1
Foreign banks and investment in developing countries
Name: Tim Parée
Student number: 10422773 Program: Economic and Business Track: Finance and Organization Supervisor: Robin Döttling
Abstract
I study the effect of foreign banks on investment in 86 developing countries during the period 1995-2005. I use an existing dataset of foreign bank presence in developing countries. This dataset enables me to search for a difference in effect of foreign banks from
developing countries and developed countries on investment. I find evidence that foreign banks have a positive effect on investment in developing countries. On the other hand I do not find clear
evidence for a difference between the two types of foreign banks. With an instrumental variable regression I show that the results of this research may be affected by simultaneous causality bias.
2
Introduction
During the last decades foreign banks from developed countries have entered the financial markets of developing countries. In the period between 1995 and 2009 the foreign bank concentration almost doubled in developing countries. This trend was caused mostly by worldwide bank deregulation and globalization (Claessens et al., 2008). Asian countries fall behind but recently China1 and Philippines2 eased their restrictions on foreign bank openings. So far most literature about foreign banks focuses on the direct effects and determinants of banks for going abroad. However, research about the indirect effects on macroeconomic factors of foreign banks is still limited. Nevertheless, this kind of literature could be of relevance for government policy of developing countries regarding bank regulation.
The purpose of this paper is to describe the effect of the entrance of foreign banks on investment in developing countries. I use an existing database of Claessens et al. (2008) as indicator for the variable of interest foreign bank assets and use data of several host macroeconomic factors as control variables. With this data I run a regression on the investment level as a percentage of GDP during the period 1995-2005. Besides that I try to find a difference in the effect between foreign banks from developing countries and developed countries.
Existing literature shows the role of investment in the economy and indicates the relevance of it. Xu (2000) has shown that financial development might lead to a higher investment level and that this could increase economic growth. In a cross country regression on 69 developing countries Borenztein et al. (1998) found similar results as Xu, but they added the condition that the level of human capital must be higher than a certain threshold. King & Levine (1993a; 1993b) found in two complementary studies that financial
development may lead to higher growth of GDP, technological innovation and productivity growth via the channel of investment. They used a cross-sectional country analysis of 80 countries from 1960 till 1989.
1 James, S. (2014, December 21). China eases restrictions on foreign bank branch openings. Bloomberg news.
Retrieved from www.bloomberg.com.
2 Martin, K.A. (2014, December 8). Entry of foreign banks seen next year. The Philippine star. Retrieved from
3 The increase of foreign banks in developing countries is well represented in the economic literature but most research has just focused on single countries. Hasan & Marton (2003) concluded for Hungary, Barajas et al. (2000) for Colombia and Kraft et al. (2006) for Croatia that foreign banks are more efficient and have higher profits. Other research documented a significant credit growth in emerging European countries partly caused by the entrance of foreign banks (Popov and Udell; 2012). Cross-country analysis from Claessens et al. (2001) and Berger & Udell (2006) concluded that the presence of foreign banks might lead to financial intermediary development by increasing the availability of funds at favorable rates and introducing new efficient financial products in developing countries. In developed countries the opposite is true and there domestic banks seem to outperform foreign banks. An analysis of 35 developing countries showed that enterprises report that they are facing lower financing obstacles in countries with higher levels of foreign bank presence (Berger et al., 2005). Lower financing obstacles and financial
intermediary development may lead to an increase of investment in developing countries. This supports the hypothesis that foreign banks positively affect investment. This hypothesis is closely related to the central research question in this paper is:
Is there an impact of foreign banks on investment in developing countries? In this paper I find clear evidence that the entrance of foreign bank has impact on investment in developing countries
Some papers contained extended amounts of collected data on the entrance of foreign banking (Claessens et al., 2008; Micco et al., 2007). This data shows the increase of foreign banks during the 90’s in Eastern Europe and Latin America. Claessens & van Horen (2012) concluded that the efficiency and success of foreign banks largely depends on the wealth and development of the home country of the foreign bank. This supports the hypothesis that the effect of foreign banks from developed countries on investment is bigger than the effect of foreign banks from developing countries. My results are not in line with this hypothesis. My findings show no clear evidence that the foreign banks from developed countries have higher effect on investment than foreign banks from developing countries.
4 The paper is structured as follows: Section 1 explains the data and the methodology. Section 2 shows the results of the regressions and explains them. Section 3 concludes.
Data & Methodology
My main source of data is the World Bank which gives open access to economic data for all countries. The dataset consists of 900 indicators for 210 nations over more than 50 years. It provides the data for the dependent variable investment and the host country’s macroeconomic variables stock market size, bond size, inflation rate, taxes, savings,
exchange rate, real interest rate, savings and GDP growth per capita. To represent the level of investment in the developing countries I use the gross fixed capital formation. It
measures the amount of acquired fixed assets or land improvements by both the public and the private sector. It does not make any adjustments to deduct consumption or
depreciations and excludes financial assets and land sales. The GFCF is available for all relevant data points and is presented as a percentage of the GDP. It gives a clear picture of the investment culture and is therefore a good representative for the investment level. To estimate the GFCF I use the following panel regression:
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑐𝑐𝑐𝑐 =
𝛽𝛽 0 + 𝛽𝛽 1 ∗ 𝐺𝐺𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹 𝑏𝑏𝑏𝑏𝐹𝐹𝑏𝑏 𝑏𝑏𝑎𝑎𝑎𝑎𝐹𝐹𝑒𝑒𝑎𝑎 𝑐𝑐𝑐𝑐 + 𝛽𝛽 2∗ 𝑆𝑆𝑒𝑒𝐹𝐹𝑆𝑆𝑏𝑏 𝑚𝑚𝑏𝑏𝐹𝐹𝑏𝑏𝐹𝐹𝑒𝑒 𝑎𝑎𝐹𝐹𝑠𝑠𝐹𝐹 𝑐𝑐𝑐𝑐 +
𝛽𝛽 3 ∗ 𝐼𝐼𝐹𝐹𝑆𝑆𝐹𝐹𝑚𝑚𝐹𝐹 𝑒𝑒𝑏𝑏𝑡𝑡𝐹𝐹𝑎𝑎 𝑐𝑐𝑐𝑐 + 𝛽𝛽 4 ∗ 𝑆𝑆𝑏𝑏𝑆𝑆𝐹𝐹𝐹𝐹𝐹𝐹𝑎𝑎 𝑐𝑐𝑐𝑐 + 𝛽𝛽 5 ∗ 𝐸𝐸𝑡𝑡𝑆𝑆ℎ𝑏𝑏𝐹𝐹𝐹𝐹𝐹𝐹 𝐹𝐹𝑏𝑏𝑒𝑒𝐹𝐹 𝑐𝑐𝑐𝑐 +
𝛽𝛽 6 ∗ 𝐼𝐼𝐹𝐹𝐼𝐼𝐼𝐼𝑏𝑏𝑒𝑒𝐹𝐹𝐹𝐹𝐹𝐹 𝐹𝐹𝑏𝑏𝑒𝑒𝐹𝐹 𝑐𝑐𝑐𝑐 + 𝛽𝛽 7∗ 𝐺𝐺𝐺𝐺𝐺𝐺 𝐹𝐹𝐹𝐹𝐹𝐹𝑔𝑔𝑒𝑒ℎ 𝑐𝑐𝑐𝑐 𝛽𝛽 8 ∗ 𝐵𝐵𝐹𝐹𝐹𝐹𝐵𝐵 𝑚𝑚𝑏𝑏𝐹𝐹𝑏𝑏𝐹𝐹𝑒𝑒 𝑎𝑎𝐹𝐹𝑠𝑠𝐹𝐹 +
𝜀𝜀 𝑐𝑐𝑐𝑐 + ƞ 𝑐𝑐𝑐𝑐
The foreign bank assets variable is provided by the dataset of Claessens, Van Horen, Gurcanlar and Mercado (2008). They make a clear distinction between to different kind of foreign banks. South-South foreign banks are banks whose home countries are other developing countries and North-South foreign banks are banks whose home countries are developed countries. Their extended dataset contains the foreign bank assets as a
percentage of the total banks assets and South-South foreign banks assets as a percentage of total foreign bank assets. The dataset shows the proportion of foreign bank assets and
5 not the absolute size. This reflects the foreign bank assets better because it depends less on a country’s size or welfare. I transformed the data of the dataset to three groups: total foreign banks assets, South-South foreign bank assets and the North-South foreign banks assets. All groups are as a percentage of the total, domestic and foreign, bank assets. Most of the banks in the dataset are commercial banks but also savings, cooperative and credit banks are included. A bank is considered to be foreign if more than 50% of the total amount of shares is held by foreign entities. The criterion for determining the ownership of the shares is direct ownership. Indirect ownership is only used if there is a clear indication that the direct ownership is just for tax purposes. Their dataset consists of 103 developing countries over the years 1995-2005. I removed the 25 countries without any data about their GFCF. I also remove the observations from Albania because their foreign bank assets rose from 0 in 1996 to 90% in 1997, dropping to 0% again one year later. This is probably caused by the Albanian rebellion of 1997. These observations are not representative and I remove them therefore. The remaining dataset consist of 86 countries from all major regions and are displayed in the appendix. The pie chart shows the spread of the 86
countries over five continents. The period I used is 1995-2005, which gives me a total of 946 data points. To test the difference between the effect of foreign South-South banks and North-South I run, besides
the regular regression with the total of foreign bank assets, three extra regressions with both types of foreign banks separately and together as variables of interest.
The dataset contains data in two
dimensions: different countries and different years. A panel data analysis is the right way to deal with such a dataset. The panel data analysis is recognizable in the regression by the 𝑐𝑐𝑐𝑐
and the extra error term ƞ 𝑐𝑐𝑐𝑐 . The 𝑐𝑐 refers to the different countries and 𝑐𝑐 refers to the
6 fixed effect model and the random effect model. To decide which model to use I run a Hausman test. This test rejects the hypothesis that the individual-level effects are
adequately modeled by a random-effects model. So I choose to use the fixed effect model. Host country macroeconomic variables are stock market size, bond market size, taxation, savings, exchange rate, inflation rate and the GDP growth. Stock market size is added to this regression because it is another way to finance investments besides bank loans. High investment levels can be a result of large stock markets (Henry, 2000). So stock market size should have a positive effect on the GFCF. The stock market size is measured by the market value of all listed domestic companies divided by the total GDP and provided by the World Bank. A company is considered to be listed when it is part of the country’s stock exchange at the end of the year. This excludes investment companies, mutual funds and all other collective investment vehicles. The dataset contains only 642 of the 946 data points. Bonds are frequently used to finance investments, besides bank loans and stock issuance. Therefore I add the variable bond portfolio investment which represents the bond market size. The bonds are defined as securities issued with a fixed rate of interest for a period of more than one year. This variable consists of both public guaranteed and private non-guaranteed bonds and is measured in U.S. dollars. I divide this by the total GDP of the countries and multiply it with 1000 to make it comparable with the other control variables. In total 799 relevant data points are available in the dataset of the World Bank. Existing literature shows that the tax rate is an important determinant of investment (Jorgensen, 1963). A High tax rate on income and profits lowers the return on investment and by that it takes down the incentive to invest. My hypothesis is that the tax rate has a negative impact on the GFCF. However, a good measurement for taxes during the period 1995-2005 is hard to find. Tax level and regulation differ for all countries. I use the taxes on income, profits and capital gains divided by the total amount of taxes. This indicator is incomplete and has only 459 useful data points. This restricts the amount of observations in my regressions. Because of this restriction I will run regressions with taxes and regressions without taxes. The
interest rate is highly negatively correlated with the investment level and so should be crucial in the regression. As an indicator for the interest rate I use the real interest rate adjusted for inflation by the GDP deflator. The World Bank has 747 relevant data points for the real interest rate available. However, the literature overview shows that foreign banks
7 may decrease the interest rate. So adding an interest rate variable to the regression will diminish the effect of foreign banks on investment via the channel of the interest rate. To solve this I try to replace the interest rate variable in the regression by a variable which represents the savings. Savings has a high correlation with both the interest rate and the investment level. My hypothesis is that savings will have a positive coefficient in the
regressions. As indicator for savings I use the gross domestic savings as a percentage of GDP. Gross domestic savings is the total GDP reduced by the total consumption. All 946 data points are available in the dataset of the World Bank. Klein & Rosengren found evidence that the exchange rate has a large effect on investment by influencing foreign direct investments. To reflect the exchange rate I use the real effective exchange rate index with base year 2010. This measures the value of a currency against a weighted average of several foreign currencies and is divided by a price deflator. The dataset is also very incomplete and consist of 462 of the 946 useful data points. According to the previous mentioned research of Jorgensen (1963), investment is also effected by the domestic price level. To reflect the price level in the regression I use the inflation rate, which is measured by the growth rate of the GDP implicit deflator. The GDP implicit deflator is the ratio of the GDP in the current local currency to the GDP in the constant local currency. The dataset contains data about every relevant country and year. Inflation reduces the real value of debt. By that high inflation reduces the price of financing with debt. This creates an incentive to invest. On the other hand inflation reduces the returns on investment. The effect of inflation on
investment is ambiguous and unclear. My results show that the coefficient of inflation is negative. So the effect of reducing the returns on investment overtakes the effect of reducing the real value of debt. The majority of the variables in this regression are divided by the total GDP. However, GDP might influence investment in an exponential way. Firms with higher profits might relatively spend more on investment than firms with lower profits. My hypothesis is that GDP has a positive impact on the GFCF. To solve this I add the variable GDP growth per capita, which is defined as the sum of gross value added by all resident producers in the economy and is divided by the midyear’s population. All data points are available in the dataset of the World Bank. Claessens & van Horen (2012) did research on determinants of success of foreign banks and they found that the distance between the home and the host country of the foreign bank is very important. However, my dataset does not provide any information about the home country of the foreign banks besides the
8 information if their home country is developed or developing. This makes it impossible to determine the distance between the home and the host country and by that to include a distance factor.
Summary variables
Variables Observations Mean St. Dev. Min Max
GFCF (% of GDP) 946 20.1 6.3 2.0 58.96
Foreign bank assets (% of total bank assets)
946 35.7 32.11 0 100
Foreign bank assets (N-S, % of total bank assets)
946 25.2 27.5 0 100
Foreign bank assets (S-S, % of total bank assets)
946 12.4 23.3 0 100
Stock size (% of GDP) 642 27.3 38.7 0.0 304.6
Bond size (% of GDP*1000) 799 4.0 16.1 -45.8 241.5
Inflation rate 946 15.6 50.3 -27.0 987.1
Firm taxes (% of total taxes) 459 29.0 13.0 2.8 93.7
Savings (% of GDP) 946 15.9 12.6 -38.5 59.3
GDP growth per capita 946 2.9 5.4 -17.2 90.9
Exchange rate index 462 93.2 26.0 40.2 270.0
Real interest rate 747 11.3 19.1 -70.4 252.1
A problem for the internal validity is simultaneous causality. If foreign banks have an effect on investment then vice versa a high level of investment might also attract more foreign banks. To show the possible presence of a simultaneous causality bias I run an instrumental variable regression. As instrument variable I use the denial rate of applications for commercial bank licenses or acquisitions of domestic banks by foreign banks. I found this variable in a dataset designed by Barth, Caprio and Levine (2004) and it is available on the World Bank website. They collected data about bank regulation and supervision in 107 countries. I combine their data about the denial rate of applications for commercial banks and acquisitions of domestic banks by foreign entities. The denial rate of application has a correlation of -0.2755 with the foreign bank assets variable but isn’t affected by investment. So it fulfills the requirements of an instrument variable. To reshape the data I first delete all data of developed countries and then I remove all the data of developing countries without
9 any application. After that there are still 48 countries left. With the denial rate and the host country’s macroeconomic factors I predict the foreign bank assets and run an IV regression.
Results & analyzes
Table 1 shows the results of the panel analysis. The explanatory variable of regression one is the total of foreign bank assets with roots in both developed and developing countries. Regression two differs from regression one in the way that the taxes variable is excluded. In regression three the variable foreign bank assets is split up into the two variables North-South foreign banks and North-South-North-South foreign banks. Regression four encloses only the North-South foreign bank assets variable while regression five encloses only the South-South foreign bank assets variable. Regression six is the IV regression with the denial rate of applications from foreign banks as instrumental variable.
Table 1: Results of panel data analysis GFCF Foreign bank
assets (1)
Foreign bank assets without taxes
(2)
North-South & South-South banks (3) North-South banks (4) South-South banks (5) IV-regression with denial rate
(6) Foreign bank assets
Foreign bank assets (North-South) 0.0371*** (0.0185) -0.0094 (0.0168) 0.0178 (0.0186) 0.0234 (0.0204) Foreign bank assets
(South-South) 0.0977 (0.0976) 0.1111 (0.1038) Predicted foreign bank assets 0.0144 (0.0295) Taxes -0.0521*** (0.0281) -0.0540*** (0.0314) -0.0538*** (0.0316) -0.0522*** (0.0319) -0.0052 (0.0391) Savings 0.1500* (0.0844) 0.0873* (0.0694) 0.1534*** (0.0816) 0.1535*** (0.0848) 0.1572** (0.0771) 0.2249 (0.0622)
10 Stock market size 0.0306**
(0.0113) 0.0391* (0.0201) 0.0312* (0.0113) 0.0315* (0.0113) 0.0316* (0.0112) -0.0169 (0.0146) Inflation rate -0.0049*** (0.0025) -0.0079** (0.0020) -0.0056** (0.0026) -0.0058** (0.0025) -0.0063** (0.0025) Exchange rate index
Number of observations 0.0571* (0.0360) 201 0.0289* (0.0146) 358 0.0579*** (0.035) 201 0.0590*** (0.0370) 201 0.0582*** (0.0365) 201 0.1611* (0.0300) 103 R-square overall 0.2451 0.11003 0.2061 0.2447 0.1849 0.2358
The variable of interest foreign bank assets in regression one is significant with a p-value of 0.053. So there is a relation between foreign banks and investment in developing countries. The coefficient has a value of 0.0371. This means that if foreign assets increase their market power with 1% the investment level will raise with 0.0306% of the total GDP. In regression one the taxes variable has a significant effect with a significance of 10%. Leaving the variable taxes out of the regression drastically changes the value of the foreign bank assets to a lower value. The same drop in coefficients holds for the other regressions, which are included in the appendix, table 3. This can be explained in two ways. Firstly, the taxes variable has a negative value in the first regression. When I leave the taxes variable out, the foreign bank assets variable can take over some of this negative value if they are correlated. The correlation between taxes and foreign bank assets is -0.1181. The second explanation is the change of observations. Because the data of taxes is limited, leaving out taxes will increase the amount in observations from 201 to 358. The new pool of observations may affect the coefficient of the foreign bank assets. Table 5 shows the countries which
observations were excluded by lack of data on taxes in regression one. There seems to be no system in these countries, so I assume they are completely random.
Regression three, four and five show that the effect of South-South banks on investment is bigger than the effect of North-South banks. The p-value is almost the same, but this is due to a bigger deviation in the effect of South-South banks on investment. Testing if both variables are equal gives a p-value of 0.4244. This means there is no
significant difference between the effects of both kinds of foreign banks on investment. This
* Corresponds to significance of 1% ** Corresponds to significance of 5% *** Corresponds to significance of 10%
11 is not in line with my hypothesis, which states that North-South banks have a bigger effect. This might be caused by a missing distance factor. Globally, the distance between two developing countries is on average smaller than the distance between a developing and a developed country. If distance has a negative effect on investment then this affects North-South banks more than North-South-North-South banks. This could probably be offset by adding a distance indicator. However, my dataset does not contain any information about the home country of foreign banks besides the distinction if their home country is developed or developing. This makes it impossible to determine the distance between the home and the host country and by that to include a distance factor.
The extended version of my regression with all the insignificant control variables removed is presented in the appendix in table 2. GDP growth per capita has a p-value around 0.154 so I remove this variable from the regressions. I also remove the control variable bond size which had a p-value of 0.317. Removing these variables increase the amount of observations from 159 to 201 and the coefficient of foreign bank assets from 0.0263 to 0.371. Table 4 in the appendix shows the regression results with savings replaced by the real interest rate. This lowers the coefficient of foreign bank assets from 0.0371 to 0.0262 and the amount of observation from 201 to 164. This is in line with my hypothesis. Surprisingly the interest variable has no significant effect on the GFCF. This is partly caused by the high negative correlation with the inflation rate. The interest rate variable is indeed adjusted for the inflation. However, even when I remove the inflation rate variable the p-value of the real interest rate is 0.156. This may be because other control variables are also effected by the interest rate and have high correlation. By this they can take over the effect of the interest rate. The table shows the correlation between the interest rate and other variables.
Correlation with interest rate
Stock market size Exchange rate index Taxes Foreign bank assets
Real interest rate -0.1849 0.1370 -0.2441 0.0767
A comparison between regression one and five shows the simultaneous causality bias. The amount observations dropped from 201 to 103. This makes the estimation of regression five less secure. At the same time using a dependent variable without any
12 correlation with the error term partly causes a drop of the coefficient from 0.037 to 0.014. This gives an indication that there may occur simultaneous causality bias in regressions one till five.
Conclusions
This research studied the effect of foreign banks on the investment as a percentage the GDP of 86 developing countries during the period 1995-2005. I found a positive relation between foreign banks and investment. If more bank assets in a developing country are possessed by foreign banks this may lead to more investments. This result is in line with my expectation following from existing literature. This finding may be an incentive for the governments of developing countries to take down there bank regulation. On the other hand no clear evidence was found for a difference in effect between the South-South and North-South foreign banks. This is in contradiction with suggestions of existing literature. Adding a distance factor could explain the equal level of North-South and South-South foreign banks. Unfortunately data about the distance between the host and home country of the foreign banks is unmeasurable. There are a few more restrictions and limitations for my results. The amount of observations is very restricted by a lack of data. Data about the exchange rate and taxes are very limited for developing countries in the relevant period. New or more extended datasets may improve the quality and validity of the results. Another limitation is the possible presence of simultaneous causality bias. The IV regression suggests that banks may be attracted to enter a foreign bank market by high investment levels. Finally, I have some suggestions for new research. Previous research found that foreign banks cause a credit growth in the host countries. I try to find an implication of this credit growth in financing investment of firms. A suggestion for new research is to find an implication of the credit growth in financing investments of households, like mortgages or car sales.
13
Appendix
Table 2: Panel analysis with all control variables
GFCF All control variables included (1)
Foreign bank assets 0.0264
(0.0163)
Taxes -0.0599***
(0.1067)
Savings 0.1340***
(0.1267)
Stock market size 0.0392*
(0.0101)
Inflation rate -0.0044***
(0.0027)
Exchange rate index 0.0502***
(0.0324)
Bond market size -0.1298
(0.1272)
GDP growth per capita 0.0565
(0.0385)
Number of observations 159
R-square overall 0.2508
Table 3: Taxes left out GFCF North-South banks (1)† North-South banks without taxes (2)† South-South banks (3)† South-South banks without taxes (4)† Foreign bank assets
(North-South)
0.0234** (0.0204)
-0.0157 (0.0163) Foreign bank assets
(South-South) 0. 1111 (0.1038) 0.0443 (0.0558) Taxes -0.0538*** (0.0316) -0.0522*** (0.0319)
14 Number of
observations
201 358 201 358
R-square overall 0.2447 0.1031 0.1849 0.0915
Table 4: Replacing savings by real interest rate
GFCF4 Real interest rate (1)† Removing inflation rate (2)†
Foreign bank assets 0.0262
(0.0195)
0.0292 (0.0197)
Real interest rate 0.0211
(0.0232) 0.0257 (0.0177) Inflation rate -0.0013 (0.0031) Number of observations 164 164 R-square overall 0.1018 0.0991
Table 5: Countries used in the regressions5 1
Algeria®ˠ
30 Ethiopia 59 Nigeria
2 Argentina® 31 Georgia 60 Pakistan®
3 Armenia®ˠ 32 Ghana® 61 Panama
4 Azerbaijan® 33 Guatemala® 62 Paraguayˠ
5 Bangladesh 34 Honduras® 63 Peru®
6 Barbados 35 Hungary® 64 Philippines
7 Belarus® 36 India® 65 Poland®
8 Benin 37 Iran, Islamic Rep. 66 Romania®ˠ
9 Bolivia® 38 Jordan® 67 Russian Federation®
10 Bosnia and Herzegovina® 39 Kazakhstan 68 Rwanda®
11 Botswana® 40 Kenya® 69 Senegal
12 Brazil® 41 Latvia® 70 Slovak Republic®ˠ
13 Bulgaria® 42 Lebanon® 71 South Africa®
† The host country macroeconomic factors (inflation rate, stock size, exchange rate, GDP and savings) are part of these regression but are not presented in this table.
* Corresponds to significance of 1% ** Corresponds to significance of 5% *** Corresponds to significance of 10%4
®These countries are also used in the IV regression ˠ Excluded in regression because of no data on taxes
15
14 Burkina Faso 43 Lithuania® 72 Sudan®
15 Burundi® 44 Macedonia, FYR®ˠ 73 Swaziland
16 Cameroonˠ 45 Madagascar 74 Tanzania
17 Chile® 46 Malawiˠ 75 Thailand
18 China®ˠ 47 Malaysia 76 Togoˠ
19 Colombiaˠ 48 Mali 77 Trinidad and Tobago®
20 Congo, Rep. 49 Mauritania 78 Tunisia®
21 Costa Rica®ˠ 50 Mauritius 79 Turkey®
22 Cote d'Ivoire 51 Mexico 80 Uganda
23 Cuba 52 Moldova® 81 Ukraine
24 Czech Republic® 53 Morocco® 82 Uruguay®
25 Dominican Republicˠ 54 Mozambique 83 Uzbekistan
26 Ecuador®ˠ 55 Namibia® 84 Venezuela, RB®
27 Egypt, Arab Rep.® 56 Nepal® 85 Zambia
28 El Salvador® 57 Nicaragua 86 Zimbabwe®
29 Estonia® 58 Niger
References
Barajas, A., Steiner, R., & Salazar, N. (2000). The impact of liberalization and foreign investment in Colombia's financial sector. Journal of Development Economics, 63(1), 157-196.
Barth, J. R., Caprio Jr, G., & Levine, R. (2004). Bank regulation and supervision: what works best?. Journal of
Financial intermediation, 13(2), 205-248.
Berger, A. N., Clarke, G. R., Cull, R., Klapper, L., & Udell, G. F. (2005). Corporate governance and bank performance: A joint analysis of the static, selection, and dynamic effects of domestic, foreign, and state ownership. Journal of Banking & Finance, 29(8), 2179-2221.
Berger, A. N., & Udell, G. F. (2006). A more complete conceptual framework for SME finance. Journal of
Banking & Finance, 30(11), 2945-2966.
Borensztein, E., De Gregorio, J., & Lee, J. W. (1998). How does foreign direct investment affect economic growth?. Journal of international Economics, 45(1), 115-135.
Claessens, S., Demirgüç-Kunt, A., & Huizinga, H. (2001). How does foreign entry affect domestic banking markets?. Journal of Banking & Finance, 25(5), 891-911.
Claessens, S., & Van Horen, N. (2012). Being a foreigner among domestic banks: Asset or liability?. Journal of
16
Claessens, S., Van Horen, N., Gurcanlar, T., & Mercado, J. (2008). Foreign bank presence in developing countries 1995-2006: data and trends. Available at the Social Science Research network web site:
http://ssrn. com/abstract,1107295.
De Haas, R., & Van Lelyveld, I. (2006). Foreign banks and credit stability in Central and Eastern Europe. A panel data analysis. Journal of banking & Finance, 30(7), 1927-1952.
Easterly, W., & Rebelo, S. (1993). Marginal income tax rates and economic growth in developing countries. European Economic Review, 37(2), 409-417.
Gropp, R., & Kostial, K. (2000). The disappearing tax base: is foreign direct investment eroding corporate
income taxes? (No. 0031).
Hasan, I., & Marton, K. (2003). Development and efficiency of the banking sector in a transitional economy: Hungarian experience. Journal of Banking & Finance, 27(12), 2249-2271.
Henry, P. B. (2000). Do stock market liberalizations cause investment booms?. Journal of Financial
economics, 58(1), 301-334.
Jorgenson, D. W. (1963). Capital theory and investment behavior. The American Economic Review, 247-259. King, R. G., & Levine, R. (1993). Finance, entrepreneurship and growth. Journal of monetary, Economics, 32(3),
513-542.
King, R. G., & Levine, R. (2004). Finance and Growth Schumpter Might Be Right. World Bank Publications. Klein, M. W., & Rosengren, E. (1994). The real exchange rate and foreign direct investment in the United
States: relative wealth vs. relative wage effects. Journal of international Economics, 36(3), 373-389. Kraft, E., Hofler, R., & Payne, J. (2006). Privatization, foreign bank entry and bank efficiency in Croatia: a
Fourier-flexible function stochastic cost frontier analysis. Applied Economics, 38(17), 2075-2088. Micco, A., Panizza, U., & Yanez, M. (2007). Bank ownership and performance. Does politics matter?. Journal of
Banking & Finance, 31(1), 219-241.
Popov, A., & Udell, G. F. (2012). Cross-border banking, credit access, and the financial crisis. Journal of
International Economics, 87(1), 147-161.
