مرکزیت هیأت مدیره و عملکرد شرکت
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|13090||2013||26 صفحه PDF||سفارش دهید||محاسبه نشده|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Journal of Accounting and Economics, Volume 55, Issues 2–3, April–May 2013, Pages 225–250
Firms with central boards of directors earn superior risk-adjusted stock returns. A long (short) position in the most (least) central firms earns average annual returns of 4.68%. Firms with central boards also experience higher future return-on-assets growth and more positive analyst forecast errors. Return prediction, return-on-assets growth, and analyst errors are concentrated among high growth opportunity firms or firms confronting adverse circumstances, consistent with boardroom connections mattering most for firms standing to benefit most from information and resources exchanged through boardroom networks. Overall, our results suggest that director networks provide economic benefits that are not immediately reflected in stock prices.
Social and economic networks are a central feature of virtually all economic activities. These networks serve as a conduit for interpersonal and interorganizational support, influence, and information flow. The links between individuals in these networks are the channels by which information is communicated, resources are exchanged, new relationships are formed, and existing relationships are leveraged. Economists and sociologists have long studied the influence of social networks on labor markets, political outcomes, and information diffusion. One important network in corporate finance is the boardroom network formed by shared board directorates. While several studies examine the structure of boardroom social networks, why they form, and their theoretical impact on firm performance, relatively few studies provide empirical evidence to assess the net economic impact of these networks on firm performance. In this paper, we directly investigate the empirical relations between a board's well-connectedness and the firm's future performance. Our empirical investigation is important because, ex-ante, there are no clear predictions on the relation between a firm's performance and its board's well-connectedness. A vast literature in organizational sociology, economics, and finance highlights both potential benefits and costs associated with being well-networked. The potential benefits of having well-connected boards can take several forms. First, directors possess a wealth of information on industry trends, market conditions, regulatory changes, and other key market data, which can flow across the boardroom network. Well-connected boards may have better access to this information and a comparative advantage in making strategic decisions ( Mizruchi, 1990 and Mol, 2001). Second, boardroom networks allow firms to leverage social relationships and reduce asymmetric information when designing contracts ( Schoorman et al., 1981). Both factors may improve the terms of contracts between firms. Third, directors possess important and useful business contacts accessible through the boardroom network, contacts that can be sources of useful business relationships (e.g., clients, suppliers) or sources of other economic benefits and resource exchange (e.g., personal and political favors) (see, for example, Mol, 2001 and Nicholson et al., 2004). Fourth, the boardroom network may be a mechanism of information transmission through which value-improving business innovations can spread ( Haunschild and Beckman, 1998). For example, firms may learn about effective corporate governance mechanisms, efficiency-enhancing technology, and innovative compensation structures through the boardroom network. Finally, the boardroom network represents a channel of communication or resource exchange between companies and can facilitate collusive competitive behavior and yield economic benefits for a set of closely linked firms ( Pennings, 1980). The existing literature also highlights several reasons why having a well-connected board may adversely affect firm performance. First, the boardroom network may propagate value-decreasing management practices. For example, the boardroom network has been found as an important explanation for the spread of options backdating (Bizjak et al., 2009, Snyder et al., 2009 and Armstrong and Larcker, 2009). Second, to the extent that having a well-connected board requires its members to serve on many board seats, directors of well-connected boards may devote limited attention to the monitoring and strategic advising of each company. Therefore, there may be a trade-off between well-connectedness and monitoring effort or intensity. This is consistent with the idea that the number of board positions a director holds (or busyness) is negatively associated with monitoring efforts and shareholder wealth (Core et al., 1999, Fich and White, 2003, Loderer and Peyer, 2002 and Fich and Shivdasani, 2006). Third, misleading or incorrect information may spread though the board network, resulting in value-decreasing strategies and investments. Finally, although collusion can have a positive impact on shareholder value, the resulting regulatory, litigation, and reputation costs can produce net losses of shareholder value. The collective arguments from the literature on boardroom networks highlight the ex-ante ambiguity regarding the net economic impact of a board's well-connectedness, and this association is therefore an open empirical question. Resolution of this ambiguity is hampered by the fact that most empirical network studies focus on interpersonal relationships between specific agents within an isolated context, such as between a firm and a lender in determining credit terms or a manager and a security analyst in determining analyst recommendations (e.g., Engelberg et al., 2012; Cohen et al., 2010). An innovation of our paper is that we take a macro-level (or “bird's eye”) view of the association between boards' well-connectedness and firm performance. As we explain below, we build the corporate network of shared directorates and measure the relative positioning of boards in the network as a means to aggregate the micro foundations established in prior research. We then assess the balance of the potential costs and benefits associated with a board's “centrality” in the networks and establish several important regularities regarding the relation between board centrality and multiple measures of firm performance. The construct of interest in our study is the “well-connectedness” of boards established by their directors' formal or professional ties. We conceptualize shared directorates between two boards as channels of information or resource exchange, and study a board's well-connectedness through such channels using standard tools of analysis developed by social network theory. A well-connected board is one that is central to the network's aggregate flow of information and resources. The concept of well-connectedness is inherently multidimensional. Network theory has developed multiple related but distinct notions of well-connectedness. First, a board may be well-connected if it possesses relatively many channels of communication or resource exchange, yielding such a board more opportunities or alternatives than otherwise comparable firms (measured by DEGREE centrality). Second, a board may be well-connected if it possesses relatively closer ties to outside boards (i.e., there are fewer steps between boards), making information or resource exchange quicker and more readily available (measured by CLOSENESS centrality). Third, a board may be well-connected if it lies on relatively more paths between pairs of outside boards, making such a company a key broker of information or resource exchange (measured by BETWEENNESS centrality). We consider a fourth and related notion, stemming from a refinement of DEGREE centrality, which recognizes that having more direct connections is more influential when such connections can reach or influence more outside boards. In other words, a board is well-connected when its direct contacts are also well-connected (measured by EIGENVECTOR centrality).1 Using a comprehensive sample of 115,411 directors from 2000 to 2007, we build the U.S. corporate boardroom network formed by shared directorates in each year. For each year, we measure each board's well-connectedness in the aggregate boardroom network using the four standard measures from the networks literature described above, as well as a composite score, which we call “N-Score,” based on the average of the four standard measures. Using these five measures of a board's well-connectedness, we find that firms with the best-connected boards on average earn substantially higher future excess returns compared to firms with the worst-connected boards. The most central firms (i.e., in the highest centrality quintile) outperform the least central firms (i.e., in the lowest centrality quintile) by an average of 4.68% per year following the portfolio formation; this association holds after controlling for the influence of industry membership, size, book-to-market, and momentum. We also demonstrate that the link between board connectedness and firm performance is robust over time and across industries and is robust to a standard set of governance controls. We interpret the positive centrality–return relation as evidence that, all else equal, firms on average experience a net benefit from having a relatively well-connected board and that equity prices under-react to this information. Alternatively, this positive relation reflects compensation for unobserved risk factors. While we can never fully rule out risk-based explanations, we design a series of tests to assess the relative merits of risk- and mispricing-based explanations. Our first test is motivated by the idea that boardroom networks provide access to resources such as shared contacts, best management practices, and improved terms of contracts. We hypothesize that the associations between boards' well-connectedness and firm performance are more pronounced among firms that stand to benefit most from such resources. Consistent with this hypothesis, the return difference between best- versus worst-networked firms is stronger among firms with high growth potential (i.e., young firms or those with low book-to-market ratios), and firms confronting adverse circumstances (with low return-on-assets or poor historical stock price performance). However, such empirical patterns could still be consistent with a risk-based hypothesis, which motivates the following two additional tests. Our second test is motivated by the hypothesis that the relation between board connectedness and future returns reflects a potential market under-reaction to an unexpected shift in a firm's fundamentals. Consistent with such a hypothesis, we document a positive association between the board's well-connectedness and changes in future profitability as measured by return-on-assets (ROA). Specifically, the changes in ROA of best-networked firms are on average two percentage points higher than those of the worst-networked firms in the year following portfolio formation. In addition, the relation between well-connectedness and future ROA improvements persists into the second year following portfolio formation. Consistent with the results of our return-based tests, we find that improvements in operating profitability associated with board connectedness are more pronounced among firms with high growth potential or firms confronting adverse circumstances. Our third test uses analyst forecast errors as a proxy for the under-reaction to the performance-relevant information embedded in board connectedness. We find that firms with relatively better-connected boards are more likely to have realized earnings that exceed the consensus forecast, and, conversely, firms with relatively less-connected boards are more likely to have realized earnings below the consensus forecast. Consistent with our earlier results, this association is particularly pronounced in firms with high growth potential or those confronting adverse circumstances. These results are consistent with analysts and the market failing to fully appreciate the net economic benefits associated with a company's well-connectedness in the boardroom network, particularly for those firms that stand to benefit the most from network resources. Collectively, the results of our paper provide evidence that firms, on average, experience a net benefit from being better connected in the board of director network and that equity prices under-react to this information. However, it is important to acknowledge that this interpretation is subject to several theoretical and measurement concerns. While studies such as Demarzo et al. (2003) and Ballester et al. (2006) provide rigorous motivation for the network measures used in our study (especially the EIGENVECTOR centrality measure), these theoretical results are based on very specific and fairly simple functional forms for network interactions between participants and for agent utilities. It remains an open question whether our standard centrality measures are consistent with agent interactions in more complex, realistic, and general environments. Our modeling of the connectedness of boards through shared directorates is also based on several important implicit assumptions. One supposition is that the linkages between firms based on shared board members represent the primary channels of social, informational, and resource exchange between the leadership of companies. It is possible that directors' networks extend beyond those associated with “formal” board appointments and are substantively influenced by “informal” social or nonprofessional connections. While we may not fully capture the breadth of directors' informal social networks, Hwang and Kim (2011) and Westphal et al. (2006) suggest that formal and informal networks are positively correlated and can be strategically complementary. However, from the perspective of documenting potential mispricing, measuring boards' formal network is more important than their informal network, since the former is more readily observable by market participants. Moreover, our network measures make implicit assumptions about the manner in which traffic flows through the corporate boardroom network. For example, BETWEENNESS and CLOSENESS centrality measures focus on the shortest paths between network nodes while ignoring other paths. The use of these measures assumes that information and resources flow through the shortest possible paths between network nodes despite longer paths potentially serving a similar role. Similarly, our measures of centrality assume that geographic proximity does not play a primary role in enhancing the social, resource, and information exchange between two companies, whereas prior research suggests that this may not be the case (e.g., Kedia and Rajgopal, 2009). As noted by Borgatti (2005) and Borgatti and Everett (2006), to the extent these assumptions do not capture the true flow of information and resources across boardroom networks, the use of standard centrality measures may not be completely appropriate. We also acknowledge that there are plausible alternative causal interpretations of our empirical findings of a positive association between board connectedness and future firm performance. One possibility is that our measures of board connectedness are correlated with some unobserved or omitted firm characteristic that is associated with future firm performance. For example, if higher-quality directors are attracted to and are more likely to accept board positions on better-connected boards (Masulis and Mobbs, 2012), our findings of a positive association between board connectedness and future performance may simply reflect endogenous matching between high-quality directors and well-connected or prestigious firms. Another possible explanation is that well-connected board members prefer to sit on the boards of well-performing firms, or firms they correctly anticipate will perform well in the future. We perform several tests to examine whether these alternative explanations are likely to drive our main results and to assess the possibility that a well-connected board leads to better future performance. First, we document a significantly positive association between recent changes in board connectedness (from the previous year to the current year) and future abnormal stock returns among our full sample as well as the subset of firms whose board composition remain unchanged from the previous year. These tests mitigate concerns that our results are driven by firm characteristics, such as board prestige, that are relatively static over time and omitted from our tests. Next, to further mitigate concerns of endogenous matching between boards and firms, we continue to document a positive association between changes in board connectedness and firm performance for the subset of firms whose board composition and outside boardroom connections remain unchanged from the previous to the current year, although with attenuated statistical significance due to the substantially smaller sample size. Finally, we examine the determinants of changes in boardroom connectedness and find no evidence that firms' past performance is associated with future changes in board connectedness. It is therefore unlikely that our results are driven by well-connected directors being attracted to firms that are expected to perform well in the future. The collective evidence from these tests is consistent with boardroom connectedness leading to better future firm performance by providing access to resources and information. However, we acknowledge that we can never fully rule out other causal explanations, but we emphasize that the endogeneity concerns described above do little to explain the paper's central finding that investors appear to misprice the implications of boards' well-connectedness for firms' future performance. The remainder of the paper is organized as follows. Section 2 reviews related literature and Section 3 describes the data, the construction of the network, and summarizes network characteristics. Section 4 discusses our empirical results and robustness tests. Section 5 examines alternative explanation for our results. Section 6 provides the summary and conclusions from our study.
نتیجه گیری انگلیسی
Since the end of the twentieth century, the electric power industry in Japan has undergone deregulation. However, the regional monopoly over power generation, transmission, and supply, which was under the control of 10 regional power utility companies, was legally permitted for a number of years. As liberalization policies evolved, the Japan Electric Power Exchange (JEPX) was established, and trading in wholesale electric power started in April 2005. The purpose of this endeavor was to ensure a fair competition and invigorate the business of transmitting and distributing electric power all over the nation. In order to maximize the total utility of the whole economy, a market economy should be allowed to function effectively. Needless to discuss the cases of failure of planned economies, full advantage of the market mechanisms should be taken to optimally allocate the limited economic resources. The market economy is a system in which the action of each economic entity is optimized by transmitting the prices formed in the market to each economic entity as the integration of a variety of information. In other words, a market economy aims to reach the state where each price reflects information on the whole economy. However, in a real economy, all information cannot be instantaneously reflected in each price, and the information gaps that exist constantly lead to arbitration. As a result, the limited economic resources become optimally allocated asymptotically. It is therefore reasonable and realistic to restate the market economy as the system asymptotically aiming to optimize the allocation of limited resources. When examining the effectiveness of a certain real market economy, it is proper to discuss how early the prices reflect other information and to what extent the information impacts the prices. Fama (1970) systematically summarized the degree of market efficiency. When prices in a certain market are formed fully reflecting an information set, the market can be understood to be efficient with respect to the information set. If a market is efficient with respect to an information set, any investment strategies based on the information set cannot consistently achieve returns in excess of the average market returns. On the contrary, unless the market prices already reflect the information set, the investors who have that information set can gain extra through the market. This concept of information efficiency has been widely accepted in general as a measure to investigate the efficiency of a market. Granger (1969) shows that one time series C Granger-causes another time series R if the previous value of C is useful to explain the current value of R. Acceptance of the Granger-causality hypothesis does not necessarily imply acceptance of a true causality hypothesis. Conversely, true causality must certainly bring about Granger-causality. In other words, the absence of Granger-causality can be understood as the absence of true causality. Therefore, a Granger-causality test is often undertaken to examine the factor of price formation. In a market economy, the more rapidly and widely markets process information, the better the improvement of performance of the whole economy become. Therefore, it is essential to examine the information processing ability of markets to evaluate the state of the market economy. This study examines the information processing capability of Japan's wholesale electricity market, which may significantly affect the overall performance of the Japanese economy. More specifically, this paper analyzes the degree by which the JEPX trading prices reflect the historical prices and the market information that impacts the cost of power generation directly. After the 2011 earthquake off the Pacific coast of Tohoku and the Fukushima Daiichi nuclear disaster, regulations on the electric power industry are being reviewed. The most important contribution of this study is the conclusion that the oil, gas, and exchange markets do not Granger-cause the JEPX, which is reached through a lag-augmented vector autoregression (LA-VAR) test after proving the inefficiency of the JEPX through a variance-ratio (VR) test to discuss how Japan's electricity market should be. The remainder of this paper is organized as follows. Section 2 explains the methodology applied, Section 3 describes the data analyzed, Section 4 presents the empirical results, and Section 5 provides a summary and statement of conclusions.