1
Foreign banks and financial stability in Europe
Bart Schooten s2243733
Faculty of Economics & Business University of Groningen
Master Thesis
MSc. International Economics and Business Supervisor: Dr. P.A. IJtsma
Co-assessor: Dr. D.J. Bezemer Final Version
July 2018
2
Table of content
ABSTRACT ... 3
1. INTRODUCTION ... 4
2. LITERATURE FRAMEWORK ... 6
2.1 FOREIGN BANKS ... 6
2.2 STABILITY OF THE FINANCIAL SYSTEM... 11
2.3 RESEARCH FRAMEWORK ... 13
3. METHODOLOGY & DATA ... 15
3.1 METHODOLOGY ... 15
3.1.1 Z-SCORE ... 15
3.1.2 NONPERFORMING LOANS TO TOTAL GROSS LOANS ... 16
3.1.3 CONTROL VARIABLES ... 16
3.2 DATA ... 18
3.2.1 WORLD BANK ... 18
3.2.2 DUTCH CENTRAL BANK (DNB) ... 19
3.3 VARIABLE OPERATIONALIZATION ... 19
3.3.1 FOREIGN BANKS ... 19
3.3.2 FINANCIAL STABILITY ... 19
3.3.4 DATA DESCRIPTIVES ... 20
4. RESULTS ... 24
4.1 Z-Score... 24
4.2 NONPERFORMING LOANS TO TOTAL GROSS LOANS ... 26
5. DISCUSSION & CONCLUSION ... 28
5.1 DISCUSSION ... 28
5.2 CONCLUSION ... 29
5.3 LIMITATIONS AND FURTHER RESEARCH ... 30
REFERENCES ... 32
APPENDICES ... 36
I. COUNTRY SAMPLE ... 36
3
Foreign banks and financial stability in Europe
Bart Schooten University of Groningen
ABSTRACT
The global financial crisis has revived the debate on the effects and influences of foreign banks on banking- and financial stability. This study examines the effect of foreign banks on financial stability by utilizing the Z-Score, a systemic risk indicator which measures the default
probability of a financial system at the country level and the financial soundness indicator nonperforming loans. The sample consists of 30 European countries between 2007 and 2013.
The results show that foreign banks have a negative, however insignificant impact on the Z- Score. Foreign banks also tend to decrease nonperforming loans. While also this relationship is not significant, foreign banks tend to increase asset quality. Overall, the results indicate that foreign banks do not have a significant impact on the host country’s financial stability.
Policymakers should however be wary of the effects of foreign banks in times of crisis.
Keywords: Foreign banks, financial stability, systemic risk JEL classification: B26, G15, G21, N24
4
1. INTRODUCTION
Financial deregulation, the opening of international capital markets and rapid advances in technology have created the path towards the globalization of the financial sector. The globalization of financial systems and the presence of foreign banks have increased rapidly due to these developments. From 1995 to 2009, the total number of banks around the globe remained the same. However, foreign banks have increased its market share from 20% to 34% in the global banking sector (Claessens & Van Horen, 2014).
Over the past decades, foreign bank have become more influential in domestic financial intermediation. The expanded presence of foreign banks has increased the need for policymakers to understand their impact and behavior. Before the global financial crisis of 2008 and several other financial crises around the globe, there was a general
consensus that financial integration was a benefit which outweighed the costs.
While cross-border lending and bank credit decreased sharply during periods of crises, the decline in foreign bank presence is limited and remains roughly stable (Claessens &
Van Horen, 2015). After the global financial crisis, several countries have seen the downsides of financial integration and its serious side effects. The crisis revived the debate on the costs and benefits of financial integration and foreign bank ownership (Claessens & Van Horen, 2013).
With the on-set of the financial crisis of 2008, policy makers and governments have put the financial sector on the top of their agendas (Beck et al., 2009). Especially with the recent financial crisis in mind, policy makers are keen to know the effects of foreign banks on the financial system in their countries. Sound financial analysis and policy advice has to be based on appropriate data by capturing different dimensions of financial sector development, where foreign banks play an important role (Beck et al., 2009).
Since foreign banks are increasingly present in domestic financial system and become more influential, their role and effects on domestic financial stability is worth examining.
Foreign banks may have the potential to strengthen financial systems, however may also be a source of contamination or aggravation of financial distress (Allen et al, 2011).
Therefore, the effect of foreign banks are relevant for policy makers. The effects of foreign banks on domestic financial stability have not been examined through the use of
5 systemic risk measures. This paper attempts to assess the relationship of foreign bank presence and the impact on financial systems by considering a systemic factor in financial stability. This study contributes to the existing literature by examining a country’s Z-Score, a systemic risk indicator that measures financial stability by
comparing the buffer of a commercial banking system to the volatility of its returns. The issue of systemic risk is seemingly the most important in assessing financial stability.
Next to the Z-Score, nonperforming loans will also be examined. This paper examines the relationship of foreign banks and financial stability and uses a sample of 30 European countries in the period of 2007-2013.
The results show that foreign banks tend to have a negative impact on a country’s Z- Score and increase the systemic risk of the financial system. This result has to be interpreted with care, as the relationship is not significant. Furthermore, foreign banks increase asset quality by decreasing nonperforming loans. Also here, the relationship is not significant. This implies that in the period of the global financial crisis, foreign banks were not a stable source of stability (De Haas & Van Lelyveld, 2014).
The main implication of the findings in this study is that foreign banks tend to decrease a country’s Z-Score and increase systemic risk. Policymakers should take knowledge of the effect of the foreign banks on financial stability during periods of financial crisis. This result, combined with the findings of this paper, shows that policymakers may try to maximize the total amount of foreign bank assets in their financial system.
The outline of this paper is structured as followed: the next chapter contains a literature review on foreign banks and financial stability from which the hypotheses will be
derived. The third chapter discusses the research methodology and the data. Chapter four provides the results from the data analysis. The final chapter provides a discussion on the results and the conclusion.
6
2. LITERATURE FRAMEWORK
This chapter discusses the literature review, which focuses on foreign banks and the stability of the financial system. The hypotheses derived from this framework will be tested and its results are discussed in chapter 4.
2.1 FOREIGN BANKS
Deregulation of the financial sector has led to several big reforms and the liberalization of the sector (Beck et al., 2009). This has resulted in the reduction of entry barriers for cross-border lending and foreign banks to enter financial systems.1 This, in combination with increasing globalization and technological developments, has transformed the global financial- and banking sector. Claessens & Van Horen (2014) show that the total number of banks in the world have remained practically the same between 1995 and 2009. However, the share of foreign banks in the global banking system went from 20%
in 1995 to 34% in 2009 (Claessens & Van Horen, 2014). The relative importance of foreign banks in global financial systems has increased. According to Claessens & Van Horen, this conclusion describes two outweighing trends. From 1995 to 2009, the number of domestic banks has decreased while the number of foreign banks has
increased. Multinational banks have become more important in financial intermediation.
Grubel (1997) was one of the first researchers to set up a theoratical framework of multinational banking. Multinational banks primarily enter foreign markets for
diversification purposes and profit maximization. Foreign banks can enhance financial stability by increasing the efficiency of capital flows. However, foreign banks may also harm financial systems through attempting to eliminate compulsory minimum reserve requirements and the threat of losing national monetary sovereignty. Gray & Gray (1981) add to this by comparing the theory of non-financial multinationals corporations to multinational banks. Their model shows that foreign banks serve to integrate national capital markets with possible endangerment to stability of national financial systems.
Another model by Claessens & Van Horen (2012) argue that banks serve clients in foreign countries and banks may achieve efficiency and scale gains, which strengthens financial stability. The design by Goetz et al. (2016) show that multinational banks’ risk diversification may be better, which may enhance their opportunities to undertake
1 Following Claessens & Van Horen (2015), a bank is considered to be a foreign bank if more than 50% of the shares are held by foreign entities.
7 higher return investments and reduce overall risk which in turn enhances financial stability. Overall, theorical models have shown that foreign banks may have positive and negative effects on financial stability.
The global financial crisis has changed the global financial landscape and has accelerated structural reforms and regulations. Banks in Europe and the United States, which held most of the subprime mortgages assets, were most affected by the financial crisis. These crises have shown that spillovers and contamination of banks on financials systems are serious issues. Agreements on capital requirements have been made before the crisis in which banks were required to meet stricter capital requirement and restore balance sheets. These regulations gave banks an incentive to downscale their international operations and cross-border activities (Claessens & Van Horen, 2015). Due to these issues, there is a large mixture of banks active in the international market. (Claessens &
Van Horen, 2014; Claessens & Van Horen, 2015). However, as mentioned previously, foreign banks have increased their presence of the past 2 decades and multinational banks have become a more important of financial systems. Foreign banks have also increasingly penetrated in developing countries and emerging markets (Buch, 1997;
Claessens et al., 2008; Claessens & Van Horen, 2013). In 1995, foreign banks had a market share of 18% in emerging markets and 24% in emerging countries, while this increased to respectively 36% and 45% in 2009. Foreign banks play a more dominant role in financial intermediation in these countries.
Several studies have examined the effect of foreign banks on the host banking system and markets. Foreign bank presence can have a positive impact on domestic financial systems in several ways. First, foreign banks may act as a buffer against negative credit supply shocks as seen in the crises originated in 2007 (Barboni, 2017). These credit supply shocks take place in the domestic credit market, which is heavily hit by a
country-specific adverse shock. Foreign banks may mitigate negative financial shocks in countries where domestic financial banks and intermediaries are unexpectedly hit by the consequences of a financial crisis. It can therefore act as a buffer and enhance financial stability.
In addition, international active banks that are more diversified can more smoothly absorb shocks occurring in the host market and can potentially be a more stable capital source (Claessens & Van Horen, 2013). This potential however, depends on the foreign
8 bank and the characteristics of the local market. Foreign banks may also transmit
financial shocks to another country (Peek & Rosengren, 1997). Foreign banks can affect the supply of credit and transmit a domestic banking crisis abroad, which potentially can induce a financial crisis and affect real activities in the host country (Peek & Rosengren, 2000). On the other hand, Honohan & Laeven (2005) argue that foreign banks have, in most cases, not played an important role in the emergence of crises. They argue that foreign banks may help in the aftermath of a crisis by acting as rehabilitators of failed banks and act as instruments in the banking system-wide reforms. Their presence and influence may mitigate the effects of the crisis and thereby enhance stability.
Banks become more active and efficient as their home countries become richer. Richer countries host more efficient banks which have lower costs and are more used to competition (Demirgüç-Kunt & Levine, 1999). As these countries are home to more efficient banks and banks with lower costs, these bank have a competitive advantage over their foreign counterparts. These banks may then make the decision to open a branche or bank in a foreign country. Next to that, foreign banks can improve the efficiency of the domestic banking market by introducing new innovative systems and decrease costs of financial intermediation (Claessens, Demirgüç-Kunt & Huizinga, 2001).
Foreign banks may also enhance financial integration between countries and regions.
This exposure to the neighboring regions is beneficial for countries, as it allows for the diversification of liquidity risk through the large network of the interbank market in Europe (Buch, Kleinert, & Zajc, 2003).
In general, diversification strategies can reduce overall risk. Goetz et al. (2016) develop a new identification strategy to evaluate the impact of geographic expansion of a bank.
For the average bank, the results show that geographic expansion materially reduces risk, while it does not affect loan quality. The results of Goetz et al. are consistent with arguments that geographic expansion lowers risk by reducing exposure to idiosyncratic local risk. However, where foreign banks may enjoy lower overall risk, risk of domestic banks increases with the presence of foreign banks in the host country. (Wu et al., 2017).
9 Before the global financial crisis of 2008, there was a consensus that financial
integration and foreign banks was an asset which outweighed the costs (Claessens &
Van Horen, 2013). The global financial crisis has revived the debate concerning foreign banks and is on the top of the agenda of policymakers and governments. Foreign banks may cause financial instability due to contagion effects and the withdrawal of credit and liquidity in host markets. Claessens & Van Horen (2012) find that foreign banks during the global financial crisis reduced credit supply more than their domestic counterparts.
Foreign banks exposed to home-country crises in 2007 and 2008 exhibit changes in lending patters that are lower by 13 and 42 percent than non-crises country banks (Adams-Kane et al., 2014). This result is also applicable to the Asian crisis of 1997-1998, where banks exposed to crisis in their home country are forced to lower their lending patterns across border considerably. Next to that, De Haas and Van Lelyveld (2014) find that parent banks were not a significant source of strength to their subsidiaries during the crisis of 2008-2009. Multinational bank subsidiaries had to slow down credit growth and supply almost three times as fast as domestic banks. Countries which have large presence of foreign banks within their banking systems thus had less credit available and were prone to financial instability.
Internationally borrowing foreign banks tighten their credit supply more during the financial crisis than domestic banks that are locally funded (Ongena et al., 2013; Dekle and Lee, 2015; Fungáčová, Herrala, & Weill, 2013). Before the global financial crisis, foreign banks provided credit supply stability, particularly in Eastern and central
European countries (De Haas & Van Lelyveld, 2006). Furthermore, foreign banks led to a sounder banking system in Latin America, where foreign banks show robust loan
growth and a more aggressive response to asset quality deterioration (Crystal et al., 2001; Crystal et al., 2002).
Financial globalization intensifies the international transmission of financial shocks with substantial real consequences for the banking sector and the real economy. When the domestic banking system has been hit by a shock, foreign banks can step in for the failing domestic banks and have the potential to be a stabilizing source (Allen et al., 2011). However, foreign banks can also transmit shocks from abroad, increasing the probability of a banking crisis. This risk is aggravated by the fact that foreign capital sources are more mobile than domestic capital (Allen et al., 2011). Foreign banks could
10 retract and reduce credit supply more easily, inducing a credit- and banking crisis which leads to financial instability. Foreign banks may facilitate capital outflows, currency crises and financial instability as a consequence (Demirgüç-Kunt et al., 1998).
Complications in local markets may force foreign banks to flee the market. These issues of commitment may enhance the fragility of the (host) financial system, especially when foreign banks have penetrated the host market to a large extent. Claessens, Demirgüç- Kunt, & Huizinga (2001) also show that foreign banks increase banking competition in the host country. Foreign entry may reduce profits for domestic banks and therefore be a destabilizing force on financial stability. This especially holds for countries where the prudential regulation and supervision are not strong.
Foreign banks take on more risk than their domestic counterparts (Chen et al., 2017).
Mainly the foreign banks’ informational disadvantages, agency problems and contagious effect the parents’ bank has on the financial condition amplify the risk-taking behavior.
Increased risk is a source of potential default of a bank. Anginer et al. (2017) find evidence of a significant correlation between parent banks’ and foreign subsidiaries’
default risk. Host country regulations are associated with the extent to which the shock to the parent affects the subsidiaries’ default risk. However, the correlation of the default risk is lower for subsidiaries in host countries that impose stricter rules, namely higher capital ratios and disclosure requirements. This increased default risk imposes higher instability of the host financial systems.
To summarize, the existing literature has shown that foreign banks may have positive effects on the domestic banking system but that these effects are subject to certain conditions. The global financial crisis has changed the global banking market
fundamentally and revived the debate regarding financial policy and regulation. Certain countries were hit more intensively by the crisis than others and the strengthening of financial systems is not certain through the entry of foreign banks. However, this study is examining the impact of foreign banks on the financial stability in European countries to find the collective impact foreign banks have on the European region, which is
connected through several key institutions (Nugent, 2017). This analysis executed by looking at the systemic risk of a the financial system within a country.
11 Based on this analysis and the discussed theory and evidence, the hypothesis is
formulated as followed:
Hypothesis 1: Foreign banks have a positive impact on the financial stability in the host country.
2.2 STABILITY OF THE FINANCIAL SYSTEM
Financial systems are important to foresee the economy with the required credit and capital. Crystal et al. (2001) examine whether foreign bank entry into emerging markets improve the condition of the institutions and the financial system. They find that foreign bank entry enhances local financial stability and that the financial strength ratings of the banks improved. Foreign ownership may provide important positive influences on the financial stability in an emerging market and the development of banking system.
Claessens & Van Horen (2013) argue that foreign banks may offer valuable
diversification services and have the potential to absorb shocks in the host market.
However, foreign banks may also be a source of contamination in times of financial crises. One striking observation of the global financial crisis of 2008 which originated in the United States is how quickly and rapid it spread out to the rest of the world
(Blanchard, Das, & Faruqee, 2010). The crisis intensified from the United States to Europe and in the fall of 2008, emerging countries where affected as well with financial instability. When faced with funding or capital shock, foreign banks may withdraw their cross-border activities and redirect to lending at their home country. De Haas & Van Horen (2012) show that cross-border bank lending contracted sharply after Lehman Brother filed for bankruptcy in 2008. Banks which had to write down on their sub-prime assets and banks with had to refinance large amounts of long term debt transmitted these shocks across borders. De Haas & Van Horen (2013) also show that foreign banks reduced overall credit supply in the host country, weakening their financial system.
Amplification of financial distress is also an issue. Contagion of banks is an issue which played a big role in recent crises (Kim, Loretan & Remolona, 2010). When a bank suffers large losses or large write-offs on its assets, other banks are likely to suffer from these losses due to linkages in the financial system. Foreign banks may thus transmit crises from their home country abroad into other financial systems.
12 The financial stability of a country concerns the security and strength of the financial system and has become an important issue for policy makers. Financial stability is defined as “the ability to facilitate and enhance economic processes, manage risk and absorb shocks” (IMF, 2018). Financial stability is however not straightforward to
measure or define as there are interdependencies and complex interactions of different elements of the financial system within themselves and the real economy (Gadanecz &
Jayaram, 2009).
The severity of the financial crisis of 2007 has caused institutions and government to implement new measures and indicators to measure the systemic risk in financial
systems. Schaeck and Čihák (2007) proposed the use of financial soundness indicators, a set of indicators of the banking sector which are supplemented by other tools and
techniques such as systemic risk measures. The International Monetary fund has compiled these indicators in the Compilation Guide on Financial Soundness Indicators (IMF, 2006). “The primary purpose of this Compilation Guide on Financial Soundness Indicators is to provide guidance on the concepts and definitions, and sources and
techniques. This is the assessment and surveillance of the strengths and vulnerabilities of financial systems, with the objective of enhancing financial stability” (IMF, 2006). Next to the Z-Score as a systemic risk measure, the financial stability indicator nonperforming loans to total loans will be used a robustness feature. The Z-Score and nonperforming loans are mostly backward looking. As these indicators are not used as early-warning indicators in this paper, this is not an issue.
Because of the globalization of the banking sector and interconnected banks around the globe, crises can spread quite rapidly and banks may contaminate banks abroad.
Therefore, it is important to take systemic risk measures into account. Systemic risk is defined by the European Central Bank as “the risk that the provision of necessary financial products and services by the financial system will be impaired to a point where economic growth and welfare may be materially affected” (ECB, 2018). Bisias et al. (2012) and Benoit et al. (2017) show that there are several factors that influence the systemic risk factors. Liquidity, stock market volatility and the interbank market are the most prominent measures which are important for systemic risk and are included in this analysis. The most important measure on the country level is the Z-score, which measures the default probability of a financial system.
13 To summarize, previous studies have not examined the impact of foreign banks on the financial stability at the country level. This paper studies a sample of 30 European countries during the period between 2007-2013, the years of the initiation of the crisis and in the years of recovery. To measure and capture the effect of foreign banks, the approaches of Čihák (2007), Čihák & Schaeck (2007) and Babihuga (2007) are followed and the systemic risk measure is Z-Score, followed by asset quality (measured by the ratio of nonperforming loans to total loans).
Table 1 provides an overview of the most important literature and the outcome of these studies which have been discussed in the literature framework.
2.3 RESEARCH FRAMEWORK
The hypothesis are graphically presented in figure 1.
Foreign banks H1 (+)
Host country financial stability
Figure 1 Research framework
This research framework is tested through the hypothesis derived in the literature review: Foreign banks have a positive impact on the financial stability in the host country.
The relevant theories and literature discussed has shown that foreign banks may have a positive impact on the financial stability in the host country. Therefore, the relationship is expected to be positive. This research is conducted for 30 European countries during 2007-2013.
14 TABLE 1
Literature overview
This table provides an overview of the most relevant literature on the relationship between foreign bank presence and the impact on the stability on the financial system that is discussed in chapter 2.
Authors (year) Independent
variables Sample Main findings
Claessens & Van Horen
(2012) Bank ownership Bank-level data from
1999-2006 Foreign banks tend to perform better when from a high income country and when regulation in the host country is relatively weak. They also perform better when larger and having a bigger market share. Foreign banks from home countries with the same language and similar regulation as the host country also perform better.
Claessens & Van Horen
(2013) Bank ownership Bank ownership data
from 1995-2009 Foreign banks tend to outperform domestic banks in developing countries. During the global financial crisis, foreign banks reduced credit more than their domestic counterparts and are a source of potential financial instability.
Claessens & Van Horen
(2014) Bank ownership Bank-level data in 135
countries in 1995- 2009
Between 1995 and 2004, the total number of domestic and foreign remained roughly the same. The number of domestic banks has decreased, driven by deregulation and technological developments while the total number of foreign banks increased. The share of foreign banks went from 20% in 1995 to 34% in 2009.
Goetz et al. (2016) Bank ownership Bank-level data Multinational banks’ risk diversification may be better, which may enhance their opportunities to undertake higher return investments and reduce overall risk which in turn enhances financial stability.
Demirguc-Kunt, Levine &
Min (1998) Foreign bank
presence Bank level data of
banks in 80 countries Foreign banks do not increase the likelihood that countries suffer a banking crisis and improve the overall efficiency of the domestic banking sector. Next to that, foreign banks accelerate long-run economic growth by boosting the domestic bank efficiency as well.
De Haas & Van Lelyveld
(2014) Multinational banks Bank-level data on the
48 large banks During the financial crisis of 2008-2009, parent banks were not a significant source of strength to their subsidiaries.
Multinational bank subsidiaries had to slow down credit growth and supply almost three times as fast as domestic banks.
Barboni (2017) Foreign bank
ownership Belgium firm-level
data Barboni (2017) finds that foreign banks may act as a buffer against negative credit supply shocks as seen in the crises originated in 2007. These credit supply shocks are in the domestic credit market, which is heavily hit by a country-specific adverse shock.
Chen et al. (2017) Bank ownership 1300 banks in 32 emerging countries during 2000-2013
Foreign banks take on more risk than their domestic counterparts. Mainly the foreign banks’ informational disadvantages, agency problems and contagious effect the parents’ bank has on the financial condition amplify the risk-taking behavior. Chen et al. find supportive evidence that these factors play a significant role in affecting the foreign banks’ risk-taking behavior.
Wu et al. (2017) Foreign bank
penetration Bank-level data from
35 countries The overall risk of domestic banks increases with the presence of foreign banks in the host country, especially when they enter the host market via M&A. The host country’s financial system is subject to instability.
Dekle & Lee (2015) Foreign bank
ownership Bank-level data on 2100 banks across the world
Dekle and Lee (2015) find that foreign affiliates tighten their internal capital markets at its headquarters. They find evidence that internal capital markets do indeed affect cross-border lending. In particular, European and U.S. bank affiliates operating in South America and Asia reduced their lending patterns by more than the domestic banks and increase financial instability.
Crystal et al. (2001) Bank ownership in
emerging markets Bank-level data in
Latin America Crystal et al. find that foreign bank entry enhances local financial stability and that the financial strength ratings of the banks improved. Foreign ownership may provide important positive influence on the financial stability in an emerging market and the development of banking system as a whole.
15
3. METHODOLOGY & DATA
This chapter starts with discussing the methodology of the research framework. Next, the operationalization, the databases and the data are described.
3.1 METHODOLOGY
In this section, the methodology of the research is described. As mentioned previously, the Z-Score will be used to measure the systemic risk of a country’s financial system.
Also, nonperforming loans to total gross loans will be examined as a robustness feature.
3.1.1 Z-SCORE
Following the work of Čihák (2007) and Čihák & Heiko (2007), the following fixed effects model is estimated:
𝒁 − 𝑺𝒄𝒐𝒓𝒆𝒊𝒕 = 𝜶𝒊+ 𝜷𝟏𝑭𝒐𝒓𝒆𝒊𝒈𝒏𝒊𝒕+ 𝜹 𝑪𝒐𝒏𝒕𝒓𝒐𝒍𝒊𝒕+ 𝜸𝒕+ 𝜺𝒊𝒕 (1) 𝑍 − 𝑆𝑐𝑜𝑟𝑒𝑖𝑡 is the aggregate Z-score for country i at year t. It captures the probability of a country’s commercial banking system. The Z-score compares the buffer of commercial banking system, which includes market capitalization and returns with the volatility of those returns. It is calculated is as the sum of the return on assets and equity over assets over the standard deviation of the return on assets, which is a proxy for return volatility.
The Z-score captures the systemic risk of a financial system. It is directly related to the bank system’s insolvency, which is the probability that the value of the assets become lower than the value of debt and thus becomes bankrupt (Čihák & Heiko, 2007).
Bankruptcy or financial distress then leads to financial instability. The higher the Z- score, the lower the probability a financial system will default and the more stable the financial system is. 𝛼𝑖 is the country specific effect.
𝐹𝑜𝑟𝑒𝑖𝑔𝑛𝑖𝑡 is the percentage of assets that are held by foreign banks, among total banks assets in country i at year t. It captures the foreign bank participation in a host country’s financial system. It is chosen over foreign bank share, as foreign assets feature a clearer pattern of how much of banking assets is in fact owned by foreign entities. Foreign banks may have better risk diversification over domestic banks and can source capital more easily. Foreign banks may also enter markets when their overall position is solid and have steady profits. Therefore, foreign bank participation is expected to have a positive influence on the Z-score of the host financial system.
16 Lastly, 𝛾𝑡 corrects for the unobserved year effect and 𝛿 𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑡 represents a set of control variables for country i at year t.
3.1.2 NONPERFORMING LOANS TO TOTAL GROSS LOANS
Building on the work of Babihuga (2007) and Kasselaki & Tagkalakis (2014), the following fixed effects model is estimated:
𝑵𝑷𝑳𝒊𝒕= 𝜶𝒊+ 𝜷𝟏𝑭𝒐𝒓𝒆𝒊𝒈𝒏𝒊𝒕+ 𝜹 𝑪𝒐𝒏𝒕𝒓𝒐𝒍𝒊𝒕+ 𝜸𝒕+ 𝜺𝒊𝒕 (2) 𝑁𝑃𝐿𝑖𝑡 is the share of nonperforming loans to total gross loans of country i in year t.
Nonperforming loans displays the asset quality of the loan portfolio within a country. It is calculated as the percentage of total nonperforming loans to total loans. A loan in considered to be nonperforming when the payment of the installments if 90 days past their due date. Nonperforming loans displays the assets quality of a banking system, which is an important feature as these assets are the main profit drivers for banks. 𝛼𝑖 is the country specific effect.
𝐹𝑜𝑟𝑒𝑖𝑔𝑛𝑖𝑡 is the percentage of assets that are held by foreign banks, among total banks assets in country i at year t. It captures the foreign bank participation in a host country’s financial system. It is chosen over foreign bank share, as foreign assets feature a clearer pattern of how much of banking assets is in fact owned by foreign entities. The
relationship between foreign banks and nonperforming loans expected to be negative.
Foreign banks, or banks in general, prefer to work in environments where assets are of high quality. Therefore, foreign banks may decided to enter market with high asset quality or markets where banks can enhance asset quality through better screening.
Lastly, 𝛾𝑡 corrects for the unobserved year effect and 𝛿 𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑡 represents a set of control variables for country i at year t.
3.1.3 CONTROL VARIABLES
With respect to the control variables, several approaches are followed. Following Önder
& Özyildirim (2012) and De Haas & Van Lelyveld (2006), 2 macroeconomic variables are used: gross domestic product (GDP) growth rate and the unemployment rate. These variables capture and control for the macroeconomic conditions in the countries. Next to that, approaches of Benoit et al. (2017) and Bisias et al. (2012) are followed. These
17 approaches capture fundamental indicators. Please see table 2 under data descriptive for the descriptive statistics.
• Gross domestic product (GDP);
• Unemployment;
• Stock market volatility;
• Concentration;
• Net interest margin;
• Overhead costs;
• Liquid assets;
• Interbank lending ratio.
The gross domestic product growth rate and unemployment are macroeconomic
indicators. These indicators measure the overall economic conditions in year t. Önder &
Özyildirim (2016) use macroeconomic variables in their analysis of the influences of foreign banks. As financial systems tend to have a procyclical character, which means that systemic risk relates to the progressive build-up of financial fragility and the evolving of aggregate risk over time. Over the macro-economic cycle, the dynamics of financial systems and the real economy tend to reinforce each other (Caruana, 2010).
Stock market volatility is the average stock price volatility of the 360-day volatility of the national stock market index and control for overall uncertainty and stability within markets and economies (Liu & Zhang, 2015). The higher the stock market volatility, the higher the uncertainty within markets in a country. Uncertainty may have cause the banking sector to become distressed and therefore have an impact on its systemic risk.
In order to control for the structure of the banking system in country i, the percentage of the three largest commercial banks as the share of total commercial banks for country i at time t is measured by the concentration of the sector.
Furthermore, the profit structure of the financial sector is measured through the net interest margin and the overhead costs. The net interest margin is the net interest a bank receives upon its loans minus the costs of the deposits. As banks pay interest on short term deposits and receive from long term loans such as mortgages, this interest margin is subject to sensitive interest changes which are subject to overall market conditions. The net interest margin is not a measure for direct profit, however it is a
18 measure of the important business of the banking system. The overhead cost measure the operating costs relative to the assets of the banking system.
Liquid assets (as a percentage of total assets) measures the asset liquidity of banks on the country level. Liquidity problems has been an issue during the crisis. When the market’s demand for illiquid assets is less than perfectly elastic and characterized by financial distress, the price for such assets are depressed. Liquid assets are therefore important within the systemic risk of a financial system, as illiquidity is was a significant problem in times on financial distress (Cifuentes, Ferrucci, & Shin, 2005).
Interbank lending ratio is the ratio of interbank loans received and provided. A rate of 100% would indicate an equal share in interbank loans received and loans provided, where a rate above 100% would mean that banks have extended more interbank loans than it has received. Interbank lending measures the interdependencies of banks on (overnight) loans and the interconnectedness of financial systems. These interbank loans have grown substantially over the past years. To the extent that interbank loans are not collateralized or insured, a bank’s failure may trigger a chain of bank failures.
Therefore, interbank loans are an important feature of financial stability (Rochet &
Tirole, 1996; Acemoglu, Ozdaglar, & Tahbaz-Salehi, 2015)
3.2 DATA
3.2.1 WORLD BANK
The World Bank has composed a database concerning key indicators of the financial sector. Their global financial stability report (GFSR) offers 111 key indicators of financial development and structures across countries, which measure the size,
efficiency and activity of banks and financial intermediaries on the country level. These indicators have been used by the World Bank to investigate the link between legal and policy environment and the indicators of financial structure. Their database also offers macroeconomic indicators, such as Gross Domestic Product growth (GDP) and
unemployment rates.
19 3.2.2 DUTCH CENTRAL BANK (DNB)
Claessens & Van Horen (2015) compiled the bank ownership database which is owned by the Dutch Central Bank (DNB). This database includes for each bank its year of establishment, years of inactivity, its ownership (domestic or foreign), and the home country of the majority shareholder if the bank is owned by a foreign entity. 2 The data is used for the measurements of the percentage of foreign bank assets among total bank assets and the percentage of foreign banks present in the host country. The database contains full bank ownership data over the period of 1995-2013 for 5498 banks in 139 countries.
3.3 VARIABLE OPERATIONALIZATION 3.3.1 FOREIGN BANKS
Foreign banks are obtained by the database of Claessens & Van Horen (2015). Foreign bank presence is split up in two parts, foreign banks present in a host country and the percentage of the assets foreign banks hold in a host country.
The presence of foreign banks is calculated as the percentage of the number of foreign owned banks to the total number of banks in an economy. Following Claessens & Van Horen (2015), a bank is considered to be foreign when 50 percent or more of its shares are owned by foreign entities. The percentage of foreign assets is calculated as the share of total banking assets held by foreign banks in a host country. The percentage of foreign assets to total assets is used in this analysis, as it reflects the foreign bank participation in the host country.
3.3.2 FINANCIAL STABILITY
Financial stability is operationalized through the Z-score. The Z-Score is used to measure the systemic risk of the financial stability in an economy. Next to that, one aggregate measure of financial stability is considered.
The Z-score is a measure on the aggregate country level that captures the probability of default of a country's commercial banking system. Z-score compares the buffer of a country's commercial banking system (the capitalization and returns) with the volatility
2https://www.dnb.nl/en/onderzoek-2/databases/bank.jsp
20 of those returns. It is a risk-adjusted measure of the stability of the host financial system (Demirgüç-Kunt, Detragiache, & Tressel, 2006).
The financial stability indicator is nonperforming loans to total gross loans. A bank loan is considered “to be nonperforming when more than 90 days pass without the borrower paying the agreed installments or interest” (ECB, 2018). A nonperforming loan is
problematic, as loans provide banks with interest income, the main source of income for banks. When loans turn into nonperforming loans, banks need to set aside more capital on the assumption that the loan will not be paid back (in full). This capital buffer reduces the banks’ capacity to provide new loans. In short, the higher non-performing loans are, the less profit banks can realize, and it reduces the provision of new loans.
Nonperforming loans are therefore crucial to financial stability, as it requires banks to hold more capital and it lowers the ability to provide economies with the required credit and stimulate economic development. It is not a systemic measure of risk, however it is an important feature of a financial system and will be used as a robustness test for the systemic measure.
3.3.4 DATA DESCRIPTIVES
This section describes the dataset, which is a combined set the World Bank and the Dutch Central Bank. The dataset consisted of 32 European countries from 2007 to 2013, which had in total 224 observations. The dataset contained14 extreme positive outliners in terms of the Z-score. These 14 observations came from Austria and Luxembourg. As these outliners were extremely disproportional to the other observations, Austria and Luxembourg were taken out of the dataset. After this, there were outliners in the
extreme low z-scores. After taking out the extreme outliners, the sample consisted of 30 European countries with 190 observations. Please see table 5 in appendix I for the countries in the sample.
21 Figure 2 Z-score and foreign bank participation
The scatterplot of the data is displayed in figure 2. As can be seen, the data is more distributed towards the extreme ends in terms of foreign bank presence. Foreign bank presence is concentrated around the 0% to 30% and 70% to 100%. Furthermore, the Z- score varies in between 3 to 15, which means that the systems vary from unstable to very stable in terms of systemic risk. As the observations are quite scattered, there might be heteroskedasticity. The White test and Breusch-Pagan test for
heteroskedasticity show that indeed the dataset suffers from heteroskedasticity (p-value of 0.02). To overcome this issue, heteroskedastic robust standard errors need to be used. There is no issue regarding heteroskedasticity within the regression on
nonperforming loans (p-value of 0.78), however robust standard errors will be used here too.
Figure 3 Nonperforming loans and foreign bank participation
22 Figure 3 shows the scatterplot for nonperforming loans and foreign banks. As can be seen, nonperforming loans varies between 0,1% and 25,81%. This difference is quite large, however the main extreme outliners have been taken out of the dataset. Overall, the scatterplot indicate a slight upwards trend towards increased nonperforming loans.
Furthermore, the Hausman test shows that the random effects model is not
appropriated to use in this dataset. The p-value for the Hausman test is 0.00, which indicates that the random effects model is not appropriate and thus the fixed effects model, as discussed in the methodology, is used.
Table 2
DESCRIPTIVE STATISTICS
This table provides the descriptive statistics of the main variables and control variables.
The statistics are based on the combined dataset of the World Bank and the Dutch Central Bank. The statistics are based on the 190 observations from 30 European countries.
I. Foreign bank participation (%)
Mean 44,51
Min 0
Median 34,54
Max 99
II. Z-Score
Mean 8,06
Min 3,17
Median 7,46
Max 16,28
III. NPL (%)
Mean 5,97
Min 0,1
Median 4,29
Max 25,81
II. Control variables (mean)
GDP growth 0,75 %
Unemployment 9,87 %
Stock market volatility 25,51 %
Concentration 66,55 %
Net interest margin 2,40 %
Overhead costs 2,55 %
Liquid assets 35,89 %
Interbank lending ratio 0,97
23 The descriptive data from the dataset is displayed in table 2. Foreign banks account for 44,51% of the banking assets in host countries in Europe. Whereas one country reported no foreign banking assets at all (Sweden, over the full 7 year period), one country has as much as 99% of banking assets owned by foreign entities (Estonia, over a 3 year period).
As can be seen, GDP growth is quite low (0,75%), which is consistent with what is expected during a financial crisis. Connecting to the GDP rate, unemployment rate is on the high side (9,87%). Furthermore, the overhead costs are higher than the net interest margin (2,55% and 2,40%). Last, the three largest commercial banks possess 66,55% of the countries’ assets.
24
4. RESULTS
This chapter provides the results of the various analyses, where the results of the estimation methods and methodology in chapter 3 are estimated. This chapter starts by presenting the results of the relationship between foreign bank presence and financial stability for the Z-score, followed by the robustness tests for nonperforming loans.
4.1 Z-Score
The results of the regression estimate of financial stability and the Z-score (equation 1) can be found in table 3.
TABLE 3
Z-SCORE AND FOREIGN BANKS, MULTIVARIATE RESULTS
This table provides estimation results of the fixed effects regression models. The dependent variable is the Z-score. The results provided are based on the sample of 190 observations over the time period of 2007-2013 for 30 European countries.
The coefficients are reported along with the corresponding robust standard errors in parentheses.
*p<.10, **p<.05, ***p<.01.
Z-Score Z-Score
Foreign banks (%) -0.62
(2.70)
Concentration (%) -0.04
(0.03) -0.04
(0.03)
Net interest margin 1.16
(0.38)*** 1.18
(0.42)***
Overhead costs 0.00
(0.00) 0.00
(0.00)
Liquid assets 0.04
(0.01)** 0.04
(0.01)**
Interbank lending ratio -0.45
(0.15)** -0.44
(0.15)**
GDP growth -0.02
(0.04)
-0.02 (0.04)
Unemployment -0.05
(0.09) -0.05
(0.09)
Stock market volatility -0.01
(0.03) -0.01
(0.03)
Constant 7.47
(2.86)**
7.65 (2.75)**
𝑅2 0.35 0.36
Observations 190 190
Time fixed effects Yes Yes
25 The results suggest that foreign banks do not have a significant impact on the Z-Score.
While the sign is negative, it has no significant effect on the Z-Score. Foreign banks thus do not significantly impact systemic risk of a host country which implies that foreign banks, however the sign being negative, do not destabilize financial stability in the host country.
The results suggest that the aspect of liquidity in assets is important for systemic risk.
Liquid assets is significantly positive related to the Z-Score, which suggests that a
financial system with more liquid assets is more stable. Liquidity has been an important issue in the financial crisis of 2007, and seems to be of importance to the systemic risk during and after the crisis.
Net interest margin also has a significant, positive relation with the Z-Score. As net interest margin is an indicator of the main profit driver of a bank, a positive margin is positively related to the Z-Score and financial stability. Higher margins indicate that banks in a financial system have positive earning on their loans over the deposits, which strengthens the capital adequacy and returns of the bank.
Interbank lending is also significantly related to the Z-Score. Recalling that the higher the ratio for interbank lending means that the banks in a financial system have extended more loans than it has received, this indicates that the more banks lend to other banks (in other financial systems), the more it has a negative impact on the systemic risk.
Lending more to other banks than receiving it therefore is deemed risky, as the
relationship with systemic risk is negative. This could indicate that the interbank loan is at risk of being paid back (in full).
The results suggest that the macroeconomic variables do not have a significant effect on the Z-Score. GDP growth, unemployment rate and the stock market volatility have no significant impact on the systemic risk of a financial system. However, the sign of GDP growth is opposite to what is expected, namely negative. The GDP growth rate is negatively associated with the Z-Score. When GDP growth rate is positive, the
relationship with the Z-score is expected to be positive, as a positive GDP rate indicates that the positive cyclical business cycles can encourage the build-up of reserves and enhance financial stability. This result suggest that the GDP growth rate has a negative affect on the Z-score and systemic risk, however, the result is insignificant. The
26 unemployment rates and stock market volatility is expected to be negatively associated with the Z-Score, as is the case here. Last, concentration of the banking system and overhead costs do not seem to have a significant impact on the Z-Score.
4.2 NONPERFORMING LOANS TO TOTAL GROSS LOANS
Table 4 reports the results for nonperforming loans to total gross loans. The results indicate that foreign banks have a negative impact on NPL and thereby improve the asset quality for countries after the crisis. Foreign bank reduce asset quality by
increasing the nonperforming loans in host countries. This result is in accordance with Barth et al. (2002), where the share of assets in foreign banks has a negative effect on nonperforming loans. This result presented here however, is not significant. This implies that foreign banks do not have a significant relationship with NPL and thus do not
influence financial stability.
TABLE 4
NONPERFORMING LOANS AND FOREIGN BANKS, MULTIVARIATE RESULTS
This table provides estimation results of the fixed effects regression models. The dependent variable is nonperforming loans. The results provided are based on the sample of 190 observations over the time period of 2007-2013 for 30 European countries.
The coefficients are reported along with the corresponding robust standard errors in parentheses.
*p<.10, **p<.05, ***p<.01.
NPL NPL
Foreign banks (%) -5.72
(6.30)
Concentration (%) 0.05
(0.07) 0.06
(0.08)
Net interest margin -2.01
(0.76)**
-1.85 (0.78)**
Overhead costs 0.07
(0.01)*** 0.07
(0.01)***
Liquid assets -0.05
(0.02)* -0.05
(0.03)*
Interbank lending ratio -1.63
(0.23)***
-1.56 (0.24)***
GDP growth -0.04
(0.09) -0.03
(0.09)
Unemployment 0.48
(0.18)** 0.48
(0.18)**
Stock market volatility -0.17
(0.05)***
-0.16 (0.05)***
Constant 7.09
(5.50) 8.86
(6.62)
𝑅2 0.76 0.76
Observations 190 190
Time fixed effects Yes Yes
27 The net interest margin is significantly negative related to nonperforming loans. This relationship is straightforward, as the relationship between interest margin and the assets where this margin is derived from, is negatively related. Next to that, overhead costs have a strong positive relationship with nonperforming loans. Overhead costs are positively related to nonperforming loans, as the costs rise with nonperforming loans and vice versa.
Stock market volatility contributes positively tot nonperforming loans, which reduces the lenders (firms) capacity to repay loans and installments as uncertainty rises and firms have reduced ability to rollover their debt positions (Klein, 2013). Next to that, unemployment has a significant, positive effect on nonperforming loans. Higher unemployment rates are a sign of a decline in economic conditions. Higher
unemployment decreases firms’ and other lenders’ ability to earn sufficient capital to pay their debts, which in turn has a positive impact on nonperforming loans.
Asset liquidity does not seem not seem to have a strong significant importance on nonperforming loans. While liquid assets do have a significant, negative relationship with nonperforming loans, the magnitude of the sign is not strong. Liquid asset are of importance when it comes to being able to easily and quickly sell of assets when systems become distressed. Higher asset liquidity could mitigate the effect of lower asset quality, as here displayed by nonperforming loans.
Last, concentration of the financial system and GDP growth do not have a relationship with nonperforming loans. While GDP growth has a negative relationship with
nonperforming loans, it has no significant impact where a significant relationship is expected. When GDP growth becomes negative or declines, economic conditions worsens and households and business are increasingly incapable of paying their
installments and principals which increases nonperforming loans. There is also evidence that interbank lending has a significantly negative impact on nonperforming loans.
28
5. DISCUSSION & CONCLUSION
This chapter starts with a discussion of the results, followed by the conclusion. This chapter ends with the limitations and recommendations for further research.
5.1 DISCUSSION
Literature has shown that foreign banks tend to strengthen the financial stability of host countries (Barth et al., 2002) and can aid countries in their financial development
(Claessens et al., 2008). Yet, it is difficult to find to find out what impact foreign banks have during and after a financial crisis. The evidence is supporting that foreign banks tend to weaken financial stability of the host country after the crisis.
Financial deregulation has increased the foreign bank penetration in financial systems in the past decades. This study finds that foreign banks account for an average of 44,56%
of the total banking assets in European countries. While foreign banks are important for financial systems and possess relative large portions the banking assets, foreign banks do not seem to have a significant relationship with the systemic risk of a financial system.
This study finds that foreign banks also have no significant relationship with asset quality. While foreign banks have a negative relationship with nonperforming loans and thus seem to improve asset quality, this relationship is not significant. The result is in accordance with Goetz et al. (2016), that show that foreign banks do not affect overall loan quality. This relationship confirms that the results of the Z-Score is robust, as the financial stability indicator, in accordance with the systemic risk, shows no significant relation with foreign banks participation.
These results have been obtained during one of the worst and most eventful financial crisis of modern history. The results and outcomes should be interpreted with care. The sample considers European countries which may have different legislation and other legal frameworks than other regions. Furthermore, the time period of 2007-2013 is a period of crisis which may be in contrast of periods of economic stability. The research and outcome are, however, valid and relevant as its demonstrates the impact of foreign banks in vulnerable economic periods in time.
29 This study has examined the relationship between foreign bank presence and financial stability. It has examined this while disregarding important aspects of financial stability, namely banking supervision and (local) banking regulation. This research has not been able to control for banking regulation in the dataset, as accurate data is difficult to acquire and incorporate in the dataset. Boudriga, Taktak & Jellouli (2009) show that the absence of the relationship between bank supervision and regulation and banking outcome. Bank regulation and supervision have no significant impact on banking outcome and financial stability, which indicates that the results in this study are robust to banking regulations.
5.2 CONCLUSION
This study uses data from the World Bank and the Dutch Central Bank to examine the relationship between foreign banks and financial stability, utilizing the systemic risk measure of the Z-Score and financial stability indicator of nonperforming loans. The sample consists of 30 European countries in the time period of the financial crisis (2007- 2013). To examine the impact of foreign banks, foreign banks is measured as the ratio of foreign bank assets among total bank assets in an economy.
The results imply that foreign banks do not have a significant impact on the systemic risk of a financial system. While foreign banks have a negative relationship with the Z- Score, it is of no significant importance. Furthermore, foreign banks also have a negative relation with nonperforming loans and thereby increase asset quality. Overall, the results suggest that foreign banks do not have a significant impact on financial systems after the crisis in terms of systemic risk. These results have to interpreted with care, as the countries in the sample have been in a period of financial crisis and the financial systems as a whole may still be unstable and the results show no significant relation between foreign banks and stability.
The results imply that policymakers should be wary of the effects of foreign banks.
While the results show that foreign banks lower the Z-score and thus increase systemic risk, the relationship is not significant. Furthermore, foreign banks increase asset quality by decreasing nonperforming loans. Also here, the relationship is not significant. In the period of the global financial crisis, foreign banks were not a stable source of stability (De Haas & Van Lelyveld, 2014). This result, combined with the findings of this paper,
30 shows that policymakers may try to maximize the total amount of foreign bank assets in their financial system.
5.3 LIMITATIONS AND FURTHER RESEARCH
The main limitation of this study is the dataset. This study considers the period of 2007- 2013, which is during and after the global financial crisis. This period could be
characterized by an overall, global economic distressed period of time. This could be an issue when comparing the outcomes of this study to other periods of time, for example without (global) crisis.
There is also the issue of endogeneity. Endogeneity is basically always a concern in empirical research. To overcome the omitted variable bias, this paper has selected relevant covariates. However, omitted variable bias is still a concern in this paper.
Furthermore, it is difficult whether the observed relationship in this paper are causal. It might be that the Z-Score may actually impact foreign bank participation, which would make foreign bank presence endogenous. Unfortunately, no suitable instruments are available in the data to address this issue.
Also, this study considers European countries. As mentioned by Claessens & Van Horen (2015), country heterogeneity is an issue in researching financial stability. This could be an issue when attempting to apply the results of this study to other regions and
continents. As other regions may have different characteristics, the results may not (fully) apply and foreign banks may have a different effect on financial stability, especially in developing countries (Crystal et al., 2002).
Secondly, several indicators are not reported by countries. For example, value at risk and derivative holdings are not reported for some countries on a continuing basis. For a valid comparison between countries or regions, these numbers need to be available.
Next to that, more extensive financial data would make cross-country comparison on financial health more comprehensive. Data on stress tests for banks (on the country level) and other sensitivity tests would make comparison on financial status and foreign banks more comprehensive. For a complete overview of the financial sector and
financial health, these indicators need to be available on a country level.
31 For further research, balance sheet contamination is an interesting aspect to study. In new research, foreign bank presence could be linked towards balance sheet
contamination and the impact foreign banks may have towards financial crisis and the possible manners in which foreign banks could affect financial stability. Next to that, crisis amplification is another interesting aspect. The various aspects on the manners foreign banks may amplify or dampen a financial crisis or financial structure provides new insights on foreign bank impact.
The bank ownership database of Claessens & Van Horen (2015) also consists of data regarding the foreign ownership and the home country of the majority shareholder. For further research, it would be interesting to consider the effects of the home country of foreign banks on financial stability and the financial characteristics.
