ASSET SALES AND PRODUCT MARKETS By KYOUNG-MIN KWON A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Finance 2012 ABSTRACT ASSET SALES AND PRODUCT MARKETS By Kyoung-min Kwon This study investigates the information content revealed by the asset sale announcement about the viability of the trading relationship with the large customer. We find the asset sale announcement reveals largely negative information about the relationship viability, and thus affects the asset selling supplier and the large customer of the asset seller negatively. We find the negative information is more evident when the post-sale servicing commitment is important and the asset seller makes intense relationship-specific investment. Consistently, the asset sale announcement also reveals reduced incentives to maintain the major trading relationship. The relationship is more likely to be terminated or shortened and the proportion of the bilateral trading is more likely to be reduced after the asset sale when post-sale servicing commitment is important. Our results show that the asset sale announcement contains negative-informationsignal on the trading relationship with the large customer. Copyright by Kyoung-Min Kwon 2012 To my wife and parents for their unreserved love and support iv ACKNOWLEDGEMENTS I am truly indebted a great deal to my dissertation committee- Dr. Edward Fee (Chair), Dr. Jun-Koo Kang, Dr. Charles Hadlock, Dr. Michael Mazzeo, and Dr. Andrei Simonov for their insightful comment, helpful guidance, and unreserved support. In particular, I am deeply indebted to Dr. Jun-Koo Kang, my lifelong mentor and role model. My dissertation would have been impossible without his endless support, encouragement, and guidance. I also owe a huge debt of gratitude to Dr. Edward Fee for all his support throughout all phases of the dissertation. I would like to thank Dr. Charles Hadlock for supporting me all throughout the doctoral program. I thank Dr. Michael Mazzeo and Dr. Andrei Simonov who have also been instrumental for me to develop as an academic. I thank Shan Yan, Min-Jung Kang, Guojun Chen, the graduate students who shared knowledge and thought for long time in the class rooms and offices at Michigan State University Finally but most, my deepest gratitude goes to my wife and parents for their warm understanding and infinite support. Without their sacrifices, this would not have been possible. v TABLE OF CONTENTS LIST OF TABLES……………………………………………………………....……………..... vii LIST OF FIGURES……………………………………………………………………………...viii 1 Introduction……………………………………………………………………………… 1 2 Literature review and hypothesis development…………………………………..……... 6 3 Data…………………………………………………………………………………….. 12 4 Asset sale announcement returns………………………………………………………. 14 4.1 Announcement returns for asset selling firms and large public customers…… 14 4.2 Multivariate analysis of announcement returns for asset selling firms……….. 15 4.3 Multivariate analysis of announcement returns for large public customers of asset selling firms……………………………………………………………... 19 5 Changes in post-asset sale relationship strength……………………………………….. 21 6 Determinant of asset sale………………………………………………………………. 24 7 Summary and conclusion………………………………………………………………. 26 APPENDIX A TABLES AND FIGURES…………………………………………….……... 28 APPENDIX B VARIABLE DEFINITIONS.........……………………………………………66 BIBLIOGRAPHY……………………………………………………………………………….. 70 vi LIST OF TABLES Table 1 Sample Distribution by Year, Industry, and Number of firms with large public customer…………………………………………………………………………….... 29 Table 2 Deal-, Seller-, Customer-, Buyer-Specific Characteristics…………………………… 31 Table 3 Cumulative abnormal returns (CARs) for asset selling firms, their large public customers, and asset buying firms……………………………………………………. 39 Table 4 OLS regressions of cumulative abnormal returns (CARs) for asset selling firms around the asset sale announcement date on explanatory variable…………………………... 43 Table 5 OLS regressions of cumulative abnormal returns (CARs) for asset selling firms with large public customers and without large customers…………………………………. 45 Table 6 OLS regressions of cumulative abnormal returns (CARs) for large public corporate customers of asset selling firms around the announcement date on explanatory variables ……………………………………………………………………………....50 Table 7 Logit regressions of relationship termination after asset sales by firms with large public customers…………………………………………………………………….... 53 Table 8 Logit regressions of change in percentage of COGS of large public corporate customers of asset selling firms after the asset sales…………………………...…...... 55 Table 9 Post-asset-sale relationship survival for asset selling firms with a large public customer and non-asset selling matching firms with a large public customer……………...…... 57 Table 10 Determinant of asset sales……………………………………………………………. 59 Table 11 Determinant of asset sales: Subsampled by technology intense industry and durable industry……………………………………………………………………………….. 61 vii LIST OF FIGURES Figure 1 Interaction Term in Logit Regression: Model (2) of Table 7.........................................62 Figure 2 Interaction Term in Logit Regression: Model (4) of Table 7.........................................63 Figure 3 Interaction Term in Logit Regression: Model (2) of Table 8………………………….64 Figure 4 Interaction Term in Logit Regression: Model (4) of Table 8………………………….65 viii 1. Introduction Many firms in the U.S. maintain durable trading relationships with major customers. Previous literature shows that the existence of such relationships can have a significant effect on firms’ financial policies and ownership structure. For example, Titman (1984) and Banerjee, Dasgupta, and Kim (2008) argue that the existence of large customer affects a firm’s ability to use debt and Fee, Hadlock, and Thomas (2006) show that product market relationships have an important implication for ownership structure of firms with major customers. Johnson, Kang, Masulis, and Yi (2011) find that a firm’s decision to issue equity has spillover effects down the supply chain. Several studies further show that the customer-supplier relationship influences other corporate important decisions, such as those related to CEO compensation (Arora and Alam (2005)), information disclosure (Almazan, Suarez, and Titman (2006)), earnings management (Raman and Shahrur (2008)), and IPO valuation (Johnson, Kang, and Yi (2008)), However, in spite of extensive research on the effect of major customer relationships on a wide spectrum of corporate decisions, relatively little is known about how customer-supplier 1 relationships influence a firm’s investment decision such as asset sales. In this study we examine the effect of asset sellers’ trading relationships with a large customer on asset sale announcement returns and the continued viability of these relationships. We explore several questions that aim to shed light on the role of the customer-supplier relationships in asset sales. For example, does a close relationship with a large customer affect an asset selling firm’s announcement returns? Does information revealed by an asset sale announcement pass along the supply chain and thus affect customer value around asset sale announcements? What are the determinants of the asset sale valuation effects for firms that 1 The exception is Fee and Thomas (2004) who examine the upstream and downstream productmarket effect of firms’ horizontal mergers and acquisitions activities. 1 maintain customer-supplier relationships? What is the impact of the asset sale on post-sale customer-supplier relationship dynamics? Customer-supplier relationships can have a significant effect on a firm’s asset sales decision and its value. There are two competing views on this effect, namely, the view that asset sales adversely affect the values of both the supplier and the customer (the “negative-informationsignal-view”) and the view that they have a positive effect on the values of both the supplier and the customer (the “positive-information-signal-view”). According to the negative-informationsignal-view, asset sales announcements by firms with a large customer convey negative information about the continued viability of sellers’ trading relationship with the customer. Shleifer and Summers (1988) argue that the supplier-customer relationships are governed by non-enforceable implicit contract and thus vulnerable to new information that can have an effect on the viability of their relationship. Because the supplier’s asset sales can involve relationshipspecific assets that are critical in continuing its relationship with major customers, they may have a significant effect on the profitability of both suppliers and customers and the continued 2 viability of their relationship. For example, if the supplier sells the assets with high levels of asset specificity that are critical in maintaining the trading relationship with its customer, its asset sale announcements should convey important information about the continued viability of its trading relationship with the customer. It is also possible that firms with a large customer engage in asset sales to prepare a potential loss of their major customers who are in financial difficulty and develop new relationships with other customers. In both cases, asset sale announcements by 2 For our sample, about 30% of suppliers’ total sales are generated from the sales to the large customers. We define a large customer as a customer who is accountable for more than 10% of a firm’s annual sales. 2 firms with a large customer can convey negative information about the customer-supplier relationship, adversely affecting the values of both the supplier and its customer. In contrast, the positive-information-signal-view suggests that asset sales announcements by firms with a large customer convey positive information about the continued viability of sellers’ trading relationship with the customer. For example, if a supplier sells its assets to increase its efficiency or to relax its financial constraint, its asset sales announcements can be good news for both the supplier and the customer since asset sales will strengthen the supplier-customer relationship by having a more reliable supplier. In this case, asset sales announcements should have a positive effect on the values of both the supplier and the customer. The negative- (positive-) information-signal-view also suggests that the negative (positive) effects of asset sale announcements on the values of the supplier and the customer are more pronounced when the suppliers are in durable goods or high-tech industries where post-sale service and product guarantee are important (Titman (1984)) or when they have made intense relationship-specific investments (Williamson (1983)). The view also suggests that asset sales by firms with a large customer changes in trading partners’ incentives to maintain their relationships and thus impacts post-sale relationship strength, particularly when the post-sale service and product guarantee support are important. We find that our results are largely consistent with the negative-information-signal-view. Specifically, we find that asset sellers with a large public customer realize less positive abnormal announcement returns when they are in industries where post-sale service and product guarantee are important: For sellers with a large public customer, their CARs (-1, 1) are significantly lowers when they operate in durable goods or high tech industries and these lower returns are more evident when these sellers are perceived to be risky (i.e., have high stock return volatility) 3 in the market. Sellers with a large public customer also realize lower returns when they have higher levels of asset specificity (i.e., higher R&D intensity) or when their post-sale relationship strength is expected to deteriorate. In contrast, we do not observe these results for sellers without a large public customer. We also find that the large customers of asset sellers experience more negative returns when sellers operate in high tech industries, particularly when they are risky, when they invest more in R&D activity, when the customer-supplier relationship is expected to be terminated after the asset sale, and when the large customers are in competitive industries, as measured by Herfindahl index. The analysis of post-sale dynamics also supports the negative-information-signal view. We find that the likelihood of terminating the major trading relationship increases after the asset sale if risky sellers operate in industries where post-sale product support is important. We also find that these risky sellers are more likely to experience reductions in their sales to the large customer after the asset sales. Our research contributes to the literature in several ways. First of all, to the best of our knowledge, this study is the first to examine the effect of customer-supplier relationship on voluntary asset sale. Previous literature shows that the existence of customer-supplier relationships can have a significant effect on a wide spectrum of corporate decisions such as capital structure, ownership structure, compensation decision, and financial decision, but no study to date examines how major customer-supplier relationship affects asset sales. Unlike Johnson, Kang, Masulis, and Yi (2011) who investigate the spillover effect of a firm’s financing decision (SEO) on large customers, we examine the effect of a firm’s investment decision on large customers. Our study is also different from Fee and Thomas (2004) who examined the spillover effect of horizontal mergers, another type of important corporate investment activities, 4 on trading partners in supply chains to find the source of value gain in horizontal mergers. We expand their analysis by examining asset sales, which enable us to investigate the change in trading relationships that does not involve the change in ownership. Second, we contribute to the asset sale literature. Previous studies examine the effect of asset sales announcements on the seller’s shareholder value and the potential source of this wealth gains. For example, Gordon, Benson, and Kempmeyer (1984), Jain (1985), and Hite, Owers, and Rogers (1987) find the positive shareholder wealth effect of asset sales. Maksimovic and Gordon (1998, 2001) and Yang (2008) argue that this positive shareholder wealth effect of asset sales comes from the increase in efficiency of selling firms. On the other hand, Lang, Poulsen, and Stulz (1995) and Schlingemann, Stulz and Walkling (2002) argue that asset sales are motivated by selling firms’ financial need, which also benefits shareholders. We extend this literature by examining the effect of asset sales on the value of customers, another important type of a firm’s stakeholders, and introducing customer-supplier relationships as another significant factor that affects firm value. Third, we contribute to the literature on product market relationships. By showing the importance of major customer relationships in asset sales we provide new evidence on the argument that the product market relationship has significant impact on corporate decisions. The study proceeds as follows. In Section II we discuss related literature and main hypotheses. Section III describes the data and sample characteristics. In Section IV we report abnormal returns for asset selling firms and their customers, and provide results from crosssectional regressions. We examine the post-asset sale changes in relationships between asset selling firms and their large public customers in Section V. In section VI we analyze the determinants of asset sale. Finally, Section VII summarizes and concludes the paper. 5 2. Literature review and hypothesis development Previous literature shows that voluntary asset sales have positive announcement effects for both asset sellers and buyers. For example, Jain (1985) finds significant positive mean excess returns of 0.70% and 0.34%, respectively, for asset sellers and buyers in a sample of over 1,000 sell-off events and Gordon, Benson, and Kampmeyer (1984) report a mean abnormal announcement returns of 8% for asset selling firms in a sample of 53 voluntary corporate selloffs. The positive returns for sellers and buyers can come from several different sources. First, the gains from asset sales may come from the increase in efficiency by reallocating assets from the inefficient sellers to the buyers who can make better use of them, thereby creating higher value for both asset sellers and buyers. Hite, Owers, and Rogers (1987) show that sellers realize gains at the bid but lose the gains at the offer termination, and interpret these results as evidence that the gains from asset sales are not from price adjustment but from efficiency gain. Similarly, Maksimovic and Phillips (2001) find that asset transactions tend to improve the allocation of resources and are originated from profit maximizing decision. The second potential source of gains is the increase in focus. John and Ofek (1995) show that asset sellers’ announcement returns and post-sale performances are greater when they increase the focus in their main business after the asset sales. The third potential source of gains is the relaxation of sellers’ financial constraint. Lang, Poulsen, and Stulz (1995) argue that the asset sale decisions are motivated by the selling firm’s financial situation and find that the gains from asset sales is dependent on the use of sale proceeds. Schlingemann, Stulz, and Walkling (2002) argue that liquidity in the market for corporate asset plays an important role in determining firms’ asset sales and show that firms are more likely to divest segments in industries with a more liquid 6 market for corporate assets, unrelated segments, poorly performing segments, and small segments. However, previous research on asset sales largely ignores the role of product market relationship in asset sales. In particular, we have almost no evidence on asset sales in which selling firms maintain the major trading relationship with their large customers. This lack of evidence is surprising given that almost 20% of asset selling firms in our sample have a close trading relationship with at least one major customer. Previous studies examine how the product market relationships affect firms’ financial decision and their performance. For example, Titman (1984) and Grinblatt and Titman (2002) argue that since the supplier’ liquidation decision affects its customers, the supplier has an incentive to implement the ex-ante value maximizing liquidation policy by choosing a certain capital structure. Titman and Wessels (1988) empirically examine the validity of this theory of optimal capital structure using supplier-customer relationship variables such as product uniqueness and indicators for durable goods and high tech industries. Similarly, Banerjee, Dasgupta, and Kim (2008) test stakeholder theory of capital structure using buyer-supplier relationship variables such as a durable goods industry classification and the sales fraction to customers and find that the effect of the customer-supplier relationship on leverage ratios of both customers and suppliers are stronger when the customers and the suppliers operate in durable goods industries and the suppliers’ sales to the customers are larger. Maksimovic and Titman (1991) model how the supplier’ financial health affects its incentive to produce high quality products for its customers and show that a financially 7 distressed firm has an incentive to cut cost and reduce the quality of its product, which lowers the price it receives for its product and thus decreases its profit. 3 These studies suggest that product market relationships, particularly supplier-customer relationships, can play an important role in explaining the causes and consequences of asset sales made by firms having relationships with major customers. The supplier-customer relationships are governed by non-enforceable implicit contract and thus vulnerable to new information that can have an effect on the viability of their relationship (Shleifer and Summers (1988)). Supporting this view, Brown, Fee, and Thomas (2009) find that suppliers with substantial relationship-specific investments are more negatively affected by customers’ LBO. Johnson, Kang, Masulis, and Yi (2011) argue that SEOs conveys negative news about the viability of supplier’s major customer relationships and document that SEO firms with major customer relationships experience more negative announcement returns than those without such relationships and the customers also realize negative returns around their suppliers’ SEO announcements. Because the supplier’s assets are usually dedicated to maintaining its bilateral trading relationships with large customers, its asset sale announcements should convey important information about the continued viability of its trading relationship with the customer as well as its financial and operational performance. These information contents revealed by the supplier’s 3 Several other studies show the importance of major customer-supplier relationships in firms’ investment decision and ownership distribution. For example, Fee and Thomas (2004) find that the improved productive efficiency achieved from the horizontal merger transfers to customers. Fee, Hadlock, and Thomas (2006) show that the variables related to supplier and customer relationship such as R&D intensity and sales fraction to customers systemically predict the presence of customers’ equity stake in suppliers. 8 asset sale should affect the customer’s judgment about the credibility of its supplier and thus change the values of both supplier and customer. 4 Asset sale news announced by a supplier with a large customer can convey either positive or negative information. If a supplier’s asset sale is motivated to increase its efficiency, to streamline its business by focusing on its core competitive industries, or to relax its financial constraint, it should be good news for the supplier as well as the customer since it will strengthen the supplier-customer relationship by having a more reliable and financially healthy supplier. In addition, the asset sale itself can reveal the news about the relationship with large customer. If the asset sale is designed to strengthen the customer-supplier relationship, for example to finance the relationship specific investment by selling the asset unrelated to the relationship with large customer, then the asset sale announcement can reveal information that the major trading relationship becomes strengthened. According to these arguments, asset sale announcements made by the suppliers with a large customer should have a positive valuation effect for both selling firms and their customers (positive-information-signal view). In contrast, asset sale announcements made by the suppliers with a large customer can convey negative information about the customer-supplier relationship, adversely affecting the values of both the supplier and its customer. Because supplier-customer relationships are largely governed mainly by implicit contracts, suppliers and customers bound by these implicit contracts may freely breach their partner’s trust and indulge in opportunistic behavior when it is profitable (Johnson, Kang, Masulis, and Yi (2011)). For example, if the supplier thinks that the relationship 4 The customers have strong incentives to monitor the condition of their supplier to secure the quality of the input for their product. The supply chain literature suggests that the financial credibility of suppliers plays an important role in achieving trust from the customers and the customers use the suppliers’ financial and operational performance to identify and measure the supply chain risk (Trkman and McCormack (2009)). 9 with its major customer is not beneficial anymore, it will decide to terminate or reduce the importance of its bilateral trading relationships with a major customer by selling assets specifically devoted to these relationships. This termination or reduction in the importance of the relationships caused by the asset sale will put the customer in the peril of sacrificing its product quality because of the disruption in supply chain as noted by Titman (1984) and Maksimovic and Titman (1991). Alternatively, asset sales by firms with large customers might reveal negative information about the sellers’ financial condition (Schlingemann, Stulz, and Walkling (2002)), which can have a negative effect on their major trading relationships. In these cases, the asset sale announcements by firms with large customers convey negative information about the viability and market values of major customer relationships, thus negatively affecting the values of both trading partners (negative-information-signal-view) These arguments lead to our first hypothesis: H1: If positive- (negative-) information-signal view holds, the abnormal returns around the asset sale announcements will be positive (negative) for both asset sellers and their large customers. However, the effects of asset sales on supplier and customer values can vary depending on the characteristics of the industries that the suppliers operate in and the importance of their trading relationship with a large customer. For example, Maksimovic and Titman (1991) argue that firms in durable goods and high-tech industries where products require specialized servicing commitment in the future and thus product guarantees are particularly important should maintain low leverage. The insight of this argument is that customers are reluctant to do business with highly levered suppliers in durable goods and high-tech industries because the concerns that 10 suppliers’ financial difficulties can affect their incentives to honor their implicit contracts with customers are particularly severe when they are in durable goods and high-tech industries. Thus, given that the relationship’s economic importance is especially large when the suppliers operate in durable goods and high-tech industries, large customers are likely to be more sensitive to asset sales made by suppliers in these industries. It is also possible that the level of asset specificity dedicated to a particular commercial relationship affects the asset sale announcement returns for both selling firms and their customers. Firms that produce the materials and intermediate products that involve high levels of asset specificity are likely to have the hard-to-replace relationship with their major customers, making it very costly for suppliers and customers to switch trading partners (Williamson (1985, 1991)). Titman and Wessel (1988) also argue that customers of firms producing unique or specialized products suffer relatively high costs in the event that they lose their supplier. Thus, the valuation effects of asset sale announcements are expected to be particularly pronounced if firms sell assets that are specialized in producing materials and intermediate products that are unique to their customer. All these arguments suggest that the positive (negative) effects of asset sale announcements for suppliers and their large customers are more pronounced when the suppliers are in durable goods or high-tech industries or when the asset specificity in the relationships is higher. Following Acs and Isberg (1991) and Santarelli (1991), we measure the degree of asset specificity using R&D intensity (R&D expenditures divided by total assets). This leads to our second and third hypotheses. 11 H2: If positive- (negative-) information-signal view holds, the abnormal returns around the asset sale announcements will be more positive (negative) for both asset sellers and their large customers if the suppliers are in durable goods or high-tech industries. H3: If positive- (negative-) information-signal view holds, the abnormal returns around the asset sale announcements will be more positive (negative) for both asset sellers and their large customers if the suppliers make large trading relationship-specific investments. The positive- (negative-) information-signal view also implies that asset sales made by firms with a large customers signal positive (negative) news about the viability of the post-sale major trading relationships. If asset sales are motivated by firms’ incentives to strengthen (weaken) their trading relationships with major customers, we expect that there is a high probability that the customer-supplier relationships continue (terminate) after the asset sales. We also expect that the asset selling firms with large customers increase (decrease) the sales to the customers after the asset sales and the large customers of the asset selling firms increase (decrease) the purchase from the asset selling firms. To the extent that the incentives of asset sellers and their major customers to make relationship-specific investments and to continue their relationships are particularly important when asset sellers are in durable goods or high-tech industries, we expect the predicted effects above to be more evident for asset sellers in these industries. Thus, our fourth hypothesis is: H4: If positive- (negative-) information-signal view holds, after the asset sales, there is a high probability that the customer-supplier relationships continue (terminate) and the sales to the 12 customers increase (decrease), particularly when sellers are in durable goods or high-tech industries. 3. Data We obtain our sample asset sale deals from Thomson Financial’s SDC Mergers & Acquisitions database during the 1993-2006 period. We exclude the cases in which asset sales are accompanied by mergers and acquisitions, recapitalizations, buybacks, exchange offers, 5 bankruptcy, equity carve-out, joint ventures, liquidations, and spin-offs. For each asset sale, we use the date an asset sale is first mentioned in Factiva as the initial announcement date. In addition, we search Factiva for involuntary asset sales and exclude such observations from the sample. Following Eckbo and Masulis (1992), we also exclude the asset selling firms that belong to financial (SIC 6000-6199) and utilities (SIC 4910-4940) industries where the investment decisions are highly regulated. To be included in our sample, we require that firms’ stock returns and financial data be available in CRSP and Compustat, respectively. These restrictions result in a final sample of 2,666 asset sale deals. We use Segment Customer tape of COMPUSTAT Database to identify the existence of large customers (customer that accounts for greater than 10% of a supplier’s sales). We follow the procedure similar to that of Fee and Thomas (2004) and Shantan, Dasgupta, and Kim (2008) to identify whether the large customer is public company. First, we match customer names’ abbreviations with names in CRSP/Compustat merged dataset. If the match is almost certain and distinct, we link the abbreviations to Compustat names and Gvkey. When a customer 5 We only use acquisition of asset (AA) and acquisition of certain asset (AC) as our asset sales sample. 13 abbreviation matches more than one company name in CRSP/Compustat merged dataset, we use business description for supplier and customer to further screen the match. Second, for all abbreviations whose links to names in CRSP/Compustat could not be confirmed, we use Lexis/Nexis Acedemic, Google, Factiva, and company website to determine if the abbreviation is that of the subsidiary of a public company. If the linkage is confirmed, we then match the abbreviations with its Compustat/CRSP parent firm. We find that out of 2,666 asset sellers, 513 have the relationship with large customers, of which 310 have publicly listed firms as their large customers, 150 have privately held firms as their large customers, and 53 have the governments as their large customers. 4. Asset sale announcement returns 4.1. Announcement returns for asset selling firms and large public customers We use the standard event study methodology to investigate the valuation effects of asset sale announcements. We estimate market model parameters by using days -301 to -46 relative to the announcement date. We use the CRSP-value-weighted index return as the proxy for the 6 market portfolio. The daily abnormal returns (residual of the market model) are accumulated to obtain the cumulative abnormal return (CAR) from day -t before the announcement date to day +t after the announcement date. Table III Panel A reports the abnormal returns for asset selling firms. The average CARs (-1, 1) for a full sample of asset sellers, a subsample of asset sellers with a large public customer, and a subsample of asset sellers without a large public customer are 2.13%, 2.49%, and 2.08%, respectively, all of which are significant at the 1% level. These results are consistent with those 6 Using the CRSP equally weighted index return gives qualitatively similar results. 14 in previous studies which show that asset sales announcements have a significant positive effect on selling firm value (Jain (1985), Gordon, Benson, and Kampmeyer (1984)). The difference in average CARs (-1, 1) between firms with and without a large public customer is not statistically significant. The median CARs show similar pattern. Panel B reports the CARS for the large customers of asset selling firms. We find that the mean and median CARs for all event windows considered are not statistically significant. Panel C reports the CARs for asset buyers. The average CARs (-1, 1) for a full sample of asset buyers, a subsample of asset buyers whose sellers have large public customers, and a subsample of asset buyers whose sellers do not have large pubic customers are 2.51%, 1.68%, and 2.63%, respectively, all of which are significant at the 1% level. The difference in average CARs (-1, 1) between the two subsamples is not statistically significant. The median CARs show similar pattern. 4.2. Multivariate analysis of announcement returns for asset selling firms To better understand the cross-sectional variation of CARs, we estimate an ordinary least squares (OLS) regression in which the dependent variable is the CAR (-1, 1). Following asset sale literature, we include the following variables in the regressions as the control variables: indicator for proceeds used to pay debt, indicator for the deal paid with cash only, proceeds amount scaled by total assets (relative deal size), firm age, long-term debt to total assets, Tobin’s q, stock return volatility, stock price run-up, log (equity capitalization), return on assets, and 7 asset liquidity. Our key variable of interest is the indicator variable for the existence of large customer that takes the value of one if an asset seller has a large customer prior to the asset sale 7 See, for example, Bates (2005), Moskowitz and Grinblatt (1999). 15 and zero otherwise (Large customer). We decompose Large customer into two indicators, the indicator for asset sellers whose relationships with the large customers continue more than one year after the asset sales (Continued large customer) and the indicator for asset sellers whose relationship with the large customers stop within one year after the asset sales (Stopped large customer). Another key variables of interest are the indicator variables for the importance of post-sale service commitment (Durable goods industry and Technology intense industry) and asset specificity (R&D intensity). Durable goods industry (Technology intense industry) takes the value of one if the asset seller operates in durable goods industry (technology intense industry) and zero otherwise. R&D intensity is R&D expenditures scaled by total assets. Detailed definitions of the variables used in Table 4 are described in Appendix. The results for a full sample of 2,666 asset sale deals are reported in Table 4. In model (1), we only include the variables used in previous studies. Consistent with previous studies, we find that asset sellers realize higher announcement returns when they use the proceed to repay debt, when the deal is paid with cash only, when relative deal size is larger, when the growth opportunity measured by Tobin’s q is smaller, when firm size is smaller, and when previous performance measured by returns on assets is larger. In model (2), we include Large customer to test whether the existence of large customer has significant effect on asset sale announcement returns. The coefficient estimate on Large customer is 0.008 but statistically insignificant. In model (3), to investigate further the effect of major customer relationship on asset sale announcement returns, we decompose Large customer into Continued large customer and Stopped large customer. The coefficient estimate on Continued large customer is positive and significant at the 5% level while the coefficient estimate on Stopped large customer is 16 insignificant. These results suggest that the asset sale announcement reveals the information about the potential changes in the viability of the major trading relationship and the stock market react positively only when selling firms maintain trading relationships with their existing large customers. In model (4) and (5), we add indicator variables for the importance of post-sale service commitment. We find that the coefficient estimates on Durable goods industry and Technology intense industry are not significant. In model (6), we include R&D intensity as an explanatory variable. This variable is not significantly related to CARs. Overall, the results in Table 4 show that the viability of the major trading relationship after the asset sales is the key determinant of asset sale announcement returns and other product market relationship variables do not have any statistically discernible effect on asset sale announcement returns for the full sample. To further examine the role of product market relationship variables in explaining the crosssectional variation of asset sale announcement returns, in Table 5, we divide the full sample into asset sellers with large public customers and asset sellers without large customers and estimate the regressions separately for these two subgroups. Models (1) through (8) show the results using a subsample of asset sellers with large public customers. In models (1) and (3), we find that the coefficient estimates on Durable goods industry and Technology intense industry are -0.02 and -0.021, respectively, both of which are significant at the 10% level. These estimates indicate that asset selling firms with large public customers that operate in durable goods (technology intense) industries realize 2% (2.1%) lower announcement returns than those with large public customers that operate in nondurable goods 17 (nontechnology intense) industries. Thus, the valuation effects of asset sale announcements are higher if firms sell assets that are specialized in producing materials and intermediate products that are unique to their customer. Given that the mean CAR (-1, 1) for asset selling firms with large public customers is about 2.49%, the impact of the post-sale service commitment on asset sale announcement returns is both statistically and economically, These results support the negative-information-signal view of H2. In models (2) and (4), we include the interaction term between Durable goods industry and stock return volatility and between Technology intense industry and stock return volatility, respectively. We find that the coefficient estimates on these interaction terms are negative and 8 significant at the 5% and 1% levels, respectively. Thus, the negative relation between Durable goods industry (Technology intense industry) and asset sales returns is particularly pronounced when sellers are perceived to be risky by the market. Model (5) investigates the effect of relationship-specific investment on the announcement return by including R&D intensity as an explanatory variable. We find that the coefficient estimate on R&D intensity is negative and significant at the 10% level, suggesting that asset sale announcement returns are more (negative) when the sellers make large trading relationshipspecific investments. This result supports the negative-information-signal-view of H3. In models (9) through (13), we reestimate the models (1) through (5) using a subsample of asset sellers without large customers. We find that none of the coefficient estimates on Durable goods industry, Technology intense industry, their interactions with stock return volatility, and R&D intensity is statistically significant. These results suggest that the factors that affect asset 8 We do not include industry fixed effects in these models because we use industry indicator variables as our explanatory variables. 18 sale abnormal returns are different between firms with and without large customers and that even among firms with large customers, asset sale abnormal returns vary depending on the importance of post-sale product support and guarantee, and the extent of relationship-specific investment. In models (6) through (8), we include the measures of ex-post changes in relationship strength with large customers. The results show that the asset sale announcement signals new information about the viability of the major trading relationship. Specifically in models (6) and (7), the coefficient estimates on Continued large customer and Post-sale relationship length (number of years from the asset sale year to relationship termination) are 0.047 and 0.012 respectively, both of which are significant at the 1% level. In model (8), the coefficient estimate on High drop in sales to large customer (indicator variable that takes the value of one if the asset selling firm has above the median decline in the percent of sales to the large customer from the year before the asset sale to after the asset sale and zero otherwise) is -0.034 and significant at 1% level. These results suggests that asset sellers who continue the major trading relationships with their exiting large public customers or strengthen their post-sale trading relationships realize more positive returns around the asset sale announcement dates than other asset sellers. These results support H4. 4.3. Multivariate analysis of announcement returns for large public customers of asset selling firms In this subsection we examine whether the factors that explain the CARs (-1, 1) for asset sellers with large public customers can also explain those for large public customers of asset sellers. The results are reported in Table 6. In models (1) through (4), we investigate the effect of the importance of post-sale service commitment on announcement returns for large public 19 customers. We find that the coefficient estimate on Technology intense industry is a significant 0.017 (model (3)), suggesting the large customers whose suppliers operate in technology intense industry realize 1.7% lower announcement return than those whose suppliers operate in nontechnology industries. In addition, in model (4), the coefficient estimate on the interaction term between Technology intense industry and Stock return volatility is negative and significant at the 10% level. Thus, when asset sellers operating in technology intense industry are perceived to be risky by the market participants, their large customers realize particularly lower returns. These results support the negative-information-signal-view of H2. In model (5), we include High R&D (indicator variable that takes the value of one if the ratio of R&D expenditures to total assets for asset selling firm is above the sample median and zero otherwise) to examine the effect of asset specificity of asset sellers on customer announcement return. The coefficient estimate on High R&D is -0.012, which is significant at 10% level, suggesting that the large customers of asset selling firms who invest more in R&D activity experience lower announcement return than those of asset selling firms who invest less in R&D activity investment. Thus, the large customers of asset selling firms experience more negative announcement returns when the asset sellers make intense relationship-specific investment, supporting the negative-information-signal-view of H3. In models (6) through (8), we investigate the relation between the ex-post relationship changes and customer announcement returns. Similar to the result for asset selling suppliers, the large public customers of asset selling suppliers who continue the major trading relationship after the asset sales realize higher announcement return than other large public customers (model (6)). These results are consistent with the negative-information-signal-vew of H4. However, However, we do not find any significant difference in abnormal returns between large public customers of 20 asset selling suppliers who strengthen their post-sale trading relationship and those of asset selling suppliers who do not strengthen their post-sale trading relationship, as measured by Postsale relationship length and High drop in % COGS (indicator variable that takes the value of one if the large customer experiences more drop than the sample median in the percent of the purchase from the asset selling supplier out of total cost of goods sold compared to the year right before the asset sale). In model (9), we include the measure of competition in the industry the large public customers belong to. The coefficient estimate on Herfindahl-Hirschman Index (HHI) in customer industry is a significant 0.285 at the 5% level, suggesting that the announcement returns of large public customers is higher when they have stronger monopolistic power in the industry. In summary, we find that there exist the spill-over effect of asset sale announcement down to the supply chain and that the ex-post viability of the major trading relationship, the importance of the major trading relationship, and the uniqueness of assets invested for the relationship are important determinants of asset sale announcement returns of both asset selling firms and their large public customers, which supports the negative-information-signal-view. 5. Changes in post-asset sale relationship strength In this section we examine how the strength of trading relationship between asset selling firms and their large customers changes after asset sales. The results are reported in Table 7 and Table 8. In the first three regressions in Table 7, we examine the likelihood of relationship termination after the asset sales in which the dependent variable is an indicator variable that takes the value of one if the relationships terminate in the year of the asset sale or in the year after the asset sale and zero otherwise. We find an insignificant coefficient estimate on Durable 21 goods industry in model (1). However, in model (2), the coefficient estimate on the interaction term between Durable goods industry and stock return volatility is positive and significant, indicating that risky asset sellers with large public customers who operate in durable goods industry are more likely to terminate the relationship after the asset sales. We examine the interaction term further for its sign and significance. Figure 1 shows that for most of the firms the interaction effect is positive and for the group of firms whose predicted probability is around 0.6 the interaction effects are significant. We also find that the major trading relationship is more likely to be terminated after the asset sales if the asset selling suppliers operate in technology intense industries. These results support the negative-information-signal-view. In model (4), we examine the likelihood of high drop in sales to the large customers after the asset sales. The dependent variable is an indicator variable that takes the value of one if the drop in the ratio of a supplier’s sales to the large customer to its total sales is higher than the sample median and zero otherwise (High drop in sales to customer). The coefficient estimate on the interaction term between Durable goods industry and stock return volatility is positive and significant at 1% level, suggesting that risky asset selling suppliers that operate in durable goods industries are more likely to experience the decrease in sales to the large customer after the asset sales. We examine the interaction term further for its sign and significance. Figure 2 shows that for most of the firms the interaction effect is positive and significant. This result is again largely consistent with the prediction of the negative-information-signal-view. In Table 8, we use as the dependent variable a high drop in % of COGS of customer that takes the value of one if the drop in the proportion of a supplier’s sales out of the customer’s COGS is higher than the sample median and zero otherwise. Model (2) shows that the likelihood of high drop in % of COGS of customer is higher when risky asset sellers with large public 22 customers operate in durable goods industry. We examine the interaction term further for its sign and significance. Figure 3 shows that for most of the firms the interaction effect is positive and significant. Model (3) shows that the likelihood of high drop in % of COGS of customer is higher when asset selling suppliers operate in technology intensive industries. Model (4) shows that the result in model (3) is only evident when asset selling suppliers operating in technology intensive industries are perceived risky by the market. We examine the interaction term further for its sign and significance. Figure 4 shows that for most of the firms the interaction effect is positive and for the group of firms whose predicted probability is around 0.3 the interaction effects are mostly significant. These results suggest that asset sales by firms with large customers indeed reveal the information about the deterioration in the viability of the major trading relationship, particularly when they are operate in industries where post-sale service commitment is important and at the same time they are perceived to be risky by the market. In Table 9, we estimate Cox proportional hazards model to examine the effect of asset sale on the survival of the major trading relationship. A non-asset-selling firm with a large public customer is matched to each asset selling firm with a large public customer by industry (Fama and French (1997) 48 industry classification), size (market value of equity), and number of years the relationship has been maintained prior to the asset sale. Our sample consists of 310 asset sellers with large public customers plus 310 non-asset-sellers with large public customers. The dependent variable is the number of years the relationship lasts after an asset sale. We found that sales dependence squared is positively related to the likelihood of relationship termination (i.e., the relationship lives shorter) while sales dependence and asset seller size are negatively related to the likelihood of relationship termination (i.e., the relationship survives longer), which is consistent with Fee, Hadlock, and Thomas (2006). More importantly, model (1) shows that asset 23 sellers are more likely to terminate the relationship after the asset sales. Model (2) shows the fact that a supplier operates in durable goods industry reduces the likelihood of relationship termination controlling asset sale event. Model (3) and (4) show that technology intense industry indicator variable and R&D intensity don’t have significant effect on the likelihood of relationship termination. These results suggest that an asset sale event averagely implies negative changes in the major trading relationship with large customer and the change is not driven by industry characteristics. 6. Determinant of asset sale In Table 10, we estimate logit regressions to examine the determinants that affect the asset sale decision. Our sample consists of 52,347 COMPUSTAT firm-year observations from 1993 to 2006. Using the COMPUSTAT Segment Customer database, 13,354 firms are identified as having at least one large customer and using Thomson Financial’s SDC Mergers & Acquisitions database, 5,681 firms are identified as engaging in asset sales. The dependent variable is an indicator that takes the value of one if a firm engages in asset sale in a certain year and zero otherwise. We find that the probability of asset sales increases with a firm’s leverage and stock return volatility. The probability of asset sales is also higher when the firm has negative free cash flows. In contrast, the likelihood of asset sales decreases as a firm has more investment opportunities and experiences better operating performance. These results are largely consistent with those in previous literature and support the negative-information-signal view. More important, we find that having large customer does not affect the probability of asset sale (model 1). However, a firm who terminated the major trading relationship prior to asset sales is more likely to sell its asset (model 2). This result suggests that the change in major trading relationship prior to asset sales affects a firm’s future asset sale decision. 24 The results from models (3) through (5) suggest that relationship-specific variables such as the indicator for durable goods industry, the indicator for technology intense industry, and the indicator for high R&D intensity do not affect the probability of asset sale. Models (6) through (8) confirms that after controlling for relationship-specific variables a firm that terminated the major trading relationship prior to asset sales is more likely to sell its asset, which suggests that asset sale announcement conveys negative information about the change in major trading relationships. In Table 11, we investigate the effect of the importance of post-sale service commitment on the likelihood of asset sale by dividing the sample into two subgroups: technology and nontechnology industries and durable and non-durable goods industries. In the first two columns, we compare firms in technology intense industries with firms in non-technology intense industries. The results show that firms in non-technology intense industries are more likely to sell their asset when the major trading relationships were terminated prior to asset sales than firms in technology intense industry. Although we find that the coefficient estimate on relationship termination is positive and significant in both regressions, its significance is stronger in regression using non-technology firms than in regression using technology firms. In the last two columns, we compare firms in durable goods industries with firms in nondurable goods industries. We find that while the coefficient estimate on relationship termination is positive and significant in regression using firms in non-durable goods industries, the corresponding coefficient estimate is not significant in regression using firms in durable goods industries. To the extent that selling assets in durable goods or technology intense industries involves higher switching costs than selling asset in non-durable goods or non-technology 25 intense industries, these results suggest that an asset sales are more likely to occur when firms face less switching costs in replacing the trading partners, further supporting the negativeinformation-signal view. 7. Summary and conclusion In this paper we examine the information content of asset sales announcements in associate with durable major trading relationships. Specifically, we suggest two competing views about the link between asset sales announcement returns and the importance of major customer-supplier relationships: the positive-information-signal view that posits that the asset sales announcements signal positive information about the major customer-supplier relationship and the negativeinformation-signal view that posits the asset sales announcements signal negative information about the major customer-supplier relationship. We find results that are largely consistent with the negative-information-signal view. Specifically, first, we find that the asset sale announcement affects the return of the large customer of the asset selling firm as well as that of the asset selling firms. The asset sale announcement affects negatively both asset selling firm and its large customer. Second, the negative effects are more prominent especially when the supplier operates in industries where post-sale servicing commitment is important and has made relationship intense investments. Third, a major trading relationship is likely terminated when the supplier implements asset sale and the relationship between asset selling supplier and large customer is more likely terminated or attenuated when post-sale servicing commitment is important and the asset selling supplier is risky. 26 Finally, we find that a firm sells its asset when the major trading relationship is terminated and selling relationship-related asset impose more costs when the firms operate in industries where post-sale servicing commitment is important. Overall, these results suggest that asset sale announcements by firms with large customers convey negative news in regard to the viability of the major trading relationship with the large customer and thus have spillover effects down to the major customer. 27 APPENDIX A TABLES AND FIGURES 28 Table 1 Sample Distribution by Year, Industry, and Number of firms with large public customer The sample consists of 2,666 asset sale deals reported in Thomson Financial’s SDC Mergers & Acquisitions database from 1993 to 2006. 310 out of the 2,666 asset sale deals involve at least one large public customer. The existence of the relationship with the large customer is found in COMPUSTAT Segment Customer database and the name of the customer in the database is used to identify whether the customer is public company. An asset seller is defined as having a large customer when the seller has at least one customer which accounts for 10% or more of the asset seller’s total sales. When an asset seller has more than one large customer, the customer that represents the largest amount of total sales is identified as the large customer. Industry classifications are based on the two-digit SIC. The numbers in parentheses are those of asset sellers with a large public customer. Agriculture , Forestry, and Fishing Year 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Total 0 0 2 0 0 0 0 2 0 0 0 0 0 0 4 (0) (0) (0) (0) (0) (0) (0) (0) (0) (0) (0) (0) (0) (0) (0) Mining and Construction Manufacturing Transportation, Communications, Electric, Gas, and Sanitary Services Number of Asset Selling Firms (Number of Asset Selling Firms with Large Customer) 7 (3) 72 (8) 8 (5) 12 (4) 65 (8) 11 (4) 24 (7) 81 (4) 13 (3) 11 (1) 144 (16) 15 (0) 11 (2) 103 (12) 37 (0) 23 (2) 136 (15) 27 (1) 19 (3) 136 (12) 40 (3) 11 (1) 111 (18) 29 (0) 10 (0) 97 (16) 15 (1) 14 (0) 95 (15) 29 (0) 17 (2) 95 (14) 24 (0) 18 (4) 85 (22) 26 (1) 22 (4) 87 (12) 18 (2) 24 (10) 120 (23) 12 (0) 223 (43) 1427 (195) 304 (20) 29 Wholesale and Retail Trade 4 19 16 17 22 15 21 20 18 8 18 13 11 13 215 (0) (1) (1) (3) (2) (0) (0) (2) (0) (1) (2) (1) (1) (2) (16) Table 1 (Cont’d) Finance, Insurance, and Real Estate Year 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Total 1 3 5 10 8 8 6 5 3 4 3 1 8 3 68 Services Others Total Number of Asset Selling Firms (Number of Asset Selling Firms with Large Customer) (0) 7 (0) 0 (0) 99 (0) 10 (3) 0 (0) 120 (0) 21 (0) 0 (0) 162 (0) 20 (0) 0 (0) 217 (0) 36 (3) 1 (0) 218 (0) 31 (1) 0 (0) 240 (0) 43 (4) 0 (0) 265 (0) 46 (3) 3 (0) 227 (0) 25 (5) 1 (0) 169 (0) 24 (5) 1 (0) 175 (0) 30 (2) 0 (0) 187 (0) 49 (7) 0 (0) 192 (0) 39 (0) 2 (1) 187 (0) 36 (2) 0 (0) 208 (0) 417 (35) 8 (1) 2666 30 (16) (20) (15) (20) (19) (19) (22) (24) (22) (21) (20) (35) (20) (37) (310) Table 2 Deal-, Seller-, Customer-, Buyer-Specific Characteristics The sample consists of 2,666 asset sale deals reported in Thomson Financial’s SDC Mergers & Acquisitions database from 1993 to 2006. In 310 out of the 2,666 asset sale deals, the asset sellers have at least one large public customer. The existence of the relationship with the large customer is found in COMPUSTAT Segment Customer database and the name of the customer, customer name, in the database is used to identify whether the customer is public company. An asset seller is defined as being with a large customer when the seller has at least one customer which accounts for 10% or more of the asset seller’s total sales. When an asset seller has more than one large customer, the customer that accounts for the largest amount of total sales is identified as the large customer. The sample also includes 310 non-asset selling matching firms with a large public customer. Panel B shows the sample of 310 large public corporate customers of asset selling firms and 310 large public customers of non-asset-selling matching firms. A non-asset-selling firm with a large public customer is matched to each asset selling firm with a large public customer by industry (Fama and French (1997) 48 industry classification), size (market value of equity), and number of years the relationship has been maintained prior to the asset sale. Panel C shows the sample of 1,090 public asset buyers, 141 of which are associated with the asset sellers who have a large customer. The symbols ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. p-values are reported in parentheses. 31 Table 2 Panel A. Asset sale and asset seller characteristics Sellers without Total sample of large public asset sellers customers (N=2,666): A (N=2,356): B Mean Median Mean Median Sellers with large public customers (N=310): C Mean Non-sellers with large public Test of difference customers (N=310): D (B-C) MannMedian Mean Median t-test Whitney z-test Deal characteristics Deal size ($ millions) Relative deal size Deals paid with cash only (indicator) Proceed used to pay debt (indicator) 223.88 77.00 227.25 0.15 0.05 0.15 0.04 0.16 0.06 0.77 1.00 0.77 1.00 0.76 1.00 0.00 0.18 0.00 67.95 0.00 32 1.55 (0.12) -3.57 *** (0.00) 0.37 (0.72) -1.74 * -1.74 * (0.08) 0.22 0.95 (0.34) -0.50 (0.62) 0.37 (0.72) 0.18 78.50 198.26 (0.08) Test of difference (D-C) t-test Wilcoxon z-test Table 2 (Cont’d) Panel A. Asset sale and asset seller characteristics (Cont’d) Sellers without Sellers with Total sample of large public large public asset sellers customers customers (N=2,666): A (N=2,356): B (N=310): C Mean Median Mean Median Mean Median Non-sellers with large public customers (N=310): D Mean Median Test of difference (B-C) Mannt-test Whitney z-test Test of difference (D-C) t-test Wilcoxon z-test Seller characteristics Large customer (indicator) Continued large customer (indicator) Stopped large customer (indicator) Asset liquidity Durable goods industry (indicator) Technology intense industry (indicator) 0.19 0.00 1.00 1.00 0.10 0.00 0.56 1.00 0.09 0.00 0.44 0.00 0.11 0.04 0.12 0.04 0.06 0.03 0.21 0.00 0.19 0.00 0.35 0.00 0.37 0.00 -6.58 *** -6.53 *** 0.42 (0.00) (0.00) (0.68) 0.42 (0.68) 0.15 0.00 0.14 0.00 0.23 0.00 0.25 0.00 -4.28 *** -4.26 *** 0.66 (0.00) (0.00) (0.51) 0.66 (0.51) 33 1.00 1.00 1.51 (0.13) 2.81 *** (0.01) Table 2 (Cont’d) Panel A. Asset sale and asset seller characteristics (Cont’d) Sellers without Sellers with Total sample of large public large public asset sellers customers customers (N=2,666): A (N=2,356): B (N=310): C Mean Median Mean Median Mean Non-sellers with large public Test of difference Test of difference customers (N=310): D (B-C) (D-C) MannWilcoxon Median Mean Median t-test Whitney t-test z-test z-test Seller characteristics (Cont’d) Total asset ($ millions) Market capitalization ($ millions) Leverage 12,281.03 2,316.66 12,941.89 2,447.08 7,258.47 1,503.88 1,644.26 324.79 2.98 *** 4.09 *** -7.02 *** -8.77 *** (0.00) (0.00) (0.00) (0.00) 13,696.40 1,603.25 14,186.70 1,757.03 9,970.16 864.27 1,605.92 379.04 1.93 * (0.05) 0.27 0.23 0.27 0.23 0.25 0.24 0.20 0.15 1.86 * (0.06) Tobin's Q 1.58 1.27 1.59 1.28 1.46 1.21 1.97 1.43 2.05 ** (0.04) R&D intensity 0.03 0.00 0.03 0.00 0.05 0.01 0.05 0.00 -5.82 *** (0.00) Stock price 0.11 0.06 0.11 0.07 0.09 0.01 0.14 0.04 0.64 run-up (0.53) Stock return 0.03 0.02 0.03 0.02 0.04 0.03 0.04 0.03 -4.36 *** volatility (0.00) ROA 0.10 0.11 0.11 0.11 0.08 0.11 0.09 0.13 3.24 *** (0.00) Firm age 25.68 21.00 26.36 23.00 20.47 14.00 17.25 11.00 5.62 *** (years) (0.00) 34 4.51 *** (0.00) 1.51 (0.13) 2.41 ** (0.02) -5.81 *** (0.00) 2.48 ** (0.01) -5.03 *** (0.00) 2.55 ** (0.01) 5.49 *** (0.00) -5.62 *** (0.00) -3.21 *** (0.00) 4.69 *** (0.00) 0.17 (0.86) 0.90 (0.37) 0.62 (0.53) 0.65 (0.51) -2.70 *** (0.01) -5.16 *** (0.00) -4.44 *** (0.00) 4.84 *** (0.00) -1.22 (0.22) 1.49 (0.14) 1.25 (0.21) 3.06 *** (0.00) -2.14 ** (0.03) Table 2 (Cont’d) Panel B: Characteristics of major supplier-customer relationships Large public customers of Large public customers of Sellers (N=224): A non-sellers (N=224): B Mean Median Mean Median Customer characteristics Customer total assets ($ millions) 47,025.54 28,728.00 43,809.21 26,676.67 Customer market capitalization ($ millions) 69,418.71 34,897.14 68,965.91 32,148.37 Customer Tobin's Q 2.08 1.68 2.14 1.69 Customer leverage 0.19 0.19 0.19 0.19 Customer ROA 0.15 0.15 0.15 0.15 Customer stock price run-up 0.16 0.08 0.10 0.07 Customer stock return volatility 0.02 0.02 0.02 0.02 % of sales to customer 0.21 0.16 0.26 0.19 % of COGS from supplier 0.06 0.01 0.02 0.00 Pre-asset-sale relationship length (years) 3.68 3.00 3.31 3.00 Post-asset-sale relationship length (years) 2.01 1.00 2.14 1.50 Test of difference (B-A) t-test Wilcoxon z-test -0.73 (0.47) -0.06 (0.95) 0.36 (0.72) -0.67 (0.50) 0.74 (0.46) -1.35 (0.18) 1.24 (0.21) -1.04 (0.30) -0.69 (0.49) -0.02 (0.99) -0.55 (0.58) 0.22 (0.83) -0.71 (0.48) 0.91 (0.36) 2.81 *** (0.01) -1.67 * (0.10) -1.22 (0.22) 0.60 (0.55) 3.42 *** (0.00) -5.25 *** (0.00) 0.07 (0.95) 0.33 (0.74) Relationship characteristics 35 Table 2 (Cont’d) Panel C: Asset buyer and deal characteristics Total Buyers (N=1,090): A Buyers whose sellers Buyers whose sellers have don't have large public large public customers customers (N=141): B (N=949): C Mean Median Mean Median Mean Median 226.96 84.00 171.87 64.00 235.14 87.90 Relative deal size 0.15 0.04 0.15 0.06 0.14 0.04 Deals paid with cash only (indicator) 0.73 1.00 0.72 1.00 0.73 1.00 Proceed used to pay debt (indicator) 0.17 0.00 0.22 0.00 0.16 0.00 Large customer (indicator) 0.21 0.00 1.00 1.00 0.09 0.00 Continued large customer (indicator) 0.11 0.00 0.58 1.00 0.04 0.00 Stopped large customer (indicator) 0.10 0.00 0.42 0.00 0.05 0.00 Asset liquidity 0.08 0.04 0.06 0.03 0.08 0.04 Test of difference (C-B) Wilcoxon t-test z-test Deal characteristics Deal size ($ millions) 1.44 (0.15) -0.19 (0.85) 0.43 (0.67) -1.77 * (0.08) 2.18 ** (0.03) -2.75 *** (0.01) 0.43 (0.67) -1.77 * (0.08) Seller characteristics 36 -37.34 *** -24.73 *** (0.00) (0.00) *** -23.72 -19.27 *** (0.00) (0.00) *** -14.79 -13.50 *** (0.00) (0.00) 1.18 1.77 * (0.24) (0.08) Table 2 (Cont’d) Panel C: Asset buyer and deal characteristics (Cont’d) Total Buyers (N=1,090): A Buyers whose sellers Buyers whose sellers have don't have large public large public customers customers (N=141): B (N=949): C Mean Median Mean Median Mean Median Durable goods industry (indicator) 0.23 0.00 0.38 0.00 0.21 0.00 Technology intense industry (indicator) 0.18 0.00 0.27 0.00 0.16 0.00 Stock return volatility 0.03 0.03 0.03 0.03 0.03 0.03 Test of difference (C-B) Wilcoxon t-test z-test Seller characteristics (Cont’d) -4.44 *** (0.00) -3.09 *** (0.00) -4.03 *** (0.00) -4.40 *** (0.00) -3.08 *** (0.00) -4.48 *** (0.00) 2.51 ** (0.01) 2.47 ** (0.01) 0.56 (0.58) 2.44 ** (0.01) -4.03 *** (0.00) 3.70 *** (0.00) 4.48 *** (0.00) 0.41 (0.68) 2.87 *** (0.00) -4.48 *** (0.00) Buyer characteristics Buyer total assets ($ millions) 6,137.44 1,056.39 5,148.67 1,462.79 6,284.34 1,015.47 Buyer market capitalization ($ millions) 6,903.44 1,237.91 7,544.41 1,598.08 6,808.21 1,125.12 Buyer leverage 0.23 0.21 0.21 0.19 0.24 0.21 Buyer Tobin's Q 2.06 1.58 2.13 1.49 2.05 1.59 Buyer stock return volatility 0.03 0.03 0.03 0.03 0.03 0.03 37 Table 2 (Cont’d) Panel C: Asset buyer and deal characteristics (Cont’d) Total Buyers (N=1,090): A Buyers whose sellers Buyers whose sellers have don't have large public large public customers customers (N=141): B (N=949): C Mean Median Mean Median Mean Median Buyer R&D intensity 0.03 0.00 0.03 0.01 0.03 0.00 Buyer stock price run-up 0.26 0.14 0.12 0.09 0.28 0.15 Buyer ROA 0.14 0.14 0.14 0.15 0.14 0.14 Buyer firm age (years) 18.31 11.00 20.37 13.00 18.00 11.00 Test of difference (C-B) Wilcoxon t-test z-test Buyer characteristics (Cont’d) 38 -3.18 *** (0.00) 1.75 * (0.08) 4.49 *** (0.00) 4.72 *** (0.00) -2.53 ** (0.01) 3.60 *** (0.00) 3.91 *** (0.00) 4.78 *** (0.00) Table 3 Cumulative abnormal returns (CARs) for asset selling firms, their large public customers, and asset buying firms The sample consists of 2,666 asset sale deals reported in Thomson Financial’s SDC Mergers & Acquisitions database from 1993 to 2006. 310 out of the 2,666 asset sale deals involve at least one large public customer. The existence of the relationship with the large customer is found in COMPUSTAT Segment Customer database and the name of the customer, customer name, in the database is used to identify whether the customer is public company. An asset seller is defined as operating with a large customer when the seller has at least one customer which accounts for 10% or more of the asset seller’s total sales. When an asset seller has more than one large customer, the customer that accounts for the largest amount of total sales is identified as the large customer. The sample also includes 310 non-asset selling matching firms with a large public customer. Panel B shows the sample of 310 large public corporate customers of asset sellers and Panel C shows the sample of 1,090 asset buyers. 141 of the asset buyers are involved with the asset sellers who have large public customers. The abnormal returns are calculated using the market model estimated with returns from days -301 to -46 relative to the asset sale announcement day. The value weighted CRSP index return is used as the proxy for the market portfolio in calculating the daily abnormal returns (AR). The cumulative abnormal return (CAR)s are calculated by accumulating the daily abnormal returns from day -10, -5, -1 before the announcement date to day 0, +1, +5, +10 after the announcement date. The symbols ***, **, and * denote values are significantly different from zero at the 1%, 5%, and 10% levels, respectively. p-values are in parentheses. 39 Table 3 Panel A: Abnormal returns for asset selling firms with and without a large public customer Asset sellers Asset sellers Asset Sellers with without Test of Difference Total large public large public customers customers (N=2,666): A (N=310): B (N=2,356): C (C-B) Mann Mean Median Mean Median Mean Median t-test -Whitney z-test 0.44%** (0.02) AR(0) 1.43%*** (0.00) CAR 1.87%*** (-1,0) (0.00) CAR 2.13%*** (-1,+1) (0.00) CAR 2.60%*** (-5,+5) (0.00) CAR 2.91%*** (-10,+10) (0.00) AR(-1) 0.00% 0.11% (0.19) (0.67) 0.34%*** 2.45%*** (0.00) (0.00) 0.45%*** 2.56%*** (0.00) (0.00) *** 0.73% 2.49%*** (0.00) (0.00) *** 0.76% 2.58%*** (0.00) (0.00) *** 1.10% 3.33%*** (0.00) (0.01) -0.22% (0.47) 0.58%*** (0.00) 0.57%*** (0.00) 1.07%*** (0.00) 0.62%** (0.02) -0.03%* (0.09) 40 0.48%** (0.02) 1.30%*** (0.00) 1.78%*** (0.00) 2.08%*** (0.00) 2.60%*** (0.00) 2.85%*** (0.00) 0.02%* (0.09) 0.32%*** (0.00) 0.45%*** (0.00) 0.69%*** (0.00) 0.81%*** (0.00) 1.20%*** (0.00) 0.65 (0.52) -2.56 ** (0.01) -1.08 (0.28) -0.50 (0.62) 0.02 (0.99) -0.36 (0.72) 1.40 (0.16) -2.37 ** (0.02) -0.86 (0.39) -0.78 (0.43) -0.11 (0.91) 0.48 (0.63) Table 3 (Cont’d) Panel B: Cumulative abnormal returns for large public customers of asset sellers Large public customers (N=310) Mean t-test Median Sign-rank test AR(-1) -0.01% -0.08 -0.20% -1.23 (0.93) (0.22) AR(0) 0.01% 0.11 -0.05% -0.52 (0.91) (0.60) CAR(-1,0) 0.00% 0.02 -0.10% -0.88 (0.99) (0.38) CAR(-1,+1) -0.18% -0.82 -0.08% -0.91 (0.41) (0.36) CAR(-5,+5) -0.30% -0.80 -0.44% -0.73 (0.42) (0.47) CAR(-10,+10) -0.22% -0.41 -0.63% -0.75 (0.68) (0.45) 41 Table 3 (Cont’d) Panel C: Cumulative Abnormal Return for Asset buyers Asset buyers Asset buyers of sellers with of sellers without Total sample large public large public customers customers (N=1,090): A (N=141): B (N=949): C Mean AR(-1) 0.09% (0.34) AR(0) 1.39%*** (0.00) CAR 1.48%*** (-1,0) (0.00) CAR 2.51%*** (-1,1) (0.00) CAR 2.49%*** (-5,5) (0.00) CAR 2.35%*** (-10,10) (0.00) Median Mean -0.03% 0.30% (0.98) (0.39) 0.40%*** 1.32%*** (0.00) (0.00) 0.51%*** 1.62%*** (0.00) (0.00) 0.98%*** 1.68%** (0.00) (0.02) 0.93%*** 1.88%** (0.00) (0.03) 0.65%*** 1.48% (0.00) (0.18) Test of Difference Median Mean Median t-test 0.24% (0.42) 0.79%*** (0.00) 0.85%*** (0.00) 0.96%*** (0.00) 0.79%** (0.04) 1.21% (0.39) 0.06% (0.53) 1.40%*** (0.00) 1.46%*** (0.00) 2.63%*** (0.00) 2.59%*** (0.00) 2.48%*** (0.00) -0.05% (0.80) 0.36%*** (0.00) 0.44%*** (0.00) 0.98%*** (0.00) 0.95%*** (0.00) 0.61%*** (0.00) (0.39) 0.15 (0.88) -0.29 (0.77) 1.08 (0.28) 0.61 (0.54) 0.71 (0.48) 42 C-B MannWhitney z-test -0.80 (0.42) -1.38 (0.17) -1.34 (0.18) -0.33 (0.75) 0.11 (0.92) 0.51 (0.61) Table 4 OLS regressions of cumulative abnormal returns (CARs) for asset selling firms around the asset sale announcement date on explanatory variable The sample consists of 2,666 asset sale deals reported in Thomson Financial’s SDC Mergers & Acquisitions database from 1993 to 2006. 310 out of the 2,666 asset sale deals are announced by the asset sellers with at least one large public customer. The existence of the relationship with the large customer is found in COMPUSTAT Segment Customer database and the name of the customer, customer name, in the database is used to identify whether the customer is public company. An asset seller is defined as being with a large customer when the seller has at least one customer which accounts for 10% or more of the asset seller’s total sales. When an asset seller has more than one large customer, the customer that accounts for the largest amount of total sales is identified as the large customer. The dependent variable is the CAR (-1, 1) for asset selling firms. The abnormal returns are calculated using the market model estimated with returns from days -301 to -46 relative to the asset sale announcement day. The value weighted CRSP index return is used as the proxy for the market portfolio in calculating the daily abnormal returns. The CAR (-1, 1) is calculated by accumulating the daily abnormal returns from day -1 before the announcement date to day +1 after the announcement date. All regressions include indicator variables for each Fama and French (1997) 48 industry and year except where noted. p-values reported in parentheses are heteroscedasticity robust and clustered by industry. Appendix summarizes the definitions of the variables used in table. The symbols ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. 43 Table 4 (1) Large customer (indicator) (2) Age of the seller Tobin's Q Log(Market capitalization) Leverage Return on assets Stock price runup Stock return volatility Asset liquidity Industry fixed effect Year fixed effect N R-squared (4) (5) (6) 0.008 (0.115) 0.017 ** 0.014 * (0.039) (0.061) -0.001 -0.004 (0.852) (0.579) -0.004 (0.197) Continued large customer (indicator) Stopped large customer (indicator) Durable goods industry (indicator) Technology intense industry (indicator) R&D intensity Proceed used to pay debt (indicator) Deals paid with cash only (indicator) Relative deal size (3) 0.013 * (0.070) -0.004 (0.550) 0.018 ** (0.040) 0.000 (0.953) 0.002 (0.490) 0.015 ** (0.013) 0.005 (0.154) 0.037 *** (0.000) 0.000 *** (0.002) -0.006 ** (0.019) -0.004 *** (0.007) -0.012 (0.355) 0.070 *** (0.004) -0.018 *** (0.000) 0.529 *** (0.007) -0.013 (0.424) 0.015 ** (0.013) 0.005 (0.174) 0.038 *** (0.000) 0.000 *** (0.001) -0.006 ** (0.018) -0.004 *** (0.010) -0.011 (0.369) 0.070 *** (0.004) -0.017 *** (0.000) 0.524 *** (0.008) -0.013 (0.421) 0.015 ** (0.011) 0.005 (0.179) 0.038 *** (0.000) 0.000 *** (0.000) -0.006 ** (0.023) -0.004 *** (0.009) -0.011 (0.370) 0.071 *** (0.003) -0.018 *** (0.000) 0.524 *** (0.008) -0.012 (0.448) 0.015 ** (0.013) 0.004 (0.183) 0.039 *** (0.000) 0.000 ** (0.011) -0.005 ** (0.033) -0.003 ** (0.013) -0.008 (0.471) 0.067 *** (0.004) -0.019 *** (0.000) 0.550 *** (0.003) -0.010 (0.457) 0.015 ** (0.012) 0.005 (0.158) 0.039 *** (0.000) 0.000 ** (0.014) -0.005 ** (0.036) -0.003 ** (0.014) -0.006 (0.600) 0.067 *** (0.005) -0.019 *** (0.000) 0.523 *** (0.006) -0.010 (0.479) -0.067 (0.131) 0.015 ** (0.012) 0.004 (0.185) 0.039 *** (0.000) 0.000 *** (0.001) -0.005 ** (0.048) -0.004 ** (0.011) -0.012 (0.351) 0.064 *** (0.008) -0.017 *** (0.000) 0.532 *** (0.007) -0.012 (0.453) Yes Yes 2,666 0.108 Yes Yes 2,666 0.109 Yes Yes 2,666 0.112 No Yes 2,666 0.095 No Yes 2,666 0.095 Yes Yes 2,666 0.112 44 Table 5 OLS regressions of cumulative abnormal returns (CARs) for asset selling firms with large public customers and without large customers The sample consists of 310 asset sales by the firms with a large public corporate customer and 2,153 asset sales by the firms without a large customer, reported in Thomson Financial’s SDC Mergers & Acquisitions database in the 1993–2006 period. We use the COMPUSTAT Segment Customer database field, customer name, to identify whether asset sellers have large corporate customers. An asset seller is defined as having a large customer when a customer represents 10% or more of an asset seller’s total sales. When an asset seller has multiple large customers, the customer purchasing the largest amount is identified as the large customer. To determine whether a customer is publicly traded or privately held, we match the names in the COMPUSTAT Segment Customer database with those in the COMPUSTAT database. The dependent variable is the CAR (-1, 1) for asset selling firms. The abnormal returns are calculated using the market model. The market model is estimated using returns from days -301 to -46 relative to the asset sale announcement date. The value weighted CRSP index return is used as the proxy for the market portfolio. The CAR (-1, 1) is calculated by accumulating the daily abnormal returns from day -1 before the announcement date to day +1 after the announcement date. All regressions include indicator variables for each Fama and French (1997) 48 industry and year except where noted. p-values reported in parentheses are heteroscedasticity robust and clustered by industry. Appendix summarizes the definitions of the variables used in table. The symbols ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. Asset sellers with large public customers (1) (2) (3) (4) (5) (6) (7) (8) Durable goods industry (indicator) Durable goods industry (indicator) * Stock return volatility Technology intense industry (indicator) -0.020 * (0.074) 0.028 (0.266) -1.290 ** (0.041) -0.021 * 0.046 ** (0.089) (0.021) -1.621 *** (0.003) Technology intense industry (indicator) * Stock return volatility R&D intensity -0.192 * (0.086) 45 Table 5 (Cont’d) (1) Asset sellers with large public customers (Cont’d) (2) (3) (4) (5) (6) (7) 0.047 *** (0.003) Continued large customer (indicator) 0.012 *** (0.000) Post-sale relationship length High drop in sales to large customer (indicator) Stock return volatility Proceed used to pay debt (indicator) Deals paid with cash only (indicator) Age of the seller Relative deal size Tobin's Q Log (Market capitalization) Leverage Return on assets (8) 0.581 (0.238) 0.020 (0.175) 0.019 (0.120) 0.000 (0.385) 0.018 (0.543) -0.003 (0.478) -0.010 ** (0.012) -0.077 ** (0.041) 0.014 (0.719) 0.886 * (0.082) 0.021 (0.151) 0.019 (0.121) 0.000 (0.574) 0.018 (0.576) -0.003 (0.574) -0.010 ** (0.018) -0.079 ** (0.038) 0.021 (0.633) 46 0.687 (0.176) 0.019 (0.185) 0.019 (0.111) 0.000 (0.490) 0.020 (0.508) -0.003 (0.519) -0.009 ** (0.023) -0.075 ** (0.034) 0.017 (0.664) 0.905 * (0.070) 0.016 (0.255) 0.018 (0.124) 0.000 (0.442) 0.016 (0.608) -0.002 (0.654) -0.010 ** (0.016) -0.080 ** (0.023) 0.014 (0.740) 0.500 (0.379) 0.023 (0.116) 0.023 * (0.070) 0.000 (0.707) 0.023 (0.395) -0.005 (0.384) -0.009 * (0.080) -0.061 (0.184) 0.007 (0.881) 0.537 (0.296) 0.028 * (0.052) 0.024 * (0.056) 0.000 (0.516) 0.033 (0.346) -0.013 ** (0.034) -0.010 ** (0.045) -0.053 (0.170) 0.054 (0.102) 0.640 (0.222) 0.030 ** (0.042) 0.023 ** (0.038) 0.000 (0.455) 0.031 (0.327) -0.008 (0.216) -0.011 ** (0.022) -0.049 (0.190) 0.030 (0.388) -0.034 *** (0.001) 0.663 (0.210) 0.027 * (0.093) 0.025 ** (0.028) 0.000 (0.584) 0.029 (0.323) -0.010 * (0.052) -0.009 * (0.084) -0.058 (0.139) 0.041 (0.176) Table 5 (Cont’d) (1) Stock price runup Asset liquidity Industry fixed effect Year fixed effect N R-squared Asset sellers with large public customers (Cont’d) (2) (3) (4) (5) (6) (7) (8) 0.001 (0.932) -0.057 (0.455) -0.003 (0.768) -0.057 (0.403) 0.002 (0.903) -0.064 (0.371) -0.003 (0.835) -0.066 (0.370) 0.001 (0.934) -0.011 (0.905) 0.004 (0.777) -0.016 (0.859) 0.004 (0.761) 0.014 (0.878) 0.004 (0.765) -0.014 (0.867) No Yes 310 0.243 No Yes 310 0.256 No Yes 310 0.241 No Yes 310 0.258 Yes Yes 310 0.352 Yes Yes 310 0.385 Yes Yes 310 0.395 Yes Yes 310 0.367 47 Table 5 (Cont’d) Asset sellers without large customers (10) (11) (12) (9) Durable goods industry (indicator) -0.002 (0.694) (13) 0.010 (0.156) -0.349 (0.108) Durable goods industry (indicator) * Stock return volatility Technology intense industry (indicator) 0.002 (0.646) Technology intense industry (indicator) * Stock return volatility R&D intensity -0.004 (0.820) 0.168 (0.717) -0.073 (0.142) Continued large customer (indicator) Post-sale relationship length High drop in sales to large customer (indicator) Stock return volatility Proceed used to pay debt (indicator) Deals paid with cash only (indicator) Age of the seller 0.541 ** (0.014) 0.012 ** (0.029) -0.001 (0.860) 0.000 (0.105) 48 0.626 *** (0.007) 0.012 ** (0.031) -0.001 (0.863) 0.000 (0.104) 0.527 ** (0.018) 0.012 ** (0.028) 0.000 (0.901) 0.000 (0.105) 0.500 ** (0.035) 0.012 ** (0.029) 0.000 (0.885) 0.000 (0.113) 0.558 ** (0.013) 0.012 ** (0.030) 0.000 (0.969) 0.000 ** (0.018) Table 5 (Cont’d) Asset sellers without large customers (Cont’d) (10) (11) (12) (9) Relative deal size Tobin's Q Log (Market capitalization) Leverage Return on assets Stock price runup Asset liquidity Industry fixed effect Year fixed effect N R-squared (13) 0.034 *** (0.001) -0.006 ** (0.014) -0.002 (0.135) -0.004 (0.714) 0.070 *** (0.009) -0.022 *** (0.000) -0.003 (0.834) 0.034 *** (0.001) -0.006 ** (0.014) -0.002 (0.153) -0.005 (0.680) 0.070 *** (0.009) -0.022 *** (0.000) -0.002 (0.861) 0.034 *** (0.001) -0.007 ** (0.014) -0.002 (0.127) -0.003 (0.762) 0.070 *** (0.009) -0.022 *** (0.000) -0.003 (0.840) 0.034 *** (0.001) -0.007 ** (0.015) -0.002 (0.130) -0.003 (0.763) 0.070 *** (0.007) -0.022 *** (0.000) -0.003 (0.830) 0.035 *** (0.001) -0.006 ** (0.021) -0.002 (0.120) -0.011 (0.452) 0.068 *** (0.008) -0.021 *** (0.000) -0.007 (0.646) No Yes 2,153 0.082 No Yes 2,153 0.083 No Yes 2,153 0.082 No Yes 2,153 0.082 Yes Yes 2,153 0.101 49 Table 6 OLS regressions of cumulative abnormal returns (CARs) for large public corporate customers of asset selling firms around the announcement date on explanatory variables The sample consists of 247 large public corporate customers of asset selling firms, reported in Thomson Financial’s SDC Mergers & Acquisitions database in the 1993–2006 period. We use the COMPUSTAT Segment Customer database field, customer name, to identify whether asset sellers have large corporate customers. An asset seller is defined as having a large customer when a customer represents 10% or more of an asset seller’s total sales. When an asset seller has multiple large customers, the customer purchasing the largest amount is identified as the large customer. To determine whether a customer is publicly traded or privately held, we match the names in the COMPUSTAT Segment Customer database with those in the COMPUSTAT database. The dependent variable is the CAR (-1, 1) for large public customers of asset selling firms. The abnormal returns are calculated using the market model. The market model is estimated using returns from days -301 to -46 relative to the asset sale announcement date. The value weighted CRSP index return is used as the proxy for the market portfolio. The CAR (-1, 1) is calculated by accumulating the daily abnormal returns from day -1 before the announcement date to day +1 after the announcement date. All regressions include indicator variables for each Fama and French (1997) 48 industry and year except where noted. p-values reported in parentheses are heteroscedasticity robust and clustered by industry. Appendix summarizes the definitions of the variables used in table. The symbols ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. (1) Durable goods industry Durable goods industry * Stock return volatility Technology intense industry Technology intense industry * Stock return volatility Stock return volatility -0.010 (0.332) (2) (3) (4) (5) -0.004 (0.872) -0.131 (0.716) -0.095 (0.510) -0.017 *** 0.009 (0.003) (0.518) -0.562 * (0.096) 0.063 (0.592) -0.012 * (0.087) High R&D (indicator) 50 (6) (7) (8) (9) Table 6 (Cont’d) (1) (2) (3) (4) (5) (6) (7) 0.001 (0.496) High drop in %COGS (indicator) HHI in customer industry Return on asset Tobin's Q Leverage Customer Tobin's Q Log (Customer market cap) Customer leverage (9) 0.009 * (0.092) Continued large customer (indicator) Post-sale relationship length Proceed used to pay debt (indicator) Deals paid with cash only (indicator) Relative deal size (8) -0.003 (0.591) -0.002 (0.830) 0.007 (0.155) 0.013 (0.545) 0.028 (0.236) 0.004 (0.223) 0.004 (0.762) -0.003 (0.423) 0.008 *** (0.004) 0.061 *** (0.009) -0.002 (0.834) 0.007 (0.148) 0.015 (0.535) 0.020 (0.351) 0.005 (0.184) 0.007 (0.667) -0.003 (0.410) 0.008 *** (0.007) 0.060 ** (0.013) -0.003 (0.726) 0.007 (0.146) 0.012 (0.609) 0.036 (0.177) 0.004 (0.225) 0.004 (0.657) -0.003 (0.356) 0.009 *** (0.004) 0.049 ** (0.017) -0.004 (0.622) 0.007 (0.199) 0.011 (0.636) 0.016 (0.491) 0.004 (0.241) 0.000 (0.995) -0.003 (0.314) 0.008 *** (0.001) 0.053 *** (0.009) 51 -0.004 (0.596) 0.006 (0.195) 0.011 (0.638) 0.020 (0.420) 0.005 (0.166) 0.003 (0.738) -0.003 (0.403) 0.008 *** (0.005) 0.057 *** (0.006) -0.002 (0.741) 0.008 (0.166) 0.016 (0.513) 0.032 (0.219) 0.003 (0.373) 0.011 (0.281) -0.003 (0.373) 0.008 *** (0.008) 0.056 *** (0.009) -0.003 (0.732) 0.008 (0.211) 0.013 (0.567) 0.016 (0.499) 0.004 (0.310) 0.006 (0.523) -0.003 (0.401) 0.007 *** (0.007) 0.053 ** (0.013) -0.003 (0.714) 0.008 (0.163) 0.014 (0.559) 0.030 (0.246) 0.004 (0.238) 0.009 (0.320) -0.003 (0.460) 0.008 *** (0.006) 0.055 ** (0.013) 0.285 ** (0.045) -0.003 (0.669) 0.007 (0.248) 0.012 (0.585) 0.013 (0.603) 0.004 (0.293) 0.003 (0.669) -0.002 (0.570) 0.007 *** (0.009) 0.050 ** (0.021) Table 6 (Cont’d) (1) Customer ROA Customer stock price runup Customer stock volatility Asset liquidity Industry fixed effect Year fixed effect N R-squared (2) (3) (4) 0.025 (0.670) -0.015 (0.166) 1.141 ** (0.049) -0.082 ** (0.050) 0.027 (0.646) -0.015 (0.163) 1.212 ** (0.031) -0.081 * (0.052) 0.020 (0.709) -0.014 (0.167) 1.240 ** (0.017) -0.082 ** (0.033) 0.020 (0.723) -0.016 (0.117) 1.428 *** (0.004) -0.087 ** (0.024) Yes Yes 247 0.266 Yes Yes 247 0.269 Yes Yes 247 0.276 Yes Yes 247 0.274 52 (5) 0.022 (0.682) -0.015 (0.162) 1.068 * (0.052) -0.089 ** (0.038) Yes Yes 247 0.265 (6) (7) (8) 0.020 (0.735) -0.013 (0.199) 1.140 ** (0.044) -0.078 * (0.054) 0.027 (0.668) -0.015 (0.161) 1.123 * (0.053) -0.078 ** (0.048) 0.022 (0.708) -0.014 (0.186) 1.095 * (0.055) -0.078 ** (0.042) Yes Yes 247 0.27 Yes Yes 247 0.241 Yes Yes 247 0.259 (9) 0.016 (0.795) -0.016 (0.122) 1.029 * (0.085) -0.078 ** (0.050) Yes Yes 247 0.253 Table 7 Logit regressions of relationship termination after asset sales by firms with large public customers. The sample consists of 310 asset selling firms with large public corporate customers, reported in Thomson Financial’s SDC Mergers & Acquisitions database in the 1993–2006 period. We use the COMPUSTAT Segment Customer database field, customer name, to identify whether asset sellers have large corporate customers. An asset seller is defined as having a large customer when a customer represents 10% or more of an asset seller’s total sales. When an asset seller has multiple large customers, the customer purchasing the largest amount is identified as the large customer. To determine whether a customer is publicly traded or privately held, we match the names in the COMPUSTAT Segment Customer database with those in the COMPUSTAT database. The dependent variable in models (1) through (3) is the indicator variable for relationship termination which takes the value of one if the customer-supplier relationship is terminated in the year of asset sale or in the year right after the asset sale. The dependent variable in model (4) is the indicator variables for high drop in sales to customer that takes the value of one if the drop in the proportion of a supplier’s sales to the large customer is higher than the sample median and zero otherwise. All regressions include year indicator variables for year. pvalues reported in parentheses are heteroscedasticity robust and clustered by industry. Appendix summarizes the definitions of the variables used in table. The symbols ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. Appendix summarizes the definitions of the variables used in table. The symbols ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. 53 Table 7 (1) (2) (3) Termination Termination Termination (Indicator) (Indicator) (Indicator) Durable goods industry (indicator) Durable goods industry (indicator) * Stock return volatility Technology intense industry (indicator) Stock return volatility Proceed used to pay debt (indicator) Deals paid with cash only (indicator) Age of the seller Relative deal size Tobin's Q Log (Market capitalization) Leverage Return on assets Stock price runup Asset liquidity Year fixed effect N -0.412 (0.347) -1.919 ** (0.012) 39.530 ** (0.042) -1.946 *** (0.004) 46.690 *** (0.001) 15.800 (0.318) 0.040 (0.895) 0.274 (0.437) -0.006 (0.616) 1.082 (0.210) -0.480 *** (0.001) 0.060 (0.540) 0.222 (0.806) 1.454 (0.108) -0.046 (0.856) 1.761 (0.382) 8.158 (0.577) 0.028 (0.930) 0.295 (0.415) -0.003 (0.818) 1.097 (0.187) -0.529 *** (0.000) 0.066 (0.479) 0.259 (0.779) 1.381 (0.148) 0.097 (0.730) 1.859 (0.304) 1.427 *** (0.002) 5.771 (0.680) 0.092 (0.745) 0.282 (0.463) 0.003 (0.846) 1.266 (0.123) -0.592 *** (0.004) -0.006 (0.942) 1.500 (0.115) 1.360 (0.112) 0.037 (0.886) 1.535 (0.413) Yes 310 Yes 310 Yes 310 54 (4) high drop in sales to customer (Indicator) 10.940 (0.367) 0.253 (0.513) 0.571 ** (0.048) 0.006 (0.496) 1.834 ** (0.041) -0.376 (0.110) 0.049 (0.719) -0.844 (0.434) -0.523 (0.657) 0.156 (0.595) 2.292 (0.428) Yes 310 Table 8 Logit regressions of change in percentage of COGS of large public corporate customers of asset selling firms after the asset sales The sample consists of 247 large public corporate customers of asset selling firms, reported in Thomson Financial’s SDC Mergers & Acquisitions database in the 1993–2006 period. We use the COMPUSTAT Segment Customer database field, customer name, to identify whether asset sellers have large corporate customers. An asset seller is defined as having a large customer when a customer represents 10% or more of an asset seller’s total sales. When an asset seller has multiple large customers, the customer purchasing the largest amount is identified as the large customer. To determine whether a customer is publicly traded or privately held, we match the names in the COMPUSTAT Segment Customer database with those in the COMPUSTAT database. The dependent variable is the indicator variable for the high drop in % of COGS of customer which takes the value of one if the drop in the proportion of a supplier’s sales out of the customer’s COGS is higher than the sample median and zero otherwise. p-values reported in parentheses are heteroscedasticity robust and clustered by industry. All regressions include year indicator variables. Appendix summarizes the definitions of the variables used in table. The symbols ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. 55 Table 8 (1) Durable goods industry (indicator) Durable goods industry (indicator) * Stock return volatility Technology intense industry (indicator) Return on asset Tobin's Q Proceed used to pay debt Deals paid with cash only Relative deal size Customer Tobin's Q Log (Customer market cap) Customer leverage Customer ROA Customer stock price runup Asset liquidity Year fixed effect N (4) -1.657 *** (0.003) 50.730 *** (0.010) 0.448 *** (0.000) -0.044 (0.973) -0.503 *** (0.003) 0.387 (0.476) -0.085 (0.692) 2.189 *** (0.010) 0.298 *** (0.008) -0.423 *** (0.002) -0.940 (0.450) 3.157 (0.280) 0.655 (0.136) 6.209 * (0.077) 12.220 (0.504) 0.538 *** (0.000) 1.050 (0.499) -0.674 *** (0.000) 0.412 (0.442) -0.099 (0.657) 2.245 ** (0.012) 0.351 ** (0.029) -0.479 *** (0.001) -0.738 (0.615) 3.780 (0.212) 0.987 ** (0.013) 6.306 * (0.064) 0.409 *** (0.000) 0.256 (0.836) -0.589 *** (0.000) 0.503 (0.329) -0.167 (0.442) 2.165 *** (0.003) 0.347 *** (0.005) -0.438 *** (0.001) -0.173 (0.894) 3.291 (0.242) 0.748 * (0.099) 6.130 * (0.073) -0.623 (0.427) 41.080 ** (0.026) 12.560 (0.494) 0.536 *** (0.001) 1.173 (0.437) -0.782 *** (0.000) 0.603 (0.225) -0.181 (0.410) 2.311 *** (0.004) 0.387 ** (0.013) -0.474 *** (0.001) -0.545 (0.682) 3.873 (0.179) 0.939 * (0.056) 6.528 * (0.050) Yes 247 Yes 247 Yes 247 Yes 247 Technology intense industry (indicator) * Stock return volatility Stock return volatility Log (Market capitalization) (3) 1.175 *** (0.005) 0.390 (0.369) (2) 56 Table 9 Post-asset-sale relationship survival for asset selling firms with a large public customer and non-asset selling matching firms with a large public customer The sample consists of 310 asset sellers with large public corporate customers and 310 nonasset-selling firms with large public corporate customers reported in Thomson Financial’s SDC Mergers & Acquisitions database in the 1993–2006 period. We use the COMPUSTAT Segment Customer database field, customer name, to identify whether asset sellers have large corporate customers. An asset seller is defined as having a large customer when a customer represents 10% or more of an asset seller’s total sales. When an asset seller has multiple large customers, the customer purchasing the largest amount is identified as the large customer. To determine whether a customer is publicly traded or privately held, we match the names in the COMPUSTAT Segment Customer database with those in the COMPUSTAT database. The dependent variable is the number of years the relationship continues after the asset sale. The estimates are from a Cox proportional hazards model. p-values reported in parentheses are heteroscedasticity robust and clustered by industry. All regressions include indicator variables for year and each of FamaFrench 48 industries. Appendix summarizes the definitions of the variables used in table. The symbols ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. 57 Table 9 Asset seller (indicator) Durable goods industry (indicator) Technology intense industry (indicator) Asset seller (indicator) * R&D intensity Asset seller with high CAR (indicator) Asset seller with low CAR (indicator) Proceed used to pay debt (indicator) R&D intensity (1) 1.199 * (0.071) (2) 1.202 * (0.067) 0.760 (0.095) Cox (3) 1.190 * (0.085) (4) 1.296 ** (0.018) (5) * 0.834 (0.351) 0.138 (0.136) 1.149 (0.224) 0.508 (0.425) 0.079 *** (0.000) 8.993 *** (0.000) 1.037 (0.601) 0.921 *** (0.008) 1.074 (0.497) 1.137 (0.259) 0.829 (0.831) 0.069 *** (0.000) 10.163 *** (0.000) 1.036 (0.610) 0.921 *** (0.008) 1.071 (0.510) 1.147 (0.229) 0.686 (0.672) 0.074 *** (0.000) 9.493 *** (0.000) 1.037 (0.602) 0.922 *** (0.009) 1.069 (0.532) 1.149 (0.224) 1.216 (0.851) 0.074 *** (0.000) 9.590 *** (0.000) 1.033 (0.646) 0.929 ** (0.020) 1.087 (0.431) 1.057 (0.642) 1.464 *** (0.001) 1.159 (0.198) 0.444 (0.338) 0.096 *** (0.001) 7.675 *** (0.000) 1.070 (0.339) 0.913 *** (0.003) 1.085 (0.438) Industry fixed effects Year fixed effects Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes N 620 620 620 620 620 -2319.90 -2318.78 -2319.66 -2318.99 -2317.65 Percent of sales to customer (Percent of sales to customer)2 Log (1+firm age) Log (total assets) Negative free cash flow supplier (indicator) Log pseudo-likelihood 58 Table 10 Determinant of asset sales The sample consists of 52,347 COMPUSTAT firms from 1993 to 2006. Using the COMPUSTAT Segment Customer database, 13,354 firms are identified as having at least one large customer and using Thomson Financial’s SDC Mergers & Acquisitions database, 5,681 firms are identified as engaging in asset sales. A firm is defined as having a large customer when a customer represents 10% or more of an asset seller’s total sales. When a firm has multiple large customers, the customer purchasing the largest amount is identified as the large customer. The dependent variable is indicator variable that takes the value of one if a firm sells its asset in a certain year and zero otherwise. p-values reported in parentheses are heteroscedasticity robust and clustered by industry. All regressions include indicator variables for year and Fama-French 48 industries except otherwise where noted. Appendix summarizes the definitions of the variables used in table. The symbols ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. (1) Company with large customer (indicator) Relationship terminated (indicator) Relationship not terminated (indicator) Durable industry (indicator) (2) Logit (3) Log (Market capitalization) (5) (6) (7) (8) 0.043 (0.326) 0.240 *** (0.001) 0.052 (0.437) 0.016 (0.818) Technology intense industry (indicator) High R&D firm (indicator) Firm age (4) -0.088 (0.381) 0.034 *** 0.029 *** 0.029 *** 0.029 *** (0.000) (0.000) (0.000) (0.000) *** *** *** 0.352 0.334 0.331 0.333 *** (0.000) (0.000) (0.000) (0.000) 59 -0.027 (0.829) 0.034 *** (0.000) 0.351 *** (0.000) 0.239 *** 0.249 *** 0.186 *** (0.001) (0.001) (0.002) 0.051 0.057 -0.017 (0.429) (0.391) (0.704) 0.006 (0.930) -0.099 (0.313) -0.029 (0.816) 0.029 *** 0.029 *** 0.034 *** (0.000) (0.000) (0.000) *** *** 0.334 0.336 0.353 *** (0.000) (0.000) (0.000) Table 10 (Cont’d) Tobin's q Leverage ROA Stock price runup Stock return volatility Negative free cash flows (indicator) Year fixed effect Industry fixed effect N -0.260 *** (0.000) 1.262 *** (0.000) -0.577 *** (0.000) -0.241 *** (0.000) 8.265 *** (0.000) 0.346 *** (0.000) -5.050 *** (0.000) Yes Yes 52,347 -0.227 *** (0.000) 1.150 *** (0.000) -0.491 *** (0.000) -0.274 *** (0.000) 9.672 *** (0.000) 0.259 *** (0.000) -4.735 *** (0.000) Yes Yes -0.225 *** (0.000) 1.142 *** (0.000) -0.488 *** (0.000) -0.275 *** (0.000) 9.719 *** (0.000) 0.263 *** (0.000) -4.709 *** (0.000) Yes No -0.222 *** (0.000) 1.119 *** (0.000) -0.487 *** (0.000) -0.277 *** (0.000) 10.020 *** (0.000) 0.262 *** (0.000) -4.704 *** (0.000) Yes No -0.258 *** (0.000) 1.258 *** (0.000) -0.579 *** (0.000) -0.241 *** (0.000) 8.287 *** (0.000) 0.349 *** (0.000) -5.034 *** (0.000) Yes Yes -0.227 *** (0.000) 1.151 *** (0.000) -0.491 *** (0.000) -0.274 *** (0.000) 9.662 *** (0.000) 0.259 *** (0.000) -4.737 *** (0.000) Yes No -0.225 *** (0.000) 1.130 *** (0.000) -0.491 *** (0.000) -0.277 *** (0.000) 9.976 *** (0.000) 0.259 *** (0.000) -4.738 *** (0.000) Yes No -0.259 *** (0.000) 1.263 *** (0.000) -0.578 *** (0.000) -0.241 *** (0.000) 8.260 *** (0.000) 0.347 *** (0.000) -5.046 *** (0.000) Yes Yes 52,347 52,347 52,347 52,347 52,347 52,347 52,347 60 Table 11 Determinant of asset sales: Subsampled by technology intense industry and durable industry The sample consists of 52,347 COMPUSTAT firms from 1993 to 2006. Using the COMPUSTAT Segment Customer database, 13,354 firms are identified as having at least one large customer and using Thomson Financial’s SDC Mergers & Acquisitions database, 5,681 firms are identified as engaging in asset sales. A firm is defined as having a large customer when a customer represents 10% or more of an asset seller’s total sales. When a firm has multiple large customers, the customer purchasing the largest amount is identified as the large customer. The dependent variable is indicator variable that takes the value of one if a firm sells its asset in a certain year and zero otherwise. p-values reported in parentheses are heteroscedasticity robust and clustered by industry. All regressions include indicator variables for year and Fama-French 48 industries except otherwise where noted. Appendix summarizes the definitions of the variables used in table. The symbols ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. Logit (1) High Tech non-High Tech Relationship terminated (indicator) Relationship not terminated (indicator) Firm age Log (Market capitalization) Tobin's q Leverage ROA Stock price runup Stock return volatility Negative free cash flows (indicator) Constant Year fixed effect Industry fixed effect N 0.176 * (0.051) 0.057 (0.240) 0.048 *** (0.000) 0.293 *** (0.000) -0.207 *** (0.000) 1.127 *** (0.000) -0.561 *** (0.000) -0.169 ** (0.012) 6.350 ** (0.010) 0.408 *** (0.000) -4.913 *** (0.000) Yes Yes 12,106 0.216 *** (0.010) -0.019 (0.745) 0.032 *** (0.000) 0.378 *** (0.000) -0.282 *** (0.000) 1.233 *** (0.000) -0.571 *** (0.001) -0.287 *** (0.000) 9.307 *** (0.000) 0.349 *** (0.000) -5.120 *** (0.000) Yes Yes 40,241 61 Durable 0.104 (0.277) 0.003 (0.965) 0.036 *** (0.000) 0.337 *** (0.000) -0.192 *** (0.000) 1.224 *** (0.000) -0.522 *** (0.000) -0.205 *** (0.000) 5.892 *** (0.000) 0.500 *** (0.000) -6.353 *** (0.000) Yes Yes 15,728 (2) non-Durable 0.223 *** (0.007) -0.037 (0.517) 0.034 *** (0.000) 0.361 *** (0.000) -0.291 *** (0.000) 1.264 *** (0.000) -0.613 *** (0.001) -0.262 *** (0.000) 9.081 *** (0.000) 0.298 *** (0.000) -4.926 *** (0.000) Yes Yes 36,619 Figure 1 Interaction Term in Logit Regression: Model (2) of Table 7 This shows the interaction effects between durable goods industry indicator and stock return volatility on the probability of major trading relationship termination after asset sale. 62 Figure 2 Interaction Term in Logit Regression: Model (4) of Table 7 This shows the interaction effects between durable goods industry indicator and stock return volatility on the probability of high drop in sales to customer after asset sale. 63 Figure 3 Interaction Term in Logit Regression: Model (2) of Table 8 This shows the interaction effects between durable goods industry indicator and stock return volatility on the probability of high drop in % of COGS from asset selling suppliers after the asset sale. 64 Figure 4 Interaction Term in Logit Regression: Model (4) of Table 8 This shows the interaction effects between technology intense industry indicator and stock return volatility on the probability of high drop in % of COGS from asset selling suppliers after the asset sale. 65 APPENDIX B VARIABLE DEFINITIONS 66 Appendix B: Variable definitions This appendix provides a detailed description of the construction of all the variables used in the tables. Variable name Definition Durable industry An indicator variable taking a value of one for asset selling firms in durable industries as defined by the US Department of Census. The Census Bureau classification of durable industry firms is contained at: http://www.census.gov/indicator/www/m3/appendixb.pdf. Firm age The number of years since the asset selling firm first reported data item AT (book value of assets) in the COMPUSTAT database. High drop in percent sales to customer An indicator variable taking a value of one if the asset selling firm has above the median decline in the percent of sales to the large customer from the year before the asset sale to after the asset sale. High drop in % COGS High R&D An indicator variable taking a value of one if the asset selling firm has above the median R&D / total assets and zero otherwise. Large customer An indicator variable taking a value of one if the firm discloses a customer that accounts for greater than 10% of its sales and zero otherwise. Large public customer An indicator variable taking a value of one if the firm discloses a customer that accounts for greater than 10% of its sales and that customer can be found in the COMPUSTAT database. Leverage COMPUSTAT data item 9 / data item 6 in the year before the offer, winsorized at the 95th percentile. Market capitalization COMPUSTAT data item CSHO* data item PRCC_F in the year before the asset sale. For firms missing COMPUSTAT observations, CRSP data using the shares outstanding and closing price for the month before the asset sale announcement is utilized. Percent of cost of goods sold The dollar amount of sales to the customer divided by the customer COMPUSTAT data item COGS. 67 Percent of firms in same industry with a large customer Number of firms in the same industry (Fama and French (1997) 48 industry) as the firm of interest that also have large customers based on the COMPUSTAT segment level data divided by the total number of firms in that same industry. Percent of sales to customer The sales made to the large customer by asset selling firm divided by COMPUSTAT data item SALES. Post-sale relationship length The number of years from the asset sale year until the relationship is terminated. Pre-asset-sale relationship length The number of years from the first time the customer shows up in filings as a large customer until the asset sale year. Proceeds The offer price as reported in the SDC Platinum Merger & Acquisition Database. Proceeds used to repay debt An indicator variable taking on a value of one if the asset selling firm reports that it will use the proceeds to repay debt and zero otherwise. R&D intensity COMPUSTAT data item XRD / data item AT in the year before the asset sale. Relationship tenure The number of years the relationship has lasted. ROA ROA (return on assets) is measured as data item OIBDP / data item AT. Stock price run-up The buy-and-hold return for the SEO firm from day -365 to day -11 before the asset sale announcement date. Stock return volatility Daily stock return standard deviation for the asset sellers from day -365 to day -11 before the asset sale announcement. Technology intensive industry Following Loughran and Ritter (2004) we define a technology firm as having one of the following 4-digit SIC codes: 3571, 3572, 3575, 3577, 3578, 3661, 3663, 3669, 3671, 3672, 3674, 3675, 3677-3679, 3812, 3823, 3825-3827. 3829, 3841, 3845, 4812, 4813, 4899, 7371-7375, 7378, and 7379. Tobin’s Q COMPUSTAT data (item LCT + item DLTT + item CSHO * item 68 PRCC_F) / item AT in the year before the asset sale. Total assets COMPUSTAT data item AT in the year before the asset sale. 69 BIBLIOGRAPHY 70 Bibliography Allen, J. W. and M. P. Gordon (2000). "Corporate Equity Ownership, Strategic Alliances, and Product Market Relationships." The Journal of Finance 55(6): 2791-2815. Amira, K., Prezas, A. P., Vasudevan, G. K. (2008). "Asset buyers and leverage." Working paper, Suffolk Univerity. Arora, A. and P. Alam (2005). "CEO Compensation and Stakeholders' Claims*." Contemporary Accounting Research 22(3): 519-547. Babich, V., A. N. Burnetas, et al. (2007). "Competition and Diversification Effects in Supply Chains with Supplier Default Risk." Manufacturing & Service Operations Management 9(2): 123-146. Banerjee, S., S. Dasgupta, et al. (2008). "Buyer-Supplier Relationships and the Stakeholder Theory of Capital Structure." The Journal of Finance 63(5): 2507-2552. 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(3): 359-399. Bates, T. W. (2005). "Asset Sales, Investment Opportunities, and the Use of Proceeds." The Journal of Finance 60(1): 105-135. Berger, P. G. and E. Ofek (1999). "Causes and Effects of Corporate Refocusing Programs." The Review of Financial Studies 12(2): 311-345. Brown, D. T., C. E. Fee, et al. (2009). "Financial leverage and bargaining power with suppliers: Evidence from leveraged buyouts." Journal of Corporate Finance 15(2): 196-211. Chan, S. H., J. W. Kensinger, et al. (1997). "Do strategic alliances create value?" Journal of Financial Economics 46(2): 199-221. Chevalier, J. A. (1995). "Capital Structure and Product-Market Competition: Empirical Evidence from the Supermarket Industry." The American Economic Review 85(3): 415-435. Cornell, B. and A. C. Shapiro (1987). "Corporate Stakeholders and Corporate Finance." Financial Management 16(1): 5-14. Cremers, K. J. M., V. B. Nair, et al. (2008). "Takeover Defenses and Competition: The Role of Stakeholders." Journal of Empirical Legal Studies 5(4): 791-818. Datta, S., M. Iskandar-Datta, et al. (2003). "Value creation in corporate asset sales: The role of managerial performance and lender monitoring." Journal of Banking & Finance 27(2): 351375. Eckbo, B. E. and R. W. Masulis (1992). "Adverse selection and the rights offer paradox." Journal of Financial Economics 32(3): 293-332. 71 Eckbo, B. E., R. W. Masulis, et al. (2000). "Seasoned public offerings: resolution of the ‘new issues puzzle’." Journal of Financial Economics 56(2): 251-291. Eisfeldt, A. L. and A. A. Rampini (2006). "Capital reallocation and liquidity." Journal of Monetary Economics 53(3): 369-399. Ellis, J. A., Fee, C. E., and S. Thomas (2009). "Product market competition and the disclosure of information about customers." Working Paper, University of Pittsburgh. Fan, Joseph P. H. and Vidhan K. Goyal (2006). "On the Patterns and Wealth Effects of Vertical Mergers." The Journal of Business 79(2): 877-902. Fan, P. H. J. P. and V. K. Goyal (2002). "On the Patterns and Wealth Effects of Vertical Mergers." SSRN eLibrary. FEE, C. E., C. J. HADLOCK, et al. (2006). "Corporate Equity Ownership and the Governance of Product Market Relationships." The Journal of Finance 61(3): 1217-1251. Fee, C. E. and S. Thomas (2004). "Sources of gains in horizontal mergers: evidence from customer, supplier, and rival firms." Journal of Financial Economics 74(3): 423-460. Fluck, Z. and Anthony W. Lynch (1999). "Why Do Firms Merge and Then Divest? A Theory of Financial Synergy." The Journal of Business 72(3): 319-346. Fuller, K., J. Netter, et al. (2002). "What Do Returns to Acquiring Firms Tell Us? Evidence from Firms That Make Many Acquisitions." The Journal of Finance 57(4): 1763-1793. Gaspar, J. M., M. Massa, et al. (2005). "Shareholder investment horizons and the market for corporate control." Journal of Financial Economics 76(1): 135-165. Gordon, J. A., P. G. Benson, et al. (1984). "Investigating the Valuation Effects of Announcements of Voluntary Corporate Selloffs." The Journal of Finance 39(2): 503-517. Harford, J. (2005). "What drives merger waves?" Journal of Financial Economics 77(3): 529-560. Healy, P. M. and K. G. Palepu (1990). "Earnings and Risk Changes Surrounding Primary Stock Offers." Journal of Accounting Research 28(1): 25-48. Hendricks, K. B. and V. R. Singhal (2003). "The effect of supply chain glitches on shareholder wealth." Journal of Operations Management 21(5): 501-522. Hite, G. L., J. E. Owers, et al. (1987). "The market for interfirm asset sales: Partial sell-offs and total liquidations." Journal of Financial Economics 18(2): 229-252. Holmstrom, B. and J. Roberts (1998). "The Boundaries of the Firm Revisited." The Journal of Economic Perspectives 12(4): 73-94. Jain, P. C. (1985). "The Effect of Voluntary Sell-Off Announcements on Shareholder Wealth." The Journal of Finance 40(1): 209-224. 72 Jensen, M. C. and W. H. Meckling (1976). "Theory of the firm: Managerial behavior, agency costs and ownership structure." Journal of Financial Economics 3(4): 305-360. Jensen, M. C. and R. S. Ruback (1983). "The market for corporate control : The scientific evidence." Journal of Financial Economics 11(1-4): 5-50. John, K. and E. Ofek (1995). "Asset sales and increase in focus." Journal of Financial Economics 37(1): 105-126. John, K. and E. Ofek (1995). "Asset sales and increase in focus." Journal of Financial Economics 37(1): 105-126. Johnson, W. C., J.-K. Kang, et al. (2011). "Supply-Chain Spillover Effects and the Interdependence of Firm Financing Decisions." SSRN eLibrary. Joskow, P. (2005). Vertical Integration. Handbook of New Institutional Economics, Springer US: 319-348. Joskow, P. L. (1988). "Asset Specificity and the Structure of Vertical Relationships: Empirical Evidence." Journal of Law, Economics, & Organization 4(1): 95-117. Kang, J.-K., J.-M. Kim, et al. (2006). "Post-takeover Restructuring and the Sources of Gains in Foreign Takeovers: Evidence from U.S. Targets." The Journal of Business 79(5): 2503-2537. Kaplan, S. N. and M. S. Weisbach (1992). "The Success of Acquisitions: Evidence from Divestitures." The Journal of Finance 47(1): 107-138. Kiefer, N. M. (1988). "Economic Duration Data and Hazard Functions." Journal of Economic Literature 26(2): 646-679. Lang, L., A. Poulsen, et al. (1995). "Asset sales, firm performance, and the agency costs of managerial discretion." Journal of Financial Economics 37(1): 3-37. Lazzarini, S., F. Chaddad, et al. (2001). "Integrating supply chain and network analyses: The study of netchains." Journal on Chain and Network Science 1(1): 7-22. Lee, H. L., V. Padmanabhan, et al. (1997). "Information Distortion in a Supply Chain: The Bullwhip Effect." Management Science 43(4): 546-558. Ling Cen, S. D., and Rik Sen (2010). "Discipline or Disruption? Stakeholder Relationships and the Effect of Takeover Threat." Working paper. Loughran, T. and J. R. Ritter (1997). "The Operating Performance of Firms Conducting Seasoned Equity Offerings." The Journal of Finance 52(5): 1823-1850. Maksimovic, V. and P. Gordon (1998). "Asset Efficiency and Reallocation Decisions of Bankrupt Firms." The Journal of Finance 53(5): 1495-1532. 73 Maksimovic, V. and P. Gordon (2001). "The Market for Corporate Assets: Who Engages in Mergers and Asset Sales and Are There Efficiency Gains?" The Journal of Finance 56(6): 2019-2065. Maksimovic, V. and S. Titman (1991). "Financial policy and reputation for product quality." Review of Financial Studies 4(1). Masulis, R. W., C. Wang, et al. (2011). "Globalizing the Boardroom - The Effects of Foreign Directors on Corporate Governance and Firm Performance." Journal of Accounting & Economics (JAE), Forthcoming. Mulherin, J. H. and A. L. Boone (2000). "Comparing acquisitions and divestitures." Journal of Corporate Finance 6(2): 117-139. Opler, T. C. and S. Titman (1994). "Financial Distress and Corporate Performance." The Journal of Finance 49(3): 1015-1040. Ortiz-Molina, H. and G. M. Phillips (2009). "Asset Liquidity and the Cost of Capital." SSRN eLibrary. Pulvino, T. C. (1998). "Do Asset Fire Sales Exist? An Empirical Investigation of Commercial Aircraft Transactions." The Journal of Finance 53(3): 939-978. Raman, K. and H. Shahrur (2008). "Relationship-Specific Investments and Earnings Management: Evidence on Corporate Suppliers and Customers." The Accounting Review 83(4): 1041-1081. Ray, S. a. W., Missaka (2009). "An Efficiency Perspective on the Gains from Mergers and Asset Purchases." The B.E. Journal of Economic Analysis & Policy Vol. 9(Iss. 1 (Contributions)): Article 43. . Reinartz, W., M. Krafft, et al. (2004). "The Customer Relationship Management Process: Its Measurement and Impact on Performance." Journal of Marketing Research 41(3): 293-305. Schlingemann, F. P., R. M. Stulz, et al. (2002). "Divestitures and the liquidity of the market for corporate assets." Journal of Financial Economics 64(1): 117-144. Se-Jik Kim, H. S. S. (2010). "Sustaining Production Chains Through Financial Linkages." Working paper. Shenoy, J. (2008). "An Examination of the Efficiency, Foreclosure, and Collusion Rationales for Vertical Takeovers." SSRN eLibrary. Sheridan, T. (1984). "The effect of capital structure on a firm's liquidation decision." Journal of Financial Economics 13(1): 137-151. Shleifer, A. and R. W. Vishny (1992). "Liquidation Values and Debt Capacity: A Market Equilibrium Approach." The Journal of Finance 47(4): 1343-1366. 74 Sibilkov, V. (2009). "Asset Liquidity and Capital Structure." Journal of Financial and Quantitative Analysis 44(05): 1173-1196. Song, M. H. and R. A. Walkling (2000). "Abnormal returns to rivals of acquisition targets: A test of the acquisition probability hypothesis'." Journal of Financial Economics 55(2): 143-171. Titman, S. and R. Wessels (1988). "The Determinants of Capital Structure Choice." Journal of Finance 43(1): 1-19. Trkman, P. and K. McCormack (2009). "Supply chain risk in turbulent environments—A conceptual model for managing supply chain network risk." International Journal of Production Economics 119(2): 247-258. Williamson, O. E. (1983). "Credible Commitments: Using Hostages to Support Exchange." The American Economic Review 73(4): 519-540. Williamson, O. E. (1985). "The Economic Institute of Capitalism." The Free Press. Williamson, O. E. (1991). "Comparative Economic Organization: The Analysis of Discrete Structural Alternatives." Administrative Science Quarterly 36(2): 269-296. Yang, L. (2008). "The Real Determinants of Asset Sales." The Journal of Finance 63(5): 22312262. 75