Sources of Financing in Different Forms of Corporate Liquidity and the Performance of M&As

Sources of Financing in Different Forms of Corporate Liquidity and the Performance of M&As Zhenxu Tong * University of Exeter Jian Liu ** University of Exeter This draft: August 2016 Abstract We examine how the performance of M&As is affected by the source of financing between two forms of corporate liquidity: bank lines of credit and corporate cash holdings. We develop two hypotheses based on agency problem and asymmetric information. We find that both the announcement return and the change in operating performance are higher for the M&As entirely financed by bank lines of credit. We also find that an acquirer is more likely to use bank lines of credit as the source of financing when institutional ownership is higher. Moreover, we find that the M&As entirely financed by bank lines of credit are associated a lower level of acquisition premium. We conclude that the findings are consistent with the agency hypothesis. JEL Classification: G32; G34 Keywords: Sources of financing; Corporate liquidity; Mergers and acquisitions * Corresponding author. Address: Xfi Centre for Finance & Investment, School of Business, University of Exeter, Rennes Drive, Exeter EX4 4ST, United Kingdom. Telephone: +44 1392 723155. E-mail address: z.tong@exeter.ac.uk. ** Address: Xfi Centre for Finance & Investment, School of Business, University of Exeter, Rennes Drive, Exeter EX4 4ST, United Kingdom. Telephone: +44 7707 919864. E-mail address: jl565@exeter.ac.uk. 1

1. Introduction Bank lines of credit and corporate cash holdings are two important forms of corporate liquidity. Previous literature shows that firms hold a large fraction of corporate liquidity to total assets. For example, Bates, Kahle and Stulz (2009) find that the mean of the ratio of cash to total assets is 23.2% for US firms in 2006. Sufi (2009) finds that 85% of the firms in his sample have bank lines of credit, and that the mean of the ratio of total lines of credit to total assets is 16%. Lins, Servaes and Tufano (2010) conduct a survey on CFOs in 29 countries and find that bank lines of credit accounts for 15% of total assets. In our paper, we examine how the performance of M&As is affected by the sources of financing between corporate cash holdings and bank lines of credit. We develop two competing hypotheses based on agency problem and asymmetric information. First, since bank lines of credit are subject to the monitoring by banks, this can reduce the agency problem. We expect that the M&As financed by bank lines of credit will outperform the M&As financed by corporate cash holdings. Second, corporate cash holdings is an internal source of financing, while bank lines of credit is an external source of financing. Since the cost of financing increases with the degree of asymmetric information, we expect that the M&As financed by bank lines of credit will underperform the M&As financed by corporate cash holdings. We construct a sample of 723 M&As from 1985 and 2013. We find that both the announcement return and the change in operating performance are higher for the M&As entirely financed by bank lines of credit. We also find that an acquirer is more likely to use bank lines of credit as the source of financing when institutional ownership is higher. Moreover, we find that the M&As entirely financed by bank lines of credit are associated a lower level of acquisition premium. We further divide the sample into sub-groups based on corporate governance. We find that when bank lines of credit are used as source of financing, both the announcement return and the change in operating performance are higher for the sub-group of 2

firms with poorer corporate governance, and that the acquisition premium is lower for such a sub-group of firms. The results are consistent with the interpretation that the performance of M&As is higher when bank lines of credit are used as a source of financing. We conclude that this is consistent with the agency hypothesis. We contribute to the literature in the following two directions. First, we complement the literature on the agency perspective about bank lines of credit. Previous literature has focused on examining how the agency problem affects the level of bank lines of credit. For example, Yun (2009) finds that after a change in takeover legislation, firms increase the fraction of corporate liquidity held in the form of corporate cash holdings relative to bank lines of credit. Sufi (2009) finds that firms with low cash flow are less likely to obtain a line of credit, and argue that firms must maintain high cash flow to remain compliant with covenants associate with bank lines of credit. Our paper differs from the previous literature in that we examine the performance brought by bank lines of credit. To our knowledge, only Nini, Smith and Sufi (2012) examine a sample of lines of credit loans, and find that a firm s operating and stock price performance improve following a violation of covenant, which implies that banks play an important governance role and their actions benefit shareholders. We differ from Nini et al. (2012) in that we conduct the research in the setting of M&As. Second, our paper contributes to the limited existing literature on how the source of financing affects the performance of M&As. To our knowledge, only two papers in the literature have examined the relation between the source of financing and the performance of M&As. Bharadwaj and Shivdasani (2003) find that tender offered that are financed by banks have better performance. Our paper extends Bharadwaj and Shivdasani (2003) in that we conduct the analysis with general M&As instead of tender offers only. Martynova and Renneboog (2009) find that acquisitions financed with internally generated funds 3

underperform those financed with debts. Our paper extends Martynova and Renneboog (2009) in that we focus on bank lines of credit which is an important source of corporate liquidity instead of the debts in general. The paper is organized as follows. Section 2 develops the hypotheses. Section 3 describes the data and the variables. Section 4 presents the results. Section 5 concludes the paper. 2. Hypotheses We develop the hypotheses in this section. 2.1. Agency problem While managers can use corporate cash holding in a discretionary way, bank lines of credit are subject to the monitoring by banks. This difference between the two forms of corporate liquidity implies an important role of the agency problem in this research setting. In terms of corporate cash holdings, for example, Harford (1999) finds that cash-rich firms are more likely to make value-decreasing acquisitions. This is consistent with the free cash flow theory proposed by Jensen (1986). In term of bank lines of credit, for example, Sufi (2009) finds that firms must maintain high cash flow to remain compliant with covenants associate with bank lines of credit, and that banks restrict firm access to credit facilities in response to covenant violations. Yun (2009) finds that firms increases the fraction of corporate liquidity held in the form of corporate cash holdings relative to bank lines of credit when the threat of takeover weakens. Since bank lines of credit reduce the agency problem, we expect that the performance will be higher for the M&As financed by bank lines of credit. Therefore, we have the following hypothesis. Hypothesis 1: The M&As financed by bank lines of credit outperform the M&As financed by corporate cash holdings from the perspective of agency problem. 4

2.2. Information asymmetry Myers and Majluf (1984) propose a pecking order for the sources of financing based on asymmetric information. They argue that the internal source of financing, such as corporate cash holding, is a cheapest source of financing. The financing cost increases when firms use debts, and equity is a most costly source of financing. From the perspective of asymmetric information, it is cheaper for firms to finance M&As by using corporate cash holdings than bank lines of credit, because bank lines of credit are an external source of financing and are associated with a higher level of asymmetric information. We expect that the performance will be lower for the M&As financed by bank lines of credit because of a higher financing cost. Therefore, we have the following hypothesis. Hypothesis 2: The M&As financed by bank lines of credit underperform the M&As financed by corporate cash holdings from the perspective of asymmetric information. 3. Data and variables In this section, we describe the data and variables. 3.1. Data We get the data from the following sources. We collect financial data from Compustat database and stock return data from CRSP. We obtain the U.S. data on mergers and acquisitions from Thomson One Banker database. We manually collect the data of bank lines of credit from 10-K annual reports. We collect the institutional ownership data from Thomson Financial/Institutional database and insider ownership data from the proxy statements. In our sample, the acquirers are public firms because we use stock market data to calculate a measure of firm performance. The targets can be either public firms or private firms. Our sample period is from 1985 to 2013. We use the following screening procedures. We choose the M&As whose sources of financing are identified by Thomson One Banker database 5

as bank lines of credit, or corporate cash holdings, or a mix of the two sources. We exclude financial firms (SIC codes between 6000 and 6999). We exclude the M&As whose deal value is less than one million dollars. We also exclude the observations with incomplete data. After the screening procedures, we obtain a final sample of 723 M&A events. Among them, 271 M&As are entirely financed by corporate cash holdings. 308 M&As are entirely financed by bank lines of credit. 144 M&As are financed by a mix of bank lines of credit and corporate cash holdings. 3.2. Variables 3.2.1. Sources of financing The sources of financing for M&As are identified in Thomson One Banker database. For example, the database records an acquisition made by Actuant Corp with the announcement date on 3 March 2008. The source of financing is recorded as Line of Credit, and its description is The transaction was financed through Actuant Corp's revolving credit facility. For another example, the database records an acquisition made by Select Comfort Corp with the announcement date on 17 January 2013. The source of financing is recorded as Corporate Funds, and its description is The transaction was financed through Select Comfort Corps existing cash reserves. We construct two dummy variables to indicate the sources of financing for M&As. Bank Lines of Credit Dummy equals ones if an M&A is entirely financed by bank lines of credit, and equals zero otherwise. Mixed Sources Dummy equals one if an M&A is financed by mixed sources of bank lines of credit and corporate cash holdings, and equals zero otherwise. 3.2.2. Announcement return We use an acquirer s announcement return, which is calculated as the cumulative abnormal return over days ( 3, +3) around the announcement date, as a measure of the stock market performance of M&As. The cumulative abnormal return is calculated using the market model 6

with the CRSP equally weighted index as the market return. To estimate the market model, we use an acquirer s daily return and the return on the CRSP equally weighted index over days -200 to -20, where day 0 is the event date. 3.2.3. Change in operating performance We use the change in ROA as a measure of the operating performance of M&As. ROA is the ratio of EBIT to non-cash assets. We calculate the Change in ROA from year t-1 to year t+1. 3.2.4. Net change in operating performance We also use the net change in ROA as another measure of the operating performance of M&As. Net Change in ROA is the difference between an acquirer s change in ROA and its matched comparable firm s change in ROA from year t-1 to year t+1. We construct a sample of comparable firms with propensity score matching. We match each acquirer firm to a non-acquirer firm within the same industry based on 2-digit SIC code, requiring that the non-acquirer firm has a minimum difference in propensity score based on firm size, market-tobook ratio, cash flow, leverage, tangibility, capital expenditure, R&D, dividends and cash flow volatility. We provide the details about the propensity score matching in Appendix A. 3.2.5. Institutional ownership We use institutional ownership as a measure of corporate governance. Institutional Ownership is the ratio of shares owned by institutional investors to the total shares outstanding at the end of a quarter prior to the announcement. Block is a dummy variable that equals one if there exists a block institutional ownership which exceeds 5% of the total shares outstanding, and equals zero otherwise. 3.2.6. Deviation from optimal insider ownership We use another measure of corporate governance based on insider ownership. Since the endogeneity problem has been discussed extensively in the literature about managerial 7

ownership, we follow Tong (2008) and construct a measure based on the deviation from optimal insider ownership. We first run a benchmark regression to get the determinants of insider ownership. Then we define the variable Deviation from Optimal Insider Ownership as the absolute value of the residuals in the benchmark regression. A higher level of Deviation from Optimal Insider Ownership indicates a lower level of corporate governance. We provide the details about the benchmark regression in Appendix B. 3.2.7. Control variables We use the following control variables. Relative Value is the ratio of deal value to the sum of the acquirer s market value of equity and deal value. Unused Lines of Credit is the ratio of unused lines of credit to assets (e.g., Sufi, 2009). Cash is the ratio of cash and marketable securities to non-cash assets, where non-cash assets are total assets less corporate cash holdings. Size is the logarithm of non-cash assets. Cash Flow is the ratio of income before extraordinary items to non-cash assets. Market-to-book ratio is defined as the market value of equity plus non-cash assets minus book value of equity, divided by non-cash assets. Leverage is the ratio of long-term debts to non-cash assets. Tangibility is the ratio of plant, property and equipment to non-cash assets. Capital Expenditure is the ratio of capital expenditures to non-cash assets. R&D is the ratio of research and development expenses to non-cash assets. Dividends is the ratio of dividends to non-cash assets. Cash Flow Volatility is the standard deviation of Cash Flow in the prior 5 years. Table 1 reports univariate statistics. 4. Results We report the results in this section. First, we report the univariate analysis on the announcement return. Next, we examine how the sources of financing between bank lines of credit and corporate cash holdings affect the announcement return and the change in operating performance. Then we demonstrate the likelihood of an acquirer s choice between bank lines 8

of credit and corporate cash holdings as the source of financing. Moreover, we report the effect of sources of financing on acquisition premium. Then we report the results of the sub-group analysis and the robustness checks. 4.1. Univariate analysis Table 2 shows the univariate analysis on acquirers announcement return across three groups. CAR (-3, +3) indicates the cumulative abnormal return over days ( 3, +3) around the announcement date. The first row shows that the mean of announcement return for the M&As entirely financed by bank lines of credit (corporate cash holdings) is 0.031 (0.002). A t-test shows that the difference is significant between the two groups. The second row shows that the mean of announcement return for the M&As financed by a mixed source of financing is not significantly different from the mean of announcement return for the group of acquirers entirely financed by corporate cash holdings. The third row compares the M&As entirely financed by bank lines of credit with the M&As financed by a mixed source of financing. The difference of the mean of announcement return between the two groups is significant. The results support the interpretation that the M&As entirely financed by bank lines of credit have the highest announcement return. This is consistent with the agency hypothesis. 4.2. Regression Table 3 reports the regression. The dependent variable is CAR (-3, +3). The coefficient of Bank Lines of Credit Dummy is 0.020 (p-value = 0.02). It implies that on average the M&As entirely financed by bank lines of credit has an additional 2% announcement return than the M&As entirely financed by corporate cash holdings. We also find the coefficient of Mixed Sources Dummy is 0.004 (p-value = 0.69). The results are consistent with the interpretation that the M&As entirely financed by bank lines of credit have the best stock market performance. This is consistent with the agency hypothesis. 9

4.3. Change in operating performance Table 4 reports the results on the change in operating performance. The dependent variable in Column 1 is the Change in ROA from year t-1 to year t+1. 1 The coefficient of Bank Lines of Credit dummy is 0.012 (p-value = 0.08), indicating that the M&As entirely financed by bank lines of credit have a higher change in operating performance than the M&As entirely financed by corporate cash holdings. The coefficient of Mixed Sources Dummy is 0.014 (p-value = 0.05). The dependent variable of the second column is the Net Change in ROA, which is the difference between the change in ROA of the event firm and the change in ROA of the comparable firm. The coefficient of Bank Lines of Credit Dummy is 0.025 (p-value = 0.05). The coefficient of Mixed Sources Dummy is 0.012 (p-value = 0.40). The results in Table 4 are consistent with the interpretation that the M&As entirely financed by bank lines of credit have the best operating performance. This is consistent with the agency hypothesis. 4.4. The choice of the source of financing We use a multinomial logistic regression to estimate the likelihood of an acquirer s choice between bank lines of credit and corporate cash holdings as the source of financing. We expect that firms with better corporate governance are more likely to choose bank lines of credit as the source of financing, because they are more likely to meet the monitoring requirements of the banks. In the multinomial logistic regression, the dependent variable is categorical. It equals 0 if an M&A is entirely financed by corporate cash holdings, equals 1 if an M&A is financed by a mixed sources of financing, and equals 2 if an M&A is entirely financed by bank lines of credit. 1 Since the dependent variable requires the availability of the data from year t-1 to year t+1, the sample size is reduced to 667 M&As in this table. 10

We use the variables at the fiscal year end of the year t-1 as the independent variables. They include corporate cash holdings and unused lines of credit 2 as the sources of financing available to a firm. They also include the variables on corporate governance, such as institutional ownership and the deviation from optimal insider ownership. We also add other control variables such as size, cash flow and so on. We report the results in Table 5. 3 Column 1 shows the likelihood of an acquirer s choice of the source of financing between bank lines of credit and corporate cash holdings. The coefficient of Cash is -11.764 (p-value = 0.01), suggesting that firms with higher level of cash are less likely to use bank lines of credit as the source of financing for M&As. The coefficient of Unused Lines of Credit is 6.230 (p-value = 0.01), suggesting that firms with higher level of unused lines of credit are more likely to use bank lines of credit as the source of financing for M&As. We also examine whether corporate governance affects an acquirer s choice of the source of financing. Column 1 shows that the coefficient of Institutional Ownership is 1.086 (p-value = 0.06). It implies that a firm with higher institutional ownership is more likely to choose bank lines of credit than corporate cash holdings as the source of financing. Moreover, we find similar results in Column 2 that an acquirer with higher institutional ownership is more likely to choose a mixed source of financing than corporate cash holdings as the only source of financing. We further examine the impact of corporate governance by adding Deviations from Optimal Insider Ownership in the regressions in Column 3 and Column 4. The variable 2 We get similar result when we use total lines of credit in the regressions. 3 Since our data of unused lines of credit starts from 1995 when the 10-K statements are available online, the sample size is reduced to 504 M&As in Column 1 and Column 2 in this table. Among them, 191 M&As are entirely financed by corporate cash holdings, 93 M&As are financed by a mix of bank lines of credit and corporate cash holdings, and 220 M&As are entirely financed by bank lines of credit. Similarly, since our data of insider ownership also starts from 1995 when the proxy statements are available online, the sample size is reduced to 424 M&As in Column 3 and Column 4 in this table. Among them, 167 M&As are entirely financed by corporate cash holdings, 82 M&As are financed by a mix of bank lines of credit and corporate cash holdings, and 175 M&As are entirely financed by bank lines of credit. 11

Institutional Ownership is still significant in the two columns, though the coefficient of variable Deviations from Optimal Insider Ownership is not significant. Therefore, some evidence in Table 5 supports the interpretation that firms with better corporate governance are more likely to choose bank lines of credit as a source of financing for M&As. This is consistent with the agency hypothesis. 4.5. Acquisition premium We examine the impact of the source of financing on acquisition premium. From the agency perspective, a manager is more likely to overpay for the M&As if there is more agency problem. Therefore, we expect that the acquisition premium is lower for the M&As financed by bank lines of credit than the M&As financed by corporate cash holdings. We collect the data of acquisition premium from Thomson One Banker database. There are three measures of acquisition premium. They are the ratios of the offer price to the target share price one day, or one week, or four weeks prior to the announcement of M&As. Table 6 shows the results. 4 Column 1 shows that the coefficient of Bank Lines of Credit Dummy is -0.053 (p-value = 0.08). It implies that M&As entirely financed by bank lines of credit are associated with lower acquisition premium. The coefficient of Mixed Sources Dummy is -0.029 (p-value = 0.26). We find a similar pattern in Column 2 and Column 3 when we measure the acquisition premium with different time intervals. Therefore, the results in Table 6 support the interpretation that a manager pay less acquisition premium when the source of financing is bank lines of credit. This is consistent with the agency hypothesis. 4.6. Sub-group analysis on corporate governance We divide the sample into sub-groups and further examine the impact of corporate governance. 4 There are 169 M&As in our sample whose data of acquisition premium are available in Thomson One Bank database. 12

Table 7 reports the results when we divide the sample into sub-groups based on whether there is a block holder of institutional ownership. Panel A shows the results for the announcement return. Column 1 shows that the coefficient of Bank Lines of Credit Dummy is 0.024 (p-value = 0.08) when there is not a block holder, and Column 2 shows that the coefficient is 0.009 (p-value = 0.37) when there is a block holder. Panel B of Table 7 shows the results about the Net Change in ROA. We find a similar pattern. Column 1 shows that the coefficient of Bank Lines of Credit Dummy is 0.060 (p-value = 0.02) when there is not a block holder, and Column 2 shows that the coefficient is 0.012 (p-value = 0.44) when there is a block holder. Table 8 reports the results when we divide the sample into sub-groups based on the deviation from optimal insider ownership 5. A higher (lower) level of deviation from optimal insider ownership indicate a lower (higher) level of corporate governance. Panel A shows the results for the announcement return. Column 1 shows that the coefficient of Bank Lines of Credit Dummy is 0.035 (p-value = 0.03) when there is a higher level of deviation from optimal insider ownership, and Column 2 shows that the coefficient is 0.008 (p-value = 0.63) when there is a lower level of deviation from optimal insider ownership. Panel B of Table 8 shows the results about the Net Change in ROA. We find a similar pattern. Column 1 shows that the coefficient of Bank Lines of Credit Dummy is 0.047 (p-value = 0.07) when there is a higher level of deviation from optimal insider ownership, and Column 2 shows that the coefficient is 0.012 (p-value = 0.61) when there is a lower level of deviation from optimal insider ownership. The results in Table 7 and Table 8 imply that as a source of financing, bank lines of credit improve the corporate governance in poorly governed firms due to the monitoring by banks. This results in a significantly positive relation between bank lines of credit as a source of 5 Since our data of insider ownership starts from 1995 when the proxy statements are available online, the sample size is reduced to 444 M&As in this table. 13

financing and stock market performance as well as the operating performance. This effect is insignificant for well-governed firms, because these firms already have good governance and there is less room for improvements in corporate governance for these firms whey they use bank lines of credit as a source of financing. This is consistent with the agency hypothesis. Table 9 reports the results on acquisition premium when we divide the sample into subgroups based on whether there is a block holder of institutional ownership. 6 Column 1 shows that the coefficient of Bank Lines of Credit Dummy is -0.120 (p-value = 0.08) when there is not a block holder, and Column 2 shows that the coefficient is 0.028 (p-value = 0.54) when there is a block holder. We find a similar pattern in Column 3 to Column 6 when we use the acquisition premium with different time intervals. We interpret the results in Table 9 in a similar way. The results in Table 9 imply that as a source of financing, bank lines of credit improve the corporate governance in poorly governed firms due to the monitoring by banks. This results in a significantly negative relation between bank lines of credit as a source of financing and the acquisition premium. This effect is insignificant for well-governed firms, because these firms already have good governance and they tend not to overpay for the M&As no matter when bank lines of credit or corporate cash holdings are used as a source of financing. This is consistent with the agency hypothesis. 4.7. Heckman two-stage estimation Since firms self-select to undertake M&As, we use Heckman two-stage estimation to control for the self-selection. In the first stage, we use the multinomial logistic model reported in Table 5, and calculate the Inverse Mills Ratio based on the estimates in the models. In the second stage, we include the Inverse Mills Ratio in the regressions as an additional control variable. 6 The number of observations is 81 in Column 1 and 88 in Column 2. Similar pattern exists for other columns. Since our data of insider ownership starts from 1995 when the proxy statements are available online, we do not conduct the sub-group analysis on acquisition premium based on the deviation from optimal insider ownership because of even a fewer number of observations available. 14

Table 10 reports the second stage of the Heckman two-stage estimation. Panel A shows the results about the announcement return. Column 1 shows that the coefficient of Bank Lines of Credit Dummy is 0.035 (p-value = 0.01). Column 2 shows that the coefficient of Mixed Sources Dummy is -0.017 (p-value = 0.23). Panel B shows the results about the Net Change in ROA. Column 1 shows that the coefficient of Bank Lines of Credit Dummy is 0.047 (p-value = 0.02). Column 2 shows that the coefficient of Mixed Sources Dummy is 0.008 (p-value = 0.76). The results are consistent with the findings in Table 3 and Table 4 in that the M&As entirely financed by bank lines of credit have both the best stock market performance and the best operating performance. Therefore, we find similar results after controlling for the self-selection. 5. Conclusion We examine how the performance of M&As is affected by bank lines of credit and corporate cash holdings as two sources of financing. We develop two hypotheses based on agency problem and asymmetric information. We find that both the announcement return and the change in operating performance are higher for the M&As entirely financed by bank lines of credit. We examine an acquirer s choice of the source of financing, and find that an acquirer is more likely to use bank lines of credit as the source of financing when institutional ownership is higher. We also find that the M&As entirely financed by bank lines of credit are associated with a lower level of acquisition premium. We conduct sub-group analysis to further examine the impact of corporate governance, and find consistent results. The results are consistent with the interpretation that the performance of M&As is higher when bank lines of credit are used as a source of financing. We conclude that this is consistent with the agency hypothesis. 15

References Bates, T., Kahle, K., Stulz, R., 2009. Why do US firms hold so much more cash than they used to? Journal of Finance 64, 1985-2021. Bharadwaj, A., Shivdasani, A., 2003. Valuation effects of bank financing in acquisitions. Journal of Financial Economics 67, 113-148. Harford, J., 1999. Corporate cash reserves and acquisitions. Journal of Finance 54, 1969-1997. Heckman, J., 1979. Sample selection bias as a specification error. Econometrica: Journal of the econometric society 153-161. Jensen, M., 1986. Agency cost of free cash flow, corporate finance, and takeovers. Corporate Finance, and Takeovers. American Economic Review 76. Lins, K., Servaes, H., Tufano, P., 2010. What drives corporate liquidity? An international survey of cash holdings and lines of credit. Journal of Financial Economics 98, 160-176. Martynova, M., Renneboog, L., 2009. What determines the financing decision in corporate takeovers: Cost of capital, agency problems, or the means of payment? Journal of Corporate Finance 15, 290-315. Myers, S., Majluf, N., 1984. Corporate financing and investment decisions when firms have information that investors do not have. Journal of financial economics 13, 187-221. Nini, G., Smith, D. C., Sufi, A., 2012. Creditor control rights, corporate governance, and firm value. Review of Financial Studies 25, 1713-1761. Opler, T., Pinkowitz, L., Stulz, R., Williamson, R., 1999. The determinants and implications of corporate cash holdings. Journal of financial economics 52, 3-46. Sufi, A., 2009. Bank lines of credit in corporate finance: An empirical analysis. Review of Financial Studies 22, 1057-1088. Tong, Z., 2008. Deviations from optimal CEO ownership and firm value. Journal of Banking and Finance 32, 2462-2470. Yun, H., 2009. The choice of corporate liquidity and corporate governance. Review of Financial Studies 22, 1447-1475. 16

Appendix A. Propensity Score Matching The table shows the results for the propensity score matching. We use a logistic regression. There are 723 observations of acquirers and 14322 observations of non-acquirers from the Compustat database. We match each acquirer firm with a non-acquirer firm by propensity score matching. We define non-acquirers as the firms that do not have any M&As in the same fiscal year as the acquirers. Matched firms are selected based on nearest propensity score and the same industry defined by the 2-digit SIC code. The dependent variable is a dummy variable that equals one if a firm is an acquirer, and equals zero otherwise. Size is the logarithm of non-cash assets. M/B is non-cash assets minus value of equity plus market value of equity, divided by non-cash assets. Cash Flow is income before extraordinary items divided by non-cash assets. Leverage is the ratio of long-term debt to non-cash assets. Tangibility is the ratio of plant, property and equipment to noncash assets. Capital Expenditure is the ratio of capital expenditure to non-cash assets. R&D is the ratio of R&D to non-cash assets. Dividends is the ratio of dividends to non-cash assets. Cash Flow Volatility is the standard deviation of cash flow to non-cash assets in the prior 5 years. The p-value is noted in the parentheses. Acquirer=1, Non-acquirer=0 Intercept 10.676 (0.01) Size t-1-0.457 (0.01) M/B t-1 0.076 (0.01) Cash Flow t-1-4.334 (0.01) Leverage t-1-0.505 (0.03) Tangibility t-1 0.611 (0.01) Capital Expenditure t-1 1.371 (0.01) R&D t-1 0.165 (0.79) Dividends t-1 7.292 (0.01) Cash Flow Volatility t-1 1.707 (0.01) Number of Observations 15045 Pseudo R-Square 0.09 17

Appendix B. Determinants of Insider Ownership The table shows the results for the determinants of insider ownership. Following Tong (2008), we run a benchmark regression to get the determinants of insider ownership. We define Deviation from Optimal Insider Ownership as the absolute value of the residual. A higher (lower) level of Deviation from Optimal Insider Ownership indicates a lower (higher) level of corporate governance. The dependent variable is Insider Ownership, which is the ownership of all executive officers and directors. Size is the logarithm of non-cash assets. Size is the natural logarithm of the book value of assets. Cash Flow is the ratio of income before extraordinary items to assets. Sales Growth is the growth rate of sales over the previous year. Plant, Property and Equipment is the ratio of PPE to assets. R&D and Advertising is the ratio of the sum of research and development expenses and advertising expenses divided by total expenses. Earnings Volatility is the standard deviation of the ratio of income before extraordinary items to assets in the prior three years. Sales Volatility is the standard deviation of the natural logarithm of sales in the prior three years. Year Dummy Variables are the dummy variables for the years in the sample and not reported in the table. Industry Dummy Variables are the dummy variables for the industries defined by two-digit SIC codes and not reported in the table. The p-value is noted in the parentheses. Insider Ownership Intercept 3.881 (0.01) Size -0.320 (0.01) Size Squared 0.007 (0.01) Cash Flow -0.062 (0.66) Cash Flow Squared 0.097 (0.83) Sales Growth 0.036 (0.18) Plant, Property and Equipment -0.055 (0.28) R&D and Advertising -0.094 (0.15) Earnings Volatility -0.185 (0.30) Sales Volatility -0.032 (0.43) Year Dummy Variables Yes Industry Dummy Variables Yes Number of Observations 444 Adjusted R-square 0.24 18

Table 1: Summary Statistics The table reports summary statistics. We use a sample of 723 U.S. M&As from 1985 to 2013. CAR (-3, +3) is the cumulative abnormal return over days (-3, +3) around the announcement date (see text for details). Relative Value is the ratio of deal value to the sum of the acquirer s market value of equity and deal value. Total Lines of Credit is the ratio of bank lines of credit to assets. Unused Lines of Credit is the ratio of unused lines of credit to assets. Cash is the ratio of cash and marketable securities to non-cash assets. ROA is the ratio of EBIT to non-cash assets. ROA is the change in ROA from year t-1 to t+1. Net ROA is the difference between an acquirer s change in ROA and its matched comparable firm s change in ROA from year t-1 to t+1. The comparable firms are matched by propensity score matching based on size, cash flow, market-to-book ratio, leverage, tangibility, capital expenditure, R&D, dividends and cash flow volatility (see text for details). Premium 1 is ratio of the offer price to the target share price one day prior to the acquisition announcement. Premium 2 is ratio of the offer price to the target share price one week prior to the acquisition announcement. Premium 3 is ratio of the offer price to the target share price four weeks prior to the acquisition announcement. Institutional Ownership is the ratio of shares owned by institutional investors to the total shares outstanding at the end of a quarter prior to the announcement. Block is a dummy variable that equals one if there exists a block institutional ownership which exceeds 5% of the total shares outstanding, and equals zero otherwise. Insider Ownership is the ownership of all executive officers and directors. Size is the logarithm of non-cash assets. Cash Flow is income before extraordinary items divided by non-cash assets. M/B is market value of equity plus non-cash assets minus book value of equity, divided by non-cash assets. Leverage is the ratio of long-term debt to non-cash assets. Tangibility is the ratio of plant, property and equipment to non-cash assets. Capital Expenditure is the ratio of capital expenditure to non-cash assets. R&D is the ratio of R&D to non-cash assets. Dividends is the ratio of dividends to non-cash assets. Cash Flow Volatility is the standard deviation of cash flow to non-cash assets in the prior 5 years. Variable Mean 25 th Percentile Median 75 th Percentile Std Dev CAR (-3,+3) 0.0161-0.0271 0.0091 0.0571 0.0827 Relative Value 0.1398 0.0371 0.0911 0.1915 0.1450 Total Lines of Credit 0.2080 0.0782 0.1693 0.3209 0.1597 Unused Lines of Credit 0.1437 0.0622 0.1233 0.2126 0.1076 Cash 0.1859 0.0213 0.0692 0.2422 0.2554 ROA 0.1241 0.0701 0.1124 0.1680 0.0836 ROA -0.0285-0.0625-0.0144 0.0166 0.0790 Net ROA 0.0103-0.0601 0.0115 0.0699 0.1285 Premium1 0.1342 0.0508 0.0976 0.1750 0.1241 Premium2 0.1278 0.0413 0.0949 0.1700 0.1200 Premium3 0.1206 0.0416 0.0850 0.1763 0.1104 Institutional Ownership 0.4733 0.1299 0.5184 0.7736 0.3376 Block 0.6418 0.0000 1.0000 1.0000 0.4798 Insider Ownership 0.1550 0.0385 0.0895 0.2105 0.1715 Size 19.9808 18.7648 19.9660 21.1084 1.8806 Cash Flow 0.0679 0.0331 0.0618 0.1042 0.0753 M/B 2.1360 1.2670 1.6582 2.3138 1.3702 Leverage 0.2318 0.0607 0.2069 0.3474 0.1957 Tangibility 0.3607 0.1377 0.2823 0.5460 0.2682 Capital Expenditure 0.0834 0.0263 0.0484 0.0979 0.0916 R&D 0.0256 0.0000 0.0000 0.0264 0.0544 Dividends 0.0125 0.0000 0.0000 0.0174 0.0195 Cash Flow Volatility 0.1090 0.0206 0.0426 0.0988 0.1808 19

Table 2: Univariate Analysis on Announcement Return The table reports the univariate analysis on announcement return. CAR (-3, +3) is the cumulative abnormal return over days (-3, +3) around the announcement date (see text for details). Corporate Cash Holdings is the group of M&As entirely financed by corporate cash holdings. Bank Lines of Credit is the group of M&As entirely financed by bank lines of credit. Mixed Sources is the group of M&As financed by a mix of bank lines of credit and corporate cash holdings. We report the mean test and the median test in the table. Bank Lines of Credit Corporate Cash Holdings Difference Mean Median Mean Median Mean p-value Median p-value CAR (-3,+3) 0.031 0.022 0.002 0.002 0.029 0.01 0.020 0.01 Mixed Sources Corporate Cash Holdings Difference Mean Median Mean Median Mean p-value Median p-value CAR (-3,+3) 0.011 0.008 0.002 0.002 0.009 0.28 0.006 0.45 Bank Lines of Credit Mixed Sources Difference Mean Median Mean Median Mean p-value Median p-value CAR (-3,+3) 0.031 0.022 0.011 0.008 0.020 0.02 0.014 0.02 20

Table 3: Sources of Financing and Announcement Return The table shows the relationship between sources of financing and announcement return. The dependent variable is CAR (-3, +3), which is the cumulative abnormal return over days (-3, +3) around the announcement date (see text for details). Bank Lines of Credit Dummy equals one if an M&A is entirely financed by bank lines of credit, and equals zero otherwise. Mixed Sources Dummy equals one if an M&A is financed by a mix of bank lines of credit and corporate cash holdings. Relative Value is the ratio of deal value to the sum of the acquirer s market value of equity and deal value. Size is the logarithm of non-cash assets. Cash Flow is income before extraordinary items divided by non-cash assets. M/B is market value of equity plus non-cash assets minus book value of equity, divided by non-cash assets. Leverage is the ratio of long-term debt to non-cash assets. Tangibility is the ratio of plant, property and equipment to non-cash assets. Capital Expenditure is the ratio of capital expenditure to noncash assets. R&D is the ratio of R&D to non-cash assets. Dividends is the ratio of dividends to non-cash assets. Cash Flow Volatility is the standard deviation of cash flow to non-cash assets in the prior 5 years. Year Dummy Variables are the dummy variables for the years in the sample and not reported in the table. Industry Dummy Variables are the dummy variables for the industries defined by two-digit SIC codes and not reported in the table. The p-value is noted in the parentheses. CAR (-3,+3) Intercept 0.168 (0.01) Bank Lines of Credit Dummy 0.020 (0.02) Mixed Sources Dummy 0.004 (0.69) Relative Value 0.049 (0.07) Size t-1-0.005 (0.04) Cash Flow t-1 0.068 (0.17) M/B t-1-0.004 (0.20) Leverage t-1 0.002 (0.93) Tangibility t-1-0.010 (0.67) Capital Expenditure t-1-0.037 (0.49) R&D t-1-0.077 (0.29) Dividends t-1-0.320 (0.08) Cash Flow Volatility t-1-0.019 (0.33) Year Dummy Variables Yes Industry Dummy Variables Yes Number of Observations 723 Adjusted R-square 0.09 21

Table 4: Sources of Financing and Change in Operating Performance The table shows the relationship between sources of financing and the change in operating performance. ROA is the ratio of EBIT to non-cash assets. ROA is the change in ROA from year t-1 to t+1. Net ROA is the difference between an acquirer s change in ROA and its matched comparable firm s change in ROA from year t-1 to t+1. The comparable firms are matched by propensity score matching based on size, cash flow, market-to-book ratio, leverage, tangibility, capital expenditure, R&D, dividends and cash flow volatility (see text for details). Bank Lines of Credit Dummy equals one if an M&A is entirely financed by bank lines of credit, and equals zero otherwise. Mixed Sources Dummy equals one if an M&A is financed by a mix of bank lines of credit and corporate cash holdings. Relative Value is the ratio of deal value to the sum of the acquirer s market value of equity and deal value. Size is the logarithm of non-cash assets. M/B is market value of equity plus non-cash assets minus book value of equity, divided by non-cash assets. Leverage is the ratio of long-term debt to non-cash assets. Tangibility is the ratio of plant, property and equipment to non-cash assets. Capital Expenditure is the ratio of capital expenditure to non-cash assets. R&D is the ratio of R&D to non-cash assets. Dividends is the ratio of dividends to non-cash assets. Cash Flow Volatility is the standard deviation of cash flow to non-cash assets in the prior 5 years. Year Dummy Variables are the dummy variables for the years in the sample and not reported in the table. Industry Dummy Variables are the dummy variables for the industries defined by two-digit SIC codes and not reported in the table. The p-value is noted in the parentheses. ROA Net ROA Intercept -0.117 0.094 (0.06) (0.44) Bank Lines of Credit Dummy 0.012 0.025 (0.08) (0.05) Mixed Sources Dummy 0.014 0.012 (0.05) (0.40) Relative Value -0.047-0.079 (0.03) (0.06) ROA t-1-0.594-0.609 (0.01) (0.01) Size t-1 0.007 0.003 (0.01) (0.40) M/B t-1 0.006 0.002 (0.03) (0.68) Leverage t-1 0.042 0.017 (0.01) (0.55) Tangibility t-1-0.067-0.045 (0.11) (0.59) Capital Expenditure t-1 0.018-0.019 (0.34) (0.59) R&D t-1 0.001 0.061 (0.99) (0.59) Dividends t-1 0.187-0.023 (0.20) (0.94) Cash Flow Volatility t-1-0.008-0.068 (0.60) (0.03) Year Dummy Variables Yes Yes Industry Dummy Variables Yes Yes Number of Observations 667 667 Adjusted R-Square 0.42 0.17 22

Table 5: Multinomial Logistic Regression The table reports the multinomial logistic regression. The dependent variable is categorical. It equals zero if an M&A is entirely financed by corporate cash holdings, equals one if an M&A is financed by a mix of bank lines of credit and corporate cash holdings, and equals two if an M&A is entirely financed by bank lines of credit. Relative Value is the ratio of deal value to the sum of the acquirer s market value of equity and deal value. Cash is the ratio of cash and marketable securities to non-cash assets. Unused Lines of Credit is the ratio of unused lines of credit to non-cash assets. Institutional Ownership is the ratio of shares owned by institutional investors to the total shares outstanding at the end of a quarter prior to the announcement. Deviation from Optimal Insider Ownership is the absolute value of the residuals based on a benchmark regression of optimal insider ownership as reported in Appendix B. Size is the logarithm of noncash assets. Cash Flow is income before extraordinary items divided by non-cash assets. M/B is market value of equity plus non-cash assets minus book value of equity, divided by non-cash assets. Leverage is the ratio of long-term debt to non-cash assets. Tangibility is the ratio of plant, property and equipment to non-cash assets. Capital Expenditure is the ratio of capital expenditure to non-cash assets. R&D is the ratio of R&D to non-cash assets. Dividends is the ratio of dividends to non-cash assets. Cash Flow Volatility is the standard deviation of cash flow to non-cash assets in the prior 5 years. Year Dummy Variables are the dummy variables for the years in the sample and not reported in the table. Industry Dummy Variables are the dummy variables for the industries defined by two-digit SIC codes and not reported in the table. The p-value is noted in the parentheses. Cash=0, Mixed=1, Line=2 Line vs. Cash Mixed vs. Cash Line vs. Cash Mixed vs. Cash Intercept 15.038 2.492 15.764-0.382 (0.01) (0.46) (0.01) (0.92) Relative Value 9.581 10.844 8.655 9.888 (0.01) (0.01) (0.01) (0.01) Cash t-1-11.764-4.769-10.673-4.867 (0.01) (0.01) (0.01) (0.01) Unused Lines of Credit t-1 6.230 3.565 5.828 4.176 (0.01) (0.07) (0.01) (0.06) Institutional Ownership t-1 1.086 1.714 1.226 2.052 (0.06) (0.01) (0.07) (0.01) Deviation from Optimal Insider Ownership t-1 1.130 3.080 (0.65) (0.22) Size t-1-0.652-0.072-0.798-0.094 (0.01) (0.61) (0.01) (0.55) Cash Flow t-1-1.415 1.308-4.274 1.956 (0.64) (0.65) (0.22) (0.55) M/B t-1 0.488 0.295 0.454 0.282 (0.03) (0.14) (0.05) (0.20) Leverage t-1 0.948-0.702 0.494-0.367 (0.37) (0.51) (0.69) (0.76) Tangibility t-1-2.068-0.779-1.371 0.715 (0.12) (0.56) (0.39) (0.65) Capital Expenditure t-1 4.136-0.324 2.844-3.034 (0.23) (0.93) (0.46) (0.47) R&D t-1-1.448-4.134-10.788-5.445 (0.82) (0.41) (0.16) (0.29) Dividends t-1 6.277 7.511 5.806 9.675 (0.53) (0.44) (0.64) (0.38) Cash Flow Volatility t-1-1.009 0.224-1.527 0.370 (0.40) (0.86) (0.20) (0.77) Year Dummy Variables Yes Yes Yes Yes Industry Dummy Variables Yes Yes Yes Yes Number of Observations 411 284 342 249 Pseudo R-Square 0.60 0.60 0.62 0.62 23

Table 6: Acquisition Premium The table reports the impact of source of financing on acquisition premium. Premium 1 is the ratio of the offer price to the target share price one day prior to the acquisition announcement. Premium 2 is the ratio of the offer price to the target share price one week prior to the acquisition announcement. Premium 3 is the ratio of the offer price to the target share price four weeks prior to the acquisition announcement. Bank Lines of Credit Dummy equals one if an M&A is entirely financed by bank lines of credit, and equals zero otherwise. Mixed Sources Dummy equals one if an M&A is financed by a mix of bank lines of credit and corporate cash holdings. Relative Value is the ratio of deal value to the sum of the acquirer s market value of equity and deal value. Size is the logarithm of non-cash assets. Cash Flow is income before extraordinary items divided by non-cash assets. M/B is market value of equity plus non-cash assets minus book value of equity, divided by non-cash assets. Leverage is the ratio of long-term debt to non-cash assets. Tangibility is the ratio of plant, property and equipment to non-cash assets. Capital Expenditure is the ratio of capital expenditure to non-cash assets. R&D is the ratio of R&D to noncash assets. Dividends is the ratio of dividends to non-cash assets. Cash Flow Volatility is the standard deviation of cash flow to non-cash assets in the prior 5 years. Year Dummy Variables are the dummy variables for the years in the sample and not reported in the table. Industry Dummy Variables are the dummy variables for the industries defined by two-digit SIC codes and not reported in the table. The p-value is noted in the parentheses. Premium1 Premium2 Premium3 Intercept -0.595-0.537-0.532 (0.01) (0.01) (0.01) Bank Lines of Credit Dummy -0.053-0.054-0.048 (0.08) (0.07) (0.07) Mixed Sources Dummy -0.029-0.022-0.016 (0.26) (0.40) (0.49) Relative Value 0.545 0.526 0.500 (0.01) (0.01) (0.01) Size t-1 0.040 0.038 0.037 (0.01) (0.01) (0.01) Cash Flow t-1-0.313-0.319-0.262 (0.09) (0.08) (0.11) M/B t-1 0.029 0.028 0.024 (0.01) (0.01) (0.01) Leverage t-1-0.165-0.162-0.135 (0.02) (0.02) (0.02) Tangibility t-1 0.189 0.171 0.156 (0.04) (0.06) (0.05) Capital Expenditure t-1-0.095-0.062-0.080 (0.68) (0.78) (0.69) R&D t-1-0.444-0.437-0.410 (0.04) (0.04) (0.03) Dividends t-1-1.540-1.540-1.423 (0.02) (0.02) (0.01) Cash Flow Volatility t-1-0.014-0.010-0.008 (0.88) (0.91) (0.91) Year Dummy Variables Yes Yes Yes Industry Dummy Variables Yes Yes Yes Number of Observations 169 169 169 Adjusted R-square 0.32 0.31 0.36 24