Reaffirming the CEO effect is significant and much larger than chance: A comment on Fitza (2014)

• Author: Timothy J. Quigley,Scott D. Graffin
• Published date: 01 March 2017
• DOI: http://doi.org/10.1002/smj.2503
• Date: 01 March 2017

Strategic Management Journal

Strat. Mgmt. J.,38: 793–801 (2017)

Published online EarlyView 29 March 2016 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/smj.2503

Received 26 October 2015;Final revision received27 October 2015

COMMENTARY

REAFFIRMING THE CEO EFFECT IS SIGNIFICANT

AND MUCH LARGER THAN CHANCE: A COMMENT

ON FITZA (2014)

TIMOTHY J. QUIGLEY*and SCOTT D. GRAFFIN

Management Department, Terry College of Business, University of Georgia, Athens,

Georgia, U.S.A.

A recent study by Fitza argued that the prior estimates of the Chief Executive Ofcer (CEO)

effect are conated with events outside the CEO’s control, are largely the result of random

chance, and that the true CEO effect is smaller than has been previously estimated. We

suggest that the empirical methodology employed by Fitza to support these claims substantially

overstates the “random chance” element of the CEO effect.We replicate Fitza’s ndings, highlight

methodological issues, offer alternative conclusions, and using multilevel modeling (MLM),

suggest that his analyses mischaracterize the CEO effect. Copyright © 2016 John Wiley & Sons,

Ltd.

INTRODUCTION

There is a long history of debate centered on how

much impact Chief Executive Ofcers (CEOs)

have over the outcomes of the organizations they

lead (Chandler, 1962; Hambrick and Finkelstein,

1987; Hannan and Freeman, 1977; Mackey, 2008;

Quigley and Hambrick, 2015; Salancik and Pfeffer,

1977). Starting with Lieberson and O’Connor

(1972), a series of studies have sought to quantify

the “CEO effect,” frequently arguing that prior stud-

ies have over- or under-estimated the “true” effect

while seeking to demonstrate that a new approach

better captures the underlying phenomenon (Fitza,

2014; Hambrick and Quigley, 2014). Like all

empirical studies, this body of work relies on

association, not necessarily causality. Neverthe-

less, the plurality of prior work is suggestive of

Keywords: CEO effect; rm performance; variance

decomposition; upper echelons; CEOs

*Correspondence to: Timothy J. Quigley, Terry College of Busi-

ness, University of Georgia, 409 Brooks Hall, Athens, GA 30602,

U.S.A. E-mail: tquigley@uga.edu

Copyright © 2016 John Wiley & Sons, Ltd.

a causal link with estimates of the CEO effect

most frequently near 15 percent (for a review, see

Hambrick and Quigley, 2014; Mackey, 2008; also

see Hambrick and Quigley, 2014 for an approach

that estimates a signifcantly larger CEO effect).

Calling this consensus into question, however,

Fitza (2014) argued that prior estimates of the CEO

effect are conated with events outside the CEO’s

control and largely the result of random chance.

Fitza suggests that when this random chance is

accounted for, “the effect CEOs have on company

performance that can clearly be distinguished from

the effect of chance is between 3.9 and 5.0 percent”

(2014: 1847).

We believe that the empirical methodology

Fitza employed, however, substantially overstates

the “random chance” element of the CEO effect.

By extension, Fitza’s conclusions underestimate

the true impact CEOs have on organizational

performance. There are two reasons for this. First,

Fitza based his conclusions on an incomplete

assessment of the quality of the ANOVA models

used in his analysis. Specically, full consideration

of model t statistics, adjusted R-squared, in