1
The impact of mergers and acquisitions on the South and Central
banking industry
Studentnr: s1955942 Name: Nerissa Windster Supervisor: Dr. J.J. Bosma
Master thesis of Finance
Abstract
This paper examines the impact of mergers and acquisitions (M&A) announcements on South and Central American bidder banks for the period of 2002 to 2012, using both accounting studies and event study. The findings for the accounting studies, using a dataset of 50 bidder banks revealed improvements in post-M&A bank performance. As performance measure the industry, size and pre-performance adjusted ratio return on equity is used. Further, improved performance is due to bank's ability to attract more loans and as a result increase profitability. The finding for the event study, using a dataset of 38 bidder banks revealed that the
shareholders of the bidder banks did not lose shareholder value.
JEL classifications: G21; G34
2 1. Introduction
Since the 1990s mergers and acquisitions (M&As) started to increase in the South and Central America region. In 1990, announced M&As were 66 deals while in 2000 there were 617 announced M&As, which leads to an increase of 835% (Metwalli and Tang, 2004). M&As in the banking industry begin due to deregulation of the banking industry in the 1990s and government processes to restructure inefficient banking systems (Hawkins and Mihaljek, 2001). Despite the M&A activities in South and Central American banks, little research has been conducted. Most researches are conducted for the regions Europe and the United States. Two reasons for this limited research are attributed to the lack of available data and the fact that the M&A activities in the region of South and Central America are more recent than the M&A activities in the United States and Europe (Goddart et al, 2012).
This paper analyzes the impact of the M&A announcements on South and Central American bidder banks for the period of 2002 to 2012 using both accounting studies and event study. Accounting study measures the impact of the M&A announcements on bank performance by analyzing changes in accounting ratios one year before and two years after the M&As. The return on equity (ROE) is used as the overall performance measure. In addition, specific bank indicators are also analyzed to examine their impact on the change in the overall performance. Next, the event study is conducted to analyze the impact of the M&A announcements by measuring the changes in the stock market prices around the announcement date. Lastly, this paper tests if the market participants anticipate post-M&A bank performance improvements at the time of the M&A announcement.
3 relation with the liquidity indicator loans to total assets. This implies the bank's ability to attract more loans and as a result of that increase profitability.
Previous literature also shows mixed results for the event studies. A group of studies find significant negative abnormal returns (Baradwaj et al., 1992; Houston & Ryngaert, 1994), while others find significant positive abnormal returns (Desai et al., 1985; Cybo-Ottone and Murgia, 2000). This paper finds insignificant results, implying that the shareholders of the bidder banks did not lose shareholder value. This result is consistent with some EU studies (Rad and Van Beek, 1999; Beitel et al., 2004; Asimakopoulos and Athanasoglou, 2013). A possible explanation is that the size of the targets compared to the size of the bidders are relatively small, to have an impact on the stock prices of the bidders (Rad and Van Beek, 1999).
Lastly, I did not find a relation between the post-M&A bank performance and the market reaction to the M&A announcement, implying that the market participants are unable to
anticipate post-M&A bank performance improvements and impound these improvements in the market prices at the time of the M&A announcement.
The remainder of the paper is organized as follows; section 2 gives a brief literature review about the motives for merger and acquisitions in the banking sector and previous empirical literature on accounting studies and event study. Section 3 provides the data and the
methodology for the two types of studies. Section 4 presents the results of the analysis and section 5 summarizes and concludes.
2. Literature review 2.1 Theories of Bank M&As
4 Agency theory explains the differences in incentives of managers and shareholders. Managers are interested in higher salaries, employment stability and growth of the firm, whereas
shareholders want to maximize their wealth (Jensen, 1986). Based on the agency theory, management of the bidder bank will engage in M&As that are beneficial for them at the expense of the shareholders (Baradwaj et al., 1992). In contrast to the wealth-maximizing theory, agency theory leads to negative returns for the shareholders of the bidder banks (Baradwaj et al., 1992).
Finally, there is the hubris theory. According to Roll, (1986) hubris theory argues that managers of the bidder banks overpay for their target acquisitions due to overconfidence in their abilities to value the target banks, which can be undervalued. Similar to the agency theory, this theory will also lead to negative returns for the shareholders of the bidder banks (Baradwaj et al., 1992).
2.1.1 Motives for M&As in South and Central American region
According to Hawkins and Mihaljek (2001) M&As in the emerging markets begin in the 1990s due to "global market and technology developments, macroeconomic pressures and the banking crisis". Because of these trends, the banks in the emerging markets implement deregulation measures such as removing the ceiling of the deposit rates. This leads to foreign entry from large international banks, most of them from developing countries. M&As in the South and Central banking industry occur due to financial liberalization, privatization of banks owned by the government and low levels of capital available after the bank crises (Hawkins and Mihaljek, 2001). These motives differ from the motives for Asia region. According to Hawkins and Mihaljek (2001), M&As in the Asia banking sector occur due to processes conducted by the government to strengthen the financial stability of the small, family-owned banks after the bank crises of 1997 and 1998. An example of a government process is the financial support the government provided to the merging banks in the form of public funds. As a results of the South and Central M&As, the total number of banks in the region decrease, the number of government-owned banks also decrease and the foreign bank, especially Spanish and American banks, market share increase (Carvallo and Kasman, 2005). Carvallo and Kasman (2005) argue that these foreign banks acquired the domestic banks and started their operation at the regional level. Andonova et al. (2013) analyze the performance of privately held companies for the period of 1995 to 2008 for the South and Central American country Colombia. Using the multiple linear regression model, they conclude that experienced M&A companies, who
5 years of the M&As. They use the ratio return on assets (ROA) as the performance measure. They characterize experience with M&As, as the number of M&A transactions the company is included divided by the total of M&A transactions in the wave. Late wave is defined as the last 10% M&A transactions that occur in the wave.
2.2 Empirical methods
Two empirical methods to analyze the impact of M&As on bank performance are accounting studies and event studies. Accounting studies analyze the changes in the accounting indicators of banks, such as efficiency changes, before and after the M&As (Diaz et al., 2004). A
limitation of accounting study is that it ignores market values, when calculating accounting indicators (Diaz et al., 2004).
Event studies examine the stock price reaction by calculating abnormal returns (AR) around the announcement date to see if the shareholders of the bidder create or lose shareholder value (Altunbaş and Marqués, 2008). A key assumption of this model is the efficient market hypothesis (EMH). It states that the stock market prices fully reflect all available information (Fama, 1970). A limitation of event study is that the period of time chosen to calculate the abnormal returns can lead to different results (Diaz et al., 2004).
Fridolfsson and Stennek (2005) suggest conducting these studies together because the two methods are complementary and do not substitute each other. Event studies do not detect when unprofitable M&As occur because it focuses only on share prices. However solely focusing on accounting indicators, as is done in accounting studies, will not capture the reasons why unprofitable M&As occur.
2.2.1 Accounting studies
6 attract loans. Cornet et al. (2006) study 134 M&As for the period of 1990 to 2000 and use industry-adjusted operating pretax cash flow return on assets (OPCFROA) as their performance measure. They find that OPCFROA increase significantly due to revenue enhancements and cost reduction activities.
European accounting studies also show mixed results. Altunbaş and Marqués (2008) examine the impact of strategic similarities on post-M&A performance using ROE of 262 bank M&As between 1992 and 2001. They find that, on average, performance improves. For domestic deals, differences between bidders and targets in terms of loans, earnings, cost, deposit and size strategy reduce performance, whereas differences in capitalization, technology, and financial innovation strategies improve performance. Ismail et al. (2009) examine 35 publicly listed banks M&As between 1992 and 1997 using mean adjusted OPCFROA. They find an insignificant change after M&As. Also, the M&As decrease profitability and capitalization significantly.
Emerging accounting studies is very limited. Khan (2011) compared pre- and post- M&A performance of the Indian banking sector by using the financial ratios Gross-profit margin, Net-profit Margin, Operating Profit Margin, Return on Capital employed (ROCE), Return on Equity (ROE) and Debt-Equity. He concludes that after the M&As the performance and efficiency of the banks have improved due to the improvement of the ratios after the M&As. Mantravadi and Reddy (2008) examine the performance for different industries in India using financial ratios as Khan (2011) and find that M&As have small impact on the profitability of firms in the banking and finance industry.
Based on these mixed results I hypothesize that:
Hypothesis 1. The M&A's announcements in the South and Central American banking industry have an impact on the operating performance of the bidder banks, where operating performance is measured by return on equity (ROE).
2.2.2 Event studies
Studies of bank M&As in the US report mixed results for the shareholders of the bidder banks. Houston & Ryngaert (1994) find significant negative abnormal returns (ARs) for a period of four days before the announcement date. Their sample consists of 153 US bank M&As.
7 US banks holding acquisitions during the period of 1976 to 1982 and conclude positive
significant ARs for the shareholders of the bidders bank holdings on the announcement date and on both the day before and the day after. James and Wier (1987) conclude, with a sample consisting of 60 bank M&As, that bidder banks create small but significant shareholder value before the announcement date for a period of five days and less.
European studies report evidence of positive or insignificant ARs. Cybo-Ottone and Murgia (2000) study 54 large EU-bank M&As during the period of 1988 to 1997 and report significant positive ARs for a period of 20 days before and after the announcement date. Rad and Van Beek (1999) analyze 56 EU bidding banks between 1989 and 1996 for a period of 40 days before and after the announcement date. The ARs for the days before the announcement date are negative and the ARs for the days after are positive. However, these ARs are small and insignificant. The reason for this is that the size of the targets are smaller in comparison with the size of the bidders and thus do not have an impact on the stock price of the bidders. The results of Beitel et al. (2004) and Asimakopoulos and Athanasoglou (2013) are in line with the
results of Rad and Van Beek (1999) showing also insignificant ARs. The empirical literature on the impact of M&As on bank performance is limited for emerging
markets. Goddart et al. (2012) analyze the success of 132 bank M&As in Latin America and Asia for the period of 1998-2009. They find that the ARs are positive but insignificant. Thus, they did not lose shareholder value due the M&A announcement. Williams and Liao (2008) analyze 73 cross-border M&As involving bidder banks from emerging markets. The abnormal returns for bidder banks in Latin America are large and significantly negative for a period of 15 days before and after the announcement date.
Based on these mixed results I hypothesize that:
Hypothesis 2. The M&A's announcements in the South and Central American banking industry have an impact on the stock price of the bidder banks.
3. Data and Methodology 3.1 Data sample
The data set is retrieved from 3 sources.
8 1. the bidders are listed and the targets are both listed and unlisted,
2. deal type is acquisition or merger,
3. the time period is from 01/01/2002 to 31/12/2012, 4. they are completed and confirmed,
5. the bidder is a South or Central American bank and the target is either a South or Central American bank or a financial services provider (FSP). "FSPs are insurance companies, asset management firms, credit institutions and brokerages" (Goddart et al., 2012),
6. the bidder bank has to have a majority of stakes in the target after the M&A. So the initial stake must have a maximum of 50% and the final stake must have a minimum of 50%. After these adjustments, the total sample data is 62 transactions.
Second, accounting data are retrieved from the database Bankscope. From the 62 transactions, 12 of the transactions are removed due to no availability of accounting data for the different ratios for the one year prior and the two years after the M&As. This leads to a dataset of 50 transactions. This is the dataset used for conducting the accounting studies.
Third, daily stock return indices (RI) for the bidders and the market index DataStream Latin America financial index are taken from DataStream. From the 50 transactions, 12 transactions do not have return information for the bidders over the estimation window. This leads to a dataset of 38 transactions for the event study.
Although the dataset for both the accounting studies and the event study is small, it is still large enough to conduct both the accounting studies and the event study. Previous studies, such as Desai et al. (1985) and Goddart et al. (2012) have 18 and 42 transactions for conducting event study, respectively. Cornett and Tehranian (1992) and Ismail et al. (2009) have 30 and 35 transactions for conducting accounting studies.
9 Table 1. Data description
Year # Transactions Bank-Bank Bank-FSP
Panel A. Transactions 2002 4 3 1 2003 7 6 1 2004 3 3 0 2005 5 5 0 2006 8 5 3 2007 3 3 0 2008 0 0 0 2009 13 4 9 2010 3 2 1 2011 1 0 1 2012 3 1 2 Total 50 32 18
Panel B. Type of country
Country Bidders Targets
Argentina 10 11 Barbados 0 1 Brazil 15 12 Chile 3 5 Colombia 8 9 Dominican Republic 0 1 Ecuador 3 1 Honduras 2 2 Mexico 2 3 Paraguay 0 1 Peru 3 2
Trinidad & Tobago 2 0
Venezuela 2 2
Total 50 50
The table shows the data description for the 50 bidder and target banks for the period of 2002 to 2012.
3.2 Methodology
3.2.1 Accounting studies Performance measure
Similar to Knapp et al. (2005) I use the profitability measure return on equity (ROE) to analyze the impact of M&A on the bank operating performance. ROE measures the overall
10 Determinants of ROE
The changes in ROE can be explained by different performance indicators. In this paper I use 5 bank performance indicators, which are the assets quality indicator; capital indicator; liquidity indicator; operating indicator; and efficiency indicator.
Assets quality indicator: it measures the changes in the loan quality. The ratio loan loss
provisions divided by net interest revenues (PROV) is used, which is also used by Ismail et al. (2009). This ratio shows the relationship between the provisions that are set aside in the profit and loss account for defaulted loans and the net interest revenue (The World bank, 2006). This ratio has to be negative in order to improve performance.
Capital indicator: it measures the amount of assets that are financed by the investors. Similar to Altunbaş and Marqués (2008) I use the equity to total assets ratio (CAPITAL). This ratio has to be positive in order to improve performance.
Liquidity indicator: it measures the ability of the bank to expand its operation, but at the same time having a buffer in the event of hard times (Asimakopoulos and Athanasoglou, 2013). The ratio used is loans divided by total assets (LIQ), as in Kwan and Eisenbeis (1999). It shows the impact of the loans on the total assets. Molyneux et al. (1998) argues that the higher the ratio, the more dependent the bank is on loans to earn profits.
Operating indicator: it measures the amount of total assets that is generated by other operating income, such as non-operating interest income. Similar to Altunbaş and Marqués (2008) I use the other operating income to total assets ratio (OPERATING). This ratio has to be positive in order to improve performance.
Efficiency indicator: it measures the relation between the overhead cost and the income, which consists of the other operating income and the net interest revenue. I use the cost to income ratio (EFFICIENCY), which is also used by Ismail et al. (2009). This ratio has to be negative. A higher cost to income ratio indicates higher costs or lower revenues.
11 Table 2. Definitions of the financial ratios
Financial ratios Symbol Formula
Performance measure ROE Net income/ equity
Assets quality indicator PROV Loan loss provisions/ net interest revenues
Capital indicator CAPITAL Equity/ total assets
Liquidity indicator LIQ Loans/ total assets
Operating indicator OPERATING Other operating income/ total assets
Efficiency indicator EFFICIENCY Overhead cost/ (other operating income + net interest revenue)
The table shows the definitions, symbols and formulas of the financial ratios used to analyze the bank performance of the 50 bidder banks for the period of 2002 to 2012. Source: Bankscope.
Performance benchmark
12 Two accounting methods
To examine changes in the financial ratios two accounting methods are used, the change model and the intercept model. In the change model, the financial ratios for each bidder bank are analyzed one year before (pre-M&A) and two years after (post-M&A) the announcement date (Powell and Stark 2005). The impact of the M&As on the financial ratios is the mean
differences between the pre-M&A financial ratios and the mean of the post-M&A financial ratios. This covers a period of 2001 trough 2014. The difference between the pre- and post- M&A financial ratios is tested using the following t-statistic (Cornet et al., 2006):
t = [ )]/( ) , (1)
where is the post- M&A financial ratios of the bidder banks, is the pre- M&A financial ratios of the bidder banks, is the standard deviation of the changes in pre- and post- M&A financial ratios and N is the number of bidder banks.
The intercept model uses OLS regressions to detect any changes in the financial ratios due to the M&A. Also, it studies the relation between pre-M&A performance and post-M&A performance (Healy et al. 1992).
The regression for the financial ratio ROE is as follows:
ISPAROE post,i = α + β ISPAROE pre,i + , (2)
where ISPAROE pre,i is the pre-M&A industry, size and pre-performance adjusted ROE and
ISPAROE post,i is the post-M&A industry, size and pre-performance adjusted ROE. The
intercept α represents the average change in the ROE due to the M&A. The slope β shows the impact of the pre- M&A performance of the ROE on the post-M&A performance and is the error term. The t-test is used to test if the difference between the post-M&A financial ratio and the pre-M&A financial ratio is significant. The same approach is also used for the five
performance indicators mentioned above.
3.2.2 Event study
The event study methodology is the market and risk-adjusted model based approach suggested by Brown and Warner (1985) and MacKinlay (1997).
13 where is the return of stock i at time t, is the return of the market portfolio at time t and
are the alpha, beta and the error term, respectively. The parameters are estimated for each stock i using the Ordinary Least Squares during the estimation window, which consists of 252 trading days. They are then referred to as and (Beitel et al., 2004). Next abnormal returns of stock i are derived by the equation:
(4)
The calculated abnormal returns are then averaged across banks using the formula:
, (5)
where n = number of stocks.
After averaging the abnormal return, the cumulative abnormal returns (CAR) are calculated for each event window by aggregating the through time:
, (6)
where is the cumulative abnormal return for the period [t1,t2].
Previous literature uses different event windows and has different results. This thesis uses 21, 11, 6 ,3 and 2 days i.e. [-10; + 10], [-5; + 5], [-10; 0], [-5; 0], [-1; + 1], [-1; 0] event windows. The event windows [-10; + 10], [-10; 0], [-1; + 1] and [-1; 0] are in line with Beitel et al. (2004) and the event windows [-5; + 5] and [-5; 0] are in line with Goddart et al. (2012).
For the cumulative abnormal returns, the statistical significance is calculated by using the t-statistic.
=
, (7)
where is the cumulative abnormal return for the period [t1,t2] and is the standard deviation of the CAR in the estimation window.
3.2.3 Robust analysis
As a robust check, the change model is conducted using median instead of mean. This implies that the impact of the M&A on the financial ratios is the median of the differences between the median of the post-M&A financial ratios and pre-M&A financial ratios. To test the statistical significance of this difference, I use the non-parametric test, Wilcoxon signed rank test. An advantage of the Wilcoxon signed rank test instead of the t-test, is that it assumes non-normality.
14 the market adjusted model assumes that the expected return of the stock is equal to the market return (Brown and Warner 1985). Thus, the parameter beta is equal to one and the parameter alpha is equal to zero. According to Dyckman et al. (1984) both of these models, detect cumulative abnormal excess returns using daily returns correctly. The formula for calculating the abnormal return using the market adjusted model is as follows:
, (8)
where is the abnormal return of stock i at time t, is the return of stock i at time t and is the return of the market portfolio at time t. For the calculation of the average abnormal
return, the cumulative abnormal return (CAR) and the t-statistic, equations (5), (6), and (7) are used, respectively.
4. Empirical results
4.1 Accounting studies results 4.1.1 Change model results
Performance measure: return on Equity (ROE)
Table 3 reports the bank mean ROE (without adjustment) and the industry, size and pre-performance adjusted ROE for the 50 bidder banks one year before and the two years after the bank M&As. With regards to the bank mean ROE, table 3 reports a significant positive change of 9,96%. This change is significant at the 10% significance level. Although it shows a
significantly positive increase, the results are not controlled for industry factors that may influence the ROE apart from the M&A. For this reason, the ROE controlled for industry, size and pre-performance (ISPAROE) is also presented in table 3. It shows an increase in ISPAROE of 11,47%, which is significant at the 10% significance level. Thus, bank performance for the 50 bidder banks improve significantly due to the M&As.
Determinants of ROE
Table 4 reports the change model results for the five performance indicators for the 50 bidder banks in the years surrounding the bank M&As. Only the industry, size, pre-performance adjusted financial ratios are reported. As mentioned above, the unadjusted financial ratios do not control for industry factors that may influence the ROE apart from the M&A. So it is difficult to draw conclusions based on the unadjusted results. Table 4 reports insignificant results for the performance indicators ISPAPROV, ISPACAPITAL, ISPAOPERATING and
15 However, the performance indicator liquidity (ISPALIQ) increase significantly with a
percentage of 2,20% at the 10% significance level.
Table 3. Change model result: return on equity (ROE)
The table reports the change model results for the industry, size, pre-performance adjusted ROE for the 50 bidder banks over the period of 2002 to 2012. Return on equity is measured as net income divided by equity. The industry, size, pre-performance adjusted financial ratio is calculated by the difference between the unadjusted financial ratio of the bidder bank and the financial ratio of the peer bank, controlled for industry, size and pre-performance. Post mean less pre is the mean of the difference between the mean of the post-M&A ROE and pre-M&A ROE. (*), (**) and (***) significantly different from zero at the 10%, 5% and 1%, respectively.
4.1.2 Intercept model results
Table 5 reports the intercept model results for the ISPAROE and the five performance indicator ISPAPROV, ISPACAPITAL, ISPALIQ, ISPAOPERATING and ISPAEFFICIENCY using equation (2). OLS regressions are used to detect any changes in performance due to the M&A and to study the relationship between pre-M&A performance and post-M&A performance. Performance measure: ISPAROE
The intercept for the ISPAROE is 3,88% and is significant at the 10% level. This is consistent with the results of the change model. The M&As result in significant bank performance
improvements. Table 5 also reports an insignificant positive association between the pre- M&A performance and the post- M&A performance. Thus, the bank performance in the pre-M&A did not continue in the post-M&A period.
Assets quality indicator (ISPAPROV)
The intercept for ISPAPROV of 1,42% is not significantly different from zero. This is consistent with the results of the change model. However, the evidence of a negative relation
Industry, size, pre-performance adjusted Year relative to M&A Bank mean (%) Mean (%)
-1 12,35 -5,97
1 22,62 2,94
2 22,01 8,06
Mean annual post-performance
for years 1 and 2 22,31 5,50
16 between the post-M&A performance and the M&A performance is present due the pre-performance coefficient of -0,377, which is significant at the 5% significance level. Thus, the continuation of the pre-M&A performance lead to a decrease in the post-M&A performance for the asset quality indicator.
Capital indicator (ISPACAPITAL)
For ISPACAPITAL table 5 reports insignificant results for both the intercept and the pre-performance. So there is no evidence of significant change in loan quality due to the M&As and also no significant association between the post-M&A performance and the pre-M&A performance for ISPACAPITAL. This is consistent with the results of the change model, which also report insignificant results.
Liquidity indicator (ISPALIQ)
Similar to the change model results, table 5 reports that for both the intercept and the pre-performance, the results are significantly different from zero. The intercept of 2,42% at the significance level of 5% implies that the M&As have a positive impact on ISPALIQ. The slope for the pre-performance of 0,543 is also highly significantly different from zero at the 1% level. It shows evidence of a positive association between the pre- M&A performance and the post- M&A performance.
Operating indicator (ISPAOPERATING)
The intercept for ISPAOPERATING of 0,19% is not significantly different from zero. This is consistent with the results of the change model. However, the evidence of a negative relation between the post-M&A performance and the M&A performance is present due the pre-performance coefficient of -0,282, which is significant at the 5% significance level. Thus, the continuation of the pre-M&A performance lead to a decrease in the post-M&A performance for the operating indicator.
Efficiency indicator (EFFICIENCY)
For ISPAEFFICIENCY table 5 reports insignificant results for both the intercept and the pre-performance. So there is no evidence of significant change in efficiency due to the M&As and also no significant association between the post-M&A performance and the pre-M&A
17 Table 4. Change model results: determinants of ROE
Industry, size, pre-performance adjusted
Indicators (%) Pre-M&A Post-M&A Difference
Assets quality indicator
(1) Loan loss provision to net interest
revenue (PROV) 4,68 -0,35 -5,03
Capital indicator
(2) Equity to total assets (CAPITAL) 0,16 -0,92 -1,08 Liquidity indicator
(3) Loans to total assets (LIQ) 0,49 2,69 2,20*
Operating indicator
(4) Other operating income to total assets
(OPERATING) -0,13 0,22 0,35
Efficiency indicator
(5) Cost to income (EFFICIENCY) 4,62 1,31 -3,31
18 Table 5. Intercept model results
Independent variables a ISPAROE post a ISPAPROV post a ISPACAPITAL post a ISPALIQ post ISPAOPRATING post a ISPAEFFICIENCY post Intercept (%) *3,88 1,42 -1,03 **2,42 0,19 1,36 (1,92) (0,42) (-0,73) (2,22) (0,66) (0,69) Pre-performance -0,272 **-0,377 0,643 ***0,543 **-0,282 -0,012 (-1,06) (-2,29) (0,98) (2,95) (-2,01) (-0,15) F-statistic ***10,36 ***36,09 2,21 ***8,79 **4,06 0,03 Adjusted R-squared 0,16 0,42 0,02 0,14 0,06 -0,02 N 50 50 50 50 50 50
The table reports the intercept model results for the industry, size and pre- performance adjusted financial ratios for the 50 bidder banks for the period of 2002 to 2012. The ISPAROE post,i , ISPAPROV post,i ,
ISPACAPITAL post,i , ISPALIQ post,i, ISPAOPERATING post,i and ISPAEFFICIENCY post,i are the post-M&A
industry, size and pre- performance adjusted ROE, PROV, CAPITAL, LIQ, OPERATING and EFFICIENCY, respectively. Numbers in the parenthesis are the t -statistics. (*), (**) and (***) significantly different from zero at the 10%, 5% and 1%, respectively. aPresence of heteroskedasticity, White's (1980) correction for standard errors is used.
4.1.3 Relation between ISPAROE and ISPALIQ
Because ISPALIQ is the only performance indicator that has significantly changed, I conduct a simple linear regression to test its impact on the improvement of the ISPAROE. In this case, I use the change in ISPAROE between the pre- and post-M&A years as the dependent variable and the change in ISPALIQ for the independent variable. The simple regression is as follows:
ISPAROE = α + ISPALIQ + , (9)
where ISPAROE is the change in the industry, size and pre-performance adjusted ROE between the pre- and post-M&A years and ISPALIQ is the change in the industry, size and pre-performance adjusted liquidity indicator between the pre- and post-M&A years. is the error term.
19 Table 6. Simple regression results
Independent variables ISPAROEa t-statistics
Intercept (%) 6,29 1,27
ISPALIQ 2,356* 1,69
F-statistic ***8,43
Adjusted R-squared 0,13
N 50
The table reports the simple regression results for the dependent variable ISPAROE and the
independent variable ISPALIQ. (*), (**) and (***) significantly different from zero at the 10%, 5% and 1%, respectively. a Presence of heteroskedasticity, White's (1980) correction for standard errors is used.
4.1.4 Discussion of the results
By comparing the tables 3, 4, 5 and 6, some observations can be made. First, both the change model and the intercept model report the same results. They both report bank performance improvements due to the M&As using the ratio return on equity, controlled for the industry, size and pre- performance. The same accounts for the performance indicator liquidity. The intercept of the intercept model and the difference estimate of the change model report significant positive results for ISPALIQ. Also for the performance indicators ISPAPROV, ISPACAPITAL, ISPAOPERATING and ISPAEFFICIENCY both models report insignificant results.
Second, I find significant positive results for ISPAROE. This result is inconsistent with Knapp et al. (2005), who finds significant negative results. This inconsistency is due to the difference in the type of financial institution studied. Knapp et al. (2005) study impact of M&As on bank holding companies while I focus on banks.
Third, the M&As do not have an impact on the performance indicators ISPAPROV and
ISPAEFFICIENCY. This result is consistent with Ismail et al. (2009), who also conducted both the change model and the intercept model.
Fourth, the M&As also do not have an impact on the performance indicators ISPACAPITAL and ISPAOPERATING. These results are inconsistent with Altunbaş and Marqués (2008), who finds significant results for both indicators. A reason for this inconsistency is the use of
unadjusted financial ratios by Altunbaş and Marqués (2008). As mentioned above it is important to control for industry factors that may influence the financial ratio apart from the M&A.
20 analysis is conducted between the change in ISPAROE and the change in ISPALIQ. It shows a significant positive relation between the ISPALIQ and the ISPAROE. This result is inconsistent with Kwan and Eisenbeis (1999), who finds a significant decrease. However, it is consistent with Cornett and Tehranian (1992) and Kuo et al. (2010). Kuo et al. (2010) analyzed the impact of loans on bank performance for listed Taiwanese banks for the period of 1998 to 2002. As their performance indicator, they also use return on equity. Using their motives, I conclude that the positive impact of the ratio loans to asset on the ISPAROE is due to loans being the most important service of the banks, making the bank more reliable on loans to earn profit (Kuo et al., 2010). Cornett and Tehranian (1992) also argues that the positive relation suggest that the bidder banks can attract more loans per dollar total assets after the M&A, which increase profitability. In sum, I conclude that hypothesis 1 is true, suggesting that the M&As
announcements in the South and Central American banking industry do have an impact on the operating performance of the bidder banks using return on equity as the performance measure. 4.2 Event study results
Table 7 reports the cumulative abnormal returns (CAR) for the different event windows. The total sample consists of 38 bidder bank M&As transactions for the period of 2002 to 2012. At the announcement date [0], the CAR of the bidder banks is 0,0016, which is positive but statistically insignificant. For all the different event windows, the CARs remain positive and statistically insignificant. Thus, the shareholders of the bidder banks did not lose shareholder value.
4.2.1 Relation between CAR and ISPROE
Similar to Powel and Stark (2005) and Ismail et al. (2009) I conduct an OLS regression
between the CAR and the ISPAROE post,i for the 38 banks, to test if the market participants can
anticipate at the time of the M&A announcement post-M&A bank performance improvements. The OLS regression for the relation between CAR and ISPAROE is as follows:
ISPAROE post,i = α + β1 ISPAROE pre,i + β2 CAR + , (10)
where ISPAROE pre,i is the pre-M&A industry, size and pre-performance adjusted ROE and
ISPAROE post,i is the post-M&A industry, size and pre-performance adjusted ROE. CAR is the
21 Table 8 reports the results of the OLS regression. First, both the intercept and the
pre-performance are not significantly different from zero. So there is no evidence of significant change in post-M&A bank performance due to the M&As and also no significant association between the post-M&A performance and the pre-M&A performance for the 38 bidder banks. It also reports an insignificant relation between the CAR and the post-M&A bank performance, with a coefficient of 0,458. This implies that the market participants are unable to anticipate post-M&A bank performance improvements and impound these improvements in the market prices at the time of the M&A announcement. These results are consistent with Powel and Stark (2005) and Ismail et al. (2009). They also find insignificant results.
Table 7. Results of the event study
Event window CAR T-test P-value
[-10;0] 0,0092 0,0294 0,9767 [-5;0] 0,0031 0,0130 0,9897 [-1;0] 0,0014 0,0105 0,9917 [0] 0,0016 0,0170 0,9865 [-1;+1] 0,0061 0,0386 0,9694 [-5;+5] 0,0151 0,0480 0,9620 [-10;+10] 0,0166 0,0386 0,9695
22 Table 8. Relation CAR and ISPAROE
Independent
variables ISPAROE post
Intercept 0,026 (1,10) Pre-performance 0,028 (0,13) CAR 0,458 (0,35) F-statistic 0,06 Adjusted R-squared -0,05 N 38
The table reports the OLS regression results for the dependent variable ISPAROE post
and the independent variable CAR for the 38 bidder banks. (*), (**) and (***) significantly different from zero at the 10%, 5% and 1%, respectively.
4.2.2 Discussion of the results
Based on the results of table 7, I conclude that the M&A announcements did not have an impact on the stock prices of the 38 bidder banks. So I reject hypothesis 2. These findings are not consistent with several US studies, which report significant positive or negative CARs for bidder banks (Desai et al., 1985; James and Wier, 1987; Baradwaj et al., 1992). Nevertheless the findings are in line with Goddart et al. (2012), who also finds positive, but insignificant CARs for a group of 42 M&A transactions of Latin American banks. Also, the results are consistent with some EU studies, which report insignificant CARs for bidder banks (Rad and Van Beek, 1999; Beitel et al., 2004; Asimakopoulos and Athanasoglou, 2013). A reason for the insignificant results is that the size of the targets are smaller in comparison with the size of the bidders and thus do not have an impact on the stock price of the bidders (Rad and Van Beek, 1999). An example is Brazilian bidder bank Banco Bradesco, who acquired the Brazilian bank Banco BMC SA in 2007. The total assets of the bidder bank is 9389% higher than the total assets of the target bank.
4.3 Robust analysis
4.3.1 Change model using medians
non-23 parametric Wilcoxon signed rank test is used to test the statistical significance of this
difference. Similar to table 3 and 4, table 9 reports for the performance measure ROE and the performance indicator liquidity significantly positive changes and for the other four
performance indicators insignificant changes. Also by using medians, the impact of outliers is controlled, which makes the estimates less volatile. This result is consistent with Ghosh, (2001), who also finds the same results for using medians instead of means. So based on the change model results using median hypothesis 1 is accepted, suggesting that the M&As announcements in the South and Central American banking industry do have an impact on the operating performance of the bidder banks using return on equity as the performance measure. 4.3.2 Event study based on the market adjusted model
24 Table 9. Change model (medians)
Industry, size, pre-performance
adjusted
Indicators (%) Pre-M&A Post-M&A Difference Panel A
Performance measure
Return on equity (ROE) 0,62 3,47 1,25*
Panel B
Assets quality indicator
(1) Loan loss provision to net interest
revenue (PROV) 1,48 -1,34 -3,48
Capital indicator
(2) Equity to total assets (CAPITAL) 0,04 0,39 0,02 Liquidity indicator
(3) loans to total assets (LIQ) -0,39 2,15 1,25*
0,12 0,18 0,02
Operating indicator
(4) Other operating income to total assets (OPERATING)
0,29 1,7 1,04
Efficiency indicator
(5) cost to income (EFFICIENCY)
25 Table 10. Market adjusted model results
Event window CAR T-test P-value
[-10;0] 0,0056 0,0178 0,9859 [-5;0] -0,0001 -0,0004 0,9997 [-1;0] --0,00001 -0,0001 0,9999 [0] 0,0016 0,0171 0,9865 [-1;+1] 0,0072 0,0450 0,9643 [-5;+5] 0,0127 0,0399 0,9684 [-10;+10] 0,0098 0,0226 0,9821
The table shows the results of the event study for the 38 bidder banks in South and Central America for the period of 2002 to 2012. The CARs are calculated using the market adjusted model approach, where the alpha is zero and the beta is one. The CARs are reported in decimals and not in percentage. (*) 10%, (**) 5% and (***) 1% significance level, respectively.
5. Conclusion
This paper analyze the impact of the M&A announcements on South and Central American bidder banks for the period of 2002 to 2012 using both accounting studies and event study. The findings for the accounting studies, using a dataset of 50 bidder banks revealed
improvements in post-M&A bank performance. As performance measure I used the industry, size and pre-performance adjusted ratio return on equity. I also find that improved bank performance is due to the improvements in the bank's ability to attract more loans and as a result increase profitability. Further, both models of accounting studies, which are the change model and intercept report the same results.
The finding for the event study, using a dataset of 38 bidder banks revealed that the M&A announcements did not have an impact on the stock prices, which imply that the shareholders of the bidder banks did not lose shareholder value. A possible explanation is that the size of the targets compared to the size of the bidders are relatively small, to have an impact on the stock prices (Rad and Van Beek, 1999) Further, I did not find a relation between the post-M&A bank performance and the cumulative abnormal returns. This implies that the market participants are unable to anticipate post-M&A bank performance improvements and impound these
26 For further research, it will be great to use a larger dataset for conducting the accounting
studies and the event study and also analyzing the changes of more performance indicators.
References
Altunbaş, Y., Marques, D., 2008. Mergers and acquisitions and bank performance in Europe: the role of strategic similarities. Journal of Economics and Business 60, 204-222.
Andonova, V., Rodrigues, Y., Sanchez, I.D., 2013. When waiting is strategic: Evidence from Colombian M&As 1995-2008. Journal of Business Research 66(10), 1736-1742.
Asimakopoulos, I., Athanasoglou, P.P., 2013. Revisiting the merger and acquisition performance of European banks. International Review of Financial Analysis 29, 237-249. Baradwaj, B.G., Dubofsky, D.A., Fraser, D.R., 1992. Bidder returns in interstate and interstate bank acquisitions. Journal of Financial Services Research 5, 261-273.
Beitel, P., Schiereck, D., Wahrenburg, M., 2004. Explaining the M&A-success in European banks. European Financial Management 10, 109-139.
Berger, A.N., Demsetz, R.S., Strahan, P.E., 1999. The consolidation of the financial services industry: causes, consequences and implications for the future. Journal of Banking and Finance 23, 135-194.
Brown, S., Warner, J., 1985. Using daily stock returns: the case of event studies. Journal of Financial Economics Vol. 14 1985, 3-31.
Carvallo, O., Kasman, A., 2005. Cost efficiency in the Latin American and Caribbean banking systems. Journal of international financial Markets, Institutions and Money 15(1), 55-72. Cornett, M.M., Tehranian, H., 1992. Changes in corporate performance associated with bank acquisitions. Journal of Financial Economics 31(2), 211-234.
Cornett, M.M., McNutt, J.J., Tehranian, H., 2006. Performance changes around bank mergers: revenue enhancements versus cost reductions. Journal of Money, Credit and Banking 38(4), 1013-1015.
27 Desai, A.S., Stover, R.D., 1985. Bank holding company acquisitions, stockholder returns, and regulatory uncertainly. Journal of Financial Research 8, 145-156.
Diaz, B., Garcia Olalla, D., Sanfilippo Azofra, S., 2004. Bank acquisitions an performance: evidence from a panel of European credit entities. Journal of Economics and Business 56, 377-404.
Dyckman, T., Philbrick, D., Stephan, J., 1984. A comparison of event study methodologies using daily stock returns: a simulation approach. Journal of Accounting Research 22, 1-30. Fama, E.F., 1970. Efficient capital markets: A review of theory and empirical work. The journal of Finance 25(2), 383-417.
Fridolfsson, S.O., Stennek, J., 2005. Why mergers reduce profits and raise share prices-a theory of preemptive mergers. Journal of The European Economic Association 3(5), 1083-1104. Ghosh, A., 2001. Does operating performance really improve following corporate acquisitions? Journal of Corporate Finance 7(2), 151-178.
Goddard, J., Molyneux, P., Zhou, T., 2012. Bank mergers and acquisitions in emerging markets: evidence from Asia and Latin America. The European Journal of Corporate Finance 18(5), 419-438.
Hawkins, J., Mihaljek, D., 2001. The banking industry in the emerging market economies: competition, consolidation and systematic stability: an overview. Bis Papers 4, 1-44. Healy, P., Palepu, K., Ruback, R., 1992. Does corporate finance improve after mergers? Journal of Financial Economics 31, 135-175.
Housten, J. H., Ryngaert, M., 1994. The overall gain from large bank mergers. Journal of Banking and Finance 18, 1155-1176.
Ismail, A., Davidson, I., Frank, R., 2009. Operating performance of European bank mergers. The Service Industries Journal 29(3), 345-366.
James, C., Wier, P., 1987. Returns to acquirers and competition in the acquisition market: the case of banking. Journal of Political Economy 95, 355-370.
28 Khan, A.A., 2011. Merger and acquisitions (M&As) in the India banking sector in post
liberalization regime. International Journal of Contemporary Business Studies 2(11), 31-45. Knapp, M., Gart, A., Becher,D., 2005. Post-merger performance of bank holding companies, 1987-1998. The Financial Review 40(4), 549-574.
Kuo, H.C., Wang, L.H., Lai, Y.H., Yu, S., Wu, C., 2010. Loan policy and bank performance: evidence from Taiwan. Bank and Bank Systems 5(2), 108-120.
Kwan , S., Eisenbeis, R., 1999. Mergers of publicly traded banking organisations revisited. Federal Reserve Bank of Atlanta Economic Review 84(4), 26-37.
Loughran, T., Ritter, R., 1997. The operating performance of firms conducting seasoned equity offerings. Journal of Finance 52, 1823-1850.
MacKinlay, A.C., 1997. Event studies in economics and finance. Journal of economic literature 35(1), 13-39.
Mantravadi, D.P., Reddy, A.V., 2008. Post-merger performance of acquiring firms from different industries in India. International Research Journal of Finance and Economics (22), 192-204.
Metwalli, A.M., Tang, R.Y., 2004. Update: M&A in Latin America. Journal of Corporate Accounting and Finance 15(2), 43-51.
Molyneux, P., Remolona, E., Seth, R., 1998. Modeling foreign bank performance and lending behavior. Financial Markets, Institutions & Instruments 7(4), 26-41.
Powell, R. G., Stark, A,W., 2005. Does operating performance increase post-takeover for UK takeovers? A comparison of performance measures and benchmarks. Journal of Corporate Finance 11(1-2), 293-317.
Rad, A.T., Van Beek, L., 1999. Market valuation of European bank mergers. European Management Journal 17(5), 532-540.
Roll, R., 1986. The hubris hypothesis of corporate takeovers. Journal of Business 59((2) Part 1), 197-216.
29 White, H., 1980. A heteroscedasticity consistent covariance matrix and a direct test for
heteroscedasticity. Econometrica 48, 817-838.
