Finance
International Business

Economic Determinants and Entry Modes of Foreign Banks Into Central Europe

Economic Determinants and Entry Modes of Foreign Banks Into Central Europe

Introduction

In the last fifteen years there has been a rapid increase in the activity of foreign banks in

several developing economies. Although, foreign bank entry occurred in many developing countries,

its pattern was not uniform (IMF 2000). In Latin America as well as in the Central European (CE)

countries, the share of foreign banks in the first half of the 1990’s was well below 20 per cent and a

decade later the foreign banks controlled almost 75 per cent of total banking assets. By contrast, in

East Asia over the same period, the average share rose only from 3 to 7 per cent (Barth 2001). The

level of development of a country seems also not to be an obvious determinant explaining foreign

bank entries. In such countries as Egypt or Bangladesh, the foreign banks hold less than 10 per cent

of banking assets; on the other hand in Cambodgia, Czech Republic or Turkey more than 60 per cent

is in the foreign hands. The differences are also meaningful among the transition countries. In

Azerbaijan or Uzbekistan, the share of foreign banks is less than 5 per cent, whereas in such

countries as Hungary or Lithuania it amounts to almost 100 per cent. The discrepancies are also in

the developed countries. In Germany or United States, the foreign-controlled banks hold less than 10

per cent of assets, whereas in Luxemburg or New Zealand they hold more than 90 per cent.

A good financial system has been shown to be an essential ingredient for sustainable

economic growth (Levine 2005, World Bank 2001). The literature on foreign banking has also

shown that foreign bank participation can help develop a more efficient and robust financial system

(Claessens et al. 2001). Most evidences show that increased foreign banking is generally positively

correlated with the improvement of the efficiency of the domestic banking sectors and helps

strengthen countries’ financial systems. Especially, studies on the developing countries have shown

that these countries have benefited from this trend at most.1 Therefore, from the policy perspectives

it is important to know what determines a favourable environment that encourages cross-border

activity and entering foreign banks. Despite the recent trends in the banking internationalization, 28

per cent of developing countries still have foreign bank participation below 10 per cent and 60 per

cent of developing countries have below 50 per cent. Among these developing countries with the

foreign bank assets below 10 per cent, the transition countries constitute almost 20 per cent and 25

per cent of the sample with the foreign bank participation below 50 per cent (Van Horen 2006 and

EBRD). As the experience of some Central and Eastern European (CEE) transition countries has

shown the foreign bank participation has turned out to be inevitable to build stable and efficient

financial system. Hence, we would expect that in other developing and transition countries, the

entering the foreign banks might also turn out to be necessary in the near future.

At the same time, by permitting foreign banks to enter, the policy makers should take care

about how the foreign banks operate in the host markets. First, letting the foreign banks come in, the

host countries may open themselves up to economic fluctuations in the entrants’ home countries.

Moreover, the organizational form may affect the competitive structure of the local banking systems,

threatening profits and market share of domestic banks and affecting the price and quality of banking

services in the host country. And finally, the entry through de novo operation involves different

levels of parent bank responsibility and financial support. This can have implications not only for

parent bank but also for local regulators, who care about the stability of the host country, and for

local depositors who care about the safety of the savings. The recent experience of Argentina, where

some foreign banks decided not to recapitalize their foreign subsidiaries presents the best example

on this.

On the other hand, the literature on international banking comes up with some arguments in

favour of one entry mode versus another. While the empirical studies show that the entry of foreign

banks through cross-border mergers and acquisitions is positively correlated with the efficiency

improvements of the acquired banks, the entry of foreign banks through subsidiaries may promote

greater access to the financial services in the host countries, as in many countries, the foreign

subsidiaries have powers identical to those of domestic banks. Hence, although, the trade-off

between one entry mode versus another is by local regulators required, from the policy perspectives,

it is also important to know what determines a foreign bank’s choice of organizational form.

Today, the banking sectors of most transition countries are among the ones with the highest

share of foreign-controlled banking asset in the world. It ranges from 70 per cent in Poland to almost

100 per cent in Slovakia (Allen et al. 2006). The change in the share of foreign participation in

banking in these countries from the early transition years to the later ones is significant.

The pattern of the banking internationalization was also not uniform in these countries. At the

beginning of the transformation, foreign banks entered the region mainly as de novo operation.

Encouraged by the fast going economical and political reforms in the region and high economic

growth, the pressure to enter the CE region has increased. With the intention of gaining rapidly share

in the local market most foreign banks used mergers and acquisitions’ (M&As) entry mode instead

of de novo operation abroad.

Generally, banks are found to be attracted to markets abroad to exploit favourable financial

system environment and to take advantage of economic opportunities in those countries (Goldberg

and Saunders, 1981). However, the question arises, what the foreign bank managements saw in the

CE countries that in the middle of the 1990s experienced in an increase of foreign bank’s entrance?

Do the theorists from the developed countries find the acceptation of their thesis in the countries

characterizing negative real economic growth, high inflation, uncertainty with the political

institution and an underdeveloped banking sector? In this paper, we try to analyze the determinants

which in the light of high uncertainties in four countries: Poland, Czech Republic, Slovakia and

Hungary, contributed to the foreign bank entries into these markets.

Our decision on the choice of this sample has been driven by some variations between these

countries. On the one hand, these countries have experienced the most significant pattern of banking

internationalization among almost all transition countries; on the other hand, they represented

various policies toward foreign bank entries as well as started their transition processes with

different initial conditions. For this reason, in our opinion, these countries constitute the best testing

ground on determinants of foreign banking participation and entry modes chosen by foreign banks.

The literature comes up with various motives for foreign banks to go abroad. In addition, the

mode of entry or the organizational form chosen by foreign banks is not only an issue of their

strategy or mission, but also depends on the entry country’s conditions and environment. Despite the

profound changes in the banking sectors of the CE economies as well as growing number of

countries embracing foreign bank entries, there is still open debate about the determinants of

banking internationalization and its modes of entries. The empirical evidences presented in the

literature come mostly from the US and developed European countries. There has been little

empirical research in this field for the developing countries so far. With our study we present the

empirical evidence on the motivation and entry vehicles of foreign banks in the CE markets.

Our contribution with respect to previous literature is twofold. First, we consider a new and

wider set of explanatory variables than previous studies, verifying different hypothesis and relative

importance of economic factors in determining banks’ choice of whether and where to expand

abroad. Second, we use a unique sample of entry models of foreign banks entering the region. Our

framework permits us to examine the relation between the relative importance of the different

country’s factors and the chosen entry model by foreign banks in the CE region.

Our major finding is that the most important factors determining foreign bank entry into CE

countries were development of the financial system and the banking sector as well as the legal origin

of the home country. We also show that most of the entries occurred in the period of a poor creditor

rights protection. Furthermore, our results find that the size of economic growth rates differentials

between host and home markets, and finally the distance between the host country and the banking

headquarters were of great economic importance. We also show that determinants of bank

internationalization have changed within development of the financial systems. Finally, our findings

present that the economic determinants had also an impact on the decision of organization form of

the foreign banks in the CE local banking markets.

The remaining of the paper is organized as follows. In the next section we present a short

overview of problems encountered by the transition from planned economy to the market economy

in the four CE countries. The third section presents the literature review about determinants related

to our main hypothesis of the banks’ expansion abroad. At the end of this section we present also the

results of the few empirical papers on foreign banking in the transition countries. In the next section,

we present the variables based on previous empirical research, which we have applied in our

regressions. In a subsection we develop also our main hypothesis related to the economic

determinants and the decision about the entry mode of the foreign banks into the CE local banking

markets. In the fifth section we present the model which investigates the incentives of foreign banks

for entering CE countries in the last decade, the period of enormous uncertainties and economic

transformation. The next section describes the results and compares them with other ones from

developed countries. Finally, the last section of the paper concludes.

2. Banking sector in early transition process

All the CE countries in our study followed the socialist financial system model, which was

designed to support the central planning economic system. Despite the centralization of financial

functions the state directed credit allocation with scant regard for repayment capacity, using the

national bank and state banks to channel funds to state owned enterprises

As a consequence of political changes in the year 1989 the creation of an effective financial

was a priority for the new governments in the CE countries. The aim was to implement a market-

oriented economy and thus fundamental changes were needed in the financial system. So the

banking industry was one of the first economic sectors, which underwent a fundamental

transformation.

Hungary was the leader among the CE countries in the banking reforms. The government began the

banking reforms even before the political changes. In the early 1980s the Hungarian government

permitted a number of foreign banks to set up operations, even though these banks competed with

state-owned banks in the areas of foreign exchange and trade-related transactions. The centralized

mono-banking system was replaced by a two-tier banking system as National Bank of Hungary

assumed the role of central bank in 1987. The new central bank was charged with pursuing monetary

policy, including exchange rate policy, and was made responsible for the supervision of the banking

sector. The second tier consisted of the specialty banks, newly created commercial banks, and the

few already operating foreign banks (Hasan and Marton 2003).

In Poland the reform of the banking system started in 1987, when the government allowed for

creation of the joint-stock banks, yet they were still owned by the state. Two years later a new

banking law was introduced, which created a two-tier banking system in Poland.

In all the CE countries as a process of creating a two-tier banking system the commercial and retail

operation was divested from the activity of national banks and transferred to new commercial banks.

In Hungary the government set up three new state-owned banks from the National Bank of Hungary,

in Poland nine banks were created out of the National Bank of Poland, while in the Czechoslovakia

through divestment form the State Bank of Czechoslovakia four banks were established. These

medium sized state-owned banks inherited segments of the old network and staff of the national

banks, household deposits and loan portfolio comprising mainly of credits granted to the state

enterprises of unknown quality. They supplemented the already existing large state-owned specialty

banks. Those specialty banks existed separately from the central bank and performed specific

functions on behalf of the government in the planned economies. A state savings bank with an

extensive branch network was responsible for collecting household deposits, although most savings

was forced and done by the state. A foreign trade bank handled all transactions involving foreign

currency. An agricultural bank provided short-term financing to the agricultural sector. A

construction bank funded long-term capital projects and infrastructure development (Bonin and

Wachtel 2003)3.

Capture

Table 1 presents the representatives of each specialization group in a particular

country.

Although three to nine new state-owned banks were set up through the divestment from

central banks, yet the banking industry remained fragmented as the three to four specialist banks still

dominated the emerging banking system. However, already in the first year of the transformation,

new banks started to operate in transition countries. The entry requirements policy of the newly

central banks and the licensing procedure for the de novo banks was very lenient at that time. The

principal motivation was to increase the competition of the four large banks, which were considered

too inertial and ineffective. The number of de novo banks was very impressive at this time. In

Hungary six new banks were established, in Poland 20 new banks and 13 new banks in

Czechoslovakia in 1990. Of these, three were foreign owned in Hungary, five in Poland and four in

Czechoslovakia.

However, this huge expansion in de novo domestic private banks later caused serious problems for

the financial system. Most of those domestic banks were in general undercapitalized and placed an

additional unwanted burden on an underdeveloped regulatory structure. In addition, some of them

have been set up either by state enterprises or by local governments in order to provide soft lending

to them. Hence, the features of banking system at the beginning of transformation were structural

segmentation, high concentration of the assets caused by few large and medium sized state-owned

banks, and an increasing number of small domestic private banks (Bonin, Hasan and Wachtel 2005).

Given the poor banking supervisory environment caused by poor accounting and financial

information, weak off-side surveillance capacity and the lack of experience with on-site

examinations, it was bound to lead to problems in the banking industry. The benevolent licensing

policy, combined with inexperienced and still weak banking supervision, caused the new private

domestic banks to take on rather unsound development strategies. In addition, the absence of

effective legal and institutional supervision also invited fraudulent behaviour by the managements of

these banks. As a consequence the new domestic banks started to have liquidity problem in very

short term. Also the former specialist banks get into trouble as they inherited a loan portfolio from

the past in which credit was granted not on commercial terms. In addition, those banks were still the

primary lending vehicle and quasi fiscal financing, usually for loss-making state-owned enterprises

that had to be either privatized or closed. The number of non-performing loans increased

significantly as the structural problems of the real economy increased caused by the ongoing

transition process in CE countries (Bonin and Wachtel 2003). Once the compliance of supervision

provision requirement was enforced, the quality of loan portfolios became apparent. As a

consequence several large state-owned banks reported huge losses and the equity adequacy ratios

were below the requirement of the banking supervision.

In Hungary, at the end of 1992, 15–28 per cent of the credits extended were nonperforming

loans and were primarily borrowed by the state-owned enterprises during the pre-1989 era (Hasan

and Marton 2003). The situation quickly became unsustainable as failing financial institutions turned

for bailout to the National Bank of Hungary. As a result the newly established national bank was in

jeopardy and the Hungarian government had to step in through a series of costly loan consolidation

programs beginning in 1992 (Várhegyi 1994, 1995, Balassa 1996). The government objective of the

bailout programs was the cleaning-up of the books of the state-owned banks, which would permit a

sell of to foreign strategic investors. The cost of the program approached close to 10 per cent of

Hungary’s GDP.

Poland was the most successful in dealing with the bad debt crisis. The success is attributable to the

design of the recapitalization program, which provided the least incentive for moral hazard. In

addition, the central bank encouraged the buyout of troubled banks by foreign strategic investors. As

a consequence, the costs of bad debt bank crisis were below 1.5 per cent of GDP and were the lowest

among the transition economies.

In Czechoslovakian a Consolidation Bank was established as a vehicle for the takeover of the

accumulated bad loans till 1991. The bank was created to take nonperforming loans from the balance

sheets of the largest state-owned banks, and the clean-up of the books of other banks in the periods

both before and after the division of Czechoslovakia (Dědek, 2001). The overall costs of the

Consolidation Bank are estimated to have reached more than 7 per cent of GDP. Nevertheless the

creation of the Consolidation Bank did not solve the problem of the banking sector and the Czech

National Bank had to intervene in the affairs of eight banks by 1996. In 1997 classified credits

reached already 32 per cent of the total banking credits in Czech Republic (Dědek, 2001). Finally the

problem was resolved through a postponed privatization of the largest banks. However, the estimates

indicate that the final cost of bank bailout in the Czech Republic may have approached 30 per cent of

GDP as compared to just 1.5 per cent in Poland or 10 per cent in Hungary.

The growing problem of bad debt was the tiger for the postponed bank privatization in all the

transition countries. In most of these countries the privatization of state-owned banks started in the

beginning of the 1990’s, yet foreign banks were entitled only to minority shares whereas controlling

stakes remained with the state treasury. However, as the problem of bad debt increased, the

government was more likely to sell controlling shares in the state owned banks to foreign strategic

investors. The governments in transition countries were in addition encouraged by privatizations

revenues as they started with the privatization of state owned banks. Thus, at the end foreign bank

made their entrance into transition countries mainly through rescuing the ailing domestic banking

sector.

In opposition to the other transition countries, in Hungary bank privatization policy from the

beginning of was aimed at selling controlling shares in state-owned banks to foreign investors.

Although the privatization required prior an initial recapitalization of the banks so that the

combination of current net worth and franchise value would attract a foreign investor. As a

consequence the Hungarian government engaged in multiple recapitalizations of its domestic banks

caused by the poor quality of loan portfolios. Thus, the government was able to attract foreign

investors and thus signal credibly the end to bailouts of these banks (Hasan and Marton 2003). At the

end of 1997, four of Hungary’s five large state-owned banks had been sold to foreign owners and by

the end 2006 the share of foreign banks was 63 per cent of total assets.

The Polish experience indicates the danger in combining the resolution of bad loans with

bank responsibility for enterprise restructuring. The main instrument used to restructure bad loans

was debt to equity swaps. Hence, weak banks with no expertise in restructuring large companies

ended up taking ownership stakes in their weak clients. Therefore bank credit was provided regularly

to ailing enterprises and no meaningful enterprise restructuring was promoted banks (Gray and Holle

1996). Poland’s program strengthened, rather than cut off the ties between weak banks and their

undesirable clients and, thus, postponed painful restructuring of ailing enterprises (Bonin and Leven

2001).

In addition, in Poland the government presented an inconsistent policy toward foreign banks. In

1993 the government attracted the first strategic foreign investor for two of the nine midsized

commercial banks, yet only minority shareholding was allowed. Thus, foreign institutions controlled

only 2.1 per cent of Polish banking assets at the end of 1994. The National Bank of Poland however

enabled foreign bank entry at the beginning as de novo operation and later through either the buyout

of failing banks or their nonperforming credit portfolio. At the same time the government arranged a

large bank merger, in which the three of the nine midsized commercial banks were merged with one

of the state savings bank to form the second largest financial group in Poland. However, the

persisting inefficiency of Polish banking system caused the government to change their attitude

toward foreign investors. So in 1997 foreign bank were allowed to take control in the initial

privatization of the state-owned banks. Since then significant strides have been made and foreign

strategic investors took control in some of the largest commercial banks. In 2004 the government

sold 30 per cent of shares in the country’s largest retail bank PKO BP through the Warsaw Stock

Exchange. It was the last state-owned commercial bank and therefore the government decided to

retain a majority stake in it. Yet, at the end of 2006, 75 per cent of total bank assets were controlled

by foreign capital.

In the Czech Republic bank privatization took place twice. In 1992 the government of the

Czechoslovakia conducted a voucher privatization transferring the shares to individual investors and

investment funds in exchange for vouchers. Three of the four large commercial banks participated in

voucher privatization, yet these banks participated on both sides of privatization as they also

sponsored the largest investment funds. As a result, Czech banks took ownership stakes in their

voucher-privatized clients, some of which continued to be loss making, while the state retained a

controlling ownership stake in the large banks. Consequently, voucher privatization in the Czech

Republic strengthened the relationship between banks and clients and left bank governance held

hostage to the legacies of the past. Thus, the privatization of the Czech banks was to little avail

because soft lending practices continued. As a consequence these banks accumulated bad debts,

which have been later transferred to the Consolidation Bank.

In Czech Republic the second round of privatization occurred from 1998 to 2001, when the

government sold holding in three major banks. Until than no Czech bank was sold to a foreign

investor. Those three banks accounted for 38 per cent of assets. Since then the proportion of foreign

owned bank assets soared to 96 per cent in 2006.

All these four transition countries took place in the enlargement process of the EU and are

members of the EU since May, 2004. Consequently they had to adapt the Second European Banking

Directive and the Single European Passport, which eliminated the last market-entry barriers into

their banking sector. Although in all the countries the deregulation of the banking sector could be

observed since 1997.

Concluding, the increasing foreign bank presence since the 1990s is one of the most striking

developments in the banking system in the transition economies. On average, foreign-owned banks

account for more two thirds of total bank assets in most transition economies at the end of 2006. The

percentage of assets in banks with majority foreign ownership in these countries ranges from 22 per

cent in Slovenia to 99 per cent in Estonia. By contrast, in EU-15, only Luxemburg and Great Britain

had more than 50 per cent of its banking sector controlled by foreign interests in 2005 (Allen,

Bartilloro and Kowalewski, 2006). Thus, banking sectors in transition countries differ significantly

from their counterparts in developing as well as from emerging market countries by the unusual high

percentage of assets held by foreign banks.

3. Literature Review

In the last decades various studies have been conducted that investigated the motivation and

location choice of banks abroad4.

The classical hypothesis (Aliber, 1984) is that banks follow their customers abroad, being

afraid of losing them once they have established relationships with banks operating in other

countries. According to the defensive expansion hypothesis, banks’ expansion enables them to retain

information on their customers.

Multinational banking hypotheses relating to the servicing and following their clients generally find

empirical support. Nigh et al. (1986) presented in their study of US banks’ overseas expansion that

the major determinant was to respond to the financial needs of US firms abroad. Their study implies

that US banks do not lead, but follow the US business sectors. Goldberg and Saunders (1980)

analyzed the factors affecting the expansion of US banks into UK, concluding that US trade is

significantly conducive to the growth of US banks in UK, while the Eurodollar interest rate and the

exchange rate are not significant factors. In a later study Goldberg and Saunders (1981) examine on

the contrary the growth of foreign banks in the US. They results provided evidence that the direct

investment made by foreign firms into the market was a significant positive determinant of growth

of foreign banks’ market share in the US. Hultman and McGee (1989) and Grosse and Goldberg

(1991) also provided results that foreign banks entered the US market to service the international

trade and direct investment needs of their home-country clients. In a recent study similar results were

presented by Magri et al. (2005) in a study on entry decisions and activity levels of foreign banks

operating in Italy. The authors report that trade influences both entry decision and activity levels of

foreign banks. However, they found also that the relative profitability of banking activity in Italy

strongly influences both entry decisions and activity levels. As a consequence the observed

correlation in several studies between proxies for foreign investment trade and the structure of a

foreign market complicates the conclusions on motivation. Thus, the motivation of bank to move

abroad may be explained by the need to follow its clients and equally by the lure of a potentially

significant new market.

The importance of new market opportunities in attracting foreign banks has been emphasized

by the eclectic theory of direct investment (Dunning, 1977). The theory was extended by Gray

(1981) to explain multinational banking. In this theory multinatinalization of banks is contingent

upon location-specific factors and ownership-specific factors.

The location-specific factors are the size and competition in the foreign market, presence of entry

restriction and other regulations. Foreign market size has been found to be a significant driver of

multinational banking by Terrell (1979) and Goldberg and Grosse (1994). While, Goldberg and

Johnson (1990) provides some support for relative lack of competition or high relative profitability

as causal factors. In contrast Nigh, Cho and Krishnan (1986) did not found that local market

opportunity to have a significant effect. In their study they analyzed the role of location-specific

factors in foreign involvement of the US banks.

Recent studies presented a new approach to multinational banking and market structures. In those

studies banks may use economic crises and distortions in the banking industry in order to enter a

foreign market. Peek and Rosengren (2000) found evidence that as a result of liberalizations and of

the worsening conditions in domestic markets, foreign banks expanded in several Latin American

countries. Consistent with this result Guille´n and Tschoegl (2000) found that Spanish banks have

increased their ownership in Argentina’s banks during the economic crisis of the last decade.

However, Engwall et al. (2001) found that foreign banks lost market share in Sweden during the

Scandinavian banking crisis there in the early 1990s. On the other hand, at the same time they found

that foreign banks increased their market share in Norway. As we see the empirical results do not

present a clear picture on market structure, yet it seems that foreign banks may use a domestic crisis

in order to increase their market share in the market.

The ownership-specific factors emphasis that banks become multinational in order to employ their

domestic strengths in foreign markets at low marginal cost and thus leverage those strengths. Such

advantages can take many forms, including large scale of operation, low cost of capital, unique

business processes or banking technology, skilled personnel and banks’ reputation (Nigh 1986;

Tschoegl 1987). Among bank-specific characteristics, size has been found to affect mainly the

patterns of foreign direct investment. Ball and Tschoegl (1982) provided evidence that the larger

banks are much more international than smaller ones.

Consistent with this result Focarelli and Pozzolo (2001) have shown that banks with foreign

shareholdings are on average larger and have headquarters in countries with a more developed and

efficient banking market. However, Berger et al. (1995) argue that larger banks have generally larger

and more internationally diversified customers, and therefore these banks have more incentives to

follow their clients when they operate abroad. If it is the case than large foreign banks would rather

follow their multinational clients than have been encouraged by their comparative advantage. In

addition, several studies have documented that foreign owned banks are not as profitable as their

domestic peers. Seth (1992) and Nolle (1995) found that foreign owned banks were not as profitable

as domestically owned banks, based on aggregate profits. DeYoung and Nolle (1996) use a profit-

efficiency model and conclude that foreign-owned banks were less profit-efficient because of their

reliance on purchased funds. Molyneux et al. (1997) applying a simultaneous equations framework

concludes that the profitability of foreign owned banks was mainly related to capital ratios,

commercial and industrial loan growth and asset portfolio composition.

Although, the presence of higher demand profit opportunities in the market of destination of the

investment seems likely to be an obvious determinant of the location choice of multinational banks,

the empirical studies are more equivocal on location-specific factors and ownership-specific factors

as motives for banks to go abroad.

Apart from leveraging existing advantages, following clients or seeking attractive markets

overseas, there are other determinants of bank expansion abroad. In the opinion of Focarelli and

Pozzolo (2001) bank internationalization depends on other factors besides the degree of economic

integration among countries. As an example Claessens et al. (2000) analyze foreign presence across

80 countries from 1988-95, and find that foreign banks are attracted to markets with low taxes and a

high per capita income. Although, the regulatory restrictions have been found to significantly affect

the pattern of bank investment abroad. Miller and Parkhe (1998) presented that US banks prefer to

expand in countries where capital requirements are less stringent and taxes are lower. Consistent

with this result, Nigh et al. (1986) and Goldberg and Johnson (1990) present that restrictions on the

entry of foreign investors significantly reduce the degree of internationalization of a country’s

banking market. According to Boot (1999) governments may wish to have the largest banks in their

countries to be domestically owned. Thus, we would expect that in high concentrated markets as the

CE are, the entry of foreign banks is more difficult. In this case a single acquisition of the former

state owned banks would imply the loss of a significant market share to the advantage of foreign

financial institution.

The literature on the restructuring and development of the financial sector in transition

economies is abundant. However, the empirical literature on banking in transition countries

concentrates mainly on the impact of foreign bank entry on banking efficiency. Yildirim and

Philippatos (2002) find that foreign banks in transition countries are more cost efficient but less

profit efficient relative to domestic banks. Hasan and Marton (2003) and Fries and Taci (2003)

demonstrate that the entry of more efficient foreign banks creates an environment that forces the

entire banking system in transition countries to become more efficient, both directly and indirectly.

Buch (2000) compares interest rate spreads in Hungary, Poland and the Czech Republic from 1995

to 1999. She finds evidence confirming the hypothesis that foreign banks create a more competitive

market environment in transition economies, but only after they have attained sufficient aggregate

market share. The results were conformed to Zajc (2002), who reported for six European transition

countries that foreign bank entry reduces net interest income and profit, and increases costs of

domestic banks. While, Bonin et al. (2003) examine the performance of banks in eleven transition

countries and show that majority foreign ownership is associated with improved bank efficiency.

On the contrary, Green et al. (2002) estimate the efficiency of domestic and foreign banks in Central

and Eastern Europe (CEE) in terms of economies of scale and scope. They find that foreign banks

are not really different from domestic banks and that bank ownership is not an important factor in

reducing bank costs. There results were in oppostition to Claessens et al. (2001), who reported that

foreign banks in CEE countries tend to have lower overhead costs and loan loss provisions and

higher profits than domestic banks.

Fries and Taci (2005) presented that costs are lower in those transition countries where foreign

owned banks have a large share of assets. While, de Hass and van Lelyveld (2003) argued that the

increase in foreign banks have contributed to credit stability in CEE by keeping up credit supply

during crisis periods, while domestic banks reduced theirs. Although their results also show that the

privatisation of domestic banking systems in CEE as such has not led to immediate positive stability

effects. They have shown that banks that are sold to foreign strategic investors do not change

immediately into more efficient banks. Additionally, they presented that the country conditions

matter for foreign bank growth, as they have reported a significant negative relationship between

home country economic growth and host country credit by foreign bank subsidiaries. Related results

were provided by Bonin et al. (2005) in a study on the impact of bank privatization in transition

countries. They have reported that state-owned banks are the least efficient and foreign de novo

banks are the most efficient of all bank types in transition countries. However, they found also that

domestic banks have a local advantage against foreign banks in pursuing fee for service business.

The effects of foreign ownership on bank efficiency have been also examined in a few

country specific studies. For Hungary, Hasan and Marton (2003) find that relatively more efficient

foreign banks created an environment that forced the entire banking system to become more efficient

in the years 1993 to 1998. Nikiel and Opiela (2002) find that foreign banks servicing foreigners and

business customers are more cost-efficient but less profit-efficient than other banks in Poland from

1997 to 2000. For Czech Republic and Poland, Weill (2003) reported that foreign owned banks were

significantly more efficient than domestically owned banks in 1997. On contrary, Matousek and Taci

(2002) observed greater efficiency in private banks in the Czech Republic for the period 1993–1998,

yet they did not found any evidence of greater efficiency of foreign owned banks in their study.

Although these single country studies provide mainly a positive relation between foreign ownership

and bank performance, yet the results are not always convincing.

Finally, Naaborg et al. (2003) present that the three largest banks in most European transition

economy are in foreign hands. However, banks from non-European countries are almost absent in

the transition countries. In addition, they report that there is a relatively strong presence in some of

the European transition economies of foreign banks from neighbouring countries.

While the empirical evidence confirms the follower relationship hypothesis, the importance

of local market opportunities requires deeper investigation. So far little research has been undertaken

in order to examine the relation between foreign bank expansion and economic and structural

characteristics of host countries. In particular, a variable measuring profit opportunities usually

mentioned in the theory is either omitted in empirical studies, because of limited data availability, or

found to be non-significant.

In addition, the validity of the foreign bank motivation and entry modes has not been yet

established for the transition countries due to the modest attention given to their empirical

verification. Our study tries to fill the existing gap in the multinational banking literature building

our study upon previous empirical work. We focus on this aspect arguing that transition countries are

an interesting testing ground for theories on multinational banking. In 1990s, the economy and

financial market were characterized by lack of competition and close regulation. The situation

changed in the 1990s due to political transformation, when the financial markets were liberalized

and competition in the market increased. As a consequence, the transition economy and thus the

banking sector offered several profit opportunities to be exploited by foreign banks. Yet, it is still

unclear which motives for foreign banks had been the leading in the decision to go into one of the

transition countries.

4. Economic determinants contributing to FDI in CE banking sectors

As have been shown in the previous section, there are various theories explaining the motives

for banks to go abroad. In this section, we review the determinants that have been provided by the

literature as the motivation of foreign banking and present the proxies we have included in our

regressions. Our main goal of this paper is to provide an answer, which determinants have been the

leading ones for foreign banks, in their decisions to open a subsidiary in one of the CE countries.

However, we have decided to organize those determinants into four major groups. Each of the group

represents a different hypothesis providing an explanation on the motives behind foreign bank

expansion into one of the CE country. In addition, we hope this way to be able to establish the

relationship between the motivation and the model of entry chosen by the foreign banks. We review

the determinants of entry modes of foreign banks in more detail in the sub-section below. Given the

above considerations, we present the following four hypotheses to be tested.

Hypothesis 1: The foreign bank involvement is positively related to client’s presence in

the CE country

Hypothesis 2: The foreign bank involvement is positively related to market opportunities

in the CE country

Hypothesis 3: The foreign bank involvement is positively related to low efficiency of

domestic banks in the CE country.

Hypothesis 4: The foreign bank involvement is positively related to favourable

regulations in the CE country.

As have been already mentioned in the past the pattern of foreign bank expansion has been

dominated by the follower relationship. Under this hypothesis banks decided to expand in order to

provide services to their home country clients in countries abroad. At the same time those banks

operating abroad have gained a growing understanding of foreign markets and have increased the

range of their operation and services. Thus, we believe that the pattern of foreign bank has some

characteristic that are peculiar to the banking industry, yet the choice of expanding abroad depends

on a wider range than just one single factor. Therefore our hypotheses should be seen with great

caution as the variables presenting them may be significant simultaneous and it is difficult to asses,

which of them may be more important on a stand alone basis.

Our first measure controls for the first hypothesis to be tested, which have been shown in

many previous studies as an important motivation for foreign bank expansion. As a proxy for the

follower hypothesis we use as proxy the stock of direct investments excluding financial industry into

one of the countries in the CE from the country of origin of the foreign bank. The variable non-

financial FDI was expressed as ratio to the domestic country GDP. We employ it as a lagged one

measure as the rationale is that home banks will follow their customers abroad so that they can

provide services for them in the foreign operations. Thus, we expected that there is a positive

relationship between foreign direct investments and the expansion of banks abroad. A strong positive

relationship has been reported in the studies of Nigh et al. (1986) and Goldberg and Johnson (1990).

They have found a positive relationship between the US banks foreign activities and the size of US

foreign direct investments abroad.

Liquid liabilities

Another common assumption in the empirical literature is that a well developed financial

market may attract foreign banks due to external agglomeration economies (Davis 1992,

Kindleberger 1974). The rationale behind is that investors consider whether to invest in foreign

banking, the size and structure of the particular financial system is likely to be one of the factors they

take into account. Thus, Konopielko (1997) formulated a hypothesis that with the economic

development of other countries the significance of the follow the client rationale for foreign entry in

banking will diminish and subsequently be replaced by search for client’s behaviour, which presents

our second hypothesis in our paper.

This claim was supported by Dopico and Wilcox (2002) who argued that the size of the host

country’s banking market is one of the significant determinants of foreign expansion. They found

that foreign banks are more pervasive in countries where banking is more profitable and where the

banking sector is smaller relative to GDP. In order to control for these characteristics, we considered

size of the financial sector and the banking sector, whereas the profit opportunities present our next

hypothesis and the proxies will be described later. In our study the size of the domestic banking

market of one of the CE countries is a location-specific determinant of foreign bank expansion.

We employ liquid liabilities, which are defined as the ratio of liquid liabilities of the financial system

to GDP. We consider this variable, as it is usual in the finance literature, as a proxy of financial

depth since it represents the size of the formal financial intermediary sector. The implicit assumption

is that the size of the financial system is positively related to the foreign bank entry. Including liquid

liabilities to GDP might also control for the effects of financial system underdevelopment that differ

systematically by income levels across countries.

In this study, similar to the study of Grosse and Goldberg (1991), the size of the banking

market is proxied by the deposits held by the domestic banks to GDP. This variable allows us to see

whether smaller and less developed domestic banking sectors attract more foreign banking. In theory

the larger the domestic banking market, the greater the number of potential customers. This would

suggest that there should be a large number of foreign banks willing to invest in large markets in

order to take advantage of the market’s potential. In our study we expect a positive relation between

the size of the banking market and the number of foreign banks. Especially in case of Poland, which

is the biggest country in the region, we anticipate to report a positive relation of foreign presense and

the size of the banking market.

Concentration

Steinherr and Huveneers (1994) provided evidence that foreign banking was less common in

countries where a smaller number of domestic banks dominated banking. They argued that greater

concentration limited the choices available to borrowers, forced domestic firms into relationships

with the dominant banks and stunted the development of an arms-length lending market. In such a

market, even though banking might be profitable, foreign banks might be unable to enter. We test for

this by including a five-bank concentration ratio in our model specifications and expect a negative

relationship with foreign banking entry.

Market capitalization and turnover ratio

Demirgüç–Kunt and Levine (1996) documented that in different countries the extent of stock

market development highly correlates with the development of banks and other financial institutions.

We use the value of domestic equities on domestic exchanges divided by GDP to measure the

development of the stock market. In addition, we use the values of equities traded to GDP, which

reflect the activity of stock markets in transition countries. The total value traded ratio is frequently

used to gauge market liquidity because it measures market trading relative to economic activity. On

one side, we would expect significant positive relationship between the development of banking

sector and capital markets in transition countries. On the other side, the more active and developed

the capital market, the greater the competition with the banking industry. Thus, we may also assume

a negative relation between stock market development and activity and foreign bank entry.

Net interest margin and overhead costs

In order to test the importance of market opportunities in the transition countries we employ

two different variables. To test whether the overall profitability of banking in the host country

influenced foreign banking, we include a profitability measure – a net interest margin (Claessens et

al. 2001, Dopico and Wilcox 2002). High net interest margins in the CEE countries in comparison to

other developed countries have been observed in the past (Allen et al., 2006). However, Lensink and

Hermes (2004) find that in developing countries, foreign entry is associated with shrinking margins.

Similar results were previously reported by Claessens et al. (2001), who demonstrated that for most

countries higher foreign ownership is associated with a reduction of costs and net interest margins

for domestically owned banks. Those results were confirmed recently by Allen et al. (2006) in a

study on the EU-25 financial system. The authors have shown a gradually decline of the interest

margins in the CEE region over the last decade and the convergence towards the levels reported in

the developed countries.

Another source of motivation to expand abroad can be the foreign banks’ efficiency relative

to that of the domestic banks. According to Tschoegl (1987), high overhead costs, low efficiency of

management and the cost of capital can increase the likelihood of foreign bank expansion into the

market. In the Czech Republic and Poland foreign owned banks were more efficient than domestic

owned banks and this was not due to scale differences or the structure of activities (Weill 2003),

which would confirm our hypothesis. Therefore, to estimate and control for inefficient domestic

banks, we include the measures of overhead costs.

We will use this two variables in order to test our third hypothesis that foreign banks expand

into those markets, where are the highest profit opportunities and the lowest efficiency of banks. We

expect that foreign banks entering the market will see an opportunity to export their knowledge,

which will give them a competitive advantage in the domestic banking markets. Thus, we assume

that the foreign banks are probably the most efficient in their home market. The combination of high

profit opportunities and the inefficiency of the domestic banks provide the motivation for the third

hypothesis on foreign bank expansion into the CE countries. Therefore we expected that those two

variables will have a positive effect on the foreign entry into the region.

 Legal origin, creditor rights and banking regulations

According to Goldberg and Saunders (1980) international expansion may be affected by both

economic and regulatory factors. In a series of influential papers La Porta et al. (1997, 1998) stress

that the cross-country differences in the legal environment and their enforcement may influence the

financial structure. Rajan and Zingales (1998) argue that bank-based financial structure prevails and

is more effective in countries with weak legal systems and poor infrastructures. While, Darby (1986)

presents that the rate of growth by particular parent countries may be stimulated by home country

regulation that reduces domestic profitability. To examine this issue, we follow La Porta et al.

(1997) and consider institutional factors that measure the quality of the legal environment both

overall and specifically for creditors.

We used the data on the legal origin from the La Porta et al. (1997, 1998) studies, the countries were

classified into five legal origin groups. With respect to legal origin, La Porta et al. (1997) distinguish

first between common law and civil law countries. The civil law comes from Roman law and relies

heavily on legal scholars to formulate its rules, whereas the common law originates from English

law and relies on judges to resolve disputes. It is common to further distinguish between French,

German and Scandinavian civil law countries. In addition, we separately control for the legal origin

of the transition economies were the legal system represents currently a combination between the

French and German civil law.

La Porta et al. (1997, 1998, 2000) argue that common law countries protect both shareholders and

creditors the most. More specifically, La Porta et al. (1998) show that countries based on the English

tradition have laws that emphasize the rights of creditors to a greater degree than the French,

German, and Scandinavian countries. French civil law countries give the weakest protection to

creditors, whereas German and Scandinavian civil law countries are somewhere in between. La Porta

et al. (1998) also examine enforcement quality. Countries with a French legal heritage have the

lowest quality of law enforcement, while countries with German and Scandinavian legal traditions

tend to be the best at enforcing contracts. In our study the variable English Legal Origin equals one

if the country has an English legal tradition and zero otherwise. Similarly, French Legal Origin,

German Legal Origin, Scandinavian Legal Origin and Socialist Legal Origin take on appropriate

values of one and zero for each country.

Legal and regulatory systems that facilitate the repossession of collateral and that grant

creditors a clear say in reorganization decisions are likely to encourage the development of banks.

As shown by La Porta et al. (1997) greater creditor right is positively associated with financial

institutions development. Thus, reforms improving creditor protection may attract foreign bank entry

into the transition countries. In terms of the specific indicators, we follow Pistor et al. (1999, 2000)

who modify the index of La Porta et al. (1997) by excluding one and including two additional

variables, referring the index to the problems of transition countries. In our analysis, the index ranges

from zero to five and aggregates creditor rights.

The creditor rights variable is described in La Porta et al. (1998) and Pistor (1999). We expect that

those countries with the legal systems that assign strong rights to creditor are more likely to support

the growth of banks including those of foreign origin.

Aliber (1984) and Hultman and McGee (1989) noted that a host country’s regulatory

environment affect foreign banking. Using the Barth et al. (2001) analysis of commercial bank

regulations, we construct an aggregate index of regulatory restrictions on bank activities in

securities, insurance, and real estate markets and restrictions on bank ownership of non-financial

firms. This measure of regulatory restrictions on bank activities gauges bank power and therefore

allow us to test whether restrictions on the range of permissible banking activities affected foreign

banking. Therefore, we anticipated a negative relation between foreign bank entry and regulatory

restriction on bank activities.

Economic growth and inflation

Weller and Scher (2001) claimed that the real economic growth and the level of development

of domestic banking determine foreign banks’ presence in the host countries. In order to control for

economic growth we include a variable representing difference in economic growth between host

and home country of the foreign bank. We expect to find a positive correlation between the

difference in economic growth rate and the presence of foreign banks.

A series of recent papers have addressed the study of the long-run influence of inflation on

growth and financial system development (Barro 1995). The main findings of this body of empirical

literature may be summarized as follows. First, inflation has a negative temporary impact upon long-

term growth rates. This effect is significant and generates a permanent reduction in the level of per

capita income. Second, inflation not only reduces the level of investment but also the efficiency with

which productive factors are used.

Exchange rate and corporate tax rate

To consider long-term economic conditions of the countries in our study, we include two

additional variables. The first is the change in foreign exchange rates of the currency of the domestic

country against the Euro currency. We use the exchange rate towards Euro as most of the foreign

banks stem from the Euro area. We will test whether fluctuations in the value of the host countries’

currencies affect the level of foreign investment in banking in CE countries.

Operating a banks subsidiary abroad will involve substantial flow of foreign currencies. A

depreciation of domestic currency may motivate foreigners to acquire the control of domestic bank.

In addition, when the host countries’ currencies depreciate, foreign banks may reduce their

repatriated income and increase their reinvestment in the host countries, as they may want to avoid

exchange rate losses. On the other hand, when the host countries’ currencies appreciate with respect

to foreign banks currencies, capital flows is expected to decrease as it becomes more expensive for

foreign investors to invest in one of the CE countries. Such a negative relation has been reported by

Goldberg and Saunders (1981) and Froot and Stein (1991).

Our second variable is the level of corporate tax in the CE countries. In the literature overseas

bank expansion is also frequently attributed to the variations in tax treatment of banks in different

countries. Thus, taxes may influence the level of foreign direct investment in banking in the region.

The corporate tax regime in use may therefore determine whether or not a country is an attractive

location for a foreign bank to establish a subsidiary. At the same time the foreign entry can be a

response to moves by the host country to attract foreign banks by offering more favourable tax

treatment than the bank’s home country or in order to increase competition in the financial services

sector.

Geographic location

The geographic differences between the home and host nations may proxy not only the

geographical, but also the cultural distance between countries. Given the importance of information

about customers as well as of knowledge of outlet markets in banking, we expected a negative

relationship between distance and foreign entry. In addition, in several studies the geographical

distance has been applied in the literature as a proxy for the degree of economic integration (Ball and

Tschoegl 1982, Grosse and Goldberg 1991).

We measure the geographic difference using the distance between banks host and home country. A

negative relationship may indicate that the difficulty of operating a subsidiary in a foreign country

grows as geographical and cultural differences increase. Focarelli and Pozzolo (2001) have reported

that the distance increases the probability of market entry by acquiring shares in a foreign bank.

While, Magri et al. (2005) presented that the likelihood of operating a foreign bank in Italy diminish

as geographical and cultural differences increased.

EU membership

Finally, following Magri et al. (2005) we introduced also a dummy in the estimates to

identify countries belonging to the EU. We assume that EU banks should have an advantage to other

foreign banks due to lower entry barriers and extended the activities that are permitted to undertake

under the EU Directive. Therefore we expected the variable to exert a positive effect, which has been

reported in Italy by Magri et al. (2005).

4.1 Economic determinants and the entry modes of foreign banks

In principle, the factors affecting the decision about entry into the CE countries may vary with

the mode of entry chosen by a bank. Since such determinants as high net interest margin or great

economic development may promote one form of entry, the others as tax relieves or high

concentration of the banking sector may influence positively the other formal structures. Hence, an

organizational form is not an arbitrary formality but rather a function of foreign bank’s strategy and

scope of its activities willing to provide in the host country. In addition, foreign bank must take into

constitute an economic environment existing both in home and host countries. The legal form chosen

by a foreign bank is also of great substantive importance from another reason. It may under certain

circumstances have effect on the stability of both home and host banking sectors. The first one may

be affected by a failure or great losses of a parent’s bank institution in a host country. From the point

of view of a host country, the regulations promoting particular modes of entry may prevent country

from a crisis or at least attenuate their effects (Tschoegl 2003).

The regulatory environment of the CE countries has changed over time. Furthermore, it was also

different among the countries themselves. In principle, the foreign banks could enter the CE

countries either by acquiring or merging with a domestic bank or through de novo operation. We

distinguish among the operational forms a subsidiary or branch of a parent company, as well as a

representative office of a bank. Since bank’s representative office can not provide any financial

services in a host country, we do not consider them in our analysis.

A branch is defined as an integral part of the parent organization and in our opinion it constitutes

the highest level of foreign banking penetration in a host country. The branch shares a parent’s credit

rating, lends and trades on the parent’s full capital base. Thus, it may have substantial advantage in a

host country banking market. However, a branch may go insolvent if its parent goes bankrupt or

other way around. Thus, this mode of entry requires a careful supervision of both home and host

country’s authorities. The Polish banking law allowed the foreign banks to enter via branches since

1989. The licensing policy was also very liberal at that time. The only requirement to be fulfilled by

a foreign bank to set up a branch was an agreement with the National Bank of Poland. However,

despite that, Poland did not experience in wave of branches. The situation has not changed

significantly until now. One of the reasons was that the Polish National Bank was not willing to

allow foreign banks to operate as branches easily.

The situation looked differently in Hungary. The Hungarian regulatory authorities abolished the

entry via branch until the 1997 and even after the implementation of the Second Banking Act

Amendment in 1997, which provided a possibility to establish a branch by a foreign institution, this

form effectively qualified as subsidiaries in terms of capital requirements and operations (Kiraly et

al. 1999). Although, the operation activities via branches are allowed, the country has not

experienced any opening of branches till 2004.

In the Czech and Slovak Republics the situation looked very similar to Poland. The banking laws

from their beginning allowed foreign banks to set up branches assumed they received a formal

approval from the host national central bank.

Since the accession into the EU, the member states has been granted a “single passport”, which

assumes that all credit institutions authorized in an EU country would be able to establish branches

or supply cross-border financial services in the other countries of the EU without further

authorization, provided that a bank was authorized to provide such services in the home state

(Dermine, 2005).

Capture

Table 2 shows that branch has been very rare mode of penetrating CE banking markets

comparing with other European countries despite any specific restrictions (excluding Hungary) per

se imposed by the regulatory authorities on this organizational form. One reason for that could be

that branches are very sensitive to the location-specific risk (Tschoegl 2003). Hence, in the course of

instable political and economic situation, the parent banks preferred to choose other organizational

forms, which could put them in the more secured position and did not require risking their

reputations once the expectations of great economic development would not have been met. Wengel

(1995) has proved it empirically concluding that the parent tends to send branches to wealthier

countries, while the less sophisticated forms to the developing ones. On the other hand, setting up a

branch of foreign bank should be justified by sufficient activities in the area for which a branch

offers an advantage (Heinkel and Levi 1992). Therefore, many studies on international banking

argue that branches are not attracted by great profit opportunities and hence they do not state in the

direct competition with other legal forms (Miller and Parkhe 1998). In the US, Heinkel and Levi

(1992) found that setting up a branch was positively correlated with the development of the domestic

money and capital markets, in which the foreign branches participate allocating the deposits of their

customers collected in the home market. Hence, we may assume that the development of the capital

markets in the CE countries as well as better creditor rights may positively affect the inflow of

branches into this region.

A subsidiary is a separate legal entity incorporated in the host country, mostly acted as wholly-

owned subsidiary company of a parent bank and often it is engaged in a broader range of financial

services than branches. Since the beginning of transformation the subsidiaries were the most

frequent forms of entering the CE banking markets. Heinkel and Levi (1992) point out that

subsidiaries differ from other forms of banking operations and thus respond differently to various

factors. First, they operate in the different area of competition than other legal forms. Second, the

parent bank has different motivations on establishing it. In the CE the history of subsidiaries can be

divided into two periods. The first, early 1990s when the subsidiaries were set up and second, the

middle of 90s when the privatization process began. The motivations of entry through this type of

organizational form have also changed across time. In the early of 1990s, the major motive driving

an establishment of a subsidiary was to provide high-quality services to these companies which had

invested in or traded with the CE countries as well as their foreign employees on the spot (Majnoni

et al. 2003). Thus, these subsidiaries were mostly engaged in the wholesale and corporate banking,

especially depositing, trade and exchange foreign operations. The best examples are Commerzbank

in Hungary (1993) and Czech Republic (1991), Bank of America (1990) and Citibank (1991) in

Poland. It should be also mentioned that many of these banks were motivated to enter by the tax

relieves which were very common practice at that time in CE countries. Unlike branches which are

subject to the home country’s regulations and tax and accounting standards, this could be an

additional motivation for setting up a subsidiary.

In the middle of 1990s, during the time of the major bank privatisations, the motivations behind

setting up a subsidiary changed. In this period foreign banks noticed an opportunity of acquiring

large domestic universal banks. Some of them acquired subsidiaries and even merged them with

already existing operation or branches. Apart from it in this period many subsidiaries of the foreign

banking institutions began to operate, especially in consumer finance sector as Porsche Bank, Opel

Bank, Fiat Bank or Sygma Bank.

Following the above argumentation, we would argue that the establishment of branches and

subsidiaries would be motivated by different factors and that they do not stay in direct competition to

each other.

As mentioned already, the most common mode of penetrating the CE banking markets which

became in the middle of 1990s was an acquisition of the existing banks. The entry through M&As

was the quickest and the simplest mode of establishing presence in the CE countries. Mostly, it took

place during the privatization process when the governments offered share in the domestic banks in

order to save them or in exchange for the takeover of bad portfolios. This process lasted till the

entrance of the CE countries into the EU. One reason for that were the administration restrictions

imposed by the governments on the acquisition of majority stakes by foreign institutions. In the

Czech Republic, for example, the acquisition of majority stakes to the strategic investors was

abolished. Thus, foreign investors were able to buy only minority interests in the domestic banks in

the first years (Bonin and Wachtel 1999). The Hungarian banking law, on the other hand, required

an agreement of President of the National Bank on acquisition of stakes in a domestic bank above 10

per cent. However, it represented the most liberal licensing policy and the privatization process with

the possibility of acquisition of majority later on. In Poland, the government started to sell majority

shares of the state-owned banks to foreign investors at the end of the 1990s (NBP 2001).

Tschoegl (2003) point out that the type of an organizational form chosen by foreign banks to

expand, is often closely connected with its strategy. He argues that the conditions which drew

foreign banks to enter developing countries erode over time and then some will have to withdraw

their local operations. Therefore, he distinguishes among others two types of banks’ strategies. First,

prospectors who enter via wholly-owned subsidiaries or joint-ventures in order to engage in

exploratory foray. Second, restructures who acquired large domestic banks in privatization process

and treat their investments rather as long-term commitment. Tschoegl (2003) also argues that as

foreign banks have no comparative advantage in retail banking vis-à-vis host country banks in the

long-run perspective, the acquisition of the domestic banks can be the only possible method to get in

this business and remain in it for certain, at least, medium term. In this sense, this mode of entry

gave the entering foreign banks much greater comparative advantage as setting up a branch or

subsidiary.

Tables 3 and 4 show the number of foreign bank entries into the CE countries in breakdown by

entry modes and entering countries during the period 1994-2004. As it can be seen, the M&As have

been the most favourite entry mode of the foreign banks into CE markets during the last years.

Capture

Capture

[Table 3] and [Table 4]

The high number of the yearly entries by M&As can be a result of the banking regulations and

restrictions imposed by the governments in the CE countries on acquisition of majority stakes in the

domestic banks and as well as other forms of entry. In the course of relaxing the restrictions, the

same foreign banks could further increase their stakes in the domestic banks. An entry via subsidiary

was the second most common mode of internationalization into the CE banking markets and

dominated over the other methods mostly at the beginning and middle 1990s.

Table 3 shows also that Poland had the highest number of foreign bank entries. However, as we

compare the assets of the foreign banks between individual banking sectors presented in the table 4,

we can observe that the Czech Republic and Slovakia are among the CE countries with the highest

share of the banking assets in the hands of the foreign banks. Furthermore, Table 5 shows that

foreign bank entries came mainly from the neighbours countries of the CE countries.

Capture

[Table 5]

5. Data and Methodology

This section describes our data set and the two econometric methods that we use to assess the

economic determinants of foreign bank expansion into the four CE countries. First, we employ

Poisson regression with our sample for the four CE countries and the OECD countries over the

1994–2004 period. Second, in order to evaluate the economic determinants and the entry mode of a

foreign bank into the CE market we use a bivariate probit model using our sample over the 1994–

2004 period. In our study we concentrate only on the OECD countries as almost all foreign banks

operating in the CE region were from the OECD member countries. All variables employed in our

analysis are presented in the Appendix.

5.1 Data

In our paper we evaluate the economic determinants of foreign bank entries and its entry

modes into the four local banking markets in CE. In order to analyze those markets we use yearly

data on countries and banks in the four CE countries, namely the Czech Republic, Hungary, Poland,

Slovakia for the period 1994-2004. These countries have shown widely different policies towards the

mode of foreign bank entry as we have presented above.

Our final sample contains 110 cross-border entries either by M&A or through setting up a

branch or subsidiary by a OECD foreign bank in one of the host countries. We established those

transactions using public information as national and international press coverage and compared it

with the list of foreign banks compiled by national bank supervisors.

In our study we define a foreign bank entry as to be followed by three forms: entry by setting up

a branch, subsidiary or/and via M&A.

We define a subsidiary or branch as a organisational form that received a domestic license or

approval by domestic bank supervisory institution. The transformations of the already existing

foreign banks, i.e. the transformations of branches into subsidiaries or vice versa are not considered

as entry and therefore are not included in our analysis. We argue that they can be driven by other

market determinants, which might not be observable for the non-existing foreign banks.

We define the entry through M&A as an acquisition of minimum of 5 per cent shares in a

domestic bank by a foreign banking institution as well as merger of domestic and foreign operation

in a host country. In our paper we are interested only in the horizontal foreign entry, which are

assumed to offer a broad potential for cost and profit efficiency improvements. Other types of

transactions, such as government owned banks or other financial institutions acquiring an bank are

excluded because they may be motivated by a different set of considerations. Moreover, our analysis

does not include mergers or acquisitions of the domestic banks with other domestic banks.

5.2 Poisson regression

In order to analyze entry decisions into the CE countries, we consider the number of entries

of foreign banks at time t into Poland, Hungary, Czech Republic and Slovak in breakdown by a

country of origin, conditioning on the specific groups of the regressors such as host-country

characteristics, physic relationship between host and home country and potential determinants of

entering. In contrast to other analysis, we are not strongly interested in the characteristics of banks

entering the CE countries as this area has been covered by many researchers whose work can also be

applicable to the four countries in our study.5 Hence, we are mainly interested in answering the

following questions:

a) How much did the host-country characteristics and in particularly macroeconomic conditions

matter in the entrance process of foreign banks into CE? Which of them did the foreign banks

consider to be the most important?

b) How much did the host-country banking regulations influence the number of foreign entries?

c) Which of the suggested in the section 2 determinants of banking internationalization did the

foreign banks mostly follow deciding on entry the CE countries?

Accordingly, we estimate the following choice model:

Capture

 

 

 

where y = number of entering banks from country i into country h at time t and Y1ht, Y2ht, Y3ht,… Y29ht

have independent Poisson distribution with parameters Capture

h = host countries (Hungary, Poland, Czech Republic, Slovak)

i = entries from the sample (OECD countries) defined together as home countries

K ht = a vector of variables specific to the host country

H hit = a vector of variables specific for the relationship between host country and home country

B it = a vector of variables specific for the home countries

We estimate a model with a Poisson specification controlling for some unobserved country-

and time-specific effects clustering the standard errors on the home country’s levels. Hence, our

error term has one or two components depending on the specification: µiht= εiht+αh or µiht= εiht+αh+θt.

We believe that a Poisson regression is the most appropriate specification of our model for

several reasons. First, most empirical studies analyzing entries of foreign banks and their activities in

the host countries use Ordinary Least Squares (OLS) as estimation methodology. However, it has

been shown that omitting the countries which do not participate in the foreign banking may lead to

inconsistent estimate parameters because of loosing information excluded from the sample. In such

cases, OLS estimates are biased towards zero (Greene 2000). Moreover, employing the OLS

regressions where the dependent variable is a count variable seems to be inappropriate as one should

explicitly account for this type of dependent variables and use the estimation techniques designed for

it (Maddala 1985). On the other hand, the non-linear methods allow us to take advantage of the

larger number of observations and reduce the biasness. It is very useful particularly, when one

investigates foreign bank entries into CE in the time-series context, where the number of individual

foreign entries is small or zero. In the cases, where there is preponderance of zero or small values

and the dependent variable is of discrete nature, we can improve on the least squares with a model

that account for these characteristics (Greene 2000).

Moreover, most studies examining the determinants of banking internationalization use due

to data unavailability either time-series or cross-section structure of the data. Both are faced to some

drawbacks, which do not allow us to take all results unambiguous. Since the cross-section studies

ignore the time-series dimension of the data which may result in the biasness of the estimates due to

omitting the country-specific effects, the time-series studies, on the other hand, besides their

attractive characteristics, suffer from the lack of availability of good-quality and sufficient length of

the data needed for the purpose of the time-series analysis.

The new panel data techniques enable us to control for these shortcomings. They allow us to

take advantage of the time dimension of the data as well as to estimate common relationships across

countries. By introducing the country dummy variables we allow for controlling for the effects of

those omitted variables that are specific either to individual CE country or are specific to each time-

period. In each regression, we test for their jointly significance.

At the end we show that our results are robust testing for significance of other explanatory

variables used in the literature examining banking internationalization.

As far as the determinants of the entry modes are concerned, we use a bivariate seemingly

unrelated probit specification. Unlike the other studies, we control explicitly for the correlation

between particular entry modes and test whether any organizational form stayed in direct

competition with others.

5.3 Bivariate Probit Regression

In our study we are also interested in the relation between economic determinants and entry

modes of foreign banks into the CE countries. In particular, we are interested in changes between the

determinants affecting particular organizational forms among the CE countries. Since the decisions

about particular mode of entry might be correlated at time t within a home country, we have chosen

a seemingly unrelated bivariate probit estimation (SURB), where the dependent variable is of the

binominal discrete nature either one or zero. Thus, the model takes a form:

Capture

where Yiht equals one when an entry via a particular form (M&A, branch or subsidiary) from country

i into country h occurs at time t versus an entry through another form from country i into country h

occurs at time t, otherwise zero.

As the equations are estimated simultaneously, we allow for the error terms to be correlated between

the entry modes. The other vectors are the same as defined in the first specification.

6. Results

This section presents the results of the Poisson regression and of the bivariate probit

regression. First, we present the descriptive statistics for our sample. Second, we discuss the results

of the Poisson regressions and we present the outcome of our robustness analysis. Finally, we show

the results of our panel analysis using the bivariate probit estimations.

6.1 Descriptive statistics

Table 6 – 9 provides summary statistics of our sample of OECD countries. Table 6 shows the

data representing economic characteristic of the 30 OECD countries in the period 1994-2004. In

addition, the table show the economic characteristics of those countries with no foreign direct

investment in CE, as well of those countries with foreign bank entry into the CE.

Capture

Capture

[Table 6]

In Table 7 we show the economic characteristics splitting the OECD countries sample using

our CE host countries: Poland, Hungary, Czech Republic and Slovakia.

Capture

Capture

[Table 7]

Similar, as in Table 6 we divided also the sample in countries with foreign entry into CE and

not. Table 8 presents the economic characteristics of those OECD countries without any foreign

direct investment in the financial services in the CE region.

Capture

[Table 8]

While, Table 9 shows the economic characteristic of the OECD countries with operation in

the the CE region.

Capture

Capture

Capture

[Table 9]

6.2 Poisson regression results

This section presents our Poisson regression results. In Table 10 we present the results for

foreign bank entry into the CE countries. We regress the dependent variable first against country

economic determinants and then progressively add our additional control variables. Table 6 shows

the results with a different set of independent variables in regressions (1) – (4). The Poisson

regressions reveal that some of our economic determinants may have a positive and statistical

significant impact on the entry decision of foreign banks into the CE countries.

In the regression (1) of the 1276 observations in the sample we lost 64 observations due to

the missing data on overheads in the home countries. In the regression (2)-(4) we lost additionally

261 observations because we missed some data on non-financial FDIs for Hungary. In all four

specifications we included dummies with respect to the host country in order to control for the

effects of those omitted characteristics which are specific to the individual CE countries. We test

also if those effects are significant and can explain the variations in the foreign banking between

these countries. Additionally, in regression (3) we added a time-effect and test if the determinants of

entering into CE have changed across time.

In the regression (1), the coefficients of two of three country-characteristics variables are

significantly different from zero. As expected, the tax rate is negatively correlated with the expected

number of foreign banks’ entries into CE countries, although, it seems not to be economic significant

in the regression. The reason might be that this variable may capture two opposite effects: (a) the

higher tax rate may discourage foreign banks to enter, especially by entry modes falling under the

local taxation; (b) the higher tax rate may encourage foreign entrants to choose particular entry

modes which gave a possibility to foreign banks to be exempted from local taxation or could

repatriate their profits to the parent banks.

As expected, the exchange rate shows a negative correlation with the expected number of

entries of foreign banks into CE and is highly significant. The negative sign of this variable may

indicate that with the depreciation of the foreign currency, the foreign banks started searching for

possibilities for great profits which occurred in the CE markets. The positive correlation between

inflation and the expected number of the foreign banks entering into CE was surprising. Yet, taking

into account that high inflation rates in the CE countries were associated with high net interest

margins, this variable may capture the effect of great profit opportunities on the CE banking markets

rather than its negative impact on the economy. This result is also consistent with other findings of

the literature on this topic. For example, Demirgüc-Kunt (1998) et al. find that inflation is associated

with higher realized interest margins and thus higher profitability of banks, especially in the

developing countries. This finding they explain by the fact that in developing countries demand

deposits frequently pay zero or below market interest rates. On the hand, in the period of high

inflation when the entries of most foreign banks occurred, the monetary policy targets of the CE

countries were already set and the specific measures to achieve them were identified.

The measure capturing the differences in the economic development between home and host

country turned out, on the other hand, not to be statistically significant, although it has an expected a

negative coefficient. It indicates that the higher the growth rate of the host country in comparison

with the home country, the higher the expected number of foreign banks entering the CE countries.

The size of the banking sector and financial sector appear to be statistical significant. Since

the bank deposits are positively correlated with the expected number of bank entries, consistent with

the hypothesis, the larger domestic banking sector, the greater number of potential clients and thus

better prospects for great profits. The second variable measuring the level of intermediation of a

country has a negative correlation with the banking internationalization into CE. We interpret this as

evidence that with the greater development of the financial sectors of the CE countries, it exists a

wider range of financial products and services outside a banking sector, and thus lower demand on

traditional banking products.

The coefficient of the concentration level of the banking sector is economically significant

and shows a negative correlation with the expected number of entering banks. Consistent with our

hypothesis, the result indicates that high concentration of the banking markets hinders new entries of

foreign banks.

The net interest margin is negatively correlated with the expected number of entries of the

foreign banks. The impact of this variable remains also statistically significant. The low interest

income of the banking sector suggested high potential of the credit markets in the CE countries and

thus great profit opportunities from the retail and wholesale banking for the new entrants.

The variable measuring the legal and regulatory structure of the CE banking markets, the

bank freedom index has, as expected, a negative sign, although it is insignificant. Possibly it is

because all CE countries were considered by foreign institutions to have similar regulatory structure

and other country-characteristics and location specific factors played a decisive role in an entry

process into a particular country.

On the other hand, the variable capturing the effects of the improvement in the creditor rights

is significant and negatively correlated with the expected number of banks entering the CE countries.

The reason could be that most of the foreign bank entries occurred in a period of poor creditor rights

protection.

The difference in the efficiency of the banking markets seemed not to be a driving factor to

an entry of foreign banks into CE countries although it indicates a negative sign. It means that the

higher the inefficiency of the banking market relative to the home market, the higher probability of

an entry. Possibly it is a result of opposite effects of this variable on an entry. Once more inefficient

banking markets may encourage M&A entries consistent with the hypothesis that foreign investors

may use their expertise in order to restructure inefficient banks, the inefficient banking markets may,

on the other hand, discourage greenfield investments.

The results of the regression provide also evidence that the law of origin of the entering

country is of great economic importance. The legal origin variables are significant even at the one

percentage significance level. This result is also consistent with other findings in the literature that

foreign banks are more willing to locate their operations in the countries which share the same legal

origin (Galindo et al. 2002).

Conditioning our regression on the dummy if a country belongs to the EU or not, we can see

that joining the EU exerts a negative effect on the number of entries of foreign banks into CE

countries. It is also statistically significant. This negative impact is possibly because of the fact that

since joining the EU creates many new opportunities, the banking markets of the CE countries had

been already penetrated by the foreign banks leaving the new entrants a limited room to step in.

Finally, consistent with other literature, our result shows that the distance between home and

host country is economic important in determining an entry decision. In the regression the variable is

statistically significant even at the one percentage level. The negative correlation with the foreign

banking suggests that banks from neighbouring countries were more expected to enter the CE’s

banking markets.

As a goodness of fit measure we perform Pearson test which in each specification was highly

insignificant suggesting that our data are indeed Poisson distributed (the results are not reported

here).

The regression (2) reports the estimates of the specification that includes non-financial FDIs

(lag), testing the hypothesis that the foreign banks were motivated to enter the CE countries by

following their customers and providing them their services on site. Although the coefficient of this

variable appears in the regression as insignificant, it has an expected positive sign. The inclusion of

the inflow of non-financial FDIs has resulted in some changes in the significance of the coefficients

as well as has changed the sign of one of the variables. We see that the estimate of the exchange rate

becomes positive, as we would expect, however insignificant. Possibly, because two different effects

may appear. In the (1) regression the exchange rate was significant at the one percentage level

because since we did not control for other motives of foreign banks’ entries into CE than motives

driven by great profit opportunities, the appreciation of the local currencies encouraged the foreign

entrants willing to take advantage from the strong currency. However, as we include the non-

financial FDIs and take into consideration the fact that depreciation of the currency creates great

prices for the foreign investors, many foreign banks followed the FDIs in the period of great

depreciation of the local currencies.

The insignificance of the concentration level of the banking sectors of the CE countries as

determinant of number of foreign bank entries can be explained possibly by two offsetting effects:

(a) the higher concentration level of the banking sector could have a negative effect on the number

of entries of foreign banks where the established retail and wholesale structure was desired; (b) the

higher concentration level could indicate under-banked and under-serviced markets and thus could

exert positive effects on foreign banks following their clients.

The economic significance of the differences in the growth rates between host and home

country after inclusion of the volume of the FDIs in the non-financial sector we interpret as the

evidence that the countries, which suffered from low economic growth were more expected to

search for the opportunities in the CE countries.

In the (2) regression the variable measuring the regulatory structure of the banking markets in

the CE countries becomes an important determinant of entry of the foreign banks into these

countries. We think that this is due to the fact that dummies with respect to the country of entry

remain jointly insignificant (compare regression (1), (2), (3) and (4)) rather than due to inclusion of

the non-financial FDIs. The reason might be that since the omitted characteristics between the CE

countries disappear, the entry into that country was more probable that imposed lower restrictions

on foreign banks’ entries.

Regression (3) presents the results of the regression after the inclusion of the time-specific

effects. We see that the results do not differ strongly from the ones of the regression (2). Interesting

is, however, the improvement of the significance level of the non-financial FDIs, which may suggest

that the foreign banks followed their customers entering the CE countries only at a certain point of

time. In order to test for it, we include in the next regression an interactive term. The parameter

estimates for three interactive terms are negative and highly significant suggesting that the “follow

the customer hypothesis” was not realized at the eve of the EU accession of the CE countries. The

parameter estimate for the non-financial FDIs on its own is, however, positive and highly significant.

It may indicate that the foreign banks followed their customers only at the beginning of the transition

process. We can also see that after inclusion of the interactive terms, two variables have changed

their significance. The exchange rate becomes significant at the one percentage significance level but

the net interest margin looses its economic significance. It may indicate that the foreign banks

following their customers could benefit also from the great depreciation of the currencies of the CE

countries. The insignificance of the net interest margin may suggest that retail and wholesale

banking activities gained an importance in the course of time. Finally, the results suggest that the

determinants motivating the foreign banks to enter the CE countries have changed across time.

CaptureCapture

[Table 10]

6.2.1 Robustness analysis

We next conduct a number of robustness tests. We test for the significance of other explanatory

variables, which may explain foreign banking entry and have been also presented in the literature.

Additionally, we check whether our results are robust to different econometric techniques. We begin

our robustness analysis with the regression (1) and include new control variables. In the regression

(6), we estimated our model with the ordered logit and in the regression (7) with Tobit. The

empirical literature points out some advantageous of these two econometric techniques for the type

of observations which we use in the regression. The results of new estimation are reported in the

Table 11. The results show that none of the covariates is significant even at 10 percentage level and

their inclusion does not affect our previous results.

CaptureCapture

[Table 11]

The national income per capita is used as a measure of a host country’s purchasing power and

thus demand for the banking services. The coefficient of this variable is negative and insignificant.

This is possibly because of the two offset effects. The measure of the country risk, as before, is

highly insignificant although shows a positive sign. It is consistent with the hypothesis the higher the

index (lower a country risk), the higher expected number of foreign banks entries. The coefficient for

the stock market as expected is positive suggesting that the foreign banks entries are positively

correlated with the stock market development. However, the variable is statistically insignificant.

Thus, we include instead the stock market capitalization and also this time the coefficient was

positive, yet insignificant. Finally, we included a measure of the size of country proxied by the

population of the host country. The coefficient of this variable was positive, but again insignificant.

6.3 Results of Bivariate Probit Regression

Capture

In the Table 12 we present results from a simple univariate probit estimation conducted on

the pooled data where we compare the coefficients for all types of entry modes. Each equation from

the previous regression (2) is estimated separately. The error terms are clustered on the home

country’s level.

To analyze the inferences in greater detail, we compare further the coefficients on

determinants affecting the presence of branches with respect to M&As and subsidiaries as well as

subsidiaries with respect to branches. This gives us a picture of comparative influences of the

different factors on the choice of entry mode. Moreover, by considering what interdependencies

between the banking forms are consistent with the empirical literature, the model shows how

different forms of banking activity compete or complement one another and what pattern of

competition or complementation between banking forms created the banking structure of the CE

countries. Unlike other studies on this presented in the literature, we employ the model which

directly control for it. The results are shown in Table 13.

In Table 14 we examine further the subsidiaries versus M&As, as these modes dominated in the

entrance process of foreign banks into CE. Moreover, as their activities could overlap to some extent

we feel that these forms could compete with each other in some areas. Moreover, in Table 14 in the

regression (2) and (3) we included also non-financial FDIs, although the literature on international

banking treats the trade variable and non-financial FDIs exchangeable as proxies for “follow the

customer hypothesis”. However, we follow Miller and Parkhe’s (1998) approach who argue that

since the bilateral trade can be positively correlated with different modes of entry, the non-financial

FDIs could explain the foreign entries through subsidiaries. We present in Table 14 in the regression

(3) the results of the regression with the country effect in order to examine if any unobserved and

omitted characteristics of the countries may explain additionally the differences in foreign bank

entries between the countries.

The results in Table 12 show that none of the entry modes stays in direct competition. It

might suggest that the foreign banks had different motives to establish their presence in the CE

markets. Hence, the organizational forms chosen by the foreign institutions could be considered as

complementary rather than as substitutes as suggesting the analysis from the developed countries

(Heinkel and Levi 1992). Moreover, the results may partly explain why the banking sectors in the

CE countries are overbanked but underserviced (Heinz 2004, OENB 2002, Bonin et al. 1998 ). It

may suggest that the level of the banking services were the same among the organizational forms of

foreign banks.

In the regression (1) in Table 12, the coefficients of country characteristics variables are in

most cases significantly different from zero. Reversely than our previous results, the tax rate appears

in the regression highly economic significant suggesting that it has an impact on bank’s

organizational form. The signs of the coefficients show, however, different signs. The positive sign

of the coefficient for branch regression indicates that since branches have an advantage in shifting

profits across borders, they were more likely in countries with the higher tax rates. This finding is in

line with the results presented in the literature (Cerulti et al. 2005). The other organizational forms

since they fall under the local tax regime, they were more likely in countries with lower corporate

tax rates. The variable proxy the risk of a country suggests interesting implications. It appears

significant only in two regressions, although of opposite signs. As we have expected and in line with

previous findings, branches were less likely in countries with high country risk as they are

considered to be the most sensitive to the local country conditions. Hence, the coefficient in a

regression where a branch is our dependent variable has a positive sign. In case of regression when a

subsidiary is a dependent variable, the country risk seems not to be economic important, although it

also shows a positive sign. The sign of the coefficient of the M&A regression is significantly

different from zero but opposite to the regressions with subsidiaries and branches it has a negative

sign. An explanation for that might be that many M&As’ deals occurred following the economic

crises in the host countries since these events provided great opportunities for favourable transaction

in terms of acquiring local banks.

The size of the banking sector and financial structure suggest very interesting implications.

Financial development matters mostly in foreign bank entries via branches and subsidiaries. In case

of regression where a M&A deal is our dependent variable only the coefficient of bank deposits is

positive and significantly different from zero. The result may confirm that as most M&As

transactions were driven by potential opportunities in the retail and wholesale banking, the

inheritance of the important clients were of great importance. On the other hand, the result might

suggest that foreign banks acquiring or merging with the domestic banks were more oriented

towards servicing large institutional and corporate clients, whom they inherited with the portfolios of

the domestic banks. The results for a branch regression are of totally different nature. The negative

signs of both coefficients may be explained by the fact that at the beginning the branches of foreign

banks entered in the early stage of countries’ development being driven by the privatisation

processes of enterprises, which partly took place via stock exchanges. Most branches of the foreign

institutions were involved in a big portion of these transactions rendering investment banking

services. With the development of the stock markets as well as growth of the private sector, the

branches of the foreign banks extended the scope of their activities offering variety of products

related to the money and capital markets. Thus, controlling also for the stock market capitalization,

the coefficient of this variable is highly significant and exerts a positive sign. This result supports

Heinkel and Levi’s (1992) hypothesis that setting up a branch of a foreign bank should be justified

by sufficient activities in the area for which a branch offers an advantage. The regression for

subsidiaries shows however different results. The signs are exactly reverse than the signs of the

coefficients of the M&As’ regression and are significantly different from zero. The positive sign of

the liquid liabilities suggests that with the development of the financial sectors the new opportunities

for subsidiaries of the foreign banks emerged. The stock market capitalization appears in the

regression as insignificant. Thus, it seems that the stock market activity did not determine the set up

of subsidiaries by a foreign banks in a local banking market. The negative sign of the coefficient for

the M&A regression may again confirm the entering foreign banks in the period following the

financial crises, where the activity of the stock markets tends to decline.

The variable measuring the concentration level of the banking sector shows in two cases

positive signs and in case of the M&A regression a negative sign. Besides the subsidiary regression

where the variable is statistically insignificant, the other coefficients are highly economic significant.

The positive sign of the coefficient for the branch regression might be explained by a different scope

of activities, mostly in investment and corporate banking, rendered by this form of the foreign

institutions. They have not stayed in the direct competition to the ones serviced by the local banks.

Higher concentration of the banking sector meant the dominance of several local institutions in the

credit market. Such a structure of the banking markets gave other foreign institutions the possibility

to gain their shares in other fields. It might be especially true for the developing economies, where

the markets are unsaturated and the development of the financial sectors forces other financial

products and services to be strongly desired. Another picture presents the regression in which a

M&A deal is considered to be our dependent variable. The coefficient of the concentration level with

a negative sign may suggest that in the markets where the local banks enhance their market power,

the states banks were less willing to sell their stakes for foreign institutions.

Many foreign institutions entered the CE banking markets in order to provide their home

clients with the services on site. In the beginning the banks restricted their activities to trade services.

The positive and significant variable for the subsidiary regression may indicate that mostly

subsidiaries were the modes of entry chosen by foreign banks to service their clients. The

insignificance of the coefficient of the M&A regression may suggest on different types of clients

followed. Since subsidiaries might follow multinational companies, entries of foreign banks through

M&A of the local banks could and hence might indicate that they might service larger companies.

The coefficient of the trade variable for the branch regression shows as expected a negative sign and

it is insignificant.

Interesting implications suggest the variable measuring the distance between the parent bank

and its presence in the home country. Although, the coefficients from all three regressions show a

negative sign, as expected, only for branch specification, it is significant. The result is not surprising

as most of the entries through M&As and subsidiaries occurred among European banks, all CE

countries can be considered to be of comparable distance.

The first conclusion that we can draw with regard to the banking regulations for all three

regressions is that higher banking regulations hinder the foreign banks’ entries. However, only in the

regression (1) for branches and (3) for M&As, the coefficients of the variable are highly significant

from zero. The reason is that the branches and M&As were the organizational forms which were

regulated differently among the CE countries. The results of the influence of the creditor rights

confirm our findings from the first regression.

The coefficients of the net interest margin present the same signs for all the regression

specifications, yet it differ in their statistical significance. The negative sign of the net interest

income may indicate increasing competition. Increasing competition may explain why we observe

the statistical significance of the coefficient for a M&A regression. Thus, this method of entry may

be preferred foreign banks with the aim to reach a strong position in the local markets in a short

period.

The literature on the international banking refers very often to the income per capita as a

variable measuring a host countries’ purchasing power and thus demand for the financial services

(Buch and Lapp 1998, Buch 2000, Sagari 1992, Yamori 1998) and largely find a positive relation to

foreign banking assets or FDIs. Our result however indicates that the impact of this variable depends

on type of the organizational form chosen by a foreign institution. We find that the branches were

more likely in the wealthier countries once M&A in the poorer. For the subsidiary regression, the

income per capita seems not to have an explanatory power.

The difference in the growth rates between home and host countries indicates that the lower

the difference (the higher the growth rate of the host country), the higher the probability of an entry

into a host country. The result is valid for all regressions, although the coefficients for the branch

and M&A regressions are not statistically significant.

Table 13 analyses the differences in a greater detail by an econometric comparison of the

presence of branches in respect to subsidiaries and M&As, while Table 14 show the comparison of

subsidiaries in respect to M&As.

Capture

[Table 13]

As we can see from the Table 13, the examination of the branches with respect to subsidiaries

and M&As supports that the branches do not stay in any direct competition to other banks’

organizational forms in the CE countries. The results fully reflect the findings from the simple probit

regression presented in the Table 12. The one difference is the insignificance of the bank freedom

coefficient in the branch regression versus M&As one. The reason might be due to two offsetting

effects: (a) the branches were more likely in countries with lower regulations on branches; (b) in

countries where the higher regulations on branches applied, the foreign banks chose an entry via

M&A instead.

The regressions in Table 14 confronting the subsidiaries versus M&As indicate, on the other

hand, more interesting implications. Already the result of Wald test suggests on some correlation

between M&As and subsidiaries. This comes from the significance of the coefficient of trade in case

we confront the entries through M&As with subsidiaries. Once we consider them separately, the

coefficient of this variable in the M&A regression seems to appear as economically unimportant.

This could suggest that subsidiaries of foreign banks could compete with the acquired local banks

for some clients, possibly larger multinational companies. After the inclusion of the non-financial

FDIs, we see that this variable is significant only in the M&A regression suggesting that since the

acquirers were large international banks, they could also follow large clients engaged in various

investments in the CE markets. We can also see that inclusion of the FDIs modify slightly our results

in regression (3) supporting the results of Miller and Parkhe (1998) that different forms of entry are

positively correlated with following specific clients. Moreover, the result suggests that since the

M&As and subsidiaries of foreign banks competed to the some extent, they reacted similarly to some

effects.

The results in Table 14 regression (3) seem to reflect the results from regression (2), however

the bank freedom index and country risk variables became insignificant for both M&A and

subsidiaries regressions. This may support our previous finding that controlling for the omitted

country characteristics and location specific factors, the differences between the CE countries in

terms of banking regulations and country risk do not explain different modes of entries chosen by

foreign banks.

Capture

[Table 14]

7. Conclusion

In the last decade we have witnessed a great influx of foreign banks into the CE countries.

The share of foreign bank assets rose from below 20 per cent to almost 80 per cent in all these

countries. We assume that the foreign banking in the CE will continue to expand, albeit at a slower

pace.

The literature on international banking has identified several factors that influence the

location choice of foreign banks. With this paper we add to this literature twofold: first, by

examining the determinants of banks’ choice – whether and where to expand abroad and second, by

determinants of organisational form chosen by foreign banks. Our empirical results show that

macroeconomic and institutional determinants influenced significantly a foreign bank’s decision to

expand the CE countries. We find that the foreign institutions were mostly attracted by large

potential of the CE banking markets and low degree of their financial sophistication. This finding

stays on the contrary to the results from the developed countries, where the foreign banks are more

likely to expand the countries with a high level of financial and banking system development.

According to these studies, only such markets offer more efficient banking product opportunities.

Our results do not support this view. They rather suggest that less developed financial systems offer

a wider range of possibilities for foreign banks to achieve great profits. Moreover, we show that in

the beginning of the transition process, many foreign banks decided to enter the CE markets simply

by following their clients. In the course of financial development occurring in the CE countries, the

new opportunities emerged and bank’s “follow the customer“ behaviour has been replaced by a

search for client’s behaviour. We also find that most foreign banks entries occurred in the poor

creditor rights protection. However, the legal origin of the home country was of great economic

importance. We show that common law countries as well as countries with German and French law

traditions were the most likely to enter the CE banking markets than other legal families. Finally, in

line with other studies, our results suggest that most banks stem from the European countries. This is

confirmed by a negative and significant coefficient of the distance between a host country and a

foreign bank’s headquarter.

We also looked on the modes of foreign bank entry and its relationship to the economic

determinants. We find that the choice of organizational form of a foreign bank depend strongly on

the economic characteristics of the host country. Moreover, consistent with previous studies, we

show that the decision on mode of entry is determined by a scope of activities a foreign bank is

going to render in a host country as well as by a type of client followed.

Our results are important from a research point of view. They expand the previous literature

on different economic factors encouraging foreign bank entries. They introduce a wider set of

explanatory variables than previous studies on this topic. They also shed a light how different factors

influence various organizational forms of entries, thus fitting in the growing literature on the role of

foreign banks in the host countries’ financial systems.

The results are also important from the policy perspectives. For one, they show that high

level of financial sophistication and strong creditors’ rights protection are not a necessary

prerequisite to attract foreign banks. Since the foreign banks have been shown to have a beneficial

influence on domestic financial systems, it seems to be good news, especially for developing and

other transition countries whose financial systems are poorly developed and the access to the

financial services is constrained. Furthermore, since the behaviour of foreign banks changes with the

development of financial systems, this suggests different implications for the supervisory authorities.

Replacing the foreign banks’ behaviour by “searching for clients’ behaviour” might encourage

foreign banks to take on excessive risks which might cause important consequences for the stability

of the host countries’ financial systems and thus might require specific policy responses.

Finally, our results suggest further agenda for research. One area for further research would

be to investigate the impact of mode of entry and organizational form on foreign bank behaviour in

the developing countries in more detail. Can the fact that the various organizational forms react

differently on the location-specific factors be explained by different behaviour of these banks in the

host countries? If yes, in what activities they engage in. Would it be true that foreign banks entering

through cross-border mergers and acquisitions are more oriented towards large retail and wholesale

clients neglecting lending to small businesses? The findings from the developed countries suggest

so. Additionally, the results from the developed countries suggest that subsidiaries of foreign banks

are more likely to engage in the retail and small-business lending and hence promoting greater

access to the financial services. The picture of the CE countries, however, shows that many

subsidiaries of foreign banks operate in the niche business providing such services as car loans or

mortgages. Thus, the scope of the activities of the foreign subsidiaries in the developing countries

requires a deeper investigation. And finally, does a lack of branches have any consequences on the

development of the financial systems in the CE countries? Several studies indicate that banking

sectors of the transition countries are overbanked but underserviced. It would be interesting to find

out the reasons for this.

Related to this, it would be interesting to look at banks’ determinants in the decision process

of whether to enter through de novo operation or cross-border mergers and acquisitions. Since the

behaviour of foreign banks may vary in the financial distress, the regulators should monitor what

banks are likely to choose subsidiaries rather than branches as an entry mode into a host country.

None of these areas has yet much been studied for developing countries.