Country Risk Analysis (CRA)

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/265295771

Country-risk measurement and analysis: A new conceptualization and

managerial tool

Article  in  International Business Review · August 2014

DOI: 10.1016/j.ibusrev.2014.07.012

CITATIONS

23 READS

2,238

3 authors, including:

Some of the authors of this publication are also working on these related projects:

Entrepreneurial market behavior and cognition View project

Regulation and Corporate Governance View project

Christopher L Brown

Georgia State University

1 PUBLICATION   23 CITATIONS   

SEE PROFILE

S. Tamer Cavusgil

Georgia State University

241 PUBLICATIONS   20,334 CITATIONS   

SEE PROFILE

All content following this page was uploaded by S. Tamer Cavusgil on 10 June 2016.

The user has requested enhancement of the downloaded file.

International Business Review 24 (2015) 246–265

Country-risk measurement and analysis: A new conceptualization and managerial tool

Christopher L. Brown a,b,d,*, S. Tamer Cavusgil c, A. Wayne Lord a,d

a World Affairs Council of Atlanta, Georgia State University, GSU Buckhead Center, Tower Place 200, Suite 500, 3348 Peachtree Road NE, Atlanta, GA 30326,

United States b Department of Political Science, College of Arts and Sciences, Georgia State University, Atlanta, GA, United States c Institute of International Business, Center for International Business Education and Research (CIBER), Robinson College of Business, Georgia State University,

United States d Robinson College of Business, Georgia State University, United States

A R T I C L E I N F O

Article history:

Received 30 March 2012

Received in revised form 28 February 2014

Accepted 18 July 2014

Available online 14 August 2014

Keywords:

Big data

Country risk

Critical thinking

Education

Index

Organizational context

State development

Strategy

A B S T R A C T

Country risk analysis has been a topic of investigation for decades, often focused on forecasting the risks

to business profitability and assets when investing in a country. While there have been gradual

improvements in the analytic techniques and overall breadth of the research, many scholars and

practitioners continue to focus on limited conceptualizations of risk, measures with a relatively small

number of variables, and/or expert analysis. Others point to the need to expand the inquiry, produce

better tools and models, take advantage of the greater availability of data and enhanced computing

techniques, and tackle puzzles differently. Advancing this discussion, we make the case for a new

conceptualization and measurement of country-level risk and introduce the Robinson Country Risk

Index (RCRI), a tool which incorporates four broad dimensions—Governance, Economics, Operations, and

Society (GEOS). Within this holistic macrostructure, the RCRI encompasses 70 sub-dimensions, 126

countries, and, at present, 8 years of data. Its ecological conceptualization, multifaceted goals, and

embedded functionalities complement and offer advantages over other risk indexes. The RCRI addresses

concerns surrounding the conceptualization and measurement of country risk and provides a dynamic

new instrument for educators, researchers, and practitioners.

� 2014 Elsevier Ltd. All rights reserved.

Contents lists available at ScienceDirect

International Business Review

jo u rn al h om epag e: ww w.els evier .c o m/lo cat e/ ibu s rev

1. Introduction

‘‘Keynes. . . used to say that his best ideas came to him from ‘messing about with figures and seeing what they must mean.’ He could be as excited as any economist at discovering correlations in the data. Yet he was famously skeptical about econometrics—the use of statistical methods for forecasting the future. He championed the cause of better statistics not to provide material for the regression coefficient, but for the intuition of the economist to play on. He believed that statistical information in the hands of the philosophically untrained was a dangerous and misleading toy.’’

* Corresponding author at: World Affairs Council of Atlanta, United States.

Tel.: +1 404 413 6198; fax: +1 404 781 0938.

E-mail addresses: [email protected] (C.L. Brown),

[email protected] (S.T. Cavusgil), [email protected] (A.W. Lord).

http://dx.doi.org/10.1016/j.ibusrev.2014.07.012

0969-5931/� 2014 Elsevier Ltd. All rights reserved.

Robert Skidelsky, ‘‘The Return of the Master’’

It is now commonplace to theorize and examine how 21st century globalization is changing the playing field for businesses, governments, and non-governmental organizations (see, for example, Atsmon, Child, Dobbs, & Narasimhan, 2012; Cusimano, 2007; Friedman, 2008; Isdell, 2009; Khanna, Palepu, & Sinha, 2005; Perez-Aleman & Sandilands, 2008; Porter & Kramer, 2011; Reich, 2007). Strategic thinkers can be overwhelmed by the many different types of state-level and other challenges they face as they try to plan, carry out operations, invest, or achieve a wide range of other goals. Intertwined with the multitude of state-level risks leaders must consider are vast differences among countries in such areas as history, size, geography, culture, language, ethnic diversity, and other contextual dynamics. To be sure, assessing country risk has been a topic of academic investigation for decades, often focused on the risks to business profits and assets when investing in a country. Yet, while there have been gradual improvements in the analytic techniques and overall breadth of the research, many researchers and practitioners continue to focus on limited conceptualizations of risk, measures with a relatively

C.L. Brown et al. / International Business Review 24 (2015) 246–265 247

small number of variables, and/or expert analysis (Bouchet, Clark, & Groslambert, 2003; Coccia, 2007; Cruces, Buscaglia, & Alonso, 2002; Funston & Wagner, 2010; Nath, 2008; Oetzel, Bettis, & Zenner, 2001).

More recently, Nath (2008) has focused on the urgency and opportunities to expand the inquiry, produce better models, and tackle new puzzles, pointing specifically to the changing global environment, greater availability of data, and enhanced computing techniques. Cukier and Mayer-Schoenberger (2013) point to the new abilities of researchers to use ‘‘all the data’’, not small subsets, as well as the opportunity ‘‘big data’’ offers to drill into the data, find nuances, and ‘‘unlock new forms of value.’’ Cavusgil, Kiyak, and Yeniyurt (2004) and Coccia (2007) address the challenges, and yet utility, in building taxonomic schemes for countries and variables. Ravallion (2012) assesses existing composite or ‘‘mash-up’’ indices and argues for stronger theoretical clarity and recognition of the ‘‘tradeoffs’’ conceptual foundations embody. He points to the sensitivity of indices to weighting and structural changes, as well as data quality, and the ongoing importance of country-specific contextual factors.

The present research seeks to advance this discussion by addressing existing conceptual issues in understanding and measuring country risk, while making the case for a new comprehensive tool, the Robinson Country Risk Index (RCRI), which puts a unique, dynamic, and integrated mix of functionali- ties in the hands of the user. By striving to be uniquely holistic, integrated, and interactive, the RCRI is complementary with many existing services and tools. Yet, by shifting the emphasis away from forecasting and prediction and toward dynamically leveraging extensive data at the country level to foster diverse strategic thinking and academic goals, the RCRI helps craft new forms of value (see comparative Table 1 in Section 2 and the discussion of results and uses to date in Section 5). The RCRI’s conceptualization rests on four broad dimensions—Governance, Economics, Opera- tions, and Society (GEOS)—across 70 sub-dimensions, 126 coun- tries, and, at present, 8 years of data. Countries are ranked according to their overall aggregate level of risk across a wide range of factors; the investigator is actively able to ‘‘drill down’’ and focus on any of the 273 variables; time-series variable and country cross referencing, weighting manipulation, and other functionalities are put at the user’s fingertips; countries are clustered by region (Africa, East Asia, Europe, Former Soviet Union, Latin America, Middle East, North America, Oceania, and South Asia) and perceived level of development (Advanced, Developed, Emerging, Frontier, and Least Developed); and a number of embedded or otherwise available modeling and statistical techniques allow the user to transform the index given a specific strategic construct or academic investigation.

Extant indices include portions of the conceptualizations or functionalities found in the RCRI. However, we argue that the RCRI helps address the concerns and opportunities outlined above, offers a new instrument for researchers, educators, and organiza- tional leaders, and serves as a robust and comprehensive alternative measure of country development. In harmony with the introductory quote about John Maynard Keynes above, this tool is best used in conjunction with comparative-historical and qualitative approaches, such as those found throughout the development and comparative political economy literature (see, for example, Acemoglu & Robinson, 2012a; Gourevitch, 1986; Haggard, 1990; Hall, 1989; Katzenstein, 2005; Kohli, 2004; Moore, 1966; Park, 1984; Rapley, 2007; Sachs, 2005; Sen, 1999) and in some existing risk services (see a sampling in Table 1).

In the remainder of this paper, we first address the relevant background and literature on country risk and discuss why we believe the RCRI adds to the understanding of country risk conceptualization and measurement. As we progress, we examine

issues surrounding the holistic approach, dynamic interactivity, variable and country clustering, rank ordering, and parsimony in the age of big data. We conclude by discussing RCRI results and use to date, as well as implications and future avenues for investiga- tion.

2. Country risk and taking a holistic, big data approach

Country risk can be broadly defined as the probability of particular future events within a state that could have an adverse effect on the functioning of a given organization (or, for that matter, an individual), whether that organization be a business, government agency, non-governmental organization (NGO), or other type of body (see, for example, Bouchet et al., 2003; Fitzpatrick, 1983; Harland, Brenchley, & Walker, 2003; Jensen & Young, 2008). The multidimensionality inherent within this definition suggests that the specific factors underlying risk change with the context of the organization involved and the specific operationalization of the dependent variable. What is a significant risk factor to one business or agency, for example logistics bottlenecks for an NGO such as CARE, might be an opportunity for another, for example a logistics supplier such as UPS. A certain type of risk, such as a prevalence of malaria or water scarcity, may be much more important to one organization than it is to another, and possibly an opportunity for yet another. A risk might be a concern because of potential short-term profit losses for one organization, because of human suffering for another, and because of national security threats to yet another. This lack of specificity has posed challenges for those seeking to conceptualize and measure country risk, both broadly as well as with respect to specific organizational profiles (Oetzel et al., 2001).

Taleb, Goldtein, and Spitznagel (2009) and Funston and Wagner (2010) point out that broad-based risk intelligence and manage- ment are key means to the end of not just organizational survival, but also organizational value creation. As such, researchers and risk services have worked to tackle the conceptual issues surrounding country risk. Broadly, these efforts can be categorized as qualitative and quantitative, though there is often overlap (Bouchet et al., 2003; Coccia, 2007; Nath, 2008). Qualitative assessments generally attempt to tackle head on the complexity of the political, economic, and social aspects of risk without sacrificing granularity and context, often weaving key statistics into their analysis. They generally rely on the perceptions of expert analysts and can sometimes lack a structured format, making it difficult for users to compare countries. Still, whether the format is structured or unstructured, the fullness of each country can be examined in extensive detail. The Economist Intelligence Unit (EIU), Stratfor, and Business Environment Risk Intelligence (BERI) offer risk analyses which fit in this category. Also in the qualitative grouping are services, such as the World Economic Forum’s Global Competitiveness Report (GCR), which ask for expert, Likert-scale, perceptive scoring from lowest to highest across a menu of variables. The benefit here is that the final average scores lend themselves to quantitative analysis.

Some quantitative measures and services, such as those provided by Political Risk Services’ (PRS), Maplecroft, and Roubini

Global Economics, attempt to rank order countries relative to each other or otherwise give them a risk rating. Rank ordering can be broad across all countries, with respect to clusters of countries (such as ‘‘Latin America’’ or ‘‘Emerging Markets’’), or at the micro- risk level (such as ‘‘infectious diseases’’ or ‘‘transportation infrastructure’’). Risk Ratings can be through qualitative expert perceptions (noted above) or based on hard data (or both). This field of inquiry includes a variety of methodologies, which vary given the different goals and conceptualizations of the investiga- tion involved. Some rank orderings are not necessarily focused on

Table 1 Comparison to relevant sources.

Source Description/Goals Structure Functionality RCRI Comparison

Economist Intelligence

Unit (EIU)

The EIU is a commercial

service that provides

extensive raw data,

qualitative country and

sector risk analysis and

forecasts, and content tied to

being part of the Economist

magazine. The EIU also offers

specialized indices, such as its

Democracy index and Where

to be born index. Its risk

briefing seeks to quantify the

risks to business profitability

and forecast the coming two

years.

The EIU’s various reports help

put the written context and

analysis around its wide-ranging

data resources. Its risk briefing is

centered on 10 dimensions

(security, political stability,

government effectiveness, the

legal and regulatory

environment, macroeconomic

risks, foreign trade and

payments issues, labor markets,

financial risks, tax policy, and the

standard of local infrastructure)

and includes 66 qualitative and

quantitative indicators used to

forecast future business risk, as

opposed to simply extrapolating

present trends into the future.

The EIU’s rich raw data tool

allows for some longitudinal

graphing for selected countries.

Its Risk Briefing includes a

customizable risk tracker which

looks at ranks and scores and

allows for the construction of

matrices (choosing countries,

risk dimensions, and industry

sectors).

The EIU’s various reports,

briefings, and articles

complement the RCRI,

helping put the written

context and analysis around

the numbers. Also, although

the RCRI does not incorporate

EIU raw data because of

copyright issues, some of the

data points the EIU provides

are the same as those

incorporated into the RCRI

from other sources. The

service’s risk briefing

includes a few broadly

similar dimensions and

variables to the GEOS

macrodimensions, but

emphasizes business risk and

forecasting. It does not share

the RCRI’s holistic, dynamic,

and academic approach

aimed at differential

diagnoses of countries

generally and given a wide

variety of risk lenses.

Maplecroft Maplecroft is a commercial

service that through its

Global Risk Portfolio (GRP)

constructs some 200 indices

across a variety of country

risk categories. The goal of

the service’s quantitative

tools and qualitative analysis

is to monitor and forecast

country risks for

multinational companies,

financial institutions,

governments, and NGOs.

Maplecroft’s products (indices,

dashboards, interactive maps,

country scorecards, etc.)

measure country risk across

political, economic, societal, and

environmental factors. The

company incorporates 6 years of

data and 1500 variables to

construct 200 risk indices

covering 200 countries. It does

not aggregate these indices into

an overall structure, but includes

tools for customization.

The GRP includes extensive

functionality through their

interactive dashboards and risk

calculators. indices, maps, and

scorecards can be tailored by

sector and geographic business

interest.

While similar to the RCRI

with extensive scope and

coverage, Maplecroft’s Global

Risk Portfolio is markedly

different in overall

conceptualization and

variable taxonomy, not as

focused on holistic,

recallable, and interwoven

longitudinal dynamics, and

often tied to its own

perceptive scoring. It is part

of a commercial service that

emphasizes forecasting for its

clients. While some functions

are similar to the RCRI,

including weighting

manipulation, distinctions

come in such areas as

methods of variable

aggregation and drill down,

as well longitudinal

interaction and index

reconstruction.

Political Risk Services

(PRS) Group

PRS Group offers qualitative

and quantitative analysis

through a monthly journal,

the International Country

Risk Guide (ICRG), and a

Political Risk Service (PRS)

System, both tied to

quantitative forecasting. The

goal of the ICRG is to provide

an early country risk warning

system for multinational

firms, banks, and equity and

currency traders over a 1–5

year time period. The goal of

the PRS System is to provide

objective risk forecasting in

three investment areas:

financial transfers (banking

and lending), foreign direct

investment (e.g., retail,

manufacturing, mining), and

exports to the host country

markets. The PRS’ System

provides industry specific 18-

month and 5-year forecasts.

The ICRG quantitative analysis

covers 140 countries and is

based on 22 weighted variables

grouped into political, financial,

and economic risk categories,

with the 12 variables that make

up the political risk category

incorporating 15 additional

variables or ‘‘sub-components.’’

The ICRG’s political risk variables

are based on expert perceptions,

and its financial and economic

risk categories are based on

transformed hard data. The more

focused PRS System forecasts

risk for 100 countries and uses 3

regime scenarios and 11 types of

government intervention which

affect business climate to

calculate scores and convert

them into letter grades.

Users can request customized

ICRG reports with different

weights of risk factors based on

industry specific needs. ICRG raw

data is also available to

subscribers. The PRS System

offers country reports based on

perceptive data and forecasts. It

also offers an optional weighting

tool that allows users to

customize the system

forecasting model to their

individual projects by adding or

subtracting variables and

adjusting the model to fit specific

firm or project needs.

The focus and structure of

both the ICRG and the PRS

System is more limited than

the RCRI given the goals of the

PRS Group. The ICRG offers

some ability for users to

interact with the data, while

the PRS system is a less

dynamic tool, fitting to its

business forecasting goals.

PRS system reports offer

qualitative ratings and

analysis that can

complement the RCRI.

Neither the ICRG or the PRS

System incorporates the

RCRI’s aggregation

techniques or extensive

functionalities, given the

RCRI’s holistic, dynamic, and

academic approach.

C.L. Brown et al. / International Business Review 24 (2015) 246–265248

Table 1 (Continued )

Source Description/Goals Structure Functionality RCRI Comparison

Roubini Global Economics Roubini Global Economics

(RGE) is commercial service

which provides qualitative

and quantitative analysis of

macro-country risks, as well

as research, scenarios, and

10-year economic forecasts.

Through RGE Country

Insights, RGE provides the

Social, Institutional, and

Regulatory Risk Index (SIRR),

a tool used to identify

investment attractiveness

and business opportunity.

The SIRR aggregates ranks and

scores for 174 countries across

four ‘‘drivers’’ (social/political,

business environment/

regulatory quality, property

rights/corporate governance,

and government effectiveness).

It includes multiple sub-

dimensions and is constructed

with 155 variables. Data is

available from 2005 and

countries are clustered into

world region and development

categories.

The SIRR is a customizable tool

which allows the user to

compare countries across

variables of interest. It includes

graphing and charting functions

and describes itself as a

‘‘recursive model which shows

the impact of interactions

between different parts of the

economy.’’

Despite our limited access to

RGE, the SIRR clearly has

similarities to the RCRI, in as

much as it includes scores

and ranks as well as drill

down and some other

comparable functionalities.

Its structure and data sources

also show some parallels, but

many significant differences.

Through the SIRR, RGE is

focused on adding

quantitative predictive

analysis to its business risk

forecasting. The RCRI’s

conceptualization is based in

the social science and

development literature,

which points to a holistic, big

data, and integrative but fully

malleable GEOS approach to

foster critical and strategic

thinking, as well as other

academic goals.

Legatum Institute The Legatum Institute is a

nonpartisan think tank that

provides both quantitative

and qualitative analysis. Its

Legatum Prosperity Index

(LPI) ranks 142 countries and

seeks to provide insight into

how prosperity is forming

and changing across the

world. Prosperity is defined

broadly to include income

and well-being.

The LPI includes eight equally

weighted dimensions which are

determined to be the foundation

of prosperity (economy,

entrepreneurship and

opportunity, governance,

education, health, safety and

security, personal freedom, and

social capital). The index

incorporates 89 variables, one-

third perceptive and two-thirds

objective, across 5 years of data.

Each dimension or sub-index is

constructed using econometric

analysis to determine impact on

income and well-being. The

index is also sub-divided into 5

world regions.

The Prosperity Index webpage

provides extensive functionality,

including weighting

manipulation, mapping, various

charting and graphing functions,

and interactive ranks with

scores. It is not focused on

adaptability toward

organizational specific interests.

The LPI is analogous to the

RCRI in its broad integration

and interactive functionality;

however, the focus or goal of

the LPI is not on country risk,

leading to markedly different

dimensions. The LPI has some

data points similar to the

RCRI and transforms its data

in a similar way. However, its

use of econometric

techniques to garner weights

and variables offers a future

avenue of research for the

RCRI. Still, the RCRI’s focus on

dynamic adaptability toward

organizational interests leads

to a need to include a broader

set of variables and, upon

index reconstruction, either

manual assignment of

weights or dynamic and

complex recalculation of

weights. Stated differently,

the LPI seeks parsimony, and

we accept messiness and

need ‘‘all the data.’’

C.L. Brown et al. / International Business Review 24 (2015) 246–265 249

risk as specifically defined, but their subject matter is often tied to risk nonetheless, such as the Legatum Institute’s Legatum Prosperi- ty Index, which looks at country wealth and well-being, or the World Economic Forum’s Global Gender Gap Report (GGR), which is focused on gender equality.

Quantitative approaches to risk analysis and forecasting sometimes incorporate econometric and statistical modeling techniques, such as regression analysis or principal component analysis, for such reasons as finding relationships between variables endogenous or exogenous to the equation at hand, moving toward parsimony in explanation, or moving on to prediction (Bouchet et al., 2003; Coccia, 2007; Nath, 2008). Others have looked to insurance markets to see if they offer better estimators of political violence risk (Clark, 1997; Jensen & Young, 2008). In the end, country risk management often combines quantitative or qualitative approaches, and generally moves on to forecasting, particularly on the adverse effects to business and finance.

Given the breadth of the field, Table 1 contrasts the RCRI to the five services and/or indexes we find most instructive to understanding our index and its contribution. In a limited way, it also samples the scope of services/indexes available. Comparison is by goals and types of analysis, conceptualization and structure, and embedded functionalities. The RCRI is focused on the multidimensionality of country risk as defined above. As opposed to the forecasting and prediction goals found in many risk services, we emphasize the goals of building critical and strategic thinking, sparking intuition, and promoting academic investigation. If taken skeptically, probabilities and prediction can of course be valuable to these latter goals, and could be a possible avenue of future research for a numbers driven project such as the RCRI. Even critics like Oetzel et al. (2001), Taleb et al. (2009), and Funston and Wagner (2010) find that risk management services sometimes do well in projecting standard trends. Still, they often fall short, and are rarely able to forecast major crises, the ‘‘black swan’’ rare events that lead to substantial loss, disruption, or even a

C.L. Brown et al. / International Business Review 24 (2015) 246–265250

‘‘punctuated equilibrium’’ that results in substantial change. These authors focus on the organic risk intelligence strategies needed to best identify the range of risks an organization faces (including black swan events) and match available risk management resources to the priority of the risk (in a form of insurance). They seek to reframe how risk management is approached away from prediction and toward better preparedness, and emphasize risk management as a value generating part of an organization, helping it to survive and thrive. We embrace their skepticism and focus on the strategic and critical thinking and academic investigation goals, but also believe the best available data, dynamically and holistically conceived, is a good place to begin the analysis.

Nath (2008) argues that while country risk services have shown gradual improvements, they generally fall short in terms of scope and coverage. In their look at the extraordinary new uses being put to the explosion of data of all kinds, or ‘‘big data,’’ Cukier and Mayer-Schoenberger (2013) take this thinking a step further, pointing to the new abilities of researchers to use ‘‘all the data,’’ not small subsets, as well as the corresponding need to accept messiness. Echoing the opening quote of this article, they stress the insights of correlation while causes remain obscure, and point to the opportunity big data offers to drill into the data, find nuances, help address problems and puzzles, and create new forms of value. They also reflect Keynes’ concerns surrounding dangerous and misleading tools in the hands of the philosophically untrained, noting how big data has become a ‘‘battleground among countries’’ for issues such as commerce, privacy, or national security, or a tool of ‘‘Big Brother.’’ We concur with these authors that big data requires ‘‘a new way of thinking,’’ and argue that by uniquely pulling an extensive amount of country-level data into a dynamic GEOS whole (as opposed to seeking parsimony given elusive, organization specific dependent variables), and centering the RCRI on strategic thinking and academic investigation, this new tool helps re-center country risk conceptualization, measurement, analysis, and mitigation. Of course, given big data’s basis in correlations, probabilities, and messiness, such tools also require:

‘‘the human element of instinct, risk taking, accidents and even error. . .(T)here will be a special need to carve out a place for the human: to reserve space for intuition, common sense, and serendipity to ensure that they are not crowded out by data and machine made answers’’ (Cukier & Mayer-Schoenberger, 2013: 40).

The subset of the social science literature which focuses on state types, state development, and state success offers the conceptual grounding for thinking of countries as human-like, organic ecologies, as the GEOS conceptualization crudely tries to do. Sachs (2005) seeks to engage in ‘‘clinical economics,’’ think like a doctor, look at countries holistically as patients, and carry out ‘‘differential diagnoses’’ to cure the plague of poverty. He comes up with a seven part checklist to carry out these diagnoses and uses an analogy of states as complex organic bodies. Also thinking clinically, Acemoglu and Robinson (2012a) use extensive com- parative historical research to argue for the centrality (higher weighting) of inclusive versus extractive political institutions as drivers for why states succeed or fail. They point out that inclusive political institutions lead to inclusive economic institutions and better economic and social outcomes. They compare societies to similar but isolated organisms, with small institutional differ- ences leading to ‘‘genetic drift’’ where ‘‘genetic mutations cumulate.’’ In other words, their differential diagnoses point to somewhat different conclusions than Sachs, as can happen with clinical doctors. These differences have led to a spirited debate between these scholars (Acemoglu & Robinson, 2012b; Sachs, 2012a; Sachs 2012b).

Also a strong presence in this theoretical space, Kohli (2004) examines ‘‘spontaneous,’’ ‘‘late,’’ and ‘‘late-late’’ developing states and focuses on the dynamics that have led some to be ‘‘cohesive,’’ others ‘‘fragmented,’’ and still others ‘‘neo-patrimonial’’. His extensive comparative historical research concludes that patterns of state authority and effective state intervention have been critical in understanding the emergence of industrialization in developing economies. Rapley (2007) reinforces the importance of effective states and points to the direct connection between a state’s governance and its socioeconomic body:

‘‘. . . a theory that separates the state from society is rather like a medical lecturer who treats the human head and body as distinct. The head may govern the body, but that does not make it independent of the body’’ (Rapley, 2007: 162).

Other scholars tie into the analogy of states as living bodies (see, for example, Park, 1984; Sen, 1999). The key theoretical point here, and as discussed below, is not just the different insights, emphases, or weights given by these scholars in their varying analyses and diagnoses. Rather, we are concerned with the development of a state-centered, robust, holistically conceived-but-adaptable tool which allows users to carry out differential diagnoses of states through their own research or organizational lens. Hence, we have developed the RCRI and its dynamic GEOS framework as a step toward the organic holism and functionality we find missing in other tools.

The central argument of the present research is that the construction of a dynamic country risk index, based on an integrated and ecological conceptualization of country risk and a broad set of data, strengthens both the understanding and measurement of country risk. The RCRI provides a wider lens than most previous quantitative measures, while at the same time allowing the researcher or practitioner to calibrate or refine the focus, depending on the specific area of interest or organizational activity. It is a broad tool that can be sharpened, through weighting changes, the addition or subtraction of data, statistics, or even complete reconceptualization, given the purposes of the organiza- tional leader or researcher. We argue that, within its limits, the index can help tackle broad and narrow puzzles, some organic to its endogenous variables and conceptualization, others with the index or its component parts serving as independent or dependent variables, and others after the index has been reconceived. Given its basis in data, one avenue for future research may be to project probability trends using the RCRI to see if it can be a tool to add to this conversation. However, we believe the better focus is to accept messiness and correlation and seek to spark intuition, risk mitigation and resilience, managerial probing, research, classroom investigation, and other forms of value. As noted above we also stress the need for expert qualitative analysis: intuitive approaches are still needed, the numbers only start the conversation. Indeed, we believe the recallable GEOS framework offers a fertile platform for this analysis (see, for example, the discussion of the RCRI White Paper series in Section 5 below). We now turn to a more specific discussion of the construction of the index, focusing on variable and group taxonomy, country ranking, and the index’s dynamic interactivity.

3. RCRI variable taxonomy and data

As noted by Coccia (2007), building a taxonomy or classification scheme is useful when it maximizes the differences among groups but reduces the complexity of a population of variables into easily recallable macro-classes. As such, many researchers and risk services posit dimensions and schemas, with the goal of focusing the investigation and providing fruitful ground for asking the right questions. For example, as noted above, Sachs (2005) uses a seven

C.L. Brown et al. / International Business Review 24 (2015) 246–265 251

part checklist to carry out country differential diagnoses. As noted in Table 1, Political Risk Services’ ‘‘International Country Risk Guide’’ is based on 22 variables grouped into political, economic, and financial risk categories, with the 12 variables that make up the political risk category incorporating 15 additional variables or ‘‘sub-components.’’ Maplecroft uses a wide variety of variables to construct some 500 indices tied to global, political, legal and regulatory, human rights, and climate risks. Roubini Global Economics’ Social, Institutional, and Regulatory Risk Index ranks 174 countries in aggregate and across four main ‘‘drivers’’ (social/ political, business environment/regulatory quality, property rights/corporate governance, and government effectiveness), with multiple sub-dimensions under the four drivers constructed with 155 variables.

Similar to variable taxonomies, Cavusgil et al. (2004) review the rich literature on country clustering, substantiating the need to cluster different types of countries which offer different oppor- tunities. Grouping countries by ‘‘region’’ or ‘‘level of business opportunity’’ can be helpful. However, they also note that country clustering is generally tied to aggregate macro-indicators or assumptions and can fall short when the analyst wants to move on to specific sector or otherwise more micro-level country analyses which belie the clusters chosen. They discuss how country clustering assumes that countries are homogenous units, and how within country heterogeneity is generally ignored. Another significant problem is the comparability, currency, and reliability of data across countries. Despite such challenges, building taxonomic schemes for variables and countries is an important first step in comparative, quantitative, country risk investigation and analysis. We argue that an important goal must be that the tool developed can then be dynamically widened or focused as the specific strategy or area of interest is made clearer.

In examining ‘‘mash-up indices of development,’’ Ravallion (2012) echoes some of the concerns outlined above, and adds clarity to the ‘‘warning signs’’ needed when constructing composite indices. Specifically, he notes the need to be clear on the conceptual foundations of any ‘‘mash-up’’ index and the tradeoffs they embody. Ravallion also points to the explicit theoretical challenges encountered when deciding on such things as weights, the data to use, and index structure. He calls for transparency and an understanding of the risks or costs, not just benefits, potentially involved for organizational decisional making. Our focus here is on taking a holistic, big data-type approach to country risk, defined with the multidimensionality outlined above, with the end goals of starting an investigation of a specific risk puzzle with the data, sparking strategic and critical thinking, and providing a platform for academic inquiries; however, we emphasize the need for dynamic customization and trusting the user to understand the pitfalls as well as the broader contextual landscape. We seek to put a new, complementary, but within its

own goals and conceptual and functional limits more effective tool in the hands of educators, practitioners, and researchers.

In this article we make the case for a broad based country risk index based on an integrated, holistic, recallable, taxonomic scheme for endogenous variables and country clusters that if necessary can be customized by users given their specific questions, risk profiles, and strategic priorities. Customization can be achieved by adjusting weights, adding new data, subtract- ing unnecessary data, re-conceptualization, or a combination of these steps. The RCRI draws Governance, Economic, Operations, and Societal—GEOS—variables into a broader whole, yet embeds or otherwise lends itself to modeling and statistical techniques which make the index dynamic and useful to a wide variety of audiences. The GEOS taxonomy draws on the social science development literature which calls for holistic, human centered views of country development and integrated, clinical, country diagnoses. A key

goal of the RCRI is to allow the user to run diagnostics on a country (or region, cluster, variable, or set of variables) within a specified strategic vision or research question. By being interactive and broadly based, the RCRI not only seeks to help the user get to a more parsimonious or focused understanding given a specific dependent variable, but also to spark that intuition and probing which leads to broader understanding and effective decision making.

3.1. Why ‘‘GEOS’’?

Building an organic, statistical ecology for a country is in one sense an impossible and dangerous proposition, given the inability to measure certain factors, the limits of data availability and reliability, and the difficulty of understanding something so complex and contextualized as a ‘‘country ecology.’’ Yet, we argue that it is worthwhile to strive to think integratively and holistically, use the significant sources of data now available at the country level, and build a broad-based and dynamic index with distinct, recallable macro-dimensions that cover major areas of country risk as if the country is a living body. Indeed, it is through building the index that we seek to address some of the methodological concerns above.

Most existing country risk indices incorporate political and economic data into their analysis, and some draw on operational and societal variables. Maplecroft’s scope and coverage possibly comes the closest to the ideas offered here, but their work is markedly different in overall conceptualization and variable taxonomy, not as focused on interwoven longitudinal dynamics, and often tied to its own perceptive scoring. By weaving GEOS together, we argue that the RCRI offers distinct taxonomic categories which crudely approximate a country’s ecology more fully than other risk indices. Indeed, in constructing GEOS a central

criteria in all the dimensional and sub-dimensional conceptualizations

is the rudimentary approximation of an integrated view of these levels

within the larger GEOS whole. Fig. 1 graphically depicts the RCRI taxonomy and weighting.

The GEOS taxonomy and weighting was initially conceived in 2009 with input from researchers across a variety of academic disciplines and practitioners from the business, government, and civil society sectors. We were also guided by the social science literature, such as Sachs’ (2005) emphasis on clinical economics and Kohli (2004), Rapley (2007), and Acemoglu and Robinson’s (2012a) classic, comparative historical reminders of the centrality of governance institutions creating the operational economic institutions that lead to economic and social outcomes. Given our goals and purposes surrounding country risk as outlined above, we sought a recallable overall framework that captured a state broadly through its governance head, economic musculoskeletal structure, operational circulatory system, and social heart. As discussed in Section 5 below, this framework has been well received by students and practitioners at a variety of levels, despite its limits and tradeoffs. While the RCRI’s Governance macro-dimension and 5 first-level sub-dimensions borrow from the structural concep- tualizations and weightings found in its more than 20 main sources because they are holistically conceived, most macro-dimensions and sub-dimensions are uniquely conceptualized.

We emphasize that the RCRI’s overall GEOS conceptualization is open to further analysis, critique, and development. For example, in 2012 we strengthened the RCRI’s Health sub-dimension by incorporating longitudinal data on diseases associated more with the developed world (heart disease, diabetes, cancer, etc.) to add to those associated more with the developing world (HIV/AIDS, tuberculosis, malaria, etc.). These variables had previously been left out because of lack of availability, but add to the holism in thinking about Health at a country level. Similarly, while the

Fig. 1. The RCRI taxonomy.

C.L. Brown et al. / International Business Review 24 (2015) 246–265252

index’s Society dimension in general gives insight into a country’s propensity toward a stronger middle class, we added a Middle Class Propensity sub-dimension (including income inequality, consumption, and poverty data) to the Society dimension in 2012 (we had previously included only income inequality data). An example of improvements moving forward might be to refine the default weighting schema through further statistical analysis or perceptive surveying.

The RCRI’s conceptualization is tied to the country risk definition stated above—namely, the probability of particular future events within a state that could have an adverse effect on the functioning of a given organization. Given the multidimen- sionality tied to this definition, the ability to dynamically adapt the tool must also be part of the core conceptualization of the project. The RCRI must be able to have ‘‘offspring,’’ solving for different strategies, risk profiles, or puzzles tied to the specific interests of the user and requiring the user to understand the importance of and even be responsible for weighting, construction, and data quality and availability, despite the Keynesian concerns noted in this paper’s introductory quote. As a tool grounded in the social science literature discussed above, we also argue the index and its GEOS framework offers a robust, experimental measure of development. In the following discussion, we examine each GEOS dimension and its sub-dimensions in turn.

3.1.1. Governance

As noted above and in much of the political economy and development literature, governments—the sovereign actors within a given territory with widely varying institutions, capacities, and influences—play the central policy or ‘‘thinking’’ role for the state. Hence, ‘‘governance’’ is central to any risk analysis, and much of the risk analysis field is focused on ‘‘political risk.’’ The RCRI’s Governance dimension is drawn from

the World Bank’s Worldwide Governance Indicators (Kaufmann, Kraay, & Mastruzzi, 2010) because of its holistic, integrative breadth across six main governance sub-dimensions. Since 1996, the WGI has examined more than 300 variables to construct six broad governance categories: Voice and Accountability, Political Stability and Absence of Violence/Terrorism, Government Effectiveness, Regularity Quality, Rule of Law, and Control of Corruption. It uses more than 30 perceptive data sources, including surveys of households and firms, NGOs, public sector organizations, and commercial business information providers, as well as its own unique methodology, to compile these six macro-variables. While we choose to rely on WGI because of its holistic, integrative breadth covering governance, the complexi- ty and country-level uniqueness of the WGI methodology prevents the RCRI users’ ability to focus the RCRI Governance lens past the six main macro-variables. As such, Governance represents only seven of the RCRI’s 273 variables (an overall Governance score and the six WGI scores), despite being composed of a robust number of indicators. It is the only GEOS macro-dimension that borrows its structure from an outside source. This limitation offers avenues for possible future improvement of the RCRI. However, another advantage of the WGI is that, even though its six macro-variables are composites of many variables, there is only minor overlap with the numerous RCRI variables in the other three macro-dimensions.

It should also be noted that there are a number of other respected data sources for governance indicators which were considered, such as those offered by Freedom House and the Polity

IV project, but these were seen as not directly amenable to the RCRI conceptualization or were already included amongst the 300-plus WGI perceptive variables. Additionally, some risk services, including Maplecroft and PRS, offer their own rich, alternative conceptualizations of political risk. In PRS’s case, this data is also

C.L. Brown et al. / International Business Review 24 (2015) 246–265 253

included in the WGI scores. These alternative conceptualizations again point to possible future avenues of research.

3.1.2. Economics

As well put by Rapley (2007) and discussed above, the state cannot be separated from its comparative historical socio- economic body, which Acemoglu and Robinson (2012a) effectively and classically frame as including economic institutions and economic and social outcomes. At times loosely tying into this type of thinking, most risk researchers and services incorporate macro- economic variables into their analysis, the economic outcomes we argue represent the musculoskeletal structure within which organizations operate.

Drawing on this literature, but also seeking a broad, holistic Economics conceptualization as part of the GEOS whole, the RCRI includes three main sub-dimensions under Economics: Macro- Economic Indicators, Market Access, and Currency Over/Underval- uation. The first two of these sub-dimensions have extensive ‘‘drill down,’’ with Macro-Economic Indicators, a country’s economic fundamentals, being broken into ‘‘Broad Economy’’ and ‘‘Finance’’ sub-dimensions, which lend themselves to further focusing across 11 additional variables. Market Access, favoring the depth of integration in the global economy, is sub-divided into ‘‘Trade Profile’’ and ‘‘Investment Profile,’’ which have 4 and 6 sub- dimensions, respectively. Market Access focuses the lens into such areas as Barriers to Trade and Foreign Direct Investment Flows (in total, the sub-dimension includes 31 variables). By centering the Economics macro-dimension on macro-economic fundamentals and the depth of integration in the global economy we place ourselves in a Western, market-oriented development mindset. Our Economics exploration, focused on representing the broad economic realities in which organizations operate, led us to this developmental bias. Of course, these economic outcomes, as well as a country’s social outcomes, are closely linked to the economic institutions and infrastructure that result from political objectives.

The Currency Over/Undervaluation variable is seen as a rough barometer for a country’s overall macro-economic health and is weighted at 2.5% percent of the Economics macro-dimension. Data here is derived from an ongoing annual assessment by the Peterson Institute (Cline and Williamson, 2008–2011), which we believe gives the best estimates of currency over/undervaluation. Howev- er, the Institute’s work only includes 48 of the RCRI’s 126 countries, again pointing to a possible future area of improvement of the RCRI. The methodology for addressing and adjusting thresholds for missing country data is covered below. Currency Over/Undervalu- ation is one of two variables where we do not treat the raw data in a linear fashion, using instead the absolute value of the data, because neither a high positive score nor a high negative score is desirable (the other variable is Inflation, as discussed in the ‘‘Data Challenges’’ section below). Overall, the Economics macro- dimension includes 46 variables within its own integrative structure. In addition to the Peterson Institute, data is retrieved from the U.S. Central Intelligence Agency’s World Factbook, the International Monetary Fund’s World Economic Outlook, the World Trade Organization’s World Trade Organization Statistics, the United Nations Conference on Trade and Development’s World Investment

Report, the World Bank’s Joint External Debt Hub and International

Trade Centre, and the World Economic Forum’s Global Competitive-

ness Report (GCR) and Enabling Trade Report (ETR).

3.1.3. Operations

The RCRI Operations’ dimension is broadly analogous to a state’s circulatory system, or the framework of economic institu- tions and infrastructure an organization must engage to be successful in a country. Indeed, one of our main Operations sources, the World Bank’s Doing Business, describes itself as a

‘‘kind of cholesterol test for the regulatory environment’’ (Doing

Business, 2012: 18). The RCRI’s Operations dimension has 4 main sub-dimensions, Business Transactions, Logistics, Operational Landscape, and Short Term Currency Fluctuation, each of which we argue provides a different key element (domestic regulatory environment, international links, broader circulatory system, and fundamental lifeblood, respectively) within a country’s overall Operation whole. The first three of these dimensions have additional sub-dimensions, allowing the user to ‘‘drill down’’ as many as five levels into the data. Importantly, each of these three main Operations sub-dimensions comes from a source whose structure lends itself to direct incorporation into the RCRI because it fits our criteria of seeking to look at that sub-dimension holistically. Also, we include ‘‘all the data’’, as well as weightings, from these sources because user-specific dependent variables remain unclear and we seek a flexible and adaptable index (and accept the messiness that engenders). The construction of a unique Operational macro-dimension is possible because of data newly available over the last several years and is a distinguishing characteristic of the Robinson Country Risk Index. Overall, Operations includes 110 of the RCRI’s variables.

3.1.3.1. Business Transactions. The Business Transactions data is drawn from The World Bank’s Doing Business project, which since 2003 has examined and measured the regulations applying to domestic small and medium-size companies, with the goal of providing an objective basis for understanding and improving the regulatory environment for business. The RCRI incorporates each of Doing Business 9 quantitative measures: Starting a Business, Construction Permits, Registering Property, Getting Credit, Pro- tecting Investors, Paying Taxes, Trading Across Borders, Enforcing Contracts, and Closing a Business. Each of these measures includes further sub-dimensions. The driving premise of Doing Business is that economic activity requires good rules, and the overall objective is that regulations are efficient, accessible, and simple to implement. To this end, Doing Business includes perceptive and hard data. The perceptive data comes from assessing laws and regulations; the hard data are ‘‘time and motion indicators that measure the efficiency in achieving a regulatory goal (such as granting the legal identity of a business)’’ (Doing Business, 2012: 17).

3.1.3.2. Logistics. If Doing Business is a cholesterol test for the domestic regulatory environment, then the World’ Bank’s Logistics

Performance Index (LPI), published bi-annually starting in 2007, is a cholesterol test for a country’s international logistics infrastructure (i.e., its international ties). The RCRI includes each of the LPI’s six indicators: Customs Logistics, Logistics Infrastructure, Internation- al Shipments Logistics, Logistics Competence, Tracking and Tracing Logistics, and Logistics Timeliness. The LPI is based on perceptive surveys of operators on the ground (global freight forwarders and express carriers). These operators provide feedback on the logistics ‘‘friendliness’’ of the countries in which they operate, and on the logistics ‘‘friendliness’’ in those countries with which they trade. Every country is ranked on a scale of one to five on each indicator, with one being the worst performance and five being the best performance.

3.1.3.3. Operational Landscape. The RCRI’s Operational Landscape sub-dimension is a composite of data drawn from the World Economic Forum’s GCR, published since 1979, and ETR, published since 2008. We include five main sub-dimensions under Opera- tional Landscape—Innovation and Sophistication, Infrastructure, Technological Readiness, Business Environment, and Market Efficiency—each of which is largely distinct from other variables in the Operations macro-dimension, but helps give a more nuanced

C.L. Brown et al. / International Business Review 24 (2015) 246–265254

view of the capillaries in a country’s cardiovascular system. For example, drilling into Innovation and Sophistication one finds such variables as ‘‘Local Supplier Quality’’ and ‘‘Availability of Scientists and Engineers’’ amongst the 19 variables; Technological Readiness includes ‘‘Availability of Latest Technologies’’ and ‘‘Internet Bandwidth’’ amongst its 9 variables; Market Efficiency includes ‘‘Effectiveness of Anti-Monopoly Policy’’ and ‘‘Reliance on Profes- sional Management’’ in its 12 variables. It should be noted that the World Economic Forum’s GCR and ETR report some of the same variables, and sometimes move variables within their various ‘‘pillars.’’ However, because of the RCRI’s conceptualization, longitudinal focus, and overall purposes, we broadly draw data from these sources but keep related but stable sub-dimensional structures. We revisit this issue as necessary.

3.1.3.4. Short-Term Currency Fluctuation. Currency is the best approximation of the blood running through the circulatory system of a country, and the daily fluctuation in the value of the currency measures the pulse. As such, we included a single variable sub-dimension in the Operations macro-dimension, Short-term Currency Fluctuation, defined as the standard devia- tion of the daily fluctuation of exchanges rates vis-à-vis the U.S. Dollar, with the United States and Ecuador expressed vis-à-vis a basket of currencies. This variable has a 2.5 percent weight within the Operations macro-dimension, with the data coming from Oanda.com and the International Monetary Fund. Of course, countries with controlled or manipulated exchange rates can score well on this indicator, just as a pulse may not help in uncovering a disease, or may even mask the disease itself. This issue is somewhat counter set against the Currency Over/ Undervaluation variable discussed above. As noted above and below, the ability to recalibrate the index is an important part of its interactivity.

3.1.4. Society

Like a heart, social outcomes pump life into a country and help spark the popular feedback loop which leads to more inclusive political and economic institutions. The RCRI includes 109 variables in its Society macro-dimension, which is sub-divided into 6 main sub-dimensions: Health, Education, Demographic Dynamic, Gender Gap, Middle Class Propensity, and Environmental Sustainability. These sub-dimensions are loosely based on broader conceptual outlines found in the Southern Center for International Studies’ World in Transition series (see for example Brown et al., 2006; Brown & Higgins Morton, 2007) and we argue that, taken together, they help capture a broader societal whole. Each, outside Middle Class Propensity, which has four variables (see Fig. 1) and is based on the academic work on middle classes by Ravallion (2009) and Cavusgil and Kardes (2013), has multiple additional sub- dimensions, drilling as many six levels into the data. All but Gender Gap and Environmental Sustainability have uniquely derived conceptualizations, with Gender Gap borrowing its structure and weighting from the World Economic Forum’s Global Gender Gap

Report (GGR), and Environmental Sustainability borrowing its structure and weighting from the Environmental Performance Index

(EPI), jointly produced by the Yale Center for Environmental Law and Policy and the Center for International Earth Science Information Network at Columbia University. We emphasize that wherever the RCRI borrows structure and weightings from another source, we believe the given source has the best available data, often available nowhere else, and a conceptualization which fits the criterion of crudely approximating that sub-dimension’s dynamics within the overall GEOS ecology.

The unique Health (18 variables across 3 sub-dimensions), Education (16 variables across 3 sub-dimensions), and Demo- graphic Dynamic (8 variables) sub-dimensions follow the same

logic and are drawn from a variety of data sources. Overall, the Society macro dimension draws from 13 sources thought to be the best available for that data. In addition to EPI and GGR, these sources include the CIA’s World Factbook, United Nations Children’s Fund’s State of the World’s Children, the World Bank’s World

Development Indicators, the United Nations Office on Drugs and Crime’s World Drug Report, the World Economic Forum’s GCR, the United Nations Education, Scientific, and Cultural Organization’s Institute for Statistics, the United Nations AIDS Statistical Database, the Inter-Parliamentary Union’s Report on Women in

National Parliaments, the United Nations Development Project’s Human Development Report, the International Labor Organization’s Labor Statistics Database, and the World Health Organization’s World Health Statistics.

3.2. Data challenges

Challenges exist with some of the RCRI data, particularly in three of the Society sub-dimensions, because the risks presented by these variables are not necessarily linear, the data is previously transformed by the data source, the data is truncated or targeted by the data source, or because the RCRI borrows structure from a source, but uniquely and generally transforms the data in its own way (see below and Appendix A concerning RCRI data transfor- mation). For example, in a much explored and debated hypothesis (including currently, given the dynamics of ‘‘globalization’’), Kuznets (1955) suggests that GINI scores—part of the Middle Class Propensity sub-dimension—represent an inverted-‘‘U’’ shape with respect to economic growth over time (income inequality first increases and then declines as liberal economic development continues). However, because of the breadth of the countries covered in the RCRI, the RCRI’s broad-based data transformation, and the strains on the working and middle classes often presented by high inequality, this variable is treated as a linear risk within the index, with higher GINI scores translating to higher risk. Similarly, high population or population growth can be a developmental risk; however, the RCRI uses a ‘‘people are good’’ principle when measuring demographic variables because for many organizations (particularly businesses), as well as for the countries themselves, the ‘‘people are good’’ principle generally holds, in as much businesses desire robust and growing markets and the size of a state’s population is tied to state power, among other factors. Also, it is difficult to treat the demographic data any differently, yet necessary to include demographics within the holistically con- ceived index. A third example ties to currencies, as we treat the Short-Term Currency Valuation linearly and use a zero target for Currency Over/Undervaluation. Arguments can easily be made that a country can at times want some short term currency fluctuation or an over or under valued currency. Inflation offers yet another example, in as much as having negative inflation, or deflation, can be problematic for a country; therefore, this variable is the second variable in the RCRI where we take the absolute value, slightly penalizing negative inflation. Discussion of source specific data challenges, particularly concerning the treatment of data by the Gender Gap Index and Environmental Performance Index, can be found in Appendix B.

3.3. Data collection

The ‘‘Golden Rule’’ for pulling RCRI data is that the data used for each year’s index is the best available data released or available during that year. In other words, the data for the 20XX RCRI is the best data released or available during 20XX, regardless of the name of the source (it might be titled with a previous or following year) or if the data is known to have been gathered in the previous year or years. However, because of the numerous sources used to

C.L. Brown et al. / International Business Review 24 (2015) 246–265 255

construct the RCRI and their individual dynamics, some caveats are still necessary. For example, some sources revise their data for previous years in their current releases. In this case, the previous years of the RCRI are revised to reflect these source revisions. Also, while pulling data forward from previous years, if it is the best data available, is used for a number of Society variables and the Logistics variables in Operations, owing to the more slow-moving nature of these variables and the fact that they are not updated annually (often both), this principle is not used for other data points, such as in carrying forward GDP Growth from a previous year. Another issue is when RCRI sources change their data collection techniques or types of data reported. Each of these technical details is tracked and regularly reviewed.

The RCRI includes 126 countries because of the limits of data availability across a wide variety of sources. We could not take the index prior to 2005 for similar reasons. Specifically, we use a 15 percent default but adjustable threshold for missing data across the index in the base calculation of the RCRI. This default position results in 119 countries being calculated into the 2012 Robinson Country Risk Index, with seven countries dropping out because of too much missing data (Hong Kong, Cote D’Ivoire, Burundi, Syria, Angola, Iran, and Lebanon). The 15 percent threshold is a somewhat natural break in the percentage of missing data in our data set, and countries not in our 126 generally have well over 30% missing data. Seventy three percent (87/119) of countries had less than 5% of missing data; 92 percent (109/119) had less than 10 percent missing data; average missing data was 3.8 percent and country-specific missing data is reported within the interactive index. Finally, it should be noted that in addition to the more than 20 main sources used to construct the RCRI, additional sources are used to fill data holes; for example, much of the Taiwan and Hong Kong data has to be retrieved directly from the Taiwan and Hong Kong government websites.

3.4. Country clusters

As noted, in addition to a variable taxonomy, the RCRI includes country clustering. This clustering is done by world region and perceived level of development. To determine regions we used standard classifications done by a variety of organizations, such as the World Bank. Sometimes countries are cross-listed in more than one region, such as Turkey (Middle East/Europe) and Mexico (North America/Latin America). For development clusters, stan- dard services such as MCSI Barra and FTSE offer delineations of advanced, developed, emerging, and frontier markets, and we adapted these classifications with minor variation based on the results of the RCRI, recognizing they are dynamic and change over time. We argue that in the future, the RCRI will be more useful in constructing developmental clusters, but for present purposes, we consulted how other organizations built their country taxonomies and adapted them as we built the RCRI country clusters.

4. RCRI data transformation and country ranking

Data transformation and aggregation, as well as country rank- ordering, are well worn methodological endeavors, but the specific steps used by different sources vary given different conceptualiza- tions and goals. In developing our methodology, the RCRI team drew on the previous work of Cavusgil (1997) and Cavusgil et al. (2004), and reviewed the methods used by other indexes, such as the Environmental Performance Index, Doing Business, the Legatum Prosperity Index, and the Gender Gap Index; however, given its broad focus, diversity of data, and multileveled, dynamic conceptualization, the RCRI’s techniques are unique. In brief, we logarithmically transform and then standardize the raw data matrix. The GEOS weights are then used for each level of upward

aggregation, with the resultant variables being standardized again after each step so that they can be aggregated upward again, using the determined weights, until all the aggregations are complete. As opposed to many risk tools, we do not impute scores for missing values, to strengthen accuracy and because of the RCRI’s focus on flexibility.

After computing all the standard scores, we rescale the data 1 to a Maximum Score (1000 by default, although this maximum score can be changed within the interactive index). After deriving the rescaled scores, we rank the countries across each variable, sub- dimension, GEOS macro-dimension, and Overall from 1 to 126, or less depending on the tolerance for missing values as set within the index (as noted above, our default threshold of 85% across all years renders 119 countries in the 2012 index). Note that this methodology means that the lowest and best rank (#1) receives the highest score (default of 1000) within the RCRI. Appendix A covers additional methodological issues, including extreme out- liers, variable directionality, and method variations with respect to region and cluster ranks and scores.

5. Overall results, use to date, and dashboarding

Appendix C reports the overall 2012 RCRI rankings and the rankings across each macro-dimension, as seen on the current RCRI dashboard. Given the uniqueness of the goals and structure of the RCRI, these rankings are unique, though it is not as surprising to see Singapore and the northern European countries clustered at the top and the African countries clustered at the bottom, as development results are often broadly diagnosed in this way for these countries. Longitudinally, many countries maintained relatively steady ranks and scores across the 2005–2012 time period. However, 14 of the 119 calculated countries show a � movement of 15 or more places, and 9 show a movement of 20 or more places. As for scores, no country gained more than 100 points, but 17 fell more than one hundred points during a period which included the Great Recession. Overall, countries showing the steepest declines included Egypt, Pakistan, Venezuela, Jordan, and Greece; top gainers included Zambia, Georgia, Albania, Kazakhstan, and Indonesia. Each of the significant movers rose or fell for different reasons, as initially told by the numbers, and, as noted, the index allows the user to drill down to the sub-dimensions and specific variables driving the change.

For example, India’s Overall RCRI Rank fell 10 places (103 points), going from 61st to 71st between 2007 and 2011, before showing a small improvement to 69th in 2012. Societal factors helped drive the downward movement. The country saw signifi- cant declines in Education and Middle Class Propensity, along with ongoing weak ranks and scores in Health, Gender Gap, and Environmental Sustainability; only in Demographic Dynamic does India score well in the Society macro-dimension, ranking at the top every year because it gets benefit within the index due to its large population. Interestingly, India saw a rise in its Economics macro- dimension in both ranks and scores between 2005 and 2008, driven by very strong gains in its Investment Profile, but showed a sharp decline in Economics after 2008, driven by a weakening on Macro-Economic Indicators and FDI Inflows. On the other hand, Venezuela’s volatile plummet of 19 places from 2005 to 2012 (from 79th in 2005 to 108th in 2010 to 98th in 2012) correlated with a volatile Economics dimension, where the country saw a steep decline between 2008 and 2010 (333 points relative to other countries on a 1–1000 scale), before showing a modest rebound. Venezuela also saw declines in its already dismal Operations and Governance rankings. However, the country improved from 49th to 44th in Society.

Two points noted above need emphasis here. First, one of the advantages of the RCRI, but beyond the scope of this paper, is that

C.L. Brown et al. / International Business Review 24 (2015) 246–265256

the user can further sharpen the focus down to the specific sub- dimensions and individual variables driving the longitudinal ranking changes for these countries. Second, the broader political economic analyses underlying cases such as India and Venezuela, analyses that tie in comparative historical distinctions and dynamics such as state-level socio-economic strategy, remain central to the index’s interpretation. To this end, in 2012 a RCRI White Paper Series was initiated, using numerical interpretation of countries and topics through the RCRI lens to start the conversa- tion, but accompanied by scholarly analysis.

Indeed, the tool’s uses to date are a central part of it current results. In addition to the White Paper Series, the RCRI has now seen broad use as a dynamic classroom tool at Georgia State University, significant investigation as a strategic thinking tool in collaborative efforts with business practitioners (generally under non-disclosure agreements), and modest use as a conduit for peer- reviewed academic research (Shea, 2014). We also expect to develop the tool’s exposure through a blog. Classroom use has been the most robust and has ranged from interactive use by professors within their classes, to student use as they carry out organization- specific scenarios, roll plays, or other course assignments. To this end, a growing body of deliverable pedagogy has been developed around the use of the tool for teaching. Courses where the index has been employed have ranged from Executive MBA and Executive Education courses (which include practitioners from the business, government, and non-governmental organizations) to undergraduate political science classes.

One of main outputs of the RCRI is to derive the explanatory charts, graphs, or tables for a clinical diagnosis, given a specific lens, and Appendix D gives a series of examples of these types of outputs, though further contextual analysis of these examples is beyond the scope of this paper. Their construction is relatively easy using the current RCRI dashboard; however, alternative dash- boarding is central to expanding the use of the instrument. The tool was originally built using Microsoft Excel and Visual Basic programming, which allowed us to address key conceptual, developmental, and functional issues, but provided some func- tional, copyright, and intellectual property protection concerns. In 2013, we started transitioning to a web-based platform to better leverage the tool inside and outside academe. The last two examples in Appendix D are of our ‘‘Build Your Own Index’’ approach (i.e., of the functional ability to re-conceptualize the index to solve for different strategic visions or puzzles), which we had to construct with a separate dashboard. This functional step will be incorporated into the main dashboard as we move forward.

Finally, while Table 1 in Section 2 above gives a comparative analysis with respect to the conceptualizations and uses of other relevant tools, we do not find empirical comparison to the rankings of other indices as useful, though a possible avenue for future research. Of the indices mentioned in Table 1, only the Legatum

Institute’s Legatum Prosperity Index offers a comparatively robust and available aggregate ranking. Yet, as noted, the LPI is not focused on dynamic country risk measurement and analysis, but rather on understanding country income and well-being (leading to markedly different dimensions and adaptability needs). The conceptualizations of other indices mentioned in Table 1 or found elsewhere are also different, given their unique, generally narrower, goals. As noted in Section 3, some are included within the RCRI. In addition to conceptualization and goals, we have argued a better comparison to other country risk tools is with respect to use and ability to generate value.

6. Implications and future research

Country-level risks and their cumulative impacts can pose vexing challenges or puzzles for businesses, governmental bodies,

NGOs, and other types of organizations, as well as for researchers and students, as these risks are often complex, integrated, and fast changing. Strategic thinkers need dynamic, interactive, well- conceptualized, and holistic tools to help identify, assess, probe, and respond to country-level challenges, given their specific queries and goals. These tools need to be combined with comparative historical, experiential, and intuitive insights. For organizations, broad-based risk management and intelligence strategies are central not just for organizational survival, but also for organizational value creation.

Drawing on greatly improved data availability and computing power, we argue the RCRI takes a step in the right direction and assists organizational leaders seeking to design and adopt mitigation strategies and build resilient, adaptive, and thriving organizations. We also posit the GEOS framework offers a robust alternative measure of country development, albeit in the Western, market-oriented, liberal tradition, and provides a dynamic and fertile research and classroom-teaching tool. Still, one must be clear on the limits and tradeoffs embedded in the RCRI’s conceptual foundations and functionalities, such as the potential costs and benefits associated with offering user choices surrounding such areas as weights, structures, and data. Other risk analysis tools, though in some ways similar, have their own specific goals, structures, and functionalities, and, as such, their own results and tradeoffs. Comparative analysis with different tools must be centered in the different conceptualizations which lead to unique measurements and tool dynamics (see Table 1). In the end, the various tools are broadly complementary if the goal is to spark strategic and critical thinking, and tool use also plays an essential but imperfect role in understanding comparative results and added value.

Initial avenues for further academic research include a richer examination of the modeling and statistical techniques which might prove useful in sharpening the RCRI’s analyses by the numbers, solving for alternative conceptualizations, and tackling challenging questions. For example, a researcher might be interested in constructing an index around the strategic orienta- tion of a beverage company, or in testing hypotheses surrounding the relationship between improved Operations and increased trade and investment, or in teasing out the key numerical correlates or drivers of social well-being, foreign direct invest- ment, or political stability across 126 embedded country ecologies.

Another interesting avenue of research, in line with much of the development and political economy literature discussed above, is in using the index as a dependent variable and asking the question ‘‘why?’’ with respect to country or cluster longitudinal profiles. What is the connection between oil exportation, economic nationalism, or IMF-style liberalization and GEOS outcomes, broadly conceived? The index offers an ecological view into countries and clusters which possibly exposes, in the aggregate as well as in detail, the results of developmental efforts, adding clarity and specificity to existing development strategies and literature. Here, the research efforts turn to the broader scholarly analyses underlying critical cases, such as the India and Venezuela cases discussed above, using a tool that allows for a data-centered, ecological starting point. A third direction of academic research is in contributing to traditional conceptualizations of ‘‘levels of development.’’ What is the line between ‘‘developed’’ and ‘‘emerging’’ economies and why is it useful? A fourth may be to test the RCRI’s capacity to generate probabilities and projections, despite the deep skepti- cism of econometric forecasting discussed above. It must be emphasized that the RCRI remains limited by such factors as data availability, data reliability, and human imagination. In the end, it is a useful place to start an investigation and spark intuition

C.L. Brown et al. / International Business Review 24 (2015) 246–265 257

and probing, and needs to be combined with all techniques at the strategic thinker, decision maker, researcher, or educator’s disposal.

Acknowledgements

The Robinson Country Risk Index is the product of a partnership between the World Affairs Council of Atlanta (WACATL) and the Center for International Business Education and Research (CIBER), both affiliated with Georgia State University’s Robinson College of Business. Dr. Christopher L. Brown serves as director of the project and is the leader and catalyst for the conceptualization and development of the index. Dr. S. Tamer Cavusgil provided important foundational work and leadership, including developing the Opportunity Index at Michigan State University and creating the Emerging Markets Risk Index within the GSU Institute of International Business. Dr. A. Wayne Lord has been pivotal in bringing the project together and offering key insights across its entire breadth. Additional critical contributions have been made by Dr. Cedric Suzman, Executive Vice President and Director of Programs at WACATL; Assistant Dean Dave Forquer and Mr. Jacobus Boers of the Robinson College of Business; Ms. Paula Reyes del Toro, Mr. Ricardo Orjuela, Mr. T.J. Ertley, Ms. Di Wang, Mr. Michael Shea, and Mr. Camilo Martinez, Graduate Research Assistants assigned to the project by Dean Fenwick Huss of the Robinson College; and Ms. Alyssa Smith, Ms. Joanne Essenwein, Mr. Claiborne Darden, and Mr. Taohang Zhang, Senior Research Associates with WACATL at various points during the project’s development. Numerous other wonderful research assistants and colleagues have contributed to the project.

Appendix A. RCRI Data Transformation and Aggregation.

As depicted in the figure above, we take the natural log of the RCRI

raw data matrix, changing the scale from the decimal scale to

logarithmic scale, which shifts the data toward a natural distribu-

tion or otherwise corrects for highly skewed distributions.

However, given that the logarithmic function works only for

values greater than 0 (you cannot take the log of a negative number

or zero), and the fact that the log of numbers between 0 and 1 is

negative, we took two additional steps. First, we shifted the

logarithmic curve by 1 so as to retain the zero values and make

values between 0 and 1 positive after transformation. Second, for

negative raw values, we took the log of the absolute value of the

variable plus 1, and then made that value negative, or a mirror of

the positive function. Mathematically:

fðxÞ ¼ �Lnðjxj þ 1Þ if x < 0

fðxÞ ¼ Lnðx þ 1Þ if x � 0

After logarithmically transforming the data, each sub-dimension,

for each year, needs to be standardized by using the formula (x � m)/

s, where m represents the annual sub-dimensional mean and s

represents the annual sub-dimensional standard deviation. This step

makes all variables comparable within a given annual sub-dimension,

with a mean 0 and standard deviation 1. Using the RCRI weights, we

then aggregated upwards. After each level of aggregation the

resulting variables are standardized again, so that they can be

aggregated upward again, using the determined weights, until all the

aggregations are complete for a given year.

A.1. Directionality, Missing Values, and Extreme Outliers

To address the directionality of the data (i.e., whether a high score

is good or bad), the standard scores are multiplied by 1 or �1. In general,

high standard scores are coded as bad and low scores are good, but if the

high score is good for a particular variable, then if it is multiplied by �1,

the country’s high score becomes a low score and the variable is

‘‘reversed.’’ As noted, for Currency Over/Undervaluation and Inflation

we use the absolute value. For Currency Over/Undervaluation, neither a

high positive score nor a high negative score is desirable. For Inflation,

this step is to slightly penalize slipping into deflation.

For missing values, because our goal is accuracy as well as to have

a flexible index, we decided not to use the other standardized values

within a level of aggregation to get a weighted average to impute for

missing values. We instead leave missing values as ‘‘N.A.’’, and then

use the weighted average of the remaining values within the level of

aggregation to aggregate upwards. When all values are missing for a

given level of aggregation, then the higher level of aggregation is left

as an ‘‘N.A.’’. This methodology influences the ranks/scores in

comparison to using imputed data for missing values as done by

some researchers, but we believe it is a more accurate reflection of

reality and allows us to construct a more dynamic index, where users

can ‘‘Build Your Own Index,’’ with their own unique conceptualization

and weighting schema.

While we use a logarithmic transformation to push the data

toward more normal distributions, some variables still have

significant outliers. So far, only in one case have we decided to

further treat the data. Zimbabwe had a reported inflation rate of 14.9

billion percent in 2009, and extremely high scores in other years

2005–2008. These reported rates so skew the inflation distributions

that we considered winsorizing at 99 percent. In other words, if 14.9

billion percent is the worst rate, then countries such as Venezuela

with 30 or 40 percent inflation get extremely good scores because the

worst score is so extreme. If you winsorize you bring your extreme

outliers back to the edges of the distribution. However, since we have

126 countries, a 99 percentile winsorization is close to the next worst

country’s raw value, so instead of winsorizing at 99 percent we

manually replaced Zimbabwe’s inflation with the next worst value for

every year 2005–2009 (but we kept the actual raw data for reportage

in the index). This manual winsorization means that we generally

have more representative RCRI scores, and Zimbabwe still gets the

C.L. Brown et al. / International Business Review 24 (2015) 246–265258

worst score. An example of the drawback is that Ethiopia, second

worst in 2009, is seen as equally bad as Zimbabwe for that particular

year. Starting in 2010, Zimbabwe’s inflation was no longer extreme

and the index was computed without any winsorization.

A.2. Ranks and Scores

After computing all the standard scores, we rescale the data 1 to a

maximum score (1000 by default, although this can be changed

within the interactive index) using the formula:

ðScoreCountry i � Sample Minimumall countriesÞ ðSample Maximumall countries � Sample Minimumall countriesÞ

� 999

þ 1

After deriving the scores, we rank them from 1 to X, depending on

the tolerance of missing values as set within the index (the 85 percent

default threshold in 2012 renders 119 countries in the index). Within

the RCRI, high scores are desirable, with the lowest rank (#1)

receiving the highest score (default 1000).

To determine a cluster or regional score for any variable, either a

simple average can be used within the index, or, by clicking a button,

country scores within a region/cluster can be weighted by population.

This is important because you may want to know what the simple

average is for Emerging Markets with respect to ‘‘Business Transac-

tions’’; alternatively, you may want to know what the European

average is of ‘‘GDP per capita,’’ but rich Germany, with a much larger

population, should count proportionally more than poor Moldova, with

a much smaller population. Whichever method is used, regions are

ranked against regions (9 total) and clusters are ranked against clusters

(5 total), again with low ranks and high scores designating lower risk.

Appendix B. RCRI Source Specific Data Challenges

The two sub-dimensions in the RCRI’s Society macro-dimension

where the RCRI incorporates source structure and weights, the

Gender Gap Index and the Environmental Performance Index,

uniquely treat their data because of their focus on gender equality

and environmental policy targets, respectively. The Gender Gap Index

truncates its measures at equality, not giving benefit or penalty to

countries for variables where female values exceed male values. We

have adopted this truncation in the RCRI, but then transform the data

in a method unique to the RCRI, leading to somewhat different, but

generally very small, differences in the aggregated country rankings

for the Gender Gap sub-dimensions. Still, there are two countries

which shift more than 20 ranks in 2012 given the RCRI methodology.

Importantly, throughout the index there are no ranking differences at

the raw data/base variable level when the RCRI borrows structure and

weights from a source, only in aggregation to higher dimensions

because of the RCRI aggregation methodology (see Appendix A).

The Environmental Performance Index data is uniquely and

selectively transformed (i.e., selective use of varying logarithmic

equations, winsorization, and imputation by variable) by its authors

before being targeted toward meeting a percentage of specific policy

goals. However, because we do not generally want to transform

already transformed data, as well as the fact that the RCRI’s focus is on

overall risk, we choose to use the raw data available from the EPI and

transform the data using the RCRI methodology, as discussed in

Appendix A. However, we do adopt the EPI selective imputation

method for missing variables in their data set, as this is a choice of

internally addressing missing raw values (see Appendix A for a

discussion of the RCRI general methodology for missing data).

Because of the decision to use the EPI raw data and the RCRI’s own

unique method of transformation, the rankings in the RCRI also show

mostly minor but some significant differences from those in the EPI.

Twelve cases rose or fell by more than 30 spots in 2012. We tested for

ranking changes using the EPI’s selectively transformed and targeted

data, but they were much more significant than going with the raw

data given the confluence of different methodologies. Outside the Society dimension, the data challenges are

generally less significant for sources where we borrow structure

and weightings. Indeed, because the rankings only change in

aggregation and not at the raw data level, similar dimensional

ranking changes do not occur with Worldwide Governance

Indicators, which does not aggregate its variables. Logistics

Performance Index, which only aggregates up one level, shows a

few ranking changes of one rank, and World Economic Forum data

only shows extremely minor variation because of a difference in

how ranks are assigned to countries with equivalent scores. Doing

Business data, which does aggregate upwards several dimensions

through a simple averaging approach, does show some minor

ranking changes given the use of the RCRI’s aggregation method; 30

countries change more than 10 places; 5 more than 20 places. Each

of these challenges points to ongoing areas of investigation as we

continuously build and evolve the dynamic, interactive, and

holistic country risk index.

Appendix C. The 2012 RCRI Overall and GEOS Rankings

C.L. Brown et al. / International Business Review 24 (2015) 246–265 259

C.L. Brown et al. / International Business Review 24 (2015) 246–265260

Appendix D. Examples of RCRI Analysis

C.L. Brown et al. / International Business Review 24 (2015) 246–265 261

C.L. Brown et al. / International Business Review 24 (2015) 246–265262

C.L. Brown et al. / International Business Review 24 (2015) 246–265 263

C.L. Brown et al. / International Business Review 24 (2015) 246–265264

C.L. Brown et al. / International Business Review 24 (2015) 246–265 265

References

Acemoglu, D., & Robinson, J. (2012a). Why nations fail: The origins of power, prosperity, and poverty. New York: Crown Business.

Acemoglu, D., & Robinson, J. (2012b). Response to Jeffrey Sachs. In Why nations fail: Blog http://whynationsfail.com/blog/2012/11/21/response-to-jeffrey-sachs.html?utm_- source=dlvr.it&utm_medium=twitter Accessed 11/21/2012.

Atsmon, Y., Child, P., Dobbs, R., & Narasimhan, L. (2012). Winning the $30 trillion decathlon: Going for gold in emerging markets. McKinsey Quarterly http://www.mckin- sey.com/insights/strategy/winning_the_30_trillion_decathlon_going_for_gold_in_emer- ging_markets Accessed 11.11.12.

Bouchet, M. H., Clark, E., & Groslambert, B. (2003). Country risk assessment: A guide to global investment strategy. Hoboken, HJ: John Wiley.

Brown, C. L., & Higgins Morton, C. (2007). Excerpts from the world in transition series: Updated for 2007. Regional studies reader (Third ed.). Atlanta: Southern Center for International Studies.

Brown, C. L., Blankenship, G., Foerster, S., Moffitt, L., Rossi, J., & Ryckeley, R. (2006). South Asia in transition. Atlanta: Southern Center for International Studies.

Cavusgil, S. T. (1997). Measuring the potential of emerging markets: An indexing approach. Business Horizons, 40(1), 87–91.

Cavusgil, S. T., & Kardes, I. (2013). Defining and measuring the middle class in emerging markets: The GSU-CIBER middle class scorecard. ESOMAR Research World, (March– April), 46–49.

Cavusgil, S. T., Kiyak, T., & Yeniyurt, S. (2004). Complementary approaches to prelimi- nary foreign market opportunity assessment: Country clustering and country ranking. Industrial Marketing Management, 33(7), 607–617.

Clark, E. (1997). Valuing political risk. Journal of International Money and Finance, 16(3), 477–490.

Cline, W. R., & Williamson, J. (2011). Estimates of fundamental equilibrium exchange rates. In Policy Brief. Washington DC: Peterson Institute of International Economics.

Coccia, M. (2007). A new taxonomy of country performance and risk based on economic and technological indicators. Journal of Applied Economics, 10(1), 29–42.

Cruces, J. A., Buscaglia, M., & Alonso, J. (2002). The term structure of country risk and valuation in emerging markets. http://papers.ssrn.com/sol3/papers.cfm?abstrac- t_id=327301 Accessed 11.02.11.

Cukier, K. N., & Mayer-Schoenberger (2013). The rise of big data: How it’s changing the way we think about the world. Foreign Affairs.

Cusimano, M. K. (2007). Beyond sovereignty: Issues for a global agenda. Belmont, CA: Thomson/Wadsworth.

Doing Business – World Bank Group. http://www.doingbusiness.org/ Accessed 11.02.11.

Fitzpatrick, M. (1983). The definition and assessment of political risk in international business: A review of the literature. Academy of Management Review, 8(April (2)), 249–254.

Friedman, T. I. (2008). Hot, flat, and crowded: Why we need a green revolution – And how it can renew America. New York: Farrar, Straus and Giroix.

Funston, F., & Wagner, S. (2010). Surviving and thriving in uncertainty: Creating the risk intelligent enterprise. Hoboken, NJ: WIley.

Gourevitch, P. A. (1986). Politics in hard times: Comparative responses to international economic crises. Ithaca, NY: Cornell University Press.

Haggard, S. (1990). Pathways from the periphery: The politics of growth in the newly industrializing countries. Ithaca, NY: Cornell University Press.

Hall, P. A. (1989). The political power of economic ideas: Keynesianism across nations. Princeton, NJ: Princeton University Press.

View publication statsView publication stats

Harland, C., Brenchley, R., & Walker, H. (2003). Risk in supply networks. Journal of Purchasing and Supply Management, 9(2), 51–62.

Isdell, N. (2009). Global corporate citizenship series: Twenty-first century capitalism: The new role business must play. CFR Corporate Conference http://www.cfr.org/ economics/global-corporate-citizenship-series-twenty-first-century-capitalism- new-role-business-must-play-video/p18722 Accessed 11.02.11.

Jensen, N. M., & Young, D. J. (2008). A violent future? Political risk insurance markets and violence forecasts. Journal of Conflict Resolution, 527–547.

Katzenstein, P. J. (2005). A world of regions: Asia and Europe in the American Imperium. Ithaca, NY: Cornell University Press.

Kaufmann, D., Kraay, A., & Mastruzzi, M. (2010). The worldwide governance indi- cators: Methodology and analytical issues. World Bank Policy Research Working Paper No. 5430.

Khanna, T., Palepu, K. G., & Sinha, J. (2005). Strategies that fit emerging markets. Harvard Business Review.

Kohli, A. (2004). State directed development: Political power and industrialization in the global periphery. New York: Cambridge University Press.

Kuznets, S. (1955). Economic growth and income inequality. American Economic Review, 45(March (1)), 1–28.

Moore, B. (1966). Social origins of dictatorship and democracy: Lord and peasants in the making of the modern world. Boston: Beacon Press.

Nath, H. K. (2008). Country risk analysis: A survey of the quantitative methods. Working Paper Series (pp. 1–31). Huntsville, TX: Sam Houstan State University.

Oetzel, J. M., Bettis, R. A., & Zenner, M. (2001). Country risk measures: How risky are they? Journal of World Business, 36(2), 128–145.

Park, H. S. (1984). Human needs and political development: A dissent to utopian solutions. Cambridge, MA: Schenkman Pub.

Perez-Aleman, P., & Sandilands, M. (2008). Building value at the top and the bottom of the global supply chain: MNC-NGO partnerships. California Management Review, 51(1), 24–49.

Porter, M. E., & Kramer, M. R. (2011). Creating shared value. Harvard Business Review, 62–77.

Rapley, J. (2007). Understanding development: Theory and practice in the third world. Boulder, CO: Lynne Rienner Publishers.

Ravallion, M. (2009). The developing world’s bulging (but vulnerable) ‘Middle class’. World Development Research Group, Policy Research Working Paper 4816.

Ravallion, M. (2012). Mashup indices of development. World Bank Research Observer, 27(1), 1–32.

Reich, R. B. (2007). Supercapitalism: The transformation of business, democracy, and everyday life. New York: Alfred A Knopf.

Sachs, J. (2005). The end of poverty: Economic possibilities for our time. New York: Penguin Press.

Sachs, J. (2012a). Government, geography, and growth: The true drivers of economic development. Foreign Affairs.

Sachs, J. (2012b). Reply to Acemoglu and Robinson’s response to my book review. http:// jeffsachs.org/2012/12/reply-to-acemoglu-and-robinsons-response-to-my-book-review/ Accessed 12/12/12.

Sen, A. K. (1999). Development as freedom. New Delhi: Oxford University Press. Shea, M. (2014). Unloosing the knot: Developing an understanding of the relationship

between war and sustainable development. Working paper presented at the 2014 Southern Political Science Association Conference in New Orleans.

Skidelsky, R. (2009). Keynes: The return of the master. New York, NY: PublicAffairs. Taleb, N. N., Goldtein, D. G., & Spitznagel, M. W. (2009). The six mistakes executives

make in risk management. Harvard Business Review, 78–81.