What Do CEOs Contribute to Firm Value? How Do CEOs ...

1 1 9-20-13 what do ceos Contribute to Firm Value? How do ceos Contribute to Firm Value? Alexander R. Bolinger Idaho State University Jeffrey T. Brookman Idaho State University Paul D. Thistle University of Nevada Las Vegas 2 ABSTRACT The question of what CEOs Contribute to firm value is important for both public policy and business practice and is the subject of ongoing debate in strategic management. In the current study, we introduce the group connection methodology to ensure that CEO effects are identified in investigating the effects of CEOs on firm value.

2 We also introduce a decomposition of performance into a market value channel, a leverage channel and a current earnings channel. We find that CEO effects are substantial and are the most strongly associated with firm value. CEOs have large effects on current earnings and leverage channels of organizations, but smaller effect on the market value channel. Keywords: CEOs; CEO effect; executive compensation; group connection method 3 what do ceos Contribute TO FIRM VALUE? HOW do ceos Contribute TO FIRM VALUE? INTRODUCTION The extent to which CEOs influence firm performance been a question at the center of a debate that has persisted among both academics and practitioners ( , Carpenter, Geletkanycz, & Sanders, 2004; Drucker, 1954; Schein, 1992).

3 On the one hand, some empirical work has shown that managerial skill is only modestly related to firm performance (Demsetz and Lehn, 1985; Bebchuck and Fried, 2004). From this perspective, exogenous factors such as environmental shocks and industry and organizational constraints are believed to limit an executive s ability to make meaningful changes to influence firm outcomes (DiMaggio & Powell, 1983; Hambrick & Finkelstein, 1987; Hannan & Freeman, 1989). Similarly, the romance of leadership perspective suggests that improvements in firm performance are often attributed to CEOs after the fact (Meindl & Ehrlich, 1987; Meindl, Ehrlich, & Dukerich, 1985).

4 On the other hand, many theorists and practitioners argue that firm leaders have an important effect on firm performance. Top managers may influence firm performance by creating an effective organizational culture (Drucker, 1973; Sanders, 2001; Schein, 1992) or by effectively navigating challenges from the organization s environment (Carpenter, Sanders, & Gregersen, 2001; Lawrence & Lorsh, 1967; Woodward, 1965). Some empirical findings support the contention that CEOs have a relatively modest effect on firm performance. For instance, Lieberson and O Connor (1972) examine sales, earnings and profit margin as performance metrics and find that CEO effects account for to of the variance in performance.

5 Similarly, Thomas (1988) finds that CEO effects account for about 4% to 7% of the variation in firm performance, whereas Crossland and Hambrick (2007) find that, 4 after accounting for variation across countries, CEO effects explain approximately 5% to 13% of the variation in ROA. Wasserman, Anand and Nohria (2010) also find considerable variation across industries, but their research suggests that on average CEO effects account for 15% of the variation in ROA and 14% of the variation in Tobin s Q. However, findings that the effects of CEOs on firm performance are modest have been criticized on methodological grounds ( , Hambrick and Mason 1984, Day and Lord, 1988).

6 In particular, these studies have relied on sequential variance decomposition. The potential problem with this methodology is that, in general, the results depend on the order in which variables are entered into the analysis. Replicating Lieberson and O Connor s study, Weiner (1978) found that, when entered last the CEO effect accounted for 9% of profit margin, but when entered first, the CEO effect accounted for 78% of the variability in profit margins. In this paper we address two related issues. First, we analyze how much CEOs Contribute to firm value. This issue is the subject of a long discussion that goes back to some of the earliest work in strategic management ( , Barnard, 1938; Selznick, 1957).

7 Recently, however, researchers have called for methodological innovations to re-examine the possibility that previous findings have underestimated CEOs contribution to firm value ( , Hambrick & Quigley, 2013; Mackey, 2008). Second, we analyze the channels through which CEOs Contribute to firm value, an issue that has received little attention in the strategic management literature. We show that firm value can be determined by three components: a market value component, a leverage component, and a current earnings component. We then analyze how much CEOs Contribute to each of these components as well as how much they Contribute to overall firm value.

8 5 Calculating CEO Effects on Firm Value Mackey (2008) raises additional methodological issues which, she argues, bias prior analyses toward finding modest effects of CEOs on performance. She points out that the firm effect or the industry effect may be nested within the other effect. More importantly, in data sets that do not track the employment changes, CEO effects are nested within either industry or firm effects. Mackey also points out that if a firm has the same CEO during the sample period then the CEO and firm effects are nested. In these situations the CEO and firm effects are, in general, not identified and cannot be estimated separately.

9 To measure the effects of CEOs on firm behavior and performance it is important that CEO effects be identified, that is, it is important that we can obtain unique estimates of the relevant parameters. There are two empirical strategies for identifying CEO effects separately from firm effects. The first strategy is to use a mover sample, that is, a sample consisting only of CEOs that have moved and been CEO at more than one company ( , Bertrand and Schoar, 2003). The second strategy is to use group connection (Abowd, Kramerz and Margolis, 1999). The mover sample strategy works by eliminating all of the nested CEOs effects in the sample.

10 Abowd, Kramerz and Margolis (1999) demonstrate that group connection is necessary and sufficient to separately indentify CEO and firm effects. To see this intuitively, consider a CEO that switches firms once during the sample period. The fixed effect for that CEO can be estimated from the fixed effects for the two firms that employed the manager. In addition, using the firm fixed effects allows the fixed effects for all other CEOs at those two firms to be estimated, even if they did not move. Removing all firms that never had a CEO move leaves groups of firms connected by CEO movement.