)

2. Marginal Interaction Effects for Country Factors in High versus Low Quartiles (Percent)

Source: IMF staff calculations. Note: Panel 1 shows country-level coefficients of changes in household debt in ordinary least squares regressions of three-year-ahead real GDP growth on changes in household and firm debt. Panel 2 shows the marginal effect of changes in household debt on GDP growth three years ahead, from panel regressions with institutional factors, evaluated at the 25th and 75th percentiles. Effects are statistically significant at the 10 percent level or higher. See also Figure 2.8 and Annex 2.2. Financial dev = financial development index from Svirydzenka 2016; Float or fix = exchange rate regime (floating, the green bar, versus fixed, the red bar); Income inequality = income inequality measures, the difference between the income share of the top 20 and bottom 20 percent income groups; KA openness = capital account openness index from the Chinn-Ito Index; Supervisory strict = measure of overall bank capital stringency from Barth, Caprio, and Levine 2013; Transparency = dummy variable indicating whether a credit registry or other form of borrower information data transparency exists.

Figure 2.10. The Impact of Household Debt by Country and Institutional Factors

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

means that the share of income of the richest households increases and the macro-level DTI ratio declines.25 On the other hand, higher-income households may decide to borrow more as a response to their relatively higher income, leading to an increase in macro-level DTI. Thus, the relationship between macro-level household debt and inequality is ambiguous. In this sample, higher inequality is associated with a slightly higher impact of changes in household debt on future growth.26 Other explanations center on behavior, arguing that higher inequality results in more people with less financial education who are more vulnerable to overlending and predatory practices.27

These results suggest that the level of household debt at which further increases are detrimental is country specific and higher for countries with better institutions. The negative effects of increases in the household debt-to-GDP ratio on future GDP growth differ by country and depend on the initial level of indebtedness and country characteristics, as outlined earlier. This means that countries can attenuate the negative effects of increased household debt that arise at high initial levels of indebtedness if they are more financially developed and have higher standards of financial information transparency (credit registries) and consumer finance protection, better regulation and supervision, less inequality, and more flexible exchange rate regimes.28 In effect, the impact on growth of a rising household debt-to-GDP ratio appears to be posi- tive in the medium term when institutions and policies are the most effective, and appears to be negative when institutions and policies are the least effective, regard- less of the initial level of household debt.

Conclusions and Policy Implications The econometric analysis clearly shows that house-

hold debt has different effects on economic growth and financial stability depending on the horizon. At business cycle frequency, high growth in household lending appears to foster above-average growth and employ-

25The macro-level DTI is the weighted average of household-level DTIs, with weights by income share.

26However, the significance of this effect varies, depending on the exact model specification.

27Along these lines, Rajan (2010) argues that household debt among lower-income households was encouraged by the political system in the United States as an easier (but riskier) way to deal with income inequality.

28While capital openness may also strengthen the association between household debt and future growth decelerations, it does so mainly in combination with less flexible exchange rate regimes.

ment at first, but tends to be followed by a period of instability and subpar GDP growth and employment. This finding is consistent with the presence of a policy trade-off between short-term and medium-term growth and financial instability. While this forecasting trade-off is a robust pattern of the data, it is stronger for advanced economies than for emerging market economies, with increases in household debt consistently signaling higher risks when initial debt levels are already high. None- theless, the results indicate that the threshold levels for household debt increases being associated with negative macro outcomes start relatively low, at about 30 percent of GDP. Therefore, although emerging market econo- mies have some space to take advantage of the positive effects of expanding households’ access to credit—in both the short and long term—with low medium-term risks, such space may be limited. Furthermore, even in countries with low macro levels of household debt, a rapid expansion in credit may lead to an increasing fraction of highly leveraged households that may be vulnerable to shocks. Finally, existing studies suggest that household debt appears positive for growth across medium- to long-term horizons, although the relation- ship weakens at high levels of indebtedness.

A country’s characteristics, institutions, and policies can mitigate the risks associated with increasing house- hold debt. The negative effects are weaker in countries with less external financing and floating exchange rates, that are financially more developed, that have better financial sector regulations and policies, and that have lower income inequality. Thus, even in countries where the level of household debt to GDP is high, the stability-growth trade-off can be attenuated by a com- bination of good policies, institutions, and regulations. On the other hand, in countries where the low initial level of household debt mitigates some of the risks, the wrong combination of institutional characteristics and policies may offset the effect of a low debt level. This indicates that the point at which further increases in household debt pose risks to future economic perfor- mance is country specific; various factors should be evaluated by country authorities to assess vulnerabili- ties arising from household leverage.

Policy action will need to calibrate the short-, medium-, and long-term benefits and risks. Policies need to carefully balance minimizing the medium-term risks of growth in household credit for financial stabil- ity without harming the potential long-term benefits of inclusion and development. Moreover, policy

C H A P T E R 2 h O u S E h O L d d E B T A N d F I N A N C I A L S T A B I L I T Y

International Monetary Fund | October 2017

action must overcome the inaction bias and political pressure generated by the very short-term positive impact of household credit on GDP growth versus the medium-term negative impact.

In any event, certain policy changes can help reduce the impact of aggregate demand externalities and behavioral biases. Some of the drag household debt places on GDP can be reduced by moving away from fixed exchange rates; introducing financial sector policies that promote financial institutions and market depth, access, and efficiency; and advancing policies that help reduce income inequality. For the most part, these policy changes may also have long-term positive effects on growth. For example, as noted by Coibion and others (2017), lower inequality may enhance lower-income households’ access to credit and their ability to smooth consumption and make long-term investments (for example, sending children to college and retraining for different careers) that benefit society. Furthermore, the reliance on foreign debt and the role of capital flows may need further attention because they expose countries to sudden stops or destabilizing capital outflows (see also IMF 2014).

Macroprudential policy can help curb household leverage. Macroprudential policies can help internalize the externality that the borrowing by each household imposes on the rest of the financial system, given that large increases in household debt are associated with a greater likelihood of financial crises and recessions. The design of targeted macroprudential measures may need to take distributional aspects into account, since certain characteristics of households are associated with a greater misalignment of debt and future income. Detailed panel regression analysis shows that various macroprudential measures can significantly reduce real household credit growth, both in advanced econo- mies and in emerging market economies (Box 2.5). Demand-side measures, such as limits on the debt-service-to-income ratio and loan-to-value ratio, seem highly effective. Supply-side measures targeted at loans, such as limits on bank credit growth, loan con- tract restrictions, and loan loss provisions, are equally effective. However, these policies would require careful calibration to maintain the balance between the short-, medium-, and long-term effects discussed.

There is also a role for policymakers to further strengthen the protection of consumer finance. The

empirical analysis found that credit registries reduce the negative effects on growth in the medium term. The development of credit registries will help improve the welfare of households vulnerable to overborrowing. Consumer financial protection not only helps unso- phisticated consumers make wiser finance decisions, it also helps enhance overall financial stability, as shown in the empirical analysis. Measures could include increasing the transparency of financial contracts, financial education, prohibition of predatory lending, and regulation of certain financial innovation products.

Similarly, good microprudential supervision can mit- igate the negative effects of household debt. As amply demonstrated during the global financial crisis, differ- ences in the quality and depth of banking supervision helped explain why some countries escaped the nega- tive externalities associated with the large increase in household debt during the preceding decade. This may reflect stronger supervisory powers or more stringent capital regulation frameworks that allowed supervi- sors to diminish the negative effect of household debt increases on future GDP.

Market solutions may also help mitigate the eco- nomic consequences of household debt in financial recessions. For example, risk sharing between mortgage lenders and borrowers could be increased, which is the aim of the shared appreciation design of mortgage contracts advocated by Shiller (2014) and Mian and Sufi (2014). In this more equity-like design of mortgage contracts, the principal is automatically written down if the local house price index falls below a specified threshold; increases in property value are shared between the homeowner and the lender. This type of mortgage loan can help price in the associated crash risk before lenders extend credit and reduce the debt overhang problem of households when house prices fall. In theory, this approach would reduce the blow to the macroeco- nomy of housing busts during episodes of household deleveraging. It would thus enhance financial stability much as nonfinancial firms or banks benefit from bail-in debt with loss-absorbing capacity vis-à-vis bondholders (see Chapter 3 of the October 2013 Global Financial Stability Report). However, more work is needed on the conditions and pricing that would entice banks to offer such contracts and to get a full understanding of the potential effects on financial stability (including banks’ ability to absorb associated losses).

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

In the long term, higher levels of credit to GDP are generally associated with higher economic growth. Financial development, including better institutions and easier access to credit by households, has been shown to be beneficial to economic growth in the long term (Levine 1998; Beck and Levine 2004). As the financial sector develops, growth-enhancing invest- ments can be more easily financed. Nonetheless, the relationship between household debt and growth is more elusive (Jappelli and Pagano 1994; De Gregorio 1996; Beck and others 2012; Sahay and others 2015a).

Recent studies have found that economies may reach a point of “too much finance.” Arcand, Berkes, and Panizza (2015) and Sahay and others (2015b) found that financial depth begins to dampen out- put growth when credit to the private sector reaches between 80 percent and 100 percent of GDP. Too much finance may increase the frequency of booms and busts because of greater risk taking and leverage, and may leave countries ultimately worse off and with lower real GDP growth. Another argument is that too much finance leads to a diversion of talent and human capital away from productive sectors and toward the financial sector (Shiller 2005).

A more detailed analysis with household credit suggests the existence of a tipping point. An empirical exercise conducted for the countries covered in the chapter finds that household debt increases long-term real GDP per capita growth, but the effects weaken at higher levels of household debt and eventually become negative. The maximum positive impact in this exer- cise is found when household debt is between 36 per- cent and 70 percent of GDP (Figure 2.1.1, panel 1). In addition, there does not appear to be an effect specific to emerging market economies, but a financial crisis seems to result in permanently lower per capita GDP growth (Figure 2.1.1, panel 2).

Box prepared by Adrian Alter and Nico Valckx.

–0.5

0.5

1.0

1.5

100 20 30 40 50 60 70 80 90 100 110

Lo ng

-t er

m p

er c

ap ita

G D

P gr

ow th

Household debt-to-GDP ratio

95 percent confidence

bound around the

turning point

(1)

Variables Per Capita GDP Growth

(2) Per Capita

GDP Growth

(3) Per Capita

GDP Growth

HHD

HHD2

Crisis

EME × HHD

Education

Constant

Observations Number of

countries AR2 Hansen Instruments

Initial per capita GDP

Source: IMF staff calculations. Note: Figure shows nonlinear effect of household debt on long-term per capita GDP growth at various levels of household debt, based on a long-term panel regression. It uses the Arellano-Bover general method of moments estimator of five-year average per capita GDP growth (shown in panel 2) on household debt to GDP (HHD), the squared ratio of household debt to GDP (HHD2), initial per capita GDP, secondary education enrollment, dummies for banking crises (Crisis), and emerging market economies’ household debt-to-GDP ratio (EME × HHD). *** p < 0.01; ** p < 0.05; * p < 0.1. 1Z-statistics in parentheses.

Figure 2.1.1. Long-Term per Capita GDP Growth and Household Debt

1. Effect of Household Debt on per Capita GDP Growth (Percent)

2. Panel Regression of per Capita GDP Growth and Household Debt, 1970–20101

0.051* 0.007 0.021 (1.726) (0.346) (0.762) –0.048** –0.024 –0.051** (–1.980) (–1.494) (–2.057) –0.017*** –0.015*** (–6.319) (–4.688) –0.000 (–0.015) 0.028 0.018* 0.017 (1.117) (1.818) (1.576) –0.012** –0.004 –0.000 (–1.973) (–1.227) (–0.078) –0.035 –0.038 –0.066 (–0.353) (–0.933) (–1.507)

278 278 278 73 73 73

0.0186 0.137 0.185 0.253 0.797 0.361

55 73 68

Box 2.1. Long-Term Growth and Household Debt

C H A P T E R 2 h O u S E h O L d d E B T A N d F I N A N C I A L S T A B I L I T Y

International Monetary Fund | October 2017

Housing assets and mortgages are important com- ponents of the balance sheets of Chinese households. High levels of ownership (about 90 percent of the population own a property) make housing the largest asset of Chinese households: more than two-thirds of their total assets (Figure 2.2.1, panels 1 and 2). On

the liability side, urban households in China have increased their borrowing. Mortgage loans from banks account for the largest share of their debt. Consistent with the life-cycle theory of debt, participation rates among urban Chinese households across age groups follow a hump shape and are highest for younger

100

Q1 Q2 Q3 Q4 Q5

Housing-to-assets ratio Homeownership Mortgage-to-debt ratio

Q 1

Q 2

Q 3

Q 4

Q 5

15 –2

30 –4

45 –5

60 +

All Income quintile Income quintile

Age (years)

Q1 Q2

Q4 Q5

Q2Q3 Q4

100

0 200 400 600 800

D eb

t- se

rv ic

e- to

-i nc

om e

ra tio

Debt-to-income ratio

2011 2015

Q1 Q2 Q3 Q4 Q5

Income quintile

2011 2015

Debt-to-income ratio = 1

Debt-to-income ratio = 5

Below 2

Between 2 and 3

Between 3 and 4

Above 4

Sources: IMF staff calculations, based on China Household Finance Survey; see Gan and others 2013 for details. Note: Data shown are mainly for urban households from different income quintiles (Q1 to Q5, lowest to highest). The housing-to-assets ratio is defined as the ratio of housing assets to total assets. The mortgage-to-debt ratio is defined as the ratio of mortgage debt to total debt. The mortgage debt participation rate is computed across age groups. Debt-to-income (multiple) and debt-service-to-income (percentage) ratios by income quintiles are scaled by the share of each household quintile in total debt. The response of consumption-to-income shocks is the coefficient in the cross-sectional regressions of the percentage change in consumption on the percentage change in income between 2013 and 2015 among households that were tracked in the survey. In panel 2, “age” refers to the age of the head of household. For panel 5, a ratio above 4 indicates a highly indebted household.

Figure 2.2.1. Characteristics of China’s Household Debt (Percent)

1. Housing-to-Assets and Mortgage-to-Debt Ratios, and Homeownership

2. Mortgage Participation Rate

3. Debt-to-Income and Debt-Service-to- Income Ratio

4. Loan Balance-to-Value Ratio

5. Distribution of Household Debt by Debt-to-Income Groups

6. Response of Consumption to Income Shocks

Box 2.2. Distributional Aspects of Household Debt in China

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

households.1 Household debt has become an increas- ingly important component of credit in China. As the household debt-to-GDP ratio rose from 18.7 per- cent to about 38 percent from 2007 to 2016, loans to households as a percentage of total loans issued by financial institutions increased from 19.4 percent to 31.3 percent over the same period.2

The debt burden of mortgage borrowers in urban areas has increased in recent years, although mortgage participation rates are still relatively low compared with advanced economies. The debt-to-income ratio increased across most income groups, especially for lower-income households. The debt service ratio, defined as total debt repayment as a percentage of total income, also increased for all income groups but espe- cially for lower-income households (Figure 2.2.1, panel 3). The loan balance-to-value ratio, defined as the remaining loan balance as a percentage of self-reported housing value, also increased over time (Figure 2.2.1, panel 4). On the other hand, mortgage loan partici- pation rates, especially for low-income households, are

1Note that not many households of those ages 45–59 borrow for mortgages because a large share of today’s housing stock still originates from the planned-economy period during which the government or state-owned enterprises distributed housing.

2Only domestic-currency (renminbi) loans are included. Data on total loans and loans to households are based on Sources and Uses of Funds of Financial Institutions published by the People’s Bank of China.

still low, which is consistent with China’s economic and financial development level.

The increased household debt could amplify the macroeconomic consequences of negative shocks. Although household debt is about 38 percent of GDP in China, more than one-third of it is held by highly indebted households, defined as those with a debt-to-income ratio greater than 4 (Figure 2.2.1, panel 5). This means that deterioration in the balance sheets of these households could have an amplified negative impact on the banking sector as well as on the macroeconomy, even though loans to house- holds, including home mortgages, in China are still a smaller fraction of banks’ total assets than in advanced economies. In addition, empirical evidence based on tracked samples of Chinese households between 2013 and 2015 shows that consumption of households with high debt to income responds more strongly to income shocks (Figure 2.2.1, panel 6). This suggests that negative shocks to household balance sheets may amplify the effect on China’s economy because of highly indebted households’ higher marginal propen- sity to consume—a pattern consistent with evidence in advanced economies (for example, Mian, Rao, and Sufi 2013).

Box prepared by Alan Xiaochen Feng, in collaboration with Feng Li and Xiaomeng Lu from the Survey and Research Center for China Household Finance at Southwestern University of Finance and Economics.

Box 2.2 (continued)

C H A P T E R 2 h O u S E h O L d d E B T A N d F I N A N C I A L S T A B I L I T Y

International Monetary Fund | October 2017

Until the global financial crisis, household debt levels evolved very similarly in the United States and Canada. US household debt increased from 56 percent in 1995 to nearly 100 percent of GDP in the first quarter of 2008 and from 62 percent to 80 percent in Canada (Figure 2.3.1, panel 1). Afterward, US household debt fell to below 80 percent by early 2017, whereas in Canada, it continued to rise to more than 100 percent. This reflects different house price and unemployment trends, as well as difference in the evo- lution of net wealth, which left Canadian households relatively better off than their US counterparts.

Box prepared by Adrian Alter, Alan Xiaochen Feng, and Nico Valckx.

The composition of household debt has changed in both countries. In response to continuously rising house prices, Canadian household debt became more tilted toward mortgage debt, which increased from 61 percent of total debt in 2005 to 66 percent of total debt in 2016 (Figure 2.3.1, panel 2). In the United States, where house prices fell by 40 percent from their peak in 2008, households’ share of mortgage debt decreased, while consumer debt increased substantially, mainly because of increased student loan debt.

Leverage is very different across households. US households’ leverage (as given by the debt-to-income ratio) remained broadly constant, except for the poorest income group, whose leverage increased slightly. In Canada, on the other hand, debt-to-income

120

160

200

240

100

130

1995 99 2003 07 11 15

In de

Pe rc

en t

Canadian debt (left scale)

US debt (left scale) Canadian real house price (right scale)

US real house price (right scale)

100

2005 16 2005 16

Canada United States

Mortgage Consumer Other

200

400

600

Q5 Income quintile

2004 2013

200

400

600

Q1 Q2 Q3 Q4Q1 Q2 Q3 Q4 Q5 Income quintile

2005 2012

Source: IMF staff calculations, based on the Luxembourg Wealth Study, US Survey of Consumer Finances, and the Canadian Survey of Financial Security. Note: Panels 3 and 4 refer to the median debt-to-income levels by income quintiles for mortgage borrowers.

Figure 2.3.1. US and Canadian Household Debt Developments and Characteristics

1. Household Debt-to-GDP Ratio and House Prices

2. Composition of Household Debt (Percent)

3. United States: Debt-to-Income Ratio Distribution (Percent)

4. Canada: Debt-to-Income Ratio Distribution (Percent)

Box 2.3. A Comparison of US and Canadian Household Debt

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

ratios increased across all income groups, resulting in an average ratio almost 50 percent higher than in the United States (Figure 2.3.1, panels 3 and 4). Moreover, highly indebted households (those with debt-to-income ratios above 350 percent) held more than Can$400 billion, or 21 percent of the total household debt in Canada at the end of 2014, up from 13 percent before the crisis (Bank of Canada 2015).

High leverage may expose households to poten- tially adverse income shocks. The past recession in the

United States showed that highly indebted households substantially reduced spending, which contributed to a significant decline in aggregate demand (Mian and Sufi 2011). Results reported in this chapter are in line with analysis by the Bank of Canada, which in its latest Financial System Review highlighted high household indebtedness and imbalances in the Canadian housing market as its two most important vulnerabilities; accordingly, it has implemented several macroprudential measures to mitigate these prob- lems (IMF 2017).

Box 2.3 (continued)

C H A P T E R 2 h O u S E h O L d d E B T A N d F I N A N C I A L S T A B I L I T Y

International Monetary Fund | October 2017

Household debt leads to higher house prices and more debt in the future, likely through reinforcing feedback effects. Dynamic panel vector autoregression analysis confirms that household debt has a short-term positive effect on real house prices and output.1 A one standard deviation shock to household debt initially leads to higher real house prices and output, but over the medium term (after about three to five years) results

Box prepared by Adrian Alter and Alan Xiaochen Feng. 1The panel vector autoregression model was conducted with a

set of 27 countries with quarterly data available starting in 1998.

in a decline (Figure 2.4.1, panels 1 and 3).2 Higher house prices are positively associated with output in the short and medium term, but negatively in the long term (Figure 2.4.2). In response to a positive shock to house prices, household debt increases steadily over the short and medium term, while reverting to its long-term mean thereafter (Figure 2.4.1, panel 4).

2These findings are consistent with Lombardi, Mohanty, and Shim 2017. See also Mian, Sufi, and Verner, forthcoming; Calza, Monacelli, and Stracca 2013; and Brunnermeier and others 2017.

–0.8

–0.4

0.0

0.4

0.8

0 4 8 12 16 20 24

Quarter

–0.5

0.0

0.5

0 4 8 12 16 20 24

Quarter

–2.0

–1.5

–1.0

–0.5

0.0

0.5

1.0

1.5

0 4 8 12 16 20 24

Quarter

–0.3

0.0

0.3

0.5

0.8

1.0

0 4 8 12 16 20 24

Quarter

Source: IMF staff calculations. Note: The figure presents impulse responses from a five-variable recursive panel vector autoregression with eight lags using quarterly data from 1998:Q1 to 2015:Q4, which includes country and time fixed effects. Shocks are identified using a Cholesky decomposition with the following order: log real GDP, corporate debt, household debt, log real house prices, and short-term interest rates. Household debt and corporate debt were scaled by GDP. The results are robust to a Nickell bias correction (using panel general method of moments techniques) and other specifications (for example, ordering, number of lags, changes instead of levels). Dashed lines represent 90 percent confidence intervals, computed using 500 Monte Carlo simulations.

Figure 2.4.1. Panel Vector Autoregression Dynamic Analysis (Percentage points)

1. Shocks to Household Debt Ratio: Effect on Real Output

2. Shocks to House Prices: Effect on Real Output

3. Shocks to Household Debt Ratio: Effect on House Prices

4. Shocks to House Prices: Effect on Household Debt Ratio

Box 2.4. The Nexus between Household Debt, House Prices, and Output

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

Micro-level panel survey data analysis confirms the impact of house prices on consumption and the role of debt. In Korea, the rise in the local house price index between 2008 and 2014 had a positive effect on household consumption, which is consistent with the initial positive response of GDP to house price shocks shown in the panel vector autoregression analysis.3

3This empirical exercise uses tracked samples of households between 2008 and 2014 and controls for changes in household income, demographic information, and city-level aggregates.

Such an effect is present only for homeowners, suggesting that the increase in house prices raises collateral value as well as perceived wealth for these households (Figure 2.4.2, panel 1). Similarly, in Australia, homeowners increased consumption in response to higher local house prices between 2012 and 2015, and the effect was stronger for house- holds with high financial leverage. This finding indicates that higher household debt reinforces the impact of house prices on the real economy (Figure 2.4.2, panel 2).

–0.1

0.1

0.2

0.3

0.4

Homeowner Non–homeowner 0

0.1

0.2

0.3

Debt-to-income ratio = 2

Debt-to-income ratio = 4

Sources: Australian Bureau of Statistics; Household, Income and Labour Dynamics in Australia; Korean Labor and Income Panel Study; Statistics Korea; and IMF staff calculations. Note: For households in Korea, regression coefficients are obtained by regressing the percentage change in consumption on changes in the local house price index between 2008 and 2014. For households in Australia, regression coefficients are obtained by regressing the percentage change in consumption on changes in the local house price index between 2012 and 2015. In both analyses, controls include the percentage change in household income, debt, and other demographic information, as well as state-level changes in income over the same period. Samples of households in both countries are restricted to those tracked over the period covered. Low leverage corresponds to a debt-to-income ratio of 2 and high leverage corresponds to a debt-to-income ratio of 4. Standard errors are clustered at the state or province level.

Figure 2.4.2. Consumption Response to House Prices (Percent)

1. Household Consumption in Korea 2. Household Consumption in Australia

Box 2.4 (continued)

C H A P T E R 2 h O u S E h O L d d E B T A N d F I N A N C I A L S T A B I L I T Y

International Monetary Fund | October 2017

This box finds that macroprudential loan-targeted measures successfully reduce the growth of real household credit in both advanced economies and emerging market economies.

Many countries introduced or tightened macropru- dential policy measures to limit systemic risk in the aftermath of the large credit boom that preceded the

Box prepared by Adrian Alter and Machiko Narita.

global financial crisis (Figure 2.5.1, panel 1). In theory, macroprudential policies reduce systemic risk by correct- ing externalities operating through the financial system. Such externalities include aggregate demand externalities and strategic complementarities among financial institu- tions, which amplify credit and asset price cycles.1

1See, for example, Hanson, Kashyap, and Stein 2011; De Nicolò, Favara, and Ratnovski 2012; and IMF 2013.

–30

–20

–10

–80 –60 –40 –20

0 20 40 60 80

1990 92 94 96 98 2000 02 04 06 08 10 12 14 16

EME macroprudential policies (cumulative sum, left scale)

AE macroprudential policies (cumulative sum, left scale)

EME real household credit (year-over-year growth, percent, right scale)

AE real household credit (year-over-year growth, percent, right scale)

–3

–2

–1

D eb

t- se

rv ic

e- to

- in

co m

e ra

tio

Lo an

-t o-

va lu

e ra

tio

Lo an

o r

bo rr

ow in

g lim

its o

r pr

oh ib

iti on

s Li

m its

o n

ba nk

cr ed

it gr

ow th

Lo an

lo ss

pr ov

is io

ns Re

se rv

e re

qu ire

m en

Le ve

ra ge

r at

Co un

te rc

yc lic

al ca

pi ta

l b uf

fe r

Li m

its o

n fo

re ig

n ex

ch an

ge p

os iti

on s

All AEs EMEs All AEs EMEs

–1.5

–1.2

–0.9

–0.6

–0.3

0.0

0.3

Al l

Lo an

D em

an d

Su pp

G en

er al

Ca pi

ta l

Lo an

Supply

Macroprudential policies

Source: IMF staff calculations. Note: In panel 1, the macroprudential policies show the cumulative sum of tightening (+) and loosening (–) policies. Panel 2 shows the estimated average effects on real household credit growth of one tightening event for each macroprudential measure, one at a time, in a panel regression of 62 countries (32 advanced economies and 30 emerging market economies). In panel 3, All comprises all 14 measures considered. Loan consists of demand-side and supply-side loans. Demand includes debt-service-to-income ratios and loan-to-value ratios. Supply measures are classified into General, Capital, and Loans. Supply (General) consists of reserve requirements, liquidity requirements, limits on foreign exchange positions, and taxes on financial institutions. Supply (Capital) consists of capital requirements, conservation buffers, the leverage ratio, and the countercyclical capital buffer. Supply (Loans) consists of limits on bank credit growth, loan loss provisions, loan restrictions, and limits on foreign currency loans. Shaded bars depict significant effects at the 10 percent confidence levels. See Annex 2.2 for estimation details. AEs = advanced economies; EMEs = emerging market economies.

Figure 2.5.1. Macroprudential Policy Tools and Household Credit Growth

1. Number of Macroprudential Policies and Real Household Credit Growth

2. Effect of Individual Macroprudential Tools (Percentage points)

3. Effect of Combined Policies, Average by Type (Percentage points)

Box 2.5. The Impact of Macroprudential Policies on Household Credit

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

In both advanced and emerging market econo- mies, targeted macroprudential measures successfully reduce real household credit growth. From a set of 14 measures, 5 measures related to credit have robust negative effects (Figure 2.5.1, panel 2). These measures are limits on the debt-service-to-income (DSTI) ratio, limits on the loan-to-value (LTV) ratio, loan contract restrictions, limits on bank credit growth, and loan loss provisions. On average, a tightening of these measures leads to a 1 to 3 percentage point decline in real household credit growth, similar to Kuttner and Shim’s (2016) results for LTV and DSTI ratio limits.2 The effects are generally stronger in emerging market economies, corroborating the findings of Cerutti and others (2017).3

On the other hand, measures that are not targeted to loans do not exhibit strong effects in contracting household credit. Reserve requirements also tend to

2Other studies, using different data and methodologies, also show that tighter LTV and DSTI ratios reduce household credit growth. See Lim and others 2011; Arregui and others 2013; Crowe and others 2013; Krznar and Morsink 2014; and Jácome and Mitra 2015.

3Loan restrictions and limits on credit growth also appear to effectively contain corporate credit growth, to the tune of 2 to 3 percentage points, while other measures have a weak or insignificant impact. The latter could reflect firms’ better access to (international) debt markets than households.

have negative effects, but they are smaller and less significant than targeted measures.4 Leverage limits, conservation buffers, and limits on foreign exchange positions are positively associated with subsequent growth in household credit. Other measures, such as capital requirements and taxes on financial interme- diaries, do not have significant effects. However, a tightening of general supply measures should increase the resilience of the financial system to aggregate shocks by building buffers. Previous studies also find weaker effects of nontargeted and capital measures and may explain their lack of effectiveness, including leakages. For example, tightening capital require- ments may have little effect when banks hold ample capital. When examining the effects of measures by type, demand-side measures (DSTI and LTV) as well as loan-targeted supply-side measures (on domestic credit growth and loan loss provisions) are found to be effective (Figure 2.5.1, panel 3).5

4See Arregui and others 2013; Crowe and others 2013; Vandenbussche, Vogel, and Detragiache 2015; and Kuttner and Shim 2016.

5Combining same-type measures allows the effects of multiple measures adjusted at the same time to be controlled for. For example, Kuttner and Shim (2016) report that changes in DSTI and LTV ratio limits are often coordinated.

Box 2.5 (continued)

C H A P T E R 2 h O u S E h O L d d E B T A N d F I N A N C I A L S T A B I L I T Y

International Monetary Fund | October 2017

Annex 2.1. Data Sources

Annex Table 2.1.1. Countries Included in the Sample for Household Debt and Data Sources Country Source Start Year Country Source Start Year Advanced Economies Emerging Market Economies Australia BIS; JST 1952 Argentina BIS 1994 Austria BIS 1995 Bangladesh Haver 2004 Belgium BIS; JST 1950 Bolivia Central Bank of Bolivia 1992 Canada BIS; JST 1956 Botswana IMF, MFS 2001 Cyprus CEIC 1995 Brazil BIS 1994 Czech Republic BIS 1995 Bulgaria ECRI 1995 Denmark BIS; JST 1951 Chile BIS; Central Bank of Chile 1983 Estonia Haver; Bank of Estonia 1993 China BIS 2006 Finland BIS; JST 1950 Colombia BIS 1996 France BIS; JST 1958 Costa Rica Central Bank of Costa Rica 1997 Germany BIS; JST 1950 Croatia Croatian National Bank 1993 Greece Haver 1980 Egypt Central Bank of Egypt 2002 Hong Kong SAR CEIC 1982 FYR Macedonia National Bank of the Republic of Macedonia 1995 Iceland Haver; IMF, MFS 1995 Georgia IMF, MFS 2001 Ireland ECRI 1998 Ghana IMF Bridge Data; IMF, MFS 2001 Israel BIS 1992 Hungary BIS 1989 Italy BIS 1950 India CEIC 1998 Japan BIS; JST 1950 Indonesia BIS 2001 Korea BIS 1962 Jordan Central Bank of Jordan 1993 Latvia Haver 2003 Kazakhstan Haver 1996 Lithuania Haver 1993 Kenya IMF, MFS 2001 Luxembourg Haver 1992 Kuwait CEIC 1997 Malta ECRI 1995 Malaysia IMF, MFS 2001 Netherlands BIS 1990 Mauritius IMF, MFS 2001 New Zealand BIS 1990 Mexico BIS 1994 Norway BIS 1975 Mongolia IMF, MFS 2001 Portugal BIS 1979 Montenegro ECRI 1995 Singapore BIS 1991 Morocco IMF, MFS 2001 Slovak Republic National Bank of Slovakia 1993 Namibia IMF, MFS 2001 Slovenia Haver; IMF, MFS 2004 Nigeria IMF, MFS 2001 Spain BIS; JST 1950 Pakistan IMF, MFS 2006 Sweden BIS; JST 1975 Panama IMF, MFS 2002 Switzerland BIS; JST 1950 Paraguay Central Bank of Paraguay; IMF, MFS 1990 United Kingdom BIS; JST 1950 Philippines Central Bank of the Philippines 1999 United States BIS; JST; CEIC 1950 Poland BIS 1995 Romania ECRI 1996 Russia BIS 1995 Saudi Arabia BIS; CEIC 1995

Serbia IMF, MFS 2003 South Africa Haver 1969 Thailand BIS 1991 Turkey BIS 1986 Ukraine IMF, MFS 2001 Uruguay BIS 2001 Venezuela BIS 2001

Sources: IMF staff. Note: BIS = Bank for International Settlements; CEIC = CEIC Data Co. Ltd.; ECRI = Economic Cycle Research Institute; Haver = Haver Analytics; IMF, MFS = Monetary and Financial Statistics database; JST = Jordà-Schularick-Taylor Macrohistory Database.

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

100

REC FIX PEN GOV NAT TAXD TAXL TRA PROT

Advanced economies Emerging market economies

Source: IMF staff calculations. Note: Figure is based on an IMF desk survey of the prevalence of certain debt characteristics in 80 countries. The desk survey reveals that a majority of countries have financial protection regulations (against predatory lending practices) and loan transparency rules and regulations (through credit registries or credit bureaus). In 80 percent of the sample, recourse is commonplace in loan agreements, whereas early prepayment restrictions feature in about 40 percent of the countries surveyed. Tax deductibility is common in half of the sample, with limitations on how much debt (or interest payments) households can deduct from their taxes. Fixed-rate mortgages (with the initial rate fixed for 10 or more years) are offered in most countries. Administrative restrictions on land supply are more prevalent in advanced economies (about 60 percent) than in emerging market economies (44 percent), whereas natural restrictions exist in about 30 percent of the countries surveyed (related to size of the country, livable land area, population density, and the like). FIX = fixed rates are offered; GOV = administrative restrictions on land supply; NAT= natural restrictions on density of development, such as topography and geography; PEN = restrictions on early payment; PROT = consumer financial protection legislation in place; REC = mortgage loans are full recourse; TAXD = debt or interest payments are tax deductible; TAXL = limits on TAXD exist; TRA = credit registry.

Annex Figure 2.1.1. Loan Characteristics, Rules, and Regulations

Annex Table 2.1.2. Household Survey Data Sources Country Name of Survey Advanced Economies Australia Household, Income and Labour Dynamics in Australia Survey Canada Luxembourg Wealth Study, Survey of Financial Security Euro Area European Central Bank’s Household Finance and Consumption Survey; Luxembourg Income Study (LIS); Luxembourg

Wealth Study (LWS) Japan Keio Household Panel Survey Korea Korean Labor and Income Panel Study; Korean Statistical Information Service Netherlands DNB Household Survey United Kingdom British Household Panel Survey United States Luxembourg Wealth Study, Survey of Consumer Finances

Emerging Market Economies China China Household Finance Survey

Source: IMF staff.

C H A P T E R 2 h O u S E h O L d d E B T A N d F I N A N C I A L S T A B I L I T Y

International Monetary Fund | October 2017

Annex Table 2.1.3. Description of Explanatory Variables Used in the Chapter Variables Description Source Macro-level Variables Nominal GDP Gross domestic product, current prices, national currency Jordà-Schularick-Taylor Macrohistory

database; Penn World Table; IMF, World Economic Outlook database

Real GDP Gross domestic product, constant prices, national currency IMF, World Economic Outlook database Real Private Consumption Private final consumption, constant prices, national currency IMF, World Economic Outlook database Consumer Price Index Consumer prices, period average, index IMF, International Financial Statistics

database Population Population, in millions of persons IMF, World Economic Outlook database Unemployment Unemployment rate (percent) IMF, World Economic Outlook database Interest Rate Three-month Treasury bill rate, money market rate, interbank market rate

(percent) Bloomberg Finance L.P.; IMF,

International Financial Statistics database; Thomson Reuters Datastream

Bank Equity Index Equity price index of the banking sector (or financial sector if banking sector price index not available)

Bloomberg Finance L.P.; Thomson Reuters Datastream

Stock Market Index Overall stock price index Bloomberg Finance L.P.; IMF, Global Data Source database; Thomson Reuters Datastream

Banking Crisis Systemic banking crisis defined as (1) significant signs of financial distress in the banking system (as indicated by significant bank runs, losses in the banking system, and/or bank liquidations); (2) significant banking policy intervention measures in response to significant losses in the banking system

Laeven and Valencia 2013

Real House Price Index House price index deflated by consumer price index Jordà-Schularick-Taylor Macrohistory database; OECD, Global Property Guide; and IMF staff calculations

Exchange Rate National currency units per US dollar, period average Thomson Reuters Datastream Real Effective Exchange

Rate Real effective exchange rate, based on consumer price index IMF, Monetary and Financial Statistics

database Exchange Rate Regime De facto exchange rate arrangement of the country Ilzetzki, Reinhart, and Rogoff 2017

data set

Institutional Variables Financial Risk Index Measure of a country’s ability to pay its way by financing its official, commercial, and

trade debt obligations; index ranges from 50 (least risk) to a low of 0 (highest risk) International Country Risk Guide, PRS

Group Financial Development Index Overall financial development index Svirydzenka 2016 Capital Account Openness

Index (Chinn-Ito Index) An index measuring a country’s degree of capital account openness Chinn and Ito 2006 data set (updated)

Official Supervisory Power Whether the supervisory authorities have the authority to take specific actions to prevent and correct problems; index ranges from 0 (no powers) to 14 (most powers)

Barth, Caprio, and Levine 2013

Overall Capital Stringency Whether the capital requirement reflects certain risk elements and deducts certain market value losses from capital before minimum capital adequacy is determined; index ranges from 0 (least stringent) to 7 (most stringent)

Barth, Caprio, and Levine 2013

Income Share Held by Highest 20 Percent

Percentage share of income or consumption is the share that accrues to subgroups of the population indicated by deciles or quintiles

World Bank, World Development Indicators

Income Share Held by Lowest 20 Percent

Percentage share of income or consumption is the share that accrues to subgroups of the population indicated by deciles or quintiles

World Bank, World Development Indicators

Source: IMF staff. Note: OECD = Organisation for Economic Co-operation and Development.

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

Annex 2.2. Methodology This annex provides a general overview of the meth-

odologies behind the various econometric exercises performed in this chapter.

Logit Analysis

The logit model analyzes how levels and changes in household debt affect financial stability. The model is given by

log P [ S it = 1 | X it ] __________ P [ S it = 0 | X it ]

= Ψ 0i + Ψ 1 X it

+ Ψ 2 X it I (HiDebt) it + ϵ it , (A2.2.1)

in which Xit refers to a vector of lagged changes and levels of household and corporate debt-to-GDP ratios, while the third term refers to interactions with an indicator I (HiDebt). The latter takes the value of one if country i experiences household debt exceed- ing 65 percent of GDP. Country fixed effects ( Ψ 0i ) were included in the estimation. The main metric to compare model performance is the area under curve. Annex Table 2.2.1 contains the underlying estimates.

Household Debt and Bank Equity Returns

This exercise provides an alternative measure of banking stress and assesses the role of household debt for future bank equity returns. According to the effi- cient market hypothesis, past household credit growth should not be correlated with future bank stock returns if investors correctly price the risks associated with the rise in household debt to the banking sector. However, downside risks may be neglected by investors during credit booms when market sentiments are high (for example, Cheng, Raina, and Xiong 2014; Baron and Xiong 2017), leading to systematic predictability of bank stock declines following increases in household debt. Following Baron and Xiong (2017), the empiri- cal specification is given by

r c,t + k − r c,t + k f = α c + γ t + β h Δ (

HHD _____ GDP

) c,t

+ β f Δ ( NFCD _____ GDP

) c,t + β d

× DivYl d c,t + X c,t δ + ϵ c,t , (A2.2.2)

in which r c,t+k is the return in year k of the bank- ing sector index in country c; is government bond

Annex Table 2.2.1. Logit Analysis: Probability of Systemic Banking Crisis

Variables (1) (2) (3) (4) (5)

Dependent Variable: Systemic Banking Crises Household Debt 4.037*** 2.501*** 1.270 2.091 (0.783) (0.925) (1.276) (1.716) Δ Household Debt 40.05*** 35.01*** 35.60*** 30.86*** (6.482) (6.334) (7.161) (8.451) Corporate Debt 0.879 0.536 (0.761) (0.743) Δ Corporate Debt 13.13*** 15.62*** (3.954) (4.220) Δ Household Debt × High HH Debt 24.41* (14.11) High HH Debt −1.355 (0.896) Constant −5.949*** −3.741*** −5.465*** −5.224*** −5.253*** (0.594) (0.150) (0.681) (0.732) (0.902) Observations 1,223 1,033 1,033 1,020 1,020 Country Cluster Yes Yes Yes Yes Yes Country Fixed Effect Yes Yes Yes Yes Yes Area under Curve 0.700 0.791 0.806 0.840 0.850 Number of Crises 46 37 37 37 37 Number of Clusters 40 34 34 34 34 Pseudo R 2 0.0612 0.142 0.153 0.204 0.218

Source: IMF staff calculations. Note: Robust standard errors in parentheses. All regressors are lagged. The third lag of household debt change was used based on significance. High household debt (High HH Debt) dummy variable is set at 65 percent of GDP, representing the top quintile of the distribution. Banking crises are taken from the updated database by Laeven and Valencia (2013). *** p < 0.01; ** p < 0.05; * p < 0.1.

C H A P T E R 2 h O u S E h O L d d E B T A N d F I N A N C I A L S T A B I L I T Y

International Monetary Fund | October 2017

yield, and DivYl d c,t is the dividend yield of the banking sector,

Δ ( HHD _____ GDP

) c,t = (

HHD _____ GDP

) c,t − (

HHD _____ GDP

) c,t − 1

and

Δ ( NFCD _____ GDP

) c,t = (

NFCD _____ GDP

) c,t − (

NFCD _____ GDP

) c,t − 1

(A2.2.3)

normalized by the standard deviation of each variable for each country, and X c,t includes control variables such as the past levels of household debt and corporate debt ratios.

The baseline model is estimated using the specifi- cation above. Two similar models are also estimated using probit analysis and quantile regressions. The probit analysis examines the relationship between past increases in the household debt ratio and the probabil- ity of bank equity crashes occurring in the next one to five years. Bank equity crashes are defined as having an annual stock return below the mean return by at least one standard deviation. In the quantile regressions, the relationship between past increases in the household debt ratio and future bank equity returns at different quantiles is examined.

Time Series Analysis of Household Debt, Income, and Consumption

Panel regressions are estimated following Mian, Sufi, and Verner, forthcoming, estimating future real GDP growth on changes in household debt and corporate debt ratios and lagged GDP growth rates. Different specifications are estimated, with changes in the debt ratio calculated over the past three years. In addition, level effects, thresholds, and nonlinearities are tested. Regression estimates are further differentiated by var- ious groupings: advanced and emerging market econ- omies, various institutional factors, and loan terms. Estimations are also performed over different time periods (before and after the global financial crisis) and were qualitatively very similar.

Specifically, the following general equation was estimated:

Δ h y i,t + h = α i h + β HH

h Δ 3 d i,t − 1 HH

+ β F h Δ 3 d i,t − 1

F + X i,t − 1 Γ h + ϵ it

h (A2.2.4)

in which α i h are country fixed effects, Δ3 refers to

three-year differences, d i,t HH and d i,t

F are the household debt-to-GDP ratio and nonfinancial firm debt-to-GDP

ratio, and h = 0, . . . ,6 is the forecast horizon. The matrix Xit includes higher-order lags of the dependent variable as additional controls. Right-hand variables are lagged by one year. Annex Table 2.2.2. provides a summary of the major panel regression estimates.

Micro Data Analysis

Euro area panel data allow the effects of household leverage on consumption, using a longitudinal house- hold panel, to be tested. Specifically, from a broader euro area household finance and consumption survey of 15 to 20 countries for 2010 and 2014, data for Belgium, Cyprus, Germany, Malta, and the Nether- lands allow testing for the effects of initial household debt-to-income and loan-to-value ratios on changes in the consumption-to-income ratio.

The following cross-sectional regression is estimated, at the household level, with change in household food consumption (percent of income) as the depen- dent variable:

Δ C i,2014 = α c + β 1 DT I i,2010

+ γControls + ϵ i , (A2.2.5)

in which debt-to-income ratio (DTIi,2010) is a proxy for past household indebtedness; household charac- teristics (such as employment, education, age of the household head, household’s net wealth and size) are considered Controls. In addition, the model includes country fixed effects ( α c ).

Macroprudential Policies and Household Credit Growth

Analysis in Box 2.5 gauged the effectiveness of macroprudential tools for reducing household credit growth. More specifically, the following panel regres- sion equation was estimated:

C i,t = ρ C i,t − 1 + β MaPP i,t − 1

+ γ X i,t − 1 + α i + μ t + ϵ i,t , (A2.2.6)

in which α i and μ t denote country and year fixed effects, i denotes country, and t the time period (quarter). The dependent variable, C i,t , refers to year-over-year growth rate of real household credit. The main independent variable, MaPP, is the policy change indicator (that is, tightening or loosening) compiled by IMF staff for each of the 14 macroprudential tools (that is, limits on the debt-service-to-income ratio, loan-to-value ratio, loan restrictions, limits on bank

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

An ne

x Ta

bl e

2. 2.

2. P

an el

R eg

re ss

io n

Es tim

at es

f or

T hr

ee -Y

ea r-

Ah ea

d G

ro w

th R

eg re

ss io

n on

H ou

se ho

ld D

eb t

an d

Po lic

y In

te ra

ct io

n Va

ri ab

le s

(1 )

(2 )

(3 )

(4 )

(5 )

(6 )

(7 )

(8 )

(9 )

Ch an

ge H

H D

/G D

P 0.

10 4

0. 14

1* **

−0 .5

88 *

−0 .0

62 −0

.2 41

* −0

.1 09

0. 01

5 −0

.7 58

** −0

.7 96

Ch an

ge F

irm D

/G D

P −0

.0 36

* −0

.0 37

* −0

.0 28

* −0

.0 34

* −0

.0 37

* −0

.0 35

* −0

.0 12

* −0

.0 32

* −0

.0 31

H H

D 30

× Δ

H H

D −0

.2 61

* −0

.3 67

* −0

.3 73

* −0

.3 60

* −0

.2 80

* −0

.4 35

* −0

.0 80

* −0

.3 10

* −0

.3 04

Fi na

nc ia

l O pe

nn es

s In

de x × Δ

H H

D −0

.1 20

−0

.1 23

* −0

.0 93

Fi xe

d FX

× Δ

H H

−0 .3

01 *

−0

.1 13

** *

0. 03

2 Fi

na nc

ia l R

is k

In de

x × Δ

H H

01 6*

0. 02

0* 0.

01 9*

In co

m e

In eq

ua lit

y × Δ

H H

0. 00

−0 .0

06 **

* −0

.0 04

Tr an

sp ar

en cy

× Δ

H H

28 5*

0. 24

6* 0.

20 2*

Fi na

nc ia

l D ev

el op

m en

t I nd

ex ×

Δ H

H D

36 9*

39 4*

** 0.

44 5*

Fi na

nc ia

l O pe

nn es

s In

de x

0. 03

−0

.5 88

× Fi

xe d

FX Fi

na nc

ia l O

pe nn

es s

In de

−0

.0 58

−0 .0

90 **

× Fi

xe d

FX ×

Δ H

H D

R 2

Ad ju

st ed

0. 58

1 0.

57 2

0. 57

5 0.

56 0.

57 0.

56 8

0. 58

5 0.

61 6

0. 61

8 O

bs er

va tio

ns 1,

00 2

1, 00

2 1,

00 2

1, 00

2 1,

00 2

1, 00

2 1,

00 2

1, 00

2 1,

00 2

N um

be r

of C

ou nt

rie s

57 57

57 Ak

ai ke

In fo

rm at

io n

Cr ite

rio n

6. 16

6. 18

6. 17

6. 2

6. 18

6. 19

3. 95

6. 08

6. 07

F- st

at is

tic 16

.1 15

.6 15

.7 14

.8 15

.4 15

.3 16

.4 16

.7 16

.6 Lo

g Li

ke lih

oo d

−2 ,9

91 −3

,0 01

−2 ,9

98 −3

,0 15

−3 ,0

04 −3

,0 06

−1 ,8

85 −2

,9 42

−2 ,9

38 So

ur ce

: I M

F st

af f e

st im

at es

. No

te : A

ll pa

ne l e

st im

at io

ns in

cl ud

e co

un try

fi xe

d ef

fe ct

s, ti

m e

fix ed

e ffe

ct s,

a nd

b as

e ef

fe ct

s. E

st im

at io

ns a

re p

er fo

rm ed

o ve

r a c

on st

an t s

am pl

e (fo

r w hi

ch d

at a

on a

ll va

ria bl

es a

re a

va ila

bl e)

. S ta

nd ar

d er

ro rs

a re

ro bu

st e

st im

a- to

rs . F

ix ed

F X

= fix

ed e

xc ha

ng e

ra te

re gi

m e

du m

m y;

H HD

= h

ou se

ho ld

d eb

t; HH

D3 0

= du

m m

y if

ho us

eh ol

d de

bt -to

-G DP

ra tio

e xc

ee ds

3 0

pe rc

en t;

in co

m e

in eq

ua lit

y =

di ffe

re nc

e be

tw ee

n in

co m

e sh

ar e

of to

p 20

p er

ce nt

a nd

e bo

tto m

2 0

pe rc

en t i

nc om

e gr

ou ps

; t ra

ns pa

re nc

y =

a du

m m

y va

ria bl

e, w

he th

er a

c re

di t r

eg is

try o

r o th

er fo

rm o

f b or

ro we

r i nf

or m

at io

n da

ta tr

an sp

ar en

cy e

xi st

s. **

* p

< 0.

01 ; *

*p <

0 .0

5; *

p <

0. 1.

C H A P T E R 2 h O u S E h O L d d E B T A N d F I N A N C I A L S T A B I L I T Y

International Monetary Fund | October 2017

credit growth, loan loss provisions, reserve require- ments, liquidity requirements, limits on foreign exchange positions, capital requirements, conservation buffers, leverage ratio, countercyclical capital buffer, limits on foreign currency loans, and taxes on financial institutions) or macroprudential group indices (that is, all MaPPs, loan MaPPs, demand, supply, supply [gen- eral], supply [capital], and supply loans). MaPPs are the cumulative sum of the number of policy changes over the past year (that is, the past four quarters) to reflect the potential delayed effects. A vector of control variables, X i,t , such as real output growth and domestic interest rates, is also included. The model is estimated with quarterly data from 62 countries (32 advanced economies and 30 emerging market economies) from the first quarter of 1990 to the fourth quarter of 2015, using both panel fixed effects and the system gener- alized method of moments technique as outlined by Arellano and Bover (1995).

References Aiyagari, S. Rao. 1994. “Uninsured Idiosyncratic Risk and

Aggregate Saving.” Quarterly Journal of Economics 109 (3): 659–84.

Angeles, Luis. 2015. “Credit Expansion and the Economy.” Applied Economics Letters 22 (13): 1064–72.

Arcand, Jean Louis, Enrico Berkes, and Ugo Panizza. 2015. “Too Much Finance?” Journal of Economic Growth 20 (2): 105–48.

Arellano, Manuel, and Olympia Bover. 1995. “Another Look at the Instrumental Variable Estimation of Error-Components Models.” Journal of Econometrics 68 (1): 29–51.

Arregui, Nicolas, Jaromír Beneš, Ivo Krznar, and Srobona Mitra. 2013. “Evaluating the Net Benefits of Macroprudential Pol- icy: A Cookbook.” IMF Working Paper 13/167, International Monetary Fund, Washington, DC.

Badarinza, Cristian, Vimal Balasubramaniam, and Tarun Ramadorai. 2016. “The Indian Household Savings Land- scape.” NCAER Working Paper, National Council of Applied Economic Research, New Delhi, India.

Badarinza, Cristian, John Y. Campbell, and Tarun Ramadorai. 2016. “International Comparative Household Finance.” Annual Review of Economics 8 (1): 111–44.

Banco Central de Chile. 2017. Financial Stability Report: First Half 2017. Santiago, Chile, June.

Bank of Canada. 2015. “Financial System Review.” Ottawa, Ontario, June.

———. 2017. “Financial System Review.” Ottawa, Ontario, June.

Bank of England. 2017. “Financial Stability Report.” Issue 41, June.

Baron, Matthew, and Wei Xiong. 2017. “Credit Expansion and Neglected Crash Risk.” Quarterly Journal of Eco- nomics 713–64.

Barth, James R., Gerard Caprio Jr., and Ross Levine. 2013. “Bank Regulation and Supervision in 180 Countries from 1999 to 2011.” Journal of Financial Economic Policy 5 (2): 111–219.

Beck, Thorsten, Berrak Büyükkarabacak, Felix K. Rioja, and Neven T. Valev. 2012. “Who Gets the Credit? And Does It Matter? Household vs. Firm Lending across Countries.” B.E. Journal of Macroeconomics 12 (1): 1–46.

Beck, Thorsten, and Ross Levine. 2004. “Stock Markets, Banks, and Growth: Panel Evidence.” Journal of Banking and Finance 28 (3): 423–42.

———, and Norman Loayza. 2000. “Finance and Sources of Growth.” Journal of Financial Economics 58 (1–2): 261–300.

Bernanke, Ben S., Mark Gertler, and Simon Gilchrist. 1999. “The Financial Accelerator in a Quantitative Business Cycle Framework.” In Handbook of Macroeconomics, edited by John B. Taylor and Michael Woodford, 1341–93. Amster- dam: Elsevier.

Blundell, Richard, Martin Browning, and Costas Meghir. 1994. “Consumer Demand and the Life-Cycle Allocation of House- hold Expenditures.” Review of Economic Studies 61 (1): 57–80.

Brunnermeier, Markus, Darius Palia, Karthik Sastry, and Christopher Sims. 2017. “Feedbacks: Financial Markets and Economic Activity.” Working Paper.

Brunnermeier, Markus K., and Yuliy Sannikov. 2014. “A Macro- economic Model with a Financial Sector.” American Economic Review 104 (2): 379–421.

Calvo, Guillermo A., and Carmen M. Reinhart. 2000. “When Capital Flows Come to a Sudden Stop: Consequences and Policy.” In Reforming the International Monetary and Financial System, edited by Peter B. Kenen and Alexander K. Swoboda. Washington, DC: International Monetary Fund.

Calza, Alessandro, Tommaso Monacelli, and Livio Stracca. 2013. “Housing Finance and Monetary Policy.” Journal of the Euro- pean Economic Association 11 (1): 101–22.

Carroll, Christopher D. 1992. “The Buffer-Stock Theory of Saving: Some Macroeconomic Evidence.” Brookings Papers on Economic Activity 2, Brookings Institution, Washington, DC.

Cerutti, Eugenio, Ricardo Correa, Elisabetta Fiorentino, and Esther Segalla. 2017. “Changes in Prudential Policy Instruments—A New Cross-Country Database.” International Journal of Central Banking 13 (1): 477–503.

Cheng, Ing-Haw, Sahíl Raina, and Wei Xiong. 2014. “Wall Street and the Housing Bubble.” American Economic Review 104 (9): 2797–829.

Chinn, Menzie D., and Hiro Ito. 2008. “A New Measure of Financial Openness.” Journal of Comparative Policy Analysis 10 (3): 309–22.

Coibion, Olivier, Yuriy Gorodnichenko, Marianna Kudlyak, and John Mondragon. 2017. “Does Greater Inequality Lead to

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

More Household Borrowing? New Evidence from Household Data.” Federal Reserve Bank of Minneapolis Working Paper 17–04, Minneapolis, MN.

Crowe, Christopher, Giovanni Dell’Ariccia, Deniz Igan, and Pau Rabanal. 2013. “How to Deal with Real Estate Booms: Les- sons from Country Experiences.” Journal of Financial Stability 9 (3): 300–19.

De Gregorio, Jose. 1996. “Borrowing Constraints, Human Capital Accumulation, and Growth.” Journal of Monetary Economics 37 (1): 49–71.

Deaton, Angus. 1991. “Saving and Liquidity Constraints.” Econometrica 59 (5): 1221–48.

Demirgüç-Kunt, Asli, and Leora Klapper. 2012. “Measur- ing Financial Inclusion: The Global Findex Database.” Policy Research Working Paper 6025, World Bank, Washington, DC.

De Nicolò, Gianni, Giovanni Favara, and Lev Ratnovski. 2012. “Externalities and Macroprudential Policy.” IMF Staff Discussion Note 12/05, International Monetary Fund, Washington, DC.

Drehmann, Mathias, and Kostas Tsatsaronis. 2014. “The Credit-to-GDP Gap and Countercyclical Capital Buffers: Questions and Answers.” BIS Quarterly Review (March).

Eggertsson, Gauti B., and Paul Krugman. 2012. “Debt, Deleveraging, and the Liquidity Trap: A Fisher-Minsky-Koo Approach.” Quarterly Journal of Economics 127 (3): 1469–513.

Elul, Ronel. 2008. “Collateral, Credit History, and the Financial Decelerator.” Journal of Financial Intermediation 17 (1): 63–88.

Fisher, Irving. 1933. “The Debt-Deflation Theory of Great Depressions.” Econometrica 1 (4): 337–57.

Friedman, Milton. 1957. A Theory of the Consumption Function. Princeton, NJ: Princeton University Press.

Fuster, Andreas, David Laibson, and Brock Mendel. 2010. “Nat- ural Expectations and Macroeconomic Fluctuations.” Journal of Economic Perspectives 24 (4): 67–84.

Gan, Li, Zhichao Yin, Nan Jia, Shu Xu, Shuang Ma, and Lu Zheng. 2013. Data You Need to Know about China. Research Report of China Household Finance Survey. Heidel- berg: Springer.

Geanakoplos, John. 2010. “The Leverage Cycle.” In NBER Macroeconomics Annual 2009, vol. 24, edited by Darron Ace- moglu, Kenneth Rogoff, and Michael Woodford. Cambridge, MA: National Bureau of Economic Research.

Gennaioli, Nicola, Andrei Shleifer, and Robert Vishny. 2012. “Neglected Risks, Financial Innovation, and Financial Fragil- ity.” Journal of Financial Economics 104 (3): 452−68.

Gourinchas, Pierre-Olivier, and Maurice Obstfeld. 2012. “Stories of The Twentieth Century for the Twenty-First.” American Economic Journal: Macroeconomics 4 (1): 226–65.

Hall, Robert E. 1978. “Stochastic Implications of the Life Cycle–Permanent Income Hypothesis: Theory and Evidence.” Journal of Political Economy 86 (6): 971–87.

Hanson, Samuel G., Anil K. Kashyap, and Jeremy C. Stein. 2011. “A Macroprudential Approach to Financial Regula- tion.” Journal of Economic Perspectives 25 (1): 3–28.

Ilzetzki, Ethan, Carmen M. Reinhart, and Kenneth S. Rogoff. 2017. “Exchange Arrangements Entering the 21st Century: Which Anchor Will Hold?” NBER Working Paper 23134, National Bureau of Economic Research, Cambridge, MA.

International Monetary Fund (IMF). 2013. “Key Aspects of Macroprudential Policy—Background Paper.” IMF Policy Paper, Washington, DC.

———. 2016a. “Capital Flows—Review of Experience with the Institutional View.” Washington, DC.

———. 2016b. “Tax Policy, Leverage and Macroeconomic Stability.” Washington, DC.

———. 2017. Canada: Selected Issues and Analytical Notes. IMF Country Report 17/211, Washington, DC.

Jácome, Luis I., and Srobona Mitra. 2015. “LTV and DTI Limits—Going Granular.” IMF Working Paper 15/154, International Monetary Fund, Washington DC.

Jappelli, Tullio, and Marco Pagano. 1994. “Saving, Growth, and Liquidity Constraints.” Quarterly Journal of Economics 109 (1): 83–109.

Jordà, Òscar, Moritz Schularick, and Alan M. Taylor. 2016. “The Great Mortgaging: Housing Finance, Crises and Business Cycles.” Economic Policy 31 (85): 107–52.

Kiyotaki, Nobuhiro, and John Moore. 1997. “Credit Cycles.” Journal of Political Economy 105 (2): 211–48.

Korinek, Anton, and Alp Simsek. 2016. “Liquidity Trap and Excessive Leverage.” American Economic Review 106 (3): 699–738.

Krznar, Ivo, and James Morsink. 2014. “With Great Power Comes Great Responsibility: Macroprudential Tools at Work in Canada.” IMF Working Paper 14/83, International Mone- tary Fund, Washington, DC.

Kumhof, Michael, Romain Rancière, and Pablo Winant. 2015. “Inequality, Leverage, and Crises.” American Economic Review 105 (3): 1217–45.

Kuttner, Kenneth N., and Ilhyock Shim. 2016. “Can Non-Interest Rate Policies Stabilize Housing Markets? Evidence from a Panel of 57 Economies.” Journal of Financial Stability 26: 31–44.

Laeven, Luc, and Fabián Valencia. 2013. “Systemic Banking Crises Database.” IMF Economic Review 61 (2): 225–70.

Laibson, David. 1997. “Golden Eggs and Hyperbolic Discount- ing.” Quarterly Journal of Economics 112 (2): 443–78.

Levine, Ross. 1998. “The Legal Environment, Banks, and Long-Run Economic Growth.” Journal of Money, Credit and Banking 30 (3): 596–613.

Lim, Cheng Hoon, Francesco Columba, Alejo Costa, Piyabha Kongsamut, Akira Otani, Mustafa Saiyid, Torsten Wezel, and Xiaoyong Wu. 2011. “Macroprudential Policy: What Instru- ments and How to Use Them?” IMF Working Paper 11/238, International Monetary Fund, Washington, DC.

C H A P T E R 2 h O u S E h O L d d E B T A N d F I N A N C I A L S T A B I L I T Y

International Monetary Fund | October 2017

Lombardi, Marco, Madhusudan Mohanty, and Ilhyock Shim. 2017. “The Real Effects of Household Debt in the Short and Long Run.” BIS Working Paper 607, Bank for International Settlements, Basel.

Meier, Stephan, and Charles Sprenger. 2010. “Present-Biased Preferences and Credit Card Borrowing.” American Economic Journal: Applied Economics 2 (1): 193–210.

Mian, Atif, Kamalesh Rao, and Amir Sufi. 2013. “Household Balance Sheets, Consumption, and the Economic Slump.” Quarterly Journal of Economics 128 (4): 1687–726.

Mian, Atif, and Amir Sufi. 2011. “House Prices, Home Equity-Based Borrowing, and the U.S. Household Leverage Crisis.” American Economic Review 101 (5): 2132–56.

———. 2014. House of Debt. Chicago: University of Chicago Press.

———, and Emil Verner. Forthcoming. “Household Debt and Business Cycles Worldwide.” Quarterly Journal of Economics.

Mussa, Michael, Giovanni Dell’Ariccia, Barry J. Eichengreen, and Enrica Detragiache. 1998. “Capital Account Liberal- ization: Theoretical and Practical Aspects.” IMF Occasional Paper 172, International Monetary Fund, Washington, DC.

Rajan, Raghuram G. 2010. Fault Lines: How Hidden Fractures Still Threaten the World Economy. Princeton, NJ: Princeton University Press.

Reserve Bank of Australia. 2017. Financial Stability Review, Sydney, April.

Sahay, Ratna, Martin Čihák, Papa N’Diaye, Adolfo Barajas, Srobona Mitra, Annette Kyobe, Yen Nian Mooi, and Seyed Reza Yousefi. 2015a. “Financial Inclusion: Can It Meet Multiple Macroeconomic Goals?” IMF Staff Discussion Note 15/17, International Monetary Fund, Washington DC.

Sahay, Ratna, Martin Čihák, Papa N’Diaye, Adolfo Barajas, Ran Bi, Diana Ayala, Yuan Gao, Annette Kyobe, Lam Nguyen,

Christian Saborowski, Katsiaryna Svirydenka, and Seyed Reza Yousefi. 2015b. “Rethinking Financial Deepening: Stability and Growth in Emerging Markets.” IMF Staff Discussion Note 15/08, International Monetary Fund, Washington DC.

Schmitt-Grohé, Stephanie, and Martín Uribe. 2016. “Down- ward Nominal Wage Rigidity, Currency Pegs, and Invol- untary Unemployment.” Journal of Political Economy 124 (5): 1466–514.

Sheedy, Kevin D. 2014. “Debt and Incomplete Financial Markets: A Case for Nominal GDP Targeting.” Brookings Papers on Economic Activity (Spring): 301–72, Brookings Institution, Washington, DC.

Shiller, Robert J. 2005. Irrational Exuberance, 2nd ed. Princeton, NJ: Princeton University Press.

———. 2014. “Why Is Housing Finance Still Stuck in Such a Primitive Stage?” American Economic Review: Papers & Proceedings 104 (5): 73–76.

South African Reserve Bank. 2017. Financial Stability Review, 1st ed. Pretoria.

Stulz, René M. 1999. “Globalization of Equity Markets and the Cost of Capital.” NBER Working Paper 7021, National Bureau of Economic Research, Cambridge, MA.

Svirydzenka, Katsiaryna. 2016. “Introducing a New Broad-Based Index of Financial Development.” IMF Working Paper 16/5, International Monetary Fund, Washington, DC.

Uribe, Martín, and Stephanie Schmitt-Grohé. 2017. Open Economy Macroeconomics. Princeton, NJ: Princeton Uni- versity Press.

Vandenbussche, Jérôme, Ursula Vogel, and Enrica Detragiache. 2015. “Macroprudential Policies and Housing Prices: A New Database and Empirical Evidence for Central, Eastern, and Southeastern Europe.” Journal of Money, Credit and Banking 47 (S1): 343–77.

Blank

Summary

C hanges in the state of the financial system can provide powerful signals about risks to future economic activity. As in the run-up to the global financial crisis, financial vulnerabilities, understood as the extent to which the adverse impact of shocks on economic activity may be amplified by financial frictions, often increase in buoyant economic conditions when funding is widely available and risks appear sub-

dued. Once these vulnerabilities are sufficiently elevated, they entail significant downside risks for the economy. Thus, tracking the evolution of financial conditions can provide valuable information for policymakers regarding risks to future growth and, hence, a basis for targeted preemptive action.

This chapter develops a new, macroeconomic measure of financial stability by linking financial conditions to the probability distribution of future GDP growth and applying it to a set of major advanced and emerging mar- ket economies.

The analytical approach developed in the chapter can be a significant addition to policymakers’ toolkit for macro-financial surveillance. The chapter shows that changes in financial conditions shift the distribution of future GDP growth. While a widening of risk spreads, rising asset price volatility, and waning global risk appetite are sig- nificant predictors of large macroeconomic downturns in the near term, higher leverage and credit growth provide a more significant signal of increased downside risks to GDP growth over the medium term.

Thus, at the present juncture, low funding costs and financial market volatility support a sanguine view of risks to the global economy in the near term. But the increasing leverage signals potential risks down the road. A sce- nario of rapid decompression in spreads and an increase in financial market volatility could significantly worsen the risk outlook for global growth. These findings underscore the importance of policymakers maintaining heightened vigilance regarding risks to growth during periods of benign financial conditions that may provide a fertile breed- ing ground for the accumulation of financial vulnerabilities.

A retrospective, real-time analysis of the global financial crisis shows that forecasting models augmented with financial conditions would have assigned a considerably higher likelihood to the economic contraction that fol- lowed than those based on recent growth performance alone.

Improvements in predictive ability of severe economic contractions, even over short horizons, can be important for timely monetary and crisis-management policies. The ability to harness longer-horizon information from asset prices and credit aggregates can also help in the design of policy rules to address financial vulnerabilities as they develop. The richness of the results obtained across countries suggests that there is significant scope for policymak- ers to further adapt the approach used in this chapter to specific country conditions including, importantly, to reflect structural changes in financial markets and the real economy.

FINANCIAL CONDITIONS AND GROWTH AT RISK3CHAPTER

91International Monetary Fund | October 2017

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

Introduction The global financial crisis was a powerful reminder

that financial vulnerabilities can increase both the duration and severity of economic recessions. Finan- cial vulnerabilities, understood as the extent to which the adverse impact of shocks on economic activity may be amplified by financial frictions, usually grow in buoyant economic conditions when investment opportunities seem ample, funding conditions are easy, and risk appetite is high. Once these vulnerabilities are sufficiently high, they can entail significant downside risks for the economy.

This interplay between shocks, financial vulnerabil- ities, and growth suggests that financial indicators can provide important intelligence regarding risks to the economic outlook. Policymakers have devoted consid- erable attention to translating the information content of financial indicators into an assessment of financial vulnerability. Approaches that have been used include expert judgment, stress tests, and heatmaps based on multiple early-warning indicators and broad financial conditions indices. These approaches all assess finan- cial vulnerability by linking the state of the financial system to the probability of a financial crisis or bank capital shortage.

Because policymakers care about the whole distri- bution of future GDP growth, linking the state of the financial system to such a distribution would enhance macro-financial surveillance. Policymakers would then be able to specify bad outcomes in terms of their risk preferences. For example, it would be possible to calculate the likelihood of output growth being below a given level and to identify thresholds for financial indicators, such as leverage, that signal heightened tail risks to growth.

This chapter develops a new analytical tool that maps financial conditions into the probability distribution of future GDP growth. In this chapter, financial conditions correspond to combinations of key domestic financial market asset returns, funding spreads, and volatility; domestic credit aggregates;

Prepared by a staff team consisting of Jay Surti (team leader), Mitsuru Katagiri, Romain Lafarguette, Sheheryar Malik, and Dulani Seneviratne, with contributions from Vladimir Pillonca, Aquiles Farias, André Leitão Botelho, Kei Moriya, and Changchun Wang, under the general guidance of Claudio Raddatz and Dong He. The chapter team has benefited from discussions with Norman Swanson, Nellie Liang, and Domenico Giannone. Claudia Cohen and Breanne Rajkumar provided editorial assistance.

and external conditions such as measures of global risk sentiment. The methodological approach extends a nascent literature that derives a direct empirical link between financial conditions and risks to the real economy and applies it to 21 major advanced and emerging market economies over the near and medium term.

The chapter examines how financial conditions provide information regarding risks to future eco- nomic growth across countries and time horizons. In advanced economies, there may be a stronger associa- tion between financial variables and future economic activity than in emerging market economies because more economic risks are traded in deeper financial markets. But, in both cases, asset prices may remain buoyant until shortly before risks materialize, as the run-up to the global financial crisis showed. Thus, incorporating information on credit aggregates such as leverage into measures of financial conditions may improve forecasts of risks to growth, especially over the medium term.

The chapter addresses the following specific questions: • Do changes in financial conditions signal risks to

future GDP growth? Are they equally informative for advanced and emerging market economies, about the intensity of recessions and the strength of booms, and over different time horizons?

• What types of financial variables are more informa- tive regarding the risks to growth at different time horizons and in different countries?

• Could we have used financial conditions to shed light on the likelihood of extremely negative growth outcomes of the past, such as the global recession following the bankruptcy of Leh- man Brothers?

• How can policymakers make use of this new tool of macro-financial surveillance?

The main findings are as follows: • Changes in a country’s financial conditions shift

the distribution of future GDP growth in both advanced and emerging market economies. A tight- ening of financial conditions, reflected in a decom- pression in spreads or an increase in asset price volatility, is a significant predictor of large macro- economic downturns within a one-year horizon. Moreover, in emerging market economies, tighter financial conditions could also portend stronger booms over the subsequent four quarters, possibly because of procyclical capital flows.

C H A P T E R 3 F I N A N C I A L C O N d I T I O N S A N d G R O w T h A T R I S k

International Monetary Fund | October 2017

• Asset prices are most informative about risks to growth in the short term, whereas credit aggregates provide more information over longer time horizons. A rising cost of funding and falling asset prices signal a greater threat of severe recession at time horizons of up to four quarters. Higher leverage signals increased downside risk to growth at horizons between one and three years.

• Movements in commodity prices and exchange rates affect the real economy in a significant, albeit complex, manner, making a simple economic inter- pretation of their predictive content challenging. On the other hand, a souring of global risk sentiment increases downside risks to growth at short time horizons of one quarter.

• In addition to these common patterns, there is heterogeneity in the information content of financial conditions for growth risks across countries. For example, while asset prices are no longer informative over horizons longer than a year for advanced econ- omies, they remain so for emerging markets.

• A retrospective real-time analysis of the global finan- cial crisis shows that forecasting models augmented by financial conditions would have assigned a much higher likelihood to the economic contraction that followed than those based on recent growth per- formance alone.

The chapter’s approach to linking financial con- ditions and risks to growth can help policymakers in numerous ways. The findings underscore the importance of policymakers maintaining heightened vigilance regarding risks to growth during periods of benign financial conditions that may provide a fertile breeding ground for the accumulation of financial vulnerabilities. Policymakers may respond to signals of an imminent near-term dire economic outcome with crisis-management-type discretionary policy actions that encompass a range of monetary and macropruden- tial tools. More broadly, this also helps in the design of policy rules to address financial vulnerabilities as they develop through the introduction of appropriate countercyclical macroprudential tools. In this regard, the output of the forecasting models could be used to calibrate parameters of structural macro-financial models used to guide such policy.1 The richness of the

1Just as estimated vector autoregression models have been used to calibrate the parameters of linear dynamic general equilibrium models used to pin down optimal monetary policy rules (for example, Christiano, Eichenbaum, and Evans 2005; Del Negro and Schorfheide 2009).

results obtained across countries suggests that there is significant scope for authorities to further adapt the broad approach used in this chapter to specific country conditions, including, importantly, to reflect structural changes in financial markets and the real economy.

The rest of this chapter is organized as follows. The next section discusses conceptual issues related to the links between macro-financial conditions, financial vulnerabilities, and risks to the outlook for economic growth. The subsequent section looks at how asset prices and financial aggregates combine to signal short- to medium-term risks to future GDP growth. The section after that provides an empirical assessment of the degree to which the information contained in measures of financial conditions can help forecast risks to economic growth in major advanced and emerging market economies over horizons up to one year. The final section discusses policy implications. Annexes explain the potential policy applications, construc- tion of financial conditions, and modeling of risks to growth in more detail.

Financial Conditions and Risks to Growth: Conceptual Issues

Economic growth has a complex and nonlinear relationship with shocks and financial vulnerabilities. Theory and recent experience both support the view that financial vulnerabilities increase risks to growth.2 When investment opportunities seem abundant and the means of financing them are easily and cheaply available, financial vulnerabilities tend to increase. Once such vulnerabilities are sufficiently high, they can amplify and prolong the impact of shocks on economic activity. GDP growth responds nonlinearly to shocks in the presence of financial vulnerabilities, which increases the likelihood of severely negative economic outcomes.3 Under such circumstances, assessments of both the baseline growth outlook and the risks to such an outlook are informed not only by the span and severity of relevant risk factors that are the source of shocks, but also by the intelligence provided by the interplay of factors that increase financial vulnerability.

2Empirical evidence shows that recessions accompanied by financial crises are typically much more severe and protracted than ordinary recessions (Claessens, Kose, and Terrones 2011a, 2011b).

3Annex 3.1 provides a framework for understanding the joint dynamics of financial vulnerabilities and growth risks in a structural macro model.

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

Several factors cause financial vulnerabilities to grow in a buoyant macro-financial environment. Ease of borrowing and high asset prices reduce the incentives to manage liquidity and solvency risks. Perceptions of high investment returns relative to the cost of funding and of the improved quality of collateral incentivize households and firms to increase their leverage without taking into account the potential negative externali- ties resulting from their collective borrowing decisions (Bianchi 2011; Korinek and Simsek 2016; Bianchi and Mendoza, forthcoming). Booming asset prices also boost the capital adequacy, lending capacity, and risk appetite of financial intermediaries (Brunnermeier and Pedersen 2009; Adrian, Moench, and Shin 2010; Adrian and Shin 2014). As intermediaries respond by increas- ing short-term wholesale funding to finance long-term credit exposures, maturity mismatches and other balance sheet weaknesses accumulate in the financial sector. For example, lenders’ incentives to invest in costly under- writing are reduced, which can result in significant mispricing of credit risk (Gorton and Ordoñez 2014).

The need to lower significant debt and correct balance sheet mismatches can clog financial interme- diation, investment, and growth for a long time once the credit cycle turns. With vulnerabilities substan- tially elevated, even small negative shocks can cause significant reversals because they force lenders to face up to the true quality of exposures and collateral. This results in a significant tightening in credit conditions. Some firms and households may be forced into default, while others may have to liquidate assets. The ensuing pressure on lenders’ profits and collateral values can then generate further rounds of contraction in credit, investment, and growth. In addition to the direct nega- tive impact of these events on lenders’ profits, rising volatility and risk spreads constrain lenders’ capacity to bear risk by increasing the capital required as a buffer against existing exposures (He and Krishnamurthy 2013; Brunnermeier and Sannikov 2014). In such cir- cumstances, risk-bearing capacity will be affected not only by capital constraints but also by funding liquid- ity concerns (Gertler, Kiyotaki, and Prestipino 2017).

A large body of empirical work has examined the information content of asset prices in forecasting the baseline growth outlook.4 Various asset prices have been found to be useful predictors of future output growth in

4Stock and Watson (2003) produce a comprehensive survey of the literature up to the early 2000s.

some countries and in some periods. Combining fore- casts obtained from models with individual asset prices appears to result in more consistent, higher-quality fore- casts. Short-term yields on risk-free securities and term spreads capture the stance of monetary policy and there- fore contain useful information about future economic activity (Laurent 1988; Estrella and Hardouvelis 1991; Bernanke and Blinder 1992; Estrella and Mishkin 1998; Ang, Piazzesi, and Wei 2006). Corporate bond spreads signal changes in the default-adjusted marginal return on business fixed investment (Philippon 2009) and shocks to the profitability and creditworthiness of financial intermediaries (Gilchrist and Zakrajšek 2012).5 There is some evidence that elevated stock-return volatility can be a useful predictor of output contraction over short horizons (Campbell and others 2001), although empiri- cal evidence for the predictive content of stock returns is weak (Campbell 1999; Stock and Watson 2003).

The key departure of this chapter is to focus on the information content of financial indicators in forecast- ing risks to growth. In addition to asset prices, credit aggregates can also be expected to provide information on the risks to growth in the short, medium, and long term. For example, a combination of low leverage and buoyant asset prices is likely to correspond, over the short term, to high expected growth (an optimistic baseline outlook) and a low likelihood of adverse out- comes (sanguine risk outlook as represented, poten- tially, by a probability density of short-term growth with relatively low variance). On the other hand, theory suggests that such an environment might be ideal for a buildup of vulnerabilities over the medium term, ultimately increasing the likelihood of low growth outcomes. As such a possibility becomes more certain, spreads and market volatility would rise and asset prices would fall.6 Other financial variables can

5Gilchrist and Zakrajšek (2012) demonstrate the superiority of their constructed bond spread over alternative proxies for the default spread investigated in the earlier literature; for example, the Baa-Aaa bond spread (Bernanke 1983), the commercial paper–Treasury bill spread (Stock and Watson 1989; Friedman and Kuttner 1998), and the so-called junk bond spread (Gertler and Lown 1999).

6Financial indicators can be classified into two types. Fast-moving asset prices tend to signal risks to growth over the near term, whereas balance sheet aggregates change gradually over time and may indicate risks over longer horizons. The evolution of aggregates and prices is not by any means independent. For example, the growth in aggregates may, beyond a point, change market expectations of risks. This would be reflected in tightening spreads, which then signal risks to growth in the near term. For a discussion, see Adrian and Liang 2016 and Krishnamurthy and Muir 2016.

C H A P T E R 3 F I N A N C I A L C O N d I T I O N S A N d G R O w T h A T R I S k

International Monetary Fund | October 2017

also be very informative in the context of small open advanced economies and emerging market economies. These variables include the nominal exchange rate and commodity prices, which may affect the cost of external funding and the availability of international collateral (Caballero and Krishnamurthy 2006).

This chapter refers to such a combination of finan- cial indicators, or an index constituted of them, as financial conditions. The term “financial conditions” often refers to the ease of funding (Chapter 3 of the April 2017 Global Financial Stability Report [GFSR]), but here it is used to refer to the combination of a broad set of financial variables that influence economic behavior and thereby the future of the economy.7

This chapter examines two alternative approaches to constructing measures of financial conditions from the information contained in several financial indicators. One attractive option is a single financial conditions index (FCI). An important advantage of such a univar- iate FCI is the parsimony with which it aggregates the information content of multiple financial indicators. Parsimony is highly desirable for forecasting because it reduces parameter uncertainty, but it may lead to suppressing the information provided by certain variables by commingling them with other, more volatile indicators in a single index. For example, the higher variability of asset prices and risk spreads may lead them to dominate univariate FCIs, with credit aggregates being assigned small factor loadings (as is indeed the case in the application described below). Since credit aggregates may carry significant infor- mation about risks to growth at longer horizons, the chapter pursues a second approach wherein financial indicators are partitioned into three separate groups based on economic similarity. The three subindices are the domestic price of risk (risk spreads, asset returns, and price volatility), credit aggregates (leverage and credit growth), and external conditions (global risk sentiment, commodity prices, and exchange rates). The separation of a large set of financial indicators into these three predetermined categories is a reasonable compromise between maintaining parsimony, allowing various classes of indicators to provide separate signals about risks to growth at different horizons, and being able to provide a more direct economic interpretation of the various subindices.

7This notion of financial conditions is similar to the definition proposed by Hatzius and others (2010). See Annex 3.2 for details on the construction of financial conditions used in this chapter.

The chapter’s empirical framework is centered on forecasts of the probability distribution of future growth outcomes based on financial conditions in a way that allows for nonlinearity and state dependence. Building on the literature on conditional density fore- casting and recent research on forecasting the distribu- tion of growth in the United States, the chapter uses financial conditions to forecast the probability distri- bution of future GDP growth in major advanced and emerging market economies for horizons of up to three years through quantile projections.8 The flexibility of this approach captures the rich nonlinear interaction between shocks, financial vulnerabilities, and economic outcomes predicted by theory. For instance, consider two combinations of financial indicators that forecast the same future median growth rate. The first combi- nation forecasts much greater downside growth risk (that is, a probability density with a significantly fatter left tail) than the second. This indicates that for a con- stant distribution of fundamental shocks, the economy is more likely to experience a very bad economic out- come in the future under the first configuration than under the second. In this sense, the first combination signals a financial system that is more vulnerable. These density forecasts can subsequently be exploited to con- struct measures of risks to economic growth associated with the state of the financial system.

Such an approach provides a natural way of assessing financial vulnerability that has several distinct advan- tages. First, the estimated link between financial condi- tions and the distribution of future economic activity would provide a close measure of financial vulnera- bility, understood as the extent to which the financial system amplifies shocks. Second, to the extent that pol- icymakers care about the whole distribution of future GDP growth, it provides a complete depiction of the risks to economic activity associated with the state of the financial system. Third, it allows policymakers to define risk tolerance in terms of GDP growth, which is more general than in terms of the probability of a financial crisis as defined under specific criteria or another ad hoc metric. For instance, this approach gives precise answers to questions such as the probabil-

8See Annex 3.3 for details on the empirical framework. Con- ditional density forecasting is surveyed by Tay and Wallis (2000); Corradi and Swanson (2006); and Komunjer (2013). The chapter’s methodology builds on some recent studies (Adrian, Boyarchenko, and Giannone 2016; De Nicolò and Lucchetta 2017) that establish a direct empirical link between financial conditions and risks to economic growth.

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

ity of GDP growth being less than –3 percent one year ahead given the current—or any hypothetical—state of the financial system.

How Do Changes in Financial Conditions Indicate Risks to Growth? Over a horizon of one to four quarters, tighter finan- cial conditions—as reflected in higher univariate FCIs—predict increased downside risks to GDP growth in most advanced economies and a more uncertain growth outlook in several emerging market economies. An increasing domestic price of risk signals an elevated threat of imminent, severe recession in advanced and emerging market economies. Rising leverage is a sig- nificant predictor of elevated downside risk over the medium term. Country-specific results vary considerably, suggesting a rich interplay of the drivers of growth risk.

What Underpins Economies’ Financial Conditions Indices?9

The drivers of economies’ FCIs vary considerably across a sample of major advanced and emerging mar- ket economies.10 An increase in the FCI corresponds to tighter financial conditions, that is, higher spreads and volatility, lower asset prices, worsening risk sentiment, exchange rate depreciation, and unfavorable commod- ity price movements. Beyond this common finding, the relative importance of these factors in determining the evolution of FCIs varies considerably across coun- tries. Higher corporate funding costs and worsening global risk sentiment (as captured by rising levels of the Chicago Board Options Exchange Volatility Index [VIX] and Merrill Lynch Option Volatility Estimate [MOVE] Index) tighten financial conditions across the board. But while sovereign spreads are clearly import- ant in emerging market economies, they are rarely so in advanced economies. And while increasing com- modity prices loosen financial conditions in exporters such as Australia, Brazil, Canada, Chile, and Russia, they tighten them in commodity-importing countries. Exchange rate appreciation uniformly loosens financial

9In this subsection, financial conditions reference the univariate FCIs described in the preceding section.

10The financial indicators that constitute a country’s FCIs may evolve over time for many reasons, including changes in risk appetite or investor risk sentiment. The methodology used to construct the FCIs, the list of financial indicators, and the sample of countries are described in detail in Annex 3.2.

conditions.11 In the case of emerging market and small open economies, this may reflect the correspondence of an appreciating exchange rate with strong capital inflows. In general, asset price shocks appear to be more important in driving changes in FCIs than credit aggregates. This pattern, however, may reflect the slower speed at which credit adjusts relative to changes in GDP at turning points in the economic cycle, especially at the end of economic booms preceding financial crises.

What Information Do Univariate FCIs Convey about Future Growth?

An increase in the FCI would signal higher down- side risks in both advanced and emerging market economies. An increase in the global FCI signals heightened downside risk to world GDP growth (Figure 3.1).12 Movements in the FCI are especially powerful signals of changes in downside tail risk to the global economy but are less informative about the baseline growth outlook and the strength of economic booms. This is reflected in the fact that the forecast of the left tail of the distribution of global GDP growth decreases significantly in response to an increase in the FCI both one quarter and four quarters ahead. In contrast, the forecasts of the central tendency of GDP growth (as captured by the median growth rate) and of the strength of booms (at the right tail of the growth distribution forecasts) are considerably less responsive to changes in the FCI, and their movement is apparent only for large changes in the FCI such as those observed in the global financial crisis. This is also the case for individual countries—the forecasts of the worst-case outcomes (at the 5th percentile of the future GDP growth distribution) are between 3 times (United States) and more than 10 times (Australia) more sensi-

11Exchange rate movements may reflect a complex combination of factors. With respect to a country’s FCI, changes in the exchange rate are most likely to be associated with changes in the ease of exter- nal financing conditions, which may relate either to evolving global funding conditions and risk sentiment or changes in the market’s perception of the country’s creditworthiness or both. Exchange rate depreciations are, in such an association, a reflection of a worsening of global conditions or in market perceptions of a country’s risk profile. Empirically, such an association appears to apply to most countries covered in the chapter, although the link has been noted in the literature as relevant primarily for emerging market economies.

12The global FCI is defined as the first principal component of the country-level FCIs.

C H A P T E R 3 F I N A N C I A L C O N d I T I O N S A N d G R O w T h A T R I S k

International Monetary Fund | October 2017

tive to changes in FCIs than the forecasts of the central tendency of economic growth.

Easing of global financial conditions through 2016 signaled reduced tail risk to global growth for 2017. This is evident in the upward movement in the bottom tail of the GDP growth density forecast (5th percen- tile) for the world economy (Figure 3.1) and a similar movement in several countries, including Australia, Brazil, South Africa, Turkey, and the United States (Figure 3.2).13

Nonetheless, FCIs do carry significant information regarding upside risks to future economic growth for emerging markets (Figure 3.3). In Brazil, Korea, and Mexico, higher levels of the FCI portend a more uncertain growth outlook at a one-year horizon, as reflected in coefficients of opposite signs at the lowest and highest quantiles (which imply fatter and longer tails at both ends of the distribution of future GDP growth). In some commodity-exporting countries, such as Chile, tightening FCIs appear to signal risk of stronger recessions as well as economic booms of lower intensity (Figure 3.3, panel 2).

Different properties of advanced and emerging market economy business cycles may account for the differing significance of the information provided by changing FCIs across countries. Some emerging market economies and commodity exporters may have a more pronounced and symmetrical boom-bust cycle that is closely tied to export-commodity prices and global risk sentiment. Positive developments in either factor can motivate significant capital inflows, relaxing domestic financial constraints on growth.14 When the risk envi- ronment reverses, capital flows may retrench, exchange rates can depreciate, and investment and growth can decline (Aguiar and Gopinath 2007). This may explain why a tightening of financial conditions can move the density of GDP growth to the left (Figure 3.3, panel 2). More broadly, increases in FCIs in emerging market economies may reflect domestic interest rate hikes targeted at attenuating overheating due to high domes- tic demand. But the higher interest rates may attract

13The exact magnitude of the movements can be improved by further country-specific calibration that, for instance, increases the number of financial indicators used in FCI construction, but the direction of the movements indicated by the model is quite robust and showcases the potential of this methodology.

14For the role of commodity prices in explaining the cyclical movements of capital flows to emerging market economies, see, for example, Chapter 4 of the April 2017 Regional Economic Outlook for the Western Hemisphere.

–2.0 –1.8 –1.6 –1.4 –1.2 –1.0 –0.8 –0.6 –0.4 –0.2

0.0

5th 25th 50th 75th 95th Percentile

–10

–20

–15

–10

–5

19 91

:Q 1

93 :Q

95 :Q

97 :Q

99 :Q

20 01

:Q 1

03 :Q

05 :Q

07 :Q

09 :Q

11 :Q

13 :Q

15 :Q

Downside and upside risks (5th and 95th percentiles) Median FCI (right scale)

16 :Q

Sources: Bloomberg Finance L.P.; Haver Analytics; IMF, Global Data Source and World Economic Outlook databases; Thomson Reuters Datastream; and IMF staff estimates. Note: Panel 1 depicts the estimated coefficients on the current quarter FCI in a quantile regression of four-quarters-ahead GDP growth on current quarter FCI and GDP growth. Panel 2 depicts the time series of estimated, conditional 5th, 50th, and 95th quantiles of four-quarters-ahead GDP growth. The median (red) line denotes the forecast of the 50th quantile of GDP growth made four quarters earlier using the methodology described in Annex 3.3. The shaded area is bound at the top and bottom by, respectively, the forecasts of the 95th and 5th quantiles of GDP growth made four quarters earlier. FCI = financial conditions index.

Figure 3.1. Tighter Financial Conditions Forecast Greater Downside Tail Risk to Global Growth

1. Quantile Coefficient Estimates (Standard deviations)

As financial conditions tighten, the probability of a large economic contraction increases ...

2. One-Year-Ahead Density Forecast (Left scale = percent; right scale = standard deviations)

... as was seen in the recent global financial and euro area sovereign debt crises.

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

–9

–6

–3

–15

–10

–5

19 91

:Q 1

93 :Q

95 :Q

97 :Q

99 :Q

20 01

:Q 1

03 :Q

05 :Q

07 :Q

09 :Q

11 :Q

13 :Q

15 :Q

16 :Q

4 –10

–5

–6

–4

–2

19 73

:Q 1

75 :Q

1 77

:Q 1

79 :Q

1 81

:Q 1

83 :Q

1 85

:Q 1

87 :Q

1 89

:Q 1

91 :Q

1 93

:Q 1

95 :Q

1 97

:Q 1

99 :Q

1 20

01 :Q

1 03

:Q 1

05 :Q

1 07

:Q 1

09 :Q

1 11

:Q 1

13 :Q

1 15

:Q 1

16 :Q

–5

–8

–4

19 91

:Q 1

93 :Q

95 :Q

97 :Q

99 :Q

20 01

:Q 1

03 :Q

05 :Q

07 :Q

09 :Q

11 :Q

13 :Q

15 :Q

16 :Q

–8

–4

–8

–4

19 81

:Q 1

83 :Q

1 85

:Q 1

87 :Q

1 89

:Q 1

91 :Q

1 93

:Q 1

95 :Q

1 97

:Q 1

99 :Q

1 20

01 :Q

1 03

:Q 1

05 :Q

1 07

:Q 1

09 :Q

1 11

:Q 1

13 :Q

1 15

:Q 1

16 :Q

–4

–20

–10

19 91

:Q 1

93 :Q

95 :Q

97 :Q

99 :Q

20 01

:Q 1

03 :Q

05 :Q

07 :Q

09 :Q

11 :Q

13 :Q

15 :Q

16 :Q

–18

–12

–6

–10

–5

19 73

:Q 1

75 :Q

1 77

:Q 1

79 :Q

1 81

:Q 1

83 :Q

1 85

:Q 1

87 :Q

1 89

:Q 1

91 :Q

1 93

:Q 1

95 :Q

1 97

:Q 1

99 :Q

1 20

01 :Q

1 03

:Q 1

05 :Q

1 07

:Q 1

09 :Q

1 11

:Q 1

13 :Q

1 15

:Q 1

16 :Q

Downside and upside risks (5th and 95th percentiles) Median FCI (right scale)

Sources: Bloomberg Finance L.P.; Haver Analytics; IMF, Global Data Source and World Economic Outlook databases; Thomson Reuters Datastream; and IMF staff estimates. Note: The country-specific financial conditions indices (FCIs) are constructed using the methodology described in Annex 3.2. The median (red) line at each point in time denotes the forecast of the 50th quantile of GDP growth made four quarters earlier using the methodology described in Annex 3.3. The shaded area is bound at the top and bottom by, respectively, the forecasts of the 95th and 5th quantiles of GDP growth made four quarters earlier.

Figure 3.2. Risk of Severe Recessions Is Especially Sensitive to a Tightening of Financial Conditions in Major Advanced and Emerging Market Economies (One-year-ahead density forecasts; left scale = percent; right scale = standard deviations)

1. Brazil 2. Australia

3. South Africa 4. Sweden

5. Turkey 6. United States

C H A P T E R 3 F I N A N C I A L C O N d I T I O N S A N d G R O w T h A T R I S k

International Monetary Fund | October 2017

capital inflows and thereby extend ongoing credit and economic booms. This may explain why tightening of financial conditions appears to be a good indicator of growing positive and negative risks around the baseline (Figure 3.3, panels 1, 3–4).

Which Asset Prices and Aggregates Best Signal Growth Risks at Various Time Horizons?

Asset prices are differentially informative regarding the domestic price of risk across countries. Term and interbank spreads, followed by corporate and sovereign spreads, are the most important risk indicators for the investment and growth outlook across advanced economies. The dynamics of house prices are particu-

larly important in countries where either the share of homeownership and floating-rate mortgages is high (such as the United Kingdom) or the mortgage market is a key node that underpins pricing and activity in systemic funding markets (as in the United States). The evidence for emerging market economies is more challenging to interpret for two reasons. First, data are much more limited and are available only for more recent years. Second, in many countries, financial mar- ket activity is often focused on equity and government bond markets. Unsurprisingly, therefore, analysis of available data suggests that for these countries, sover- eign spreads and equity returns are most significant.

Domestic asset prices are the dominant driver of growth risks in the short term, while credit aggregates

–0.8

–0.6

–0.4

–0.2

0.0

0.2

0.4

0.6

–1.6

–1.4

–1.2

–1.0

–0.8

–0.6

–0.4

–0.2

0.0

5th 25th 50th 75th 95th Percentile

–2.5

–2.0

–1.5

–1.0

–0.5

0.0

0.5

1.0

5th 25th 50th 75th 95th Percentile

–1.0

–0.5

0.0

0.5

1.0

1.5

5th 25th 50th 75th 95th Percentile

Sources: Bloomberg Finance L.P.; Haver Analytics; IMF, Global Data Source and World Economic Outlook databases; Thomson Reuters Datastream; and IMF staff estimates. Note: The panels depict estimated coefficients on the current quarter financial conditions index (FCI) from quantile regressions of four-quarters-ahead GDP growth on current quarter FCI and GDP growth. The coefficients are standardized to depict the impact of a one standard deviation increase in current quarter FCIs on four-quarters-ahead GDP growth (also expressed in standard deviations). 1In line with Morgan Stanley Capital International (MSCI) markets classification criteria, Korea is classified as an emerging market economy in panel 4.

Figure 3.3. In Emerging Market Economies, Changes in Financial Conditions Also Affect Upside Risks (Quantile regression estimates for selected emerging market economies: four quarters ahead)

1. Brazil 2. Chile

3. Mexico 4. Korea1

100

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

are the dominant drivers in the medium term. Results from a panel quantile regression with country fixed effects, estimated separately for advanced and emerging market economies, highlight some common patterns in the relationship between these FCI components and risks to growth. • Domestic price of risk: Tightening of financial condi-

tions caused by a rising price of risk is a significant predictor of downside growth risks over horizons of up to one year. This inverse relationship between the price of risk and the growth forecast is stronger in the left tail of the distribution of future growth and is more significant for advanced economies (Figure 3.4, panels 1–4). The price of risk becomes uninformative over longer horizons in advanced economies. In emerging market economies, an inter- esting pattern arises—a higher price of risk signals lower downside (tail) risks at two- to three-year horizons. One possible explanation is the negative impact of tighter domestic financial conditions on leverage and balance sheet expansion, which appears to be associated with lower risks to growth in both the short and medium term (Figure 3.4, panels 5–6).

• Leverage: Higher credit growth and credit to GDP signal greater downside risk to growth at horizons of one year and longer. The relationship is eco- nomically more significant at the lower quantiles of GDP growth and in advanced economies than in emerging market economies (Figure 3.5, panels 1–2). Over shorter time horizons (one quarter), however, the information content differs across countries, with rising leverage continuing to signal higher downside risks in emerging market and large advanced economies but signaling lower downside risks in small open advanced economies.

• External conditions: While changing external con- ditions convey statistically significant information regarding risks to future growth, their informa- tion content represents a complex combination of forces. For example, movements in exchange rates can reflect different risk implications through real and financial channels, each of which may be more potent at different horizons. And the impact of changes in commodity prices on risks to growth will differ depending on whether a country is a commodity exporter or importer. Consequently, the signal given by changes in external conditions proved difficult to interpret in a straightforward

manner. Nonetheless, a clearer interpretation arises when isolating changes in global risk sentiment from the other external variables.15 Higher global risk aversion, reflected in a higher VIX, signals greater downside risks to growth in the short term, includ- ing a larger threat of an imminent recession (Fig- ure 3.6). However, increases in the VIX also signal lower downside risks to growth at longer horizons of one to two years, possibly because, in most cases, tighter global financial conditions slow the growth of leverage and balance sheet mismatches, which may lessen medium-term growth risks.

The view that emerges from these results is that the prevailing low funding costs and financial market volatility support a positive view of risks to the global economy in the short term, but increasing lever- age signals potential risks down the road. In such circumstances, a scenario of a rapid decompression in spreads and increase in financial market volatil- ity could significantly worsen the risk outlook for global growth.

How Well Do Changes in Financial Conditions Forecast Downside Risks to Growth? Severely adverse growth performance during the global financial crisis is used to demonstrate the potential use of measures of financial conditions in improv- ing forecasts of risks to growth at horizons of up to one year. Augmenting growth forecast models based on past growth performance with financial condi- tions significantly improves forecasting ability. This is reflected in the greater likelihood that is assigned to the actual negative growth outcomes during that period.

Applying the univariate FCIs to historical episodes highlights the index’s power to help predict future economic downturns over short horizons. Notably, the model was used to predict the distribution of growth for the first quarter of 2009, broadly corresponding to the peak of the global financial crisis. • At a one-quarter horizon (that is, in the fourth

quarter of 2008), conditioning the risk forecast of future growth on financial conditions (besides economic growth) adds significantly to capturing

15Formally, a separate model of the kind described in Annex 3.2 was examined with the external conditions subindex defined as a global risk sentiment index (equal to the change in the VIX).

101

C H A P T E R 3 F I N A N C I A L C O N d I T I O N S A N d G R O w T h A T R I S k

International Monetary Fund | October 2017

–0.4

–0.3

–0.2

–0.1

0.0

0.1

0.2

0.10 0.20 0.25 0.40 0.50 0.60 0.75 0.80 0.90

Quantile

–0.4

–0.3

–0.2

–0.1

0.0

0.1

0.2

–0.4

–0.3

–0.2

–0.1

0.0

0.1

0.2

0.10 0.20 0.25 0.40 0.50 0.60 0.75 0.80 0.90

Quantile

–0.4

–0.3

–0.2

–0.1

0.0

0.1

0.2

0.10 0.20 0.25 0.40 0.50 0.60 0.75 0.80 0.90

Quantile

0.10 0.20 0.25 0.40 0.50 0.60 0.75 0.80 0.90

Quantile

–0.1

0.0

0.1

0.2

0.10 0.20 0.25 0.40 0.50 0.60 0.75 0.80 0.90

Quantile

–0.1

0.0

0.1

0.2

0.10 0.20 0.25 0.40 0.50 0.60 0.75 0.80 0.90

Quantile

Sources: Bloomberg Finance L.P.; Haver Analytics; IMF, Global Data Source and World Economic Outlook databases; Thomson Reuters Datastream; and IMF staff estimates. Note: The panels depict coefficient estimates on the price of risk index in pooled quantile regressions of one-quarter-ahead, four-quarters-ahead, and eight-quarters- ahead GDP growth for advanced economies (left column) and emerging market economies (right column). The coefficients are standardized by centering and reducing (zero mean, unit variance) both the dependent variable and the regressors to enable comparison across quantiles, across time horizons, and between advanced and emerging market economies. The coefficient estimate for a given quantile should be read as the impact of a one standard deviation change in the price of risk on the future quantile of GDP growth also expressed in terms of standard deviations. The vertical lines in the green bars denote confidence intervals at 10 percent and, where they cross the x-axis, correspond to absence of statistical significance of the regressor.

Figure 3.4. Higher Price of Risk Is a Significant Predictor of Downside Growth Risks within One Year (Quantile regression coefficients)

1. Advanced Economies: One Quarter Ahead

Economic significance is highest over one quarter ...

2. Emerging Market Economies: One Quarter Ahead

... albeit less so in emerging market economies.

3. Advanced Economies: One Year Ahead

It remains so over one year in advanced economies ...

4. Emerging Market Economies: One Year Ahead

... and in emerging market economies.

5. Advanced Economies: Two Years Ahead

Price of risk becomes uninformative over longer horizons in advanced economies ...

6. Emerging Market Economies: Two Years Ahead

... but, in emerging market economies, higher funding costs signal lower risk over longer horizons.

102

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

imminent tail risks to growth, both at the epicenter of the crisis (that is, the United States) and in a commodity-exporting emerging market economy (Chile). Notably, the likelihood attached to poor growth outcomes around the actual realization is significantly higher if rapidly tightening financial conditions are incorporated into the growth forecast (the density in red) as opposed to a model whose only information for forecasting is the growth

outcome (the density in blue) in the fourth quarter of 2008 (Figure 3.7).16

16GDP growth exhibits a high degree of persistence in the sample of advanced and emerging market economies covered by this chap- ter’s analysis. Consequently, from a forecasting perspective, a quantile autoregression model of GDP growth represents a conservative and hard-to-beat benchmark against which to assess the marginal con- ditioning information content of financial conditions. The quantile autoregression model is unlikely to forecast rare (severe) recessions

–0.4

–0.3

–0.2

–0.1

0.0

0.1

0.2

0.10 0.20 0.25 0.40 0.50 0.60 0.75 0.80 0.90 Quantile

–0.4

–0.3

–0.2

–0.1

0.0

0.1

0.2

0.10 0.20 0.25 0.40 0.50 0.60 0.75 0.80 0.90 Quantile

Sources: Bloomberg Finance L.P.; Haver Analytics; IMF, Global Data Source and World Economic Outlook databases; Thomson Reuters Datastream; and IMF staff estimates. Note: The panels depict coefficient estimates on the credit aggregates index in pooled quantile regressions of three-years-ahead GDP growth for advanced and emerging market economies. The coefficients are standardized by centering and reducing (zero mean, unit variance) both the dependent variable and the regressors to enable comparison across quantiles, across time horizons, and between advanced and emerging market economies. The coefficient estimate for a given quantile should be read as the impact of a one standard deviation change in leverage on the future quantile of GDP growth also expressed in terms of standard deviations. The vertical lines in the green bars denote confidence intervals at 10 percent and, where they cross the x-axis, correspond to absence of statistical significance of the regressor.

Figure 3.5. Rising Leverage Signals Higher Downside Growth Risks at Longer Time Horizons (Quantile regression coefficients)

1. Advanced Economies: Three Years Ahead

2. Emerging Market Economies: Three Years Ahead

–0.4

–0.3

–0.2

–0.1

0.0

0.1

0.2

0.10 0.20 0.25 0.40 0.50 0.60 0.75 0.80 0.90 Quantile

–0.4

–0.3

–0.2

–0.1

0.0

0.1

0.2

0.10 0.20 0.25 0.40 0.50 0.60 0.75 0.80 0.90 Quantile

Sources: Bloomberg Finance L.P.; Haver Analytics; IMF, Global Data Source and World Economic Outlook databases; Thomson Reuters Datastream; and IMF staff estimates. Note: The panels depict coefficient estimates on the VIX index in pooled quantile regressions of one-quarter-ahead GDP growth for advanced and emerging market economies. The coefficients are standardized by centering and reducing (zero mean, unit variance) both the dependent variable and the regressors to enable comparison across quantiles, across time horizons, and between advanced and emerging market economies. The coefficient estimate for a given quantile should be read as the impact of a one standard deviation change in the VIX on the future quantile of GDP growth also expressed in terms of standard deviations. The vertical lines in the green bars denote confidence intervals at 10 percent and, where they cross the x-axis, correspond to absence of statistical significance of the regressor. VIX = Chicago Board Options Exchange Volatility Index.

Figure 3.6. Waning Global Risk Appetite Signals Imminent Downside Risks to Growth (Quantile regression coefficients)

1. Advanced Economies: One Quarter Ahead (External conditions = VIX)

2. Emerging Market Economies: One Quarter Ahead (External conditions = VIX)

103

C H A P T E R 3 F I N A N C I A L C O N d I T I O N S A N d G R O w T h A T R I S k

International Monetary Fund | October 2017

• These results remain robust in a broader cross sec- tion of countries. Among countries that experienced a significant growth downturn during the crisis, adding FCIs to an autoregressive growth forecast- ing model significantly increases the conditional likelihood of a GDP growth outcome less than or equal to the actual growth outturn one quarter ahead (Table 3.1).17 In addition to predicting a fat- ter left tail for the growth distribution, the average growth forecasts including FCIs are closer to the actual severe economic contraction experienced by these countries in the first quarter of 2009, and well below the market consensus, which remained rela- tively optimistic even after the collapse of Lehman Brothers (Table 3.2).

The exercise also shows that conditioning on uni- variate FCIs may not work as well at longer horizons. This possibility is evident when comparing the relative predictive ability of the autoregressive growth model with the model augmented with FCIs at one- and four-quarter horizons for the first quarter of 2009. In the case of the global financial crisis, examining the behavior of sampled countries’ FCIs through 2008 is revealing. Close examination shows why the forecast- ing gain differs once the information set is augmented with FCIs at different time horizons. In the first quarter of 2009, GDP growth for most countries was among the worst in their recent economic history. The Lehman Brothers bankruptcy, at the beginning of the fourth quarter of 2008, was the bellwether for a swift and severe deterioration in financial conditions. Risk spreads and market volatility increased steeply, and asset values crashed. The information emanating from FCIs throughout the fourth quarter of 2008 clearly sig- naled potential negative fallout for economic activity. By contrast, economic indicators took additional time to catch up to the actual magnitude of the decline.

and macroeconomic crises well. A good test of the predictive contribution of financial indicators for such growth episodes would be to examine how their addition to the conditioning information set would change the likelihood assigned to the realized (bad) growth outcome at various horizons.

17Results are presented for a selection of advanced and emerging market economies in Tables 3.1–3.3, even though similar results are obtained for other sampled countries that experienced a recession at the time of the global financial crisis. Results for countries that did not experience an economic contraction suggest that the model augmented with FCIs does not generate false alarms—that is, significantly lower conditional probability of a recession at one- and four-quarter forecast horizons.

0.05

0.10

0.00

0.15

0.20

–15.0 –12.5 –10.0 –7.5 –5.0 –2.5 0.0 2.5 5.0 7.5 10.0 12.5 15.0

0.02

0.04

0.06

0.08

0.10

0.12

0.14

–15.0 –12.5 –10.0 –7.5 –5.0 –2.5 0.0 2.5 5.0 7.5 10.0 12.5 15.0

GDP growth (quarter-over-quarter annualized percent change)

One-quarter-ahead conditional forecast density (at 2008:Q4): without FCI

One-quarter-ahead conditional forecast density (at 2008:Q4): with FCI

Realized value

Sources: Bloomberg Finance L.P.; Haver Analytics; IMF, Global Data Source and World Economic Outlook databases; Thomson Reuters Datastream; and IMF staff estimates. Note: The figure displays conditional probability distributions of one-quarter-ahead GDP growth based on a parametric, T-skew density, fitted over quantile regression estimates as described in Annex 3.3. In particular, it includes two conditional distributions of growth based on two forecasting models that use either growth or growth and financial conditions indices (FCIs) to predict future growth (in 2009:Q1). The figure also includes the realized values of GDP growth (black vertical line). Blue density = model with single regressor (one-quarter-lagged GDP growth); red density = model with two regressors (one-quarter-lagged GDP growth and one-quarter-lagged FCI).

Figure 3.7. Probability Densities of GDP Growth for the Depths of the Global Financial Crisis (Probability)

1. United States

Accounting for financial conditions generates a more pessimistic outlook for risks to growth one quarter before 2009:Q1.

2. Chile

104

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

Table 3.1. Forecast of GDP Growth Distribution for the Global Financial Crisis with and without Financial Conditions Indices (Cumulative probability of actual 2009:Q1 growth outturn, percent)

Selected Advanced Economies Selected Emerging Market Economies

Real-time FCI

Augmented FCI

Augmented Autoregressive

Real-time FCI

Augmented FCI

Augmented Autoregressive Germany Brazil

One quarter ahead for 2009:Q1

5.4 2.4 0.0 One quarter ahead for 2009:Q1

35.5 39.6 7.5

Four quarters ahead for 2009:Q1

0.1 0.4 0.0 Four quarters ahead for 2009:Q1

4.2 5.0 5.5

Sweden Chile One quarter ahead

for 2009:Q1 6.5 5.9 4.8 One quarter ahead

for 2009:Q1 6.4 8.0 2.6

Four quarters ahead for 2009:Q1

0.0 0.8 0.5 Four quarters ahead for 2009:Q1

4.0 1.7 2.0

United Kingdom South Africa One quarter ahead

for 2009:Q1 29.8 29.5 5.8 One quarter ahead

for 2009:Q1 7.2 4.6 0.8

Four quarters ahead for 2009:Q1

0.8 2.8 1.5 Four quarters ahead for 2009:Q1

5.3 6.2 1.6

United States Turkey One quarter ahead

for 2009:Q1 46.7 30.3 8.5 One quarter ahead

for 2009:Q1 31.5 27.1 5.3

Four quarters ahead for 2009:Q1

2.6 4.0 4.2 Four quarters ahead for 2009:Q1

3.5 2.3 2.8

Sources: Bloomberg Finance L.P.; Haver Analytics; IMF, Global Data Source and World Economic Outlook databases; Thomson Reuters Datastream; and IMF staff estimates. Note: The table depicts the cumulative probabilities of a growth outcome in 2009:Q1 of less than or equal to the actual growth outturn (quarter over quarter, annualized) in that period drawn from conditional density forecasts of GDP growth made four quarters earlier (that is, in 2008:Q1). The left column depicts probabilities from the model with financial conditions indices (FCIs) estimated with information available in real time. The middle column depicts probabilities from the model with FCIs estimated with full in-sample information. The right column depicts probabilities from the autoregressive model of GDP growth. Autoregressive = quantile regression of one-year-ahead GDP growth on current quarter GDP growth; FCI augmented = quantile regression of one-year-ahead GDP growth on current quarter GDP growth and FCI.

Table 3.2. Market Consensus Forecasts for the Global Financial Crisis Were Considerably More Optimistic Than Forecasts Based on Financial Conditions

Growth Forecasts Conditional on Lagged GDP and FCI Consensus Growth Forecasts Growth Outturn in

2009:Q112008:Q1 2008:Q4 2008:Q1 2008:Q4 Brazil 3.1 −4.3 4.6 2.1 −6.9 Canada 1.7 −5.3 1.7 −0.1 −8.8 France 1.9 −1.2 1.6 −0.6 −6.4 Mexico 2.6 −3.6 2.8 −0.1 −14.7 South Africa 2.7 −2.0 4.7 2.7 −6.1 Switzerland 1.9 −2.0 2.8 −1.6 −5.5 Turkey 3.4 −7.4 4.8 0.8 −15.2 United States 1.9 −3.8 1.6 −1.3 −5.4

Sources: Bloomberg Finance L.P.; Consensus Economics; Haver Analytics; IMF, Global Data Source and World Economic Outlook databases; Thomson Reuters Datastream; and IMF staff estimates. Note: Columns 2 and 3 of the table denote, respectively, the conditional mean forecasts for (quarter over quarter, annualized) GDP growth in 2009:Q1 made one quarter and one year earlier based on an ordinary least squares regression of future GDP growth on current quarter FCI and GDP growth. Columns 4 and 5 denote market consensus forecasts for 2009:Q1 made one quarter and four quarters earlier, respectively. Column 6 depicts the actual growth outturn. FCI = financial conditions index. 1Based on data available as of August 3, 2017.

105

C H A P T E R 3 F I N A N C I A L C O N d I T I O N S A N d G R O w T h A T R I S k

International Monetary Fund | October 2017

This explains why autoregressive-conditional quantile forecasts were behind the curve, even at the end of 2008. A few quarters earlier, in early 2008, FCIs had risen from their boom-time lows but were only at their historical averages (for emerging market economies) or at levels corresponding to recessions significantly milder than the outturn of the first quarter of 2009 (for advanced economies). Consequently, one year ahead, conditioning on FCIs does not result in signifi- cantly different predictions of growth during the global financial crisis relative to either consensus forecasts or autoregressive-conditional quantile forecasts.

Partitioning the FCI constituents into subindices enables the forecasts conditioned on financial indi- cators to regain relative predictive gains over longer time horizons in several countries (Table 3.3).18 One-year-ahead conditional forecasts for annual growth assign significantly higher likelihood to growth outcomes less than or equal to the outturn of the first quarter of 2009 when the forecasts are based on infor- mation in financial indicators than when based only on

18The contribution of each financial indicator to its group subin- dex is determined according to a methodology designed to improve forecast performance as discussed in Annex 3.2.

lagged GDP growth. This is the likely consequence of separating credit aggregates from asset prices, thereby allowing their information to gain greater weight at horizons beyond one quarter.

Real-time conditional density forecasts of economic growth are almost identical to those reported above for in-sample forecasts (Figures 3.8 and 3.9). Hence, using information in FCIs and in partitioned financial indicators available only up to one to four quarters earlier than the first quarter of 2009 would result in conditional likelihoods being assigned to the actual growth outcomes that are very similar to those obtained through in-sample forecasts using financial indicators (Tables 3.1 and 3.3).19

19This is implied by the fact that real-time forecasts of the quan- tiles of future GDP growth obtained through recursive estimation are almost identical to (or, below the median quantile, often lower than) those obtained through the in-sample forecasts. The fact that a majority of financial indicators are available only from the mid-1990s to the mid-2000s, especially for emerging market econ- omies, prevents backtesting of the model’s forecasting ability relative to earlier crisis-related recessions, for example, in Sweden (1990–92), Mexico (1994), east Asia (1997), and Turkey (2000–01), among others. More generally, low-frequency and limited time series data on real and financial variables preclude implementation with suffi- cient power of appropriate out-of-sample forecast evaluation tests described in Corradi and Swanson 2006 and Komunjer 2013.

Table 3.3. Forecast of GDP Growth Distribution for the Global Financial Crisis: Comparing Partitioned and Univariate Financial Conditions Indices with Autoregressions (Cumulative probability of actual 2009:Q1 growth outturn, percent)

Selected Advanced Economies Selected Emerging Market Economies Real-time Partitioned Financial Variables

Partitioned Financial Variables

FCI Augmented Autoregressive

Real-time Partitioned Financial Variables

Partitioned Financial Variables

FCI Augmented Autoregressive

Germany Brazil Four quarters ahead

for 2009:Q1 0.8 0.7 0.4 0.0 Four quarters ahead

for 2009:Q1 14.0 6.7 5.0 5.5

Sweden Chile

Four quarters ahead for 2009:Q1

7.1 5.7 0.8 0.5 Four quarters ahead for 2009:Q1

12.7 10.4 1.7 2.0

United Kingdom South Africa

Four quarters ahead for 2009:Q1

6.4 5.0 2.8 1.5 Four quarters ahead for 2009:Q1

5.4 7.3 6.2 1.6

United States Turkey

Four quarters ahead for 2009:Q1

24.7 19.1 4.0 4.2 Four quarters ahead for 2009:Q1

7.4 4.4 2.3 2.8

Sources: Bloomberg Finance L.P.; Haver Analytics; IMF, Global Data Source and World Economic Outlook databases; Thomson Reuters Datastream; and IMF staff estimates. Note: The table depicts the cumulative probabilities of a growth outcome in 2009:Q1 of less than or equal to the actual growth outturn (quarter over quarter, annualized) in that period drawn from conditional density forecasts of GDP growth made four quarters earlier (that is, in 2008:Q1) according to the four alternative methodologies. Autoregressive = quantile regression of one-year-ahead GDP growth on current quarter GDP growth; FCI = financial conditions index; FCI augmented = quantile regression of one-year-ahead GDP growth on current quarter GDP growth and FCI; partitioned financial variables = quantile regression of one-year-ahead GDP growth on current quarter GDP growth and subindices of financial indicators.

106

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

–20

–10

19 73

:Q 1

75 :Q

2 77

:Q 3

79 :Q

4 82

:Q 1

84 :Q

2 86

:Q 3

88 :Q

4 91

:Q 1

93 :Q

2 95

:Q 3

97 :Q

4 20

00 :Q

1 02

:Q 2

04 :Q

3 06

:Q 4

09 :Q

1 11

:Q 2

13 :Q

3 15

:Q 4

–30

–15

19 91

:Q 1

92 :Q

2 93

:Q 3

94 :Q

4 96

:Q 1

97 :Q

2 98

:Q 3

99 :Q

4 20

01 :Q

1 02

:Q 2

03 :Q

3 04

:Q 4

06 :Q

1 07

:Q 2

08 :Q

3 09

:Q 4

11 :Q

1 12

:Q 2

13 :Q

3 14

:Q 4

16 :Q

–30

–20

–10

19 73

:Q 1

75 :Q

2 77

:Q 3

79 :Q

4 82

:Q 1

84 :Q

2 86

:Q 3

88 :Q

4 91

:Q 1

93 :Q

2 95

:Q 3

97 :Q

4 20

00 :Q

1 02

:Q 2

04 :Q

3 06

:Q 4

09 :Q

1 11

:Q 2

13 :Q

3 15

:Q 4

–20

–10

19 91

:Q 1

92 :Q

2 93

:Q 3

94 :Q

4 96

:Q 1

97 :Q

2 98

:Q 3

99 :Q

4 20

01 :Q

1 02

:Q 2

03 :Q

3 04

:Q 4

06 :Q

1 07

:Q 2

08 :Q

3 09

:Q 4

11 :Q

1 12

:Q 2

13 :Q

3 14

:Q 4

16 :Q

1 –15

–10

–5

19 81

:Q 1

83 :Q

86 :Q

88 :Q

91 :Q

93 :Q

96 :Q

98 :Q

20 01

:Q 1

03 :Q

06 :Q

08 :Q

11 :Q

13 :Q

16 :Q

–40

–20

19 91

:Q 1

92 :Q

2 93

:Q 3

94 :Q

4 96

:Q 1

97 :Q

2 98

:Q 3

99 :Q

4 20

01 :Q

1 02

:Q 2

03 :Q

3 04

:Q 4

06 :Q

1 07

:Q 2

08 :Q

3 09

:Q 4

11 :Q

1 12

:Q 2

13 :Q

3 14

:Q 4

16 :Q

–20

–15

–10

–5

19 73

:Q 1

75 :Q

2 77

:Q 3

79 :Q

4 82

:Q 1

84 :Q

2 86

:Q 3

88 :Q

4 91

:Q 1

93 :Q

2 95

:Q 3

97 :Q

4 20

00 :Q

1 02

:Q 2

04 :Q

3 06

:Q 4

09 :Q

1 11

:Q 2

13 :Q

3 15

:Q 4

–10

–5

19 91

:Q 1

92 :Q

2 93

:Q 3

94 :Q

4 96

:Q 1

97 :Q

2 98

:Q 3

99 :Q

4 20

01 :Q

1 02

:Q 2

03 :Q

3 04

:Q 4

06 :Q

1 07

:Q 2

08 :Q

3 09

:Q 4

11 :Q

1 12

:Q 2

13 :Q

3 14

:Q 4

16 :Q

Sources: Bloomberg Finance L.P.; Haver Analytics; IMF, Global Data Source and World Economic Outlook databases; Thomson Reuters Datastream; and IMF staff estimates. Note: This figure shows the estimates of the 5th (bottom), 50th (middle), and 95th (top) quantiles of GDP growth based on the quantile regression model where one-quarter-ahead GDP growth is regressed on current date financial conditions index and GDP growth.

In-sample estimation Real-time estimation

Figure 3.8. In-Sample and Recursive Out-of-Sample Quantile Forecasts: One Quarter Ahead (Percent)

1. Germany 2. Brazil

3. United Kingdom 4. Chile

7. United States 8. South Africa

5. Sweden 6. Turkey

107

C H A P T E R 3 F I N A N C I A L C O N d I T I O N S A N d G R O w T h A T R I S k

International Monetary Fund | October 2017

–6

–4

–2

2001 02 03 04 05 06 07 08 09 10 11 12 13 14 15 –10

–5

2001 02 03 04 05 06 07 08 09 10 11 12 13 14 15

–4

–2

2001 02 03 04 05 06 07 08 09 10 11 12 13 14 15 –5

2001 02 03 04 05 06 07 08 09 10 11 12 13 14 15

–10

–5

2001 02 03 04 05 06 07 08 09 10 11 12 13 14 15 –10

–5

2001 02 03 04 05 06 07 08 09 10 11 12 13 14 15

–8

–6

–4

–2

2001 02 03 04 05 06 07 08 09 10 11 12 13 14 15 –4

–2

2001 02 03 04 05 06 07 08 09 10 11 12 13 14 15

In-sample estimation Real-time estimation

Sources: Bloomberg Finance L.P.; Haver Analytics; IMF, Global Data Source and World Economic Outlook databases; Thomson Reuters Datastream; and IMF staff estimates. Note: This figure shows the estimates of the 25th (bottom), 50th (middle), and 75th (top) quantiles of GDP growth based on the quantile regression model with partitioned financial indicators replacing the univariate financial conditions index.

Figure 3.9. In-Sample and Recursive Out-of-Sample Quantile Forecasts: Four Quarters Ahead (Percent)

1. Germany 2. Brazil

3. United Kingdom 4. Chile

7. United States 8. South Africa

5. Sweden 6. Turkey

108

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

This augurs well for the parameter stability of the chapter’s forecast model, demonstrating that its fore- casts and relative predictive ability are not an artifact of incorporating events such as the global financial crisis into estimates of its parameters.

Policy Implications The chapter’s findings underscore the importance of policymakers maintaining heightened vigilance regarding risks to growth during periods of benign financial condi- tions that may provide a fertile breeding ground for the accumulation of financial vulnerabilities. Changes in the domestic price of risk appear to be potent signals of immi- nent threats to growth and can be useful for swift deploy- ment of monetary easing and crisis-management policy actions. Incorporating information in slower-moving indi- cators could help better calibrate countercyclical policies, even though doing so systematically would require com- bining the information derived from the models described in this chapter with appropriate structural models.

This chapter develops a new macroeconomic measure of financial stability by linking financial conditions to the probability distribution of future GDP growth. Since policymakers care about the whole distribution of future GDP growth, linking the state of the financial system to such a distribution would enhance macro-financial surveillance. Policymakers would be able to specify bad outcomes in terms of their risk preference or tolerance and undertake appro- priate action based on the information provided by financial conditions. Thus, the new modeling approach can be a powerful tool for forecasting and policy development.

Financial conditions contain useful information with which to help forecast risks to economic growth at short- and medium-term horizons. Thus, the tools used and developed in this chapter can help policy- makers assess the risks to the real economy associ- ated with various states of the financial system. For example, at the current juncture, elevated leverage signals downside risks to growth in the medium term, although in the short term, this risk is mitigated by the low price of risk. However, a scenario of rapid decom- pression in spreads and an increase in financial market volatility would add to the risks arising from leverage, significantly worsening the growth outlook.

Policymakers could use the information provided by such a surveillance framework to identify immi-

nent threats and take swift countervailing action over very short horizons. If a rapid increase in the price of risk at a time of elevated leverage or balance sheet mismatches indicates an imminent threat to the economy, policymakers can quickly ease monetary policy and deploy a wide range of crisis-management and -prevention measures to prevent tail events or reduce their magnitude. During the global financial crisis, bilateral and multilateral swap lines, general creditor guarantees, asset purchase programs, and emergency liquidity facilities, among others, were marshalled by a number of countries at relatively short notice.

The framework developed in this chapter could potentially help policymakers design policy actions to respond in a timely manner to threats to financial stability indicated by changes in financial conditions. It is natural to think of calibrating policy actions on the state of financial conditions—much as monetary policy action is calibrated to information on inflation and output under standard Taylor rules. For example, countercyclical macroprudential tools, such as bank capital buffers and limits on loan-to-value ratios, could be designed and calibrated to contain the growth of financial vulnerabilities in the presence of loose finan- cial conditions. In this regard, the estimated forecast relationships from the GDP growth-at-risk model of this chapter can also be used to calibrate structural models that are amenable to counterfactual analysis and policy development.20

Practical implementation of forecasting of risks to growth based on financial conditions will require data gaps to be closed. This need strengthens the case for greater data-gathering efforts. It also points to a need for continuous calibration of these types of models as data gaps gradually close and for incorporation of country-level information that may substitute for the lack of standard financial indicators. In this way, policymakers and others could significantly improve on the forecasting power of the models presented here by incorporating rich country-level information to com- plement the models’ broad financial indicators. As local financial markets undergo structural developments, and authorities consider certain financial indicators to

20One option could be to use the conditional density forecasts of GDP growth to calibrate the higher moments (for example, conditional volatility or skewness) of structural models that embed financial accelerator mechanisms such as the one described in Annex 3.1.

109

C H A P T E R 3 F I N A N C I A L C O N d I T I O N S A N d G R O w T h A T R I S k

International Monetary Fund | October 2017

be increasingly relevant, these could also be gradually incorporated into the analysis.21

Annex 3.1. Financial Vulnerabilities and Growth Hysteresis in Structural Models22

An Illustrative Simulation

A simulation exercise of a structural model is con- ducted to illustrate the nonlinear response of output growth to shocks depending on the level of financial vulnerabilities. The exercise shows that embedding an occasionally binding funding constraint on borrowers in an otherwise standard New Keynesian (NK) open economy structural model is sufficient to generate two key stylized facts. These are, first, that the steady-state probability distribution of GDP growth is negatively skewed and, second, that asset prices and credit aggre- gates are leading indicators of risks to GDP growth.

In the presence of financial frictions, the response of output growth to shocks is highly nonlinear. Recent advances in macroeconomic theory have clarified the importance of financing constraints on borrowers and intermediaries in generating this response. In their seminal contributions, Bernanke and Gertler (1989); Kiyotaki and Moore (1997); and Bernanke, Gertler, and Gilchrist (1999) clarified the role of credit market frictions in determining fluctuations in real economic activity. Their linear real business cycle models embed a financial accelerator mechanism in which endogenous developments in credit markets propagate and amplify shocks to the real economy. Although these models explain how financial frictions increase the amplitude of real business cycles, they do not shed light on how and when they can increase the duration of those cycles or generate extreme, unlikely negative outcomes (asymmetry, or tail risk). The key insight of recent advances in business cycle theory is that this outcome depends on individual financial decisions of banks, firms, and households that fail to take into consider- ation dynamic credit supply externalities implied by their decisions. That is, individual borrowers fail to

21The methodology developed in this chapter is used to model the impact of financial vulnerabilities on GDP growth. It is flexible in the inputs it can receive. In countries where risks to the real economy posed by amplifiers, whether real or fiscal, are not traded in deep financial markets, corresponding nonfinancial indicators could also be used as inputs.

22Prepared by Mitsuru Katagiri. (This annex is a summary of Katagiri, forthcoming.)

take into account the fact that once aggregate leverage is sufficiently high, shocks can activate occasionally binding collateral constraints (OBCCs). This, in turn, can generate a vicious cycle of deleveraging and nega- tive asset price spirals that clog credit intermediation, consumption, investment, and growth.23

The simulation exercise embeds an OBCC into an NK open economy dynamic general equilibrium model. The OBCC is modeled as in Kiyotaki and Moore 1997. To tease out implications for optimal policy, nominal frictions based on an open economy NK model are incorporated in the spirit of Galí and Monacelli 2005. The main features of the model are as follows: Households are endowed with trad- able goods as in Bianchi 2011, while they produce nontradables using capital and labor. Households maximize their lifetime utility by choosing an inter- temporal portfolio of tradable and nontradable goods for consumption and supplying labor to the produc- tion process. Their borrowing must be lower than a fixed fraction of their capital value (that is, there is a collateral constraint). The nontradables sector is monopolistically competitive, and price setting is sub- ject to nominal frictions. Asset prices are determined under a fixed supply of capital. Nominal interest rates are set under a standard Taylor rule responding to inflation and output. The exchange rate is pinned down by the uncovered interest parity condition. The parameters are calibrated based on standard values in the literature of an OBCC model and an open econ- omy NK model, including Bianchi 2011 and Galí and Monacelli 2005.

The simulated density of future output is shown to be negatively skewed; that is, it has a fat left tail, indicating a greater risk of severe recession. The unconditional distribution of future output (Annex Figure 3.1.1, panel 3) is negatively skewed—the skew- ness measure, at –1.51, is statistically significant. In the simulation, as in reality, the collateral constraint does not typically bind. Thus, the evolution of all economic variables, including output, is standard for the most part. However, when the OBCC binds (a rare event), output and asset prices decline significantly because

23For models embedding OBCCs on end-borrowers, see Bianchi 2011; Korinek and Simsek 2016; and Bianchi and Mendoza, forthcoming. For OBCCs or value-at-risk constraints on interme- diaries, see He and Krishnamurthy 2013 and Brunnermeier and Sannikov 2014.

110

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

of the vicious cycle of asset fire sales and tighter credit conditions, and output suffers.

The simulation exercise clearly indicates the utility of conditioning the growth outlook on asset prices. Risk premiums in the simulation exercise are defined as the return on capital minus the inverse of the stochastic discount factor, as is standard.24 Annex Figure 3.1.1 shows the conditional density of output in period t, given that the risk premium in period t − 1 is less than 30 basis points (the case of high asset prices depicted in panel 1) and more than 30 basis points (the case of low asset prices depicted in panel 2). Those two panels indicate that when risk premiums rise (equivalently, when asset prices fall), the conditional density of one-period-ahead output shifts to the left and becomes negatively skewed. Higher risk premiums predict a lower average value of one-period-ahead output and a more pessimistic risk outlook (fatter left tail).

Asset prices and credit aggregates can also be useful leading indicators of recessions or financial crises. The relationship between one-period-ahead output and risk premiums (Annex Figure 3.1.2, panel 1) indicates that the lower quantile of output declines significantly with rising risk premiums, whereas its upper quantile is significantly less sensitive. The relationship between one-period-ahead output and the credit-to-GDP ratio shows that a financial crisis occurs only when the ratio is at a historically high level (Annex Figure 3.1.2, panel 2). Finally, risk premiums and credit-to-output ratios are significantly higher than their steady-state values for several peri- ods before a crisis (Annex Figure 3.1.3).

Calibrating Policy Rules to Attenuate Risks to Growth from Financial Vulnerability

Macroprudential policy contingent on the state of financial conditions can mitigate the adverse real effects of financial crises. The decentralized equilibrium described in the previous section of this annex is not socially optimal because agents fail to take into consid- eration the negative systemic externalities of their lever- age choices on asset prices. Borrowers’ resulting excess leverage increases the frequency of financial crises.

24Note that risk premiums based on this definition are not directly observable in the data, but are conceptually close to the excess return of risk assets as defined in Gilchrist and Zakrajšek 2012 and hence can be calculated from financial market data.

0.94 0.95 0.96 0.97 0.98 0.99 1.01 1.02 1.03 1.04 1.05

One-period-ahead output (Normalized; steady state = 1.0)

0.94 0.95 0.96 0.97 0.98 0.99 1.01 1.02 1.03 1.04 1.05 One-period-ahead output

(Normalized; steady state = 1.0)

0.94 0.95 0.96 0.97 0.98 0.99 1.01 1.02 1.03 1.04 1.05

One-period-ahead output (Normalized; steady state = 1.0)

8 × 104

× 104

Source: IMF staff estimates.

Annex Figure 3.1.1. Conditional Densities of Growth with High and Low Asset Prices—One-Period-Ahead Forecasts (Frequency)

1. High Asset Prices

2. Low Asset Prices

3. Unconditional

111

C H A P T E R 3 F I N A N C I A L C O N d I T I O N S A N d G R O w T h A T R I S k

International Monetary Fund | October 2017

Bianchi (2011) and Bianchi and Mendoza (forthcom- ing) show that a macroprudential tax (that is, a tax on debt before the crisis) that is contingent on the state of financial conditions can prevent excess leverage and implement the socially optimal outcome as a decen- tralized equilibrium. This socially optimal outcome can also be implemented by a regulation on loan-to-value (LTV) ratios.

Once the optimal state-contingent macroprudential policy (taxes on debt or LTV regulation) is intro- duced, vulnerability to a recession (as measured by the negative skewness of the output distribution) is significantly mitigated. In the baseline simulation of the equilibrium without optimal macroprudential policy,

0.90

0.94

0.98

1.02

1.06

0 1 32

O ne

-p er

io d-

ah ea

d ou

tp ut

O ne

-p er

io d-

ah ea

d ou

tp ut

Risk premium (percent)

0.88

0.92

0.96

1.00

1.04

1.08

0.08 0.10 0.12 0.14

Credit-to-output ratio

Source: IMF staff estimates.

Annex Figure 3.1.2. One-Period-Ahead GDP and Financial Conditions (Normalized; steady state = 1.0)

1. Risk Premium

Increasing risk premiums signal a more pessimistic growth outlook ...

2. Credit-to-Output Ratio

... as does elevated leverage.

0.94

0.95

0.96

0.97

0.98

0.99

1.00

0.85

0.90

0.95

1.00

1.05

1.10

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

t – 4 t – 3 t – 2 t – 1 t t + 1 t + 2 t + 3 t + 4

Source: IMF staff estimates. Note: The crisis happens in period 5 (t ) in the figures. The crisis is defined as a period in which output declines by more than 3 percent. The red dashed lines denote steady-state values.

Annex Figure 3.1.3. Asset Prices and Credit Aggregates before and after a Financial Crisis

1. Output (Normalized; steady state = 1.0)

Severe economic contractions are preceded by several periods of excessive leverage and, shortly before the crisis, by sharply rising risk premiums.

2. Credit-to-Output Ratio (Normalized; steady state = 1.0)

3. Risk Premium (Percent)

112

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

the probability of a recession driven by a financial crisis is 1.3 percent, and the skewness of the density of future GDP growth at –1.51 is statistically significant. Implementation of the state-contingent debt tax or state-contingent LTV regulation reduces these values to, respectively, 0.5 percent and –0.66.

A simple policy rule conditioned on financial indicators comes close to implementing the optimal macroprudential policy. The optimal policy itself is a complex nonlinear function of state variables and is probably too complicated to implement in practice.25 Fortuitously, a simple rules-based macroprudential policy responding to vulnerability measures does a good job of mitigating the harmful effects of finan- cial crises. Risk premiums are used to improve the

25The nonlinearity stems from the fact that policymakers should raise borrowing costs through taxes or LTV regulations only when a crisis is predicted.

performance of a simple rules-based macroprudential policy because they have predictive power for the crisis. Annex Figure 3.1.4 compares the evolution of real and financial indicators under a simple policy rule whereby debt taxes are a linear function of risk premiums to the baseline equilibrium. Policy based on a simple linear rule delivers almost the same performance as the optimal policy, implying that financial conditions such as risk premiums are useful for conducting macropru- dential policies in practice.26

26There are two caveats. First, all crises in the OBCC model are caused by a simple collateral constraint, whereas many other factors can contribute to financial crises. Second, the model assumes that policymakers can immediately respond to vulnerabilities. If there is a delay in policy reactions or their transmission to the real economy, the policy implications may be different.

0.94

0.95

0.96

0.97

0.98

0.99

1.00

t – 4 t – 3 t – 2 t – 1 t t + 1 t + 2 t + 3 t + 4

0.70

0.75

0.80

0.85

0.90

0.95

1.00

1.05

0.85

0.90

0.95

1.00

1.05

1.10

1.4

1.6

1.8

2.0

2.2

Baseline Simple

Source: IMF staff estimates.

Annex Figure 3.1.4. Simple Debt Tax Ameliorates Risk of Leverage-Induced Recessions

1. Output (Normalized; steady state = 1.0)

2. Asset Prices (Normalized; steady state = 1.0)

3. Credit-to-Output Ratio (Normalized; steady state = 1.0)

4. Inflation (Percent)

113

C H A P T E R 3 F I N A N C I A L C O N d I T I O N S A N d G R O w T h A T R I S k

International Monetary Fund | October 2017

Annex 3.2. Estimating Financial Conditions Indices27

Univariate Financial Conditions Indices

A simple way to build a summary measure of financial conditions is to construct univariate financial conditions indices (FCIs) following the approach in the April 2017 GFSR, although with some important modifications. The main change is that the coverage of financial indicators is expanded to include additional information relevant to assessing domestic financial vulnerabilities. FCIs will therefore also include variables that summarize global risk sen- timent (Chicago Board Options Exchange Volatility Index [VIX], Merrill Lynch Option Volatility Esti- mate [MOVE] Index), credit aggregates that directly indicate the level of financial vulnerability in the economy, and commodity prices and exchange rates that may influence and reflect the ease of funding and financial constraints—for example, by altering borrowers’ net worth.28

Following the methodology presented in Annex 3.1 of the April 2017 GFSR, FCIs are reestimated for 11 advanced economies starting in 1973 and for 10 emerging market economies starting in 1991. A set of 19 financial indicators is used to capture both domestic and global developments influencing a coun- try’s financial conditions (see Annex Table 3.2.1 for country coverage and Annex Table 3.2.2 for variables included and data sources). The FCIs are estimated based on Koop and Korobilis 2014 and build on the estimation of the time-varying parameter vector autore- gression model of Primiceri (2005) and dynamic factor

27Prepared by Romain Lafarguette and Dulani Seneviratne. 28An important reason to expand coverage to aggregates is that

beyond a few advanced economies, it is unlikely that developments in asset prices provide an adequately encompassing and timely sum- mary of the information regarding vulnerabilities that is contained in these financial aggregates. Thus, conditioning directly on the information content of the aggregates may improve the accuracy of forecasts of the risk outlook for growth.

models of Doz, Giannone, and Reichlin (2011).29 This approach has two advantages. First, it can control for current macroeconomic conditions. Second, it allows for dynamic interaction between the FCIs and macro- economic conditions, which can also evolve over time. The model takes the following form:

x t = λ t y Y t + λ t

f f t + u t ,

[ Y t f t ] = B 1,t [

Y t – 1 f t – 1

] + B 2,t [ Y t – 2 f t – 2

] + . . . + ε t , (A3.2.1)

in which x is a vector of financial indicators, Y is a vector of macroeconomic variables of interest (includ- ing real GDP growth and inflation), λ t

y are regression coefficients, λ t

f are the factor loadings, and f t is the latent factor, interpreted as the FCI.

Univariate FCIs offer a parsimonious way of sum- marizing the information in several financial indica- tors, which could be advantageous from a forecasting perspective because it can help reduce parameter uncertainty. However, the weight of each variable is not necessarily driven by economic considerations of relative importance as suggested either by theory or by country-specific characteristics. For example, movements in asset prices may be effective in pin- pointing risks at short horizons, but slower-moving credit aggregates are likelier to yield more infor- mation at longer time horizons. Moreover, while asset prices are likely to be an adequate summary of financial vulnerabilities in some advanced economies, credit aggregates may possess significantly greater information content in emerging market economies. Consequently, financial indicators need not receive the same weight across different time horizons and countries; therefore, as described in the second sec- tion of this annex, the chapter also uses an approach that seeks to exploit the information content of

29The FCIs are estimated using Koop and Korobilis’ (2014) code (https:// sites .google .com/ site/ dimitriskorobilis/ matlab).

Annex Table 3.2.1. Country Coverage

Australia Germany Mexico Turkey Brazil India Russia United Kingdom Canada Indonesia South Africa United States Chile Italy Spain China Japan Sweden France Korea Switzerland

Source: IMF staff.

114

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

Annex Table 3.2.2. Data Sources Variables Description Source

Domestic-Level Variables Term Spreads Yield on 10-year government bonds minus yield on

three-month Treasury bills Bloomberg Finance L.P.; IMF staff

Interbank Spreads Interbank interest rate minus yield on three-month Treasury bills

Bloomberg Finance L.P.; IMF staff

Change in Long-Term Real Interest Rate

Percentage point change in the 10-year government bond yield, adjusted for inflation

Bloomberg Finance L.P.; IMF staff

Corporate Spreads Corporate yield of the country minus yield of the benchmark country; JPMorgan CEMBI Broad is used for emerging market economies where available

Bloomberg Finance L.P.; Thomson Reuters Datastream

Equity Returns (local currency)

Log difference of the equity indices Bloomberg Finance L.P.

House Price Returns Log difference of the house price index Bank for International Settlements; Haver Analytics; IMF staff

Equity Return Volatility Exponential weighted moving average of equity price returns Bloomberg Finance L.P.; IMF staff Change in Financial

Sector Share Log difference of the market capitalization of the financial

sector to total market capitalization Bloomberg Finance L.P.

Credit Growth Percent change in the depository corporations’ claims on private sector

Bank for International Settlements; Haver Analytics; IMF, International Financial Statistics database

Sovereign Spreads Yield on 10-year government bonds minus the benchmark country’s yield on 10-year government bonds

Bloomberg Finance L.P.; IMF staff

Banking Sector Vulnerability

Expected default frequency of the banking sector Moody’s Analytics, CreditEdge; IMF staff

Exchange Rate Movements

Change in US dollar per national currency exchange rate; for the United States, Bloomberg Finance L.P.’s DXY index is used

Bloomberg Finance L.P.; IMF, Global Data Sources and International Financial Statistics databases

Domestic Commodity Price Inflation

A country-specific commodity export price index constructed following Gruss 2014, which combines international commodity prices and country-level data on exports and imports for individual commodities; change in the estimated country-specific commodity export price index is used

Bloomberg Finance L.P.; IMF, Global Data Sources database; United Nations, COMTRADE database; IMF staff

Trading Volume (equities)

Equity markets’ trading volume, calculated as level to 12-month moving average

Bloomberg Finance L.P.

Market Capitalization (equities)

Market capitalization of the equity markets, calculated as level to 12-month moving average

Bloomberg Finance L.P.; Thomson Reuters Datastream

Market Capitalization (bonds)

Bonds outstanding, calculated as level to 12-month moving average

Dealogic; IMF staff

Change in Credit to GDP Change in credit provided by domestic banks, all other sectors of the economy, and nonresidents (in percent of GDP)

Bank for International Settlements; Haver Analytics; IMF staff

Real GDP Growth Percent change in GDP at constant prices IMF, World Economic Outlook database Inflation Percent change in the consumer price index Haver Analytics; IMF, International Financial

Statistics database

Global-Level Variables VIX Chicago Board Options Exchange Market Volatility Index Bloomberg Finance L.P.; Haver Analytics MOVE Merrill Lynch Option Volatility Estimate Index Bloomberg Finance L.P.

Source: IMF staff. Note: CEMBI = Corporate Emerging Markets Bond Index; DXY = Dollar Index Spot; MOVE = Merrill Lynch Option Volatility Estimate Index; VIX = Chicago Board Options Exchange Volatility Index.

115

C H A P T E R 3 F I N A N C I A L C O N d I T I O N S A N d G R O w T h A T R I S k

International Monetary Fund | October 2017

financial indicators in a manner that is more sensitive to countries and time horizons.

Data Partitioning Based on Linear Discriminant Analysis

The individual financial indicators are aggregated into groups using linear discriminant analysis (LDA), a data-reduction technique (Annex Table 3.2.3). LDA aims to project a data set onto a lower-dimensional space while ensuring adequate separation of data into categories. LDA is similar to principal components analysis (PCA) in the sense that it maximizes the common variance among a set of variables, but it diverges from PCA by also ensuring that the linear combination of the variables discriminates across the classes of another categorical variable of interest. In the framework of the chapter, this categorical variable is a dummy variable, defined at the country level, equal to one when future GDP growth at a one-year horizon is below the 20th percentile of historical outcomes and equal to zero otherwise. Consequently, the loading on each individual financial indicator in the LDA is determined in a way that maximizes its contribu- tion to discriminating between periods of low GDP growth and periods of normal GDP growth. This is convenient from the chapter’s perspective because it allows for a link between financial indicators and GDP growth in the data-reduction process. By contrast, the PCA approach aggregates only information about the common trend among financial indicators.30

30LDA assumes independence of normally distributed data and homoscedastic variance among each class, although LDA is consid- ered robust when these assumptions are violated. See Duda, Hart, and Stork 2001. See Izenman 2013 for a thorough exposition of the LDA technique.

Annex 3.3. The Conditional Density of Future GDP Growth31

Quantile Regressions

The estimation of the conditional density forecast is conducted through quantile projections.32 This approach starts by using quantile regressions to directly estimate the conditional quantiles (q) of the forecast distribution of GDP growth ( y ) h quarters ahead, as a function of both its current level and current financial conditions (FC ):

y t + h,q = β f,q h FC t + β y,q

h y t + ϵ t,q h . (A3.3.1)

In the baseline approach, FC corresponds to a pre- determined univariate financial conditions index (FCI) constructed in the manner described in Annex 3.2.

The empirical model is subsequently modified to investigate the relative significance of asset prices, credit aggregates, and global or foreign factors in signaling risks to GDP growth in the near to medium term:

y t + h,q = α p,q h p t + β a,q

h Agg t + γ y,q h y t + ϕ f,q

h f t + ϵ t,q h ,

(A3.3.2) in which p, Agg, and f correspond to the principal com- ponents of the price of risk (asset prices and risk spreads),

31Prepared by Sheheryar Malik and Romain Lafarguette. 32For an introduction to quantile regression, see Koenker 2005. As

highlighted by Komunjer (2013), quantile regressions rely on specific functional form assumptions and have some important advantages in forecasting the conditional distribution of the variable of interest. These include the desirability of the conditional quantile estimator as a predictor of the true future quantile; robustness of the estimation to extreme outliers and violations of normality and homoscedasticity of the errors; flexibility, allowing for time-varying structural parame- ters and the optimal weighting of predictors depending on country, horizon, and the relevant portion of the distribution; and the ability to avoid overfitting (compared with more complex models such as copulas and extreme value theory).

Annex Table 3.2.3. Partitioning of Financial Indicators into Groups Price of Risk Leverage Foreign Shocks Persistence

Financial and Real Indicators (when available)

Term spread Credit to GDP Bilateral exchange rate (US dollar to local currency)

GDP growth Corporate spread Credit growth (quarterly) Short-term rate Commodity prices Real long-term rate VIX1

Sovereign spread Interbank spread Equity returns Equity historical volatility House price returns

Source: IMF staff. 1 Except for the United States, for which VIX enters as a price-of-risk variable. VIX = Chicago Board Options Exchange Volatility Index.

116

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

credit aggregates, and global or foreign variables (com- modity prices, exchange rates, and global risk sentiment). This approach disentangles the contribution of changes in the price of risk from evolving credit aggregates and shocks to the external environment when it comes to forecasting risks to GDP growth. It thereby provides insight into which variables signal growth tail risks over various time horizons. This can help policymakers and others design a surveillance framework that seeks to embed information flowing in at different frequencies.

Deriving the Density Forecast

The quantile regression in equation (A3.3.1) delivers an estimate for the conditional quantile function (or inverse cumulative distribution function) h quarters ahead—that is, y ˆ t + h,q ( = β̂ f,q h FC t + β̂ y,q h y t ) . Given the noisiness of such estimates in practice, recovering the corresponding predictive probability density function will inevitably require smoothing of the quantile func- tion. In this chapter, this is accomplished via fitting a parametric form skewed t distribution.33

For each quarter, the analysis attempts to pin down four parameters of the predictive den- sity {μ t + h , s t + h , v t + h , ξ t + h } by minimizing the squared distance between the estimated quantile function, y ˆ t + h,q , and (theoretical) quantile func- tion y q

f ( μ t + h , s t + h , v t + h , ξ t + h ) corresponding to the above skewed t distribution (see Giot and Laurent 2003). The four parameters (μ, s, v, ξ ) are, respectively, the location, scale, degrees of freedom, and the shape of skewed t distribution. Specifically, the 5th, 25th, 50th, 75th, and 95th percentiles are matched via

{μ t + h , s t + h , v t + h , ξ t + h } = μ t + h , s t + h , v t + h , ξ t + h

argmin

∑ q { y ˆ t + h,q − y q f ( μ t + h , s t + h , v t + h , ξ t + h ) } 2 ,

in which μ t + h ∈ ℝ , s t + h > 0 , v t + h ≥ 2, and ξ t + h > 0 . Notwithstanding the skewness property,

33There are many choices for fitting a conditional density on the set of conditional quantiles. Adrian, Boyarchenko, and Giannone (2016) adopt a parametric approach focusing on a distribution family chosen a priori (t skewed), whereas De Nicolò and Lucchetta (2017) use a nonparametric approach. The functional form for the skewed t distribution is motivated by Fernandez and Steel (1998) and further explored and refined in Giot and Laurent 2003 and Lambert and Laurent 2002; see also Boudt, Peterson, and Croux 2008. Alternative specifications for the skewed t distribution are present in literature—for example, as put forth by Hansen (1994) and Azzalini and Capitanio (2003). These are essentially equivalent given a nonlinear transformation of the skewness parameter.

choice of a skewed t functional form is advantageous from the perspective of flexibility. For example, v → ∞, f ( y; μ, s, v, ξ) is characterized by tail proper- ties resembling a Gaussian distribution. Moreover, the density is symmetric for ξ = 1 .

References Adrian, Tobias, Nina Boyarchenko, and Domenico Giannone.

2016. “Vulnerable Growth.” Federal Reserve Bank of New York Staff Report 794.

Adrian, Tobias, and Nellie Liang. 2016. “Monetary Policy, Financial Conditions, and Financial Stability.” Federal Reserve Bank of New York Staff Report 690.

Adrian, Tobias, Emanuel Moench, and Hyun-Song Shin. 2010. “Macro Risk Premium and Intermediary Balance Sheet Quantities.” IMF Economic Review 58 (1): 179–207.

Adrian, Tobias, and Hyun-Song Shin. 2014. “Procyclical Leverage and Value-at-Risk.” Review of Financial Studies 27 (2): 373–403.

Aguiar, Mark, and Gita Gopinath. 2007. “Emerging Market Business Cycles: The Cycle Is the Trend.” Journal of Political Economy 115 (1): 69–102.

Ang, Andrew, Monika Piazzesi, and Min Wei. 2006. “What Does the Yield Curve Tell Us about GDP Growth?” Journal of Econometrics 131 (1–2): 359–403.

Azzalini, Adelchi, and Antonella Capitanio. 2003. “Distributions Generated by Perturbations of Symmetry with Emphasis on a Multivariate Skew t Distribution.” Journal of the Royal Statisti- cal Society: Series B 65: 367–89.

Bernanke, Ben S. 1983. “Nonmonetary Effects of the Financial Crisis in the Propagation of the Great Depression.” American Economic Review 73 (3): 257–76.

———, and Alan S. Blinder. 1992. “The Federal Funds Rate and the Channels of Monetary Policy Transmission.” Ameri- can Economic Review 82 (4): 901–21.

Bernanke, Ben S., and Mark Gertler. 1989. “Agency Costs, Net Worth, and Business Fluctuations.” American Economic Review 79 (1): 14–31.

———, and Simon Gilchrist. 1999. “The Financial Accelerator in a Quantitative Business Cycle Framework.” In Handbook of Macroeconomics, edited by J. B. Taylor and M. Woodford. Amsterdam: Elsevier.

Bianchi, Javier. 2011. “Overborrowing and Systemic External- ities in the Business Cycle.” American Economic Review 101 (7): 3400–26.

———, and Enrique Mendoza. Forthcoming. “Optimal Time-Consistent Macroprudential Policy.” Journal of Polit- ical Economy.

Boudt, Kris, Brian G. Peterson, and Christophe Croux. 2008. “Estimation and Decomposition of Downside Risk for Portfolios with Non-Normal Returns.” Journal of Risk 11 (2): 79–103.

117

C H A P T E R 3 F I N A N C I A L C O N d I T I O N S A N d G R O w T h A T R I S k

International Monetary Fund | October 2017

Brunnermeier, Markus K., and Lasse Heje Pedersen. 2009. “Market Liquidity and Funding Liquidity.” Review of Finan- cial Studies 22 (6): 2201–38.

Brunnermeier, Markus K., and Yuliy Sannikov. 2014. “A Macro- economic Model with a Financial Sector.” American Economic Review 104 (2): 379–421.

Caballero, Ricardo J., and Arvind Krishnamurthy. 2006. “Bub- bles and Capital Flow Volatility: Causes and Risk Manage- ment.” Journal of Monetary Economics 53 (1): 35–53.

Campbell, John Y. 1999. “Asset Prices, Consumption and the Business Cycle.” In Handbook of Macroeconomics, edited by John B. Taylor and Michael Woodford. Amsterdam: Elsevier.

———, Martin Lettau, Burton G. Malkiel, and Yexaio Xu. 2001. “Have Individual Stocks Become More Volatile? An Empirical Exploration of Idiosyncratic Risk.” Journal of Finance 56 (1): 1–43.

Christiano, Lawrence J., Martin Eichenbaum, and Charles L. Evans. 2005. “Nominal Rigidities and the Dynamic Effects of a Shock to Monetary Policy.” Journal of Political Economy 113 (1): 1–45.

Claessens, Stijn, Ayhan Kose, and Marco Terrones. 2011a. “Financial Cycles: What? When? How?” CEPR Discussion Paper 8379, Centre for Economic Policy Research, London.

———. 2011b. “How Do Business and Financial Cycles Inter- act?” CEPR Discussion Paper 8396, Centre for Economic Policy Research, London.

Corradi, Valentina, and Norman Swanson. 2006. “Predictive Density Evaluation.” In The Handbook of Economic Forecast- ing, edited by Graham Elliott, Clive W. J. Granger, and Allan Timmermann. Amsterdam: Elsevier.

De Nicolò, Gianni, and Marcella Lucchetta. 2017. “Forecasting Tail Risks.” Journal of Applied Econometrics 32 (1): 159–70.

Del Negro, Marco, and Frank Schorfheide. 2009. “Monetary Policy Analysis with Potentially Misspecified Models.” Ameri- can Economic Review 99 (4): 1415–50.

Doz, Catherine, Domenico Giannone, and Lucrezia Reich- lin. 2011. “A Two-Step Estimator for Large Approximate Dynamic Factor Models Based on Kalman Filtering.” Journal of Econometrics 164 (1): 188–205.

Duda, Richard O., Peter E. Hart, and David G. Stork. 2001. Pattern Classification. New York: Wiley-Interscience.

Estrella, Arturo, and Gikas A. Hardouvelis. 1991. “The Term Structure as a Predictor of Real Economic Activity.” Journal of Finance 46 (2): 555–76.

Estrella, Arturo, and Frederick S. Mishkin. 1998. “Predicting U.S. Recessions: Financial Variables as Leading Indicators.” Review of Economics and Statistics 80 (1): 45–61.

Fernandez, Carmen, and Mark F. J. Steel. 1998. “On Bayesian Modelling of Fat Tails and Skewness.” Journal of the American Statistical Association 93 (441): 359–71.

Friedman, Benjamin M., and Kenneth N. Kuttner. 1998. “Indi- cator Properties of the Paper-Bill Spread: Lessons from Recent Experience.” Review of Economics and Statistics 80 (1): 34–44.

Galí, Jordi, and Tommaso Monacelli. 2005. “Monetary Policy and Exchange Rate Volatility in a Small Open Economy.” Review of Economic Studies 72 (252): 707–34.

Gertler, Mark, Nobuhiro Kiyotaki, and Andrea Prestipino. 2017. “Wholesale Banking and Bank Runs in Macroeco- nomic Modeling of Financial Crises.” NBER Working Paper 21892, National Bureau of Economic Research, Cambridge, MA.

Gertler, Mark, and Cara S. Lown. 1999. “The Information in the High Yield Bond Spread for the Business Cycle: Evidence and Some Implications.” Oxford Review of Economic Policy 15 (3): 132–50.

Gilchrist, Simon, and Egon Zakrajšek. 2012. “Credit Spreads and Business Cycle Fluctuations.” American Economic Review 102 (4): 1692–720.

Giot, Pierre, and Sébastien Laurent. 2003. “Value-at-Risk for Long and Short Trading Positions.” Journal of Applied Econo- metrics 18 (6): 641–64.

Gorton, Gary, and Guillermo Ordoñez. 2014. “Collateral Cri- ses.” American Economic Review 104 (2): 343–78.

Gruss, Bertrand. 2014. “After the Boom—Commodity Prices and Economic Growth in Latin America and the Caribbean.” IMF Working Paper 14/154, International Monetary Fund, Washington, DC.

Hansen, Bruce E. 1994. “Autoregressive Conditional Density Estimation.” International Economic Review 35 (3): 705–30.

Hatzius, Jan, Peter Hooper, Frederic S. Mishkin, Kermit L. Schoenholtz, and Mark M. Watson. 2010. “Financial Conditions Indexes: A Fresh Look after the Financial Crisis.” NBER Working Paper 16150, National Bureau of Economic Research, Cambridge, MA.

He, Zhiguo, and Arvind Krishnamurthy. 2013. “Intermediary Asset Pricing.” American Economic Review 103 (2): 732–70.

International Monetary Fund (IMF). 2017. “Drivers of Capital Flows and the Role of Investor Base in Latin America.” Regional Economic Outlook: Western Hemisphere, May, Washington, DC.

Izenman, Alan J. 2013. Modern Multivariate Statistical Tech- niques: Regression, Classification, and Manifold Learning. New York: Springer-Verlag.

Katagiri, Mitsuru. Forthcoming. “Macroprudential and Mone- tary Policy in a Small Open Economy.” IMF Working Paper, International Monetary Fund, Washington, DC.

Kiyotaki, Nobuhiro, and John H. Moore. 1997. “Credit Cycles.” Journal of Political Economy 105 (2): 211–48.

Koenker, Roger. 2005. Quantile Regression. New York: Cam- bridge University Press.

Komunjer, Ivana. 2013. “Quantile Prediction.” In Handbook of Economic Forecasting, edited by Graham Elliott and Allan Timmermann. Amsterdam: Elsevier.

Koop, Gary, and Dimitris Korobilis. 2014. “A New Index of Financial Conditions.” European Economic Review 71 (C): 101–16.

118

G L O B A L F I N A N C I A L S T A B I L I T Y R E P O R T : I S G R O w T h A T R I S k ?

International Monetary Fund | October 2017

Korinek, Anton, and Alp Simsek. 2016. “Liquidity Trap and Excessive Leverage.” American Economic Review 106 (3): 699–738.

Krishnamurthy, Arvind, and Tyler Muir. 2016. “How Credit Cycles across a Financial Crisis.” Stanford University Working Paper.

Lambert, Philippe, and Sébastien Laurent. 2002. “Modelling Skewness Dynamics in Series of Financial Data Using Skewed Location-Scale Distributions.” Discussion Paper, Institut de Statistique, Université Catholique de Lou- vain, Belgium.

Laurent, Robert D. 1988. “An Interest Rate–Based Indicator of Monetary Policy.” Federal Reserve Bank of Chicago Economic Perspectives 12 (1): 3–14.

Philippon, Thomas. 2009. “The Bond Market’s q.” Quarterly Journal of Economics 124 (3): 1011–56.

Primiceri, Giorgio E. 2005. “Time Varying Structural Vector Autoregression and Monetary Policy.” Review of Economic Studies 72 (3): 821–52.

Stock, James H., and Mark W. Watson. 1989. “New Indexes of Coincident and Leading Economic Indicators.” In NBER Macroeconomics Annual 1989, edited by O. J. Blanchard and S. J. Fischer. Cambridge, MA: National Bureau of Eco- nomic Research.

———. 2003. “Forecasting Output and Inflation: The Role of Asset Prices.” Journal of Economic Literature 41 (3): 788–829.

Tay, Anthony S., and Kenneth F. Wallis. 2000. “Density Fore- casting: A Survey.” Journal of Forecasting 19 (4): 235–54.

To order, visit bookstore.imf.org/gf1017

Highlights from IMF Publications

FREE DOWNLOAD

AVAILABLE

COMING SOON

Blank

Global Financial Stability Report, October 2017

G lobal Financial Stability R

eport Global Financial Stability Report

Wo r l d E c o n o m i c a n d F i n a n c i a l S u r v e y s

I N T E R N A T I O N A L M O N E T A R Y F U N D

17OCT

IM F

O C T

Is G row

th at R isk?

Is Growth at Risk?

106735_Ch 00 FM_web
106735_Ch 01_web
106735_Ch 02_web
106735_Ch 03_web
106735_Ch 04 Ad_web
Blank Page