Corporate Governance and Investment Efficiency of Diversified firms: Evidence from Corporate Asset Purchases

Transcription

1 Corporate Governance and Investment Efficiency of Diversified firms: Evidence from Corporate Asset Purchases I-Ju Chen Department of Finance, Yuan Ze University, Taiwan and Sheng-Syan Chen Department of Finance, National Taiwan University, Taiwan JEL classification: G31; G34 Keywords: corporate governance, investment efficiency, diversification discount Address correspondence to I-Ju Chen, Department of Finance, College of Management, Yuan Ze University, 135, Far-East Rd., Chung-Li, Taoyuan, Taiwan, ROC. Tel : ext.3664 Fax : ; ijchen@saturn.yzu.edu.tw. We wish to thank Dosoung Choi, Kim Wai Ho, Frank C. Jen, and Cheng-few Lee for helpful comments and suggestions. 1

2 Corporate Governance and Investment Efficiency of Diversified firms: Evidence from Corporate Asset Purchases Abstract This study examines the relationship among the corporate governance, investment efficiency, and the excess value of diversified firms. Using a sample of diversified firms that announced asset purchases between 1988 and 2003, we find that diversified firms with better governance structure improve their investment efficiency subsequent to asset purchases. The characteristics of well-governed diversified firms include larger board size, higher board independence, less busy board, higher institutional ownership, higher outside director ownership, higher CEO equity-based pay, better audit quality, and better shareholder s protection. We also find that the changes of investment ratio are associated with the problem of cross-subsidization among divisions within diversified firms. The inefficient cross-subsidies in internal capital market as discussed in Rajan et al. (2000) or Scharfstein and Stein (2000) do not exist in well-governed diversified firms. The investment allocations among divisions within the well-governed diversified firms have improved after asset purchases. We further show that diversified firms with better governance mechanisms, either internal or external, have higher efficiency of investment allocations among divisions of diversified firms. Our results are generally consistent with the hypothesis that governance has a substantial impact on the excess value of diversified firms through its impact on investment allocations. JEL classification: G31; G34 Keywords: corporate governance, investment efficiency, diversification discount 2

3 1. Introduction Managers tend to waste corporate resources. This is the implication of the extensive literature on agency costs introduced by Jensen and Meckling (1976), who argue that the agency problem is crucial to firm performance due to the separation of ownership and control. This paper examines how such agency problems result in investment inefficiency and how corporate governance structure affects the efficiency of investment allocations. Our study focuses on one particular event: asset acquisition. We examine the event of asset acquisition for three reasons. First, asset acquisition is an important corporate investment activity in the United States 1 and the potential wealth destruction to firm shareholders is large as documented by Moeller et al. (2005). It has often been recognized as inefficient investment decision for managerial self-interest as indicated by Jensen (1986, 1993) and Freund et al. (2003). 2 Second, unlike mergers and acquisitions, the transaction of asset purchases does not involve in corporate control issue (Warusawitharana, 2007), is more invest-oriented (Maksimovic and Phillips, 2001), and has less concern of information asymmetry to assess the value of acquired asset (Graham, Lemmon, and Wolf, 2002). Therefore, the sample of asset acquisition enables us to investigate how resource allocates among divisions of diversified firms around transactions and how governance mechanism affects diversification discount of diversified firms. 3 Lastly, while firm-level governance itself is only slowly changing, there is substantial variation in firm-level investment outcome over time. This provides us with a good opportunity to investigate the effect of governance structure on firm value and the use of resource allocation in individual firms. 1 There is a large and active market for asset purchases in the United States. For example, more than $100 billion worth of assets were transacted between firms in 2004 (Warusawitharana, 2007). 2 The agency problem is most emphasized in financial literature that has the most direct implications for investment (Baumol, 1959; Marris, 1964; Williamson, 1964; Jensen,1986, 1993; among others). 3 Some studies investigate the mechanism of corporate governance and link the difference of corporate governance between diversified firms and focused firms to the value discount (Denis et al., 1997; Palia, 1999; Lins and Servaes, 1999; Anderson et al., 2000; Aggarwal and Samwick, 2003; Singh et al., 2004). Their results do not unanimously assist that the failure of governance mechanism of diversified firms explains the reason of diversification discounts. 3

4 We propose that corporate governance plays an important role in efficient resource allocation among divisions of diversified firms and thus affects firm valuation. Shleifer and Vishny (1997), Tirole (2001) and Becht et al. (2002) argue that corporate governance works to alleviate the agency problem and enhances firm performance. As documented by Rajan et al. (2000), Scharfstein (1998), Scharfstein and Stein (2000), and Stein (2003), agency problem leads to inefficient investment behavior and attributes to the major reason of the diversification discount. Hence, comparing to a poorly governed diversified firm, a well-governed diversified firm has greater improvement in investment efficiency and has higher market valuation subsequent to asset purchases. A sample of asset purchase announcements during the period of is collected to test our predictions. Consistent with the literature, we find that governance structures are significantly different between single-segment firms and multi-segment firms. Our evidence shows that there is a significant association between the improvement in investment efficiency and 1) board characteristics, 2) both institutional ownership (public pension funds holdings) and outside director ownership, 3) CEO equity-based pay, 4) audit quality, and 5) shareholder s protection. The analysis of composite governance index, either internal or external, consistently shows that the changes of investment efficiency are associated with the problem of cross-subsidization among divisions within diversified firms and it is found that the investment allocations among divisions improve for well-governed diversified firms after asst purchases. Finally, our cross-sectional regression analysis indicates that the increase in the excess value of diversified firms is associated with the improvement in investment efficiency for those with better governance. Our study makes at least three valuable contributions to the literature. First, using the sample of asset purchases, we document that a well-governed diversified firm is effective to alleviate the agency problems between the headquarter and division-managers and to enhance firm valuation by efficient capital allocations. To our best knowledge, no study so far has investigated how governance structure influences the capital allocation of diversified firms surrounding their investment decision and how it affects the resulting changes in value discounts. Several studies investigate the mechanism of corporate governance and link the difference in 4

5 corporate governance between diversified firms and focused firms to the value discount (Denis et al., 1997; Palia, 1999; Lins and Servaes, 1999; Anderson et al., 2000; Aggarwal and Samwick, 2003; Singh et al., 2004). Their results do not provide consistent conclusion regarding whether the failure of governance mechanism of diversified firms explains the reason of diversification discounts. For example, Denis et al. (1997) and Anderson et al. (2000) show that there is a negative relationship between the excess value of diversification and managerial incentives. However, Aggarwal and Samwick (2003) find that the diversification is increasing in incentives. The interacting or counteracting effect among unobserved factors is a possible reason for unclear relation between the agency conflicts and diversification discounts 4 by a large sample study. Our study uses the event of asset purchase to highlight the role of corporate governance in capital allocation among divisions of diversified firms and provides the evidence that governance structure is crucial in determining value discount of diversified firms. Second, our findings have important implications for improving the investment efficiency for corporate asset acquisition process. We provide the evidence that well-governed diversified firms make more efficient investment decision and enhance firm valuation. Prior studies argues that firms with better governance mechanisms help to mitigate agency problems and result in more efficient investment decisions (Fama, 1980; Fama and Jensen, 1983; Agrawal and Mandelker, 1987; Bizjak et al., 1993; Datta et al., 2001; among others) and improve firm performance. They have examined board characteristics (Fama and Jensen, 1983; Monks and Minow, 1995; Beasley et al., 2000; Farber, 2005; Brickley, et al. 1997; among others), ownership structure (Morck et al., 1988; McConnell and Servaes,1990; Hermalin and Weisbach, 1988; Loderer and Martin, 1997; Himmelberg et al., 1999; Demsetz and Lehn,1985; among others), CEO compensation (Fama and Miller, 1972; Jensen and Murphy, 1990; Bizjak et al., 1993; Datta et al., 2001; among others), and takeover provisions or investor s protection (Jensen, 1988; Scharfstein, 1998; La Porta et al., 2000; Gompers et al., 2003; among others). However, there is no study so far examining the 4 Some studies document that diversification discount could in fact be an artifact of data selection and empirical approaches. Gramham, Lemmon, and Wolf (2002) and Campa and Kedia (2002) report that the diversification discount could be related to firm characteristics. Whited (2001) and Chevaler (2004) argue that the discount may be partially explained by measurement errors. Villalonga (2004a,b) by using new segment dataset find that there exists a diversification premium. 5

6 relationship between the investment efficiency and firm valuation under different governance mechanisms. Our study investigates the importance of each governance mechanism by linking to the capital allocation and valuation of diversified firms. Third, our analysis adds a new perspective to the asset acquisition literature by examining whether the change in firm valuation can be explained by the acquiring firm s governance structure prior to the asset acquisition. Previous studies document that post-asset-acquisition performance is affected by synergy consideration (Maksimovic and Phillis, 2001; 2002) or managerial discretion 5 (Freund et al., 2003). Similar to the studies of Fruend et al. (2003), we address the agency issue of asset acquisitions. However, we provide the counter evidence that asset buyers experience value improvement when they have well-governed structure. Good governance helps firms implement efficient capital allocations and improves the valuations of diversified firms. The reminder of the paper is organized as follows. In section 2, we discuss the related literature and describe the research hypotheses in more detail. In section 3, we describe the sample and research design. The empirical results are provided in section 4 to section 7. The conclusions are shown in section Literature and hypotheses statements 2.1 Governance structure and investment decision Stein (2003) documents that the problems of asymmetric information and agency have significant impact on the efficiency of corporate investment decisions. Literature, such as Shleifer and Vishny (1997), Tirole (2001), and Becht et al.(2002), states the importance of corporate governance which is the set of mechanisms designed to align the interest of managers to shareholders and assure themselves of getting a return on their investment. Many studies find that corporate governance mechanisms deal with agency 5 Freund et al. (2003) find that firms with fewer growth opportunities and large amounts of free cash flow experience poorer operating performance following the asset purchase. The evidence is consistent with the managerial agency hypothesis suggested by the authors. 6

7 problems and help to implement efficient investment decisions. Agrawal and Mandelker (1987) find that managerial incentives have significant impact on firm investment decisions. Similar to Agrawal and Mandelker (1987), Hill and Snell (1988) find that investment strategy, such as diversification, is affected by the governance and the divergence between stockholder and management interests is minimized when stock concentration and management stockholding are substantial. Gibbs (1993) shows that outside director s power to monitor and control manager s behavior has a significant impact on corporate restructuring decisions and the role of equity compensation for top management works as a device to align management interests with shareholder interests. Bizjak et al. (1993) stress the importance of the structure of managerial compensation to induce optimal investment choices. Seward and Walsh (1996) find the evidence that those voluntary spun-off companies usually have efficient internal control and governance which are usually characterized with the presence of stock options in the CEO s compensation plan, CEO payment in the form of performance-contingent compensations, minority of inside directors in the board. By linking the agency cost explanation for corporate diversification, Denis et al. (1997) provide similar evidence that the level of diversification is negatively related to managerial equity ownership and to the equity ownership of outside blockholders. Dattta et al. (2001) find that managers with high equity-based compensation pay lower acquisition premiums, acquire targets with higher growth opportunities, and have higher stock price performance around and following acquisition announcement. 2.2 Governance structure and firm valuation Better governance structure is expected to align the interest of managers and shareholders and improve firm performance. Shleifer and Vishny (1986) theoretically prove that the existence of large shareholders can improve performance of firms. Earlier studies focus on the effect of ownership structure on firm valuation find that there exists a non-monotonic relation between the inside ownership and firm performance (Morck et al., 1988; McConnell and Servaes, 1990; Hermalin and Weisbach, 1988; Loderer and Martin, 1997; among others). However, Demsetz and Lehn (1985) Demsetz and Villalonga (2001) find that there is no statistically significant relation between ownership structure and firm performance after 7

8 considering the endogeneity issue between the inside ownership and firm valuation. Lemmon and Lins (2003) investigate a sample in eight East Asian countries to study the effect of ownership structure on firm valuation during Asian financial crisis. They find that ownership structure plays an important role in determining the incentives of insiders to expropriate minority shareholders during Asian financial crisis. Rather than emphasizing on the internal governance mechanism, recent studies focus on the effect of external monitoring mechanism to firm valuation. Gompers et al. (2003) construct a governance index to proxy for the level of shareholder rights to examine the relationship between shareholder rights and corporate performance. Their empirical results show that firms with stronger shareholder rights have higher firm value. Durnev and Kim (2005) use another firm-level governance dataset from 27 countries and find that firms with higher governance and transparency rankings are valued higher in stock markets. Since governance structure is helpful to reduce the destruction of market value caused by agency conflicts of interest between the firm s stakeholders, some studies investigate the corporate governance structure of diversified firms and ask whether the differences in governance are associated with the value discount from diversification. By focusing on internal governance mechanism, Denis et al. (1997) and Anderson et al.(2000) find no compelling evidence to show that governance structure of diversified firms explain the value discount from diversification. Singh et al. (2004) further suggest the internal governance differences between diversified and focused firms are due to their being at different stages of corporate evolution. However, Aggarwal and Samwick (2003) consider the endogenous relationships among firm performance, diversification and managerial incentives. Their results indicate that diversification is positively related to managerial incentives and conclude that diversification is indeed related to the agency problem of a firm. Although it has not reached an agreement that the failure of corporate governance attributes to the value discounts of diversified firms, Lins and Servaes (1999) find that the effect of diversification on firm value is different across countries and the value of diversification is associated with the governance structure and institutional environment by using international data from Germany, Japan, and the 8

9 United Kingdom. No consistent result on the link of corporate governance and value discounts of diversified firms is probably due to the compounding and unobserved effects among large sample study. The sample of asset purchase facilitates the investigation of capital expenditure made by different governance structure and the effect to the firm performance. This allows us to directly gauge the benefit from well-governed diversified firms to implement an investment decision such as asset purchase. 3. Data and methods 3.1 Sample Our initial sample consists of asset purchases announced during and is taken from the Securities Data Corporation (SDC) database. From the data source, we require that the buyer is a publicly traded firm, and information on the asset purchase is available in the database maintained by the SDC. We exclude firms with segments operating in the financial industry (SIC ) for unavailability of segment data. We also limit the sample to include the largest transaction per firm per year. For each transaction, the asset purchased has a value of at least $100 million. We also eliminate firms with other corporate news announcement such as earnings, dividends, or capital structure changes three days before or three days after the initial announcement date. To be included in the sample, accounting data on operating performance variables must be available on both the COMPUSTAT annual industrial database and the COMPUSTAT Business Segment Information database. Stock price data must be available in the Center for Research in Security Prices (CRSP) database. Because our governance data are drawn from several sources, we do not require complete data availability for all variables to maximize the sample size. Our final sample consists of 1148 announcements of asset purchases. Table 1 shows the distribution of asset purchases by calendar year. Among the 1148 sample of asset purchases, 645 are undertaken by firms with multiple business segments as listed on Compustat s Industry Segment Files and 503 are undertaken by single-segment firms. [Insert Table 1 near hear] 9

10 3.2 Governance data We use multiple measures of corporate governance including mechanisms on board characteristics, ownership structure, CEO compensation, audit quality, and the degree of managerial entrenchment due to takeover protection (market for corporate control). We employ four measures of board characteristics, such as board size, board independence, leadership structure, and busy board. We clarify the board characteristics from Investors Responsibility Research Center (IRRC) Directors database beginning in The data on board characteristics prior to 1997 is collected from Compact D/SEC database. We employ four measures of ownership structure, such as inside ownership, blockholder ownership, institutional ownership, and outside director ownership. The data on ownership structure is collected from Compact D/SEC database. We also use CEO equity-based pay and CEO ownership as the measures for CEO compensation. Compustat Executive Compensation database provides the detail elements of CEO compensations beginning in We also investigate the role of audit committee. Two measures are used to as the proxy for audit quality: outside directors on audit committee and number of audit committee meetings. Those data is collected from proxy statements. The measure of Gompers, Ishii, and Metrick (2003) corporate governance index is used to proxy for external governance. The Gompers et al. (2003) index measures the numbers of antitakeover provisions in a firm s charter and in the legal code of the state in which the firm in incorporated. The data of the index is assembled and reported about every two years (1990, 1993, 1995, 1998, 2000, and 2002) by the IRRC and the index varies between zero and 24. Following the concept of Gompers et al. (2003), higher score of governance index are referred to as having the higher management power or the weaker shareholder rights. Therefore, a firm is usually thought to be a better external-governed firm when it has lower score of governance index. For those years without governance data, we use the data from closest year. 3.3 Research design To measure the impact of governance on firm investment and the associated diversification discount, we mainly conduct two analyses. First, by using regression 10

11 analysis we examine whether different level of governance variables leads to a change in firm investment efficiency and the associated change in diversification discounts. Second, we use the principal component analysis to have a composite index of internal governance and this index gives us a great opportunity to compare the changes in investment allocations among divisions of diversified firms for different level of internal governance. The following equation describes the regression, which we estimate separately for each measure of corporate governance: Δ MINV i = β 0 + β 1 GM i + k1dumdi i + k2dumfiti + k3dumagencyi + k4δcapex i where + k5 ΔOPI i + k6δleverage + k7δinverseqi + k8δmispricingi + k9 k ΔHerfindahl i + 10 (1) ΔSIZE ΔMINVi indicates the change in measure of investment efficiency from year -1 to year +1 for the sample of multi-segments firms i. We follow the techniques introduced by Rajan et al.(2000) and Ahn and Denis (2004) and compute three measures of overall investment efficiency of diversified firms: relative investment ratio (RINV), the relative value added by allocation (RVA) and the absolute value added by allocation (AVA). Relative investment ratio (RINV) is defined as the sales-weighted sum of firm-and industry-adjusted investment in high-q segments minus the sales-weighted sum of firm and industry adjusted investment in low-q segments. More precisely, Relative Investment Ratio (RINV) = I j S j ( S I ( ) S I wj ( S I ( ) S TS I j S j ( S I ( ) S k ss n j ss n ss n j j )) j j j n k j = 1 = 1 = + 1 j = 1 j j j I wj ( S j j i I ( ) S ss j )) where S is the sales of segment j, w is the sales of segment j divided by the firm s j j (2) ss total sales, I j is the capital expenditures of segment j, ( I ) S j is the capital expenditure-to-sales ratio of the median single-segment firm operating in the same three-digit SIC industry as firm j, and TS is the total sales of the firm. For j =1 k, 11

12 the firm s segments have an industry median q greater than the firm s sales-weighted average q, while j=(n-k+1) n indicates that the firm s segments have an industry median q less than the firm s sales-weighted average q. This measure implies that RINV will be greater when firms invest more in their high-q segments, that is, when they are more efficient. The second measure is called relative value added by allocation (RVA). We weight firm and industry-adjusted segment investment by the difference between the industry median Tobin s q for that segment and the sales weighted average q for the firm. More precisely, Relative Value Added by Allocation (RVA) = I j I I ss n j I ss S j ( q j q)( ( ) j w j j ( ( ) j )) 1 S S S S n j= 1 = j TS j (3) where q j is the median Tobin s q for single-segment firms operating in the same three-digit SIC code as the given segment, q is the sales-weighted average q of the firm. This measure takes into account the extent to which the firm allocates more (less) capital to divisions with better (poorer) investment opportunities. If the industry median q is a good proxy for the marginal q of segment investment, RVA can be viewed as a measure of the overall value added (subtracted) by the firm s investment allocation process. The third measure of overall investment efficiency of diversified firms we compute is the absolute value added by allocation (AVA) in which we weight industry-adjusted segment investment by the difference between the segment s q and 1. More precisely, Absolute Value Added by Allocation (AVA) = n j = 1 S ( q I 1)( S TS I ( ) S j ss j j j ) j (4) We also calculate asset-based measures of investment efficiency. When we calculate the asset-based measures, we replace the sales with the asset to measure all 12

13 the elements in defining the investment efficiency. GM i represents 13 governance mechanisms of diversified firm i: (1) board size, (2) board independence, (3) leadership structure, (4) busy board, (5) inside ownership, (6) blockholder ownership, (7) institutional ownership, (8) outside directors ownership, (9) CEO equity-based pay, (10) CEO ownership, (11) outside directors on audit committee, (12) number of audit committee meetings and (13) market for corporate control. The control variables in the regression include DumDI i, DumFit i, DumAgency i, Δ CAPEX i, Δ OPI i, Δ LEVERAGEi, Δ Inverseqi, Δ Mispricingi, Δ SIZE Δ Herfindahl i, and. Most of these are motivated by other studies and control for firm-specific characteristics that may be correlated with both firm investment efficiency and firm valuation. DumDI is the dummy variable indicating diversity-increased or diversity-unchanged of diversified firms around asset purchase to control for diversity cost effect (Rajan et al.,2000). We define the multi-segment firms as diversity-unchanged asset buyers for the similarity of 3-digit SIC codes between the target asset and acquirer s industry. If the asset acquired by the firms has no common 3-SIC codes with firm s segments, then this purchase is categorized as diversity-increased. If the asset buyer belongs to diversity-increased buyers, then we denote DumDI equal to one and zero otherwise. Maksimovic and Phillips (2001) find that the comparative advantage of diversified firms is in their main industries and the overall productivity increases when firms purchase asset of low productivity. To control for the fit or synergy effect, we use DumFit to indicate the asset acquired belongs to main segments of the diversified buyer or not. If the asset has the common 3-digit SIC with the main segments of the buyers, we denote DumFit as one and zero otherwise. Similar to Maksimovic and Phillips (2001), we define that main segments are segments which are at least 25 percent of firm s total sales. Freund et al. (2003) find that operating performance of asset buyers following the transaction is negatively related to the amount of free cash flow. DumAgency is 13

14 used to control for the agency effect. If the sample diversified firms have low growth opportunities and high free cash flow, we denote DumAgency as one and zero otherwise. It is concerned that the link between the change in investment policy or in valuation and governance may be the byproduct of changes in financial resources (see Campa and Kedia, 2002; Ahn and Denis, 2004). To control this possibility, we control for firm investment level using the change in the industry-adjusted ratio of capital expenditures to sales, denoted as Δ CAPEX, for the buyer firms. We also control for firm profitability level by using the change in the industry-adjusted ratio of operating income-to-sales, denoted as Δ OPI, for the buyer firms. Leverage is also concerned to have negative impact on investment for diversified firms (see Ahn et al., 2006). Hence, Δ LEVERAGE controls for the leverage effect., defined as the change in the ratio of debt to asset, A possible concern is that the relation between investment efficiency or value changes and governance is affected by whether the acquired asset with particularly good (or poor) opportunities. To control for this possibility, we follow Rajan et al. (2000) to include the change in the inverse of Tobin s q, denoted as Δ Inverseq, as a control variable. Some studies document that investment decisions might be influenced by market valuation to the firm, especially when stock prices deviate from fundamental (Stein, 2003; Polk and Sapienza, 2006). Hence, we control the mispricing effect, denoted as Δ MISPRICING. We also control the size effect, denoted as Δ SIZE. The change in firm size ( Δ SIZE ) is the square root of the change in the firm s sales. We also concern that level of firm focus may have different impact on the firm performance (John and Ofek, 1995; Comment and Jarrell, 1995; Campa and Kedia, 2002). Hence, the sales-based Herfindahl index is used to control the focus effect. This index is calculated across all business segments as the sum of the squares of each segment s sales as a proportion of total sales 6. The close Herfindahl index is to 6 According to John and Ofek (1995), a sales-based Herfindahl index, H, is reflects the degree to which sales are concentrated in a few of a company s business segments. It is computed across n business segments as the sum of the squares of each segment i s sales, S i, as a proportion of total sales. 14

15 one, the more concentrated are the firm s sales within a few of its segments. 4. Univariate comparison of corporate governance variables In this section, we discuss the governance variables and examine their relation with the investment efficiency of diversified firms. According to Denis and McConnell (2003) and Cremers and Nair (2005) and others, our governance includes five different perspectives: board characteristics, ownership structure, CEO compensation, audit committee, and market for corporate control. We measure changes in investment allocations from the fiscal year prior to the announcement of the asset purchases (Year -1) through the fiscal year following completion of the purchase (Year +1) for the sample of multi-segments firms. Table 2 summarizes the definitions to each governance variables and predictions and empirical results to the change in investment efficiency and firm excess value. [Insert Table 2 near hear] 4.1 Board characteristics Financial literature stresses the role of board (Monks and Minow, 1995; Fama and Jensen,1983) to improve the quality of manager s decision. Jensen (1993) suggests the smaller boards are more effective monitors than larger boards, because smaller boards mitigate the free rider problem and effectively discipline the management (Yermack,1996). Besides the monitoring function, the board may play its role as to serve the management team (Monks and Minow, 1995; Chaganti et al.,1985). Boone et al. (2007) find that firms with multiple segments have larger board size. Table 3 reports descriptive statistics on the mean and median change in investment efficiency for multi-segments firms by different level of governance mechanisms surrounding the asset purchase. Panel A through panel C of Table 3 investigate the mean and median change in investment ratio by different measures of investment efficiency. Board characteristics variables include board size, board H = n i= 1 [ Si n i= 1 ] S i 2 15

16 independence, leadership structure, and busy board. The results show that the diversified firms with relatively larger board size experience much improvement in investment efficiency post asset purchase. Fama (1980) and Fama and Jensen (1983) suggest the importance of board independence to effectively monitor management decisions. Many studies provide the supportive evidence (Gibbs,1993; Beasley et al., 2000; Anderson et al., 2000; Farber, 2005; Fich and Shivdasani, 2006; Boone et al., 2007). On the other hand, the concern of cost delivering information to outside directors may deter board independence (Baysinger and Hoskisson,1990). Board independence is defined as the number of the outside directors in the board. We define outside directors as directors who not have an executive position in the firm, not had such a position in the past, or not are related to an executive. The mean and median differences in changes in relative investment ratio between the firms with higher board independence and lower board independence are significant at the 1% level in Table 3. Fama and Jensen (1983) and Jensen (1993) stress the importance of separating the CEO and chairman positions to have effective board. Their argument has been supported by the evidence of Goyal and Park (2002) and Lehn and Zhao (2006). However, as suggested by Brickley et al. (1997), combining these two positions in one person may be efficient for a diversified firm because the board chairman is not expertise in all industrial knowledge for the firm operating in multiple-businesses. In our study leadership structure is a dummy variable equal to one for the chairman of the board of diversified firms serving as chief executive officer and zero otherwise. The results in Table 3 indicate combining both CEO and chairman within a diversified firm does not deter the investment behavior. Another variable related to the board characteristics is the busyness of board. Previous studies raise the concern that too busy directors to effectively monitor the management of so many firms (Ferris et al., 2003; Fich and Shivdasani, 2006). We follow the study of Fich and Shivdasani (2006) and consider directors busy if they serve on three or more boards. Then we construct a (0, 1) indicator that takes the value of one if 50% or more of the board s directors are busy. Throughout the paper, we refer to this variable as the busy board indicator. The mean and median 16

17 differences in changes in relative investment ratio between the firms with a busy board and the firms without a busy board are significant at the 1% level in Table 3. This indicates that the investment allocation of a diversified firm with busy board is not as efficient as one with less busy board. 4.2 Ownership structure Although there is no consensus, some studies indicate that ownership structure is an important factor of firm decisions and have a casual relation with firm value Leland and Pyle, 1977; Hoskinson et al., 1994; Cho, 1998). Inside ownership can align the interests of the shareholders and management but can also have an entrenchment effect allowing management to grab private benefits at the expense of all shareholders (Stulz, 1988).Hence, the ownership structure of a firm may be chosen endogenously (Demsetz and Lehn, 1985). Denis et al. (1997) find that alignment effect of inside ownership is crucial to the diversified firms. They provide the evidence that the level of managerial ownership has significantly negative impact on firm diversification. The study of Lehn and Zhao (2006) also provide an instance that managerial incentive has important impact on the efficient investment decision and firm performance. Our ownership variables include ownership by insiders, blockholders, 18 largest public pension funds, and outside director s ownership. The results in Table 3 indicate that the change in investment ratio is not significantly different between the firms with high inside ownership and those with low inside ownership. Shleifer and Vishny (1986, 1997) argue that blockholder ownership is the most direct way to align cash flow and control rights of outside investors and effectively limit management inefficient investment decisions. Shome and Singh (1995) and Allen and Phillips (2000) show that there is positive relationship between blockholder ownership and financial performance. The results in Table 3 indicate that the change in investment ratio is not significantly related to different level of blockholder ownership. The result is consistent with the idea of Demsetz and Lehn (1985) that a firm may endogenously choose its optimal inside and blockholder ownership. Bushee (1998) argues that large stockholding and sophistication of institutional investors are more likely to monitor and discipline the managers so that managers 17

18 invest efficiently to maximize the firm value. However, different institutions have different incentives to monitor 7. Cremers and Nair (2005) discuss that public pension funds are generally more free from conflicts of interest and corporate pressure than other institutional shareholders, so they are easier to align the managerial interests. Followed Cremers and Nair (2005), we measure the percentage of equity ownership of 18 largest public pension funds 8 to proxy for institutional ownership. Throughout this paper, the percentage of equity ownership held by 18 largest public pension funds is used to represent the institutional ownership. The results in Table 3 indicate that diversified firms with higher cumulative percentage of common stock shares of the firm held by the 18 largest public pension funds experience larger improvement in investment ratio surrounding the transaction. The difference is all significant at least 5% level. Jensen (1993) states that lacking equity holdings of board members is one of the reasons that attributes to the failure of internal control system. Yermack (2004) document that higher performance incentives to outside directors in terms of equity and stock options significantly enhance firm performance. We define that outside directors ownership is the percentage of equity ownership held by outside directors. Our results in Table 3 show there is a significant difference of change in investment efficiency between the diversified firms with higher outside director ownership and those with lower outside director ownership. 4.3 CEO compensation Fama and Miller (1972) address the importance of managerial incentive to investment decision. Well compensation design may align the interest of managers 7 Pound (1988) documents that institutions such as banks and insurance companies are more likely to support management in proxy contests due to conflicts of interest. Gillan and Starks (2000) stress the important role of institutional ownership in protecting the shareholder rights. Thomsen and Pedersen (2000) find that financial investor ownership is associated with the shareholder value and profitability. 8 We follow Cremers and Nair (2005) to calculate the percentage of shares held by 18 largest public pension funds. The list of the public pension funds includes California Public Employees Retirement System, California State Teachers Retirement, Colorado Public Employees Retirement Association, Florida State Board of Administration, Illinois State Universities Retirement System, Kentucky Teachers Retirement System, Maryland State Retirement and Pension System, Michigan State Treasury, Montana Board of Investment, New Mexico Educational Retirement Board, New York State Common Retirement Fund, New York State Teachers Retirement System, Ohio Public Employees Retirement System, Ohio School Employees Retirement System, Ohio State Teachers Retirement System, Texas Teachers Retirement System, Virginia Retirement System, State of Wisconsin Investment Board. 18

19 with shareholders and obtain optimal investment (Bizjak et al.,1993; Datta et al., 2001). On the other hand, compensation could also reflect the CEO s bargaining power to extract more private from the board and thus could indicate managerial entrenchment (Ahn and Walker, 2007). The study of Harford and Li (2007) provide the counter evidence that CEO compensation schemes are not so effective to control the agency conflicts. We define that CEO equity-based pay is the percentage of equity-based compensation in CEO s total compensation, with equity-based pay defined as the value of stock option and restricted stock grants. It is found from Table 3 that the diversified firms with higher ratio of CEO equity-based pay experience larger improvement in investment ratio. Our results is consistent with Anderson et al. (2000) that diversified firms with higher CEO equity-based pay generally have higher excess value of firms. This may be due to the reason that CEO of diversified firms have higher incentive to choose profitable projects for the reason that the degree of unsystematic risk of CEO wealth is fully diversified compared with the focused firm (Gibbs, 1993). Another variable of CEO compensation is CEO ownership. Jensen and Meckling (1976) propose that increase of management ownership will enhance the firm valuation through alignment the interest of managers and stockholders. Many studies provide the evidence that investment or financing decisions of firms can be influenced by the level of CEO ownership (Agrawal and Mandelker, 1987; Kim et al., 1988; Lewellen et al., 1985, 1989; Morck et al., 1988). Similarly, CEO ownership structure can be determined endogenously to maximize firm performance. We do not find there is a significant difference of change in investment efficiency between the diversified firms with high CEO ownership and those with low CEO ownership. 4.4 Audit committee Accounting literature emphasizes the importance of auditing regime to enhance the internal corporate control by reducing the information asymmetry between the management and shareholders (Healy and Palepu, 2001; Bushman and Smith, 2002; Adams and Ferreira, 2007). Therefore, the role of audit committee is aid to align the interest between the agent and principal. The studies, including Spira (1999), Carcello and Neal (2000), Klein (2002), Beasley (1996), and Beasley et al. (2000), document that outside directors are able to independently monitor the management and improve 19

20 the effectiveness of the board. In addition, McMullen and Raghunandan (1996) find that the effectiveness of audit committee can by improved by increasing the number of meetings. Beasley et al. (2000) find that there is higher possibility of company fraud if the audit committee does not hold the meeting frequently. The results in Table 3 show that those diversified firms with more independent audit committee and with more audit committee meetings show significant improvement in investment ratio following asset purchases. 4.5 Market for corporate control A great deal of theory and evidence suggests that takeovers address governance problems (Jensen,1988; Jensen and Ruback, 1993; and Scharfstein, 1988; and among others). A poorly performed firm can resist a takeover by adopting antitakeover provisions (ATPs) in its charter. If a firm is protected against a takeover, it is viewed a firm with poor external governance. Gompers et al.(2003) develop a Governance Index to proxy for the level of shareholder rights and find that firms have higher valuation when they have stronger shareholder rights. The studies of La Porta et al. (2000) and Dittmar and Mahrt-Smith (2007) indicate that effective external governance has a substantial impact on firm valuation. We follow the concept of Gompers et al. (2003) and use the Gompers et al. (2003) index as the proxy for external governance. A firm is thought to be a better to be a better external-governed firm when it has lower score of governance index. The results in Table 3 indicate that different degree of managerial entrenchment due to takeover protection has different impact on the change in investment ratio and the difference between these two groups of firms is significant at the 1% level. [Insert Table 3 near hear] Panel B and Panel C of Table 3 present the results by different measures of investment efficiency. It shows similar results that diversified firms with larger board size, higher board independence, less percentage of busy directors, higher shareholding of 18 largest public pension funds, higher outside director ownership, higher quality of audit committee and less takeover protection experience greater mean and median improvement in investment ratio surrounding asset purchases. 20

21 5. Changes in investment allocations and corporate governance To provide the further evidence on how the change in investment ratio of diversified firms is related to corporate governance structure, we examine board characteristics, ownership structure, CEO compensation, the role of audit committee and the market for corporate control in a multivariate framework. Table 4 reports the estimates of ordinary least squares (OLS) regressions of the change in investment allocations on the measures of corporate governance. [Insert Table 4 near hear] The dependent variable for all of the regression is the change in investment ratio for diversified firms. As control variables we include DumDI, DumFit, DumAgency, ΔCAPEX, ΔOPI, ΔLEVERAGE, ΔInverse, ΔMispricing, ΔSIZE and Herfindahl in the regression. Panel A shows the results using the measure of relative investment efficiency as the dependent variable, panel B and panel C show the results using the measure of relative value-added by allocation and absolute value-added by allocation as the dependent variable, respectively. In each model we include important control variables which are shown to have impacts on the firm investment decision in the literature. Panel A of Table 4 shows the regression results by the measure of change in relative investment efficiency on governance variables. Model 1 to 4 of panel A presents the relation between the change in relative investment efficiency of diversified firms and board characteristics. Model 1 documents the relation between the change in relative investment efficiency of diversified firms and Board size. Similar to the studies of Monks and Minow (1995) and Chaganti et al. (1985), we find that board size is positively related to the change in relative investment efficiency of diversified firms. Although it is inconsistent to the concept of Jensen (1993), our finding provides an evidence that multi-segment firms need more sophisticated board to serve the management team so as to implement a better asset acquisition decision. Model 2 examines the relation between the change in relative investment efficiency of diversified firms and Board independence. Consistent with Beasley et al. 21

22 (2000) and Farber (2005), model 2 indicates that higher board independence of diversified firms tend to effectively monitor top management and implement a better investment decision. Our result provides one of explanations to the result of Anderson et al. (2000) which shows diversified firms utilize more outside directors and finally lead to higher excess value on diversified firms. Model 3 tests the effect of Leadership structure on the change in relative investment efficiency of diversified firms. Inconsistent with Fama and Jensen (1983) and Jensen (1993), we do not find the evidence that combining the role of CEO and chairman of the board in one individual significantly reduces investment efficiency. This maybe due to the reason suggested by Brickley et al. (1997) that drawback of ineffective monitoring role combing two positions in one person may be offset by cost-saving benefit of transferring information. Model 4 addresses the role of Busy board documented in Ferris et al. (2003) and Fich and Shivdasani (2006). Consistent with Fich and Shivdasani (2006), we find that the extent of board busyness is negatively related to the change in relative investment efficiency. Our study provides the evidence that directors either inside directors or outside directors who hold directorships in multiple companies have less effectiveness on monitoring and cause negative impacts on firm investment efficiency. Model 5 to 8 present the relation between the change in relative investment efficiency of diversified firms and ownership structure. In each model we use different measures proxy for ownership mechanisms. Model 5 and model 6 present the effect of Insider ownership and Blockholder ownership on the change in relative investment efficiency of diversified firms. Similar to the results in Denis et al. (1997), we do not find strong evidence that insider ownership plays an important role in determining effective investment decisions. Our finding also provides the evidence that insider ownership structure may be chosen endogenously by diversified firms in value-maximizing way, as suggested by Demsetz and Lehn (1985). Model 7 presents the relation between the change in relative investment efficiency of diversified firms and Institutional ownership. We find that ownership of 18 largest public pension funds is positively related to the change in relative investment efficiency. Our evidence shows that diversified firms with higher percentage of institutional ownership which is more free from conflicts of interest and 22

23 corporate pressure experience larger improvement in investment efficiency post asset purchases. Jensen (1993) stresses the importance of alignment effect of outside director and management team. Model 8 presents the result of the relation between the change in relative investment efficiency of diversified firms and Outside director ownership. Consistent with Yermack (2004), we find that higher percentage of outsider ownership is effectively to enhance firm investment efficiency. Model 9 and 10 examine the effect of CEO compensation on the change in relative investment efficiency. Model 9 documents that the relation between the change in relative investment efficiency of diversified firms and CEO equity-based pay. We find that CEO equity-based pay is positively related to the change in relative investment efficiency. Although inconsistent with Anderson et al. (2000), our result support the finding of Datta et al. (2001) that managerial equity-based compensation plays a crucial role in terms of investment efficiency. Besides, we do not find that there is significant relation between the CEO ownership and the change in relative investment efficiency. Model 11 and 12 examine the relation between the audit quality and the change in relative investment efficiency. Model 11 uses the dummy of Outside directors on audit committee to proxy for audit quality. Model 12 uses the Number of audit committee meetings to proxy for audit quality. We find that the role of audit committee is important in terms of independent monitoring the management and effective decision making. Finally, model 13 presents the relation between the level of shareholder protection (Market for corporate control) and the change in relative investment efficiency. Consistent with Gompers et al. (2003), we find that firms with stronger shareholder rights adopt better asset acquisition decision and enhance their investment efficiency. Among all the models in Table 4, we also find that the coefficients of DumDI and DumFit are significant and their sign are consistent with our prediction. In 23