The financial crisis and abnormal returns caused by
mergers and acquisitions in the banking industry.
Lars Theunissen
Student Number: 10252738
Faculty of Economics and Business June 26th, 2018
BSc Economics and Business
Specialization: Finance and Organization
Bachelor Thesis, 12 EC’s Supervisor: Robin Döttling
2
Statement of originality
This document is written by Lars Theunissen who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.
The faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3
Abstract
This paper studies the influence of the recent financial crisis on abnormal returns around the announcement date of mergers and acquisitions in the banking industry in the United States. The resolution of failing target banks is investigated to see the effect of the financial crisis. An event study in combination with a cross-sectional regression model is used to infer that the abnormal returns of the target companies are positively influenced by the recent financial crisis. Whereas the results for the relationship between bidder abnormal returns and the financial crisis are insignificant.
4 Table of contents Statement of originality ... 2 Abstract ... 3 Table of contents ... 4 1. Introduction ... 5 2. Literature review ... 7
2.1 Mergers and acquisitions ... 7
2.2 Financial crisis ... 8
2.3 Other variables ... 9
2.4 Hypotheses ... 11
3. Data and methodology ... 12
3.1 Data ... 12 3.2 Methodology ... 12 3.3 Control variables ... 14 4. Empirical results ... 16 4.1 Targets ... 16 4.2 Acquirers... 20
5. Conclusion and discussion ... 25
5.1 Conclusion ... 25
5.2 Discussion ... 25
Reference list ... 27
5
1. Introduction
Mergers and acquisitions are a common feature worldwide which also applies to the banking sector. Besides, the recent financial crisis has had a substantial influence in this industry, for example the collapsing of banks, banks that became ‘Too-Big-To-Fail’, etc. Therefore, when investigating the abnormal returns caused by mergers and acquisitions it is interesting to see what the effect is of such a crisis, especially the effect of the resolution of failing banks. During the recent financial crisis banks faced more problems, especially liquidity problems (Diamond and Rajan, 2009). Actual returns are expected to be lower during an economic downturn, because banks are experiencing such problems. Therefore, does this imply that the crisis will have a negative influence on abnormal returns? Or is it because of the financial constraints which banks are facing in such periods that we are only observing the ‘good’ deals? In a recession most banks have a shortage of capital, so when mergers and
acquisitions do take place this can be a signal of firms performing well. If this is the case the financial crisis can have a positive effect on abnormal returns. To find out what the actual influence will be the following research question is composed:
Does the recent financial crisis have an effect on the abnormal returns caused by mergers and acquisitions in the banking industry in the United States?
DeYoung et al. (2009) review the literature on mergers and acquisitions in the financial industry after the year 2000 and they discuss more than 150 papers. Their conclusion is that European bank mergers increase shareholder value whereas the wealth effects for shareholders involved in M&A deals in North America still remain uncertain. This makes the United States an interesting area for this research to investigate the abnormal returns caused by mergers and acquisitions. Together with taking the financial crisis into account and even the effect of M&A deals used to resolve the financial problems of target companies this paper will contribute to the existing literature.
To answer the research question the cumulative abnormal returns are calculated first using the value weighted market model. These cumulative abnormal returns are calculated for 77 target and 116 acquiring banks. Next, a regression model is used to find the effect of the recent financial crisis on the abnormal returns of both target and acquiring companies. The influence of the resolution of failing banks on abnormal returns will be used to answer the research question. A positive relationship is found between the resolution of distressed
6
banks and the abnormal returns of these target companies, whereas the results for this relationship for the acquiring companies are insignificant.
The next section of this paper reviews the most relevant literature in this area and uses the findings of these papers to state the hypotheses. Section 3 describes where the data comes from, how the sample has been selected and which model has been used. The empirical results are shown and analysed in section 4. The last section states the conclusions which are based on the results from section 4 together with a discussion on improvements for this research and recommendations for further research.
7
2. Literature review
2.1 Mergers and acquisitions
Mergers and acquisitions are a common phenomenon in the economic world and they happen in waves. Takeover activity is greatest during a period of economic expansion when stock markets are booming (Betton et al., 2008; Martynova and Renneboog, 2008). This makes sense because mergers and acquisitions are part of growth strategies. There are several reasons for firms to involve in M&A deals. First of all, the merging companies can have expectations to create synergies. These synergies could result from cost reductions and an increase in efficiency (Houston and Ryngaert, 1994; Betton et al., 2008; Ismail, 2011). This can be realized through economies of scale and economies of scope which merging
companies can achieve. Second, firms taking part in a takeover can benefit from monopoly rents and therefore increase their revenue, which is referred to as market power effects (Pilloff, 1996; Betton et al., 2008; Brewer and Jagtiani, 2013). These two reasons can result in an increase of profits and they can create value for the shareholders of the merging parties. On the other hand, some takeover motives will destroy (shareholder) value. These motives are mostly related to managers acting in their own interest, for instance managers
entrenching themselves (Shleifer and Vishny, 1989), managerial hubris (Roll, 1986) and other agency costs (Jensen, 1986). Finally, mergers and acquisitions can be used to resolve the financial problems which target companies are facing, especially in times of a financial crisis (White and Yorulmazer, 2014).
Martynova and Renneboog (2008) review the overall M&A literature of a whole century and conclude that abnormal returns for target shareholders are (nearly) always positive and significant. Besides, they infer that the returns for shareholders of acquiring firms are insignificant and that previous literature does not agree on whether the combined value of merging firms will increase or not. In contrast, Roll (1986) and Moeller et al. (2004) state that abnormal returns to bidders are often negative. Furthermore, Betton et al. (2008) show significantly positive returns for acquirers and targets together for the period 1980-2005.
When looking at the literature for mergers and acquisitions in the financial industry the drawn conclusions about abnormal returns around the announcement date are more or less the same. Most studies find that target shareholders earn significantly strong positive
8
abnormal returns, acquirers earn slightly negative returns and the combined value of the targets and acquirers together is insignificant, but some studies do find positive total stockholder wealth effects as well. First of all, findings for target returns are always positive and significant, but differences in percentages are substantial, e.g. 1.18% (Frame and
Lastrapes, 1998) and 33.58% (Neely, 1987). Kolaric and Schiereck (2014) review the empirical evidence of M&A deals and state that abnormal returns to targets are mostly between 10% and 15%. At the other hand, they also conclude that shareholders of the acquiring firm often earn slightly negative abnormal returns, which is supported by studies as Houston and Ryngaert (1994), DeLong (2001), Campa and Hernando (2006) and Brewer and Jagtiani (2013). However, a few studies do report positive returns to bidders (e.g. Desai and Stover, 1985; James and Wier, 1987). The overall results for stockholder value of merging firms is often insignificant for the banking industry (Houston and Ryngaert, 1994; Pilloff, 1996; DeLong, 2001), but some studies also find significantly positive results (Becher, 2000; Anderson et al., 2004). DeYoung et al. (2009) review more than 150 studies published since 2000 and they conclude that for the European banking industry the wealth effects for stockholders are positive, whereas this conclusion cannot be made for the M&A deals in North America.
2.2 Financial crisis
The recent financial crisis which started in 2007 had a tremendous impact in the banking industry. Banks faced more problems and even became insolvent, so governments
intervened to prevent the financial systems from collapsing. Fahlenbrach et al. (2012) states that banks with more short-term debt, more leverage and more growth potential have a higher chance to underperform during a recession. During the recent financial crisis banks with lower leverage and more capital performed better, as did smaller banks. This is also supported by Diamond and Rajan (2009) claiming that the capital structure of banks should rely on capital instead of short-term leverage in times of an economic downturn to prevent liquidity problems.
Martynova and Renneboog (2008) show that the combined value gains of takeovers are higher during takeover waves. These waves happen during economic expansion and end with the collapse of stock markets. The effect of the 1997 financial crisis on the performance of bank mergers in Asia is explained by Crouzille et al. (2008). Their findings demonstrate a
9
negative and significant effect on the total value of banks during a crisis. The explanation for this effect is that in times of a recession mergers and acquisitions are not only used to improve efficiency, but also to strengthen the financial system and to solve financial problems which banks are facing. Governments and central banks also play a role in these mergers. The effect the resolution of distressed banks through mergers and acquisitions has on abnormal returns for the shareholders of these merging firms remains uninvestigated. That is what this paper adds to the existing literature. According to White and Yorulmazer (2014) it can still be profitable to acquire failing banks, despite the approaching insolvency. This can be achieved through new customers and relationships. Furthermore, Beltratti and Paladino (2013) investigated the effect of the recent financial crisis (2007-2010) on the abnormal returns of bidding firms in Europe, but they failed to find significant results around merger announcements.
2.3 Other variables
It is possible that the recent financial crisis had an influence on abnormal stock returns generated by mergers and acquisitions in the financial sector. Following the existing literature there are other factors affecting these returns as well. To begin with, the size of merging firms and transactions can have an influence. When investigating corporate takeovers most studies find that bidder returns around the merger announcement are smaller for larger acquiring firms (Loderer and Martin, 1990; Moeller et al., 2004). Betton et al. (2008) support these findings with significantly positive results for smaller acquiring firms and significantly negative results for larger acquiring firms. Houston and Ryngaert (1994) conclude after reviewing the literature on M&A deals in the banking industry that larger acquisitions often lead to negative shareholder returns for the acquiring firm. This negative relationship is not supported by Campa and Hernando (2006), stating that larger
transactions lead to an increase in shareholder value in the European financial sector. Besides the absolute size of M&A deals, the relative size of target to bidder firm is often used in researches to investigate the relationship with abnormal returns. James and Wier (1987) and DeLong (2001) find a positive relationship between abnormal returns to shareholders and the relative size of merging firms. On the opposite, Brewer and Jagtiani (2013) find significantly negative results for the relative size of the target to the acquiring company.
10
Another factor which has an effect on the value creation for shareholders of companies involved in takeovers is the method of payment, according to the relevant literature. On this influence the findings of previous studies are less diverse. M&A deals financed with only stock payments underperform compared to takeovers using all-cash as their method of payment. This holds for target returns as well as bidder returns. Bidder returns are significantly negative when the financing method is all-stock. The same holds for combined abnormal returns. These findings are based on Martynova and Renneboog (2008), reviewing the empirical evidence of the performance of corporate takeovers and Betton et al. (2008), investigating corporate takeovers in the period 1980-2005. In addition, DeLong (2001) and Kolaric and Schiereck (2014) reach the same conclusion when reviewing the empirical evidence of bank M&As.
Also, the financial situation of the target firm before the takeover is an important factor which can affect the abnormal returns associated with these takeovers. This financial situation can best be described by stock performance and exposure to risk. Namely, target firms which underperform prior to a merger have a higher chance to be acquired and will lead to an increase in shareholder wealth effects (DeLong, 2001). The exposure to risk is best explained by the amount of capital or leverage. Banks with more capital or lower leverage performed better during the recent financial crisis, according to Fahlenbrach et al. (2012). This means banks with higher leverage during the crisis had an increased probability to be acquired.
Most studies find that a larger deal size or larger firms involving in mergers result in lower abnormal returns around the announcement date. Actually, in the banking industry this does not hold for very large banks. If banks are taking part in M&A activity the resulting firms can become ‘Too-Big-To-Fail’ after the takeover. This implies that the government will provide very large banks with subsidies. The government is obliged to give this support to these large banks, because if these banks will collapse the damage to the economic system will be even more disastrous. Firms involving in these megabank mergers can exploit the benefits associated with safety net subsidies. Therefore, the abnormal returns around the merger announcement of banks becoming ‘Too-Big-To-Fail’ are positive and significant (Kane, 2000; Brewer and Jagtiani, 2013).
11 2.4 Hypotheses
Based on the relevant literature the expectation is that mergers and acquisitions in the financial industry will lead to value creation or destruction for shareholders upon merger announcements. First of all, target abnormal returns are expected to be positive. In contrast, the returns for the acquiring company are expected to be (slightly) negative. To test these expectations the statistical hypotheses will be that target and bidder abnormal returns are both equal to 0.
During a financial crisis more takeovers are used for other reasons than to improve efficiency, for example to solve financial problems which banks are experiencing. Investors will take this into account and therefore the effect will be reflected in the abnormal returns around the announcement date. For this reason, the recent financial crisis is expected to have a negative impact on the abnormal returns to shareholders of acquiring firms. In contrast, target firms are often underperforming in times of a financial crisis and therefore abnormal returns of these companies are expected to be positive around announcement dates. The statistical hypothesis to test this will therefore be that the parameter measuring the effect of the resolution of failing banks on the abnormal rate of return will be equal to 0 and this is the null hypothesis. The main hypothesis which will be tested in this thesis, therefore, is:
H₀: The recent financial crisis has no influence on abnormal returns caused by mergers and acquisitions in the banking industry in the United States (dummy variable RESOLVE = 0). H₁: The recent financial crisis has an influence on abnormal returns caused by mergers and acquisitions in the banking industry in the United States (dummy variable RESOLVE ≠ 0).
12
3. Data and methodology
3.1 Data
The data which is needed for this research is collected from three databases, namely Zephyr, the Center for Research in Security Prices (CRSP) and Compustat. Zephyr is a database which is often used for mergers and acquisitions. Information on M&A deals is gathered here, such as the names of target and acquiring companies, announcement dates, deal values and the method of payment of each transaction. In order to select the sample for this research only mergers and acquisitions are included. The announcement dates of these mergers and acquisitions are between 01/01/2004 and 30/06/2014. Another requirement is that both companies, target and bidder, are situated in the United States at the time of the takeover announcement. Furthermore, the deal needs to be completed and the banking industry is selected for both target and acquiring companies. Finally, the companies involved in the mergers and acquisitions have to be listed companies during the estimation and event windows. This last criteria is necessary to find the stock returns of the firms.
After meeting all of the requirements the sample consists of 150 M&A transactions. The choice to include more than 100 M&A deals in this research is done to get more reliable results. The stock returns of the merging firms are then collected from CRSP together with the CRSP value weighted index. After collecting all stock price data, cleaning the data and looking for enough observations within the estimation and event windows for each company, there are 116 acquiring and 77 target companies left. Other firm-specific
information, such as the total assets of a company and the Tier 1 risk-adjusted capital ratio, is collected from the Compustat database. Finally, to determine if the takeover was carried out to resolve the possible financial problems of the target bank, the merger and acquisition announcements of the takeovers announced during the recent financial crisis and the years afterwards were checked manually.
3.2 Methodology
An event study is used to determine whether the mergers and acquisitions in the banking industry in the United States create or destroy value for the shareholders during the period investigated in this research. According to Kolaric and Schiereck (2014), an event study is the best methodology when defining the success of a takeover. They also conclude that the
13
market model is definitely the most used model in empirical studies. First of all, Cable and Holland (1999) show that the market model should be preferred to the mean and market adjusted return models in an event study. Besides, there are some multifactor models which could be used to calculate the expected returns in an empirical study, such as the Fama and French three factor model or the Carhart four factor model. These models would probably have a higher explanatory power than the single factor models, but significant differences have not been found in the overall results (Kolaric and Schiereck, 2014). For these reasons, the market model is used in this study and in particular the value weighted market model which is also used by Pilloff (1996), DeLong (2001) and Betton et al. (2008).
First the value weighted market model is used to calculate the cumulative abnormal returns (CARs). Afterwards, a cross-sectional regression model is explained to test these CAR estimates. The market model uses an estimation window and an event window. The
estimation window is chosen from day -293 until day -42, where day 0 is the announcement date of the takeover. This window already stops 42 days before the announcement date, which is done to exclude the effects of rumours going on about the mergers and acquisitions before the actual announcement is made (Schwert, 1996). The event window is chosen from day -2 until day 2. This event window is often used in recent studies covering the latest takeover wave (Martynova and Renneboog, 2008). The expected stock returns can be calculated with the following formula:
Rit = αi + βi*Rmt + εit
The CRSP value weighted index is used as benchmark for the market index (Rmt). The expected returns can be estimated by a regression which is done in Stata. After that the abnormal return (AR) can be calculated by subtracting the expected return (Rit) from the actual return (Ra). Then the cumulative abnormal return (CAR) can be calculated to see if the actual return is different from the expected return during the event window for all
observations together:
CAR = ∑ Ra – Rit
Also the average cumulative abnormal return (ACAR) can be calculated by taking the average of the CAR. To test if the ACAR is significantly different from 0 t-tests are used. In addition, a regression is done to calculate the cumulative abnormal returns for all companies together. The p-value of the constant from this regression is used to determine the significance. If these p-values are significant then the conclusion can be made that the mergers and
14
acquisitions do indeed have an effect on the stock returns of these companies. All tests have been running with a significance level of 10%, 5% and 1% and robust standard errors have been used.
Then a cross-sectional regression model is used to test if the recent financial crisis has an effect on the abnormal stock returns. The dependent variable is the cumulative abnormal return (CAR) and the main explanatory variable is the resolution of failing banks (RESOLVE), which is a dummy variable indicating 1 if the M&A deal was completed to resolve the financial problems which a target company was facing. Other variables included in this regression are the deal value, the relative size of target to bidder, a dummy variable
indicating whether the deal was fully paid in cash, a dummy variable indicating whether the deal was financed with a mix of cash and shares, the financial crisis period, the period following the crisis, the capital ratio of the target bank and a dummy variable indicating whether the combined firm becomes ‘Too-Big-To-Fail’. In section 3.3 further information is given on these control variables. The model therefore is:
CAR = β0 + β1 * LNDEALVALUE + β2 * RELATIVESIZE + β3 * CASH + β4 * MIX + β5 * CRISIS + β6 * POSTCRISIS + β7 * RESOLVE + β8 * TIER 1 + β9 * TBTF + ε
If β8 ≠ 0 then the financial crisis and the CAR are related and we can conclude that the recent financial crisis has an influence on abnormal returns caused by mergers and acquisitions in the banking industry in the United States.
3.3 Control variables
In section 2.3 of the literature review, some variables are described which may also have an influence on shareholder wealth effects, other than the resolution of distressed banks. To control for these influences variables have been included in the cross-sectional regression model. Some of these control variables are:
- the deal value (LNDEALVALUE), which is the natural logarithm of the total amount of the M&A deal.
- the relative size (RELATIVESIZE), which is the total assets of the target firm divided by the total assets of the bidder before the takeover announcement was made.
- the method of payment (CASH), which is a dummy variable indicating whether the deal was fully paid in cash or not.
15
- the method of payment (MIX), which is a dummy variable indicating whether the deal was financed with a combination of cash and shares or not.
- the recent financial crisis period (CRISIS), which is a dummy variable indicating 1 if the M&A deal was announced between July 2007 and December 2010.
- the period following the crisis (POSTCRISIS), which is a dummy variable indicating 1 if the M&A deal was announced between January 2011 and June 2014.
Furthermore another control variable will be added to look at the ‘collapsing’ of banks. This will be done by using the Tier 1 risk-adjusted capital ratio of the target bank (the total adjusted capital divided by the risk-weighted assets) before the merger announcement was made. Namely, this shows the condition of the target bank before the takeover. Banks with more leverage have a higher chance to face more problems during a financial crisis. Therefore, banks with more leverage and less capital were otherwise (if they did not take part in a merger) more likely to collapse. Moreover, the Tier 1 ratio values give an indication of the financial condition of the target bank, especially during a financial crisis.
The last control variable in the regression model will look at the effect of banks that become ‘Too-Big-To-Fail’ as a consequence of a takeover. This will be done by adding the total assets of the merging firms. If this total amount exceeds the threshold of 100 billion dollars, the dummy variable will be 1 and otherwise 0. Brewer and Jagtiani (2013) also use the threshold of 100 billion dollars to investigate the effect of banks becoming too big to fail. For this control variable the values of the total assets are from before the announcement date as well.
16
4. Empirical results
In this section the results of the event study and the regression model will be discussed. First, the results of the target companies are analysed in section 4.1 and in the following section the results of the acquiring companies are studied.
4.1 Targets
First of all, the results of the event study are analysed to see if the abnormal returns are significantly different from 0. Using the value weighted market model the cumulative abnormal returns have been calculated for 77 target companies. Afterwards, a regression was done to calculate the cumulative abnormal return of all these companies treated as a group. The results of this regression are shown in table 1 below.
Table 1. Regression CARs targets
Regression of the cumulative abnormal returns of 77 target companies.
Coefficient Robust Std. Err.
P-value
Constant 0.2172 0.019 0.000
The p-value of the constant from this regression gives the information about the statistical significance across all companies. The advantage compared to a t-test is that robust standard errors are used in this case. The coefficient 0.2172 indicates that the cumulative abnormal returns of the target companies increase with 21.72%. The corresponding p-value 0.000 means that this increase is significant at the 10%, 5% and 1% levels. This confirms the expectation that target abnormal returns would be positive. Therefore, the statistical hypothesis, stating that the abnormal returns of the target companies will be equal to 0, is rejected.
The increase of 21.72% is in line with previous literature in terms of target returns always being significantly positive, but varying between 1.18% (Frame and Lastrapes, 1998) and 33.58% (Neely, 1987). According to Kolaric and Schiereck (2014), the abnormal returns to targets are mostly between 10% and 15% for the banking sector. This means the result found in this paper is above average. A possible reason for this result is the resolution of failing target banks during the recent financial crisis and the years after the crisis when there still were many failures. Namely, DeLong (2001) states that target firms which underperform
17
prior to a merger are more likely to be acquired and this results in an increase in the abnormal returns for these target shareholders.
Next, the regression model explained in section 3.2 has been used to see if and how the variables included in this model influence the cumulative abnormal returns. In table 7 (see section Appendices at the end of this paper) the correlation matrix for the variables used in the regression model for the target companies is given. The highest correlation coefficient in this table is 0.529 between the variables Resolve and Crisis. This means there is no multicollinearity in this model.
Table 2. Descriptive statistics targets
In this table a summary is given of the variables used in the regression model for the target companies. The dependent variable CAR is the cumulative abnormal return. The lndealvalue is the natural logarithm of the transaction value of a takeover. The relativesize is the total assets of the target company divided by the total assets of the acquiring company. Cash is a dummy variable and equals 1 if only cash was used as the method of payment. Mix is a dummy variable and equals 1 if a mix of cash and shares was used as the method of payment. Crisis is a dummy variable and equals 1 if the takeover was announced between July 2007 and December 2010. Postcrisis is a dummy variable and equals 1 if the takeover was announced between January 2011 and June 2014. Resolve is a dummy variable and equals 1 if the takeover was used to resolve the financial problems of the target company. Tier 1 is the Tier 1 capital ratio of the target company. TBTF is a dummy variable and equals 1 if the total assets of both companies exceeds 100 billion dollars.
Variables Obs Mean Std. Dev. Min Max
CAR 77 0.217 0.164 -0.053 0.701 Lndealvalue 74 18.677 1.136 16.455 22.197 Relativesize 75 0.335 0.320 0.002 1.084 Cash 75 0.08 0.273 0 1 Mix 75 0.587 0.496 0 1 Crisis 77 0.130 0.338 0 1 Postcrisis 77 0.442 0.500 0 1 Resolve 77 0.052 0.223 0 1 Tier 1 67 11.708 2.975 5.8 18.55 TBTF 75 0.053 0.226 0 1
In table 2 above the descriptive statistics can be found of the variables which have been used in the regression model for the target companies. In this table the number of
observations, the mean, the standard deviation, the minimum value and the maximum value are given for the dependent variable (the cumulative abnormal return) and the independent variables. To see if there are any outliers a possible method is applying winsorizing. This method implies adding and subtracting a certain threshold to the mean value. All values which are more extreme can be put at the point of the threshold. The dummy variables are excluded from this method. Namely, in table 2 the minimum and maximum values are 0 and 1 for all dummy variables which are the correct values. The variable lndealvalue is already in
18
levels because of the natural logarithm, so it is not needed to winsorize this variable as well. Using a threshold of three standard deviations there are no extreme values for the other variables (the cumulative abnormal returns, the relative size ratio and the Tier 1 capital ratio). This means applying winsorizing is not needed here.
To see if the explanatory variables included in the regression model do influence the cumulative abnormal returns of the target companies five regressions were done. The results of these regressions are presented in table 3 below.
Table 3. Regression results targets
In this table the results from the OLS regressions for the target companies are shown. The dependent variable is the cumulative abnormal return. The lndealvalue is the natural logarithm of the transaction value of a takeover. The relativesize is the total assets of the target company divided by the total assets of the acquiring company. Cash is a dummy variable and equals 1 if only cash was used as the method of payment. Mix is a dummy variable and equals 1 if a mix of cash and shares was used as the method of payment. Crisis is a dummy variable and equals 1 if the takeover was announced between July 2007 and December 2010. Postcrisis is a dummy variable and equals 1 if the takeover was announced between January 2011 and June 2014. Resolve is a dummy variable and equals 1 if the takeover was used to resolve the financial problems of the target company. Tier 1 is the Tier 1 capital ratio of the target company. TBTF is a dummy variable and equals 1 if the total assets of both companies exceeds 100 billion dollars. The coefficients of each regression are shown below together with the standard errors in parentheses. All standard errors are robust standard errors. (1) (2) (3) (4) (5) Lndealvalue -0.0156 (0.014) -0.0116 (0.015) -0.0095 (0.013) -0.0219 (0.015) -0.0297* (0.018) Relativesize -0.2647*** (0.046) -0.2861*** (0.053) -0.2721*** (0.046) -0.2765*** (0.053) -0.2871*** (0.046) Cash 0.0405 (0.041) 0.0138 (0.052) 0.0333 (0.040) 0.0223 (0.053) 0.0479 (0.064) Mix 0.0333 (0.036) 0.0221 (0.041) 0.0324 (0.037) 0.0201 (0.040) 0.0396 (0.039) Crisis 0.0466 (0.053) 0.0402 (0.068) 0.0902* (0.052) -0.0541 (0.060) 0.0548 (0.062) Postcrisis 0.0895** (0.036) 0.0712 (0.046) 0.0969*** (0.036) 0.0494 (0.047) Resolve 0.1332* (0.068) 0.2128*** (0.060) 0.1895*** (0.068) Tier 1 0.0064 (0.009) 0.0086 (0.009) 0.0051 (0.008) TBTF -0.2643*** (0.070) -0.1940*** (0.064) -0.2106*** (0.065) -0.2733*** (0.066) -0.3250*** (0.074) Constant 0.5376** (0.249) 0.4171 (0.277) 0.4227* (0.246) 0.5962** (0.278) 0.6929** (0.312) Year fixed effects No No No No Yes Obs 72 63 72 63 63 R² 0.3918 0.3744 0.3729 0.4187 0.5032
19
If year fixed effects were used in the regression, the number of observations and the R-squared are given at the bottom of the table as well. Regression 4 in table 3 is the regression where all variables are included, whereas in the other regressions one or two variables are omitted. In regression 5 the variable postcrisis is omitted and instead year fixed effects were used in this regression. When looking at the R-squared regression 5 has the highest value (0.5032) indicating that this regression model has the highest explanatory power.
The main explanatory variable is the dummy variable Resolve. The regression coefficient for this variable is significant at the 10% level in the first regression and even significant at the 1% level in regressions 4 and 5. In all three cases the coefficient is positive. This means if the reason of the takeover is to resolve the financial problems which a target bank is facing then this will increase the abnormal returns for the target shareholders of these banks. This relationship is supported by the literature. DeLong (2001) states that underperforming target firms are more likely to be acquired and as a result this will increase the shareholder wealth effects of these companies. This seems logical because the
announcement that a healthy bank will acquire a bank in financial distress is positive news for these target banks.
Furthermore, the dummy variable Crisis is only significant in regression 3 where the variables Resolve and Tier 1 are excluded. That the main explanatory variable Resolve is omitted in regression 3 may be the reason that the variable Crisis is significant in this regression and not in the other regressions. In regression 3 the coefficient for Crisis is significantly positive at the 10% level. This indicates there is a positive relationship between the cumulative abnormal return and the dummy variable Crisis for target companies. This is in contradiction with the existing literature, but not too much value should be attached to this result. Namely, in regressions 4 and 5 the results for the variable Crisis are not significant and these regressions have the highest explanatory power.
Besides, significant results are shown in table 3 for the variables Relativesize, TBTF,
Lndealvalue and Postcrisis. In all regressions the coefficient for Relativesize shows a
significantly negative relationship with the cumulative abnormal returns. In addition, these results are all significant at the 1% level. Brewer and Jagtiani (2013) did also find significantly negative results for the relationship of the abnormal returns to shareholders and the relative
20
size of merging firms. At the other hand, there are also studies which found positive results for the coefficient of the relative size of target to bidding firm.
The results for the regression coefficient of the variable TBTF are in all regressions negative and significant at the 1% level as well. These results contradict the existing literature, because relevant studies show that there are ‘Too-Big-To-Fail’ advantages upon merger announcements for shareholders. Sometimes these advantages could be reduced if the merging companies are already too big to fail (DeLong, 2001), but in this research there is even a negative relationship.
In contrast to the first four regressions, the coefficient for the variable Lndealvalue shows up statistically significant at the 10% level in regression 5 which has the highest explanatory power. This significant coefficient is negative. A negative relationship between the size of a M&A transaction and shareholder returns for the merging firms is also found by the majority of the existing literature. This holds for corporate mergers and acquisitions as well as takeovers within the banking industry.
The last significant variable in table 3 is the dummy variable Postcrisis. In regression 1 and 3 the coefficient for this variable is positive and significant at the 5% and 1% level, respectively. These results do agree with the literature on this subject. Namely, Martynova and Renneboog (2008) show that the value gains of takeovers increase during takeover waves. These waves happen during economic expansion such as the periods following a financial crisis. Another possible reason why the variable Postcrisis shows up significantly in these regressions is because there still were many bank failures in the years after the financial crisis.
4.2 Acquirers
Next, the results of the acquiring companies are analysed to find out if shareholder wealth effects are present for these companies. To begin with, the cumulative abnormal returns of 116 acquiring companies have been calculated during the event study. Also in this case a regression was done to calculate the cumulative abnormal return of all these acquiring companies together. In table 4 on the next page the results of this regression are presented. The regression coefficient is 0.0096 which means that the cumulative abnormal returns of the acquiring companies increase with 0.96% as a consequence of the merger and
21
Table 4. Regression CARs acquirers
Regression of the cumulative abnormal returns of 116 acquiring companies.
Coefficient Robust Std. Err.
P-value
Constant 0.0096 0.005 0.042
The p-value given in the table is 0.042 which means the increase of the abnormal returns is statistically significant at the 10% and 5% levels. Again, the statistical hypothesis can be rejected, because the abnormal returns of the acquiring companies are not equal to 0. In contrast to the target companies, the shareholders of the acquiring companies were expected to have negative abnormal returns. Actually, the opposite is the case with a positive and significant increase of 0.96%.
The increase of the cumulative abnormal returns of 0.96% for the acquiring
companies seems economically not very relevant compared to the 21.72% increase for the target companies. Actually, the significantly positive result for acquiring companies is economically relevant, because it contradicts previous literature. Namely, there are a few studies which do report positive returns to bidders, but most of the existing literature find negative abnormal returns to shareholders of acquiring banks (or insignificant results). Possible reasons for the positive and significant abnormal returns for acquirers could be increased access to information for acquiring companies or better market estimations of the deal values of mergers and acquisitions than in previous decades (Houston et al., 2001). Furthermore, Alexandridis et al. (2017) show that after the recent financial crisis takeovers create more value for shareholders of acquiring companies than ever before. This could also be an explanation for the results in table 4.
When analysing the results of the regression model for the acquiring companies the correlation matrix and the descriptive statistics need to be discussed first. Table 8 in section Appendices shows the correlation coefficients between the variables used in the regression model for the acquiring companies. The highest value in this table is 0.582 between the variables Resolve and Crisis, so multicollinearity is also in this model not a problem. The descriptive statistics of the variables used in the regression model for the acquiring companies are presented in table 5 on the next page.
22
Table 5. Descriptive statistics acquirers
In this table a summary is given of the variables used in the regression model for the acquiring companies. The dependent variable CAR is the cumulative abnormal return. The lndealvalue is the natural logarithm of the transaction value of a takeover. The relativesize is the total assets of the target company divided by the total assets of the acquiring company. Cash is a dummy variable and equals 1 if only cash was used as the method of payment. Mix is a dummy variable and equals 1 if a mix of cash and shares was used as the method of payment. Crisis is a dummy variable and equals 1 if the takeover was announced between July 2007 and December 2010. Postcrisis is a dummy variable and equals 1 if the takeover was announced between January 2011 and June 2014. Resolve is a dummy variable and equals 1 if the takeover was used to resolve the financial problems of the target company. Tier 1 is the Tier 1 capital ratio of the target company. TBTF is a dummy variable and equals 1 if the total assets of both companies exceeds 100 billion dollars.
Variables Obs Mean Std. Dev. Min Max
CAR 116 0.010 0.051 -0.119 0.162 Lndealvalue 113 18.179 1.320 14.950 22.197 Relativesize 78 0.339 0.314 0.002 1.084 Cash 113 0.088 0.285 0 1 Mix 113 0.637 0.483 0 1 Crisis 116 0.095 0.294 0 1 Postcrisis 116 0.5 0.502 0 1 Resolve 116 0.043 0.204 0 1 Tier 1 69 12.06 3.253 6.72 21.73 TBTF 78 0.051 0.222 0 1
Again, when looking for any possible outliers in the dataset the method of applying winsorizing needs to be considered. The minimum and maximum values of the dummy variables are also in this case the correct values, namely 0 and 1 respectively. The deal values are already levelled by means of the natural logarithm. Again, for the other variables the threshold of three standard deviations will be applied. Similar to the descriptive statistics of the target variables there are no extreme values for the variables of the acquiring
companies. For this reason applying winsorizing is also not used in the case of the variables used in the regression model for the acquiring companies.
Next, the results of the regression model for the acquirers can be analysed. For this model the same regressions were done as with the model for the target companies. The results of these regressions are shown in table 6 on the next page. Regression 5, where the variable Postcrisis is omitted from the regression model and year fixed effects were added instead, has the highest explanatory power. Namely, the R-squared of this model is highest with a value of 0.3360.
23
Table 6. Regression results acquirers
In this table the results from the OLS regressions for the acquiring companies are shown. The dependent variable is the cumulative abnormal return. The lndealvalue is the natural logarithm of the transaction value of a takeover. The relativesize is the total assets of the target company divided by the total assets of the acquiring company. Cash is a dummy variable and equals 1 if only cash was used as the method of payment. Mix is a dummy variable and equals 1 if a mix of cash and shares was used as the method of payment. Crisis is a dummy variable and equals 1 if the takeover was announced between July 2007 and December 2010. Postcrisis is a dummy variable and equals 1 if the takeover was announced between January 2011 and June 2014. Resolve is a dummy variable and equals 1 if the takeover was used to resolve the financial problems of the target company. Tier 1 is the Tier 1 capital ratio of the target company. TBTF is a dummy variable and equals 1 if the total assets of both companies exceeds 100 billion dollars. The coefficients of each regression are shown below together with the standard errors in parentheses. All standard errors are robust standard errors. (1) (2) (3) (4) (5) Lndealvalue -0.0101** (0.005) -0.0095** (0.005) -0.0104** (0.005) -0.0091* (0.005) -0.0096* (0.005) Relativesize 0.0478*** (0.017) 0.0510*** (0.019) 0.0483*** (0.017) 0.0504*** (0.019) 0.0494** (0.020) Cash -0.0036 (0.015) -0.0068 (0.015) -0.0035 (0.015) -0.0069 (0.015) -0.0011 (0.015) Mix 0.0047 (0.010) 0.0066 (0.010) 0.0047 (0.010) 0.0067 (0.011) 0.0075 (0.009) Crisis -0.0147 (0.012) -0.0218* (0.013) -0.0181* (0.010) -0.0170 (0.016) -0.0025 (0.017) Postcrisis 0.0270** (0.011) 0.0291** (0.012) 0.0264** (0.010) 0.0300** (0.013) Resolve -0.0095 (0.018) -0.0098 (0.021) -0.0231 (0.024) Tier 1 -0.0017 (0.002) -0.0018 (0.002) -0.0023 (0.002) TBTF 0.0280** (0.012) 0.0223** (0.011) 0.0242** (0.010) 0.0258* (0.013) 0.0173 (0.015) Constant 0.1593* (0.092) 0.1650* (0.089) 0.1659* (0.086) 0.1579 (0.097) 0.1669 (0.105) Year fixed effects No No No No Yes Obs 75 65 75 65 65 R² 0.2967 0.2910 0.2955 0.2922 0.3360
*, ** and *** imply significance at the 10%, 5% and 1% levels, respectively.
In table 6 the results for the main explanatory variable Resolve are all negative. When companies acquire weak targets this will have a negative influence on the abnormal returns of the acquiring companies in most cases. However, none of the results for the variable
Resolve are significant. This implies no conclusions can be made about the effect of
takeovers executed to resolve failing banks on the abnormal returns of the acquiring
companies. Therefore, the null hypothesis that the recent financial crisis has no influence on abnormal returns around the announcement date cannot be rejected for acquiring banks in the United States.
24
In regression 2 and 3 where the main explanatory variable Resolve is excluded, a significant result can be found for the variable Crisis. The coefficient for this variable is in both regressions negative and significant at the 10% level. This indicates that the cumulative abnormal returns of the acquiring companies are negatively influenced by the mergers and acquisitions announced within the crisis period. The negative relationship between the cumulative abnormal returns of acquiring companies and the financial crisis agrees with the existing literature.
Furthermore, in table 6 the variables Lndealvalue, Relativesize, Postcrisis and TBTF have significant results. In all regressions the coefficient for the variable Lndealvalue is negative and statistically significant at the 5% or 10% level. Similar to the target companies, the size of the takeover deal and the shareholder returns for the acquiring companies are negatively related. Again, this is supported by most of the existing literature.
In contrast to the regression coefficient of the variable Relativesize in the target model this coefficient is significantly positive in the acquirer model. This positive relationship with the cumulative abnormal returns of the acquiring companies is significant at the 1% level in all regressions, except for regression 5 where the result is significant at the 5% level. The relevant literature is not unanimous on the effect of the relative size of merging firms on abnormal returns, but James and Wier (1987) and DeLong (2001) did find a positive
relationship as well.
As explained in section 4.1 the shareholder wealth effects increase during a period following an economic downturn. This means there should be a positive relationship
between the value gains of takeovers and the post-crisis period. The regression coefficients for the variable Postcrisis for acquiring companies are positive and significant at the 5% level. Therefore, these results correspond to the literature on this subject.
The coefficient for the variable TBTF is the last significant result in table 6. In regressions 1-4 these results are both positive and significant. This indicates that the advantages of becoming ‘Too-Big-To-Fail’ are present for shareholders of the acquiring banks which is in consensus with the literature. Namely, Kane (2000) and Brewer and Jagtiani (2013) present similar conclusions in their studies. Note that the result for the coefficient of the variable TBTF is insignificant in regression 5 which has the highest explanatory power.
25
5. Conclusion and discussion
5.1 Conclusion
First of all, to answer the research question of this paper an event study has been done. During this event study the value weighted market model was used to calculate the
cumulative abnormal returns for both target and acquiring companies. Afterwards, a cross-sectional regression model has been used to see if bank resolutions along with eight other variables influenced the cumulative abnormal returns of the merging firms.
Mergers and acquisitions do create value for shareholders of both target and acquiring banks in the United States between January 2004 and June 2014. Therefore, the statistical hypothesis that abnormal returns caused by takeovers will be equal to 0 can be rejected. Namely, target abnormal returns increased with 21.72% around the announcement date (significant at the 1% level) whereas the abnormal returns of the acquiring companies increased with 0.96% (with a significance level of 5%). In addition, the effect of takeovers completed to resolve failing banks on abnormal returns was studied in this paper. A significantly positive result (at the 1% level) was found for shareholders of these failing target banks, whereas the relationship between bank resolutions and the cumulative
abnormal returns to acquiring banks gives insignificant results. This means the answer to the research question of this paper is that the recent financial crisis does have an influence on abnormal returns around the announcement date of mergers and acquisitions in the banking industry in the United States, but only for the target companies.
Furthermore, looking at the regression results of the regression with the highest explanatory power the conclusion can be made that the deal value, the relative size of target to bidding firm and banks becoming ‘Too-Big-To-Fail’ as a consequence of a takeover also affect the abnormal returns of target shareholders. The abnormal returns for shareholders of acquiring banks are only influenced by the deal value and the relative size of merging firms in the regression with the highest R-squared.
5.2 Discussion
The significantly positive abnormal returns for the shareholders of acquiring companies are not anticipated by the existing literature. There are various possible reasons to explain the unanticipated results of this research. For example the positive abnormal returns for
26
acquiring banks could be explained by the post-crisis period which has a significantly positive effect on the cumulative abnormal returns. In addition, Alexandridis et al. (2017) infer that in the period following the recent financial crisis the shareholder wealth effects for acquiring companies were higher than ever before. This is an interesting topic for further research. Namely, which factors cause the abnormal returns of acquiring banks to increase after the most recent economic turmoil?
The positive result for the influence of the mergers and acquisitions announced during the financial crisis on the abnormal returns of target companies is more difficult to explain. It could be due to the financial constraints and shortage of capital during a recession that we only observe the ‘good’ deals. During an economic downturn takeover activity is lower and therefore the mergers and acquisitions which are completed may be a signal of well-performing acquiring banks. These signals can then be reflected in the abnormal returns of the target companies around the announcement date of the takeover, but further
research on this subject is needed as well. Note that the main explanatory variable Resolve was omitted in the only regression where this result was significant.
27
Reference list
Alexandridis, G., Antypas, N., & Travlos, N. (2017). Value creation from M&As: New evidence. Journal of Corporate Finance, 45, 632-650.
Anderson, C., Becher, D., & Campbell, T. (2004). Bank mergers, the market for bank CEOs, and managerial incentives. Journal of Financial Intermediation, 13(1), 6-27.
Becher, D. (2000). The valuation effects of bank mergers. Journal of Corporate Finance, 6(2), 189-214.
Beltratti, A., & Paladino, G. (2013). Is M&A different during a crisis? Evidence from the European banking sector. Journal of Banking & Finance, 37(12), 5394-5405.
Betton, S., Eckbo, B., & Thorburn, K. (2008). Corporate Takeovers. In B. Eckbo, Handbook of
Empirical Corporate Finance (Vol. 2, pp. 288-429). Elsevier/North-Holland.
Brewer, E., & Jagtiani, J. (2013). How Much Did Banks Pay to Become Too-Big-To-Fail and to Become Systemically Important? Journal of Financial Services Research, 43(1), 1-35. Cable, J., & Holland, K. (1999). Modelling normal returns in event studies: a model-selection
approach and pilot study. The European Journal of Finance, 5(4), 331-341.
Campa, J., & Hernando, I. (2006). M&As performance in the European financial industry.
Journal of Banking & Finance, 30(12), 3367-3392.
Crouzille, C., Lepetit, L., & Bautista, C. (2008). How did the Asian stock markets react to bank mergers after the 1997 financial crisis? Pacific Economic Review, 13(2), 171-182. DeLong, G. (2001). Stockholder gains from focusing versus diversifying bank mergers. Journal
of Financial Economics, 59(2), 221-252.
Desai, A., & Stover, R. (1985). Bank holding company acquisitions, stockholder returns, and regulatory uncertainty. The Journal of Financial Research, 8(2), 145-156.
DeYoung, R., Evanoff, D., & Molyneux, P. (2009). Mergers and Acquisitions of Financial Institutions: A Review of the Post-2000 Literature. Journal of Financial Services
Research, 36(2), 87-110.
Diamond, D., & Rajan, R. (2009). The Credit Crisis: Conjectures about Causes and Remedies.
American Economic Review, 99(2), 606-610.
Fahlenbrach, R., Prilmeier, R., & Stulz, R. (2012). This Time Is the Same: Using Bank
Performance in 1998 to Explain Bank Performance during the Recent Financial Crisis.
The Journal of Finance, 67(6), 2139-2185.
Frame, W., & Lastrapes, W. (1998). Abnormal Returns in the Acquisition Market: The Case of Bank Holding Companies, 1990–1993. Journal of Financial Services Research, 14(2), 145–163.
Houston, J., & Ryngaert, M. (1994). The overall gains from large bank mergers. Journal of
Banking and Finance, 18(6), 1155-1176.
Houston, J., James, C., & Ryngaert, M. (2001). Where do merger gains come from? Bank mergers from the perspective of insiders and outsiders. Journal of Financial
Economics, 60(2/3), 285-331.
Ismail, A. (2011). Does the Management's Forecast of Merger Synergies Explain the Premium Paid, the Method of Payment, and Merger Motives? Financial Management, 40(4), 879-910.
28
James, C., & Wier, P. (1987). Returns to Acquirers and Competition in the Acquisition Market: The Case of Banking. Journal of Political Economy, 95(2), 355-370.
Jensen, M. (1986). Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers. The
American Economic Review, 76(2), 323-329.
Kane, E. (2000). Incentives for Banking Megamergers: What Motives Might Regulators Infer from Event-Study Evidence? Journal of Money, Credit and Banking, 32(3), 671-701. Kolaric, S., & Schiereck, D. (2014). Performance of bank mergers and acquisitions: a review of
the recent empirical evidence. Management Review Quarterly, 64(1), 39-71.
Loderer, C., & Martin, K. (1990). Corporate Acquisitions by Listed Firms: The Experience of a Comprehensive Sample. Financial Management, 19(4), 17-33.
Martynova, M., & Renneboog, L. (2008). A century of corporate takeovers: What have we learned and where do we stand? Journal of Banking & Finance, 32(10), 2148-2177. Moeller, S., Schlingemann, F., & Stulz, R. (2004). Firm size and the gains from acquisitions.
Journal of Financial Economics, 73(2), 201-228.
Neely, W. (1987). Banking Acquisitions: Acquirer and Target Shareholder Returns. Financial
Management, 16(4), 66-74.
Pilloff, S. (1996). Performance Changes and Shareholder Wealth Creation Associated with Mergers of Publicly Traded Banking Institutions. Journal of Money, Credit and
Banking, 28(3), 294-310.
Roll, R. (1986). The Hubris Hypothesis of Corporate Takeovers. The Journal of Business, 59(2), 197-216.
Schwert, G. (1996). Markup pricing in mergers and acquisitions. Journal of Financial
Economics, 41(2), 153-192.
Shleifer, A., & Vishny, R. (1989). Management entrenchment: The case of manager-specific investments. Journal of Financial Economics, 25(1), 123-139.
White, P., & Yorulmazer, T. (2014). Bank Resolution Concepts, Tradeoffs, and Changes in Practices. Economic Policy Review, 153-173.
29
Appendices
Table 7. Correlation coefficients targets
In this table the correlation between the variables used in the regression model for the target companies is given. The dependent variable CAR is the cumulative abnormal return. Lndeal is the natural logarithm of the transaction value of a takeover. Relsiz is the total assets of the target company divided by the total assets of the acquiring company. Cash is a dummy variable and equals 1 if only cash was used as the method of payment. Mix is a dummy variable and equals 1 if a mix of cash and shares was used as the method of payment. Crisis is a dummy variable and equals 1 if the takeover was announced between July 2007 and December 2010. Postcr is a dummy variable and equals 1 if the takeover was announced between January 2011 and June 2014. Resol is a dummy variable and equals 1 if the takeover was used to resolve the financial problems of the target company. Tier 1 is the Tier 1 capital ratio of the target company. TBTF is a dummy variable and equals 1 if the total assets of both companies exceeds 100 billion dollars.
Observations = 63
CAR Lndeal Relsiz Cash Mix Crisis Postcr Resol Tier 1 TBTF
CAR 1 Lndeal -0.223 1 Relsiz -0.438 0.145 1 Cash 0.042 -0.103 -0.188 1 Mix 0.112 0.187 -0.126 -0.339 1 Crisis 0.014 -0.059 -0.056 -0.104 0.102 1 Postcr 0.164 -0.145 0.196 -0.054 -0.110 -0.348 1 Resol 0.045 0.227 -0.081 -0.076 0.094 0.529 -0.126 1 Tier 1 0.164 -0.063 0.185 0.023 0.094 0.060 0.378 -0.103 1 TBTF -0.257 0.139 -0.136 0.164 -0.038 0.115 -0.126 0.466 -0.186 1
30
Table 8. Correlation coefficients acquirers
In this table the correlation between the variables used in the regression model for the acquiring companies is given. The dependent variable CAR is the cumulative abnormal return. Lndeal is the natural logarithm of the transaction value of a takeover. Relsiz is the total assets of the target company divided by the total assets of the acquiring company. Cash is a dummy variable and equals 1 if only cash was used as the method of payment. Mix is a dummy variable and equals 1 if a mix of cash and shares was used as the method of payment. Crisis is a dummy variable and equals 1 if the takeover was announced between July 2007 and December 2010. Postcr is a dummy variable and equals 1 if the takeover was announced between January 2011 and June 2014. Resol is a dummy variable and equals 1 if the takeover was used to resolve the financial problems of the target company. Tier 1 is the Tier 1 capital ratio of the target company. TBTF is a dummy variable and equals 1 if the total assets of both companies exceeds 100 billion dollars.
Observations = 65
CAR Lndeal Relsiz Cash Mix Crisis Postcr Resol Tier 1 TBTF
CAR 1 Lndeal -0.210 1 Relsiz 0.279 0.210 1 Cash -0.012 -0.288 -0.139 1 Mix -0.059 0.193 -0.175 -0.332 1 Crisis -0.234 0.015 -0.017 -0.092 0.063 1 Postcr 0.384 -0.178 0.109 0.027 -0.031 -0.355 1 Resol -0.165 0.234 -0.085 -0.074 0.094 0.582 -0.157 1 Tier 1 0.042 -0.019 0.141 0.136 0.107 -0.042 0.395 -0.121 1 TBTF -0.023 0.153 -0.140 0.166 -0.036 0.140 -0.157 0.467 -0.196 1
