University of Groningen Faculty of Economics and Business Master Thesis The Effects of Foreign Bank Presence on Banking System Efficiency

(1)

25

University of Groningen

Faculty of Economics and Business

Master Thesis

The Effects of Foreign Bank Presence on Banking

System Efficiency

Groningen 2011

Nándor Nagy (s2076799)

dr. Robert C. Inklaar

(2)

1

The Effects of Foreign Bank Presence on Banking System

Efficiency

Abstract

This thesis investigates the effects of foreign banks on the efficiency of the banking sector in a long time period and also try to find answer for the role of financial development in terms of the strength of the effects. The data covers 20 years from 1990 to 2009 and 29 countries (15 developed and 14 transition countries). A strong support was found that foreign bank presence is associated with higher level of efficiency and this effect is stronger for countries with lower levels of financial development. Competition and spillover effect that was brought by foreign banks were beneficial for host countries.

(3)

2

1. Introduction

This thesis investigates the effects of foreign bank presence on the efficiency of the host country’s banking system. This topic is stemming from the widely researched area of financial development and growth. Majority of the publications from this field found evidence that financial development cause economic growth1. Since higher level of

economic growth is preferred, this fact leads to the question that how can a country reach a higher level of financial development? There are multiple strategies for policy makers in a country to gain the desired effects of a higher level financial development. One way is to raise the efficiency of the overall banking system. In this research I interpret financial

development as the banking system development although financial system comprises from institutions other than banks (insurance companies, investment funds, stock market etc.). Efficiency rise can be aimed by liberalizing the financial (banking) markets which policy leads to a raised number of foreign banks in the sector. The mechanism through these foreign direct investments enhancing the performance of the sector is the same as in any other sector of the economy. “One is that the investments make the market more competitive, eroding monopoly power of local firms. Another is that the investment might arise the productivity of local firms through some spillover effect. This will happen if increased competitive pressure induces firms to reduce internal inefficiencies, or if there are direct knowledge spillovers or learning effects by the channels of raised competition or spillover effects.” 2 According this foreign banks can have efficiency enhancing effect on the domestic banking system. Several papers investigated this question and found evidence that foreign bank presence is associated with higher level of efficiency. Most of these papers analyzed the foreign bank effects on domestic banks and not on sector level so it is interesting to find out what happens if the whole banking system is the subject of the analysis.

There is a huge gap between the ratio of foreign banks in developed and developing or transition countries. The share of foreign banks are much higher in developing/transition countries than in developed countries (average share in developing ones is around 60% in developed ones is below 40%)3. This suggests that probably higher profit ratio attracts foreign banks to start business in market environments where competition is not that intense yet. That is why it is reasonable to suppose that the effects are different for

countries with distinct financial development and so income levels. Researchers4 who were interested in that how does the financial development and income level determines the effect of foreign banks concluded that in a short time period at lower level of financial development foreign banks are associated with stronger effects than at higher levels. The database is constructed of mixed income level countries to discover these diverse effects.

1

Levine, Loayza and Beck (2000), Levine (2001), Rioja and Valev (2007), Akimov, Wijeweera and Dollery (2009)

2_{Navaretti and Venables (2006)} 3

author’s calculation

(5)

4

In the following the thesis introduces the relevant literature on the topic and define the concepts according which the hypothesizes were set up. This is followed by a discussion and description of the database where the efficiency indicators and other variables are

presented. Then the applied methodology is introduced which is followed by the estimations and the discussion of the results. In the end a brief conclusion, the bottlenecks of the model and some implications are presented.

2. Literature review

2.1 Financial system and efficiency

In this thesis I investigate the changes in the financial systems in response to the presence of foreign institutions. I am doing this by focusing on the efficiency change of the banking sector in relation to the presence of foreign banks. In the past thirty years liberalization had a great and not questionable effect on the competition among economic actors and this is true for the liberalized financial sector also. Liberalized financial markets will lead to greater cross-border competition, this competition will lead to greater efficiency and a crowding out of the less efficient financial firms (Claessens et al. 2001, Berger et al. 2000, Hasan et al. 2000). A financial system consists of institutional units and markets that interact, typically in a complex manner, for the purpose of mobilizing funds for investment, this typically involves transforming and managing risk and providing facilities, including payment systems, for the financing of commercial activity5. A financial system is built up from three parts: 1.) financial intermediaries (banks, insurance companies and other actors that are mobilizing savings to investors), 2.) financial markets (the place where borrowers and lenders meet, like money market and stock exchange), 3.) financial markets infrastructure (the physical system through which the transactions happen)6.In this thesis I do not investigate the changes of the whole system but particularly the effects on financial intermediaries, among them only the banks are going to be analyzed. As from the definition a more developed banking system means a better system that can better allocate and mobilize savings, ease risks and facilitate transactions. It can reduce adverse selection7 and moral hazard8 which can lead to banking crises Mishkin (1996). In this sense financial development (referred as banking system development in the later) is crucial for a country’s economic performance because financial development is the capability of one country to channel savings into investment efficiently and effectively within its own borders owing to (i) the quality of its institutional and

regulatory framework, (ii) the size of its financial markets, the diversity of its financial instruments and private agents’ ease of access to them and (iii) the financial markets’

5

The OECD Glossary of Statistical Terms

6_{European Central Bank, Financial Stability Report (2010)} 7

Definition: „Adverse selection is an asymmetric information problem that occurs before the transaction occurs when potential bad credit risks are the ones who most actively seek out loan.” Mishkin (1996)

8_{Definition: „Occurs after the transaction occurs because the lender is subject to hazard that the borrower has}

(6)

5

performance, e.g. in terms of efficiency and liquidity9. From the perspective of this thesis the third point (efficiency) is the substantive factor. Efficiency is defined as “A level of

performance that describes a process that uses the lowest amount of inputs to create the greatest amount of outputs. Efficiency relates to the use of all inputs in producing any given output, including personal time and energy”10. Efficiency is strongly linked to financial development. In this work I am measuring the changes in efficiency of the banking sector in response to foreign bank presence. The system that works with higher efficiency can gain higher performance so it can reach a higher level of development. Beside the above mentioned reasons why financial development is important there is another substantial reason why countries should pay attention to their financial systems. It has been argued by many researchers that financial development can be the cause of economic growth.

Countries with a more developed financial system more likely to reach higher level of economic growth in the long run. Although some evidence was found the question is still opened, if there is any positive and significant relationship between financial development and economic growth?

2.2 Financial development and economic growth

According to Levine, Loayza and Beck (1999) yes there is. In their paper they were especially looking for casualty links with a cross country and panel data regression, the two kinds of regressions had the same results “financial intermediary development exerts a statistically significant and economically large impact on economic growth”11. In one of their later paper Levine, Loayza and Beck (2000) finds that better developed legal and accounting systems lead to more developed financial systems which foster economic growth. Levine (2001)finds further evidence that by international financial liberalization letting foreign banks in the country will enhance the efficiency of the domestic banking system and this better developed banking system can foster economic growth through productivity growth.

Eschenbach and Francois (2004) investigated the link between financial sector openness and financial sector competition and linked this to economic growth. They found a string positive link either between the openness or competition and economic growth. Rajan and Zingales (1998) have the same conclusion from a different perspective. They investigated this link through the industrial sector which is more dependent on external financing so a better developed financing sector can boost the economy by easier access to resources. Other studies do not always find a positive relationship between financial development and growth. Rioja and Valev (2004) found that financial development does not always have a positive effect and if it does it varies across the level of development. Shan and Morris (2002) also found meager evidence on the growth impact of financial development, they even questioned if financial development is necessary at all! Arestis and Demetriades (1997)

9_{European Central Bank, Occasional Paper Series, Domestic Financial Development in Emerging Economies} 10

Investopedia

(7)

6

had a negative conclusion as well: “it is by no means universal that financial development can lead to economic growth”. A recent study Lartley (2010) found the opposite results of Rioja and Valev (2004) and had a conclusion that financial development has a positive effect on growth and it does not vary across development levels. Using an endogenous growth model and a panel data analyzing technique over 27 countries Akimov, Wijeweera and Dollery (2009) found a robust positive relationship between financial

development and economic growth.

Another doubtful debate is about the direction of causality, if financial development Granger cause12 economic growth?Chang and Caudill (2005) by using a Granger causality technique found an unidirectional relationship in the case of Taiwan. An opposite opinion of this was carried out by Dawson (2003) who investigated if financial development13 Granger caused economic growth, applied also an endogenous growth model for 13 CEE countries14 he concluded that there were no positive relationship between them which means that CEE countries probably are not constrained because of underdeveloped financial system. These papers found evidence that either financial development caused economic growth or either not but higher development is never negatively associated with economic growth so it is never a drawback for a country to make steps and develop their financial systems especially if a country is a developing or a transition one. In their work Blossome and Promisel (1998) collected the possible ways to improve the financial system. Their suggestions are around eight areas: balancing competition in banking with incentives to create franchise

value, developing capital markets and an investor base, strengthening prudential regulation and supervision, adopting incentive-based safety nets, encouraging self-policing, enhancing production and dissemination of information and strengthening informal finance. In this enumeration the second point is in the focus of this research. According to them by

developing capital markets and investor base the financial system will be stronger and better developed.

2.3 Effects of Foreign Banks

One way to develop the capital markets and the investor base is to open and liberalize the financial markets. This means that foreign capital and financial institutions should be welcomed in the economies. Letting foreign financial institutions in the country means physical presence of those which has two main implications: raising competition and introducing new management, supervisory techniques and new products to the market which can spread over the whole system. These newly introduced techniques can lead to a more stable and efficient banking system.

12

A variable x is said to Granger-cause a variable y when a prediction of y on the basis of its history can be improved by further taking into account the previous period’s x. Yi Wen (2007).

13

He defined financial development with one variable: liquid liabilities to GDP

(8)

7

It is an interesting research area if domestic banking system as a whole can have advantages from the presence of foreign banks. Benefits could stem from different factors (more stabile banking system, stabile credit supply, reducing the possibility of banking crises, reducing the effects of a recession, more efficient banking system and as a result an overall welfare gain for customers). Although high rate of foreign bank participation in the sector have

disadvantages also.

General advantages and disadvantages are summarized in Table 115:

Table 1

Advantages Disadvantages

Bringing new innovation for financial products and new technique for banking

Foreign control of banking assets

Introducing new fields of financing Foreign banks can be interested in promoting their home country projects in host country

More developed financial markets in general so firms do not have to import financial services from abroad

Regulatory differences

Advanced banking infrastructure (regulations, know how, transparency)

Home country effects: "There is significant negative relationship between home country economic growth and host country credit by greenfields" (Haas and Lelyveld 2005)

Attracting other sector FDI

Competitive effects due to raised number of actors

In this table the advantages are around two effects: the general competitive effect and the spillover effects. Competition is the natural attendant of market economy. It is said to improve the effectiveness of the whole economy Boone (1998) because: reduces total industry profits, reduces the profit level of the least efficient active firm and increases that of the most efficient one, increases total industry output and reduces the output level of firms that feature much higher costs than the leader. These features absolutely meet with the factors (input, output) that determine efficiency. Raised competition should enhance efficiency in the banking sector as well. Spillover effect also has efficiency raising effect but through a different channel. Knowledge, knowhow and technological experiences can be exchanged between companies on purpose because they are making business together so it is intended. If this knowledge transfer is not intended it is spillover Fallah and Ibrahim (2004). Jaffe (1986) wrote if knowledge is accumulated by one firm it tends to help the

(9)

8

development of other technologically close firms. According to this if foreign banks enter to a market with a technique or knowledge which was not present on the market dissipation can happen to local banks that helps them to raise their efficiency by reducing costs or selling new products in the markets. Eller, Hais and Schnitzer (2006) analyzed efficiency gain from this perspective. They distinguished between spillover and competitive effects, they found that these two effects are hardly existing together because higher competition

reduces the likelihood of spillover. Distinguishing between the two effects in practice is hard and not the purpose of this thesis, efficiency gain can stem from either effect.

Disadvantages are centered around the fear of credit volatility thus the defenselessness to crises periods as Anderson and Kegels (1998) said “Liberalizing the banking sector is unlike liberalizing other economic activities banks are the conduits of monetary control.” This thesis is focusing on the first issue that through competitive or spillover effects which was brought by the foreign institutions a better efficiency can be reached which leads to better performance of the sector. The second issue is important as well although it is not part of this research. One part of research papers are interested in if foreign ownership can be linked to a more stabile banking system which guarantees continuous credit supply during recessions or banking crises. These papers find that the fear of foreign banks effect on credit volatility is baseless. Demirguc-Kunt, Levine, Min (1998) have the conclusion that “foreign bank participation lowers the probability that a country will experience a banking crises”16 . Concerning these theories and findings we can conclude there is enough evidence that despite the possible threats foreign banks can play a stabilizing or moreover a developing positive role in the economies.

2.3.1 Efficiency of the banking system

Stabilizing effect is just one side of the coin. The overall performance of the whole financial sector is at least as important if not more as the stabilizing effect. As in the definitions of financial development and efficiency a better performance is associated with higher level of efficiency. Is the presence of foreign banks in a particular economy lead to higher efficiency of domestic banking sector as a whole? Demirguc-Kunt, Levine and Min (1998) found foreign bank presence leads to lower overhead costs and profits of banks (lower values of these indicator implies higher level of efficiency), further they also link this efficiency gain to economic growth and they conclude that foreign banks accelerate overall economic growth by boosting efficiency in the banking sector. Eller, Hais and Schnitzer (2006) have the same conclusion that letting foreign banks in the country will lead the economy from a high fixed and transaction cost station to a lower cost station which is interpreted as higher level of efficiency so countries should let foreign banks in the countries. Claessens, et al. (2001) have a result that larger share of foreign banks put a downward pressure on local banks’ interest margin and profitability so enhancing efficiency of the whole system. This efficiency gain

16

(10)

9

leads to higher level of crediting to small and medium sized enterprises which boosts the economy as well. Clarke, et al. (2006) also finds that foreign banks tend to lend small enterprises more than local banks. De Haas and Naaborg (2006) also have the positive conclusion for small enterprises. Fries and Taci (2005) argues there are three main factors that promotes cost efficiency: lower nominal interest rates, higher share of foreign banks in total number of banks and higher intermediation ratio. Claeys and Hainz (2006) analyzed 10 eastern European countries and found that foreign bank presence is associated with lower interest rates. Drakos (2003) investigated the impact of the transition process on 283 banks in 11 CEE countries over the time 1993-1999 pooling this to a panel dataset. One of the two links that he proved was the impact of foreign bank entry on the interest margins of the bank systems. He used the interest margins of foreign banks as dependent variable and liquidity risk, default risk, interest rate risk and leverage were his explanatory variables. He proved that foreign bank presence leads to an increased competition which has a positive impact on sector efficiency which was measured by interest margins of the system (lower level is more efficient). McFadden (1994) found that the liberalization of the Australian financial markets caused a very aggressive response from the Australian domestic banks. They started to cut on costs and invested in new technologies to maintain their market shares. As a result foreign banks only managed to acquire less part of the market than it had been expected. Uibopin (2004) used information of 219 banks from 10 CEE countries

between 1995 and 2001. This paper investigated the link between the performance of the banking system and foreign bank presence in a short time period. He found that foreign bank entry increases the competition and causing a rise in the overhead costs, this means that efficiency is reduced. In the same time raised competition caused a decrease on revenue and interest margins. Papi and Revoltella (2001) investigated the link between financial sector foreign direct investments and efficiency level. In their database there were 112 bank from 9 transition countries between 1993 and 1997. They defined efficiency as the level of ROA17 and overhead costs. They found that foreign ownership is positively linked to profitability and cost efficiency. Zajc (2004) investigated the link between foreign bank entry and their impact on the performance of the banking industry in 6 CEE countries in the time period of 1995-2000. Their dependent variables were: interest margins, non interest income, profit, overhead costs and loan-loss provisions. He has the same conclusion as Uibopin (2004) that foreign bank presence spurs the competition which leads to higher overhead costs but lowers net interest rates, income and profit. Demirguc-Kunt Huizinga (2000) used a database which included information on banks in 44 countries over the period 1990-1997 including developed and transition countries also. They presented empirical evidence that underdeveloped banking systems have higher profits and net interest margins so greater level of banking system development reduce the profits and margins.

(11)

10

Although these findings are consistent in that foreign bank presence is associated with higher levels of efficiency and through financial development the case is not that simple because of two reasons. First these papers were mainly investigated the links in developing or transition countries and not in developed countries. The effects are different in developed and developing countries. Second there are two ways for a foreign bank to start operating in a country. It can either break into the market via new investments (greenfield investment) or via buying up a local institution (M&A) or just buying shares in it. Although the different effects of the modes entry is not subject of investigation and therefore also a limitation of this thesis but it is important to highlight that the effects can be modified because of the mode of entry. The most important difference between the two modes is their effect on competition. In the case of greenfield investment there is an extra actor in the market that spurs competition better than an M&A where the number of actors are constant. A

greenfield investment can have a greater effect on the efficiency of the banking system18. 2.3.2 Different effects for countries with different income level

The effects of foreign banks on efficiency is said to be different in countries with different income level. First of all it is important to highlight the fact that foreign bank penetration is much higher in developing or transition countries than in developed counties. Claeys and Hainz (2006) emphasize that foreign bank market share was above 55% in new member states in 2003 this ratio is very low for the EU 15. This is surprising because there are no restrictions on market entry in neither of the countries. Effects are different for countries with levels of financial development. Countries with a better developed financial system can reach faster economic growth so they will have higher income levels as well. This idea is derived from the theory of financial development and economic growth where financial development spurs economic growth so countries with higher level of development possess higher level of income due to the faster economic growth. Different market environments react differently for the presence of foreign banks. Demiguc-Kunt and Huizinga (2000) found evidence that banks have higher profits and interest margins in underdeveloped bank systems which are in developing countries. Their results also suggest that greater

development of the banking system lowers the profits and interest margins of the banks. “Greater bank development brings about tougher competition, higher efficiency and lower profits”19 . This implies that foreign banks must have different effects in markets with different levels of financial development. Lensink, Meesters and Naaborg (2008) supports this theory. They investigated foreign bank efficiency in countries with different qualities of institutions. Although their findings are not consistent with the mainstream theory because they found that foreign banks are less efficient in general but they concluded that foreign banks are more efficient in countries with a higher quality of institutional system. Hermes and Lensink (2004) investigated the effects of foreign banks in 48 countries over 1990-1996.

18_{Literature that investigates the different effects and purposes of the mode of entry in detail are: Haas and}

Lelyveld (2006), Peria, Soledad and Mody (2004), Claeys and Hainz (2006)

(12)

11

They found that foreign bank presence reduced the efficiency of domestic banks. They also investigated the strength of the effect according to the financial development level. They have the conclusion that “foreign bank presence is associated with higher costs and margins of domestic banks at lower levels of financial development, while foreign bank presence is associated with falling costs and margins of domestic banks at higher levels of financial development”. The effect in a short time period is efficiency destroying in developing or transition countries but efficiency enhancing in developed countries. Uibopin (2004) has exactly the same conclusion as Hermes and Lensink (2003). Claessens, Demirguc-Kunt and Huizinga (2000) researched the same issue in their paper. Their results are the opposite of Hermes and Lensink (2003) because one of the main findings of the paper is the reduction in profitability and interest margins of the domestic banks in response to larger share of

foreign bank presence. One possible explanation for the contradiction is the different

databases they used. Claessens, Demirguc-Kunt and Huizinga (2000) had a longer and wider database (1988-1996 and 80 countries). That is why their database supports the theory in a long time period “foreign bank entry improves the functioning of domestic banking

markets.” Another main finding of the paper is that foreign banks have higher profits and interest margins in developing countries while the opposite is true in developed countries. The findings of these authors obviously suggests that foreign banks have different effects in countries with different levels of financial (economic) development. In more developed markets the competition is more fierce so banks have already reduced costs and profits that is why the effect of foreign banks is not that strong.

(13)

12 Table 2 Authors, year of publication Coverage of database (countries, years)

Dependent variables Explanatory variables Mayor findings

Bonin, Hasan and Wachtel (2005) 6 CEE countries, 1994-2002 ownership status: foreign greenfield, domestic de novo, state owned, privatized

ROA, interest margin, commission income, costs, loan, deposit, liquid assets, equity and loan loss provision ratios

The most efficient banks in terms of profit and costs are the foreign greenfield banks while the least efficient ones are the state owned ones. Timing and mode of

privatization matters. Claessens Demirguc-Kunt and Huizinga (2001) 88 countries, 1988-1995 ownership status: foreign and domestic, developed or

developing countries

interest margin, non interest income, before tax profits, net profit, loan loss provision, overhead costs

Foreign banks are more profitable in emerging markets. The increase in foreign bank share leads to a

reduction of profitability and interest margins of domestic banks so enhancing the systems efficiency but possibly crowding out the least efficient domestic banks

Claeys and Hainz (2006) 10 CEE countries, 1995-2003 lending rate, ownership dummy and mode of entry dummy, interest rate

foreign bank share and control variables for macroeconomic and financial development conditions

Interest rates decreased after foreign bank entry, this fall is more

pronounced in the case of greenfield investments. Foreign de novo banks charge higher interest rates than acquired banks. Clarke, Cull, and Peria (2006) 35 developing countries, 1995-1998

interest rate, access to long term loans

Foreign bank

participation, GDP, M2, GDP growth, bank concentration, controls for enterprise type

Falling interest rates on long term loans are associated with foreign bank participation. Foreign bank presence diminishes the financial burdens of small and medium sized enterprises

Demirguc-Kunt and Huizinga (2000) 44 countries, 1990-1997 before tax profit/total assets, net interest margin/ total assets

equity/ta, loan/ta, GDP/ca, growth, inflation, tax rate, bank credit/GDP, bank assets /GDP, central bank assets/GDP, stock market cap…

Banks have higher profits in underdeveloped banking systems. Greater bank sector development lowers the profits and interest margins of banks. The reason is the tougher competition that increases efficiency and lowers profits. Stock market development also plays an important role in the same way.

Demirguc-Kunt Levine and Min (1998)

80 countries, 1988-1995

before tax profit/ta, overhead costs/ta

Foreign bank share, equity/ta, non-interest earning assets/ta, overhead/ta, GDP/ca, growth, inflation, real interest rate, costumer funding/ta

Foreign bank share is associated with reduced profits and overhead costs. The number of foreign banks counts not the their share in the total banking assets. Indirect link to economic growth. Foreign bank share per se is not associated with economic growth but the efficiency gain that they brought is.

Drakos (2003) 11 CEE countries, 1993-1997

interest margin rates of foreign, domestic and state owned banks

liquid assets/tl, loan loss provisions/loans, net short term

assets/equity, equity/ta

(14)

13 Authors, year of publication Coverage of database (countries, years)

Dependent variables Explanatory variables Mayor findings

Eller, Hais and Schnitzer (2006) 11 CEE countries, 1996-2003 Financial sector FDI/GDP or /human capital, cross border financial sector M&A

GDP growth, GDP, capital stock, human capital stock, employment, government

spending/GDP, total FDI inflow,

There is relationship between FSFDI and growth but the effect is conditional on human capital stock. Foreign owned banks have impact on financial sector development.

Fries and Taci (2005)

15 CEE countries, 1994-2001

total costs, cost efficiency

country level factors, banking reform index, foreign banks interest rates

More foreign assets market share is associated with lower interest rates. Cost efficiency is not linearly related to reforms, it is greater in the early stages of the reforms. Foreign privatized banks are the most efficient.

Hermes and Lensink (2004)

48 countries, 1990-1996

Net interest margin, profitability,

overhead costs, loan loss provision, non interest income

Foreign bank number, private credit/GDP, GDP/ca, inflation, equity/ta, overhead costs, GDP growth, short & long term deposits

Foreign bank presence is associated with higher costs and margins of domestic banks at lower levels of financial development, whereas the opposite is true at higher levels of financial development. Profits are reduced but all the other dependent variables increase. Lensink, Meesters and Naaborg (2008) 105 countries, 1998-2003

total costs, price of funds and price of labor, ln TC/price of labor

ln (pf/pl), ln(loans), ln(securities), interactive variables of this, year dummy, nominal interest rate, private credit/GDP, loan loss reserve/gross loan

Foreign banks are less efficient in general but the degree depends on the quality of the institutions. The better the quality of the institutions in a country the higher the efficiency of foreign banks. Papi and Revoltella (2003) 9 transition economies, 1995-2002

ROA, overhead costs/ total assets

total assets, net loans/total assets, operating income/net interest revenue, ownership dummy

Foreign ownership is positively associated to profitability. To improve operating efficiency foreign majority of shares are required for cost efficiency.

Uibopin (2004) 10 CEE countries, 1995-2001

Interest

income/interests earning assets, ROA, non-interest income/ta, total operating expenses/ta, loan loss provision

foreign bank number, foreign bank market share, private

credit/GDP, equity/ta, GDP growth, GDP/ca, inflation, interest rate

Foreign bank entry is associated with lower before tax profits, non-interest income and loan loss provision but with higher costs. Foreign bank entry enhances competition on the market. The more developed the banking market the less foreign banks associated with decreasing income and loan loss provision.

Zajc (2004) 6 CEE countries, 1996-2000

interest margin, non-interest income, before tax profit, overhead costs, loan loss provisions

foreign bank number and share, equity/ta, non interest earning assets, customer and short term funding, country dummies, GDP/cs, GDP growth, inflation

(15)

14

3. Hypotheses

H1: Foreign bank presence is positively correlated with the efficiency level of the banking

system in a given country.

The goal of this thesis is to discover the effects of foreign bank presence on the efficiency of the banking systems. Related literature suggets that foreign bank presence is efficiency enhancing. The finding ofDemirguc-Kunt, Levine and Min (2001), Claessens et al. (2001), Claeys and Hainz (2006), Drakos (2003) are consistent with this theory. Even though they all used different databases, techniques and efficiency measurements to estimate the effects they found a positive relationship between foreign bank presence and efficiency of the domestic banks. It is important to highlight that these papers investigated the effects of foreign banks on domestic banks and not on systemic level. Their findings give the idea that foreign banks influences not just the performance of domestic banks but the whole sector. In this interpretation the banking system consists of domestic and foreign banks (number of state owned banks are limited), the efficiency growth of one group enhances the efficiency level of the whole system. Furthermore according to the international trade theory only the most efficient firms are going multinational20. This implies that foreign banks in a given country have higher rate of efficiency which also raises the overall efficiency of the banking system.

H2: The strength of foreign banks’ effect is negatively correlated with the financial

development of a given country.

The more developed the banking system the less effect of foreign bank presence has on the efficiency of the sector. Effects of foreign banks are not the same for developed and

transition countries. Developed countries have more developed banking system than transition countries. This fact claims that a more developed banking system is functioning with higher level of efficiency already. So in the case of developed countries foreign banks arrive to a market where the competition is already intense and the technology is already at high level so efficiency indicators are already low. This means that their effect will not be that strong or even opposite, reducing the efficiency of the sector. The opposite is the case for transition counties. Their banking system is less developed so less efficient. In this way foreign bank presence can have a greater effect on system efficiency through competition or spillover effect. In a more developed market banks have already shifted their activity

towards non-interest income because of the risen competition on the market that is why foreign bank presence cannot push this indicator lower that much. The same is true for overhead costs. In developed economies as a result of competition the costs are already

(16)

15

reduced and necessary investment have already been made in the infrastructure so the effect cannot be that strong. The findings of Hermes and Lensink (2004) gives support for this hypothesis because: “At lower levels of financial development foreign bank presence has a strong effect on domestic banks…”.

4. Data and Methodology

4.1 The database

In this thesis I am investigating the relationship between foreign bank presence and the efficiency of the local banking system as a whole. For this I am using a panel dataset which includes aggregated data for 29 European countries over 20 years 1990-2009. There are 15 developed countries (Austria, Belgium, Denmark, France, Germany, Greece, Italy,

Luxemburg, Nederland, Norway, Portugal, Spain, Sweden, Switzerland, United Kingdom) and 14 transition countries (Albania, Bulgaria, Croatia, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Russia, Slovakia, Slovenia and Ukraine) in the dataset. With this partition I can control for different effects of foreign bank presence in developed and

developing countries which is said to be different according to the literature.

In the following I am describing my outcome and explanatory variables and providing a description of the data as well.

4.1.1 Dependent variables21

These variables measure the efficiency of the banking system:

Bank overhead costs/total assets: Accounting value of a bank's overhead costs as a share of

its total assets.

ROA: Average return on assets (Net Income/Total Assets)

Net interest margins: Accounting value of bank's net interest revenue as a share of its

interest-bearing (total earning) assets.

Researchers measure the efficiency of a bank with a number of variables and the same measurements can be used on aggregated level as well. The most common one is the costs to revenue ratio, it is the cost required to generate each dollar of revenue.22 I would not use this one because this is too general and can change due to many reasons. Instead of this I am choosing a cost efficiency, a profit efficiency ratio and a third kind of indicator which denotes that how much income does a bank have from their lending and borrowing activity relative to their interest earning assets. I am doing this because according to Berger and Mester (1997) both cost and profit efficiency is needed to be investigated to discover the efficiency of the system. Cost efficiency gives a measure of how close a bank’s cost is to what a

21

Source: The World Bank, Financial Structure Indicators, updated in 2010

(17)

16

practice bank’s cost would be for producing the same output bundle under the same conditions23. Profit efficiency measures how close a bank is to producing the maximum possible profit given a particular level of input prices and output prices24. According to these I chose bank overhead costs/total assets for cost efficiency ratio and ROA for profit efficiency ratio.

Efficiency is often measured in terms of spending on overheads relative to the amount of earning assets Berger and DeYoung (1997) so reducing overheads are a desired goal for the institutions. This is true for system level also, the lower the ratio of bank overhead

costs/total assets the more cost efficient is the system working. “Imperfect and high overhead costs can signal excessive managerial perquisites and market power that

contradict the notions of sound governance of banks and efficient intermediation.”25 Lack of competitive market can be the reason behind high levels of overheads. That is why I expect to have this ratio lower for developed countries and higher for transition countries at least in the early years. Demirguc-Kunt, Levine and Min (1998) found evidence that foreign bank presence is related to lower values of overhead costs. Claessens et. al (2001) found evidence that costs of the banking system are falling due to foreign bank presence but remarks the possible threats of crowding out the least efficient domestic banks. Uiboupin (2004) in the same time investigates the effects of foreign banks on banking systems with the help of a database that covers a shorter time period. One of his hypothesis was that overhead costs are increasing because banks have to invest to maintain their market share due to the raised competition. Although he found meager evidence for this assumption and concluded that foreign banks generally enhances competition in the market Figure 1 supports his idea. The figure shows the average bank overhead costs/total assets of developed and transition countries for each year. In the early years overhead costs for transition countries raised at a significant26 rate and it started to decline in 1999 but it never reached the overhead costs level of developed countries. This could happen because of the layoffs that foreign banks did in their acquired banks or because of the higher costs to break into the market for greenfield foreign banks Haas and Lelyveld (2005). Banks in developed countries are still operating with a higher cost efficiency ratio this supports the idea that foreign banks can have less effect on overhead costs in developed countries. In all bank overhead costs should be negatively correlated to foreign bank presence.

23

Laeven (1999)

24_{Laeven (1999)} 25

Barth, Caprio and Levine (2005)

(18)

17

Figure 1

Source: author’s calculation

The second dependent variable that measures efficiency is ROA which is the most popular efficiency measure in the literature. This variable captures the level of costs that the banking system is working with and also the income producing ability. So it is a very important

indicator also for the institutions and they attempt to raise it as high as it is possible. In the case of reduced competition this ratio is higher because banks can sell their products at a higher price so the higher the ratio the less efficient the system is working. For the first sight it might be strange because everyone would expect higher profits from a bank that works with higher rate of efficiency but in my interpretation this higher profit reflects to lack of competition in the market and not to cost efficiency. Hermes and Lensink (2004) used ROA as an efficiency variable and found negative relationship with foreign banks and return on assets at domestic banks. Claessens, Demirguc-Kunt and Huizinga (2000) found that foreign banks have higher profits than domestic banks in developing countries and the opposite is true for developed. This means that profit rate in general should be higher in developing countries. Figure 2 give evidence to this idea. Although transition countries are less developed the return on assets in their banking systems are higher than in developed

countries (except the crises period27 from 1999-2001 where is a deep trough). This cannot be because of better cost saving techniques but can happen due to the fact that competition in those countries are milder. Although foreign bank’s effects on profitability is not that clear a small increase is expected in profitability but not unambiguously.

27

(19)

18

Figure 2

The third variable measures efficiency is the net interest margins in the banking system. It is an important variable but not clearly a cost efficient nor a profit efficient index. It shows how much profit do banks make with the difference in their lending and borrowing rate. This measure is an important factor of a country’s economic life because lower lending rates are desirable as they have a positive influence on the investments and develop the competition in the market so supporting the development of the economy. That is why this is a widely used variable to measure financial efficiency28. The higher this value the more reason we have to suppose that the banking system is inefficient. Brock and Rojas-Suarez (2000) suggests that high values of net interest margins are usually associated with bank

inefficiency. This can happen due to the lack of competition so the banking institutions can work with higher interest rates which can be an unnecessary burden on the whole system’s efficiency. The averaged data by year on net interest margin is illustrated in Figure 3. The outcome supports the finding of other authors that interest margins are higher in less developed countries. In transition countries it is interesting to see that in 1990 the interest margin levels of the two country groups were almost the same. This could happen due the financial system of the previous regime where banks were all state owned and high interest and profit margins were not a goal. Although the rate levels converged to each other the difference between the two country groups are still around 2%. A modest decline also happened in developed countries. In their research Demirguc-Kunt and Huizinga (1999) used net interest margin as one of their efficiency variable. They found that net interest margins are higher in developing and transition countries while the opposite is true for developed countries. Demirguc-Kunt et al. (2003) drew a conclusion that banks that are operating in less competitive markets are able to work with higher interest margins. Drakos (2003)

(20)

19

investigated the links between foreign bank presence and efficiency. He found evidence that foreign bank presence is associated with lower level of net interest margins.

Figure 3

These three variables are measuring “the pure operational efficiency and the competitive structure of the banking market.”29

4.1.2 Explanatory variables

The key explanatory variable in this research is which measures for foreign bank presence. In this sense there are two kinds of variables that could be used:

· number of foreign banks/total number of banks · foreign bank asset share/total bank assets

I chose not to use both variables due to two reasons. The first is theoretical one. I am

interested in the effects of foreign banks in a given country for which according to Demirguc-Kunt, Levine and Min (1998) the presence itself is more important than the market share. They did not find any significant relationship with foreign bank market share so they concluded that foreign bank presence, per se, spurs efficiency growth. Claessens and Huizinga (2000) did not find any significant results with foreign bank market share either. This means that the channels (competition, spillover) through which the effects are

operating are coming with the presence also. The second is a technical one. Finding data for the number of foreign bank in total number of banks is easier and not all the databases that I used to construct this variable hold information for foreign bank assets as a share of total bank assets.

All the databases that I used calculated their variables from the Bankscope database30. I could not use it because the time period it covers is too short. As a result I decided to use

(21)

20

completed databases by different authors who used the Bankscope database for the same purpose to measure foreign bank presence in a given country. The main sources for

transition countries that I am relying on is the database from Claessens, et al. (2008)31. This paper have a database on the number of foreign banks in total banks in developing countries from 1995 till 2006 including the transition economies from Europe that I need. For the missing periods 1990-1994 and 2007-2009 I used two databases. To find data for the period 1990-1994 I used a dataset from Beck, Demirguc-Kunt and Levine(2001) which is a World Bank database and next to many other financial variables contains data for all countries of the world regarding to foreign bank presence as well. For the period 2007-2009 I used the database of OECD Banking Statistics which has information on the structure of the financial systems of OECD countries in the period of 2000-2009. I also used some control databases to check if the authors calculations were punctual comparing to each other. Other databases32 on transition banking systems which supported my numbers, although the numbers were not precisely the same. If the difference was more than 5% I constructed the ratio from the averaged value of the two databases, this was not typical but happened in some cases. Central bank databases should provide some information about foreign institutions in their countries but usually they do not have precise statistics about foreign bank penetration in their databanks, one good exception was Croatia. Despite this they usually have annual reports where sometimes they mention sometimes not this ratio. Using all these databases finally I managed to set up a reliable dataset for foreign bank presence in 14 transition countries (Albania, Bulgaria, Croatia, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Russia, Slovakia, Slovenia and Ukraine) over the period 1990-2009. The data collection for the developed countries was not much different. For the period 1990-1997 I used the same database from Beck , Demirguc-Kunt and Levine (2001) as for the transition countries for the period 1990-1994. For the period 2000-2009 I used the OECD Banking Statistics. There are two years which are not covered by these two database for each developed countries. To fill this gap I used central bank data and different country reports which were issued by them. For most of the cases I found reliable data that fitted in between the two periods. In developed countries there was not any heavy change in the ratio of foreign bank number to the number of all banks. Figure 4 illustrates the changes in foreign bank on average by year by the two country groups. A great rise in foreign banks among transition countries are obvious but this rise have been stopped by the current financial crisis in 2008 that caused a mild decrease in the ratio. In developed countries33 a slow but continuous rise happened in the past twenty years and the foreign bank share grew from 30% to 40%.

30

Bankscope contains information on more than 30.000 banks worldwide, including ± 6200 European banks for the last 8 years.

31_{Claessens, Horen, Gurcunlar and Mercado (2008)} 32

Cleys and Hainz, (2008), EBRD Structural Change Indicators

(22)

21

Figure 4

4.1.3 Control variables

To avoid omitted variables I need indicators that control for effects that can have an impact on my dependent variables other than the presence of foreign banks. I am using two sets of control variables. One set is controlling for the changing financial structure in the bank system, the other set is controlling for the overall macro-economic environment. I used a further developed version of the World Bank database that is available on World Bank. This database contains all the needed financial structure information over the time for the chosen countries that are in my dataset. The World Bank’s World Development Indicators and Global Development Finance database provide information for all macroeconomic variables that I am using in my research.

Control variables that are measuring the overall development of the banking system: · bank credit/bank deposits: Private credit by deposit money banks as a share of

demand, time and saving deposits in deposit money banks.

· deposit money bank assets/central bank + deposit money bank assets: Ratio of deposit money bank claims on domestic nonfinancial real sector to the sum of deposit money bank and Central Bank claims on domestic nonfinancial real sector. · deposit money bank assets/GDP: Claims on domestic real nonfinancial sector by

deposit money banks as a share of GDP.

· private credit by deposit money banks/GDP: Private credit by deposit money banks and other financial institutions to GDP.

(23)

22

amount of its deposits at any given time. The higher the ratio, the more the system is relying on borrowed funds34. This means that if this ratio is too high the system is more vulnerable to shocks like banking crises and this also can lead to higher interest rates. If this ratio is too low it means that there are too munch unused resources in the banking system which does not imply a well functioning and efficient banking system. The second financial development control variable is the deposit money bank assets/central bank + deposit money bank assets ratio. This is a ratio that measures the relative size of the banking system. The ratio should be changing heavily among the transition countries over the time periods because they used to have a one tier banking system where most of the assets were owned by the central banks and their branches. The ratio in a market economy should be very close to one in order to have a competitive and efficient banking system. The third financial control variable deposit money bank assets/GDP is an absolute size measure that shows the relative

importance of the banking sector comparing to the whole economy. In more developed countries this ratio is higher so this is a substantial measure of the overall development of the banking system. My forth control variable private credit by deposit money banks/GDP, the ratio shows the activity level of the private money banks relative to the GDP. This measures the activity of the sector in terms of its most important function, the allocation of savings to investors. In developed countries this ratio is higher so it is rational to include this variable in the model. Figure 5 shows the change in this indicator. The movement of the lines is almost parallel so the difference did not shrink in the past twenty years. If there is money abundance in the world economy it is going to flow in every country in an equal rate.

Figure 5

(24)

23

Control variables that control for other factors that possibly have an effect on efficiency from macroeconomic point of view:

· GDP growth · inflation35

· real interest rate36

GDP growth is the most commonly used activity indicator for a country. Each country strives to reach a higher level of growth. This is inevitably have an effect on the banking system also. Since higher economic growth reflects a higher demand for money, banks in prosper periods lend more and this increased supply can lead to lower interest rates and increased revenue. These effect could be detected on two of my dependent variables (ROA and net interest margin). Inflation can have an impact on all of my dependent variables. ROA and Bank overhead costs/total assets affected because costs and revenues are sensitive for inflation. Net interest margin is possibly changing in response to a higher level of inflation which was anticipated so banks change their interest rates to realize the same amount of profit. Real interest rate as its own naturally have an impact on net interest margin and ROA. The higher the real interest rate the higher will be the interest rates that the banks can charge on their credits so the nominal value of ROA will raise as well. In this sense both outcome variable will increase if the real interest rate is high reflecting a lower level of efficiency.

My last two variables help me to control for the effects of foreign banks in countries with different income and financial development level:

· dummy variable which takes the value zero if it is a developed country and takes the value one if it is a transition country. Its aim is to spilt up the panel into two

subsections for developed and transition countries.

· private credit by deposit money banks/GDP * number of foreign banks/total

number of banks is an interaction variable that is generated by the multiplication of a

financial development indicator and the foreign bank share in a given country. This term shows how much of the reaction of efficiency on foreign bank presence depends on the level of financial development. If this value is negative for instance but the foreign bank share variable is positive it implies that at higher level of

financial development the effect of foreign banks are weaker than at lower level and vice versa.

35

GDP deflator

(25)

24

The following two tables give a summary statistics about the variables. Table 2 shows the correlation matrix of the variables. High correlations between the variables are highlighted in the table. Macroeconomic control variables are not correlated with any other variables. Foreign bank share variable is only correlated with the interaction variable because it is a component of it. The most correlated variables are deposit money bank assets/ GDP and private credit by deposit money bank assets/GDP. The main reason behind the high correlation is the high value of assets comparing to the GDP which allows deposit money banks to put more credit on the market since the greater value of their assets means greater collaterals for their credits. Since high correlation was detected I tried to omit one or the other variable from the model to check if this changes the results. The outcomes were about the same as including both variables this implies that multicollinearity is not a problem in the model. This technique was applied for the other highly correlated variables and the results were the same. Both of the variables also highly correlated with the bank credit/bank deposits ratio. The higher the assets of deposit money banks and the more they lend the higher will be the bank credit/bank deposits ratio due to the leverage37. Both variables are negatively correlated to net interest margins at a high level. The explanation is the law of supply and demand, the more credit is on the market the lower will be the price of it so banks cannot work with high margins so the correlation between them is high. The last high value of correlation is between bank overhead costs and net interest margins. Possible reason is that banks try to push their high costs to the customers trough high interest rates. Table 3 shows the summary of statistics for the whole sample including all the 29 countries and also separated for the two country groups. The number of observations show that neither of the variables have the maximum amount of data (maximum would be 580) but the available data covers approximately on average the 90% of the sample. Comparing the two tables of the separated groups according to the averages it is obvious that the financial structure and macro level indicators were on a higher level in the past twenty years in developed countries. This cannot be surprising since the decade of 1990 for the transition countries was chaotic because of the transition process. The average of my dependent variables also suggests a higher level of efficiency in the banking sector in the past twenty years for developed countries.

37

(26)

25 Table 2 Foreign bank share Deposit money bank assets/ GDP Bank credit/bank deposits Deposit money bank assets/ central+dep money bank assets Bank overhead costs Net interest margin Return on assets Private credit by deposit money bank /GDP GDP growth Inflation Real interest rate Private credit by deposit money bank /GDP * Foreign bank share Foreign bank share 1.0000

Deposit money bank assets/ GDP -0.0265 1.0000 Bank credit/bank deposits -0.2101 0.5574 1.0000 Deposit money bank assets/ central+dep money bank assets 0.0296 0.3856 0.3706 1.0000 Bank overhead costs -0.1112 -0.3230 -0.0638 -0.1599 1.0000

Net interest margin -0.1459 -0.5645 -0.1607 -0.3027 0.6666 1.0000

Return on assets 0.1058 0.1050 0.0256 -0.0228 -0.0564 0.0001 1.0000 Private credit by deposit money bank /GDP -0.0041 0.9717 0.6223 0.4034 -0.2918 -0.5222 0.1120 1.0000 GDP growth 0.1994 -0.2955 -0.1929 0.0890 -0.1043 -0.1471 -0.0597 -0.2924 1.0000 Inflation -0.0938 -0.1671 -0.1318 -0.2070 0.3067 0.3194 0.0181 -0.1639 -0.1379 1.0000

Real interest rate -0.0161 -0.0723 -0.1268 -0.1489 -0.0469 0.0674 0.0341 -0.1082 -0.1340 -0.3697 1.0000

Private credit by deposit money bank /GDP * Foreign bank share

(27)

25

Table 3

Variable Obs Mean Std. Dev. Min Max

Foreign bank share 550 0,5415007 4,084913 0 96 Deposit money bank

assets/ GDP

506 0,8414637 0,492031 0,100541 2,66468 Bank credit/bank

deposits

538 1,117996 0,535423 0,059602 3,14499 Deposit money bank

assets/ central+dep money bank assets

507 0,9314838 0,122923 0,39498 1,04565

Bank overhead costs 532 0,0398497 0,025485 0,004847 0,25938 Net interest margin 531 0,0393257 0,032156 0,006679 0,3296 Return on assets 534 0,0146111 0,052921 -0,36475 0,929 Private credit by deposit

money bank /GDP

500 0,7085372 0,48352 0,028037 2,61446 GDP growth 575 1,749515 5,440458 -32,1186 13,3 Inflation 575 39,25408 205,5596 -4,02769 3334,8 Real interest rate 535 4,512479 20,56177 -91,7244 374,309 Private credit by deposit

money bank /GDP * Foreign bank share

478 0,7981616 11,47037 0 250,988

Variable developed Obs Mean Std. Dev. Min Max

Foreign bank share 282 0,6781021 5,699981 0,076923 96 Deposit money bank

assets/ GDP

283 1,155382 0,41099 0,375988 2,66468 Bank credit/bank

deposits

270 0,9701 0,052194 0,661888 1,0033

money bank /GDP

(28)

27

4.2 Applied technique

Following my hypothesizes foreign bank penetration affects the efficiency of the whole banking system in a particular country. To set up an empirical model the following equations are used:

1.)

EFFICIENCY

it

= α

it

+ β

1

FOREIGN

it

+ β

2

X

it

+ ε

it

2.)

EFFICIENCY

it

= α

it

+β

1

FOREIGN

it

+ β

2

(FOREIGNFINDEV)*

it

+ β

4

X

it

+ ε

it

Efficiency is the dependent variable in country i in time t. As in the data description country i’s banking system efficiency is measured by three variables: bank overhead costs/total assets, interest margins and ROA. α shows the intercept which shows if all other variables are zero including the foreign banks then what is the level of efficiency. Variable FOREIGNit

shows the share of foreign banks in total number of banks in a country in a given year.

X

it is a

set of control variables that controls for effects that are supposed to have influence on banking system efficiency. The set of control variables has two subgroups one which controls for macroeconomic effects and one which controls for financial structure effects other than foreign banks. Macroeconomic control variables are: real GDP annual growth rate, inflation and real interest rate. Financial structure control variables: bank credit/bank deposits, deposit money bank assets/central bank + deposit money bank assets, deposit money bank assets/GDP, private credit by deposit money banks/GDP. In the second equation

Variable transition Obs Mean Std. Dev. Min Max

Foreign bank share 268 0,3977634 0,276867 0 0,94 Deposit money bank

assets/ GDP

223 0,4430835 0,226057 0,100541 1,18398 Bank credit/bank

deposits

237 0,8874906 0,160132 0,39498 1,04565

money bank /GDP

(29)

28

(FOREIGNFINDEV)*

is an interaction variable that allows to measure the strength of the effect at different financial development level.

ε

is the error term. The first and second equation is used to estimate the effects of foreign banks. With the second equation it is possible to estimate the distinct effect at different financial development level and also to find out if the effects are different in developed and transition countries by splitting up the panel into two subsections with a dummy38.

Fixed effects model was chosen to estimate the model. Its most important advantage is the flexibility because it is able to control for stable characteristics of each individuals (countries) so eliminating a large share of possible bias. With this ability it can explore the relationship between the dependent and independent variables. In this research I suppose that each county has a special characteristic that has an impact on the outcome variables (such as culture, politics, regime, economic freedom). Within effects are supposed to be stronger than the between effects of the cross section units because foreign bank presence have impact in the countries and not between the countries although if there were any cross-border externalities fixed effect estimator could capture that also. Technically it creates a dummy variable for each country and year so letting to control for country and time fixed effects in the same time. Another reason why fixed effect model is going to be used is because of the special characteristics of the sample. Countries of this database were not chosen randomly. Developed countries are all OECD countries and the transition countries are all from the eastern-European region (most of them are OECD countries). This model allows me to control for time fixed effects39 and country fixed effects in the same time. Other possible threats that can bias the estimation is heteroskedasticity and

autocorrelation40. Both were present in the model for which clustered standard error option for country groups was used which treats the disturbance of these problems by increasing the value of the standard errors.

38

Fixed effect model does not allow to include time-invariant variables like dummies in the model. A separate equation cannot be set up.

39_{Time fixed effects are present in the model according to the testparm test} 40

(30)

29

5. Discussion of results

5.1 Net Interest Margin

The first efficiency variable that is estimated is the net interest margin. Table 4 shows two regressions. In the first one the results are insignificant for all variables but the GDP growth. This implies that economic growth has a significant negative impact on net interest margins. The possible explanation can be related to better economic condition which can increase the number of competitors and the banks are able to sell more credit on the markets which raise competion so margins are needed to be reduced. Including the interaction variable changes the results of the estimation. Foreign bank share appears statistically significant with a negative sign in the model suggesting that increasing level of foreign bank presence

(31)

30

Table 4

Foreign bank presence and the change in net interest margin, the role of financial development

1 2

VARIABLES Net interest margin Net interest margin

Foreign bank share 0.00493 -0.0344**

(0.0169) (0.0140)

Deposit money bank assets/ GDP -0.0288 -0.0174

(0.0218) (0.0231)

Bank credit/bank deposits -0.00220 0.00116 (0.00427) (0.00486) Deposit money bank assets /

central+dep money bank assets

0.0250 (0.0299)

0.0152 (0.0272) Private credit by deposit money

bank /GDP 0.0226 (0.0234) -0.0141 (0.0295) GDP growth -0.00204*** -0.00163*** (0.000685) (0.000516)

Inflation 6.95e-05 7.04e-05

(4.32e-05) (4.25e-05)

Real interest rate 0.000177 0.000195

(0.000211) (0.000211) Private credit by deposit money

bank /GDP * Foreign bank share

0.0787** (0.0308) Constant 2.384*** 2.335*** (0.858) (0.811) Observations 429 429 Number of country2 28 28 R-squared 0.387 0.446

Robust standard errors in parentheses

(32)

31

University of Groningen Faculty of Economics and Business Master Thesis The Effects of Foreign Bank Presence on Banking System Efficiency