Explore the link between Financial Structures and Economic Growth

 

SN: 15085442

SSEESGS51: Financial Development Economic Growth and Financial Development in the Post-Soviet Region

Word Count: 2,999

  1  

Using a sample of 11 post-Soviet countries over a timespan of 17 years (from 1996-2013), we will perform a regression analysis using a fixed-effects estimator to determine the association that financial development – measured with proxies such as credit provided to the private sector by banks (a measure of the size of the banking sector) and the difference between the lending and deposit rate (the “interest rate spread”, which is a measure of the efficiency of the banking sector) – has with economic growth, measured by the growth rate of real GDP per capita. Our findings indicate that financial development measured by the interest rate spread is negatively and significantly associated with economic growth. This confirms our hypothesis that more efficient financial systems play a role in channelling economic growth. However, we simultaneously find that there is no statistically significant association between domestic credit provided to the private sector by banks and economic growth, and that the coefficient signs for the current and lagged versions of this indicator are inconclusive. This, however, is likely a result of the fact that the region has endured a series of banking crises related to debt and increased levels of leverage during the time period of our study, especially throughout the 1990’s and then later during the financial crisis of 2008-9. I. Literature Review on the Relationship Between Financial Development and Economic Growth and as it relates to the Post-Soviet Region There is a large amount of literature, both empirical and theoretical, on the link between financial development and economic growth. From the theoretical angle, Schumpeter (1911) argued that the services provided by financial intermediaries – namely in mobilizing savings, managing and allocating risk, providing oversight over company managers, and evaluating projects (Levine and King, 1993) – are crucial for technological innovation and economic development. Researchers in later years would seek to prove this empirically. Goldsmith (1969), for example, was one of the first to record the link between financial development and economic growth with a monumental study of 35 countries over 150 years, using the ratio of the market value of financial claims (i.e. bank/financial intermediary assets) to GNP (i.e. national wealth) as the proxy for financial development. While his findings were generally inconclusive due to the difficulty in obtaining reliable data, he was able to observe relatively conclusively that the role that financial intermediaries play in the financing process is indeed linked to the development of the economy and future economic growth: “periods of rapid economic growth have been accompanied…by the above-average rate of financial development (Goldsmith, 1969).” Although groundbreaking in both the size/diversity of his sample and topic of focus, there were nevertheless several deficiencies in Goldsmith’s research, with scholars improving on his work in later years. Levine and Renelt (1992) noted, for example, that Goldsmith did not control for other factors that influence economic growth, potentially leading him to overestimate the influence that financial development. King and Levine (1993), in an attempt to build off of Schumpeter’s theory and improve on Goldsmith’s work, performed a cross-country, time series study on 80 countries, measuring four indicators of financial development and their impact on four growth indicators. More importantly, the four indicators of financial development identified by King and Levine are the following:

1) the size of the formal financial intermediary system, measured by the ratio of liquid liabilities to GDP (otherwise known as the classical definition of “financial depth”);

  2  

2) the importance of private financial intermediaries in the provision of credit over that of

central banks (measured by the ratio of credit issued by private banks to total credit issued (which includes credit issued by the central bank). The argument here is that private banks are more effective at evaluating risk and collecting/analysing information than central banks;

3) credit provided to non-financial private firms in terms of total credit; and

4) credit provided to non-financial private firms as ratio to GDP.

The notion behind indicators (3) and (4) is that banks that issue credit to private firms are likely better able to provide the services identified by Schumpeter (1911) than banks that primarily channel money to the government. King and Levine, ultimately found that the level of financial development was a predictor for future economic growth, measured in terms of either of GDP per capita, domestic investment as %GDP, rate of physical capital accumulation, or improvements in the efficiency of physical capital allocation (King and Levine, 1993). Which way, however, does causality run? Does financial development cause economic growth or vice versa? Studies using panel data, which is considered more robust given that it accounts for large differences between countries, have generally found that causality runs from financial development to economic growth. Using GMM dynamic panel estimators, Levine et al (2000) found that there is a positive, casual link between and running from financial intermediary development (represented mostly by the indicators described above from his 1993 paper) and economic growth from a sample of 74 countries over the period of 1960-1995 (using 5-year intervals). As it relates to our region of study, Koivu (2002) used a fixed-effects panel model estimator to analyse the link between banking sector efficiency and the size of the banking sector and economic growth using a sample of 25 EBRD transition countries (i.e. the countries of the former Soviet Union, Yugoslavia, and post-communist Eastern Europe) during a 7-year timeframe (1993-2000). Using the interest rate margin (i.e. lending rate minus deposit rate) as well as domestic credit to private sector (%GDP) to assess the level of association of financial development with economic growth (real GDP growth), Koivu found that there is negative and significant association between the interest rate margin and economic growth (especially in the group of CIS countries); however, domestic credit to the private sector is only weakly positively correlated with economic growth and is even negatively correlated at a lagged-1 level. This observation was attributed to the fact that previous increases in domestic credit caused serious debt problems for these transition countries when economic crises hit, which affected future growth. Little research on financial development in the post-Soviet region followed Koivu (2002) until Cojocaru (2011), who used panel data and employed GMM and fixed effects model estimators on the same group of countries for the period of 1990-2008 to determine the relationship between financial development and growth. Ultimately, Cojocaru (2011) found that domestic credit to the private sector provided by banks does exhibit a positive and significant association

  3  

with economic growth, in contrast with Koivu’s (2002) findings. Of note is that Cojocaru’s analysis ends at 2008, before the effects of the financial crisis of 2008-9 could be adequately incorporated. Our timeframe will include the most recent financial crisis, and we expect this to have an impact on the results of our analysis, especially as they relate to credit to the private sector. II. Data and Preliminary Observations This paper will mostly build off of the work of Koivu (2002), and therefore we will be using many of the same variables in our model. Our dependent variable will be real GDP per capita growth (GROWTH). The group of countries we are analysing consists of: Armenia, Azerbaijan, Belarus, Georgia, Kyrgyzstan, Latvia, Lithuania, Moldova, Russia, and Ukraine. Kazakhstan, Uzbekistan, Tajikistan, and Turkmenistan were excluded due to lack of crucial data. Our dataset is strongly balanced and suffers from very few missing observations. Financial Development Explanatory Variables The variable MARGIN, which is the difference between the lending and deposit rates (the interest rate spread), will allow us to measure the effectiveness of the banking sector. A larger, positive interest rate spread, which would indicate a greater discrepancy between a low deposit rate and high lending rate, indicates that there are high transaction costs in the banking sector that hamper the effectiveness of banks to serve as efficient financial intermediaries and channels for investment. High transaction costs, which are reflected in a high lending rate compared to the deposit rate (thereby accentuating the rate spread), are a product of a banking system where risk is more difficult and costly (in terms of time and money) to ascertain due to opaque corporate structures, lack of reliable and public information, or weak regulatory enforcement, to name a few factors that contribute to banking sector inefficiency. Given that high transaction costs translate into higher borrowing costs, lending rates rise and SMEs, which form the backbone of any healthy and developed economy become, or remain, unable to secure financing for investment. The problem is accentuated in the post-Soviet region by the fact that banks dominate the financial sector, with financial markets remaining relatively undeveloped. High MARGIN is an issue that almost all countries in the reason faced in the 1990’s and continue to face, with the exception of the Baltic states, which have historically had low spread rates after having committed to financial sector reforms in the 90s in the lead-up to joining the EU in 2004. From a preliminary view of Graph-1, it appears that the interest rate spread amongst most countries was the most volatile in the early 1990’s into the early 2000’s with some upwards movement during the financial crisis of 2008-2011. We see some extreme values in Russia, for example, where there was a huge decline in the interest rate spread between 1996-1997, followed by a slight increase afterwards coinciding with the 1998 debt crisis, and then a gradual decline over the 2000’s as the Russian economy grew.

  4  

Graph-1: Interest Rate Margin by Country (%) vs. Real GDP per capita Growth (%)

Source: World Bank As we look closer at these panels, we notice that there are some trends, especially in the 1990’s, which may produce results that run counter to our initial hypothesis. Looking at Russia again in Graph-2, we see that from 1998–1999, the interest rate spread grew simultaneously with economic growth, which runs counter to our original hypothesis on the relationship between economic growth and the interest rate spread. This may weaken the predicted negative association in our regression model. Another reason why MARGIN is so important for our measure of financial development against GROWTH is because the width of the interest rate spread tends to also indicate low deposit interest rates, which can be a significant problem during times of high inflation. As Ledeneva and Seabright (2000) point out, the costs of saving money are high when inflation is high, and especially so when inflation is higher than the deposit rate. In this case, there is no incentive for people to save their money at a bank (or save at all), as their cash would only quickly lose its worth. When people would rather spend their money rather than save it in a bank, banks become starved of the money they would use for lending, and potential domestic investors would face higher costs for borrowing – a cycle which naturally should lead to less economic development. Graph-3 of Kyrgyzstan exemplifies this phenomenon, with inflation consistently outpacing deposit rates. Unsurprisingly, Kyrgyzstan also exhibits some of the slowest on average rates of growth in our sample.

  5  

Graph-2: Interest Rate Spread (%) vs. Real GDP per capita Growth (%) – Russia

Source: World Bank Graph-3: Annual Inflation (%) vs. Deposit Interest Rates (%)

Source: World Bank The second explanatory variable in this set is CREDIT, which is domestic credit provided to the private sector by banks as a %GDP. This variable measures the size of the banking sector (as it relates to private sector). Countries with higher levels of financial development tend to exhibit

  6  

large values of CREDIT. The OECD average during the timeframe of this study (1996-2013) was 97%, (World Bank), whereas for Belarus, for example, this number was a paltry 18.1%. The only countries exhibiting near-OECD levels of CREDIT, as we can see from Graph-4 below, are Lithuania, Latvia, and Estonia, the last of which became an OECD member in 2010. When it comes to our sample region, we anticipate that CREDIT will hardly exhibit an association with economic growth. As we can see below, CREDIT has seen gradual upwards growth almost regardless of the direction of GROWTH in most panels (Estonia stands out in particular, however), except for pronounced movements (both up and down) during financial crises and particularly during the last one. Graph-4: Domestic Credit to Private Sector (%GDP)

Source: World Bank Macroeconomic Control Variables Our control variables are listed on Table-3 and were selected so as to control for other factors that are associated with GROWTH. Our variables GOVERNMENT, DOMESTIC, OPENNESS, INFLATION, and POPULATION are all standard control variables for GROWTH that have been used exhaustively in literature. L_GDP is also included as it controls for economic convergence, the phenomenon where transition countries starting from lower levels of national income grow rapidly at the start (Ayadi et al., 2015). OECD is a control variable used by Koivu (2002) to control for economic growth from linkages to and from developed countries (from

  7  

where most FDI originates). INDEX, is an average of five indices produced by EBRD that represents a score of a country’s economic development (the higher the index, the more developed the economy). It is used in Koivu (2002) with a lag of one year to account for the fact that the effects of reforms are usually not felt immediately. Comprehensive List and Important Summary Statistics Table-1: Dependent Variables

Dependent Variable Description Source GROWTH Real GDP per capita growth

(logged first-difference)

World Bank

Table-2: Independent Variables for Financial Development (FINANCE)

Independent Variable Description Source MARGIN Difference between the

lending and deposit rates

World Bank

CREDIT Domestic credit provided to private sector by banks

(%GDP)

World Bank

Table-3: Control Variables for Macroeconomic Performance (CONTROL)

Control Variable Description Source INFLATION Annual CPI Inflation (%)

World Bank

GOVERNMENT Government Expenditures as %GDP

World Bank

DOMESTIC Domestic Investment as %GDP

World Bank

POPULATION Annual Population Growth

World Bank

OPENNESS Sum of Imports and Exports of Goods and Services as

%GDP

World Bank

L_INDEX Simple Average of EBRD Transition Indicator Indices

EBRD

OECD Average Real GDP per capita Growth of OECD member

countries (logged first-difference)

OECD

  8  

L_GDP Lagged Real GDP per capita (logged)

World Bank

V. Estimation Given the above, our model is the following:

GROWTHi,t = β0,i + β1FINANCE + β2CONTROLS + ui,t,

where FINANCE represents our financial development explanatory variables and CONTROLS represents the set of controls listed above. β0,i is the individual dummy for each country (time invariant) and ui,t, is the error term. Following Koivu (2002), we will use a fixed-effects estimator for our regression. We use the contemporaneous and lagged value together of each financial development variable in our regression. This is because improvements in the financial system can take time to affect economic growth.

Table-4: Regression Output (1)

(2)

(3)

MARGIN -.182*

(-2.96)

- -.186* (-2.75)

MARGIN (L1) -.032 (-1.18)

- -.038 (.124)

CREDIT - 0.028 (1.05)

.020 (.433)

CREDIT (L1) - -.066 (-2.01)

-.038 (.057)

INFLATION -.007 (1.22)

-.018* (-2.98)

-.003 (-.46)

GOVERNMENT .018 (.08)

-.048 (-.21)

-.002 (-.01)

DOMESTIC .127 (1.220

.114 (1.07)

.094 (.88)

POPULATION -.151 (-.08)

-.684 (-.45)

-.764 (-.48)

OPENNESS .064* (2.58)

.073* 2.18

.071* (2.15)

L_INDEX 3.30 (.67)

6.59 (1.18)

3.91 (.71)

OECD 1.64*** (6.97)

1.59*** (5.45)

1.58*** (5.58)

L_GDP -5.97*** (.018)

-2.62 (-1.20)

-4.58 (-2.29)

  9  

Observations 181 181 181

R-Squared Within

.4347 .3797 .2590

R-Squared Between

.0427 .0716 .0036

R-Squared Overall

.1133 .0861 .0203

MARGIN and its lag are negatively associated with economic growth, with the contemporaneous value being significant as we see in the first row in columns (1) and (3). This means that a one level (equivalent to 1% given the measurements) decrease in MARGIN is associated with a .183 percentage point increase in GROWTH. That is to say that if Kyrgyzstan’s 2002 MARGIN (22%) was equivalent to Latvia’s (4.73%) during the same period, its GROWTH would have been 3.23 percentage points higher. Nothing concrete can be said about CREDIT, as expected. In its contemporaneous state, it is positively associated with GROWTH; however, when lagged, it is negatively associated. This is most likely given the fact CREDIT has a tendency in our sample to rise (in some cases significantly) prior to a financial crisis, especially prior to 2008/9 for many panels. VI. Conclusions and Reservations These results, which were checked for robustness, have a number of implications for policymakers, as well as for researchers studying financial development in the post-Soviet region. First, given the impact that a lower level of MARGIN could have for GROWTH, as demonstrated briefly above, policymakers in the region need to act to reduce transaction costs in the banking sector. This would include making the sector more transparent (such as in improving accounting standards or requiring that certain information be kept in the public realm) and enforcing clearer regulations on a regular and impartial basis, among other possible actions. As it relates to CREDIT, policymakers need to be wary on how to channel economic growth from this area of financial development. Encouraging more private credit without enforcing better lending and borrowing practices can lead to the accumulation of toxic non-performing loans and greater banking sector instability during financial crises when lending rates soar. For researchers, this study furthers Koivu’s (2002) conclusion that CREDIT, for the time being, is not a proper means of measuring financial development against economic growth in this particular region. This could be interpreted as saying that expanding credit to the private sector does not affect economic growth when the private sector is not ready to channel that finance into investments contributing to economic growth. Reservations These results are not without their reservations. We used fixed-effects estimations on a static panel in order to derive our coefficients despite the fact that models with time series elements containing economic components – especially as they relate to growth – tend to exhibit dynamic qualities (i.e. we may have needed to include the lagged GROWTH variable on the right side of the equation as a regressor). A GMM estimator, which could also control for endogeneity issues

  10  

with instruments, may have been more appropriate. Nevertheless, the results produced above are sensible and roughly correlate with other studies. References

• Ayadi, Rym, Emrah Arbak, Sami Ben Naceur, and Willem Pieter De Groen. 2015. “Financial Development, Bank Efficiency, and Economic Growth Across the Mediterranean.” In Economic and Social Development of the Southern and Eastern Mediterranean Countries, 219–33. Springer.

• Beck, Thorsten, Ross Levine, and Norman Loayza. 2000. “Finance and the Sources of

Growth.” Journal of Financial Economics 58 (1-2): 261–300. doi:10.1016/S0304- 405X(00)00072-6.

• Cojocaru, Laura, Evangelos M. Falaris, Saul D. Hoffman, and Jeffrey B. Miller. 2016.

“Financial System Development and Economic Growth in Transition Economies: New Empirical Evidence from the CEE and CIS Countries.” Emerging Markets Finance and Trade 52 (1): 223–36. doi:10.1080/1540496X.2015.1013828.

• Goldsmith, R.W. 1969. Financial Structure and Development. Studies in Comparative Economics. Yale University Press.

• King, Robert G, and Ross Levine. 1993. “Finance and Growth: Schumpeter Might Be Right.” The Quarterly Journal of Economics, 717–37.

• Koivu, Tuuli. 2002. “Do Efficient Banking Sectors Accelerate Economic Growth in

Transition Countries.”

• Levine, Ross, and David Renelt. 1992. “A Sensitivity Analysis of Cross-Country Growth Regressions.” The American Economic Review 82 (4): 942–63.

• Schumpeter, J.A. 1934. The Theory of Economic Development: An Inquiry Into Profits,

Capital, Credit, Interest, and the Business Cycle. Economics Third World Studies. Transaction Books.

• Seabright, Paul, and Alena Ledeneva. 2000. The Vanishing Rouble: Barter Networks and Non-Monetary Transactions in Post-Soviet Societies. Cambridge University Press.