Does the Location of Directors Matter?

s delegated monitors of top management on behalf of shareholders, corporate boards of directors rely critically on information about the firm in making governance decisions. Theoretical research in corporate governance shows how a board’s ability to obtain and use information is closely related to key aspects of board structure, such as size and independence (Raheja (2005), Harris and Raviv (2008)). Complementing the theoretical literature, a number of empirical studies find evidence to suggest that board size and board independence can be explained to some extent by the complexity of firms’ operations and outside directors’ costs of acquiring information (see, e.g., Boone, Field, Karpoff, and Raheja (2007), Coles, Daniel, and Naveen (2008), and Linck, Netter, and Yang (2008)).

Motivated by these studies, in our paper, Does the Location of Directors Matter? Information Acquisition and Board Decisions, forthcoming in the Journal of Quantitative and Financial Analysis published by Cambridge University Press, we examine empirically a new dimension of board structure that we expect to influence the costs of information acquisition, namely, geographic distance between directors and headquarters. Geographic distance has been shown to matter for the gathering of information in a wide variety of financial contexts, including bank lending (Petersen and Rajan (1994, 2002)), venture capital financing (Lerner (1995)), equity analysis (Malloy (2005), Bae, Stulz, and Tan (2008)), bond underwriting (Butler (2008)), investment management (Coval and Moskowitz (1999, 2001)), and regulatory enforcement (Kedia and Rajgopal (2011)). Recent research has also begun to examine the effects of geography on the structure and decisions of boards (Masulis, Wang, and Xie (2011); Knyazeva, Knyazeva, and Masulis (2012)). To date, the lack of detailed data on where individual directors reside vis-à-vis corporate headquarters has been an obstacle to extending research on board geography. In this paper, we overcome this obstacle by constructing a database of over 4,000 residential addresses of outside directors at S&P 1500 firms during 2004-2007.

We use the director-level residential data to construct several measures of the board’s distance from headquarters. These measures allow us to investigate how board distance relates to information acquisition and board decisions. Our analysis reveals that when a firm’s assets are more intangible and thus quantitative performance measures are less informative about managerial effort, the board tends to be located closer to headquarters (e.g., there is a larger fraction of unaffiliated directors who reside within 100 kilometers). We also find that more remote boards tie CEO dismissal decisions and CEO incentive compensation more strongly to stock price performance.

Our analysis builds upon the premise that residing farther from headquarters increases directors’ costs of obtaining certain types of information. Some kinds of information about management performance (e.g., stock prices) can be easily acquired by remote directors, but other kinds can only be obtained by directors who are in close proximity to the information source. Petersen (2004) defines “soft” information as information that cannot be codified and transferred across geographic distance. Indeed, soft information can only be acquired from personal observation or face-to-face interactions (Stein (2002), Petersen (2004)). Examples of soft information that directors might obtain include personal assessments of employees or operations, impressions from face-to-face meetings with management, or inferences about local business conditions.

Two implications follow from the idea that distance increases directors’ costs of obtaining soft information. First, in equilibrium, a board’s distance from headquarters should be negatively related to firm characteristics that proxy for directors’ need to acquire soft information. Second, more distant boards will tend to rely more heavily on hard, public information in making key monitoring decisions, such as whether to dismiss a CEO or how to set the CEO’s compensation.

Consistent with the first implication, we find in multivariate analyses that board distance—measured as the fraction of unaffiliated directors who reside more than 100 kilometers from firm headquarters—is negatively related to the fraction of the firm’s assets consisting of intangibles. Supporting the second implication, we find that board distance is significantly related to CEO dismissal and CEO compensation decisions. In multivariate logit regressions, we document that the occurrence of non-routine CEO turnover events is more sensitive to poor industry-adjusted stock performance when boards are farther from headquarters. In other tests, we find that greater board distances are associated with higher levels of CEO equity-based pay, more equity pay relative to other forms of compensation, and higher pay-for- performance sensitivities.

Our results are robust to the use of several alternative measures of board distance including (i) an indicator for whether or not at least 50% of the unaffiliated directors on the board live farther than 100 kilometers from headquarters; (ii) a continuous measure based on the median distance among a board’s unaffiliated directors; and (iii) the fraction of unaffiliated directors who are geographically separated from headquarters by substantial driving times. We also conduct tests to rule out the possibility that our results merely reflect the influence of other factors, such as social ties and social interactions between the CEO and directors, CEO power and influence over director selection, the Sarbanes-Oxley Act, regional effects, or the particular location of headquarters. Our main results hold in all of these tests, providing strong support for the view that board distance not only reflects information-gathering costs, but also shapes how directors use different types of information in their governance decisions.

A small number of contemporaneous studies also examine aspects of board geography. Masulis, Wang, and Xie (2011) investigate the implications of foreign directors for corporate governance and firm performance. They document that firms with foreign directors on the board pay CEOs higher compensation and have lower sensitivity of CEO turnover to performance. In addition to providing evidence on how very large geographic distances can affect monitoring by the board, their work suggests the importance of cross-country differences in accounting rules, regulations, and social and cultural norms. Knyazeva, Knyazeva, and Masulis (2012) show that geographic proximity of a firm’s headquarters to large pools of director talent strongly influences the firm’s use of independent directors and directors with specialized expertise. Our results complement these findings by establishing that distance-related costs of information acquisition, together with proximity of headquarters to an available supply of director talent, helps to determine a board’s overall location relative to headquarters.

Our work is related to the broader literature on the determinants and implications of board structure. Recent work shows that a single board size or fraction of outsiders is unlikely to be optimal for all firms (Boone et al. (2007), Linck, Netter, and Yang (2008), and Coles, Daniel, and Naveen (2008)). We add to this strand of literature by showing that board location reflects a tradeoff between director expertise and information-gathering costs, and thus no single board distance is likely to be best for all firms. Our research on board geography also extends a growing literature that explores new dimensions of board structure, such as social ties between directors and CEOs (Hwang and Kim (2009), Fracassi and Tate (2012)), links between corporate directors and mutual fund managers (Cohen, Frazzini, and Malloy (2008)), busy directors (Ferris, Jagannathan, and Pritchard (2003), Fich and Shivdasani (2006)), and directors’ outside career opportunities (Booth and Deli (1996), Mobbs (2009), Masulis and Mobbs (2011)).

The full paper is available for download here.