Asset Buyers and Leverage. Khaled Amira* Kose John** Alexandros P. Prezas*** and. Gopala K. Vasudevan**** October 2009

Transcription

1 Asset Buyers and Leverage Khaled Amira* Kose John** Alexandros P. Prezas*** and Gopala K. Vasudevan**** October 2009 *Assistant Professor of Finance, Sawyer Business School, Suffolk University, **Charles William Gerstenberg Professor of Finance, Stern School of Business, New York University. ***Professor of Finance, Sawyer Business School, Suffolk University, ****Associate Professor of Finance, Charlton College of Business, University of Massachusetts, Dartmouth. Corresponding Author: Kose John Charles William Gerstenberg Professor of Finance Stern School of Business New York University NY, NY 10012

2 Asset Buyers and Leverage Abstract We examine a sample of 414 firms announcing asset purchases. Overall, buyer announcement-period returns increase with buyer leverage, but are higher only for buyers with private debt. Consistent with the hypothesis that private debt reduces the agency costs of managerial discretion we find that buyers with higher private debt make more efficient purchase decisions. Announcement-period returns are inversely related with buyer size, number of the buyer s anti takeover provisions, and relative size of the acquisition. Also, announcement-period returns increase with industry-adjusted seller leverage, suggesting that buyers benefit from the lower liquidity of assets sold by sellers with lower debt capacity. We also find that buyer announcement-period returns are directly related to their operating performance in the post-purchase year. Key Words: Asset Purchases, Capital Structure, Corporate Governance, Acquisitions JEL Classification: G31, G32, G34 2

3 Asset Buyers and Leverage Asset purchases occur frequently, especially in manufacturing. For example, Sabic purchased GE s plastics division for $11.6 billion and Univision Communications paid $1.1 billion for USA s television stations. Empirical examinations of asset purchase announcements find that the announcement-period returns for buyers of assets are positive or zero on average. The literature provides several explanations for both buyer and seller announcement-period returns. The fit hypothesis suggests that asset sales promote efficiency by deploying assets to better use. The focus hypothesis asserts that firms sell unrelated assets to improve the performance of their remaining more-focused operations. These explanations implicitly assume that managers act in the best interests of shareholders. However, Jensen (1986) indicates that managers serve their self-interests and are willing to pay in excess of economic value when they are using a firm s free cash flow to finance investments. The free cash flow hypothesis argues that firms can restrain managerial discretion by issuing debt to finance investments. The ensuing debt service obligation forces managers to disgorge future cash flows rather than invest them to enhance their benefits. In this study we focus on the monitoring hypothesis of Diamond (1984, 1991), Boyd and Prescott (1986), Berlin and Loeys (1988), and Warner (1977). These studies argue that banks and other private lenders tend to be more efficient at monitoring than public debt holders. This monitoring hypothesis implies that private debt may serve as a more efficient mechanism in reducing the agency costs of managerial discretion than public debt. 3

4 In our examination of the stock price reaction to asset purchase announcements we focus primarily on the impact of financial leverage on the announcement-period returns of the buyers. We classify a buyer firm as high (low) debt when its total debt-toassets ratio is higher (lower) than the median ratio for all the firms that bought assets in the same year. We further classify a buyer firm as public (private) debt when its publicto-private debt ratio is more (less) than one. We report several results that add to the literature. Overall, the univariate results show that announcement-period returns are positive and significant. Consistent with the free cash flow hypothesis, when the buyer is in the high debt group, buyer announcement-period returns are higher, and seller announcement-period returns are lower. Consistent with the monitoring hypothesis our univariate results also show that buyer and seller announcement-period returns are higher when the buyer has higher private debt. Cross-sectional regressions extend these results while controlling for other factors affecting stock price reaction such as the method of payment, relatedness of purchased asset to assets in place, buyer size, purchase price relative to buyer s assets, the corporate governance of the buyer, and the industry-adjusted leverage of the seller. Results show a significant relation between announcement-period returns and leverage. Consistent with the free cash flow hypothesis announcement-period returns are directly related to the debt of the asset buyer. However, announcement-period returns are higher for buyers with higher private debt, but not buyers with higher public debt. Announcement-period returns are lower for firms with more anti takeover provisions in place. The results also show a significant positive relation between announcement-period returns of buyers and 4

5 the industry-adjusted leverage of sellers. Furthermore, results show no relation between stock price reaction and the method of payment used for the purchase or the fit of the acquired asset with the buyer s assets in place. We also examine the operating performance of the asset buyers around the acquisition period. We use two ways to measure the operating performance of buyers around the time of an acquisition: 1) pretax operating cash flow scaled by the book value of assets; and 2) industry-adjusted performance, which adjusts raw performance by the performance of an industry portfolio. By both measures, performance is positive and significant, suggesting that buyers outperform their industry in all seven years surrounding the asset purchase. Cross-sectional regressions show a significantly positive relation between the announcement-period return and operating performance of buyers in the year following the purchase. This indicates that the market anticipates the improved operating performance following the asset purchase when it reacts positively to the acquisition announcement. Overall, our findings are consistent with the prediction of the monitoring hypothesis that buyers with private debt make more efficient purchase decisions. The results lend support to the notion that private debt is a better monitoring mechanism, thereby reducing the agency costs of managerial discretion and increasing firm value. Further, our findings also indicate that firms with higher anti takeover provisions (ATPs) in place create less value through these acquisitions perhaps due to their higher agency costs. The findings also support the view that the lower liquidity frequently attributed to assets sold by more leveraged sellers leads to better purchase decisions by the buyers of 5

6 these assets. We also find that transactions where the buyer is larger or the relative size of the acquisition is higher create less value. Furthermore, the increase in value induced by asset purchases reflects the buyer s post-purchase operating performance. The remainder of the paper is organized as follows. Section I reviews related research and introduces the testable hypotheses. Section II describes the data and methodology. Section III presents the findings. Section IV summarizes the results. I. Literature and hypotheses Section A reviews related papers and Section B develops the hypotheses. A. Literature Several studies examine the stock price reaction around the time of announcement of the transfer of asset ownership between firms. Most authors find positive announcement-period returns for sellers and positive or zero returns for buyers. 1 There are several hypotheses that seek to explain the wealth effects associated with these transfer announcements. The fit hypothesis suggests that sales promote efficiency by deploying assets to better uses. Certain assets have a better fit with the buyer s organization than the seller s. Since these assets are worth more to buyers, they will pay higher prices, and the sellers will capture some of the gains from the acquisition. In accordance with this hypothesis, Alexander, Bensen, and Kampmeyer (1984) document significantly positive abnormal returns for asset sellers, while Hite, Owers, and Rogers (1987) find positive returns for both sellers and buyers. Further, Maksimovic and Phillips (2001) find that asset sales tend to improve the productivity of the assets sold. 1 Buyers often sell off the acquired assets. For example, Kaplan and Weisbach (1992) find that 40% of acquisitions were sold by buyers in about seven years. 6

7 The focus hypothesis asserts that firms sell unrelated assets to improve the performance of their remaining more-focused operations. John and Ofek (1995) find that focus-increasing sellers experience significantly improved performance in the three-year period following the asset sale. They also find that announcement period stock returns for such sellers and rise with the operating performance of the remaining assets. They do not find significant stock returns for the buyers of the assets. 2 These explanations implicitly assume that managers act in the best interests of shareholders. However, managers value size and control. Jensen (1986) indicates managers are willing to pay in excess of economic value when they are using a firm s free cash flow to finance investments that serve their self-interests. Consistent with this free cash flow hypothesis, Lang, Stulz, and Walking (1991) find that bidders with high free cash flow and poor investment opportunities (low Tobin s q) experience negative stock returns when they announce acquisitions. Freund, Prezas, and Vasudevan (2003) provide additional supporting evidence for this hypothesis in the case of asset buyers. They document that announcement-period returns are inversely related to the amount of free cash flow for buyers with low Tobin s q. They also find that operating performance worsens in the three post-purchase years, and changes in performance are negatively related to the amount of free cash flow, especially for buyers with fewer growth opportunities. 3 2 Berger and Ofek (1995) find that the value of firms that diversify across different business lines is less than the sum of the imputed values of their segments. The loss of value is substantially less for focusincreasing diversification. 3 Asset sales proceeds retained by sellers represent free cash flow managers may use to pursue their interests. Lang, Poulsen, and Stulz (1995) document that asset sellers experience positive announcement returns only when the asset sales proceeds are paid out. 7

8 The preceding explanations of wealth effects associated with the announcement of asset purchases do not relate to the financial leverage of buyers or sellers. Our objective is to study the impact that leverage may have on these announcement returns. The explanations below provide the intuitive justification for the impact the debt level of the asset buyers and/or sellers may have on the announcement returns. Internal funds enable managers to avoid the monitoring by capital markets. This gives them the discretion to invest the firm s free cash flow in a self-serving manner. In the control hypothesis, Jensen (1986) argues that the firm can restrain managerial discretion by issuing debt (and not retaining the proceeds). The ensuing debt service obligation forces managers to disgorge future cash flows rather than invest them to enhance their benefits. 4 In essence, to avoid bankruptcy and the associated loss of their control rents, managers of firms with more debt make more efficient investment decisions. We examine whether this hypothesis is supported by the empirical evidence from asset purchases by studying the relation between announcement-period stock returns and financial leverage of asset buyers. Existing evidence indicates that the efficiency of monitoring by lenders varies with the type of debt employed. Diamond (1984, 1991), Boyd and Prescott (1986), Berlin and Loeys (1988), and Warner (1977) argue that banks and other private lenders tend to be more efficient at monitoring than public debt holders. Their large holdings enable them to exert more influence and pressure on management than public debt holders. This monitoring hypothesis implies that private debt may serve as a more efficient mechanism in reducing the agency costs of managerial discretion than public 4 Stulz (1990) and Hart and Moore (1995) formalize Jensen s argument for short-term and long-term debt, respectively. 8

9 debt. We examine whether asset purchases lend support to this hypothesis by studying the relation between announcement-period returns and private debt financing of asset buyers. Shleifer and Vishny (1992) offer an explanation of the price buyers are willing to pay to buy industry-specific assets of distressed firms. A distressed firm operating in a depressed industry may find few of its industry peers are able to pay the fundamental value to buy its assets, resulting in an illiquid market. 5 Hence, asset sellers may be forced to sell at lower prices (prices below value in best use) to well-financed buyers from other industries. Consistent with this liquidity hypothesis, Pulvino (1998) finds that in the airline industry asset sellers with lower debt capacity tend to receive lower prices for their planes than competitors with higher debt capacity. Similarly, Brown (2000) finds that real estate investment trusts with higher leverage are forced to sell assets at fire-sale prices to real estate competitors with lower leverage. For asset buyers, the liquidity hypothesis implies that returns will be higher when they buy assets from highleverage sellers. Hence, we examine the relation between the announcement returns of buyers and the financial leverage of asset sellers. Gompers, Ishii, and Metrick (2003) find that firms with more anti takeover provisions (ATPs) tend to have lower value and long-run stock price performance. They argue that firms with more ATPs will potentially have higher agency costs, because these provisions protect the managers by preventing their firms from being taken over. Managers of such firms are more likely to indulge in acquisitions that reduce firm value since they are less likely to be disciplined by the takeover market. Masulis, Wang, and 5 Even if these industry competitors are not financially constrained, antitrust and government regulations may prevent them from buying these assets. 9

10 Xie (2007) examine takeovers and find that acquirers with higher ATPs experience significantly lower announcement-period returns because these firms are more likely to engage in acquisitions that potentially destroy shareholder wealth. B. Hypotheses Against this background, we examine whether debt impacts the stock returns of buyers upon announcement of asset purchases. We first examine whether announcement -period returns depend on the total debt level of buyers. Then we determine whether returns depend on the type of debt, private or public, of the buyers. Finally, we ascertain whether stock returns depend on the debt level of the asset sellers. Our predictions are summarized in the following testable hypotheses. Hypothesis 1: The announcement-period returns of asset buyers are positively related to their financial leverage. Hypothesis 1 follows from the control hypothesis developed in Jensen (1986), Stulz (1990), and Hart and Moore (1995). Asset purchases are a form of corporate investment. Debt may alleviate the agency costs of managerial discretion by preventing selfinterested managers from making value-dissipating investments. We expect to find that buyers with more debt realize more gains from asset purchases and have higher announcement-period returns than buyers with less debt. Hypothesis 2: The announcement-period returns of asset buyers are positively related to their level of private debt. Our first testable hypothesis predicts that firms with more debt make better asset purchases, but it does not differentiate between private and public debt. Such a differentiation could be constructive, because, according to the monitoring hypothesis, 10

11 private debt may provide a better monitoring mechanism than public debt. We expect a direct relation between announcement-period returns and the private debt of buyers. Hypothesis 3: The announcement-period returns of asset buyers are positively related to the industry-adjusted leverage of the asset sellers. Hypothesis 3 follows from the liquidity hypothesis developed in Shleifer and Vishny (1995), which suggests that asset fire sales are more likely when sellers are in financial distress. This implies that asset buyers realize greater gains when they buy assets from distressed sellers. Hence, we expect a positive relation between the stock price reaction of the buyers and the degree of distress of the sellers. As in Pulvino (1998), our measure of distress is the industry-adjusted debt ratio of the seller in the year prior to the asset purchase. II. Sample and methodology We first describe our data and then the methodology. A. Sample We sample 414 U.S. industrial firms that acquired assets between 1986 and Each firm in the sample was chosen to meet the following criteria: 1. The buyer is an industrial firm, and information on the asset purchase is available in the database maintained by the Securities Data Company (SDC). 2. Only the first asset purchase is included per firm per year. 3. The asset purchased has a value of at least $100 million. 4. The buyer and the seller are publicly traded firms. 11

12 5. Accounting data on operating performance variables are available in the Compustat annual database for three years before and three years after the acquisition year. 6. Stock price data must be available in the Center for Research in Security Prices (CRSP) database with sufficient returns to estimate the market model. 7. News of the announcement is available on the Dow Jones News Retrieval Service, and the announcement is not contaminated by release of other information such as dividend or earnings announcements, or capital structure changes around the announcement date of the acquisition. Table I reports the distribution of the 414 acquisitions by calendar year. There is a substantial increase in the number of asset purchases over the period. Most of the purchases occur between 1996 and 2001 and the highest number is 65, in 1998 and The lowest number of asset purchases, 1, occurs in The table also reports the median total debt-to-book value of assets and the median amount of private and public debt-to-book value of assets by year for the asset buyers. All sample buyers had public and private debt. We follow COMPUSTAT definitions of debt. The private debt and public debt are obtained by examining the debt financial footnotes attached to the financial statements in the annual reports of the firms 10-K annual filings. These footnotes report the detailed distribution of the firm s debt. We include in the public debt all bonds and notes that are publicly placed. Private debt consists of all bank debt, privately placed bonds, and capitalized leases. When the SEC filings do not provide sufficient details regarding the debt, we search the SDC Platinum 12

13 and the Dow Jones interactive news wires for announcements about debt financing and the types of placement. Table II reports summary statistics for the 414 firms in the sample. All values shown are for the year prior to the acquisition (year 1). In the year before the asset purchase, the mean (median) book value of assets of the buyers is $19, ($2,593.45) million. The mean (median) purchase price is $ ($250) million. Purchased assets represent approximately 3.2% of the buyer s book value, 67% of the buyers paid with cash, and 56% of the time the purchased assets belonged to one of three main business lines of the buyer. The governance index we use is similar to Gompers et al. (2003). They obtain the anti takeover provisions from a publication by the Investors Responsibility Research Center (IRRC). The IRRC publication covers 24 unique ATPs from which Gompers et al. (2003) construct their governance index by adding one point for each provision that enhances managerial control. Firms with higher (lower) indices are expected to have lower (higher) shareholder rights since these ATPs make it more difficult for outsiders to remove the existing management. The mean of the governance index in our sample is indicating that these firms had close to ten anti takeover provisions in place. The mean pretax operating cash flow of the buyers is $1,891.3 million, with a mean cash flow-to-book value of assets ratio of 14.2%. Asset buyers have a mean total debt-to-book value of assets ratio of 28.4%. Their mean public debt scaled by the book value of assets is only 12.9%, and their mean private debt scaled by the book value of assets is 10.6%, indicating that the sample firms financed approximately the same percentage of their assets with public and private debt. 13

14 B. Methodology We examine the stock price reaction to the announcement of asset purchases, using the standard event-study method of Brown and Warner (1985) to compute daily excess returns. Average daily abnormal returns are computed by a two-step procedure, using stock price data from CRSP. We use the CRSP equal-weighted index as the market portfolio proxy. First, we estimate the parameters of a single-factor market model for each firm. We use the returns from day 255 to day 46 to estimate each firm s alpha and beta coefficient. Second, we compute the excess return by subtracting a firm s expected daily return from its actual return. Cumulative abnormal returns are calculated by summing the abnormal returns over the period from day 1 to day +1, where day 0 is the date of the asset purchase announcement. We classify the buyers into two groups, high debt and low debt, based on the total debt-to-book value of assets ratio. A firm is classified as high debt when its ratio of total debt-to-assets is higher than the median ratio for all the firms that bought assets in the same year and low debt when its ratio of debt-to-total assets is lower than the median for that year. For each group, we calculate the abnormal period return for the buyers and for the sellers. We also classify the buyers into two groups, public debt and private debt, based on the public-to-private debt ratio. Each year, a firm is classified as public debt when its ratio of public-to-private debt is more than one, and private debt when its public-toprivate debt ratio is less than one. 14

15 We then regress the buyer announcement-period returns against several buyer and seller characteristics. First we account for the buyer s total and private debt, both scaled by the book value of assets. Specifically, a dummy HDEBT set equal to 1 if the asset buyer has a ratio of total debt-to-assets that is higher than the median ratio for all the firms that bought assets in the same year. A dummy PRIDEBT, set equal to 1 if the buyer is a private debt firm 6. Hypotheses 1 and 2 predict positive coefficients for both these dummy variables. We also control for several buyer-specific characteristics that may affect the announcement-period returns. One is the method of payment for the purchase; a dummy CASH, set equal to one if the offer is made in cash, and zero otherwise. Travlos (1987) finds more positive stock price reaction for cash acquisitions, because overvalued bidders would pay for an acquisition using stock. Loughran and Vijh (1997) find firms paying cash for mergers have positive announcement returns during the five years after the acquisition, while firms paying with stock experience significantly poorer returns. We expect a positive coefficient for this dummy. John and Ofek (1995) and others have found the stock price reaction to be related to firm size, hence we use size as a control variable. We use SIZE, the natural log of the book value of assets, as our proxy for the size of the buyer. The relative size of the acquisition is a third characteristic that may affect announcement-period returns. We use RELSIZE, the asset purchase price scaled by the buyer s book value of the assets, which represents the proportion of the buyer s assets spent for the asset purchase as our proxy for the relative size of the acquisition. Travlos (1987) finds a negative coefficient, while Moeller, Schlingemann, and Stulz 6 We use also the dummy PUBDEBT for public debt, set equal to 1 if the buyer is a public debt firm. 15

16 (2004) find a positive (negative) coefficient for acquisitions made by smaller (larger) firms. We also control for the impact of several seller-specific attributes on the buyer s announcement-period returns using: IEBITDA, the industry-adjusted EBITDA/total assets in year 1; EBIDTAS, the EBITDA/total assets in year 1; IDEBT, the industry adjusted debt ratio in year 1; and DEBT, the total debt/total assets ratio in year 1, which are proxies for the operating performance and financial health of the asset seller. We use dummy variables for the industry-adjusted EBITDA/total assets ratio and industry-adjusted debt ratio to avoid multicollinearity between these variables and their unadjusted counterparts. To determine the industry-adjusted debt ratio in year 1, we subtract the median operating performance of all firms with the same two-digit SIC code as the seller firm from the sample seller s debt ratio. 7 If the difference is positive, the industry-adjusted debt ratio dummy is equal to 1, otherwise it is zero. This dummy is a better indicator of financial distress than the debt ratio because a high debt ratio does not necessarily imply that the seller is in distress, especially in a high-leverage industry. Pulvino (1998) uses a similar industry-adjusted measure. The liquidity hypothesis predicts a negative coefficient for the industry-adjusted debt ratio but not necessarily for the debt ratio of the seller. The dummy IEBITDA is set equal to 1 if the industryadjusted EBDITA/total assets is less than zero, otherwise it is zero. The liquidity hypothesis suggests a negative coefficient for EBITDA/total assets and a positive coefficient for the industry-adjusted EBITDA/total assets of the seller, as better operating performance reduces the probability of financial distress. 7 The industry-adjusted EBITDA/total assets in year 1 are determined in a similar way. 16

17 Next we examine the operating performance of the buyers over a seven-year period. This includes the three years before the asset purchase (years 3 through 1), the year of the asset purchase (year 0), and the three-year period after the asset purchase (years +1 through +3). The measure of operating performance is pretax operating cash flow scaled by the book value of total assets. Pretax operating cash flow is defined as net sales, less cost of goods sold, and less selling and administrative expenses before deducting interest, depreciation, and amortization expenses (Compustat item #13). The book value of total assets is the total value of assets (liabilities and net worth) from the balance sheet (Compustat item # 6). Barber and Lyon (1996) note that one problem in using the book value of assets to scale cash flows is that changes in this variable could create a bias in post acquisition performance. We address this issue by following the acquirer s performance for three years after the acquisition. We use pretax operating cash flow rather than earnings for two reasons. First, earnings include interest expense, income taxes, and special items that can obscure operating performance, the focus of our research. Second, operating cash flow represents the economic benefits generated by the firm, and as a pretax measure it is unaffected by the changes in capital structure or tax rates that can accompany asset purchases (Barber and Lyon, 1996). Since the level of these economic benefits depends on the total value of the firm s assets, we scale pretax operating cash flow by the book value of total assets. This gives us a performance measure we can compare across firms and through time (henceforth, operating performance). 17

18 We evaluate the levels of operating performance in two ways: raw and industryadjusted. To determine industry-adjusted performance, we subtract the median of the operating performance of all firms with the same two-digit SIC code as the buyer from the sample firm s operating performance. John and Ofek (1995), Kaplan (1989), and others have used this measure. To test the statistical significance of levels of operating performance we apply Wilcoxon signed-rank tests and parametric t-tests. III. Results First, we report the overall announcement-period returns for buyers and sellers during the three-day period around the asset purchase announcement. We also report the announcement-period returns for the buyers and corresponding sellers for the subsamples of high debt and low debt buyers, as well as the subsamples of public debt and private debt buyers. We then report the impact of asset buyer and seller characteristics on the buyers three-day cumulative abnormal returns. Finally, we report the operating performance of asset buyers over the seven-year period around the purchase. A. Announcement-Period Returns Three-day cumulative abnormal stock returns (days 1 through +1) to buying firms are reported in Table III. The first row in both panels reports the cumulative abnormal returns for the overall sample. The mean (median) cumulative abnormal return for the buyers is 1.238% (0.505%), significantly different from zero at the 1% (1%) level. The market perceives that asset purchases create value. The second and third rows in Panel A report the abnormal returns for the subsamples of high and low debt buyers. The mean (median) return for the high debt sample is positive, 1.813% (0.962%), significantly different from zero at the 1% (1%) level. The mean (median) return for the 18

19 low debt sample is 0.686% (0.453%), both not significantly different from zero at customary levels. The means of the two groups are significantly different at the 10% level. The second and third rows in Panel B report the abnormal returns for the subsamples of public and private debt buyers. The mean (median) return for the private debt sample is positive, 2.343% (1.220%), significantly different from zero at the 1% (1%) level. The mean (median) return for the public debt sample is % (0.150%), both not significantly different from zero at customary levels. The means (medians) of the two groups are significantly different at the 1% (1%) level. Our analysis indicates that the market expects asset purchases to create value for the buyers. Freund, Prezas, and Vasudevan (2003), and Sicherman and Pettway (1992) report similar findings. Consistent with the control hypothesis, our results suggest that the market expects this value to be higher for buyers with higher levels of debt. Furthermore, consistent with the monitoring hypothesis, our results imply that the market expects buyers with more private debt to create more value. Our univariate results, however, do not account for other factors that may impact the stock price reaction of the buyers. Our regression results in Table V address this issue. Table IV reports the three-day abnormal returns for asset sellers. The first row in both panels reports the cumulative abnormal returns for the overall sample. The mean (median) return for the overall sample is 0.949% (0.385%), significant at the 5% (5%) level. This is consistent with the fit and focus hypotheses. The second and third rows in Panel A report the abnormal returns for the subsample of sellers who sold their assets to high and low debt buyers. The mean (median) return for the sample of firms that sold their assets to high debt buyers is 0.598% (0.383%), significant at the 10% (10%) level. 19

20 The mean (median) returns for the sample of firms which sold their assets to low debt buyers is 1.279% (0.445%), significant at the 10% (10%) level. The second and third rows in Panel B report the abnormal returns for the subsample of sellers who sold their assets to private debt and public debt buyers. The mean (median) return for the sample of firms that sold their assets to private debt buyers is 1.035% (0.532%), significant at the 10% (10%) level. The mean (median) returns for the sample of firms which sold their assets to public debt buyers is 0.890% (0.298%), both not significant at customary levels. 8 Table V reports the results of our cross-sectional regressions relating buyer announcement-period returns to buyer and seller characteristics. The independent variables are: the dummy CASH for the type of payment; the dummy FIT, for fit of the asset with the existing assets of the buyer; the dummy HDEBT set equal to 1 if the buyer is a high debt firm; the dummy PRIDEBT set equal to 1 if the buyer is a private debt firm; the variable SIZE for buyers book value of assets; the variable RELSIZE the relative purchase price which is the purchase price of the asset scaled by the book value of assets of the buyer; and the variable GOVERNANCE which is the GIM index for the asset buyer in the year of the purchase. The seller variables are: the dummy IEBITDA which is set equal to 1 if the industry-adjusted EBITDA/total assets of the seller is positive and zero otherwise; the dummy IDEBT for the industry-adjusted debt ratio of the seller which is set equal to 1 if this value is positive and zero otherwise; the variable EBITDAS which is the EBITDA/total assets of the seller; and the variable DEBT which 8 Since our interest is in asset buyers, we do not derive cross-sectional results for the relation between asset seller stock returns and the leverage of buyers. 20

21 is debt ratio of the seller. Except for the payment and fit dummy variables and the relative purchase price, all explanatory variables are determined as of year 1. The first column in Table V reports the results including only the HDEBT dummy and other control variables. Consistent with our hypothesis 1, the coefficient of HDEBT is positive (0.010) and significant at the 5% level. The coefficient of SIZE is negative ( 0.005) and significant at the 1% level. Consistent with the findings of Moeller, Schlingemann, and Stulz (2004), this indicates that the market perceives smaller firms purchase assets that create more value. The coefficient of RELSIZE is negative ( 0.013) and significant at the 1% level. This implies that the higher the purchased asset costs relative to the buyer s assets in place, the less value it creates. The coefficient of IDEBT industry-adjusted debt ratio of the asset seller in year 1 is positive (0.010) and significant at the 10% level. This is consistent with the liquidity hypothesis which predicts that firms with higher leverage are more inclined to sell their assets at lower prices, enabling buyers to capture more of the transaction gains. The coefficients of the remaining explanatory variables are not significant at customary levels. The adjusted R- squared is 13.40%, while the F-value is 6.78 and significant at the 1% level. The second column in Table V reports the results of the buyer announcementperiod returns for the same variables except that the HDEBT dummy is replaced by the private debt dummy, PRIDEBT. The coefficient of PRIDEBT is positive, 0.012, and significant at the 5% level. This indicates that the announcement-period returns are 1.2% higher for buyers which belong to the private debt group. Similar to the first regression, the coefficients of SIZE and RELSIZE are negative and significant. None of the other 21

22 variables are significant. The adjusted R-squared is 14.10%, while the F-value is 7.11 and significant at the 1% level. The third column in Table V repeats the regression in the first column with the addition of the GOVERNANCE variable. The coefficient of HDEBT is not significant. The coefficient of GOVERNANCE is negative, , and significant at the 1% level indicating that the market perceives acquisitions by high ATP firms as creating lower value. Similar to the regression in the first column, the coefficients of SIZE, RELSIZE and IDEBT are all significant. The adjusted R-squared is 5.6% and the F-value is 2.27, significant at the 1% level. The fourth column repeats the regression in the second column with the addition of the GOVERNANCE variable. The coefficient of PRIDEBT is positive, and significant at the 5% level. The coefficient of GOVERNANCE is negative, and significant at the 1% level. From the remaining variables, only the coefficients of SIZE and RELSIZE (both negative), and the coefficient of IDEBT (positive) are significant. The adjusted R-squared is 8.5% and the F-value is 2.98, significant at the 1% level. The results in Table V indicate that the market perceives acquisitions by firms with higher debt create more value. However, consistent with our hypothesis 2, the increase in value is significant for buyers with private debt. 9 These results support the monitoring hypothesis, as they imply that holders of private debt are better monitors ensuring that the buyer firms acquire more efficient assets, thereby lowering their managerial discretion costs. We also find that the announcement-period returns are higher for buyers with fewer anti takeover provisions. Further, announcement-period returns are lower when the buyer 9 We have repeated the regressions using the dummy for public debt and found it to be negative and significant. 22

23 is larger and when the relative size of the acquisition is larger. Our findings of positive coefficients for the industry-adjusted debt ratio of the seller are consistent with the liquidity hypothesis of Shleifer and Vishny (1992). B. Operating Performance Table VI presents the results for the operating performance of asset buyers. In Panel A the mean (median) cash flow-to-book value ratio is (0.147) in year 3 and (0.145) in year 1, but declines to (0.126) in year +1. This ratio improves in year +3, with a mean (median) value of (0.123). In Panel B the mean (median) values for the industry-adjusted cash flow-to-book value ratio are positive and significant in all years, indicating that the sample asset buyers are performing better than their industry peers overall. The mean (median) industryadjusted cash flow-to-book value ratio is (0.035) in year 3, (0.043) in year 1 and (0.030) in year +1. This ratio improves to (0.036) in year We have reported the operating performance of the asset buyers during the seven year period around the announcement. In efficient capital markets, prices should reflect the operating performance resulting from the asset purchase. We should expect to find a positive relation between the change in firm value at the time of announcement and the operating performance following the acquisition. To check for the presence of such a relation in our sample, we provide regressions of the announcement-period return of buyers against EBITDAP1, the EBITDA to assets ratio in year All our regressions control for the cash flow-to-book value of assets in year 1. The other variables we use 10 Although not reported here, changes in industry-adjusted operating performance are insignificant for the periods 1 to +1 and 1 to +2, but significantly positive for the period 1 to Repeating the regressions using the EBITDA/Assets in year +2 and EBITDA/Assets in year +3 produces insignificant results. 23

24 in the regression are CASH, FIT, HDEBT, PRIDEBT, and SIZE defined in Table V. Table VII reports the findings. In column 1, we report the regression results with these variables and the HDEBT dummy. The coefficient of the EBITDAP1 is positive, 0.092, and significant at the 1% level. Other significant variables are the HDEBT dummy and SIZE. The adjusted R- squared is 3.80%, while the F value is 3.60 and significant at the 1% level. The second column repeats the regression substituting debt with private debt. The coefficient of PRIDEBT is positive, and significant at the 1% level. The coefficient of the EBITDAP1 is positive, 0.092, and significant at the 1% level indicating that the market is able to anticipate some of the improvements in performance following the acquisition. The coefficient of SIZE is negative, and significant at the 1% level. The adjusted R-squared is 4.1% and the F-value is 3.82, significant at the 1% level. The third column reports the results with none of the debt dummies. The coefficient of the EBITDAP1 is positive, and significant at the 1% level. The coefficient of SIZE is negative, and significant at the 1% level. The adjusted R-squared is 3.2% and the F-value is 5.32 and significant at the 1% level. The results in Table VII thus imply that the market anticipates a firm s operating performance following the asset purchase when reacting positively to the announcement of the acquisition. It further reinforces our finding that buyers in the private debt sample make acquisitions that increase firm value. IV. Summary and Conclusions Our research provides evidence on the impact of leverage on the stock market s reaction to asset purchase announcements. Our hypothesis is that buyer firms with higher 24

25 debt, especially private debt, make better acquisition decisions. We also hypothesize that buyers realize better returns when they purchase assets from sellers with lower debt capacity. We further examine the effect of anti takeover provisions on the announcementperiod returns to asset buyers. We find that both the mean and median cumulative abnormal returns for our sample of 414 buyers are positive and significant. This indicates that asset purchases enhance firm value. We obtain several cross-sectional results regarding the impact of debt on asset buyers. First, there is a positive and significant relation between the announcement-period returns and buyers overall leverage. This suggests that the market perceives acquisitions by firms with higher leverage as creating more value, which is consistent with the control hypothesis that managers of firms with more debt make more efficient investment decisions. Second, there is a positive and significant relation between the announcement-period returns and private debt. This is consistent with the hypothesis that private debt provides better monitoring, thereby reducing the agency costs of managerial discretion. On the other hand, we find that the market perceives buyers with higher amounts of public debt as creating less value through these acquisitions. Third, consistent with the liquidity hypothesis, there is a significantly positive relation between the stock price reaction of buyers and the industry-adjusted leverage of sellers. Further, our cross-sectional results indicate that announcement-period returns are lower when the buyer has in place fewer anti takeover provisions. This suggests that the market perceives these firms as making acquisitions that create less value. Also, announcement-period returns are lower when the buyer is larger and when the relative size of the acquisition is larger. 25

26 We find that sample buyer firms experience significantly better mean and median operating performance than the industry in the seven-year period around the asset purchase. Our results also suggest that the market anticipates some of the post-purchase changes in the operating performance of the buyers; there is an overall positive and significant cross-sectional relation between the announcement-period returns and the operating performance of buyers in the year after the asset purchase. Our study indicates that the structure of debt and the corporate governance of the asset buyer are important determinants of the value created by these asset purchases. 26

27 References: Alexander, G., G. P. Bensen, and J. Kampmeyer, 1984, Investigating the Valuation Effects of Announcement of Voluntary Corporate Selloffs, Journal of Finance 39, Barber, B. M. and J. D. Lyon, 1996, Detecting Abnormal Operating Performance: The Empirical Power and Specification of Test-Statistics, Journal of Financial Economics 41, Barclay, M.J. and C.W. Smith, Jr, 1995, The Maturity Structure of Corporate Debt, Journal of Finance 50, No 2, Berger, P.G and E. Ofek, 1995, Diversification's Effect on Firm Value, Journal of Financial Economics 37, Berlin, M. and J. Loeys, 1988, Bond Covenants and Delegated Monitoring, Journal of Finance 43, Boyd, J. and E.C. Prescott, 1986, Financial Intermediary Coalitions, Journal of Economic Theory 38, Brown, D.T., 2000, Liquidity and Liquidation: Evidence from Real Estate Investment Trusts, Journal of Finance 55, Brown, S.J. and J.B. Warner, 1985, Using Daily Stock Returns: The Case of Event Studies, Journal of Financial Economics 14, Diamond, D.,1984, Financial Intermediation and Delegated Monitoring, The Review of Economic Studies 3, Diamond, D., 1991, Monitoring and Reputation: The Choice Between Bank Loans and Publicly Traded Debt, Journal of Political Economy 99, Freund, S., A.P. Prezas, and G.K. Vasudevan, 2003, Operating Performance and Free Cash Flow of Asset Buyers, Financial Management 32, Gompers, P.A, J. Ishii, and A. Metrick, 2003, Corporate Governance and Equity Prices, Quarterly Journal of Economics 118, Hart, O. and J. Moore, 1995, Debt and Seniority: An Analysis of the Role of Hard Claims in Constraining Management, American Economic Review 85, Hite, G.L., J.E. Owers, and R.C. Rogers, 1987, The Market for Interfirm Asset Sales, Partial Sell-offs and Total Liquidations, Journal of Financial Economics 18,

28 Jensen, M.C., 1986, Agency Costs of Free Cash Flow, Corporate Finance and Takeovers, The American Economic Review 76, John, K. and E. Ofek, 1995, Asset Sales and Increase in Focus, Journal of Financial Economics 37, Kaplan, S., 1989, The Effects of Management Buyouts on Operating Performance and Value, Journal of Financial Economics 24, Kaplan, S. and M. Weisbach, 1992, The Success of Acquisitions: Evidence from Divestitures, Journal of Finance 47, Kruse, T.A., 2002, Asset Liquidity and the Determinants of Asset Sales by Poorly Performing Firms, Financial Management 31, Lang, L., A. Poulsen, and R.M. Stulz, 1995, Asset Sales, Firm Performance and the Agency Costs of Managerial Discretion, Journal of Financial Economics 37, Lang, L., R. M. Stulz, and R. A. Walkling, 1991, A Test of the Free Cash Flow Hypothesis: The Case of Bidder Returns, Journal of Financial Economics 29, Loughran, T. and A. Vijh, 1997, Do Long Term Shareholders Benefit from Corporate Acquisitions? Journal of Finance 52, Masulis, R.W, C. Wang, and F. Xie, 2007, Corporate Governance and Acquirer Returns, Journal of Finance, Vol 62 (4), Maksimovic, V. and G. Phillips, 2001, The Market for Corporate Assets: Who Engages in Mergers and Are There Efficiency Gains? Journal of Finance 56, Moeller, S. F. Schlingemann, and R. Stulz, 2004, Firm Size and the Gains from Acquisitions, Journal of Financial Economics 73, Pulvino, T., 1998, Do Asset Fire Sales Exist? An Empirical Investigation of Commercial Aircraft Transactions, Journal of Finance 53, Shleifer, A. and R.W. Vishny, 1992, Liquidation Values and Debt Capacity: A Market Equilibrium Approach, Journal of Finance 47, Sicherman, N. W. and R. H. Pettway, 1992, Wealth Effects for Buyers and Sellers of the Same Divested Asset, Financial Management 21, Stulz, R., 1990, Managerial Discretion and Optimal Financing Policies, Journal of Financial Economics 26,

29 Travlos, N.G., 1987, Corporate Takeover Bids, Methods of Payment and Bidding Firms Stock Returns, Journal of Finance 42, Warner, J.B., 1977, Bankruptcy, Absolute Priority, and the Pricing of Risky Debt Claims, Journal of Financial Economics 4,

30 Table I. Distribution of Asset Buyers by Calendar Year The distribution of 414 asset buyers by year of the purchase during the period The sample is taken from the Securities Data Company database. Only the first asset purchase is included per firm per year. Sample asset purchases were by industrial firms and had a value of at least $100 million. Except for the number of firms, values are for the year prior to the purchase. Total debt is defined as the book value of debt divided by the book value of total assets. Public debt is defined as the book value of public debt divided by the book value of total assets. Private debt is defined as the book value of private debt divided by the book value of total assets. Year Number of buyers with Private and Public Debt Median Buyer Total Debt-to- Book Value of Assets Median Buyer Public Debt-to- Book Value of Assets Median Buyer Private Debt-to- Book Value of Assets Total number of firms

31 Table II. Summary Statistics for Asset Buyers Summary statistics for selected variables for 414 asset buyers during the period The sample is taken from the Securities Data Company database. Only the first asset purchase is included per firm per year. Sample asset purchases were by industrial firms and had a value of at least $100 million. Except for the purchase price, values are for the year prior to the purchase. Pretax operating cash flow is defined as net sales, minus cost of goods sold, minus selling and administrative expenses, but before deducting interest, depreciation, and amortization expenses. Cash flow-to-book value is pretax operating cash flow divided by the book value of total assets and measures the sample firm s raw performance. Industryadjusted cash flow-to-book value is the buyer firm s cash flow-to-book value ratio minus the median cash flow-to-book value ratio of all firms with the same two-digit SIC code as the buyer. Median industry cash flow-to-book value is the median cash flow-to-book value ratio of all firms with the same two-digit SIC code as the buyer. Total debt is defined as the book value of total debt divided by the book value of total assets. Public debt is defined as the book value of public debt divided by the book value of total assets. Private debt is defined as the book value of private debt divided by the book value of total assets. Variable Obs. Min. Mean Max. Median Std. Dev. Panel A. Firm Variables Book Value ($MM) Purchase Price ($MM) Dummy Variable Set Equal to 1 if Purchase is for Cash Dummy Variable Set Equal to 1 if Asset Purchased Belongs to One of Three Main Lines of Business of the Buyer Governance Index of the Buyer in Year Panel B. Firm Operating Performance and Leverage Variables Pretax Operating Cash Flow ($MM) Cash Flow-to-Book Value of Assets Industry-Adjusted Cash Flow-to- Book Value of Assets Median Industry Cash Flow-to- Book Value of Assets Total Debt-to-Book Value of Assets Total LT Debt-to-Book Value of Assets Public Debt-to-Book Value of Assets Private Debt-to-Book Value of Assets