INTERNATIONAL DIVERSIFICATION OF SPANISH BANKS

Transcription

1 INTERNATIONAL DIVERSIFICATION OF SPANISH BANKS Isabel Argimón (*) Banco de España (Preliminary draft. Do not quote.31 Oct 2016) ABSTRACT The aim of this research is the analysis of the effects of banks international diversification on bank s risk. The less than perfect correlation of country specific risks suggests that the internationalization of the banking business can isolate the results of the consolidated banking group which operates in different countries from parent s country idiosyncratic shocks. On the other hand, diversification can involve an underestimation of risk as it may allow crosssubsidization across activities carried out in different countries, or because distance can hinder the ability of a bank s headquarters to monitor its subsidiaries. If such risk materializes, it can destabilize the bank or the banking group. The effects of such diversification could be embedded in capital requirements and should be taken into consideration when carrying out stress-tests. We provide empirical evidence of the relevance of international geographic diversification and business comovements between the Spanish and the host economy on bank s ex-post risk using detailed individual data of Spanish banks and banking groups. The results are consistent with the view that geographic diversification reduces risk. JEL classification xx, xx Key words: geographic diversification, internationalization of banks, cross-border activity, cross-subsidization I appreciate the help of C. Luna, E. Llorente, P. Núñez-Lagos and C. Sierra with the data (*) Corresponding author: Isabel Argimon, Banco de España; phone isabel.argimon@bde.es 1

2 Introduction Before the crisis, there was a strong debate on the need to incorporate diversification effects in regulatory risk metrics. The prevailing capital requirements framework did not properly account for the possibility of diversification benefits. Moreover, the use of bank stress tests since the crisis, both as a crisis management tool and as an early warning mechanism are based on consolidated balance sheets. The top-down stress-tests do not take into account differences in the geographical structure of banking groups, in the sense that the scenario of simultaneous shocks in different economies has a low probability. Banks see diversification as an opportunity for an upward shift in the risk-return trade-off, so that risk-adjusted returns should be higher at more diversified banks, in accordance with portfolio theory. The question of whether diversified financial institutions outperform their more concentrated peers has been widely researched (see Saunders and Walters (1994), DeYoung and Roland (2001) and Stiroh (2004) for a literature review). While business diversification benefits are expected from the reduction in the effect of idiosyncratic shocks on cash flow variance, the benefits of geographic diversification for internationally active banks and banking groups are expected to arise from non-synchronized fluctuations in economic activity across markets. However, diversification may also lead banks to assume higher risk and increase their market power (while it may generate international spillover and contagion effects that may spread through direct exposures or asset prices). A lower variability of net income to assets may incentivize banks to adjust downwards their capital to assets ratio or their expected net income to assets. In fact, diversification can be carried out into higher risk activities, thus increasing the overall portfolio risk even if the returns are not highly correlated. That is: global banks functional and geographical diversification can have both costs and benefits in terms of financial stability. During the crisis, risk seemed to spread across countries and recent trends in banking show increased bank concentration in terms of both of business lines and geographic activity. In fact, there seems to be a return to traditional banking and a home bias as far as geographic destination of activity. In particular, direct cross-border lending as a share of total banking assets has declined, while the share of local lending by foreign bank affiliates has remained steady (IMF, 2015). The reduction in cross-border lending may have reduced risk-sharing for banking groups as they have increased their relative exposure to domestic shocks. Moreover, from the perspective of recipient countries, the volatility of domestic credit may also have increased as it now has a lower foreign component, a component which is less exposed to domestic shocks. However, this trend has also reduced the cross-border transmission of foreign shocks, thus reducing overall financial volatility. The stability of local claims from banking subsidiaries has also contributed to isolate these recipient markets from shocks arising in parent s country or in general, abroad. Spanish banks stand out as the ones with the largest share of local activity (i.e. activity developed with the residents in a given country by the operative units of the consolidated group that reside in such a jurisdiction; that is, intra-boundaries activity) among the major banking systems (McCauley et al (2010)). The evolution of Spanish banks since the crisis has been characterised by an increasing weight of foreign activity in the consolidated results of banking groups. This does not imply that Spanish banks have been alien to the generalised tendency of global banks since the crisis to retrench from some geographical regions and/or business lines. In fact, factors such as strategy, business models, profitability, efficient allocation of scarce resources, political and social instability, the technological revolution and regulation may explain the decisions on the internationalization of banks. The overall financial stability effect of these change in unclear a priori. 2

3 Spanish banks follow the so-called multinational banking model (McCauley et al. 2010), which is characterized by the predominant role played by subsidiaries, in contrast to the relatively low weight of cross-border activity. The Spanish international banks have a business model focused on traditional commercial (mostly retail) banking, following the parent s business model. In the Spanish case, local claims are financed with local liabilities, so that subsidiaries have large funding autonomy. That is, each subsidiary must be funded independently, mainly through retail deposits, in their corresponding markets, and do not receive support from their parent, thus minimizing intra-group transactions. They are structured by incorporated local subsidiaries, directly regulated and supervised by local authorities, which often have a significant market-share. We would like to provide empirical evidence on what the effects of geographic diversification have been in terms of banks risk and return. We assess whether there has been an increase in terms of banks resilience resulting from the trend to increase the weight of business away from home in the Spanish case. The role of the specific country or economic area where foreign activity is directed would certainly have an important role. The first thing we want to look at is the effect of diversification through subsidiaries on banks performance. We do not analyse the determinants of geographic expansion, but take external presence as given. We assess the effects on risk of the degree of bank s diversification abroad and of the synchronization between economies. The relevance of the paper can be justified on different grounds: most available empirical evidence on the costs and benefits of diversification date from pre-crisis years. Moreover, most studies of geographic diversification have focused on local diversification (between regions in a given country). The international dimension has not been so extensively explored (Buch et al 2013; García-Herrero et al. 2013; Berger et al 2015.). Moreover, the international dimension of most national banking systems is characterized either by large cross-border exposures, so that they are aligned with the so-called international banking model or by diversification in the business line, so that they do not follow the Spanish approach of focusing their client base on local households and SMEs while retail deposits are the main source of financing. Additionally, we make use of individual bank data that cover the period before the financial crisis and up to Such information allows us to assess whether internationalization compensates or exacerbates the results of banks during macroeconomic downturns in the home country. This proposal focus on the time and international dimension, addressing the questions: what have been the effects of changes in international geographic diversification on banks performance? What have been the effects under recessionary periods in the home country? 1. Why diversify There are three main arguments that from a theoretical point of view justify diversification in the banking sector: increase market power, improvement in resource management and reduction of agency problems (Gulamhussen et al.(2011)). The improvement in resource management is linked to efficiency gains from scale and scope. It can be extended to arguments of cross-selling financial products or following their clients abroad (Focarelli and Pozzolo, 2005) and to the financial aspect of firm management linked to the isolation from idiosyncratic shocks and its positive impact in terms of the reduction of the cash flow variance. Risk diversification arising from operating in a larger number of geographic areas renders banks more resilient to shocks and reduces their costs 1. 1 Portfolio diversification theory shows that when returns are not perfectly correlated, diversification reduces risk (Markovitz (1959)), so that diversifying investment portfolios internationally may facilitate a decline in risk (Levy and Sarnat (1970), Driessen, Laeven (2007)) 3

4 However, diversification can lead to negative outcomes. On the one hand, it can be accompanied by increased complexity that can affect the organizational and operational level (Liang and Rhodes (1988)). It can lead to increased risk-taking as the lower costs of funding may be used to take additional risk. Moreover, the quality of bank loan portfolios is endogenous (Acharya et al (2002)) in the sense that it is determined by the level of monitoring that can depend on the diversification degree 2. In particular, less expertise in monitoring can be associated to lending in a new sector or location. It can also contribute to increased systemic risk: although diversification can benefit a bank individually it can generate interlinkages that could increase spillover effects and generate contagion in international markets. In a way, widespread diversification makes banks more similar to one another, exposing them to similar risks and thus increasing the probability of a joint failure. Finally, foreign exchange risk may increase risk as part of the assets is denominated in foreign currency. A part of the literature has focused on the analysis of the effects of geographic diversification on bank s market valuation addressing the issue of whether banks that diversify are valued at more or less than the sum of their constituting parts 3. In particular, Gulamhussen et al (2010) in an international sample of more than 500 large banks across 56 countries between 2001 and 2007 find that internationally diversified banks trade at a premium, but with a threshold, suggesting that the benefits of geographic scale and scope economies more than offset the agency costs. Another strand has assessed the effects of geographic expansion through mergers and acquisitions. In their study of the effect of geographic diversification on risk, Amihud et al. (2002) show that cross-border mergers and acquisitions have no net effect on the risk (and returns) of the acquiring banks. As for the effects of geographic diversification on risk and on return performance, the evidence is somewhat mixed. Liang and Rhoades (1988) using a sample of US banking organizations operating during the period show that geographic diversification across Metropolitan Statistical areas or counties provides opportunities for banks to reduce risk. However, they also find that as diversification increases, the level of earnings declines, which can result from increased operating risk resulting from changes in strategy, management or facilities. Achyara et al (2002) also find that that increased diversification reduces return at very high levels of risk, but increases returns at moderate levels of risks. Deng and Elyasiani (2008), analysing Bank Holding Companies in the US, find that geographic diversification is associated with risk reduction (and lower stock price variability), along with value enhancement. However, the magnitude of such reduction depends on the distance between the holding company and its branches, a result also found in Berger and DeYoung (2001) as diseconomies associated with distance are limiting gains from geographic diversification. The work of Deng et al (2007) gathers evidence that geographic diversification can provide a funding advantage as they show with data of over 60 US BHC from 1994 to 1998 that the diversification of deposits reduces the bond-yield spread. Goetz, Laeven and Levine (2015) find that geographic expansion reduces risk when banks expand into areas with low macroeconomic correlation. In particular, they analyse the effect on risk of Bank Holding Companies expanding into U.S metropolitan statistical areas with asynchronous business cycles, taking advantage for the identification strategy of the different calendar in the removal of the prohibition to enter other states. Acharya, Hasan and Saunders (2002), with information from 105 Italian banks over the period , disaggregated by industrial sector, asset decomposition and geographic region (Italy, other EU countries and Rest of the World), find that geographic diversification results in an 2 Acharya et al (2002) point at three reasons why diversification can reduce monitoring or monitoring efficiency: less expertise in monitoring can be associated to lending in a new sector or location; entrance into a highly competitive market can lead to adverse selection and increase in size can lead to agency-based scale inefficiencies. 3 We will not deal with the large strand of literature that has been devoted to the analysis of financial stability issues linked to the effects on the home or host economies of the presence of foreign banks or cross border activity. 4

5 improvement in the risk-return trade-off, but only for banks with low levels of risk. Positive diversification benefits are also found in Fang and van Lelyveld (2014). They use a correlation matrix approach to calculate international diversification effects and treat bank subsidiaries as individual assets of the banking group portfolio. The approach is applied to 49 large banking groups over the period and concludes that banks credit risk could be reduced by 1.1%, spanning from negligible to 8% Meslier et al. (2015) using an unbalanced panel of US BHC and banks from 1994 to 2006 analyse the effects of intrastate and interstate geographic diversification for bank risk and return. They assess the role of bank size and disparities in economic conditions within states or across states on these effects. They find that the effects of diversification depend on bank size and the size of economic disparities. For large banks both intrastate and interstate diversification is beneficial in terms of risk-adjusted returns while for small banks only the former is beneficial, but always with a threshold. Moreover, they also find that expanding activities towards markets with different economic conditions impacts the benefits of diversification. In particular, greater disparities in economic conditions amplify the effect of intrastate diversification on small banks risk. One of the main challenges in the empirical literature is the identification of an exogenous source of variation in geographic expansion and accounting for where the expansion takes place. If the banking group increases the riskiness of its asses when it expands geographically, an OLS regression will yield upwardly biased estimates of the impact of geographic diversification on risk. Moreover, the specific economies where the expansion takes place and its degree of dissimilarity with the home economy will also contribute to affect the impact. Goetz et al., 2016 address these issues using a gravity-deregulation methodology, and with data for BHC in the U.S. find that geographic diversity reduces risk As for the empirical evidence on negative effects of banks geographic diversification, Morgan and Samolik (2003) find for U.S. Bank Holding Companies for the period that domestic geographical diversification is not associated with higher returns and lower risk. Berger et al. (2015) with a population of U.S. commercial banks estimate for different measures of risk the role played by internationalization for the period 1989:Q1 to 2010:Q4. They find that international banks have higher risk than purely domestic banks and that such risk is increasing with the degree of internationalization. Gulamhussen, Pinheiro and Pozzolo (2014) using a sample of 384 listed banks from 56 countries for the period find that international diversification increases bank risk proxied either by the expected default frequency and by the z-score. 2. Basic approach to internationalization of bank s portfolio We follow the model presented in Berger et al (2015) of a bank s portfolio with two risky assets: a foreign asset with expected return µ f and standard deviation δ f and a domestic asset with expected return µ d and standard deviation δ d. The correlation between the two assets is ρ fd and the bank invests a proportion w in the foreign asset Two hypothesis can emerge from this simple approach: a) Benefits hypothesis (diversification hypothesis): banks will have lower risk when they diversify, as they are less exposed to domestic shocks. It implies ρ fd is low and δ f is not too large relative to δ d and the return µ f is not too low relative to µ d. b) Cost hypothesis (market risk hypothesis). Banks that diversify geographically have higher risk due to market-specific factors that make foreign assets relatively riskier (δ f is high relative to δ d and/or the return µ f is low relative to µ d) unless the higher risk is offset by a low correlation ρ fd. 5

6 We propose determining empirically which hypothesis dominates in the case of Spanish banks and whether there are differences before the crisis and since Data used 3.1. Sample banks We use quarterly balance sheet and income statements reports submitted by Spanish banks to the Banco de España. We use the whole population of banks and we consider only as international activity the activity of foreign subsidiaries that are classified as credit institutions or financial credit institutions 4. We link bank subsidiaries to their parent bank by using the anonymised reported identity of the entity that holds at least 50% of a bank s equity or has its control. The measure of geographic diversification could thus be affected by the possible transformation of branches into subsidiaries, although we do not observe any bank changing from domestic to international in our database, as international expansion of Spanish banks took place mostly at the end of the nineties, although some multinational banks increased their geographical scope in the twenty first century. The initial sample of banks is an unbalanced panel of banks spanning from 1999Q4 to 2014Q4, although the risk variable that we use starts in 2002Q4, because of the lag structure of the model that we estimate. The maximum number of banks in a given quarter in the initial database is 83 and the minimum is 26. We measure all variable at the company level and not at the individual bank level, as we want to capture the diversification provided via affiliation with banks in other locations. To compute the different measures of diversification, we use balance sheet and income statements data from consolidated financial statements that includes the activity corresponding to the subsidiaries abroad and we merge it with the statements provided for each country where the group has a subsidiary. If the bank does not belong to a group, we use its individual bank data. The initial dataset has information on 100 subsidiaries of 13 banks operating in 51 different jurisdictions 5. We keep detailed information on those subsidiaries whose weight on any of the main balance sheet items (total assets, total liabilities, deposits, loans or wealth) is above 5% of total foreign activity. We add the rest in an aggregate which we call other which includes activity whose weight on the consolidated balance is minor or which is carried out in subsidiaries in small jurisdictions such as Andorra or Panama, for which we do not have data on their economies. We remove any bank-quarter observations with missing or incomplete financial data on accounting variables such as total assets, equity, loans and deposits, remove those observations for which the annual asset growth rate is above 30% and, finally, omit observations for those banks for which we have information for less than three consecutive years (12 quarters). Our final database consists of 64 banks, seven of which have been engaged in international activity through subsidiaries in any given time during the period of analysis. We distinguish among 15 different jurisdictions. During the period of analysis, bank mergers and acquisitions took place, and not only associated to the financial crisis. The literature has followed different approaches to consider them. Meslier et al (2015) identify banks whose total assets has grown by more than 30% between any two consecutive years (t-1 and t) and construct a dummy variable which is equal to 1 in year t and the two following years and zero elsewhere. All the estimates exclude these three-year windows as the time dimension is used to compute measures of risk. Liang and Rhoades (1988) run two regressions: one with the 5509 banking organizations that existed 4 The so-called «Filiales entidades de crédito y filiales establecimientos financieros de crédito. 5 The gross database contains 153 subsidiaries of 17 banks operating in 53 different jurisdictions. Howevert there is no net data for 53 of these subsidiaries, only gross data that cannot be directly compared to consolidated information. The maximum number of subsidiaries a bank has is 26 and the minimum 1. 6

7 over the entire period and another one with observations that included all banking organizations that were in existence in 1976, but had merged, been acquired, failed, or changed their name by In spite of the fact that they only report the results for the complete panel, they state that the results were robust across different samples. Martín Oliver et al. (2014) analysing the links between productivity and social welfare in the banking industry applied to Spanish banks over the pre-crisis period of , use unconsolidated balance sheets and income statement data for each bank and year, with merged banks treated as a new entity. Doing the same implies that you only look at the cross section dimension, so that you can assess whether more or less geographic diversification has positive or negative effects on risk adjusted results. We have initially carried out the analysis in such a way that the acquiring bank s code is maintained and the target bank drops from the sample. Following Meslier et al (2015) we present the results obtained when we drop the observations of banks whose total assets has grown by more than 30% between any two consecutive years Risk measure We propose using the Z-score as our measure of risk. The Z-score should be interpreted as a distance-to-default measure, i.e. as the number of standard deviations the Return on Assets (ROA) defined as net income divided by total assets can diverge from its mean before the bank defaults. A higher Z-score indicates a safer bank. It is thus defined as: Z-score =(ROA +E/A) / δ ROA So it is calculated as the sum of the bank s mean ROA and mean capitalization ratio, E/A (the average equity to assets ratio over the same period as the average ROA) divided by the standard deviation of ROA (δ ROA), which proxies its volatility. A higher Z-score signals a lower probability of bank insolvency 6. We also use the log of z-score to mitigate the impact of outliers. As we use data from consolidated statements, these series reflect the ex-post risk generated by all the operations of the bank in both their home countries plus those of their subsidiaries and branches abroad. In the robustness checks, we also employ two alternative measures of risk: the standard deviation of ROE, where ROE is net income over equity and the standard deviation of ROA International diversification Internationalization can be addressed with different identification strategies: the number of countries where banks are active (the extensive margin), the share of foreign relative to total assets and liabilities (the intensive margin), and the different modes of entry into foreign markets (cross-border asset holdings versus foreign branches or subsidiaries) 7. We measure geographic diversification taking only into account the international activity carried out through subsidiaries, using the financial statements that bank groups provide in relation to their foreign activity by country with the operations of their subsidiaries overseas. This does not imply that the expansion carried out through cross-border lending from the home country activity or through the establishment of physical branches is not relevant, but it may have different effects and motivations, so that we initially focus on expansion through subsidiaries 8. 6 Following Berger et al (2015) we compute Z-score over a 12-quarter period. 7 Initially, we only consider international activity through subsidiaries. 8 A branch typically has no legal personality separate from its parent. Thus, in principle, under a branch structure, all of the parent bank s assets are available to cover all of its liabilities in case of resolution or liquidation, regardless of their geographic distribution. A subsidiary is locally incorporated and the parent bank s liability is 7

8 In particular, we construct a Hirsch-Herfindhal index to capture international concentration, thus taking into account both the number of countries in which a bank has a subsidiary, including its activity in Spain and the weight of each activity. We define it so that the lower bound corresponds to a non-diversified (domestic) bank and the upper bound corresponds to the most internationally diversified bank: HH j,t= 1- Rj j=1 (Assets in subsidiaries or parent bank j / total assets j) 2 Where R j is the total number of jurisdictions where bank j operates, including the home country. Again, a value close to 0 indicates very low geographic diversification while as it grows close to 1 it reflects higher diverse international presence. Stiroh and Rumble (2006), Gulamhussen et al (2014), García-Herrero and Vázquez (2013), Meslier et (2015) make use of this index, that we also define using deposits as the relevant variable. Also, following Cetorelli and Golberg (2014) in their research on bank complexity, we normalize the Herfindhal index by R, the number of countries/regions where the bank operates and define: NHH j,t= R j/(r j-1) ( 1- Rj j=1 (Assets in subsidiaries or parent bank j / total assets j) 2 ) We also define international share as the proportion of assets in foreign subsidiaries over total assets of the banking group: share j,t= foreign subsidiaries assets j,t/total assets j,t Again values close to one indicate high international activity Business cycle synchronization According to Markowitz s portfolio theory, diversification implies that a banking group s or a bank s total credit risk resulting from subsidiaries activities is lower than the simple sum of risks from each one of the individual subsidiaries. Such result arises because of the less than perfect correlation of credit risk drivers in different countries. We propose making use of the concept of business cycle synchronization, developed in the literature that studies conditions on the optimality of currency areas (Busl and Kappler (2013)) to proxy the correlation between the fluctuations in economic activity across markets. In particular, the bilateral synchronization between country k and country j at time t, η kjt is measured as the negative absolute difference between the two countries' real GDP growth rates. As an alternative measure of economic activity, we use the change in the unemployment rate. η kjt = - Y kt - Y jt It is defined so that higher values of η kjt (the closer to zero) indicate a higher degree of bilateral synchronization between country k and j in year t. As Busl and Kappler (2013) point out this approach has several advantages over a traditional time-invariant correlation measure of business cycle synchronization. On the one hand, it allows synchronization to vary over time. On the other hand, η ijt is independent of the underlying sample period for each t. Finally, it is based on observed growth rates instead of filtered rates, which can be subject to measurement errors. We use the OECD statistics of quarterly seasonally adjusted GDP growth and the quarterly unemployment rate, not seasonally adjusted and import prices from the World Economic limited to the capital it holds in it. In principle, subsidiaries have an independent board as well as local liquidity and risk management practices; however, many international banks have centralized practices in which the boundaries of the individual legal entity may be blurred. While managing risks on a consolidated basis makes sense from an economic perspective, the existence of different legal entities may limit the free movement of resources between institutions belonging to the same group. (IMF,2015) 8

9 Survey, IFO. As we have diversified banks operating in different jurisdictions, we construct the index as a weighted average of the one to one synchronization where the weights are the former quarter share of the foreign country assets on total assets. (See other synchronization indices in by Kalemli-Ozcan, Papaioannou and Peydró (2013) Financial regulation, financial globalization, and the synchronization of economic activity. Journal of Finance 68, pp Control variables We include also a set of control variables, which are time-varying bank-level, as they affect bank s risk outcomes. In particular, we follow Berger et al and define a measure of Income diversification as a greater reliance on non-interest income is linked to higher volatility in returns. Interest_income= net interest income/(net interest income + net fees and commissions income) We also include size measured as the ln of total assets as it is an important determinant of the capacity to internationalize as they may have economies of scale in foreign exchange management. On the other hand, size may affect risk taking, as larger banks have a greater capacity to absorb risk (Berger et al 2014) and may be affected by moral hazard associated to the too big to fail hypothesis. To capture the cost structure, we construct the variable overhead cost as the ratio of bank s operating expenses over total assets. Demirgüc-Kunt and Huizinga, 2010 found that banks with higher costs were less stable. 4. Empirical results 4.1. Some stylized facts We first look for differences in the ex-post risk attained by banks that geographically diversify through subsidiaries and those that do not. We measure returns by the before tax return on assets (ROA) calculated over a three-year period and risk by its standard deviation and by z- score. We also compute a measure of ex-post risk-adjusted returns dividing mean ROA over its standard deviation. The figures recorded in Table A.1 in the Appendix distinguish between the group of individual banks that are purely domestic and those that are international and Table A.2 distinguishes between the years up to 2007 and since then. The results show that for the overall period, banks that did not diversify had a lower return (in terms of ROA), but also a lower risk, both in terms of standard deviation and in terms of z-score. Therefore, we observe the traditional trade-off between return and risk and find that international banks are riskier than purely domestic banks. We also observe that domestic banks are smaller than those that diversify and that both their interest income and their overhead costs are higher Empirical specification of baseline model Following previous work in this area, we postulate estimating equations of the form: Y i,t-k+1,t = α 1i + α 2t +β 1 DIV i,t-k +β 3 BS i,t-k + γ X i,t-k + λ DIV i,t-k *X i,t-k +ε i,t-k+1,t 9

10 where Y it is a measure of performance (profitability, risk-adjusted return, risk-taking or default risk of bank i at time t), which is computed over the k periods from t-k+1 to t, while the independent variables, proxied as discussed in the previous section, are measured at period t-k to ensure that they are predetermined in relation to the dependent variable. In particular, DIV it is an indicator of international diversification, BS it is the weighted index of synchronization or correlation between the business cycle of Spain and the country where the subsidiary has been set and X it are control variables, both at the bank and the time dimension (bank s efficiency, size,..) Baseline results We first compute our diversification and synchronization measures at the country level. Table 1 provides regression results on the relationship between the risk of the bank, its overall international geographic diversification and the role that macroeconomic synchronization between the home and the host country can play, while controlling for time-varying bank characteristics and time fixed effects. We present the results using our main measure of bank risk, z-score, computed over a three year period, following Beck et al, We report robust t-ratios adjusted for bank clustering so as to account for unobserved time-invariant features at the bank level. The findings in Table 1 provide support to the hypothesis that internationalization contributes to reducing risk. The relevance of the correlation between economic cycles and the diversification of assets seems to depend on whether we are considering domestic and multinational banks or we limit the analysis to those banks that have international subsidiaries. We find evidence that synchronization affects ex-post risk with a negative sign, so that the higher the correlation between the Spanish business cycle and that of the host country the lower the risk, when we consider the overall sample or only large banks (those above 1000 million ). Therefore, international diversification seems to reduce banks exposure to idiosyncratic local market risks. When we limit the sample to multinational banks synchronization does not seem to play a role on risk. However, we find evidence of a positive relationship between international geographic diversification and stability for international banks (col (3)), supporting evidence that the more internationalized banks are, the lower the ex-post risk that they show. A we do not find evidence that lower economic synchronization leads to lower risk in internationalized banks, all the benefits of being multinational arise because of the geographic diversification of assets. In fact, when we include in the analysis those banks that are domestic we do not find statistical evidence that banks risk differs depending on the amount of geographic diversification, but only on synchronization. Turning to bank controls, we find that firm size has a negative statistically significant coefficient, which is not consistent with larger banks having better risk management skills or greater capacity to absorb losses through risk diversification. This finding is more aligned with the toobig-to.-fail hypothesis. We also find that overhead costs have a negative impact on risk, which is consistent with the findings of Demirgüç-Kunt and Huizinga, and Berger et al, 2015 that banks with higher costs are less stable. Finally, the sign of the variable interest income is not statistically significant so that the weight of traditional business does not seem to affect bank s ex-post risk Internationalization before and after the crisis We next examine the effect of internationalization, synchronization and bank risk before and after the crisis as there may be differences in how internationalization affects risk during financial crisis. Two explanations compete in relation to what to expect from internationalization 10

11 in a crisis context. On the one hand, more internationalised banks may be subject to less risk than domestic banks as their exposure to shocks is mitigated as they hold assets and liabilities in both the domestic and the foreign market, which tend to be imperfectly correlated. On the other, internationalised banks may increase their risk as they may face higher costs associated to managing risk in different jurisdictions or because parents and subsidiaries show risk correlation because of their business model or their funding interrelations 9. Their sources of financing may be more volatile during crises as they have higher dependence on capital and inter-bank markets than purely domestic banks. We first split the sample in two, with observations up to 2007 and observations from 2007 onwards and run separate regressions. We then define the dummy variable from 2007 that takes value 0 up to 2007 and 1 since the first quarter of 2008 and include the interaction terms herfindhal x from 2007 and syncro_u x from 2007 to test whether there is a difference in the relationship of diversification and risk before and after the crisis. To contrast whether the effect does not depend so much on whether it is before the crisis or not, but whether it depends on the economic cycle we also construct multiplicative variables herfindhal x growth and syncro_u x growth, where growth is the change over same quarter, previous year. The results reported in Table.2 suggest that the negative impact of synchronization on risk is statistically significant both before and after the beginning of the crisis (cols(1) to (3)), so that internationalization to jurisdictions with low economic synchrony seems to contribute to isolate banks, in general, and from the crisis, in particular. This could result from international banks lowering more their risk than domestic banks or receiving different government support in the crisis. We control for the possibility of this last hypothesis by taking into account that saving banks (cajas), which are not publicly traded and which do not have international subsidiaries, were the recipients of most public support. We thus run the regressions without saving banks (col(4)), so that we isolate the role of capital markets in monitoring risk and avoid distortions caused by direct public support. We find again that higher synchronization is associated with lower risk, and we find evidence that such relationship changes after the crisis, when it becomes positive. We also find that before the crisis the Herfhindal index has a negative impact, suggesting that geographical diversification before the crisis leads to higher ex-post risk, a result that does not hold after the crisis. When we test if the cycle (growth) is reinforcing the relationship between international diversification and risk and between synchronization and risk, the coefficient is not statistically significant (col (5)). We thus do not find evidence that the economic cycle contributes to the impact of synchronization or geographic diversification on risk. Therefore, these findings suggest that it is not internationalization per se that affects banks ex-post risk, but the asynchronicity in economic cycles that reduces risk. 5. Taking into account differences in economic areas Up to here, we have defined the variables treating each country where Spanish banks have subsidiaries equally. So, for instance, the calculated Herfindhal index is the same whether a bank is diversifying into advanced economies or into emerging economies, as far as the weight in total assets were the same. We can think that the risks may differ in both areas. By the same 9 Levine et al, 2016 find that the geographic expansion of banks across U.S. states lowered their funding costs, especially if the headquarter of the bank is located in a state with low correlation with the overall U.S. economy. On the other hand, Anginer et al find evidence of a positive correlation between parent bank and foreign subsidiaries default risk, which is lower for subsidiaries that have a higher share of retail deposit funding and that are more independently managed from their parents. 11

12 token, we have assumed that the effect of synchronization does not depend on the economic area, which is not necessarily the case. We next try to provide a disaggregated view to establish the relevance for banks risk of the specific economic areas where international activity takes place. Grouping countries by geographic regions that share some common features should allow establishing with greater clarity the contribution of synchronicity and diversification to risk. In particular, we distinguish between three regions that we denote advanced (adv), Latin American (lat) and other countries (other). We define two sets of international share variables for each group of countries, which will allow considering the differential effect of having subsidiaries in these three different regions versus not diversifying and the effect of having diversified in a specific region instead of another one. We compute the first set as in the previous section so that each one of them is the proportion of assets in foreign subsidiaries located in each geographic region over total assets of the banking group (share_lat, share adv, share_other). So the sum of these three shares provides the overall share of foreign assets in subsidiaries for the Spanish banks. We compute the second set only for those banks that have diversified geographically. We define these new variables as the weight that each geographic region has on the total foreign assets in subsidiaries for each banking group. We include in the regressions the corresponding share of LatinAmerica and other countries (intshare_lat, intshare_other), so that the comparison is with the diversification that has been carried out in advanced economies. We also decompose the aggregate Herfindhal index that we built into three local Herfindhal indexes (hh_lat, hh_adv and hh_other) to measure the concentration of the assets of each international bank within each one of these groups of countries. These three indices allow us to capture the contribution of each of these groups of countries to the effect of international diversification on risk. The sum of these three Herfindhal indexes equals the Herfindhal index for a given parent bank calculated in the previous section. We can interpret these new Herfindhal indexes as the contribution of the region to the overall effect of diversification, comparing them to a purely domestic bank. We also decompose the synchronization index into three synchronization indexes for each one of the regions where we only include, in each one of them, those countries that are located in the specific country group and weight it using the weight on total assets lagged one period. Summary information on the distribution of international bank assets across regions is presented in Table A.3 of the Appendix. On average, banks that diversify geographically maintain 82% of its assets at home, against 12% in subsidiaries. These banks maintained, on average, 20% of these foreign assets in Latin America and 70 % in advanced economies, other than Spain. As economic cycles are more synchronized in industrial countries (Griffith-Jones et al, 2002), we can expect that benefits will be greater if diversification is into emerging countries. However, as synchronization can be expected to be stronger within a given region, we can also expect that the greater the diversification among regions, the greater the benefits. The average return abroad will certainly influence the final result. We run similar regressions as before and present the results in Table 3. Again, we find evidence that overall it is the lower synchronization of economic cycles that reduces risk (col (1) and (2)), and that the distribution of assets among different regions does not affect ex-post risk. However, when only multinational banks are analyzed, geographic areas in terms of diversification matter for risk, both when considering only the total weight of the different regions on total assets or when considering the relative weights as captured by the Herfhindal index. So, again, we find evidence that the more diversified in different regions are multinational banks the lower their ex-post risk. 12

13 The differences before and after 2007 for these multinational banks, as recorded in Table 4. We find evidence that before the crisis (cols (1) to (3)) it is the synchronization with Latin American countries which negatively contributes to risk, while after 2007, the synchronization with other non- advanced countries also has a negative effect. We do not find evidence that the synchronization with advanced economies affects risk. We also find evidence that before 2007, the higher the proportion of assets into Latin American or other non-advanced economies the lower the risk, while after the crisis it was the proportion of assets into Latin America and into advanced economies that contributed to a reduction in risk. Therefore, when we decompose the international activity of multinational banks by regions we find evidence that the less synchronized are the economies where multinational banks have subsidiaries and the higher the volume of assets to these economies the lower the ex-post risk. We also find evidence that while synchronization with advanced economies does not seem to affect risk, the weight that assets in these economies contributes to the reduction of risk after Decomposition of z-score We next analyse the contribution of each one of the components of the z-score (ROA, capital ratio and sd of ROA) to the results that we have obtained and present the results in Table 5. We find that ROA and the capital ratio (cols (1) and (2)) are positively affected by synchronization, so that the higher the synchronization the higher the return and the capital ratio, while the standard deviation of ROA (col(3)) is not affected by the correlation between economies. So it seems that it is the positive effect of synchronization on profitability and solvency that explains the negative effect that synchronization has on risk as captured by z- score. As for the effect of diversification, we find that the contribution of Latin American countries is positive for profitability while the contribution of other emerging countries is negative for capital, while no region has a contribution on the standard deviation of ROA. It is therefore either the lack of statistical significance of diversification into different regions or the rather mixed results that they provide for the different components of z-score which accounts for the null effect of diversification on risk. When we consider only international banks, we find evidence that synchronization with Latin American countries positively affects profitability, while synchronization with advanced economies reduces it (col (4)). Synchronization with different economic regions does not seem to affect capital ratios, but it affects the standard deviation of ROA. In particular, the higher the synchronization with Latin American or with other non-advanced economies the higher the contribution to the variability of profitability. All these results may support the finding that risk as captured by z-score increases as the synchronization with Latin American and other nonadvanced economies increases. On the other hand, the diversification into different regions does not seem to affect profitability, but it affects capital ratios, so that the higher the diversification into Latin American or advanced economies the higher the leverage ratio. Moreover, the higher the diversification into advanced or not-advanced-non Latin American economies the lower the volatility of profits. So it seems that the positive role that diversification into Latina American economies and other nonadvanced economies have in the reduction of risk, stems from capital ratios and the volatility of profits. 13

14 6. Summary and preliminary conclusions This paper provides empirical evidence on how international geographic diversification of Spanish banks assets through subsidiaries affects ex-post risk. It also provides evidence on the relevance of business synchronization for risk. It is the first paper that focuses on internationalization with Spanish data covering the crisis period and which assesses the role played on ex-post risk by the correlations between the home and the host country. The results suggest that geographic diversification per se does not affect ex-post risk when we do not discriminate between domestic and international banks. We provide evidence that it is the level of synchronization that matters for risk, thus supporting the view that benefits of geographic diversification mostly arise because of the lack of correlation between the foreign and the domestic asset. On the other hand, when we only consider multinational banks, the higher the international diversification, the lower the risk. For these set of banks, ex-post risk is reduced by increasing the number of jurisdictions in which they have subsidiaries and the weight their assets represent for the group. There are no major differences in these relationships between the period up to 2007 and the period after that year, but we find that diversification becomes positive and relevant after When we group countries by economic regions and distinguish among them we find that the specific areas where there are subsidiaries are important. We find again that, overall, it is the lower synchronization of economic cycles that reduces risk. When only multinational banks are analyzed, economic areas where international activity through subsidiaries takes place matter, both in terms of synchronization and diversification. In particular, we find that low economic synchronization with Latin America reduces risk both before and after 2007 and that the low economic synchronization with other non-advanced economies also contributes to risk reduction after We do not find evidence that higher synchronization with advanced economies affects risk. As for diversification, the results differ before and after Before the crisis, diversification into LatAm or other non-advanced countries reduced risk, while after the crisis, it was the diversification into LatAm and into advanced economies which reduced risk. Moreover, before the crisis, the higher the weight that LatAm and non-advanced economies had on foreign assets, the lower the risk, while after the crisis, the distribution of foreign assets into different geographic areas does not affect risk. The evidence gathered supports the diversification hypothesis that provides for benefits in the international diversification of banks, which is mostly channeled through the lower synchronization with the countries in which subsidiaries operate. It also provides evidence that the specific regions are important to account for these benefits, which reinforces the finding on the relevance of synchronization. 14