WPS8478 Policy Research Working Paper 8478 Benchmarking Costs of Financial Intermediation around the World Pietro Calice Nan Zhou Finance, Competitiveness and Innovation Global Practice June 2018 Policy Research Working Paper 8478 Abstract The costs of financial intermediation have important it uses regression analysis to examine the underlying consequences for financial development. Using bank- bank-level, structural, macroeconomic, and institutional level data for 160 countries during 2005–14, this paper determinants of net interest margins. Finally, the paper analyzes the composition and sources of bank net inter- uses the results of the econometric analysis to construct est margins. First, it uses an accounting decomposition country-level bar charts of relative contributing factors framework to provide summary statistics on the size of net to financial intermediation costs. The results provide interest margins and highlight the cost and profit compo- evidence-based guidance on key areas of structural nents in countries, regions, and income groups. Second, reforms to reduce the costs of financial intermediation. This paper is a product of the Finance, Competitiveness and Innovation Global Practice. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/research. The authors may be contacted at pcalice@worldbank.org and nzhou@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Benchmarking Costs of Financial Intermediation around the World Pietro Calice and Nan Zhou† JEL Classification Numbers: G21, G28 Keywords: Net Interest Margins, Costs of Financial Intermediation, Banks Authors email addresses: pcalice@worldbank.org; nzhou@worldbank.org † World Bank Group. We are indebted to Augusto de la Torre, Erik Feyen, Alfonso Garcia Mora, Davide Mare, Zafer Mustafaoglu and Martin Raiser for their comments on earlier versions of this paper. We would also like to thank all participants to the seminar held on May 1st, 2018 at the World Bank for their valuable suggestions. Errors and omissions are only ours. Benchmarking Costs of Financial Intermediation around the World 1. Introduction Bank financial intermediation, that is, channeling funds from units in surplus to units in deficit, plays a critical role in sustainable and inclusive growth. There is a considerable body of evidence showing that the extent to which an economy is making use of banking intermediation is not only associated with economic growth (Figure 1) and broader access to financial services (Figure 2) but it is a causal factor in explaining overall economic performance (see, for example, Levine, 2005), poverty reduction (e.g., Beck et al., 2007) and reduced inequality (e.g., Demirgüç-Kunt and Levine, 2009). The costs associated with financial intermediation have an important bearing on the depth and breadth of the banking system. High costs of financial intermediation are associated with credit rationing and thus a lower level of credit channeled to borrowers (Stiglitz and Weiss, 1981). The interest spread, i.e., the difference between the lending rate and the deposit rate, is a commonly accepted measure of how costly financial intermediation is for a society. Countries with lower intermediation spreads experience higher levels of financial development (Figure 3) and higher penetration in the use of financial services (Figure 4). Understanding the determinants of intermediation spreads is, therefore, important to inform policy for improving overall bank efficiency of intermediation and achieving financial deepening and inclusion, especially for those countries that in the absence of developed capital markets heavily rely on bank financing as a source of external funding. The starting point for analyzing the determinants of banks’ intermediation spreads is the seminal work by Ho and Sanders (1981). In their pioneering study, Ho and Sanders (1981) model banks as mere intermediaries between lenders and borrowers where intermediation spreads depend on four basic components: (i) the degree of bank risk aversion; (ii) the degree of competition in the banking market; (iii) the average size of bank transactions; and (iv) interest rate risk. This model has been subsequently extended to incorporate additional factors explaining net interest margins. McShane and Sharpe (1985) incorporated money market rates. Allen (1988) considered loan heterogeneity in the model (banks offer different types of loans and deposits) and showed that margins can be reduced as a result of diversification of banking services. Angbazo (1997) introduced credit risk (and its interaction with interest rate risk). Maudos and Fernandez de Guevara (2004) considered the importance of operating costs as a determinant of banks’ margins. More recently, Carbó Valverde and Rodriguez (2007) included both traditional and nontraditional activities in order to study the effect of specialization on net interest margins. While the model by Ho and Sanders (1981) and its subsequent extensions remains the workhorse of the theoretical literature, its empirical verification is challenging due to the influence of macroeconomic and institutional variables which are difficult to capture in the theoretical model. To circumvent this problem, some authors (see, for example, Ho and Saunders, 1981; Saunders and Schumacher, 2000) use a two-stage estimation procedure. In the first stage, the effect of the explanatory variables of the interest margin not explicitly introduced in the theoretical model is controlled for in order to estimate a “pure” margin. In the second stage, the relationship between the “pure” spread and the factors posited by the theoretical model is analyzed. The application of a two-stage approach has the advantage that it allows a “pure” net interest margin to be estimated, although it requires long time series. On the other hand, other authors (see, for example, McShane and Sharpe, 1985; Angbazo, 1997; Maudos and Fernandez de Guevara, 2004; Carbó Valverde and Rodriguez, 2007) employ a single-stage estimation procedure, where all the variables of the theoretical model along with the other additional variables used to estimate the “pure” spread are included in the explanation of the net interest margin. The empirical literature using cross-country samples or focusing on a single developed or developing country broadly confirms the predictions of the theory. Spreads are found to be positively related to banks’ operating costs, suggesting that banks pass on to their customers their higher transformation costs (see, for example, Maudos and Fernandez de Guevara, 2004; Carbó Valverde and Rodriguez, 2007; Williams, 2007; Beck and Hesse, 2009). There is also a consensus that intermediation spreads tend to increase with market power and low competition (see, for example, Angbazo, 1997; Saunders and Schumacher, 2000; Dabla-Norris and Floerkemeier, 2007; Williams, 2007; Almarzoqi and Ben Naceur, 2015). Finally, research shows a consistently positive relationship between bank margins and interest rate risk (see Maudos and Fernandez de Guevara, 2004; and Maudos and Solis, 2009, among others), indicating that banks facing uncertainty and volatility require a higher premium to compensate for reinvestment risk and refinancing risk. With regard to other determinants of interest margins, findings are more mixed. For example, some studies (Ho and Saunders, 1981; and Maudos and Solis, 2009, among others) find a positive association between net interest margins and the size of the operations (proxied by the log of total assets or total loans), suggesting that the larger the transaction the larger the expected loss, while others (for example, Angbazo, 1997; and Maudos and de Guevara, 2004) report a negative relationship, pointing to the presence of economies of scale in financial intermediation. Similarly, while most of the papers find that credit risk exerts a positive effect on margins (see, for example, Angbazo, 1997; Maudos and Fernandez de Guevara, 2004), suggesting that banks charge a risk premium to compensate for credit risk, others (for example, Williams, 2007; Fungáčová and Poghosyan, 2011) find a negative relationship between credit risk and interest margins, interpreting this result as evidence that banks are unable to accurately price risk or that depositors require a higher risk premium for putting their money in riskier banks. Finally, Saunders and Schumacher (2000) and Brook and Rojas (2000), among others, find that risk aversion, proxied by the capitalization ratio, positively determines interest margins, implying that those banks that are more risk averse demand a higher premium to compensate for the higher costs of equity financing compared to external financing. However, for example, Dabla-Norris and Floerkemeier (2007) and Horvath (2009), find a negative association between capitalization and net interest margins, suggesting that less capitalized banks have to take on more risks (translating into higher margins) to receive higher returns and internally boost their reserves. In recent years, studies have begun exploring the potential impact of bank ownership on net interest margins. Micco et al. (2007) and Fungáčová and Poghosyan (2011) find that the form of bank ownership has a strong impact on bank margins in developing countries and the Russian Federation, respectively. Martinez and Mody (2004) find that in Latin America foreign banks can charge lower spreads relative to local banks, while the opposite conclusion is obtained by Schwaiger and Liebig (2008) for a sample of CEE economies. No effect of foreign ownership is found by Dabla-Norris and Floerkemeier (2007) in Armenia, and by Beck and Hesse (2009) in Uganda. Few papers examine the role of macroeconomic conditions and the regulatory framework in explaining banks’ intermediation spreads. For example, Demirgüç-Kunt et al. (2004) and Claeys and Van der Vennet (2008) find a positive relationship between inflation and bank margins, while Brock and Rojas (2000), Saunders and Schumacher (2000) and Gelos (2006) show that higher reserve requirements are associated with higher spreads. Others (for example, Demirgüç-Kunt et al., 2004, Gelos, 2006, and Poghosyan, 2012) control for the quality of the institutional and legal setup and find that in general less supportive environments are associated with higher margins. This paper adds to existing studies on two main fronts. First, using bank-level panel data on more than 14,000 commercial banks in 160 countries for the period 2005-2014, and controlling for country-level 3 structural, macroeconomic and institutional data, this paper provides an empirical analysis of the determinants of intermediation spreads in all countries in the world for which data are available, representing to the best of our knowledge the most comprehensive attempt to study net interest margin determinants in a large sample of countries since the seminal contribution by Demirgüç-Kunt and Huizinga (1999). Second, and most importantly, this paper uses the results of the empirical analysis to benchmark country-level costs of financial intermediation with a view to highlight the factors that most affect bank intermediation spreads. Our main concern is not to uncover new findings about the drivers of intermediation spreads but rather to provide an easy-to-interpret tool to highlight the most important factors by country. To this end, empirical results are used to decompose the difference between each country’s net interest margins and those of the average banking system in the world as well as of the average banking system of the regional grouping to which the country belongs. Simple bar charts of relative contributing factors are then constructed to sharpen the intuition behind the estimates to show “what determines a country’s net interest margin.” Benchmarking costs of financial intermediation around the world is a useful exercise in detecting deficiencies at the country level; it can also help policy makers identify areas to prioritize through reform. For example, if high intermediation costs are especially driven by high operating costs in the banking system, efforts to reduce the opportunity cost of holding reserves are unlikely to bring down spreads. Improving productivity by, for example, incentivizing an increased use of technology and automation in the production and distribution of banking services would be a more effective strategy. On the other hand, if macroeconomic volatility creates uncertainty that banks need to compensate for by charging higher risk premia, reducing information asymmetries through the promotion of credit bureaus or improving contract enforcement through more effective liquidation procedures will do little to eliminate the underlying drivers of high spreads. Understanding the main determinants of intermediation costs is also relevant to the agenda of many international financial institutions, especially multilateral development banks, which are increasingly concerned with maximizing the additionality of development finance by crowding in private capital. High costs of financial intermediation represent an important obstacle to leveraging private sector investments for growth and sustainable development. Therefore, understanding the factors driving bank spreads is important for the implementation of this agenda. Along the same lines, analysis of bank intermediation efficiency is an increasingly common component of IMF/World Bank-led Financial Sector Assessment Programs (FSAPs) and other country-level diagnostics, hence developing an international benchmark in this area can help focus the analysis. The remainder of this paper is organized as follows. The next section presents the methodology and the data used to analyze the drivers of bank intermediation costs around the world. Section 3 presents an accounting decomposition of the intermediation spread in its constituent parts, i.e., cost and profit components. Section 4 presents the results of the econometric model and evaluates the importance of bank-specific, structural, macroeconomic and institutional factors affecting spreads. Section 5 introduces the methodology used to decompose econometric results to build bar-charts of relative contributing factors of financial intermediation costs. The last section concludes. 2. Methodology and data Bank intermediation costs can be measured using both ex-ante and ex-post spreads. The ex-ante spread is the difference between contractual rates charged on borrowers and those paid on depositors. The ex- post spread is the difference between banks’ actual interest revenue and their actual interest expense, usually divided by earning assets or total assets. The ex-post spread differs from the ex-ante spread by 4 the amount of loan defaults. Therefore, ex-post spreads allow for a broader examination of the costs of financial intermediation. Moreover, ex-post spreads allow for an analysis of the driving factors of efficiency in a general equilibrium setting, taking into account all operations of a bank. For example, banks may compensate high taxation of a certain form of intermediation by charging lower rates on other forms. This effect would not be captured by ex-ante spreads. Finally, ex-post spreads are easy to calculate from banks’ financial statements and, though accounting principles may still differ across countries, recent convergence towards intentional standards makes comparability a relatively minor issue. Ex-ante spreads are typically calculated at the aggregate level and put together from a variety of sources, which may be different from country to country. For all these reasons, this paper uses ex-post spreads as a measure of bank intermediation costs. We begin our analysis of the drivers of bank intermediation costs by providing an accounting decomposition of ex-post spreads around the world. The decomposition of net interest margins can be a useful exercise to get to the factors that drive intermediation costs in a banking market. An accounting decomposition of bank net interest margins can be derived from a straightforward accounting identity: BTP/TA ATP/TA TX/TA. (1) where BTP/TA is before-tax profits to assets, ATP/TA is after-tax profits to assets, and TX/TA is taxes to assets. From a bank’s income statement, before-tax profits must satisfy the following accounting identity: BTP/TA NI/TA NII/TA – OV/TA – LLP/TA (2) where NI is net interest income, NII refers to noninterest income, OV stands for overhead costs, and LLP refers to loan loss provisioning, all scaled by total assets, TA. The identities above allow for a decomposition of net interest margins, NI/TA, into its components: NI/TA ATP/TA TX/TA – NII/TA OV/TA LLP/TA (3) The above accounting identity suggests a useful breakdown of the realized spreads in its constituent parts, providing initial indications of differences across countries. However, comparing accounting ratios without controlling for differences in the structure of the banking market as well as in the macroeconomic and institutional environments in which banks operate may be misleading. The next step of the paper is therefore to provide an econometric analysis of the determinants of net interest margins, contributing to shed light on the main drivers of intermediation costs. Based on the framework developed by Ho and Saunders (1981) and its subsequent extensions, we employ a single-stage estimation procedure and estimate the following basic model: , , , , , , , , , (4) where NIM is the net interest margin for bank in country at time ; B is a vector of bank-level variables; S and I are a set of country-level macroeconomic and structural variables, and institutional variables, respectively; , is the per capita income, and , , the error term. NIM, B, S, and I are standardized at country level to generate sensitivities of the net interest margin with respect to the main dependent variables. The empirical specification (4) is estimated with country fixed effects and time fixed effects, allowing for heteroscedasticity by applying robust standard errors. To correct for the fact that the number of banks varies considerably across countries, we use the weighted least square 5   technique, with the weight given by the inverse of the number of banks for the country in each year. Finally, as an extension, the paper also estimates a model to explore potential variations in how bank, macroeconomic and structural, and institutional variables affect net interest margin at different levels of economic development: , , α , , , , , , , , , , , , , , (5) The sample is formed by an unbalanced panel of more than 14,000 banks in 160 countries for the period 2005-2014. Bank-level data are taken from the Bankscope database. We focus on depository financial institutions, including commercial banks, savings banks, and cooperative banks. To address issues of redundant financial statements in Bankscope, we adjust for mergers and acquisitions among banks and retain unconsolidated statements only for those with concurrent consolidated reports, using the procedure discussed in Duprey and Lé (2016). Structural, macroeconomic and institutional variables are drawn from the World Bank Global Financial Development Database, World Development Indicators, and Ease of Doing Business Indicators, respectively. Our dependent variable, the net interest margin, is calculated as the ratio of total interest revenue minus total interest cost divided by average earning assets. In line with previous studies, this is our baseline measure of cost of financial intermediation. As a robustness check, however, we use the ratio of total interest revenue minus total interest cost divided by total assets. We use a variety of control variables to explain variations in net interest margins. Specifically, we use the following set of bank-level variables:  Size of operations: this variable is proxied by the log of gross loans. While ideally we would like to measure this variable by the average lending transaction size, data are not available for our sample so, in line with the previous studies, we resort to the whole loan book taken in log. The benchmark theoretical model posits a positive relationship between the size of bank operations and margins since, for a given credit and market risk, the larger the operation the larger the potential loss and hence the higher the margin the bank will demand. However, the presence of economies of scale stemming from size or diversification would suggest that the fixed costs associated with the transaction are spread over a larger base, enabling the bank to achieve a smaller margin. Therefore, we do not have a particular prior on the expected sign of this covariate.  Risk aversion: as imperfect as it is, we proxy the bank’s degree of risk aversion by the ratio of equity to total assets, again, in line with similar studies. As equity is more costly than other sources of funding, a higher proportion of equity in the bank’s capital structure indicates greater risk aversion and is expected to be reflected in higher margins. Hence, the estimated coefficient is expected to be positive.  Opportunity cost of bank reserves: this represents a regulatory and opportunity cost for the bank and is proxied by the ratio of cash and balances held at the central bank over total assets. The opportunity cost arises from the foregone interest the bank can earn from investing in higher-yielding assets the money it keeps in cash reserves. A positive association is therefore expected between this variable and net interest margins because banks must compensate for the missing interest. 6    Overhead: this variable captures cross-bank differences in the organization and operation of the bank. Banks incurring higher staff and administrative costs are likely to pass on to their clients these costs by increasing margins. Therefore, a positive sign is expected for this variable. We proxy operating costs by the ratio of operating expenses to total assets.  Credit risk: the risk of default on a credit requires the bank to apply a risk premium to the interest charged to clients. In the absence of better alternatives for our sample and in line with previous studies, we proxy credit risk by a stock measure: the ratio of loan loss reserves to loans. Higher provisions are expected to be associated with higher credit risk and therefore higher margins; hence, we anticipate a positive relationship between this covariate and our dependent variable.  Income diversification: a diversified bank is able to attract new customers and offer a wider array of products and services, benefiting from fee-based income. By doing so, the bank is expected to provide interest-dealing services with lower margins as these will be compensated for by higher fees and commissions due to cross-selling and cross-subsidization. We proxy bank product diversification by the ratio of noninterest income to total revenue and expect a negative sign for its coefficient. Bank-level variables are complemented by the following country-level macroeconomic and structural variables:  Inflation: like with interest rates, macroeconomic instability introduces uncertainty that needs to be compensated by higher spreads. We measure macroeconomic instability by the inflation rate proxied by the CPI variable, which is expected to have a positive relationship with net interest margins.  Interest rate risk: uncertainty in market rates is expected to translate into higher margins as banks will try to hedge interest rate risk by applying a risk premium. We use the standard deviation of the monthly money market rate (or policy rate when the latter is not available) as a proxy for interest rate risk. A positive sign for this variable is expected.  Competition: according to our benchmark theoretical model (Ho and Saunders, 1981), a competitive market structure is expected to put pressure on margins and therefore a negative relationship is expected between competition and net interest margins. In line with recent studies that use behavioral measures of competition, we measure the degree of competition in the banking market by the country-level Lerner index. The latter is the difference between the price and the total marginal costs as a proportion of the price. A lower Lerner index indicates a strong degree of competition in the banking sector, which results in lower margins; therefore, a positive relationship between the Lerner index and the interest margins is expected: banks with greater market power can charge higher spreads than they could in a more competitive market.1 We also include the following country-level institutional variables:  Creditor rights: the degree to which collateral and bankruptcy laws protect the rights of lenders is expected to facilitate lending and ultimately reduce bank margins. We measure this by the                                                              1 As a robustness check, we use the concentration ratio and show that the latter is not a significant predictor of competition, in line with recent studies (see, for example Beck and Hesse, 2009). Results are available upon request. 7   Doing Business Indicator strength of legal rights index, which tracks changes related to secured transactions and insolvency every year. The index ranges from 0 to 12 for the period 2013-14 and from 0 to 10 during 2005-12, with higher scores indicating that collateral and bankruptcy laws are better designed to expand access to credit. In order to mitigate the differences in the Doing Business methodology, we use a factor of 1.2 to scale the 2005-12 raw data values. A negative sign is expected as stronger creditor rights are expected to translate into lower spreads.  Information environment: more credit information sharing on borrowers lowers the cost of screening and monitoring, reduces adverse selection and reduces loan losses, thus contributing to lower bank spreads. We proxy the quality of the information environment by the Doing Business Indicator depth of credit information index, which measures rules and practices affecting the coverage, scope and accessibility of credit information available through either a credit bureau or a credit registry. The index ranges from 0 to 8 during 2013-14 and from 0 to 6 during 2005-12. For similar reasons discussed above for creditor rights, we scale the 2005-12 value by a factor of 4/3. Higher values indicate the availability of more credit information, from either a credit bureau or a credit registry, to facilitate lending decisions. A negative sign is expected on the coefficient of this variable.  Contract enforcement: the enforceability of creditor rights is important in determining the premium that banks will demand on their intermediation activity. We proxy contract enforcement by the Doing Business Indicator recovery rate, which is recorded as cents on the dollar recovered by secured creditors through judicial reorganization, liquidation or debt enforcement (foreclosure or receivership) proceedings. Higher recovery rates are expected to reduce bank margins; hence, we expect a negative coefficient for this variable. Lastly, we include one broad measure of economic development in our analysis:  GDP per capita: real per capita income measured in thousands of constant 2010 US dollars. More advanced economic development tends to associate with deeper and more efficient financial systems, which are conducive to lower net interest margins. A negative sign for this variable is therefore expected. In a modeling extension, this variable is also used as a simple proxy for testing heterogeneous effects of bank-level, macroeconomic, structural, and institutional variables along different stages of economic and financial development. Bank-level variables are trimmed at the 1 percent level while macroeconomic variables are winsorized at 0.5 percent to eliminate multiple outliers, implausible negative values and extreme outliers. Table 1 presents descriptive statistics while Table 2 summarizes the correlation matrix. 3. Accounting decomposition of intermediation spreads We begin our analysis by presenting a breakdown of the net interest margin into its components: operating costs minus noninterest income, loan loss provisions, taxes and net profit. The average net interest margins (weighted by total assets) for the period 2005-14 by country are presented in Table 3, column 2. Columns 3-6 report its components as a share of the net interest margin while columns 7-11 present accounting ratios as in equation (3) above, i.e., components scaled by total assets. Figure 5 8   presents similar statistics on net interest margins for groupings by region and income, while Figures 6-7 show trends in net interest margins by region and income group.2 Several countries, especially advanced European economies hit hard by the global financial crisis, display very low net interest margins (below 1 percent), signaling the prevalence of difficult operating conditions since 2007, characterized by a combination of low interest rates and low demand, which inevitably impacted bank margins. At the other end of the distribution (top decile), we mostly find emerging markets and developing economies from Sub-Saharan Africa and Latin America, where average spreads are between 4 percent and 5 percent (Table 3, column 2). This is confirmed by data at the aggregate level: the net interest margin is the highest for banks operating in the Latin America region and for banks in low income countries (Figure 5). Unsurprisingly, operating costs represent a significant component of the net interest margin in many jurisdictions (Table 3, column 8). Commercial and retail banking involve the establishment of a large number of branches, equipment and personnel to serve and monitor clients. Inefficient organizational structures and a low level of automation of loan and deposit production can exacerbate the importance of overhead in contributing to intermediation costs. In this context, we observe significant variation in operating costs across jurisdictions, with overhead as a share of total assets ranging from 0.8 percent in Japan to 16.4 percent in Russia. Interestingly, we find that in the top quartile (i.e., the most virtuous countries in terms of cost efficiency) of the distribution, along with many advanced economies, characterized by a relatively high degree of technological adoption offsetting high wages and high density branch network, there are also several, small, densely populated emerging economies such as Bahrain, Lebanon and Mauritius, which may be able to exploit economies of scale from a geographically concentrated distribution network. The economies of Sub-Saharan Africa stand out with high shares of overhead, followed by countries in Europe & Central Asia and Latin America (Figure 6a). This is the case also for low income countries (Figure 6b). Table 3, column 4, shows the share of taxation absorbed by the net interest margin. Bank taxation includes several forms of explicit and implicit taxes. The former includes mostly corporate income taxes while the latter refers to reserve and liquidity requirements and other restrictions on credit. Only explicit taxes are reflected in column 4, as implicit taxes lower directly the net interest income of the bank. We can observe significant cross-country variation in the extent to which banks are explicitly taxed. This reflects not only differences in corporate income tax rates but also the treatment of bad and doubtful loans and the non-application in many cases of thin capitalization rules (i.e., interest deductions) to the financial sector. In this regard, the very low if not zero taxation observed in the countries of the Gulf Cooperation Council (GCC) reflects the absence of significant income taxation in those economies, whereas the low tax share found in Portugal, Serbia or Spain is presumably explained by the tax deductibility of nonperforming loans. Loan loss provisions are also part of the intermediation spread as banks have to take into account historic losses when contracting new loans. The share of provisioning over the net interest margin can be interpreted as a crude measure of asset quality (Table 3, column 5); however, the data are heavily influenced by the different accounting and regulatory treatments of provisioning and nonperforming loans in various jurisdictions. Unsurprisingly, the top quartile is dominated by European countries,                                                              2 Regional groups exclude OECD high-income economies: Australia, Austria, Belgium, Canada, Chile, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Israel, Italy, Japan, Korea, Rep., Latvia, Luxembourg, Netherlands, New Zealand, Norway, Poland, Portugal, Slovak Republic, Slovenia, Spain, Sweden, Switzerland, United Kingdom, and United States. 9   which have experienced a significant rise in nonperforming loans in the aftermath of the global financial crisis (Figure 6a). Countries such as Cyprus, Greece and Iceland all present high levels of loan loss provisions as a percentage of net interest margins. Other countries such as Kazakhstan and Tunisia, where problems in the banking sector emerged recently, also show a relatively high level of provisioning. Table 3 also presents statistics on the share of noninterest income over total assets, as in accounting identity (3) above. Though not a direct component of spreads, the relative importance of fee-based services reveals the extent to which banks are able to diversify their income sources away from the lending business, which can allow banks to operate with low margins while maintaining adequate levels of profitability (Table 3, column 7). Banks in some Latin American countries such as Colombia and Peru, and banks operating in Europe & Central Asia such as Croatia and Slovak Republic seem to rely heavily on fee-based activities. At the regional level, this pattern seems to be strongest in Sub-Saharan Africa (Figure 5). In contexts characterized by low traditional intermediation activity, noninterest sources of income may provide a boost to bank profitability. As a residual, Table 3, column 6 presents the extent to which net interest margins translate into bottom line profitability, i.e., the profit margin of the lending business. Again, there are large differences across countries. Disproportionately high profit margins may signal competitiveness issues in the banking system. Entry barriers, competition from other financial institutions, the ownership structure of the banking sector, market segmentation, the presence of switching costs and the competition law system can all have a significant impact on the ability of banks to exploit market power and achieve high profits. From this perspective, the fact that, for example, GCC countries populate the top quartile is in line with recent research that highlights low competition in these banking systems (see, for example, Calice et al., 2016; and Anzoategui et al., 2010). On the other hand, large profits may also indicate high country risk, with especially foreign banks insisting on large returns to compensate for a high degree of country-level economic and political uncertainty. This may contribute to explain the high profit margins of countries such as Iraq or Republic of Congo. At a regional level, banks operating in Sub-Saharan Africa and Latin America have the highest ratio of profits to total assets, while the lowest is found in high income economies, probably reflecting higher levels of competition. A final breakdown, presented in Figure 7, is by bank size. Larger banks tend to be more profitable and efficient, facing lower overhead. However, they tend to pay more taxes and need to account more for bad loans. 4. Econometric estimates of intermediation spreads The accounting decomposition framework presented in the previous section allows us to identify the items in the banks' income statement that make up the net interest margin. However, this rather mechanical exercise is only the first step towards ascertaining the main drivers behind the costs of financial intermediation since it does not incorporate how banks respond to country-level structural, macroeconomic and institutional features. Therefore, we complement our analysis with standard econometric techniques. Specifically, we use cross-country averages and standard deviations to standardize the dependent and control variables discussed in Section 2 (see Table 1 for summary 10   statistics), and estimate the sensitivities of net interest margins with respect to dependent variables, as specified in equation (4).3 Let xi,j,t denote a variable for bank i in country j and year t. At the country-level, this variable is represented as xj, the average generated for j over i and t. Then, we use the mean mx and standard deviation sx of xj to measure the cross-country central tendency and dispersion of x. The dependent variable, as well as bank-level, macroeconomic, structural, and institutional variables are standardized as xi,j,t–mx /sx for the regressions. This procedure allows the estimated coefficients to be interpreted as the standard deviation change in the net interest margin per one standard deviation change in the independent variable. We first consider bank-level variables: size of operations, risk aversion, opportunity cost of bank reserves, operating costs, credit risk, and income diversification (Table 4a, column 1). The results show that banks with larger operations charge a higher margin, suggesting that any potential benefit arising from economies of scale is offset by higher risk, since for a given value of credit and market risk larger operations are expected to translate into a higher potential loss, for which banks demand a risk premium. We also find that more risk averse and better capitalized banks require higher margins, consistently with theory and previous empirical evidence. This can be explained by the reluctance of risk-averse banks to engage in more profitable but riskier lending activities. Our results also show that reserve requirements are a monetary or regulatory policy tool that could affect the cost of intermediation. Its proxy, the opportunity cost of bank reserves variable, is positively and significantly associated with spreads. Higher reserve requirements are translated into higher interest spreads to compensate for the missing incomes resulting from zero or low return on reserves. Next, we turn to operating costs. As expected, less efficient banks operating with higher costs charge higher margins, passing onto depositors and borrowers their higher overhead. As expected, higher credit risk is also associated with higher margins, as banks require higher profits to compensate for risk, and this effect is stronger in wealthier countries. Finally, as predicted by theoretical models, the noninterest income variable has a negative and significant effect, showing that banks engaging more in non-lending activities have lower intermediation spreads. This may reflect a strategy of cross-subsidization with traditional activities (see Carbó Valverde and Rodriguez, 2007; and Lepetit et al., 2008). For example, banks may reduce lending rates to borrowers who also use other bank services, generating fees and commissions such as payment services or underwriting of securities. Table 4a, column 2, evaluates the impact of macroeconomic and structural variables. The results show that macroeconomic and competitive conditions do affect net interest margins. Bank net interest margins increase with inflation, suggesting that macroeconomic instability creates uncertainty to which banks respond by lifting margins upward. We also find that the volatility in the money market rate is positive and significant: banks respond to increased uncertainty in benchmark rates by raising spreads. However, interest rate risk is not significant at standard confidence intervals. We finally find that the Lerner index, our measure of market competition, enters positively and significantly, suggesting that banks commanding market power can charge higher lending rates and offer lower deposit rates, resulting in higher costs of financial intermediation. Table 4a, columns 3-4, evaluates the impact of institutional variables when controlling for bank-level and country-level macroeconomic and structural variables. Though the coefficients for the institutional                                                              3 The estimations are based on a sample of 128 countries on which all specifications can be implemented to arrive at fully comparable coefficient estimates. 11   variables come with the expected negative sign, including that of GDP per capita, signaling that a less developed credit and institutional infrastructure is associated with higher margins, the results show that these variables have no distinct significant impact on net interest margins though they are jointly significant (estimated with the standard F-test). The lack of individual significance of these regressors may suggest that the bulk of their variability is absorbed by country fixed effects in our relatively short panel. Considering the evidence that the effect of changes in covariates on the net interest margin differs between rich and poor countries (see, for example, Demirgüç-Kunt and Huizinga, 1999; Pogosyan, 2012), we next address this potential heterogeneity by interacting our independent variables with GDP per capita, measured in thousands of constant 2010 US dollars, as specified in equation (5). The purpose of this exercise is to further explore whether our main results hold using different models. Table 4a, columns 5-8, presents the results. Overall, our results are confirmed, and as expected we find that the impact of our variables is generally stronger in less wealthy economies. There are nonetheless two notable exceptions: first, the impact of our credit risk measure becomes greater as income level increases, suggesting better risk-pricing capabilities by banks in higher income economies. Second, the impact of income diversification decreases at higher income levels, suggesting reduced opportunities for product bundling and tying in more effective competition law systems. To further validate our results and explore systematically potential influence by institutional variables on bank intermediation costs, we estimate groups of additional equations that introduce fixed effects incrementally, as presented in Table 4b: columns 1–2 present results for weighted cross-sectional models; columns 3–4 present results with additional time fixed effects, columns 5–6 show results with country fixed effects, and, finally, columns 7–8 show results with bank fixed effects. Each group contains a main and income-interactive model, with columns 5–6 being our benchmarks discussed previously. We find that bank-level and macroeconomic and structural variables broadly behave similarly across specifications, with the exception that the sign on size of operations changes from negative to positive when country or bank fixed effects are introduced. We conjecture that although there are overall economies of scale in lending operations in the cross-section, risks arising from large exposures contribute positively to the net interest margin once differences in average bank sizes across countries are considered. With respect to institutional variables, the effects are more mixed. In the cross-section, there is evidence that depth of credit information and enforceability of contracts are conducive to improving bank intermediation efficiency. The strength of collateral and bankruptcy laws seems positively correlated with higher net interest margins, but the effect becomes diminished and reverses direction once country and bank level fixed effects are introduced. Finally, as a further robustness test, Tables 5a and 5b use the ratio of net interest income to total assets as an alternative dependent variable to replicate the regression exercise, and arrive at largely similar conclusions. 5. Bar-charts of relative contributing factors of financial intermediation costs The final step of our empirical exercise is to use the results from our econometric estimates to illustrate relative contributing factors to financial intermediation costs. To this end, we use parameters from our benchmark estimated model to decompose the differences between each country’s net interest margin and that of the average banking system in the world to highlight the relative contribution of bank-level, macroeconomic, structural, and institutional variables. 12   Specifically, we summarize the covariates for each country by computing the pooled averages for each bank-level, macroeconomic, structural, and institutional variable x, then subtracting from it the cross- country mean mx and dividing the result by the cross-country standard deviation sx (as shown in Table 1). This produces a standardized score xj – mx /sx that can be multiplied directly with the corresponding coefficient estimate cx from the equation. The contribution from a particular variable on the net interest margin, compared to the average banking system, can then be generated by multiplying its standard score with the cross-country standard deviation of the dependent variable, sy. Formally, for country j, the contribution of variable x toward the country’s difference from the net interest margin in the average country is: dx,j sy cx xj – mx /sx (6) For the current exercise, we opt for the model in Table 4, column 4, for its parsimony, although the approach can be easily adapted to incorporate additional variables and alternative specifications deemed proper. Visualization for within- and between-country comparison is presented in Table 6a. Values in the cell represent dx,j , the percentage contributions from variables to the difference in net interest margin against the average banking system. The dashed line across each column indicates the level of the cross-country mean, representing the corresponding value in Table 1 and hereby normalized to 0. A conducive, below- average contribution from a variable is indicated with a blue bar, and an unfavorable, above-average contribution in red. The left and right ends of a cell represent respectively the most and least favorable contributions across countries. Table 6b presents results at the region level by aggregating average contributions for each variable across economies belonging to regional groups. Building on the same regression results, Tables 7a-g replicate the exercise by benchmarking each country’s variables against the averages of its regional group. Figure 8 compares the relative explanatory powers of covariates implied by our model. Overall, most of variation in net interest margin is captured by bank-level variables, with income diversification and overhead in the lead, followed by cash reserves of banks. We also identify size of operations, bank capitalization, as well as macroeconomic conditions including inflation and interest rate level to be influential correlates of bank financial intermediation costs. The impact of institutional variables is somewhat muted from our analysis. To illustrate how to use the bar-charts presented in Tables 6-7, we take the decomposition results for Saudi Arabia. Saudi banks in our sample of 2005-14 have an average ratio of net interest income to total assets of 2.63 percent, and a net interest margin (i.e. net interest income to average earning assets) of 2.86 percent. These values compare favorably against world-wide averages (3.61 percent and 4.54 percent, respectively) yet they stand above averages in high income economies (2.03 percent and 2.30 percent, respectively). The accounting decomposition exercise in Table 3 shows that Saudi banks, although not highly diversified in terms of non-lending income-generating business, benefit from relatively low overhead costs, smaller than average levels of loan loss provisions, zero corporate tax, and are able to convert most of net interest margin into profit. Comparing the Saudi Arabian banking system against the average system in the world, the econometric exercise confirms the patterns above by attributing a large difference (-64 basis points, bps hereafter) to lower noninterest expenses, as well as smaller amounts to lower provisions for loan impairments (-2bps) 13   and cash reserve holdings (-9bps) by banks. It also points out that stable macroeconomic conditions have been conducive to reducing costs of bank financial intermediation through inflation (-5bps) and interest rate (-3bps) channels. Our empirical exercise also highlights that high credit concentration (as indicated by size of operations, +51bps), relatively low price competition (as measured by Lerner index, +12bps), and relatively weak environment of contract enforcement (as proxied by recovery rate, +11bps) can be areas worthwhile of consideration when policies to improve banking intermediation efficiency are designed. 6. Conclusions The costs of financial intermediation have important consequences for financial development, including financial deepening and financial inclusion. Based on a comprehensive cross-country data set with bank- level data, this paper analyzed net interest margin determinants around the world. First, we use an accounting decomposition framework to provide summary statistics on the size of net interest margins and to highlight its cost and profit components in countries, regions and income groups. Second, we use regression analysis to examine underlying bank-level, structural, macroeconomic and institutional determinants of net interest margins. Finally, we use the results of our econometric analysis to illustrate the relative contribution of each variable, so as to provide guidance on key areas of structural reforms to reduce the costs of financial intermediation. We broadly confirm findings of previous research. The analysis provides evidence that the costs of financial intermediation are negatively associated with income levels: net interest margins are significantly higher in lower income countries. On the other hand, a regional focus shows that costs of financial intermediation are especially high in Latin America and Sub-Saharan Africa. Accounting decomposition of intermediation spreads indicates that higher margins in low income economies, Latin America and Sub-Saharan Africa are mostly explained by higher overhead (probably reflecting low economies of scale and low productivity), higher credit risk (which may reflect a weaker environment) and greater bank profitability (which may be due to low market competition). The econometric analysis highlights that all variables have the expected impact on net interest margins, with magnitude conditional upon levels of economic development. Higher intermediation spreads are significantly associated with larger operations, higher risk aversion, higher opportunity costs arising from reserve requirements, higher overhead, higher credit risk, lower competition, higher interest rate risk and higher inflation. In general, the impact of these variables is stronger in less wealthy economies. Finally, we find suggestive evidence that a relatively underdeveloped credit infrastructure negatively affects net interest margins across countries: deficiencies in the contractual and information frameworks are associated with higher intermediation spreads. The bar-charts of relative contributing factors to financial intermediation costs, which are constructed using a decomposition of the difference between each country’s net interest margin and the average banking system in the world, help visualize deficiencies and focus policy makers’ attention. While the usefulness of the bar-charts is to highlight country-level areas of focus, some interesting patterns emerge. Market-developing policies aimed at addressing deficiencies in the contractual and information frameworks and at maintaining macroeconomic stability can have important repercussions for the costs of financial intermediation in lower income economies and should be the main concern of policy makers in these countries. These policies should be complemented by market-enabling policies aimed at increasing contestability in the banking system, and by policies aimed at promoting technological innovation in the production and distribution of banking services to reduce overhead and increase 14   efficiency. On the other hand, further promoting income diversification and reducing credit concentration in the system should be the key focus of policy makers in higher income economies concerned with lowering the costs of financial intermediation. 15   References Allen, L. (1988). The determinants of bank interest margins: A note. Journal of Financial and Quantitative Analysis, 23(2), pp. 231–235. Almarzoqi, R. and S. Ben Naceur (2015). Determinants of bank interest margins in the Caucasus and Central Asia. IMF Working Paper 15/87. Washington DC. Angbazo. L. (1997). Commercial bank net interest margins, default risk, interest-rate risk, and off- balance sheet banking. Journal of Banking & Finance, 21(1), pp. 55-87. Anzoategui, D., M. Martinez and R. Rocha (2010). Bank competition in the Middle East and Northern Africa Region. Review of Middle East Economics and Finance, 6(2), Article 2. Beck, T. and H. Hesse (2009). Why are interest spreads so high in Uganda? Journal of Development Economics, 88(2), pp. 192-204. Beck, T., A. Demirgüç-Kunt, and R. Levine (2007). Finance, Inequality and the Poor. Journal of Economic Growth, 12(1), pp. 27-49. Brock, P. L. and L. Rojas (2000). Understanding the behavior of bank spreads in Latin America. Journal of Development Economics, 63(1), pp. 113-134. Calice, P., P. Buccirossi, and R. Cervone (2016). Competition in the GCC SME lending markets: An initial assessment. GCC Knowledge Series. Washington, DC: World Bank Group. Carbó, S. and F. Rodriguez (2007). The determinants of banks’ margins in European banking. Journal of Banking & Finance, 31, pp. 2043–2063. Claeys, S. and R. Vander Vennet (2008). Determinants of bank interest margins in Central and Eastern Europe: A comparison with the West. Economic Systems, 32(2), pp. 197-216. Dabla-Norris, E. and H. Floerkemeier (2007). Bank efficiency and market structure: what determines banking spreads in Armenia?, IMF Working Paper 07/134. Washington, DC. Demirgüç-Kunt, A. and H. Huizinga (1999). Determinants of commercial bank interest margins and profitability: some international evidence. World Bank Economic Review, 13(2), pp. 379-408. Demirgüç-Kunt, A., L. Laeven, and R. Levine (2004). Regulations, market structure, institutions, and the cost of financial intermediation. Journal of Money, Credit, and Banking, 36(3), pp. 593–622. Demirgüç-Kunt, and R. Levine (2009). Finance and inequality: theory and evidence. World Bank Policy Research Working Paper 4967. Duprey, T. and M. Lé (2016). Bankscope dataset: getting started. Available at SSRN: https://ssrn.com/abstract=2191449. Fungáčová, Z. and T. Poghosyan (2011). Determinants of bank interest margins in Russia: Does bank ownership matter?. Economic Systems, 35(4), pp. 481-495. Gelos, M. G. (2006). “Banking Spreads in Latin America”. IMF Working Paper 06/44. Washington DC. Ho, T. S. and A. Saunders (1981). The determinants of bank interest margins: theory and empirical evidence. Journal of Financial and Quantitative analysis, 16(4), pp. 581-600. 16   Horvath, R. (2009). Interest margins determinants of Czech banks. Institute of Economic Studies, Faculty of Social Sciences, Charles University, WP 11/09 (Prague). Lepetit, L., E. Nys, P. Rous, and A. Tarazi (2008). Bank income structure and risk: An empirical analysis of European banks. Journal of Banking & Finance, 32(8), pp. 1452-1467. Levine, R. (2005). Finance and growth: theory and evidence. Handbook of Economic Growth, 1, pp. 865-934. Martinez Peria, M. S. and A. Mody (2004). How foreign participation and market concentration impact bank spreads: evidence from Latin America. Journal of Money, Credit and Banking, 36(3), pp. 511-537. Maudos, J. and F. Fernandez De Guevara. (2004). Factors explaining the interest margin in the banking sectors of the European Union. Journal of Banking & Finance, 28(9), pp. 2259-2281. Maudos, J. and L. Solís (2009). The determinants of net interest income in the Mexican banking system: An integrated model. Journal of Banking & Finance, 33(10), pp. 1920-1931. McShane, R. W., and I. G. Sharpe (1985). A time series/cross section analysis of the determinants of Australian Trading bank loan/deposit interest margins: 1962–1981. Journal of Banking & Finance, 9, pp. 115–136. Micco, A., U. Panizza, and M. Yanez (2007). Bank ownership and performance. Does politics matter? Journal of Banking & Finance, 31(1), pp. 219-241. Poghosyan, T. (2013). Financial intermediation costs in low income countries: The role of regulatory, institutional, and macroeconomic factors. Economic systems, 37(1), pp. 92-110. Saunders, A. and L. Schumacher (2000). The determinants of bank interest rate margins: an international study. Journal of International Money and Finance, 19(6), pp 813-832 Schwaiger, M.S. and D. Liebeg (2008). Determinants of bank interest margins in Central and Eastern Europe. Financial Stability Report 14, pp. 68-84, Österreichische Nationalbank (Vienna). Stiglitz, J. E. and A. Weiss (1981). Credit Rationing in Markets with Imperfect Information. American Economic Review, 71(3), pp. 393-410. Williams, B. (2007). Factors determining net interest margins in Australia: domestic and foreign banks. Financial Markets, Institutions and Instruments, 16, pp. 145–165. 17   Figure 1: Financial depth and economic growth Source: World Development Indicators. Figure 2: Financial depth and financial inclusion Source: World Development Indicators; Financial Access Survey. 18   Figure 3: Costs of financial intermediation and financial depth Source: Global Financial Development Database. Figure 4: Costs of financial intermediation and financial inclusion     Source: World Development Indicators; Financial Access Survey. 19   Figure 5: Accounting decomposition of net interest margins by region and income level, 2005-14 average 6 5 4 % of total asset 3 2 1 0 EAP ECA LAC MNA SAS SSA OECD HIC UMI LMI LIC HIC Overhead - noninterest income Tax expenditure Loan loss provision After-tax profit Note: EAP: East Asia and Pacific; ECA: Europe and Central Asia; LAC: Latin America and Caribbean; MNA: Middle East and North Africa; SAS: South Asia; SSA: Sub-Saharan Africa. HIC: High-income countries; UMI: Upper-middle-income countries; LMI: Lower- middle-income countries; LIC: Low-income countries. Regional groups exclude high-income OECD countries. 20   Figure 6a: Accounting decomposition of net interest margins by region, 2005-14 East Asia and Pacific Europe and Central Asia 6 6 5 5 % of total assets % of total assets 4 4 3 3 2 2 1 1 0 0 2005 2008 2011 2014 2005 2008 2011 2014 Overhead - noninterest inc. Tax expenditure Overhead - noninterest inc. Tax expenditure Loan loss provision After-tax profit Loan loss provision After-tax profit     Latin America and Caribbean Middle East and North America 6 6 5 5 % of total assets % of total assets 4 4 3 3 2 2 1 1 0 0 2005 2008 2011 2014 2005 2008 2011 2014 Overhead - noninterest inc. Tax expenditure Overhead - noninterest inc. Tax expenditure Loan loss provision After-tax profit Loan loss provision After-tax profit     South Asia Sub-Saharan Africa 6 6 5 5 % of total assets % of total assets 4 4 3 3 2 2 1 1 0 0 2005 2008 2011 2014 2005 2008 2011 2014 Overhead - noninterest inc. Tax expenditure Overhead - noninterest inc. Tax expenditure Loan loss provision After-tax profit Loan loss provision After-tax profit   21   Figure 6a: Accounting decomposition of net interest margins by region, 2005-14, continued OECD High-income 6 5 % of total assets 4 3 2 1 0 2005 2008 2011 2014 Overhead - noninterest inc. Tax expenditure Loan loss provision After-tax profit Figure 6b: Accounting decomposition of net interest margins by income level, 2005-14 High-income economies Upper-middle-income economies 6 6 5 5 % of total assets % of total assets 4 4 3 3 2 2 1 1 0 0 2005 2008 2011 2014 2005 2008 2011 2014 Overhead - noninterest inc. Tax expenditure Overhead - noninterest inc. Tax expenditure Loan loss provision After-tax profit Loan loss provision After-tax profit Lower-middle-income economies Low-income economies 6 6 5 5 % of total assets % of total assets 4 4 3 3 2 2 1 1 0 0 2005 2008 2011 2014 2005 2008 2011 2014 Overhead - noninterest inc. Tax expenditure Overhead - noninterest inc. Tax expenditure Loan loss provision After-tax profit Loan loss provision After-tax profit   22   Figure 7: Composition of net interest margins by decile of bank asset size, 2005-14 100 5 80 4 % of total assets 60 3 Share, % 40 2 20 1 0 0 Smallest 2 3 4 5 6 7 8 9 Largest Bank asset size decile Overhead - noninterest income Tax expenditure Loan loss provision After-tax profit Net interest margin (right axis) Figure 8: Cross-country variations in net interest margin attributions, 2005-14 Standard deviation, percentage point 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Size of operations 0.20 Risk aversion 0.17 Overhead 0.60 Reserve opportunity cost 0.38 Credit risk 0.03 Income diversification 0.66 Inflation 0.15 Interest rate risk 0.03 Competition 0.06 Information environment 0.01 Creditor rights 0.02 Contract enforcement 0.07 Note: Results are based on equation in Table 4a, column 4. The standard deviation of aggregate contribution from real GDP per capita, country and time fixed effects, and regression residual is 1.83 percentage point. 23   Table 1: Summary statistics Variable Obs Mean Std Dev Min Max Mean Std Dev Source At observation level At country level Net interest margin, % 147,761 3.662 1.891 0.310 14.100 4.729 2.242 Bankscope Net interest income to total assets, % 152,997 3.137 1.556 0.000 11.146 3.459 1.486 Bankscope log(gross loans, million USD) 150,345 5.316 1.812 1.142 10.934 5.688 1.417 Bankscope Equity / total assets, % 152,641 11.743 8.620 1.606 77.778 13.470 4.717 Bankscope Noninterest expense / total assets, % 154,624 4.267 6.821 0.000 50.000 4.313 2.670 Bankscope Cash / total assets, % 154,460 5.948 7.670 0.000 42.857 11.991 8.347 Bankscope Loan loss provisions / gross loans, % 127,296 2.346 3.011 0.000 23.140 4.394 2.680 Bankscope Noninterest income / revenue, % 152,770 24.171 23.122 0.000 100.000 38.888 13.481 Bankscope Inflation, % 155,175 3.040 2.898 -1.674 23.642 5.304 3.188 WDI Money market rate, %, monthly average 153,381 2.523 2.868 0.001 13.560 5.446 3.602 WDI Country Lerner Index 152,164 0.217 0.235 -1.750 0.480 0.272 0.101 GFDD Depth of credit information 155,171 7.071 1.884 0.000 8.000 3.521 3.083 Doing Business Legal right index 155,171 8.693 2.669 0.000 12.000 6.200 2.705 Doing Business Recovery rate, % 154,994 68.431 20.842 0.000 92.900 33.916 24.369 Doing Business GDP per capita (thousand, USD) 155,028 40.040 18.324 0.219 141.165 13.887 20.983 WDI Note: The mean and standard deviation at country level are based on pooled bank-level averages for each country, for the stable sample of 128 countries reflected in the regression analysis. Table 2: Correlation matrix [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [1] Net interest 1.00 margin [2] Net interest 0.88 1.00 income / TA [3] log(gross loans) -0.26 -0.27 1.00 [4] Equity / TA 0.32 0.22 -0.32 1.00 [5] Noninterest 0.34 0.27 -0.21 0.41 1.00 expense / TA [6] Cash / TA 0.19 0.08 -0.18 0.13 0.10 1.00 [7] Loan loss provisions / gross 0.26 0.22 0.08 0.18 0.42 0.07 1.00 loans [8] Noninterest 0.06 -0.08 0.12 0.24 0.56 0.11 0.39 1.00 income / revenue [9] Inflation 0.50 0.38 -0.10 0.26 0.37 0.13 0.31 0.34 1.00 [10] Money market 0.35 0.27 -0.07 0.20 0.33 0.01 0.26 0.28 0.67 1.00 rate volatility [11] Country Lerner 0.09 0.08 -0.06 0.03 -0.03 0.13 -0.04 -0.10 0.05 -0.01 1.00 Index [12] Depth of credit -0.16 -0.07 -0.10 -0.12 -0.06 -0.11 -0.30 -0.29 -0.44 -0.28 -0.07 1.00 information [13] Legal right index -0.03 0.04 -0.26 -0.08 -0.15 0.04 -0.40 -0.34 -0.28 -0.22 0.17 0.45 1.00 [14] Recovery rate -0.38 -0.25 -0.07 -0.24 -0.25 -0.19 -0.41 -0.38 -0.60 -0.49 -0.02 0.60 0.54 1.00 [15] GDP per capita -0.41 -0.29 -0.08 -0.22 -0.29 -0.19 -0.47 -0.36 -0.60 -0.44 -0.01 0.45 0.52 0.75 1.00 24   Table 3: Accounting decomposition of net interest income to total assets, 2005-14 As a percentage of net interest income As a percentage of total assets [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] Net interest Overhead - Tax Loan loss After-tax Noninterest Overhead Tax Loan loss After-tax [1] Economy income to noninterest expenditure provisions profit income expenditure provisions profit total assets income Afghanistan 3.917 63.9 5.4 18.4 12.3 1.860 4.364 0.210 0.720 0.482 Albania 3.501 38.7 6.6 20.5 34.2 0.652 2.006 0.231 0.719 1.196 Algeria 2.098 4.5 17.9 29.4 48.2 0.971 1.066 0.375 0.617 1.011 Angola 4.161 6.0 10.1 24.9 59.0 3.320 3.570 0.420 1.038 2.453 Antigua and Barbuda 3.150 55.8 0.7 50.1 -6.6 1.936 3.693 0.023 1.579 -0.209 Argentina 4.013 9.4 25.6 14.8 50.2 5.559 5.936 1.025 0.596 2.014 Armenia 4.782 33.9 11.7 14.0 40.3 2.587 4.208 0.561 0.672 1.928 Australia 1.765 25.3 14.5 11.5 48.7 0.676 1.122 0.256 0.204 0.860 Austria 1.869 43.0 6.9 26.3 23.8 1.100 1.903 0.130 0.491 0.445 Azerbaijan 4.585 22.0 10.3 37.4 30.2 2.660 3.671 0.472 1.716 1.386 Bahrain 1.532 14.2 5.6 22.1 58.1 0.879 1.096 0.086 0.338 0.890 Bangladesh 2.756 8.9 31.3 34.9 24.9 2.296 2.541 0.862 0.963 0.686 Barbados 3.597 36.7 10.2 9.7 43.5 1.647 2.965 0.366 0.350 1.563 Belarus 3.155 32.7 17.9 25.6 23.8 4.810 5.842 0.566 0.806 0.752 Belgium 0.919 63.5 5.8 11.2 19.4 0.356 0.940 0.053 0.103 0.179 Belize 5.974 30.9 17.0 11.4 40.7 2.812 4.658 1.017 0.681 2.430 Benin 3.514 34.6 8.4 30.0 27.0 2.721 3.937 0.296 1.054 0.948 Bhutan 3.215 10.8 22.7 13.7 52.7 0.818 1.166 0.731 0.441 1.696 Bolivia 4.217 44.0 7.0 14.5 34.5 2.754 4.608 0.297 0.611 1.455 Bosnia and Herzegovina 3.347 46.7 3.2 32.2 17.8 1.915 3.479 0.108 1.078 0.596 Botswana 5.234 19.0 13.7 11.4 55.9 2.721 3.714 0.716 0.598 2.928 Brazil 5.338 41.7 6.3 25.9 26.0 2.025 4.252 0.339 1.385 1.387 Bulgaria 3.691 30.4 4.8 30.2 34.6 1.486 2.608 0.178 1.115 1.276 Burkina Faso 3.454 17.2 13.8 38.2 30.8 4.190 4.786 0.475 1.321 1.063 Burundi 5.845 25.9 19.3 17.6 37.2 4.568 6.080 1.129 1.028 2.175 Cabo Verde 3.386 50.8 5.2 20.5 23.4 1.353 3.074 0.178 0.695 0.792 Cambodia 4.529 32.4 12.0 11.8 43.8 1.420 2.888 0.542 0.536 1.983 Cameroon 3.643 26.7 23.9 17.1 32.3 3.119 4.091 0.872 0.623 1.176 Canada 1.679 33.3 11.9 11.1 43.7 1.494 2.053 0.199 0.186 0.734 Central African Republic 5.729 36.0 17.9 23.9 22.1 5.100 7.164 1.026 1.370 1.269 Chad 4.434 13.9 23.9 12.3 49.9 4.344 4.960 1.061 0.547 2.210 Chile 3.474 28.9 8.2 23.4 39.4 1.435 2.440 0.286 0.814 1.370 Table 3: Accounting decomposition of net interest income to total assets, 2005-14, continued As a percentage of net interest income As a percentage of total assets [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] Net interest Overhead - Tax Loan loss After-tax Noninterest Overhead Tax Loan loss After-tax [1] Economy income to noninterest expenditure provisions profit income expenditure provisions profit total assets income China 2.288 29.2 18.5 15.4 36.9 0.409 1.077 0.423 0.353 0.843 Colombia 5.161 20.5 12.3 28.7 38.5 4.313 5.372 0.634 1.482 1.987 Congo, Dem. Rep. 6.252 53.3 10.6 16.7 19.4 6.394 9.725 0.665 1.045 1.212 Congo, Rep. 2.945 -1.8 26.0 8.8 66.9 4.299 4.247 0.766 0.260 1.970 Costa Rica 4.928 56.8 7.1 10.2 25.9 1.659 4.456 0.350 0.503 1.277 Croatia 2.860 35.9 8.3 25.1 30.7 1.261 2.287 0.238 0.719 0.877 Cyprus 2.168 38.9 8.0 54.8 -1.7 1.091 1.933 0.174 1.188 -0.036 Czech Republic 2.639 23.0 11.1 12.9 53.0 1.296 1.904 0.292 0.341 1.398 Côte d'Ivoire 3.860 31.5 11.1 24.7 32.7 4.368 5.584 0.429 0.953 1.261 Denmark 1.572 39.5 8.0 30.0 22.6 0.777 1.397 0.125 0.471 0.355 Djibouti 3.737 41.6 11.5 10.6 36.3 1.327 2.883 0.430 0.395 1.356 Dominican Republic 7.206 56.8 6.2 11.6 25.4 2.859 6.951 0.445 0.836 1.832 Ecuador 4.904 48.5 6.4 18.3 26.8 3.160 5.539 0.312 0.898 1.315 Egypt, Arab Rep. 2.392 10.1 19.4 33.7 36.8 1.318 1.560 0.464 0.805 0.880 El Salvador 4.473 26.1 10.9 31.4 31.6 1.969 3.137 0.485 1.405 1.415 Estonia 2.288 0.6 7.1 29.2 63.1 1.574 1.587 0.163 0.668 1.445 Ethiopia 3.262 -34.2 34.6 14.3 85.4 3.118 2.002 1.127 0.466 2.786 Finland 0.993 16.2 16.9 8.1 58.9 0.965 1.125 0.168 0.080 0.584 France 1.000 43.8 11.5 13.8 30.8 0.776 1.214 0.115 0.138 0.308 Gabon 2.762 -15.5 32.7 19.0 63.8 4.959 4.530 0.903 0.525 1.763 Gambia, The 6.228 25.7 22.0 9.9 42.4 5.972 7.572 1.368 0.619 2.641 Georgia 5.929 42.0 6.5 25.6 25.9 3.304 5.791 0.387 1.516 1.538 Germany 1.462 60.7 10.3 16.4 12.5 0.756 1.643 0.151 0.240 0.183 Ghana 7.676 37.3 13.9 14.7 34.0 4.266 7.131 1.070 1.129 2.612 Greece 2.240 52.8 5.3 68.4 -26.4 0.503 1.685 0.118 1.532 -0.592 Grenada 4.414 73.1 2.5 8.9 15.5 0.963 4.188 0.111 0.394 0.684 Guatemala 7.885 38.7 4.8 37.0 19.6 0.932 3.980 0.379 2.916 1.542 Guinea 5.033 16.0 24.9 25.9 33.2 6.721 7.524 1.254 1.304 1.673 Guyana 4.196 26.4 22.1 6.3 45.3 1.720 2.826 0.926 0.264 1.900 Haiti 4.495 55.8 7.7 5.6 30.9 3.048 5.556 0.344 0.254 1.388 Honduras 6.421 52.5 9.2 16.4 21.9 1.964 5.335 0.589 1.055 1.406 Hungary 2.931 38.7 7.8 42.8 10.7 3.002 4.135 0.227 1.256 0.314 26   Table 3: Accounting decomposition of net interest income to total assets, 2005-14, continued As a percentage of net interest income As a percentage of total assets [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] Net interest Overhead - Tax Loan loss After-tax Noninterest Overhead Tax Loan loss After-tax [1] Economy income to noninterest expenditure provisions profit income expenditure provisions profit total assets income Iceland 3.108 -9.4 14.6 38.9 55.8 2.845 2.553 0.455 1.210 1.736 India 2.409 27.1 15.7 20.2 37.0 1.088 1.740 0.379 0.488 0.890 Indonesia 5.088 36.2 12.8 14.7 36.3 1.400 3.241 0.652 0.747 1.848 Iraq 2.531 -60.2 22.4 29.5 108.3 3.529 2.004 0.568 0.747 2.741 Ireland 0.480 10.6 38.6 151.3 -100.5 0.820 0.872 0.185 0.726 -0.482 Israel 2.248 45.8 12.6 13.1 28.6 1.247 2.275 0.283 0.294 0.642 Italy 1.957 45.3 10.9 34.0 9.8 1.204 2.091 0.213 0.666 0.191 Jamaica 6.916 49.9 10.7 7.7 31.6 2.560 6.013 0.743 0.536 2.184 Japan 1.042 53.2 12.4 12.1 22.2 0.252 0.807 0.130 0.126 0.232 Jordan 2.891 25.0 16.5 11.6 46.9 1.119 1.843 0.476 0.335 1.356 Kazakhstan 3.670 12.0 9.0 60.2 18.8 2.233 2.673 0.329 2.210 0.689 Kenya 6.515 28.0 17.1 10.1 44.8 3.732 5.559 1.111 0.657 2.920 Korea, Rep. 2.256 37.7 9.9 22.0 30.4 0.789 1.640 0.223 0.496 0.686 Kosovo 5.747 47.6 5.9 19.5 27.0 1.436 4.174 0.338 1.118 1.554 Kuwait 2.495 -3.1 4.3 35.8 63.1 1.152 1.073 0.107 0.893 1.573 Kyrgyz Republic 6.213 37.8 5.3 10.0 46.8 3.173 5.523 0.331 0.623 2.909 Lao PDR 2.637 17.7 15.9 10.7 55.8 1.820 2.285 0.418 0.282 1.471 Latvia 1.982 11.5 9.5 50.9 28.2 1.820 2.048 0.187 1.009 0.558 Lebanon 1.882 29.7 10.9 7.6 51.8 0.838 1.398 0.205 0.143 0.974 Lesotho 5.261 20.5 20.7 6.4 52.3 3.672 4.752 1.091 0.336 2.753 Liberia 4.414 12.8 18.6 36.2 32.4 7.153 7.717 0.821 1.599 1.431 Libya 1.461 3.5 27.6 28.9 40.0 0.793 0.844 0.403 0.422 0.584 Lithuania 1.876 24.6 8.6 36.9 29.9 1.259 1.722 0.161 0.692 0.561 Luxembourg 0.876 4.5 15.1 10.7 69.7 0.944 0.984 0.132 0.094 0.611 Macedonia, FYR 3.573 36.1 2.4 30.6 30.9 1.839 3.130 0.085 1.094 1.104 Madagascar 5.613 27.3 15.4 12.4 44.9 2.543 4.077 0.863 0.698 2.518 Malawi 5.296 11.8 22.0 10.4 55.8 7.833 8.459 1.167 0.549 2.954 Malaysia 2.189 17.3 17.1 13.2 52.5 0.897 1.274 0.373 0.288 1.149 Mali 4.257 33.6 11.3 29.6 25.5 3.435 4.864 0.480 1.261 1.087 Malta 2.385 29.9 21.7 6.7 41.7 0.758 1.472 0.517 0.160 0.994 Mauritania 3.852 11.9 8.4 53.3 26.4 3.480 3.938 0.324 2.053 1.016 Mauritius 2.459 9.7 9.4 14.3 66.5 1.223 1.463 0.232 0.352 1.636 27   Table 3: Accounting decomposition of net interest income to total assets, 2005-14, continued As a percentage of net interest income As a percentage of total assets [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] Net interest Overhead - Tax Loan loss After-tax Noninterest Overhead Tax Loan loss After-tax [1] Economy income to noninterest expenditure provisions profit income expenditure provisions profit total assets income Mexico 5.748 34.5 9.6 27.2 28.7 2.341 4.327 0.554 1.562 1.647 Moldova 3.878 16.0 6.4 23.7 53.9 3.609 4.229 0.247 0.919 2.092 Mongolia 2.879 33.8 10.8 20.3 35.1 1.178 2.150 0.312 0.585 1.010 Montenegro 3.150 55.8 0.7 50.1 -6.6 1.936 3.693 0.023 1.579 -0.209 Morocco 2.881 30.5 19.2 14.4 35.9 1.257 2.137 0.553 0.414 1.035 Mozambique 5.926 32.8 10.1 11.4 45.7 3.900 5.846 0.600 0.674 2.706 Myanmar -0.689 143.0 -33.0 0.0 -10.1 1.828 0.843 0.227 0.000 0.069 Namibia 4.387 30.3 19.8 6.0 43.8 2.940 4.271 0.871 0.263 1.922 Nepal 3.836 28.5 17.8 8.1 45.6 1.079 2.172 0.685 0.311 1.748 Netherlands 1.157 47.4 6.1 22.8 23.8 0.337 0.886 0.070 0.264 0.275 New Zealand 1.854 28.0 17.7 10.2 44.2 0.854 1.373 0.328 0.189 0.819 Nicaragua 6.323 36.0 14.0 18.5 31.4 2.601 4.879 0.888 1.169 1.988 Niger 4.362 25.0 16.4 24.6 34.0 3.714 4.804 0.717 1.074 1.482 Nigeria 5.621 48.1 7.0 18.1 26.9 3.069 5.771 0.391 1.015 1.513 Norway 1.526 25.8 17.0 7.7 49.5 0.734 1.128 0.259 0.118 0.756 Oman 3.006 21.0 8.9 9.5 60.6 1.208 1.838 0.268 0.286 1.822 Pakistan 4.036 32.1 17.5 17.9 32.5 1.218 2.515 0.705 0.723 1.312 Panama 3.500 33.6 9.7 16.3 40.3 1.596 2.772 0.341 0.572 1.412 Papua New Guinea 4.454 -19.9 32.5 5.6 81.8 4.690 3.803 1.446 0.250 3.645 Paraguay 5.758 38.3 5.2 12.1 44.5 4.684 6.889 0.298 0.694 2.560 Peru 5.208 27.3 14.9 16.8 40.9 2.387 3.811 0.776 0.876 2.132 Philippines 3.127 41.7 7.6 13.2 37.5 1.575 2.879 0.237 0.414 1.172 Poland 3.074 24.8 10.7 18.6 45.9 1.979 2.741 0.328 0.572 1.412 Portugal 1.595 56.4 3.6 39.7 0.3 0.817 1.717 0.057 0.633 0.004 Qatar 2.463 -8.1 1.2 8.0 99.0 1.159 0.959 0.029 0.196 2.437 Romania 3.765 31.0 5.3 41.8 21.9 2.444 3.611 0.199 1.576 0.823 Russian Federation 4.144 29.4 13.7 22.9 34.0 15.224 16.443 0.568 0.949 1.409 Rwanda 7.299 49.3 12.0 14.4 24.4 4.216 7.811 0.876 1.048 1.779 Saudi Arabia 2.629 -0.4 0.0 13.2 87.1 1.353 1.343 0.000 0.348 2.291 Senegal 4.549 41.4 9.0 18.9 30.6 2.907 4.792 0.408 0.861 1.393 Serbia 4.405 40.8 2.4 35.6 21.2 7.253 9.049 0.104 1.570 0.936 Seychelles 2.946 -13.1 32.9 7.1 73.2 2.438 2.050 0.969 0.209 2.155 28   Table 3: Accounting decomposition of net interest income to total assets, 2005-14, continued As a percentage of net interest income As a percentage of total assets [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] Net interest Overhead - Tax Loan loss After-tax Noninterest Overhead Tax Loan loss After-tax [1] Economy income to noninterest expenditure provisions profit income expenditure provisions profit total assets income Sierra Leone 7.440 45.2 12.9 16.9 25.1 5.978 9.337 0.959 1.255 1.866 Singapore 1.487 2.6 11.7 11.3 74.5 0.840 0.879 0.173 0.168 1.107 Slovak Republic 2.962 40.4 9.4 16.5 33.7 1.180 2.377 0.277 0.489 0.999 Slovenia 1.872 42.6 8.3 69.5 -20.4 1.239 2.037 0.156 1.302 -0.383 South Africa 2.678 18.8 13.5 26.0 41.7 2.438 2.940 0.362 0.696 1.117 Spain 1.523 34.1 4.5 45.4 16.0 0.686 1.206 0.068 0.691 0.244 Sri Lanka 3.909 42.8 20.3 8.8 28.1 1.289 2.962 0.794 0.342 1.100 Sudan 3.394 35.5 6.4 29.8 28.4 4.240 5.444 0.216 1.011 0.963 Suriname 3.895 33.4 18.1 10.3 38.2 1.962 3.263 0.704 0.402 1.488 Swaziland 5.440 22.9 22.8 10.2 44.1 4.875 6.119 1.242 0.555 2.399 Sweden 1.382 33.5 13.6 6.6 46.3 0.822 1.285 0.188 0.091 0.640 Switzerland 0.753 61.0 9.4 13.6 16.0 1.391 1.850 0.071 0.102 0.121 Syrian Arab Republic 1.758 -13.6 24.1 59.6 29.9 1.799 1.560 0.423 1.048 0.526 Tajikistan 4.169 31.6 14.9 32.6 20.8 5.058 6.377 0.621 1.361 0.867 Tanzania 5.451 36.9 14.9 13.0 35.3 2.970 4.982 0.812 0.706 1.921 Thailand 2.870 32.9 11.3 19.5 36.3 0.911 1.856 0.324 0.561 1.041 Togo 2.952 15.3 16.8 23.2 44.7 4.800 5.251 0.496 0.685 1.320 Trinidad and Tobago 4.142 40.0 10.1 9.2 40.7 1.852 3.508 0.417 0.381 1.687 Tunisia 2.620 30.6 7.7 39.9 21.7 1.405 2.208 0.202 1.046 0.568 Turkey 4.097 22.1 13.2 17.5 47.2 1.713 2.618 0.540 0.718 1.934 Uganda 7.265 33.2 12.9 15.6 38.3 3.627 6.037 0.939 1.135 2.781 Ukraine 4.536 39.8 6.3 53.5 0.4 2.802 4.606 0.288 2.427 0.018 United Arab Emirates 2.465 -6.0 0.8 28.6 76.6 1.344 1.196 0.020 0.706 1.887 United Kingdom 0.973 22.4 13.4 40.8 23.4 1.261 1.479 0.131 0.397 0.227 United States 2.796 39.1 12.8 21.4 26.7 1.653 2.747 0.358 0.599 0.746 Uruguay 3.797 55.5 10.9 9.0 24.7 2.205 4.311 0.413 0.341 0.938 Uzbekistan 2.775 21.2 8.5 26.8 43.5 4.013 4.600 0.237 0.744 1.207 Venezuela, RB 6.041 42.3 3.1 11.9 42.6 2.271 4.829 0.190 0.721 2.572 Vietnam 2.854 33.1 9.7 25.8 31.3 0.752 1.698 0.278 0.736 0.894 West Bank and Gaza 3.411 27.7 14.5 4.0 53.8 1.688 2.634 0.493 0.138 1.834 Yemen, Rep. 4.181 31.4 13.5 24.3 30.7 1.248 2.562 0.565 1.016 1.285 Zambia 6.760 41.9 17.9 12.0 28.2 4.652 7.481 1.213 0.811 1.907 29   Table 3: Accounting decomposition of net interest income to total assets, 2005-14, continued As a percentage of net interest income As a percentage of total assets [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] Net interest Overhead - Tax Loan loss After-tax Noninterest Overhead Tax Loan loss After-tax [1] Group income to noninterest expenditure provisions profit income expenditure provisions profit total assets income East Asia & Pacific 3.081 11.0 17.3 17.3 54.5 2.475 2.813 0.532 0.532 1.678 Europe & Central Asia 3.478 32.0 7.6 31.2 29.3 2.868 3.980 0.263 1.084 1.019 Latin America & Caribbean 4.333 36.8 10.6 16.2 36.4 2.342 3.936 0.461 0.703 1.575 Middle East & North Africa 2.640 14.9 12.4 19.9 52.8 1.314 1.707 0.327 0.524 1.395 South Asia 3.601 25.8 18.4 20.4 35.3 1.605 2.533 0.664 0.736 1.273 Sub-Saharan Africa 4.475 28.4 16.0 19.4 36.2 4.319 5.591 0.718 0.867 1.619 OECD high income 1.745 32.4 10.7 28.6 28.4 1.101 1.666 0.186 0.498 0.495 High income 2.031 25.7 10.8 23.7 39.8 1.285 1.808 0.219 0.481 0.807 Upper middle income 4.000 32.1 11.1 20.6 36.2 2.705 3.991 0.443 0.825 1.446 Lower middle income 4.148 30.3 13.4 20.2 36.0 2.493 3.751 0.557 0.838 1.495 Low income 5.034 31.1 15.5 17.9 35.5 4.258 5.826 0.782 0.899 1.785 Note: The component variables, as percentages of total assets, are pooled weighted averages over the 2005-14 period. Net interest income to total assets is reported as the country- level sum of overhead, tax expenditure, loan loss provision, after-tax profit, less noninterest income, all in percentage of total assets. Region and income level aggregates are based on simple averages of relevant countries. 30   Table 4a: Regression results: determinants of net interest margins Net interest margin [1] [2] [3] [4] [5] [6] [7] [8] Size of operations 0.0847*** 0.0853*** 0.0854*** 0.0855*** 0.122*** 0.125*** 0.126*** 0.126*** (12.23) (12.33) (12.33) (12.34) (11.65) (11.94) (11.93) (11.93) … x GDP per capita -0.00260*** -0.00273*** -0.00274*** -0.00274*** (-8.62) (-9.02) (-9.04) (-9.03) Risk aversion 0.0866*** 0.0871*** 0.0871*** 0.0872*** 0.0883*** 0.0899*** 0.0899*** 0.0899*** (15.31) (15.43) (15.45) (15.46) (12.47) (12.73) (12.72) (12.72) … x GDP per capita -0.0000668 -0.000181 -0.000180 -0.000178 (-0.24) (-0.65) (-0.65) (-0.64) Overhead 0.281*** 0.280*** 0.280*** 0.280*** 0.329*** 0.328*** 0.329*** 0.329*** (21.20) (21.06) (20.99) (20.99) (16.82) (16.76) (16.72) (16.71) … x GDP per capita -0.00435*** -0.00430*** -0.00430*** -0.00432*** (-4.53) (-4.48) (-4.47) (-4.47) Reserve opportunity cost 0.224*** 0.223*** 0.223*** 0.224*** 0.268*** 0.271*** 0.271*** 0.271*** (22.03) (22.01) (22.00) (22.00) (20.32) (20.58) (20.58) (20.57) … x GDP per capita -0.00380*** -0.00407*** -0.00410*** -0.00410*** (-7.15) (-7.70) (-7.73) (-7.75) Credit risk 0.0168** 0.0180** 0.0180** 0.0178** -0.00451 -0.00263 -0.00217 -0.00211 (2.82) (3.01) (3.00) (2.97) (-0.56) (-0.33) (-0.27) (-0.26) … x GDP per capita 0.00212*** 0.00200*** 0.00201*** 0.00199*** (5.06) (4.64) (4.75) (4.71) Income diversification -0.360*** -0.359*** -0.359*** -0.359*** -0.445*** -0.443*** -0.443*** -0.443*** (-45.07) (-45.01) (-45.02) (-45.03) (-40.05) (-39.88) (-39.88) (-39.86) … x GDP per capita 0.00622*** 0.00617*** 0.00616*** 0.00617*** (18.12) (17.89) (17.85) (17.83) Inflation 0.0658*** 0.0650*** 0.0649*** 0.0739*** 0.0726*** 0.0725*** (5.90) (5.72) (5.71) (5.84) (5.64) (5.62) … x GDP per capita -0.00155*** -0.00162*** -0.00161*** (-3.92) (-3.69) (-3.66) Interest rate risk 0.0111 0.0114 0.0116 0.0219* 0.0235* 0.0234* (1.42) (1.45) (1.48) (2.30) (2.43) (2.42) … x GDP per capita -0.00158** -0.00178** -0.00174** (-2.61) (-2.90) (-2.80) Competition (Lerner) 0.0248*** 0.0251*** 0.0251*** 0.0403*** 0.0404*** 0.0405*** (4.90) (4.92) (4.93) (4.25) (4.25) (4.25) … x GDP per capita -0.000810*** -0.000821*** -0.000824*** (-3.77) (-3.80) (-3.81) Information environment -0.00404 -0.00273 -0.0161 -0.0167 (-0.22) (-0.15) (-0.70) (-0.72) … x GDP per capita -0.000710 -0.000550 (-0.68) (-0.51) Creditor rights -0.00626 -0.00690 -0.00888 -0.00877 (-0.28) (-0.31) (-0.36) (-0.36) … x GDP per capita -0.000621 -0.000634 (-0.99) (-1.01) Contract enforcement -0.0322 -0.0308 0.0102 0.00905 (-0.89) (-0.85) (0.19) (0.17) … x GDP per capita -0.00303 -0.00289 (-1.46) (-1.40) GDP per capita -0.00735 -0.00327 (-1.42) (-0.65) Constant -0.431*** -0.369*** -0.299*** 0.0542 -0.146*** -0.204*** 0.171 0.310 (-15.13) (-11.89) (-3.57) (0.21) (-4.21) (-5.54) (1.18) (1.13) Time FE Yes Yes Yes Yes Yes Yes Yes Yes Country FE Yes Yes Yes Yes Yes Yes Yes Yes N 114284 114284 114284 114284 114284 114284 114284 114284 adj. R-sq 0.720 0.722 0.722 0.722 0.731 0.734 0.734 0.734 Note: t-statistics in parentheses. * p<0.05, ** p<0.01, *** p<0.001. Variables in deep blue rows are standardized at country-level (see Table 1 country-level statistics). Table 4b: Regression results: alternative specifications Net interest margin [1] [2] [3] [4] [5] [6] [7] [8] Size of operations -0.0667*** -0.0623*** -0.0593*** -0.0526*** 0.0855*** 0.126*** 0.146*** 0.123** (-9.75) (-5.63) (-8.73) (-4.71) (12.34) (11.93) (3.71) (2.79) … x GDP per capita 0.000486 0.000338 -0.00274*** 0.00206 (1.46) (0.99) (-9.03) (1.64) Risk aversion 0.0751*** 0.0614*** 0.0772*** 0.0639*** 0.0872*** 0.0899*** 0.129*** 0.133*** (13.01) (8.39) (13.45) (8.79) (15.46) (12.72) (11.15) (8.62) … x GDP per capita 0.00131*** 0.00130*** -0.000178 -0.000995 (4.45) (4.41) (-0.64) (-1.87) Overhead 0.380*** 0.497*** 0.384*** 0.501*** 0.280*** 0.329*** 0.130*** 0.169*** (26.02) (21.83) (26.03) (21.63) (20.99) (16.71) (11.22) (6.99) … x GDP per capita -0.0116*** -0.0116*** -0.00432*** -0.00347* (-10.63) (-10.59) (-4.47) (-2.40) Reserve opportunity cost 0.234*** 0.276*** 0.244*** 0.287*** 0.224*** 0.271*** 0.164*** 0.189*** (22.71) (20.65) (23.39) (21.38) (22.00) (20.57) (12.31) (10.77) … x GDP per capita -0.00480*** -0.00485*** -0.00410*** -0.00233*** (-8.14) (-8.26) (-7.75) (-4.08) Credit risk -0.00368 -0.00971 -0.00153 -0.00903 0.0178** -0.00211 -0.00859 -0.0153 (-0.58) (-1.15) (-0.24) (-1.08) (2.97) (-0.26) (-1.04) (-1.42) … x GDP per capita 0.00132*** 0.00147*** 0.00199*** 0.000829 (4.05) (4.51) (4.71) (1.83) Income diversification -0.345*** -0.427*** -0.348*** -0.427*** -0.359*** -0.443*** -0.303*** -0.385*** (-46.94) (-38.59) (-47.20) (-38.37) (-45.03) (-39.86) (-23.34) (-21.97) … x GDP per capita 0.00632*** 0.00621*** 0.00617*** 0.00605*** (18.97) (18.43) (17.83) (11.96) Inflation 0.138*** 0.145*** 0.149*** 0.153*** 0.0649*** 0.0725*** 0.0699*** 0.0762*** (14.36) (13.25) (14.52) (13.55) (5.71) (5.62) (8.08) (8.20) … x GDP per capita -0.00104* -0.00100* -0.00161*** -0.00153*** (-2.30) (-2.17) (-3.66) (-5.66) Interest rate risk 0.0525*** 0.0507*** 0.0481*** 0.0492*** 0.0116 0.0234* 0.0172** 0.0254** (6.62) (5.33) (5.88) (5.14) (1.48) (2.42) (2.81) (3.10) … x GDP per capita -0.000324 -0.000797 -0.00174** -0.00129* (-0.54) (-1.28) (-2.80) (-2.43) Competition (Lerner) 0.0213*** 0.0459*** 0.0208*** 0.0469*** 0.0251*** 0.0405*** 0.0219*** 0.0319*** (4.26) (4.60) (4.21) (4.75) (4.93) (4.25) (5.49) (4.28) … x GDP per capita -0.000839*** -0.000900*** -0.000824*** -0.000594*** (-3.61) (-3.90) (-3.81) (-3.59) Information environment -0.124*** -0.111*** -0.113*** -0.0984*** -0.00273 -0.0167 0.0285 0.0245 (-12.38) (-9.30) (-11.30) (-8.29) (-0.15) (-0.72) (1.62) (1.09) … x GDP per capita 0.00172** 0.00160** -0.000550 -0.000423 (3.27) (3.02) (-0.51) (-0.41) Creditor rights 0.0196* 0.0108 0.0218** 0.0145 -0.00690 -0.00877 -0.0552** -0.0538* (2.50) (1.02) (2.66) (1.36) (-0.31) (-0.36) (-2.77) (-2.34) … x GDP per capita 0.000646 0.000485 -0.000634 0.000273 (1.70) (1.28) (-1.01) (0.55) Contract enforcement -0.176*** -0.229*** -0.173*** -0.226*** -0.0308 0.00905 -0.0438 -0.0147 (-13.64) (-12.14) (-13.34) (-11.97) (-0.85) (0.17) (-1.19) (-0.25) … x GDP per capita 0.00441*** 0.00435*** -0.00289 -0.00218 (8.98) (8.80) (-1.40) (-1.05) GDP per capita -0.00923*** -0.0226*** -0.00907*** -0.0225*** -0.00735 -0.00327 -0.00741 -0.00504 (-19.18) (-22.51) (-18.61) (-22.25) (-1.42) (-0.65) (-1.11) (-0.77) Constant 0.278*** 0.301*** 0.385*** 0.407*** 0.0542 0.310 0.361*** 0.307*** (22.30) (22.50) (13.30) (14.12) (0.21) (1.13) (4.06) (3.55) Time FE No No Yes Yes Yes Yes Yes Yes Country FE No No No No Yes Yes - - Bank FE No No No No No No Yes Yes N 114284 114284 114284 114284 114284 114284 114284 114284 adj. R-sq 0.566 0.592 0.570 0.596 0.722 0.734 0.308 0.333 Note: t-statistics in parentheses. * p<0.05, ** p<0.01, *** p<0.001. Variables in deep blue rows are standardized at country-level (see Table 1 country-level statistics). 32   Table 5a: Robustness test: alternative dependent variable Net interest income / TA [1] [2] [3] [4] [5] [6] [7] [8] Size of operations 0.131*** 0.132*** 0.131*** 0.131*** 0.201*** 0.203*** 0.203*** 0.203*** (16.73) (16.80) (16.77) (16.78) (17.08) (17.32) (17.27) (17.28) … x GDP per capita -0.00442*** -0.00454*** -0.00452*** -0.00453*** (-12.29) (-12.58) (-12.50) (-12.54) Risk aversion 0.103*** 0.104*** 0.104*** 0.104*** 0.105*** 0.107*** 0.107*** 0.107*** (16.35) (16.50) (16.54) (16.53) (13.35) (13.62) (13.64) (13.64) … x GDP per capita 0.000225 0.000109 0.000114 0.000111 (0.70) (0.34) (0.36) (0.34) Overhead 0.321*** 0.318*** 0.318*** 0.318*** 0.372*** 0.369*** 0.369*** 0.369*** (18.24) (18.09) (18.00) (18.00) (14.41) (14.28) (14.24) (14.20) … x GDP per capita -0.00441** -0.00433** -0.00433** -0.00430** (-2.86) (-2.81) (-2.80) (-2.76) Reserve opportunity cost 0.0140 0.0126 0.0124 0.0123 0.0192 0.0205 0.0202 0.0198 (1.33) (1.20) (1.18) (1.18) (1.42) (1.52) (1.50) (1.47) … x GDP per capita -0.000384 -0.000619 -0.000603 -0.000590 (-0.78) (-1.27) (-1.23) (-1.20) Credit risk 0.0538*** 0.0553*** 0.0547*** 0.0547*** 0.0356*** 0.0377*** 0.0369*** 0.0367*** (7.18) (7.35) (7.26) (7.27) (3.76) (3.96) (3.89) (3.87) … x GDP per capita 0.00180*** 0.00168*** 0.00176*** 0.00180*** (3.82) (3.49) (3.74) (3.83) Income diversification -0.458*** -0.456*** -0.456*** -0.456*** -0.558*** -0.555*** -0.555*** -0.555*** (-51.09) (-51.02) (-51.01) (-51.01) (-45.45) (-45.20) (-45.14) (-45.09) … x GDP per capita 0.00741*** 0.00734*** 0.00732*** 0.00731*** (18.26) (18.04) (18.03) (17.90) Inflation 0.0747*** 0.0743*** 0.0744*** 0.0806*** 0.0803*** 0.0806*** (6.16) (6.06) (6.06) (5.97) (5.87) (5.90) … x GDP per capita -0.00173*** -0.00198*** -0.00202*** (-3.77) (-4.01) (-4.12) Interest rate risk 0.0265** 0.0265** 0.0265** 0.0376*** 0.0390*** 0.0393*** (3.24) (3.24) (3.23) (3.84) (3.93) (3.96) … x GDP per capita -0.00154* -0.00169* -0.00181** (-2.32) (-2.49) (-2.63) Competition (Lerner) 0.0281*** 0.0294*** 0.0293*** 0.0450*** 0.0468*** 0.0466*** (4.67) (4.84) (4.84) (3.98) (4.11) (4.09) … x GDP per capita -0.000895*** -0.000911*** -0.000901*** (-3.49) (-3.53) (-3.49) Information environment 0.0197 0.0195 0.0136 0.0154 (0.94) (0.93) (0.54) (0.60) … x GDP per capita -0.00161 -0.00208 (-1.37) (-1.69) Creditor rights -0.00658 -0.00651 -0.0181 -0.0183 (-0.27) (-0.27) (-0.67) (-0.67) … x GDP per capita 0.000757 0.000777 (1.06) (1.09) Contract enforcement -0.0831 -0.0833 -0.0885 -0.0847 (-1.95) (-1.96) (-1.43) (-1.37) … x GDP per capita 0.000592 0.000142 (0.26) (0.06) GDP per capita 0.000715 0.00932 (0.12) (1.52) Constant -0.249*** -0.173*** -0.0403 -0.0747 0.160*** 0.106* 0.257 -0.134 (-7.69) (-5.03) (-0.43) (-0.25) (3.90) (2.38) (1.60) (-0.42) Time FE Yes Yes Yes Yes Yes Yes Yes Yes Country FE Yes Yes Yes Yes Yes Yes Yes Yes N 116326 116326 116326 116326 116326 116326 116326 116326 adj. R-sq 0.675 0.677 0.677 0.677 0.689 0.692 0.692 0.692 Note: t-statistics in parentheses. * p<0.05, ** p<0.01, *** p<0.001. Variables in deep blue rows are standardized at country-level (see Table 1 country-level statistics). 33   Table 5b: Robustness test: alternative dependent variable, alternative specifications Net interest income / TA [1] [2] [3] [4] [5] [6] [7] [8] Size of operations -0.0475*** -0.0296* -0.0429*** -0.0241 0.131*** 0.203*** 0.290*** 0.315*** (-5.75) (-2.32) (-5.20) (-1.86) (16.78) (17.28) (7.18) (7.01) … x GDP per capita -0.000632 -0.000700 -0.00453*** -0.00307* (-1.63) (-1.78) (-12.54) (-2.32) Risk aversion 0.0865*** 0.0682*** 0.0873*** 0.0693*** 0.104*** 0.107*** 0.166*** 0.166*** (13.60) (8.56) (13.83) (8.76) (16.53) (13.64) (13.79) (10.86) … x GDP per capita 0.00192*** 0.00192*** 0.000111 -0.000576 (5.73) (5.71) (0.34) (-1.02) Overhead 0.435*** 0.558*** 0.436*** 0.560*** 0.318*** 0.369*** 0.191*** 0.242*** (23.03) (20.34) (22.96) (20.21) (18.00) (14.20) (11.24) (6.63) … x GDP per capita -0.0116*** -0.0116*** -0.00430** -0.00470* (-9.03) (-9.04) (-2.76) (-2.17) Reserve opportunity cost 0.00737 0.00911 0.0124 0.0159 0.0123 0.0198 -0.0117 -0.0189 (0.73) (0.70) (1.21) (1.22) (1.18) (1.47) (-1.01) (-1.30) … x GDP per capita -0.000639 -0.000746 -0.000590 0.000741 (-1.23) (-1.44) (-1.20) (1.46) Credit risk 0.0347*** 0.0317** 0.0357*** 0.0319*** 0.0547*** 0.0367*** 0.0377*** 0.0341** (4.65) (3.28) (4.80) (3.31) (7.27) (3.87) (4.18) (2.88) … x GDP per capita 0.000821* 0.000914* 0.00180*** 0.000460 (2.17) (2.41) (3.83) (0.87) Income diversification -0.431*** -0.524*** -0.433*** -0.523*** -0.456*** -0.555*** -0.389*** -0.487*** (-52.54) (-42.81) (-52.49) (-42.55) (-51.01) (-45.09) (-27.28) (-24.96) … x GDP per capita 0.00693*** 0.00679*** 0.00731*** 0.00734*** (18.90) (18.44) (17.90) (13.58) Inflation 0.149*** 0.155*** 0.167*** 0.167*** 0.0744*** 0.0806*** 0.0723*** 0.0770*** (14.08) (12.38) (14.54) (12.88) (6.06) (5.90) (7.71) (7.53) … x GDP per capita -0.00149** -0.00129* -0.00202*** -0.00189*** (-2.80) (-2.40) (-4.12) (-5.99) Interest rate risk 0.0494*** 0.0475*** 0.0452*** 0.0454*** 0.0265** 0.0393*** 0.0263*** 0.0316*** (5.64) (4.46) (5.01) (4.25) (3.23) (3.96) (4.35) (4.16) … x GDP per capita -0.000149 -0.000463 -0.00181** -0.000658 (-0.22) (-0.68) (-2.63) (-1.18) Competition (Lerner) 0.0211*** 0.0435*** 0.0213*** 0.0449*** 0.0293*** 0.0466*** 0.0243*** 0.0337*** (3.66) (3.87) (3.73) (4.04) (4.84) (4.09) (5.36) (3.86) … x GDP per capita -0.000854** -0.000911*** -0.000901*** -0.000578** (-3.24) (-3.48) (-3.49) (-2.97) Information environment -0.0916*** -0.0876*** -0.0829*** -0.0784*** 0.0195 0.0154 0.0388* 0.0350 (-8.46) (-6.73) (-7.63) (-5.99) (0.93) (0.60) (2.19) (1.57) … x GDP per capita 0.00132* 0.00124* -0.00208 -0.000491 (2.29) (2.13) (-1.69) (-0.45) Creditor rights 0.0360*** 0.0204 0.0429*** 0.0266* -0.00651 -0.0183 -0.0318 -0.0381 (4.17) (1.74) (4.76) (2.24) (-0.27) (-0.67) (-1.59) (-1.64) … x GDP per capita 0.00102* 0.000994* 0.000777 0.000987 (2.37) (2.32) (1.09) (1.81) Contract enforcement -0.180*** -0.222*** -0.179*** -0.219*** -0.0833 -0.0847 -0.116** -0.145* (-12.21) (-10.45) (-12.02) (-10.26) (-1.96) (-1.37) (-3.21) (-2.57) … x GDP per capita 0.00366*** 0.00351*** 0.000142 0.00235 (6.69) (6.38) (0.06) (1.17) GDP per capita -0.0110*** -0.0190*** -0.0107*** -0.0187*** 0.000715 0.00932 0.0000393 0.00941 (-19.61) (-18.66) (-18.90) (-18.26) (0.12) (1.52) (0.01) (1.43) Constant 0.304*** 0.300*** 0.409*** 0.403*** -0.0747 -0.134 0.283*** 0.144 (23.24) (21.60) (12.30) (12.08) (-0.25) (-0.42) (3.38) (1.74) Time FE No No Yes Yes Yes Yes Yes Yes Country FE No No No No Yes Yes - - Bank FE No No No No No No Yes Yes N 116326 116326 116326 116326 116326 116326 116326 116326 adj. R-sq 0.513 0.535 0.515 0.537 0.677 0.692 0.338 0.364 Note: t-statistics in parentheses. * p<0.05, ** p<0.01, *** p<0.001. Variables in deep blue rows are standardized at country-level (see Table 1 country-level statistics). 34   Table 6a: Contributing factors to net interest margin, 2005-14 [1] [3] [6] [8] [12] [14] [4] [5] [7] [9] [10] [11] [13] [15] Net interest margin [2] Economy Size of Reserve Income Information Contract Risk aversion Overhead Credit risk Inflation Interest rate risk Competition Creditor rights Other factors (v.s. world) operations opportunity cost diversification environment enforcement 4.01 Afghanistan -0.248 -0.021 0.061 1.119 -0.025 0.659 0.096 0.038 0.007 0.009 -0.004 0.039 2.279 -0.76 Albania -0.110 -0.014 -0.318 0.246 0.016 0.793 -0.121 -0.012 -0.007 -0.002 -0.021 0.001 -1.209 0.98 Algeria 0.031 0.375 -0.230 0.785 -0.007 -0.685 -0.019 -0.004 0.107 0.009 0.021 -0.031 0.630 1.64 Angola -0.054 0.039 0.169 0.584 0.011 -0.839 0.318 0.049 0.076 0.009 0.020 0.112 1.150 1.14 Argentina -0.047 0.130 0.921 0.256 -0.012 -1.629 0.255 0.030 0.009 -0.007 0.014 0.027 1.199 2.51 Armenia -0.243 0.466 0.112 0.472 -0.032 0.305 -0.003 0.015 0.013 -0.005 -0.001 0.010 1.396 -2.52 Australia 0.279 -0.157 -0.436 -0.399 -0.054 0.238 -0.111 -0.012 -0.071 -0.005 -0.023 -0.113 -1.652 -3.14 Austria 0.233 -0.203 -0.489 -0.304 -0.019 0.076 -0.150 -0.020 0.005 -0.005 -0.004 -0.104 -2.156 3.04 Azerbaijan -0.141 0.198 0.159 0.291 0.033 0.194 0.045 -0.014 0.052 -0.001 0.009 0.002 2.212 -1.26 Bahamas, The -0.065 0.190 -0.352 0.191 -0.029 0.237 -0.141 -0.020 0.044 0.009 -0.018 -0.065 -1.236 -1.90 Bahrain 0.196 0.157 -0.534 -0.254 0.011 0.244 -0.132 -0.015 0.019 -0.001 0.022 -0.001 -1.610 -0.86 Bangladesh 0.041 -0.112 -0.369 -0.178 -0.002 -0.555 0.108 0.068 -0.022 0.009 0.000 0.042 0.105 3.04 Belarus -0.066 0.295 0.893 0.305 0.006 -0.876 0.312 0.047 -0.011 -0.001 0.022 0.061 2.052 -3.78 Belgium 0.472 -0.339 -0.695 -0.496 -0.047 1.343 -0.147 -0.017 -0.080 -0.002 0.007 -0.135 -3.642 3.19 Belize -0.206 0.075 0.192 0.709 -0.033 0.152 -0.172 -0.025 0.006 0.009 -0.006 -0.035 2.524 -0.35 Benin -0.092 -0.174 0.041 -0.098 0.054 -0.611 -0.070 -0.018 -0.007 0.009 0.011 0.060 0.547 0.52 Bolivia -0.094 0.178 0.300 -0.106 0.014 -0.428 0.046 0.024 0.012 -0.007 0.038 0.004 0.541 0.30 Bosnia and Herzegovina -0.113 0.196 0.003 0.690 0.038 -0.039 -0.125 -0.023 -0.012 -0.003 0.003 0.013 -0.327 0.48 Botswana -0.127 -0.220 -0.140 -0.221 -0.019 -0.143 0.170 -0.008 -0.040 -0.002 -0.003 -0.050 1.280 1.15 Brazil 0.050 0.149 0.105 -0.502 -0.003 0.634 0.007 0.012 -0.015 -0.005 0.020 0.067 0.628 -0.45 Bulgaria 0.039 -0.009 -0.238 0.199 0.011 0.146 -0.036 -0.019 0.021 -0.003 -0.021 0.021 -0.559 0.06 Burkina Faso -0.044 -0.192 0.231 0.061 0.009 -0.888 -0.108 -0.015 0.019 0.009 0.006 0.062 0.913 4.20 Burundi -0.334 0.056 1.342 0.106 0.061 -0.652 0.233 0.025 0.051 0.009 0.019 0.091 3.197 0.41 Cambodia -0.242 0.587 -0.162 0.688 -0.007 0.308 -0.027 -0.027 0.037 0.004 -0.011 0.080 -0.818 -0.64 Cameroon -0.072 -0.233 0.265 -0.085 0.048 -1.063 -0.120 -0.024 0.035 0.008 0.008 0.073 0.519 -2.67 Canada 0.150 -0.173 -0.392 -0.393 -0.050 0.177 -0.164 -0.021 -0.192 -0.007 -0.011 -0.140 -1.452 -0.80 Chile 0.151 0.147 -0.245 0.010 -0.030 -0.100 -0.088 0.001 -0.032 -0.004 0.008 0.028 -0.645 -1.66 China 0.265 -0.154 -0.656 0.069 -0.024 1.151 -0.107 -0.027 0.051 -0.003 0.008 0.012 -2.246 1.71 Colombia 0.149 0.108 0.368 -0.200 0.002 -0.395 -0.063 -0.004 0.048 -0.005 0.002 -0.058 1.763 3.61 Congo, Dem. Rep. -0.238 0.064 1.376 0.777 0.008 -1.259 0.136 0.068 -0.055 0.009 0.009 0.112 2.605 2.00 Costa Rica -0.118 0.234 0.114 -0.242 -0.035 1.009 0.113 -0.006 -0.019 -0.005 0.020 0.048 0.890 -1.07 Croatia -0.045 0.016 -0.151 0.257 0.043 0.293 -0.122 0.024 0.009 -0.001 -0.005 0.028 -1.420 -1.74 Cyprus 0.156 -0.186 0.304 -0.210 0.060 0.155 -0.165 -0.020 0.018 0.004 -0.017 -0.088 -1.750 -2.23 Czech Republic 0.207 -0.084 -0.468 -0.320 -0.016 0.406 -0.139 -0.021 0.054 -0.005 -0.003 -0.003 -1.837 0.02 Côte d'Ivoire -0.034 -0.221 0.231 -0.273 0.040 -0.962 -0.126 -0.020 -0.008 0.009 0.009 0.022 1.351 -1.31 Denmark -0.049 -0.007 -0.150 -0.289 0.020 0.414 -0.156 -0.018 0.012 -0.002 -0.020 -0.132 -0.933 4.11 Dominican Republic -0.255 0.103 0.828 0.410 0.005 0.733 0.011 0.022 -0.067 -0.007 0.021 0.088 2.221 1.17 Ecuador -0.124 0.013 0.664 -0.228 0.014 -0.235 -0.054 -0.017 -0.020 -0.006 0.026 0.065 1.071 -1.32 Egypt, Arab Rep. 0.148 -0.057 -0.458 -0.106 0.056 0.347 0.227 -0.010 -0.097 -0.006 0.020 0.049 -1.436 1.29 El Salvador -0.051 0.151 -0.093 0.381 -0.007 1.239 0.017 0.006 0.031 -0.007 0.004 0.026 -0.402 Table 6a: Contributing factors to net interest margin, 2005-14, continued [1] [3] [6] [8] [12] [14] [4] [5] [7] [9] [10] [11] [13] [15] Net interest margin [2] Economy Size of Reserve Income Information Contract Risk aversion Overhead Credit risk Inflation Interest rate risk Competition Creditor rights Other factors (v.s. world) operations opportunity cost diversification environment enforcement -1.98 Estonia -0.081 -0.025 -0.480 -0.227 -0.021 -0.036 -0.026 -0.006 -0.012 -0.005 -0.005 0.004 -1.066 -0.44 Ethiopia -0.148 -0.055 -0.395 -0.160 0.025 -0.824 0.363 -0.023 0.117 0.009 0.011 0.019 0.622 -3.35 Finland 0.271 -0.228 -0.592 -0.396 -0.051 -0.322 -0.168 -0.013 -0.116 -0.002 -0.017 -0.138 -1.581 -2.71 France 0.273 -0.131 -0.368 -0.460 -0.014 -0.351 -0.177 -0.020 -0.051 -0.002 0.007 -0.029 -1.389 3.75 Gambia, The -0.431 -0.043 0.976 0.144 0.033 -0.954 -0.036 0.027 -0.026 0.009 0.004 0.034 4.014 4.26 Georgia -0.160 0.315 0.566 0.351 0.034 0.225 0.005 0.023 0.014 -0.003 -0.010 0.025 2.874 -2.34 Germany -0.041 -0.162 -0.403 -0.452 -0.036 0.572 -0.171 -0.026 -0.312 -0.007 -0.002 -0.119 -1.183 4.40 Ghana -0.183 0.021 0.708 0.168 0.034 -0.276 0.282 0.062 0.049 0.004 -0.012 0.042 3.505 -2.17 Greece 0.337 -0.195 -0.458 -0.386 -0.004 0.596 -0.096 -0.016 -0.023 -0.003 0.011 -0.014 -1.921 1.80 Guatemala -0.087 -0.072 0.081 0.258 -0.018 1.020 0.030 -0.026 0.020 -0.006 -0.008 0.033 0.578 4.16 Haiti -0.164 -0.217 0.458 1.665 -0.023 -0.379 0.126 0.008 -0.037 0.009 0.020 0.112 2.583 2.81 Honduras -0.086 -0.011 0.469 -0.141 -0.013 0.596 0.073 -0.005 -0.014 -0.007 -0.008 0.057 1.896 -2.66 Hong Kong SAR, China 0.092 0.281 -0.627 -0.026 -0.048 0.285 -0.116 -0.013 -0.029 -0.005 -0.027 -0.116 -2.315 -1.20 Hungary 0.274 -0.117 -0.162 -0.347 0.049 -0.364 -0.070 -0.001 -0.037 -0.003 -0.009 0.003 -0.413 -1.43 Iceland -0.072 0.143 -0.237 -0.235 0.011 -1.126 0.053 0.024 -0.014 -0.005 -0.010 -0.114 0.154 -1.55 India 0.280 -0.178 -0.483 -0.210 -0.032 0.393 0.143 0.048 -0.001 -0.004 -0.010 0.042 -1.541 0.64 Indonesia 0.066 0.033 -0.173 -0.186 -0.027 0.852 0.077 0.004 0.047 -0.003 0.007 0.035 -0.089 -4.22 Ireland 0.325 -0.318 -0.891 -0.494 -0.055 1.489 -0.096 -0.014 -0.014 -0.005 -0.023 -0.135 -3.986 -2.35 Israel 0.345 -0.249 -0.358 0.197 -0.032 -0.168 -0.146 -0.015 -0.027 -0.005 -0.018 -0.042 -1.829 -2.06 Italy -0.019 -0.064 -0.370 -0.491 -0.005 0.440 -0.157 -0.003 -0.094 -0.005 0.020 -0.059 -1.253 2.48 Jamaica -0.065 0.022 0.303 -0.140 -0.032 0.664 0.279 0.045 0.041 0.009 -0.017 -0.070 1.444 -3.25 Japan 0.133 -0.287 -0.647 -0.419 -0.028 1.699 -0.231 -0.028 0.048 -0.005 0.002 -0.150 -3.335 -0.85 Jordan 0.109 0.067 -0.415 0.137 0.016 0.263 -0.012 -0.012 0.061 0.009 0.027 0.035 -1.132 2.40 Kenya -0.153 0.193 0.285 -0.144 -0.001 0.046 0.193 0.001 0.057 0.009 -0.023 0.028 1.906 -1.63 Korea, Rep. 0.332 -0.257 -0.514 -0.303 -0.027 0.960 -0.108 -0.018 0.027 -0.006 -0.003 -0.118 -1.598 -1.77 Kuwait 0.404 0.011 -0.686 -0.078 0.020 0.154 -0.030 -0.009 0.116 -0.004 0.019 0.026 -1.710 3.75 Kyrgyz Republic -0.356 0.200 0.360 0.144 0.008 -0.066 0.140 0.033 0.085 0.004 -0.019 0.017 3.202 -2.10 Latvia -0.032 -0.091 -0.309 -0.050 0.018 -0.724 -0.018 0.007 0.001 0.000 -0.027 0.004 -0.877 -2.24 Lebanon 0.025 -0.152 -0.538 0.467 0.032 0.253 -0.095 -0.027 -0.085 -0.002 0.020 0.025 -2.158 -2.50 Lithuania 0.080 -0.119 -0.358 -0.014 0.007 -0.464 -0.072 -0.018 -0.014 -0.007 0.003 -0.025 -1.497 -3.08 Luxembourg 0.102 -0.077 -0.562 -0.396 -0.037 -0.223 -0.151 -0.020 -0.001 0.009 0.014 -0.010 -1.727 0.61 Macedonia, FYR -0.164 0.217 0.007 0.667 0.070 0.227 -0.122 -0.019 0.008 -0.002 -0.003 -0.006 -0.275 2.41 Madagascar -0.216 0.009 0.029 0.237 0.012 0.366 0.172 -0.009 0.009 0.009 0.026 0.065 1.704 5.38 Malawi -0.349 0.142 1.320 -0.172 -0.022 -1.074 0.201 -0.004 0.011 0.009 -0.010 0.070 5.255 -1.84 Malaysia 0.181 -0.051 -0.657 0.847 -0.007 0.023 -0.119 -0.022 -0.062 -0.006 -0.030 0.001 -1.942 0.44 Mali -0.166 -0.070 0.411 -0.121 0.023 -1.066 -0.092 -0.018 0.000 0.009 0.009 0.052 1.468 1.02 Mauritania -0.271 0.552 0.111 0.283 0.087 -0.797 0.039 -0.016 0.092 0.009 0.020 0.112 0.800 -1.77 Mauritius -0.032 -0.084 -0.509 0.044 -0.017 0.170 0.007 0.008 0.061 -0.001 -0.002 -0.056 -1.361 -0.17 Mexico -0.043 0.050 0.162 -0.320 -0.010 -0.148 -0.069 -0.022 0.108 -0.007 -0.002 -0.079 0.207 1.78 Moldova -0.250 0.441 0.441 0.364 0.013 -0.563 0.135 0.056 0.011 0.009 -0.016 0.028 1.116 36   Table 6a: Contributing factors to net interest margin, 2005-14, continued [1] [3] [6] [8] [12] [14] [4] [5] [7] [9] [10] [11] [13] [15] Net interest margin [2] Economy Size of Reserve Income Information Contract Risk aversion Overhead Credit risk Inflation Interest rate risk Competition Creditor rights Other factors (v.s. world) operations opportunity cost diversification environment enforcement 1.10 Mongolia -0.240 0.073 -0.136 0.062 0.001 0.633 0.183 0.058 0.117 0.000 -0.003 0.066 0.285 0.45 Montenegro -0.125 0.138 0.120 0.366 0.021 -0.042 -0.067 -0.019 -0.145 0.000 -0.030 -0.014 0.244 -1.47 Morocco 0.364 -0.155 -0.407 -0.276 0.006 0.141 -0.176 -0.024 0.016 -0.002 0.020 0.027 -1.007 3.31 Mozambique -0.258 0.066 0.744 0.254 0.001 -0.638 0.173 0.047 -0.011 0.009 0.018 0.041 2.863 -0.84 Nepal -0.141 -0.096 -0.504 -0.077 -0.023 0.649 0.169 -0.028 -0.057 0.009 -0.011 -0.011 -0.714 -3.40 Netherlands 0.217 -0.125 -0.572 -0.206 -0.027 -0.043 -0.161 -0.022 -0.106 -0.005 0.005 -0.135 -2.223 -2.34 New Zealand 0.274 -0.147 -0.397 -0.416 -0.049 0.185 -0.094 -0.013 -0.044 -0.005 -0.030 -0.108 -1.496 1.12 Niger -0.221 -0.175 0.247 0.155 -0.010 -0.494 -0.103 -0.017 0.009 0.009 0.011 0.071 1.641 2.66 Nigeria 0.111 0.191 0.326 0.012 0.033 -0.237 0.219 0.003 -0.028 0.007 -0.015 0.034 2.002 -2.48 Norway -0.020 -0.126 -0.553 -0.375 -0.048 0.985 -0.145 -0.004 0.050 -0.002 -0.003 -0.143 -2.097 -1.12 Oman 0.215 0.044 -0.477 -0.179 0.017 0.331 -0.014 -0.013 0.082 0.001 0.011 0.013 -1.147 -0.67 Pakistan 0.077 -0.075 -0.289 -0.130 0.050 0.327 0.266 0.024 -0.071 0.006 -0.003 0.001 -0.851 -1.36 Panama -0.022 -0.065 -0.322 -0.347 -0.037 0.294 -0.045 -0.017 0.038 -0.007 0.005 0.036 -0.870 3.26 Paraguay -0.138 -0.052 1.136 1.107 -0.017 -0.547 0.039 0.025 -0.050 -0.007 0.020 0.065 1.681 1.75 Peru 0.138 -0.074 0.230 -0.231 -0.003 0.341 -0.112 -0.008 0.036 -0.007 -0.009 0.036 1.411 -0.38 Philippines 0.093 -0.035 -0.138 -0.207 0.021 0.119 -0.041 -0.014 -0.028 0.002 0.012 0.094 -0.262 -1.48 Poland 0.245 -0.078 -0.250 -0.310 0.009 -0.161 -0.123 -0.013 0.019 -0.007 -0.017 -0.002 -0.792 -2.35 Portugal -0.061 -0.045 -0.386 -0.446 0.027 0.192 -0.166 -0.022 -0.022 -0.005 0.021 -0.093 -1.344 -1.98 Qatar 0.350 0.063 -0.661 -0.238 -0.027 0.125 -0.035 -0.012 0.106 0.003 0.020 0.031 -1.711 0.31 Romania 0.083 -0.006 0.147 0.521 0.034 -0.179 0.008 0.026 -0.016 -0.004 -0.022 0.037 -0.318 1.54 Russian Federation -0.275 0.268 3.646 -0.219 0.041 -1.987 0.170 0.028 -0.057 -0.002 0.007 -0.006 -0.075 4.13 Rwanda -0.231 0.130 0.903 0.024 0.038 -0.093 0.088 0.009 0.001 0.002 -0.008 0.098 3.168 -2.10 Saudi Arabia 0.510 0.052 -0.644 -0.088 -0.019 0.169 -0.050 -0.027 0.117 -0.007 0.011 0.112 -2.234 -1.11 Senegal -0.111 -0.114 0.278 -0.156 0.006 -1.457 -0.181 -0.018 0.023 0.009 0.004 0.047 0.560 6.03 Sierra Leone -0.433 0.095 1.330 -0.018 0.064 -0.532 0.221 -0.022 -0.028 0.009 -0.005 0.088 5.257 -2.77 Singapore 0.176 0.050 -0.600 -0.011 -0.019 -0.889 -0.115 -0.018 0.117 -0.001 -0.029 -0.142 -1.291 -1.72 Slovak Republic 0.164 -0.126 -0.376 -0.225 -0.004 0.381 -0.121 -0.018 0.003 -0.002 -0.014 -0.028 -1.359 -2.50 Slovenia 0.166 -0.157 -0.471 -0.317 0.039 0.195 -0.140 -0.019 -0.029 0.005 0.012 -0.024 -1.755 0.36 South Africa -0.010 -0.041 0.207 0.183 -0.033 -0.364 0.033 -0.010 -0.053 -0.007 -0.007 0.016 0.448 -2.87 Spain 0.202 -0.160 -0.541 -0.462 -0.001 0.375 -0.151 -0.020 0.023 -0.005 0.000 -0.092 -2.035 0.42 Sri Lanka 0.001 -0.176 -0.057 -0.095 -0.008 0.483 0.204 0.045 -0.029 0.000 0.013 -0.007 0.044 2.08 Swaziland -0.169 0.056 0.631 -0.030 0.009 -0.707 0.087 0.001 -0.024 -0.002 -0.003 0.010 2.224 -2.04 Sweden -0.093 0.096 -0.358 -0.515 -0.044 0.306 -0.190 -0.015 0.026 -0.002 -0.010 -0.101 -1.137 -3.51 Switzerland 0.000 -0.271 -0.614 -0.320 -0.052 0.691 -0.250 -0.021 -0.061 -0.004 -0.007 -0.020 -2.584 1.96 Tanzania -0.245 -0.041 0.497 0.189 -0.011 -0.311 0.190 0.036 0.024 0.009 -0.007 0.051 1.583 -1.81 Thailand 0.316 0.122 -0.484 -0.447 -0.004 0.755 -0.106 -0.014 0.062 -0.004 0.007 -0.038 -1.978 -0.61 Togo -0.168 -0.160 0.354 -0.326 0.006 -1.458 -0.077 -0.017 -0.021 0.009 0.009 0.035 1.201 -0.17 Trinidad and Tobago 0.011 0.023 -0.215 0.237 -0.037 0.112 0.164 -0.011 0.040 -0.001 -0.023 0.041 -0.504 -1.71 Tunisia 0.149 -0.143 -0.403 -0.326 0.077 0.022 -0.051 -0.020 0.081 0.001 0.014 -0.035 -1.071 0.50 Turkey 0.279 0.027 -0.136 -0.118 -0.018 0.547 0.141 0.024 -0.019 -0.004 0.008 0.053 -0.287 37   Table 6a: Contributing factors to net interest margin, 2005-14, continued [1] [3] [6] [8] [12] [14] [4] [5] [7] [9] [10] [11] [13] [15] Net interest margin [2] Economy Size of Reserve Income Information Contract Risk aversion Overhead Credit risk Inflation Interest rate risk Competition Creditor rights Other factors (v.s. world) operations opportunity cost diversification environment enforcement 4.39 Uganda -0.241 0.196 0.860 0.111 -0.018 -0.030 0.166 0.048 0.018 0.009 -0.006 0.001 3.277 1.51 Ukraine 0.027 0.084 0.163 0.118 0.051 -0.004 0.262 0.075 -0.009 0.004 -0.023 0.088 0.674 -2.68 United Kingdom 0.017 0.037 -0.390 -0.258 -0.016 -0.167 -0.120 -0.020 0.017 -0.007 -0.023 -0.134 -1.615 -1.01 United States -0.207 -0.058 -0.201 -0.166 -0.041 1.242 -0.137 -0.017 0.001 -0.007 -0.023 -0.111 -1.281 -0.67 Uruguay -0.056 -0.009 0.398 0.303 0.017 -0.747 0.114 0.039 -0.027 -0.007 0.012 -0.009 -0.698 3.70 Venezuela, RB 0.056 -0.054 0.302 0.454 -0.002 0.198 0.398 0.070 0.015 0.004 0.025 0.095 2.143 -1.38 Vietnam 0.058 0.026 -0.490 -0.277 -0.041 0.722 0.160 0.012 -0.023 -0.001 -0.007 0.063 -1.579 3.38 Zambia -0.248 0.112 0.952 0.410 0.005 -0.690 0.157 0.025 -0.074 0.001 -0.017 0.023 2.720 Note: Values of contribution in percentage points of net interest margin. results based on equation in Table 4a, column 4. The dashed line for each column indicates the cross-country mean for a variable, as listed in Table 1 and normalized to 0 in the regression analysis. The bars indicate relative distances from world averages, with the left and right sides of the cell representing minimum and maximum values across countries. Column “Other factors” reflects contributions associated to real incomes, country and time fixed effects, and regression residuals. Table 6b: Contributing factors to net interest margin, 2005-14 [1] [3] [6] [8] [12] [14] [4] [5] [7] [9] [10] [11] [13] [15] Net interest margin [2] Region Size of Reserve Income Information Contract Risk aversion Overhead Credit risk Inflation Interest rate risk Competition Creditor rights Other factors (v.s. world) operations opportunity cost diversification environment enforcement -1.04 East Asia & Pacific 0.076 0.093 -0.412 0.051 -0.016 0.396 -0.021 -0.006 0.029 -0.002 -0.007 0.006 -1.223 0.95 Europe & Central Asia -0.077 0.140 0.318 0.246 0.024 -0.074 0.021 0.012 -0.003 -0.001 -0.008 0.014 0.336 1.55 Latin America & Caribbean -0.058 0.041 0.288 0.167 -0.012 0.129 0.048 0.006 0.009 -0.003 0.006 0.023 0.908 -1.41 Middle East & North Africa 0.227 0.024 -0.496 -0.014 0.017 0.124 -0.035 -0.016 0.047 0.000 0.019 0.023 -1.326 0.08 South Asia 0.002 -0.110 -0.274 0.071 -0.007 0.326 0.164 0.033 -0.029 0.005 -0.002 0.018 -0.113 1.87 Sub-Saharan Africa -0.178 0.003 0.475 0.067 0.016 -0.615 0.087 0.006 0.009 0.006 0.003 0.047 1.945 -2.38 OECD high income 0.136 -0.122 -0.434 -0.323 -0.019 0.278 -0.131 -0.014 -0.032 -0.004 -0.006 -0.076 -1.638 Note: See note for Table 6a. Contribution for each group is represented by the average of contributions for countries within each group (see Table 7a-g for country groupings). Left and right sides of cells represent minimum and maximum values across regions. 38   Table 7a: Contributing factors to net interest margin, East Asia & Pacific Comparisons, 2005-14 [1] [3] [6] [8] [12] [14] [4] [5] [7] [9] [10] [11] [13] [15] Net interest margin [2] Economy Size of Reserve Income Information Contract Risk aversion Overhead Credit risk Inflation Interest rate risk Competition Creditor rights Other factors (v.s. world) operations opportunity cost diversification environment enforcement 1.45 Cambodia -0.318 0.494 0.250 0.637 0.008 -0.087 -0.006 -0.020 0.008 0.006 -0.004 0.074 0.405 -0.62 China 0.188 -0.247 -0.243 0.018 -0.009 0.755 -0.086 -0.021 0.022 -0.001 0.015 0.007 -1.022 -1.63 Hong Kong SAR, China 0.016 0.188 -0.215 -0.077 -0.033 -0.111 -0.095 -0.007 -0.058 -0.003 -0.020 -0.122 -1.092 1.68 Indonesia -0.011 -0.060 0.239 -0.237 -0.012 0.456 0.098 0.010 0.018 -0.001 0.014 0.030 1.135 -0.81 Malaysia 0.105 -0.144 -0.244 0.796 0.009 -0.373 -0.097 -0.016 -0.091 -0.004 -0.023 -0.004 -0.719 2.13 Mongolia -0.316 -0.020 0.276 0.011 0.016 0.237 0.204 0.064 0.088 0.002 0.004 0.060 1.508 0.65 Philippines 0.016 -0.128 0.274 -0.258 0.037 -0.277 -0.019 -0.008 -0.057 0.004 0.019 0.088 0.961 -1.74 Singapore 0.099 -0.043 -0.187 -0.063 -0.004 -1.285 -0.094 -0.012 0.088 0.000 -0.021 -0.147 -0.067 -0.78 Thailand 0.239 0.028 -0.072 -0.498 0.012 0.359 -0.085 -0.008 0.034 -0.003 0.015 -0.044 -0.754 -0.34 Vietnam -0.018 -0.067 -0.078 -0.328 -0.025 0.326 0.181 0.018 -0.052 0.000 0.000 0.058 -0.356 Note: See note for Table 6a. The bars indicate relative distances from group averages (see Table 6b), with the left and right sides of the cell representing minimum and maximum values across countries. Table 7b: Contributing factors to net interest margin, Europe & Central Asia Comparison, 2005-14 [1] [3] [6] [8] [12] [14] [4] [5] [7] [9] [10] [11] [13] [15] Net interest margin [2] Economy Size of Reserve Income Information Contract Risk aversion Overhead Credit risk Inflation Interest rate risk Competition Creditor rights Other factors (v.s. world) operations opportunity cost diversification environment enforcement -1.71 Albania -0.033 -0.155 -0.636 0.000 -0.008 0.867 -0.143 -0.024 -0.004 -0.001 -0.013 -0.013 -1.546 1.56 Armenia -0.166 0.325 -0.206 0.226 -0.056 0.379 -0.024 0.003 0.017 -0.004 0.007 -0.004 1.060 2.09 Azerbaijan -0.064 0.058 -0.159 0.045 0.009 0.268 0.023 -0.026 0.055 0.000 0.017 -0.012 1.876 2.09 Belarus 0.011 0.154 0.575 0.059 -0.018 -0.801 0.291 0.036 -0.008 0.000 0.030 0.047 1.715 -0.65 Bosnia and Herzegovina -0.036 0.056 -0.315 0.444 0.014 0.036 -0.146 -0.034 -0.009 -0.002 0.010 -0.001 -0.663 -1.40 Bulgaria 0.116 -0.150 -0.556 -0.047 -0.013 0.220 -0.057 -0.030 0.024 -0.002 -0.014 0.007 -0.895 -2.02 Croatia 0.032 -0.125 -0.469 0.011 0.019 0.367 -0.143 0.012 0.013 0.000 0.002 0.015 -1.756 -2.69 Cyprus 0.233 -0.326 -0.013 -0.456 0.036 0.229 -0.186 -0.031 0.021 0.005 -0.009 -0.101 -2.086 3.31 Georgia -0.084 0.175 0.248 0.105 0.010 0.299 -0.016 0.012 0.017 -0.002 -0.002 0.011 2.537 2.80 Kyrgyz Republic -0.279 0.060 0.042 -0.102 -0.017 0.008 0.118 0.021 0.089 0.005 -0.011 0.004 2.866 -3.45 Lithuania 0.157 -0.259 -0.676 -0.260 -0.017 -0.390 -0.094 -0.030 -0.011 -0.006 0.010 -0.038 -1.833 -0.34 Macedonia, FYR -0.087 0.077 -0.310 0.421 0.046 0.301 -0.144 -0.030 0.011 -0.001 0.005 -0.019 -0.611 0.84 Moldova -0.173 0.300 0.123 0.118 -0.012 -0.489 0.114 0.045 0.014 0.010 -0.009 0.015 0.779 -0.50 Montenegro -0.048 -0.002 -0.198 0.120 -0.003 0.032 -0.089 -0.030 -0.141 0.001 -0.023 -0.028 -0.092 -0.64 Romania 0.160 -0.146 -0.171 0.274 0.009 -0.105 -0.013 0.015 -0.013 -0.003 -0.014 0.023 -0.655 0.59 Russian Federation -0.198 0.127 3.329 -0.465 0.017 -1.913 0.148 0.017 -0.054 -0.001 0.014 -0.020 -0.411 -0.45 Turkey 0.356 -0.113 -0.454 -0.364 -0.042 0.621 0.120 0.012 -0.015 -0.003 0.015 0.039 -0.623 0.56 Ukraine 0.104 -0.056 -0.155 -0.128 0.027 0.070 0.241 0.063 -0.006 0.005 -0.015 0.074 0.338 Note: See note for Table 7a. 39   Table 7c: Contributing factors to net interest margin, Latin America & Caribbean Comparison, 2005-14 [1] [3] [6] [8] [12] [14] [4] [5] [7] [9] [10] [11] [13] [15] Net interest margin [2] Economy Size of Reserve Income Information Contract Risk aversion Overhead Credit risk Inflation Interest rate risk Competition Creditor rights Other factors (v.s. world) operations opportunity cost diversification environment enforcement -0.41 Argentina 0.011 0.088 0.633 0.089 0.000 -1.759 0.207 0.024 0.000 -0.004 0.008 0.004 0.291 -2.81 Bahamas, The -0.007 0.149 -0.640 0.024 -0.017 0.107 -0.190 -0.026 0.035 0.012 -0.025 -0.088 -2.144 1.64 Belize -0.148 0.034 -0.096 0.542 -0.021 0.023 -0.220 -0.031 -0.003 0.012 -0.013 -0.058 1.616 -1.03 Bolivia -0.036 0.137 0.012 -0.274 0.026 -0.557 -0.002 0.018 0.002 -0.004 0.032 -0.019 -0.367 -0.41 Brazil 0.108 0.108 -0.183 -0.669 0.010 0.504 -0.042 0.006 -0.025 -0.002 0.014 0.044 -0.280 0.16 Colombia 0.207 0.066 0.080 -0.367 0.015 -0.525 -0.111 -0.010 0.038 -0.002 -0.005 -0.081 0.855 0.45 Costa Rica -0.060 0.192 -0.174 -0.409 -0.023 0.879 0.064 -0.011 -0.028 -0.002 0.013 0.025 -0.018 2.56 Dominican Republic -0.197 0.061 0.540 0.242 0.018 0.604 -0.038 0.016 -0.077 -0.004 0.015 0.065 1.313 -0.39 Ecuador -0.066 -0.029 0.376 -0.395 0.027 -0.365 -0.102 -0.022 -0.030 -0.003 0.019 0.042 0.163 -0.26 El Salvador 0.007 0.110 -0.381 0.213 0.005 1.109 -0.032 0.000 0.021 -0.004 -0.002 0.003 -1.310 0.25 Guatemala -0.029 -0.114 -0.207 0.091 -0.006 0.891 -0.019 -0.032 0.011 -0.003 -0.014 0.011 -0.330 2.61 Haiti -0.106 -0.258 0.169 1.498 -0.010 -0.509 0.078 0.002 -0.047 0.012 0.013 0.090 1.675 1.25 Honduras -0.028 -0.052 0.181 -0.309 0.000 0.466 0.025 -0.010 -0.023 -0.004 -0.014 0.034 0.988 0.93 Jamaica -0.007 -0.019 0.015 -0.307 -0.020 0.535 0.230 0.039 0.031 0.012 -0.023 -0.093 0.535 -1.73 Mexico 0.015 0.008 -0.126 -0.487 0.002 -0.277 -0.117 -0.028 0.099 -0.004 -0.009 -0.102 -0.701 -2.91 Panama 0.036 -0.106 -0.610 -0.515 -0.024 0.164 -0.094 -0.022 0.029 -0.004 -0.002 0.013 -1.778 1.71 Paraguay -0.080 -0.093 0.848 0.940 -0.004 -0.676 -0.009 0.019 -0.060 -0.004 0.013 0.042 0.773 0.19 Peru 0.196 -0.116 -0.058 -0.398 0.009 0.211 -0.160 -0.014 0.026 -0.004 -0.016 0.013 0.503 -1.72 Trinidad and Tobago 0.069 -0.019 -0.504 0.069 -0.025 -0.018 0.116 -0.017 0.030 0.001 -0.030 0.018 -1.412 -2.23 Uruguay 0.002 -0.051 0.110 0.135 0.029 -0.877 0.065 0.034 -0.036 -0.004 0.006 -0.032 -1.606 2.15 Venezuela, RB 0.114 -0.095 0.014 0.287 0.010 0.068 0.349 0.064 0.006 0.007 0.019 0.072 1.234 Note: See note for Table 7a. Table 7d: Contributing factors to net interest margin, Middle East & North America Comparison, 2005-14 [1] [3] [6] [8] [12] [14] [4] [5] [7] [9] [10] [11] [13] [15] Net interest margin [2] Economy Size of Reserve Income Information Contract Risk aversion Overhead Credit risk Inflation Interest rate risk Competition Creditor rights Other factors (v.s. world) operations opportunity cost diversification environment enforcement 2.39 Algeria -0.197 0.352 0.266 0.799 -0.023 -0.809 0.017 0.012 0.059 0.009 0.002 -0.054 1.956 -0.49 Bahrain -0.032 0.134 -0.038 -0.240 -0.006 0.120 -0.097 0.001 -0.028 -0.001 0.003 -0.024 -0.284 0.08 Egypt, Arab Rep. -0.080 -0.081 0.037 -0.091 0.039 0.224 0.262 0.006 -0.144 -0.006 0.002 0.026 -0.110 0.56 Jordan -0.118 0.044 0.081 0.151 0.000 0.139 0.023 0.004 0.014 0.009 0.009 0.012 0.194 -0.36 Kuwait 0.177 -0.013 -0.190 -0.063 0.004 0.030 0.005 0.007 0.068 -0.004 0.000 0.003 -0.384 -0.83 Lebanon -0.202 -0.176 -0.042 0.482 0.015 0.129 -0.060 -0.011 -0.133 -0.002 0.001 0.002 -0.833 -0.07 Morocco 0.136 -0.179 0.089 -0.262 -0.011 0.017 -0.141 -0.008 -0.031 -0.003 0.002 0.004 0.319 0.29 Oman -0.012 0.020 0.019 -0.165 0.001 0.207 0.021 0.002 0.034 0.001 -0.008 -0.010 0.179 -0.58 Qatar 0.123 0.039 -0.165 -0.224 -0.043 0.001 0.000 0.004 0.059 0.003 0.001 0.008 -0.385 -0.69 Saudi Arabia 0.283 0.029 -0.149 -0.074 -0.036 0.045 -0.014 -0.011 0.069 -0.007 -0.007 0.090 -0.908 -0.30 Tunisia -0.079 -0.167 0.093 -0.312 0.060 -0.102 -0.015 -0.005 0.033 0.001 -0.005 -0.058 0.255 Note: See note for Table 7a. 40   Table 7e: Contributing factors to net interest margin, South Asia Comparison, 2005-14 [1] [3] [6] [8] [12] [14] [4] [5] [7] [9] [10] [11] [13] [15] Net interest margin [2] Economy Size of Reserve Income Information Contract Risk aversion Overhead Credit risk Inflation Interest rate risk Competition Creditor rights Other factors (v.s. world) operations opportunity cost diversification environment enforcement 3.93 Afghanistan -0.250 0.089 0.335 1.047 -0.018 0.333 -0.069 0.005 0.036 0.004 -0.001 0.022 2.392 -0.95 Bangladesh 0.039 -0.002 -0.096 -0.250 0.004 -0.881 -0.056 0.036 0.007 0.004 0.002 0.025 0.218 -1.64 India 0.278 -0.068 -0.210 -0.282 -0.025 0.067 -0.021 0.016 0.028 -0.009 -0.008 0.024 -1.428 -0.92 Nepal -0.143 0.014 -0.230 -0.148 -0.016 0.323 0.004 -0.060 -0.028 0.004 -0.008 -0.029 -0.601 -0.75 Pakistan 0.076 0.034 -0.016 -0.201 0.057 0.001 0.102 -0.009 -0.042 0.001 0.000 -0.017 -0.738 0.33 Sri Lanka 0.000 -0.067 0.216 -0.166 -0.001 0.157 0.040 0.012 0.000 -0.005 0.016 -0.025 0.157 Note: See note for Table 7a. Table 7f: Contributing factors to net interest margin, Sub-Saharan Africa Comparison, 2005-14 [1] [3] [6] [8] [12] [14] [4] [5] [7] [9] [10] [11] [13] [15] Net interest margin [2] Economy Size of Reserve Income Information Contract Risk aversion Overhead Credit risk Inflation Interest rate risk Competition Creditor rights Other factors (v.s. world) operations opportunity cost diversification environment enforcement -0.23 Angola 0.124 0.036 -0.306 0.517 -0.005 -0.224 0.230 0.043 0.067 0.002 0.017 0.065 -0.795 -2.22 Benin 0.086 -0.177 -0.434 -0.165 0.037 0.004 -0.158 -0.023 -0.016 0.002 0.009 0.013 -1.398 -1.40 Botswana 0.051 -0.224 -0.615 -0.288 -0.036 0.472 0.082 -0.014 -0.050 -0.008 -0.006 -0.097 -0.666 -1.81 Burkina Faso 0.135 -0.196 -0.245 -0.006 -0.007 -0.273 -0.196 -0.021 0.010 0.002 0.003 0.015 -1.032 2.33 Burundi -0.156 0.052 0.867 0.039 0.045 -0.037 0.145 0.020 0.041 0.002 0.016 0.043 1.252 -2.51 Cameroon 0.106 -0.236 -0.210 -0.152 0.031 -0.448 -0.208 -0.030 0.025 0.001 0.005 0.026 -1.427 1.74 Congo, Dem. Rep. -0.059 0.061 0.901 0.710 -0.008 -0.644 0.049 0.062 -0.065 0.002 0.006 0.065 0.659 -1.86 Côte d'Ivoire 0.144 -0.225 -0.244 -0.339 0.023 -0.347 -0.214 -0.026 -0.017 0.002 0.006 -0.025 -0.595 -2.31 Ethiopia 0.031 -0.059 -0.870 -0.227 0.009 -0.209 0.276 -0.029 0.107 0.002 0.008 -0.028 -1.323 1.88 Gambia, The -0.253 -0.047 0.501 0.077 0.017 -0.339 -0.123 0.021 -0.036 0.002 0.002 -0.013 2.069 2.53 Ghana -0.004 0.017 0.233 0.101 0.018 0.339 0.195 0.056 0.040 -0.002 -0.015 -0.005 1.559 0.52 Kenya 0.025 0.190 -0.190 -0.211 -0.017 0.661 0.106 -0.005 0.047 0.002 -0.026 -0.019 -0.039 0.54 Madagascar -0.038 0.005 -0.446 0.170 -0.004 0.981 0.084 -0.014 0.000 0.002 0.023 0.018 -0.242 3.50 Malawi -0.171 0.139 0.845 -0.239 -0.038 -0.459 0.113 -0.009 0.001 0.002 -0.012 0.023 3.309 -1.43 Mali 0.012 -0.073 -0.064 -0.188 0.007 -0.451 -0.179 -0.024 -0.009 0.002 0.007 0.005 -0.477 -0.85 Mauritania -0.093 0.549 -0.365 0.216 0.071 -0.182 -0.048 -0.022 0.083 0.002 0.017 0.065 -1.145 -3.65 Mauritius 0.146 -0.088 -0.985 -0.023 -0.034 0.785 -0.080 0.002 0.052 -0.008 -0.004 -0.103 -3.306 1.44 Mozambique -0.080 0.062 0.269 0.187 -0.015 -0.023 0.086 0.042 -0.020 0.002 0.015 -0.006 0.918 -0.75 Niger -0.043 -0.178 -0.228 0.088 -0.026 0.121 -0.191 -0.023 -0.001 0.002 0.008 0.024 -0.305 0.79 Nigeria 0.289 0.188 -0.149 -0.055 0.016 0.378 0.131 -0.003 -0.037 0.001 -0.018 -0.013 0.056 2.26 Rwanda -0.053 0.127 0.428 -0.043 0.022 0.522 0.000 0.003 -0.009 -0.004 -0.010 0.051 1.222 -2.98 Senegal 0.067 -0.117 -0.197 -0.222 -0.010 -0.842 -0.268 -0.024 0.014 0.002 0.002 0.000 -1.386 4.15 Sierra Leone -0.255 0.092 0.855 -0.085 0.048 0.083 0.134 -0.028 -0.038 0.002 -0.008 0.041 3.312 -1.51 South Africa 0.168 -0.045 -0.269 0.116 -0.050 0.251 -0.054 -0.016 -0.063 -0.013 -0.009 -0.031 -1.497 0.21 Swaziland 0.009 0.052 0.156 -0.097 -0.008 -0.092 -0.001 -0.005 -0.034 -0.008 -0.006 -0.037 0.278 0.09 Tanzania -0.067 -0.045 0.022 0.122 -0.027 0.304 0.102 0.030 0.014 0.002 -0.010 0.004 -0.362 -2.49 Togo 0.010 -0.164 -0.121 -0.393 -0.010 -0.843 -0.165 -0.023 -0.030 0.002 0.007 -0.012 -0.745 2.52 Uganda -0.063 0.192 0.385 0.044 -0.034 0.585 0.079 0.042 0.009 0.002 -0.008 -0.046 1.331 1.50 Zambia -0.069 0.109 0.477 0.343 -0.011 -0.075 0.069 0.019 -0.084 -0.005 -0.020 -0.024 0.775 Note: See note for Table 7a. 41   Table 7g: Contributing factors to net interest margin, OECD High Income Economies Comparison, 2005-14 [1] [3] [6] [8] [12] [14] [4] [5] [7] [9] [10] [11] [13] [15] Net interest margin [2] Economy Size of Reserve Income Information Contract Risk aversion Overhead Credit risk Inflation Interest rate risk Competition Creditor rights Other factors (v.s. world) operations opportunity cost diversification environment enforcement -0.13 Australia 0.142 -0.035 -0.001 -0.076 -0.034 -0.040 0.020 0.002 -0.039 -0.001 -0.017 -0.037 -0.015 -0.76 Austria 0.097 -0.081 -0.054 0.019 0.000 -0.202 -0.019 -0.006 0.037 -0.001 0.001 -0.028 -0.518 -1.40 Belgium 0.336 -0.217 -0.261 -0.173 -0.028 1.065 -0.016 -0.003 -0.048 0.002 0.012 -0.059 -2.004 -0.28 Canada 0.014 -0.050 0.043 -0.070 -0.031 -0.101 -0.033 -0.007 -0.160 -0.004 -0.005 -0.064 0.186 1.59 Chile 0.015 0.269 0.189 0.334 -0.011 -0.378 0.043 0.015 0.000 0.000 0.014 0.104 0.992 0.16 Czech Republic 0.070 0.038 -0.034 0.003 0.003 0.128 -0.008 -0.007 0.086 -0.001 0.003 0.073 -0.199 1.07 Denmark -0.186 0.115 0.284 0.034 0.040 0.136 -0.025 -0.004 0.044 0.001 -0.014 -0.056 0.704 0.40 Estonia -0.217 0.098 -0.046 0.097 -0.002 -0.314 0.105 0.008 0.020 -0.001 0.001 0.080 0.572 -0.97 Finland 0.135 -0.105 -0.157 -0.073 -0.032 -0.600 -0.038 0.001 -0.084 0.002 -0.011 -0.062 0.056 -0.33 France 0.137 -0.009 0.067 -0.137 0.005 -0.629 -0.046 -0.006 -0.019 0.001 0.013 0.047 0.248 0.04 Germany -0.177 -0.039 0.032 -0.128 -0.017 0.294 -0.040 -0.012 -0.281 -0.004 0.004 -0.043 0.454 0.21 Greece 0.201 -0.073 -0.023 -0.063 0.015 0.318 0.035 -0.002 0.009 0.000 0.017 0.062 -0.283 1.19 Hungary 0.137 0.005 0.273 -0.024 0.068 -0.641 0.060 0.013 -0.005 0.001 -0.003 0.079 1.224 0.96 Iceland -0.208 0.265 0.198 0.089 0.031 -1.404 0.183 0.038 0.018 -0.001 -0.004 -0.038 1.791 -1.83 Ireland 0.189 -0.196 -0.457 -0.171 -0.036 1.211 0.034 -0.001 0.018 -0.001 -0.017 -0.059 -2.349 0.04 Israel 0.209 -0.126 0.077 0.520 -0.013 -0.446 -0.016 -0.001 0.005 -0.001 -0.012 0.034 -0.192 0.32 Italy -0.155 0.058 0.064 -0.168 0.014 0.162 -0.026 0.011 -0.062 -0.001 0.026 0.017 0.384 -0.86 Japan -0.003 -0.165 -0.212 -0.095 -0.009 1.421 -0.100 -0.014 0.080 -0.002 0.007 -0.074 -1.698 0.75 Korea, Rep. 0.195 -0.135 -0.080 0.020 -0.008 0.682 0.022 -0.004 0.059 -0.003 0.003 -0.042 0.040 0.29 Latvia -0.169 0.031 0.125 0.274 0.038 -1.002 0.113 0.021 0.033 0.004 -0.021 0.080 0.761 -0.70 Luxembourg -0.035 0.045 -0.128 -0.073 -0.018 -0.501 -0.020 -0.006 0.030 0.012 0.020 0.066 -0.090 -1.02 Netherlands 0.081 -0.003 -0.138 0.117 -0.008 -0.321 -0.030 -0.008 -0.074 -0.001 0.011 -0.059 -0.586 0.05 New Zealand 0.138 -0.024 0.038 -0.093 -0.030 -0.093 0.037 0.001 -0.012 -0.001 -0.024 -0.032 0.142 -0.10 Norway -0.156 -0.004 -0.119 -0.052 -0.029 0.707 -0.015 0.009 0.082 0.002 0.003 -0.067 -0.459 0.90 Poland 0.109 0.044 0.184 0.013 0.029 -0.439 0.007 0.001 0.050 -0.003 -0.011 0.074 0.846 0.03 Portugal -0.197 0.077 0.048 -0.123 0.046 -0.086 -0.036 -0.009 0.010 -0.001 0.027 -0.017 0.293 0.66 Slovak Republic 0.028 -0.004 0.058 0.098 0.015 0.103 0.010 -0.004 0.035 0.002 -0.008 0.048 0.279 -0.11 Slovenia 0.030 -0.035 -0.036 0.007 0.058 -0.083 -0.009 -0.005 0.003 0.008 0.018 0.052 -0.117 -0.48 Spain 0.066 -0.037 -0.106 -0.139 0.018 0.097 -0.020 -0.006 0.054 -0.001 0.006 -0.016 -0.397 0.35 Sweden -0.229 0.218 0.076 -0.192 -0.024 0.028 -0.059 -0.001 0.058 0.002 -0.004 -0.026 0.501 -1.13 Switzerland -0.136 -0.149 -0.180 0.003 -0.033 0.413 -0.119 -0.007 -0.029 0.000 -0.001 0.056 -0.947 -0.30 United Kingdom -0.119 0.159 0.044 0.065 0.003 -0.445 0.011 -0.006 0.049 -0.004 -0.017 -0.058 0.023 1.38 United States -0.343 0.065 0.233 0.157 -0.022 0.964 -0.006 -0.004 0.033 -0.004 -0.017 -0.035 0.356 Note: See note for Table 7a. 42