1
The influence of banking competition on mergers and acquisitions
Kevin Wilms S1985620
Abstract
This paper studies whether countries with higher banking competition have fewer targets in international M&A-transactions using a panel data set of 27 EU-countries in the period 2005-2013. Both the effect of banking competition over time and the difference in individual coun-tries is analyzed. The hypothesis that an increase in banking competition decreases the num-ber of target companies in acquisition-deals is not confirmed by empirical evidence. The dif-ference in the effect banking competition has on old and young firms, which is found in other banking competition studies, is not found either with respect to mergers and acquisitions. JEL: G15 G21 G34
Keywords: Banking competition; Mergers and acquisitions; Access to finance
Master Thesis for MSc Finance and MSc International Economics and Business
June 26, 2015
Supervisors:
Dr Padma Rao Sahib
2 Introduction
In a world with perfect capital markets, a firm can always get capital for a business opportunity which has a positive net present value. However, since this world does not have perfect capital markets, firms sometimes face financial constraints for business opportunities that otherwise would be beneficial for the firm. An acquisition can mitigate these financial constraints (Liao, 2014). If the new parent firm has better access to capital, it can reallocate funds towards the daughter firm, so all positive net present value opportunities can be taken. Erel et al. (2015) confirm empirically that financial constraints are lowered for target firms when they are ac-quired. If, for some reason, the capital markets in an economy would become better and more efficient, we would see a decrease in the number of acquisitions due to dampened financial constraints.
Banks are an important source of capital for firms. Several studies show evidence that in-creased competition among banks increases credit availability for firms (Black and Strahan, 2002; Cetorelli, 2004). The argument is that increased competition among banks spurs innova-tion and lowers prices towards marginal costs. This process should be beneficial for borrowing firms. However, other studies find that increased competition among banks has the opposite effect for small firms (Petersen and Rajan, 1995; Beck et al, 2004; Zarutskie, 2004). The ar-gument here is that relationship lending is better possible when banking competition is low. Banks are better able to provide loans to young or distressed firms, because they are able to get extra profits in the future. The question thus remains what the effect of increased banking competition is on loan availability and therefore on the acquisition-activity in an economy. Banking competition has become a more prominent issue in the EU area than it is in most oth-er areas. Thoth-ere has been a process of integration among countries in the EU banking industry. This has taken place through: deregulation of capital flows, creating a single banking license through the Second Banking Directive and the creation of a single currency (Weil, 2013). One would expect an increase in banking competition after the implementation of these regulations. However, empirical results show otherwise. The euro area experienced a decline in the bank competition after the introduction of the EMU and following the global financial crisis. Com-petition levels in euro countries seem to have converged in the wake of the EMU (Sun, 2011).
3 This paper is organized as follows; Section 2 presents a literature review of previous articles regarding motives for mergers and acquisitions and banking competition. Section 3 describes the methodology. Section 4 highlights how the data is collected and presents some descriptive statistics of this data. The empirical results will be presented in section 5 and provides some extra analysis. Section 6 contains the conclusion of this paper.
Literature
This study mainly contributes to the field of research that investigates the drivers of mergers and acquisitions and the field which researches the real effects of banking competition. I shall first address previous articles on banking competition and the second part describes relevant literature about the motives for mergers and acquisitions.
Banking competition
Traditional models of banking competition such as Klein (1971) predict that lower levels of competition would lead to higher interest rates on loans, leading to a decrease in the equilibri-um supply of loans. Many empirical studies in the field of banking competition use banking regulations as a source of exogenous variation to investigate banking competition. Jayaratne and Strahan (1996) find that the increase of competition through deregulation has positive real effects on per capita real income and output growth. Other studies find that this deregulation of the banking industry promotes entrepreneurship (Black and Strahan, 2002), increases personal bankruptcy rates (Dick and Lehnert, 2010) and allows firm entry and access to bank credit (Cetorelli and Strahan, 2006). Rice and Strahan (2010) find that deregulation of the banking industry expands credit supply by reducing the cost of credit. Francis et al. (2014) show a neg-ative relation between intrastate banking deregulation and corporate cash holdings, which points towards the view that banking competition decreases financial constraints for firms. Other studies about banking competition investigate the effects of consolidations in the bank-ing industry. Karceski et al. (2005) find that reduced competition in the bankbank-ing industry through mergers of banks negatively influences the stock price of the borrowers of the merging banks. Dolar and Yang (2013) document a negative relationship between small business credit and decreased competition through a consolidation in the banking industry of California after the 2007-2008 crisis.
4 is forced to charge high interest rates. A monopolistic bank could, however, subsidize the firm when it is young or distressed and extract the rents later. The monopolistic bank can charge higher than competitive interest rates in a later period because of the relationship between bank and firm. Petersen and Rajan also test this idea empirically and find that young firms do get more institutional finance in concentrated banking markets than in competitive banking mar-kets. They also find that older firms are more indebted in competitive markets than in concen-trated markets in their research. Zarutskie (2004) confirms this finding by looking at young and older firms following a banking deregulation. She finds that new firms have lower levels of debt following a banking deregulation and old firms have higher levels of debt. Beck et al. (2004) find that the relationship between banking concentration and access to finance is influ-enced by other country specific variables. In their study of 74 developed and developing coun-tries they find that banking concentration increases obstacles to finance only in councoun-tries with low levels of economic and institutional development. Foreign ownership of banks and an efficient credit registry dampen the effect of banking concentration on financing obstacles. Cornaggia et al. (2015) find that banking competition decreases innovation by public corpora-tions and increases innovation by private corporacorpora-tions that are dependent on external finance and have limited excess to bank loans. They argue that banking competition enables growing firms to secure bank loans instead of being acquired by public corporations. Therefore banking competition reduces the number of innovative takeover targets, which reduces the portion of state-level innovation attributable to public corporations. The first mechanism through which this is achieved is that banking competition increases loan availability and the supply of credit. Small firms that need capital for growth would get a bank loan instead of selling equity to achieve growth when the supply of loans is higher due to increased banking competition. Alt-hough this idea is contradicting the model of Petersen and Rajan (1995) the idea that an in-crease in banking competition would inin-crease the supply of credit and therefore dein-crease the incentive to be acquired by another firm is also incorporated in research on the motives for mergers and acquisitions.
5 Motives for Mergers and Acquisitions.
Studies have found that efficiency (Schoar, 2002), market synergies (Hoberg and Phillips, 2010), transfers of wealth from employees to merging firms (Shleifer and Summers, 1988) and transfers from the government to merging firms (Auerbach and Reishus, 1988) are motives for mergers and acquisitions. Another motive for firms to acquire other firms could be their over-valued equity (Shleifer and Vishny, 2003). There is new evidence that points towards the view that capital constraints of the target company could be a motive for acquisitions as well. Erel et al. (2015) document that subsequent to an acquisition target companies have lower cash hold-ings, lower sensitivity of cash holdings to cash flow, lower sensitivity of investment to cash flow, and an increase in investments. These results point towards the view that financial straints are lowered for target firms when they are acquired. This means that the financial con-straints could be a motive for a takeover.
Liao (2012) finds that firms benefit from relieving their financial constraints through selling minority equity stakes. Almeida et al. (2010) formulate a model in which banking competition could lower the number of mergers and acquisitions. They work out a model in which finan-cially distressed firms are acquired by liquid firms in their industries even in the absence of operational synergies. These studies point towards the mechanism that financially constrained firms are more likely to be taken over. If these financial constraints could be mitigated by, for example, an increased loan supply in a certain country, a decrease of targets in that country would be expected. A source for an increase in the loan supply could be an increase in the competition among banks. The hypothesis of this paper is as follows: an increase (decrease) in banking competition leads to an increase (decrease) in loan supply which leads to an decrease (increase) in the financial constraints for firms and this would lead to a decrease (increase) in target firms. A flow-chart of these relationships is provided in Figure 1. As literature suggests different effects for young and old firms, it is expected that the mechanism between banking competition and M&A-activity will be more pronounced for older firms.
6 Methodology
In order to measure the effect of banking competition on M&A-activity, cross-border acquisi-tions will be analyzed. I use cross-border acquisiacquisi-tions to see what the effect is of the difference in banking competition of the countries over time. The advantage of cross-border acquisitions is that the composition of these acquisitions can be different for each country. Many other home-country variables are influencing the M&A-activity in a country, so fixed effects are used to capture these effects. The main interest of this study is the influence of differences in banking competition a country has over time on the composition of international M&A-activity. The empirical model that I estimate has the form:
𝐴𝑐𝑞𝑟𝑎𝑡𝑖𝑜𝑖,𝑡 = 𝛼 + 𝛽1 𝐻𝐻𝐼𝑖,𝑡+ 𝛾𝑍𝑖,𝑡+ Country𝑖+ ε𝑖,𝑡 , (1) where 𝑖 indexes country and 𝑡 indexes time. The dependent variable (Acqratio) is the number of firms that are taken over in country 𝑖 and year 𝑡, divided by the total number of international acquisitions done in that country that year. The acquisition-ratio is calculated as follows:
𝐴𝑐𝑞𝑟𝑎𝑡𝑖𝑜𝑖,𝑡 = ∑ 𝑇𝑎𝑟𝑔𝑒𝑡𝑠𝑖,𝑡
∑ 𝑇𝑎𝑟𝑔𝑒𝑡𝑠𝑖,𝑡+∑ 𝐴𝑐𝑞𝑢𝑖𝑠𝑖𝑡𝑜𝑟𝑠𝑖,𝑡 , (2)
where 𝑖 indexes country and 𝑡 indexes time. The analysis is focused on target companies be-cause the hypothesis is that there will be fewer target companies when banking competition is high. I do, however, scale this variable by dividing the number of targets a country has in a year by total number of M&A deals that companies of that country are involved with in a that year. For example, 168 UK firms were taken over by another EU member and 215 EU firms (non-UK) were taken over by a UK firm in 2013. The acquisition ratio of the UK in 2013 is 168/(215+168) = 0.439. An acquisition-ratio higher than 0.5 thus implies that there were more target firms in a country-year than there were acquiring firms in that country-year.The variable HHI is the banking competition variable and will be the Herfindahl-Hirschman Index. The Herfindahl index is calculated as follows:
𝐻𝐻𝐼𝑖,𝑡= ∑𝑛 𝑀𝑖,𝑗,𝑡2
𝑗=1 (3)
7 factor costs of a bank. The advantage of these variables is that they are better in measuring banking competition than variables like the Herfindahl index (Claessens and Laeven, 2004). The disadvantage is that many observations are needed to accurately measure banking compe-tition with the Lerner index or the H-statistic. Measures like these are mostly used when meas-uring competition using multiple years of data. However, since I need to know the differences in the banking competition each year separately, it is impossible to measure the Lerner index or H-statistic for each country-year. As an alternative measure of banking concentration, I calculate the HHI using the total assets of the bank as well. The variable 𝑀 in formula (3) would in this case be the market share of assets of the bank. For robustness tests, the Her-findahl index will also be calculated using the market-shares of the deposits of the largest three and five banks in a country.
The Z in formula (1) represents the three control variables used: GDP per capita, a regulatory index and a measure of stock market performance. As Beck et al. (2004) show, economic and institutional development influences the effect banking concentration has on the availability of finance and therefore it might influence the effect banking concentration has on M&A-activity as well. I add GDP per capita (divided by $1000) as a control variable to the equation to cap-ture the effect of economic development of a country. Since companies search for growth op-portunities in developing countries, I expect, like the results of Beck et al. (2004) indicate, that countries with a high GDP per capita have a lower acquisition-ratio. This effect, however, could also be incorporated in the fixed effects of the model. For institutional development I follow Beck et al. (2004) and use a summary variable from Kaufmann, Kraay, and Zoido-Lobaton (1999) that averages six indicators proxying for voice and accountability, regulatory quality, political stability, rule of law, control of corruption, and effectiveness of government. In this variable, the maximum a country can get is 2.5 points, the minimum is zero points and a higher value indicates better regulations. Since Beck et al. (2004) show that firms in countries with a high score for institutional development have a lower chance of financing obstacles, we expect that the regulatory index is negatively related to the acquisition-ratio.
8 As suggested in the previous literature review, the age of the target firm could affect the im-pact banking competition has on loan supply for these firms and therefore on the expected effect on M&A-activity. Zarutskie (2004) finds that an increase in banking competition de-creases debt levels in firms of 5 years or younger and inde-creases debt levels in firms of 16 years or older. I test whether the relationship between banking competition and M&A-activity is opposite for firms of different age as well.
Data
Balance-sheet and income-statement information of European banks can be found on
Bankscope of Bureau van Dijk. All of the banks in the European Union have balance-sheet and income-statement information from 2005 onwards except for Greece; this is the reason why Greece is excluded in this research. Because information on 2014 and 2015 is not fully availa-ble yet for most of the European banks, data up to 2013 is used. The total number of banks in the sample is 4722, which includes all commercial and cooperative banks that were active in this period. Only these banks are chosen since these are the main suppliers of capital for com-panies (apart from selling equity).The Zephyr database contains information on mergers and acquisitions on the individual deal level. I only use M&A transactions between firms which are located in EU countries. After filtering out all transactions by firms in non-EU countries (and Greece) I have 14758 M&A transactions left. This data on individual firms is used to calculate the ratio of international targets of acquisitions to total international acquisitions in a country for the period 2005-2013.
The data on GDP per capita and the regulatory variable are collected from the World Bank-statistics. The average values of these two variables can be found in Table 1 for each country included in this sample. The data on the yearly returns of the stock market indexes are taken from Yahoo Finance and the stock market indexes of Malta, Luxembourg and Cyprus were found on the website of their stock exchanges1. The average of the yearly returns and the name of each index can be found in Table 1. Table 1 also depicts some other summary statistics of the main variables in this study. All of the statistics which are not a standard deviation are averages over the 9 years we use. As can be seen, larger countries are associated with a lower Herfindahl index, and thus a higher banking competition. This makes sense, since larger mar-kets can sustain more banks. A higher GDP is associated with a higher score for the regulato-ry-Index.
1
9 Country HHD St. dev HHD Change HHD #International M&A Acquisition ratio St. dev acqui-sition ratio Stock-Index Stock Return GDP per capita ($) Regulatory index Belgium 0.241 0.048 -0.048 173.0 0.539 0.031 Bel 20 4.0% 44200 1.310 Bulgaria 0.118 0.025 -0.001 45.7 0.924 0.053 SOFIX -3.0% 6296 0.196 Czech Republic 0.160 0.014 -0.017 68.7 0.763 0.114 PX -0.5% 18901 0.878 Denmark 0.291 0.035 -0.028 123.3 0.503 0.059 OMXC 20 12.9% 57538 1.837 Germany 0.094 0.009 0.001 401.3 0.527 0.039 DAX 12.3% 41988 1.465 Estonia 0.578 0.098 -0.072 40.9 0.532 0.112 OMXTGI 6.7% 15570 1.030 Ireland 0.199 0.026 0.028 83.1 0.468 0.051 ISEQ -3.8% 52990 1.499 Spain 0.219 0.031 -0.035 165.4 0.621 0.068 IBEX 35 3.8% 30799 0.880 France 0.073 0.010 0.008 335.8 0.445 0.072 CAC-40 3.9% 40897 1.226 Croatia 0.159 0.013 -0.015 11.8 0.853 0.086 CROBEX -0.7% 13337 0.372 Italy 0.105 0.012 0.010 157.4 0.612 0.029 FTSE-MIB -9.1% 36203 0.550 Cyprus 0.288 0.057 0.059 34.3 0.290 0.087 CFTSE-20 31.1% 27134 1.064 Latvia 0.133 0.013 -0.016 27.9 0.797 0.084 OMXL 6.5% 12375 0.647 Lithuania 0.220 0.023 0.013 34.0 0.746 0.096 OMXV 12.8% 12400 0.714 Luxembourg 0.065 0.011 0.004 81.8 0.251 0.054 LuxX 13.3% 101889 1.690 Hungary 0.161 0.014 0.021 35.0 0.790 0.054 BUX 2.6% 13104 0.773 Malta 0.444 0.050 0.008 7.0 0.494 0.157 DWML 2.0% 19503 1.193 Netherlands 0.257 0.057 -0.037 302.6 0.402 0.046 AEX 5.3% 49823 1.661 Austria 0.116 0.015 -0.010 108.2 0.362 0.063 ATX 7.7% 46764 1.583 Poland 0.081 0.012 -0.006 87.7 0.697 0.093 WIG 20 6.2% 11800 0.686 Portugal 0.241 0.016 -0.023 40.9 0.685 0.131 PSI-20 2.5% 21941 0.997 Romania 0.171 0.030 -0.032 35.9 0.917 0.085 BETI 4.4% 7974 0.105 Slovenia 0.148 0.006 -0.002 9.9 0.637 0.279 SBITOP -21.1% 23115 0.941 Slovakia 0.188 0.017 0.008 23.0 0.751 0.124 SAX -4.3% 16175 0.759 Finland 0.375 0.035 0.034 129.9 0.436 0.071 OMXH 25 9.5% 46919 1.861 Sweden 0.444 0.029 -0.034 240.8 0.399 0.043 OMX S 10.7% 52645 1.759
Great Britain 0.111 0.006 0.002 474.3 0.397 0.051 FTSE-100 5.1% 41680 1.401
Average 0.210 0.026 -0.007 121.5 0.587 0.083 - 4.5% 31998 1.077
10 Figure 2: Herfindahl-index of banks using total bank deposits for the 6 largest economies in the EU.
Figure 2 depicts the Herfindahl index over time of the six largest economies in the EU in terms of GDP. The Herfindahl index in Figure 2 is calculated using the total deposits of the banks over time. Although the average of the Herfindahl-index in the EU does not change that much in this time period, the bank-concentration does vary across countries. It can also be seen that there is quite some variation over time in some countries. This variation over time is used to identify the influence of banking concentration on the composition of international M&A-activity. Although the average banking concentration has declined slightly, Table 1 shows that the Herfindahl index of banking deposits did rise in 12 of the 27 countries.
11 Figure 3: Nationality of firms that acquire a firm in one of the six largest economies of the EU.
12 Figure 4: International M&A composition over time of the six largest economies of Europe
Results
The dependent variable in the regressions in Table 2 is the ratio of the total number of interna-tional targets in a country-year to the total number of internainterna-tional M&A deals in that same country-year (both targets and acquisitors). All these regressions have the Herfindahl index calculated with total deposits of banks as an independent variable. Since normal heteroskedas-ticity-tests are not applicable to panel-data, a likelihood test to test for heteroskedasticity in panel data was performed. The Chi-square(26) statistic is 138.52, so robust standard errors are used in the regressions. Although a Hausman-test for panel data that was performed has a p-value of 0.4610, a fixed effects model is chosen since the yearly observations of each country are assumed to be correlated with that country. The observations are not drawn from a larger population. Column (2) and (3) show the regressions using the acquisition ratio calculated with data which contains only deals with target firms of more than 5 years old and target firms of less than 5 years old, respectively. It can be seen that some observations are lost when the firm-age is accounted for. This is due to the fact that some countries do not have enough inter-national M&A-transactions in some years to calculate an acquisition-ratio. The countries that lose some observations in these specifications are Bulgaria, Hungary, Malta and Slovenia.
13 (1) (2) (3) VARIABLES Acquisitions-ratio Acq-ratio >5y old Acq-ratio <5y old
Herfindahl index deposits 0.243 0.535** -0.0810 (0.162) (0.249) (0.387) Stock Market index -0.00790 0.0271 -0.0598**
(0.0214) (0.0192) (0.0264) GDP per capita -0.0199 0.0164 -0.0445* (0.0179) (0.0198) (0.0242) Regulatory index -0.139 -0.227* 0.0666 (0.106) (0.124) (0.258) Constant 0.749*** 0.652*** 0.663** (0.125) (0.123) (0.306) Observations 243 242 229 R-squared 0.027 0.038 0.018 Number of countries 27 27 27
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
Table 2: Fixed effects regression with different dependent variables. (1) has the ratio between the num-ber of M&A targets a country has in a year to the total numnum-ber of international M&A deals a country has in a year as an dependent variable. (2) has the ratio between the number of M&A targets with an firm-age of at least 5 years old a country has in a year to the total number of international M&A deals of firms of at least 5 years old a country has in that year as an dependent variable.(3) has the ratio
be-tween the number of M&A targets with an firm-age of at most 5 years a country has in a year to the total number of international M&A deals of firms of at most 5 years old a country has in that year as an
dependent variable.
As can be seen in Table 2, column (1) shows that there is a positive relationship between the Herfindahl indexes of countries and the ratio of targets to the total of international M&A deals. This positive relationship is expected since a decrease in banking competition means an in-crease of the Herfindahl index. This inin-crease of the Herfindahl index is associated with an increase in the acquisition-ratio, which means more targets relative to the total of M&A deals in that country-year (in layman’s terms: more banking competition means less targets). This result is, however, not statistically significant.
14 all firms that are older than 5 years are excluded, the direction of the relationship reverses, although this result is not statistically significant. It can also be seen that the stock-market and GDP control variables are statistically significant in specification (3) at the 5%-level and the 10%-level respectively. The effect is, as we expected, negative for both variables.
Auto-correlation and omitted variables
A problem with panel-data is the possibility of auto-correlation. A test derived by Woolridge is used to test whether auto-correlation is apparent in the relationship. This test was used since other tests are not appropriate for panel data. An F-value (1,26) of 16.182 shows that auto-correlation is a problem in this dataset. A simple AR(1) variable is added as an independent variable in the regressions of Table 3. The direction of the relationship does not change and is still as expected. None of the relationships are, however, statistical significant. Column (2) shows the same regression with the difference that only firms older than 5 years are included the dependent variable. The effect the Herfindahl Index of deposits has on the acquisition ratio is positive, but not statistically significant. In column (3) is shown that when only firms younger than 5 years are included, the relationship has the opposite direction, which is ex-pected when accounting for the results of Petersen and Rayan (1995). These results are howev-er, not statistically significant.
15 (1) (2) (3) (4) (5) (6) VARIABLES Acquisitions-ratio Acq-ratio >5y old Acq-ratio <5y old Acquisitions-ratio Acq-ratio >5y old Acq-ratio <5y old HHI deposits 0.290 0.355 -0.0919 0.282 0.369 -0.246 (0.232) (0.312) (0.491) (0.237) (0.324) (0.495) Stock Market -0.00362 0.0284 -0.0612* 0.00195 0.0269 -0.0259 (0.0136) (0.0200) (0.0330) (0.0137) (0.0207) (0.0338) GDP per Capita -0.0201 0.00856 -0.0637 -0.0363 -0.0147 -0.0274 (0.0258) (0.0327) (0.0525) (0.0386) (0.0481) (0.0749) Regulatory index -0.0827 -0.296 0.0338 -0.0817 -0.269 -0.125 (0.147) (0.181) (0.288) (0.151) (0.190) (0.291) d2006 0.0388 0.00664 0.163** (0.0272) (0.0426) (0.0663) d2007 0.0581** 0.00407 0.125** (0.0263) (0.0366) (0.0585) d2008 0.0666** -0.0102 0.203*** (0.0282) (0.0379) (0.0605) d2009 0.0584** -0.0314 0.155*** (0.0272) (0.0364) (0.0571) d2010 0.0172 -0.0539 0.198*** (0.0280) (0.0377) (0.0605) d2011 0.0781*** 0.0541 0.159*** (0.0260) (0.0359) (0.0575) d2012 0.0569** -0.00291 0.123** (0.0235) (0.0346) (0.0547) Constant 0.675*** 0.789*** 0.768** 0.682*** 0.838*** 0.711** (0.136) (0.208) (0.322) (0.145) (0.224) (0.346) Observations 216 215 202 216 215 202 R-squared (within) 0.0166 0.0324 0.0238 0.0995 0.0893 0.1141 Number of countries 27 27 27 27 27 27
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
Table 3: Fixed effects regressions with an AR(1) variable added as an independent variable. (4), (5) and (6) have year-dummies added as independent variables. (1) and (4) have the ratio between the number of M&A targets a country has in a year to the total number of international M&A deals a country has in
a year as an dependent variable. (2)and (4) have the ratio between the number of M&A targets with an firm-age of at least 5 years a country has in a year to the total number of international M&A deals of
firms of at least 5 years old a country has in that year as an dependent variable.(3)and (6) have the ratio between the number of M&A targets with an firm-age of at most 5 years a country has in a year to the total number of international M&A deals of firms of at most 5 years old a country has in that year as
a dependent variable.
Robustness tests
16 Appendix B shows regressions with the same Herfindahl index as in columns (4), (5) and (6) of Table 3, but with different dependent variables. We test whether the relationship is robust with respect to the definition of an acquisition. In the former specifications, we included all acquisitions, such as institutional acquisitions, management buy-outs, minority acquisitions etc. In specifications (1), (2) and (3) in Appendix B, all minority acquisitions are excluded, and in columns (4), (5) and (6), all minority acquisitions and institutional acquisitions are exclud-ed. The direction of the relationship changes when the age of the target firm is not accounted for and all minority acquisitions and institutional acquisitions are excluded. This result is, again, not statistically significant. When the age of the firm is accounted for, the effect of banking concentration on M&A-activity does not change direction in comparison with former specifications but is not statistically significant.
The regressions in Appendix C test to what extent the relationship is robust towards the num-ber of years old a firm can be old to be considered a “young” firm. Next to the formerly used 5 years, 4 years (regression 1 and 2) and 6 years (regression 3 & 4) are tested. The direction of the relationship does not change for the 6-year specification but the direction of the 4 year specifications do change. These results are not statistically significant.
Country-difference analysis
In the former analysis the difference in banking competition of a country over time and M&A-activity of a country over time was used to analyze the relationship. Since none of the expected relationships are found, it could be the case that there is not enough variation in the variables. To further research the relation between banking competition and mergers and acquisitions activity between countries, I will look at the differences between countries. Because the differ-ences over time of the Herfindahl index might not be large enough to obtain significant results, the differences in banking concentration of specific countries might explain more about the relationship between banking concentration and M&A-activity. The targets and acquirers of the six biggest economies in the EU are used since only these countries have enough yearly M&A-activity with other countries in this dataset to properly analyze this. This does however mean that the following results might not appropriately show the relationship for
less-developed European countries. The difference in Herfindahl index between a target-firms’ country and the acquiring firms’ country will be analyzed. Six countries are in this dataset, so 30 target-country and acquirer-country combinations can be found. The difference in the Her-findahl index of two countries is calculated as follows:
17 Where 𝑘 stands for the target-country, 𝑙 for the acquirer-country and 𝑡 for time. 𝐻𝐻𝐼𝑘𝑙𝑡 is thus the difference between the Herfindahl index of deposits for the target-firms’ country and the acquirer-firms’ country. When the variable 𝐻𝐻𝐼𝑘𝑙𝑡 is positive, this means the Herfindahl index of the country is higher. This means that banking concentration is higher in the target-country than it is in the acquirer-target-country. For example, the difference in the Herfindahl in the combination of Spain-France has a value of 0.1153 in 2013. Since the variable is positive, this means that the target country, Spain, has a higher banking concentration than the acquirer country, France. This analysis does not focus on the composition of international M&A-activity but we rather look at the actual number of targets a certain country-combination has. We add the number of target firms of the country combinations as follows:
𝑇𝑎𝑟𝑔𝑒𝑡𝑠𝑘𝑙𝑡 = ∑𝑛 𝐷𝑒𝑎𝑙𝑘𝑙𝑡
𝑖=1 (5)
Where 𝑘 stands for the target-country, 𝑙 for the acquirer-country and 𝑡 for time. 𝑇𝑎𝑟𝑔𝑒𝑡𝑠𝑘𝑙𝑡 is the number of firms that are taken over in country 𝑘 by firms in country 𝑙 at time 𝑡. As a ro-bustness test, we do look at the target/total-ratio of the different country combinations. This target/total ratio is the number of target firms in a country combination divided by the total number of deals in the same combination. It is approximately the same acquisition ratio as was used in the previous analysis, but with country-combinations instead of countries themselves. The acquisition-ratio is analyzed to see what effect scaling of the number of targets to the countries M&A-activity has on the relationship.
The empirical model that is estimated has the following form:
𝑇𝑎𝑟𝑔𝑒𝑡𝑠𝑘𝑙𝑡 = 𝛼 + 𝛽1 𝐻𝐻𝐼𝑘𝑙𝑡+ 𝛾𝑍𝑘𝑙𝑡+ CountryCountry𝑚+ ε𝑘𝑙𝑡 , (6) Where the 𝑍 –variable represents the control variables. The control variables are the difference between the control variables in the target firms’ country and the acquiring firms’ country. These control variables are calculated the same way as the difference in Herfindahl index in formula (4) but instead of the HHI, the variables for GDP per capita, stock market index and the regulatory index are used.
The results are shown in table 4. Tests for Heteroscedasticity, fixed or random effects and au-to-correlation are performed in the same way as in the previous section. The likelihood-ratio test for Heteroskedasticity has a Chi-square(29)-statistic of 239.87. Robust-standard errors are thus used. A fixed effects model is used since the Hausman statistic for panel data (Sargan-Hansen statistic) has a p-value of 0.0146. A fixed effects are represented by the
combina-18 tions. Auto-correlation appears to be present according to the Woolridge-test for
auto-correlation which has an F-value of 7.537.
VARIABLES (1) Target-firms (2) Target-firms (3) Target-firms (4) Acq-ratio Difference in HHI -36 .93** -18.70 -20.92* -0.572** (13.75) (13.51) (11.72) (0.260) Diff. Regulatory Index 7.920 0.766 0.651 0.0810 (6.517) (8.353) (7.494) (0.185) Diff. GDP per Capita -9.175** -6.208*** -6.489*** -0.101***
(3.980) (2.164) (1.819) (0.0364) Diff. Stock market Index 1.086 -0.383 -0.379 -0.0102
(2.923) (2.078) (1.840) (0.0439) 2006 4.837*** 1.28e-09 (1.190) (0.0283) 2007 9.385*** 2.15e-09 (1.399) (0.0309) 2008 5.079*** 2.14e-10 (1.469) (0.0314) 2009 -0.573 -2.40e-09 (1.486) (0.0315) 2010 1.502 8.12e-10 (1.469) (0.0314) 2011 3.644*** -2.53e-09 (1.399) (0.0309) 2012 4.327*** -2.07e-10 (1.190) (0.0283) Constant 18.60*** 17.38*** 14.03*** 0.500*** (0) (0.406) (0.660) (0.0179) Observations 270 240 240 240 R-squared 0.151 0.0528 0.3213 0.0725 Number of DID 30 30 30 30
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
Table 4: (1) Is a fixed-effects regression with the number of target-firms of different combinations of a target-firm country and acquirer-firm country a year as dependent variable. (2) has the same variables
and an AR(1) variable. (3) has the same variables as specification (2) and extra year-dummy variables. (4) has the same independent variables as specification (3), but with the ratio between the total number of target firms are taken over by firms in one specific other country in a certain year divided by the total
number of international deals the same country-combination has in that year.
19 (0.113) would mean a decrease in the number of targets of about 4. When autocorrelation is accounted for (specification 2), the relationship is negative as well, but not statistically signifi-cant.
20 Conclusions
The objective of this paper is to investigate what effect banking competition has on the mer-gers and acquisitions market. I use a dataset containing statistics on 27 EU-countries during the years 2005-2013 on banking competition and international merger and acquisitions for the analysis. The hypothesis is that the number of targets of acquisitions in a country would reduce when loan supply is increased due to an increase in banking competition. When loan supply is increased, this would mean that the financial constraints are lowered for the potential target firms, so that there is no necessity for an acquisition anymore.
The empirical results cannot confirm the hypothesized relationship. When looking at a coun-tries’ individual banking concentration over time, I do find that the direction of the relationship is as expected, but these results are not statistically significant. Banking competition literature is inconclusive about the relationship between banking competition and loan supply for young firms. To account for this, the relationship between banking competition and M&A-activity is analyzed for old and young firms separately as well. The supposed opposite direction of the relationship as expected by looking at the results of Petersen and Rajan (1995), are not clearly found.
The relationship between the differences in the Herfindahl index between two countries and the number of international targets in these countries a negative relationship is found. This result contradicts with the hypothesis. The results are still quite visible when auto-correlation and omitted variables are accounted for. Since only six countries could be included in this analysis, these results could be driven by other factors such as the financial crisis or a banking system not functioning as expected. Another reason for this result could be that other variables are influencing the relationship between Banking competition and the mergers and acquisitions market besides financial constraints. It could be argued that banks in some countries are more eager to provide funding for mergers or acquisitions in good times than banks in other coun-tries.
21 Literature
Almeida, H., Campello, M., and Hackbarth, D., 2011, Liquidity mergers, Journal of Financial Economics 102, 526-558.
Auerbach, A., and Reishus, D., 1988, The Impact of taxation on mergers and acquisitions, in: Auerbach, A., Mergers and Acquisitions. University of Chicago Press, Chicago, pp. 69 – 86. Beck, T., Demirgüç-Kunt, A., and Maksimovic, V., 2004, Bank competition, financing obsta-cles and access to credit, Journal of Money, Credit and Banking 36(3), 627–648, 2004. Black, S., and Strahan, P., 2002, Entrepreneurship and bank credit availability,
Journal of Finance 57, 2807-2833.
Cetorelli, N., 2004, Real effects of bank competition, Journal of Money, Credit, and Banking 36, 543-558.
Cetorelli, N., Strahan, P., 2006, Finance as a barrier to entry: bank competition and industry structure in local U.S. markets, Journal of Finance 61 Issue 1, 437-461.
Claessens, S., and L. Laeven, 2004, What drives bank competition? Some international evi-dence. Journal of Money, Credit, and Banking 36 (3), pp. 563-583.
Cornaggia, J., Mao, Y., Tian, X., Wolfe, B., 2015, Does banking competition affect innova-tion?, Journal of Financial Economics 115, 189-209.
Dolar, B. and Yang, F., 2013, The 2007-2008 financial crisis and the availability of small business credit, Journal of finance and accountancy volume 12,
Erel, I., Jang, Y,. Weisbach, M., 2015. Do acquisitions relieve target firms’ financial con-straints?, Journal of Finance 70, 289-328.
Francis, B., Hasan, I., Wang, H., 2014, Banking deregulation, consolidation and corporate cash holdings: U.S. evidence, Journal of Banking & Finance 41, 45-56.
Hoberg, G., and Phillips, G., 2010, Product market synergies and competition in mergers and acquisitions: A text-based analysis, Review of Financial Studies 23(10), 3773-3811.
22 Karceski, J., Ongena, S., Smith, D.C., 2005. The impact of bank consolidation on commercial borrower welfare, The Journal of Finance 60, 2043–2082.
Kaufmann, Daniel, Aart Kraay, and Pablo Zoido-Lobaton, 1999, Governance Matters, World Bank Policy Research Working Paper 2196.
Liao, R. C., 2014, What drives corporate minority acquisitions around the world? The case for financial constraints, Journal of Corporate Finance 26, 78-95.
Panzar, J., Rosse, J., 1987, Testing for monopoly equilibrium, Journal of Industrial Economics 35, 443-456.
Petersen, M. and Rajan, R.G., 1995, The effect of credit market competition on lending rela-tionships, Quarterly Journal of Economics110, 407–443.
Rice, T., Strahan, P., 2010, Does credit competition affect small-firm finance? Journal of Fi-nance 65, 861-889
Schoar, A., 2002, Effects of corporate diversification on productivity, Journal of Finance 57, 2379-2403.
Shleifer, A. and Vishny, R., 2003, Stock market driven acquisitions, Journal of Financial Eco-nomics 70, 295-311.
Shleifer, A., and Summers, L. H., 1988, Breach of trust in hostile takeovers. In: Auerbach, A., Corporate takeovers: causes and consequences, University of Chicago Press, 33-68
Sun, Y., 2011, Recent developments of European banking competition, IMF working paper, WP 11/146.
Weil, L., 2013, Banking competition in the EU: how has it evolved?, Journal of international financial markets, institutions and money 26, 100-112.
23 Appendix A (1) (2) (3) (4) (5) (6) (7) (8) (9) VARIABLES Acquisitions-ratio Acq-ratio >5y old Acq-ratio <5y old Acquisitions-ratio Acq-ratio >5y old Acq-ratio <5y old Acquisitions-ratio Acq-ratio >5y old Acq-ratio <5y old
Herfindahl index Assets 0.0656 0.223 -0.363
(0.239) (0.319) (0.485)
Herfindahl index deposits of biggest 3
0.124 0.161 -0.111
(0.149) (0.206) (0.312)
Herfindahl index deposits of biggest 5
0.146 0.189 -0.122
(0.157) (0.217) (0.329)
Stock market index 0.00177 0.0263 -0.0268 0.00194 0.0266 -0.0266 0.00202 0.0267 -0.0267
(0.0140) (0.0208) (0.0340) (0.0139) (0.0208) (0.0341) (0.0139) (0.0208) (0.0341) GDP per Capita -0.0383 -0.0174 -0.0250 -0.0338 -0.0134 -0.0254 -0.0331 -0.0126 -0.0256 (0.0378) (0.0479) (0.0730) (0.0382) (0.0486) (0.0741) (0.0382) (0.0486) (0.0741) Regulatory index -0.110 -0.281 -0.110 -0.105 -0.286 -0.0888 -0.105 -0.286 -0.0882 (0.148) (0.189) (0.288) (0.147) (0.188) (0.287) (0.147) (0.188) (0.286) Constant 0.762*** 0.888*** 0.712** 0.732*** 0.897*** 0.634* 0.724*** 0.886*** 0.638* (0.148) (0.224) (0.354) (0.143) (0.215) (0.339) (0.143) (0.216) (0.340)
Year-dummies Yes Yes Yes Yes Yes Yes Yes Yes Yes
Observations 216 215 202 216 215 202 216 215 202
Number of countries 27 27 27 27 27 27 27 27 27
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
Fixed effects regressions with year dummies and AR(1) variables. (1),(2),(3), have the Herfindahl index calculated using Total Assets of banks as independent variable. (4),(5),(6) have the Herfindahl index calculated using the total deposits of the three biggest banks in the country as independent variable and (7),(8),(9) have the Her-findahl index calculated using the total deposits of the five biggest banks in the country as independent variable. (1),(4),(7), have the ratio between the number of M&A
targets a country has in a year to the total number of international M&A deals a country has in a year as an dependent variable. (2),(5),(8), have the ratio between the number of M&A targets with an firm-age of at least 5 years a country has in a year to the total number of international M&A deals of firms of at least 5 years old a country has in that year as an dependent variable. (3),(6),(9), have the ratio between the number of M&A targets with an firm-age of at most 5 years a country has in a
24 Appendix B (1) (2) (3) (4) (5) (6) VARIABLES Acq-ratio excl. minority Acq-ratio excl. minority >5y old Acq-ratio excl. minority <5y old Acq-ratio excl. minority and institutional Acq-ratio excl. minority and
institu-tional >5y old
Acq-ratio excl. minority and institutional <5y old
Herfindahl index deposits 0.0443 0.479 -0.382 -0.00482 0.377 -0.563 (0.264) (0.311) (0.668) (0.267) (0.322) (0.664) Stock Market index 0.00529 0.0249 0.0293 0.00318 0.0236 0.0290 (0.0165) (0.0195) (0.0432) (0.0164) (0.0200) (0.0449) GDP per Capita -0.0325 0.00231 -0.0905 -0.0290 0.00614 -0.0916 (0.0424) (0.0499) (0.104) (0.0430) (0.0517) (0.102) Regulatory index -0.244 -0.307 -0.255 -0.241 -0.308 -0.200 (0.166) (0.196) (0.398) (0.168) (0.202) (0.398) Constant 0.912*** 0.796*** 1.127*** 0.906*** 0.802*** 1.120** (0.176) (0.209) (0.430) (0.175) (0.213) (0.464)
Year fixed effects Yes Yes Yes Yes Yes Yes
Observations 216 216 194 216 216 194
Number of countryid 27 27 27 27 27 27
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
Fixed effects regression with year dummies and an AR(1) variable and different dependent variables. (1) has the ratio between the number of M&A targets excluding minority acquisitions a country has in a year to the total number of international M&A deals excluding minority acquisitions a country has in a year as an dependent variable . (2) has the ratio between the number of M&A targets excluding minority acquisitions with a firm-age of at least 5 years a country has in a year to the total number of international M&A deals excluding minority acquisitions of firms of at least 5 years old a country has in that year as an dependent variable.(3) has the ratio
between the number of M&A targets excluding minority acquisitions with a firm-age of at most 5 years a country has in a year to the total number of international M&A deals excluding minority acquisitions of firms of at most 5 years old a country has in that year as an dependent variable. (4) has the ratio between the number of
M&A targets excluding minority and institutional acquisitions a country has in a year to the total number of international M&A deals excluding minority and institu-tional acquisitions a country has in a year as an dependent variable . (5) has the ratio between the number of M&A targets excluding minority and instituinstitu-tional
acqui-sitions with a firm-age of at least 5 years a country has in a year to the total number of international M&A deals excluding minority and institutional acquiacqui-sitions of firms of at least 5 years old a country has in that year as an dependent variable.(6) has the ratio between the number of M&A targets excluding minority and institu-tional acquisitions with a firm-age of at most 5 years a country has in a year to the total number of internainstitu-tional M&A deals excluding minority and instituinstitu-tional
25 Appendix C (1) (2) (3) (4) VARIABLES Acq-ratio >4y old Acq-ratio <4y old Acq-ratio >6y old Acq-ratio <6y old
Herfindahl index deposits 0.304 0.0586 0.369 -0.572 (0.304) (0.531) (0.325) (0.475) Stock Market index 0.0182 -0.0407 0.0271 -0.0233 (0.0195) (0.0356) (0.0208) (0.0328) GDP per Capita -0.0114 -0.0836 -0.0122 -0.0233 (0.0451) (0.0813) (0.0482) (0.0716) Regulatory index -0.226 0.168 -0.229 -0.270 0.304 0.0586 0.369 -0.572 Constant 0.794*** 0.505 0.783*** 0.955*** (0.210) (0.367) (0.225) (0.332)
Year-dummies Yes Yes Yes Yes
Observations 215 195 215 206
Number of countries 27 27 27 27
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
Fixed effects regression with year dummies and an AR(1) variable. (1) has the ratio between the number of M&A targets with an firm-age of at least 4 years a country has in a year to the total number of
inter-national M&A deals of firms of at least 4 years old a country has in that year as an dependent varia-ble.(2) has the ratio between the number of M&A targets with an firm-age of at most 4 years a country
has in a year to the total number of international M&A deals of firms of at most 4 years old a country has in that year as an dependent variable. (4) has the ratio between the number of M&A targets with an
firm-age of at least 6 years a country has in a year to the total number of international M&A deals of firms of at least 6 years old a country has in that year as an dependent variable.(5) has the ratio be-tween the number of M&A targets with an firm-age of at most 6 years a country has in a year to the total number of international M&A deals of firms of at most 6 years old a country has in that year as an
