Prior research on CEO succession has omitted consideration of a critical institutional reality: some exiting CEOs do not fully depart the scene but instead remain as board chairs. We posit that predecessor retention restricts a successor's discretion, thus dampening his or her ability to make strategic changes or deliver performance that deviates from pre-succession levels. In short, a predecessor's continuing presence suppresses a new CEO's influence. Based on analysis of 181 successions in high technology firms, and with extensive controls (for circumstances associated with succession, the firm's need and capacity for change, and for endogeneity), we find substantial support for our hypotheses. In supplementary analyses, we find that retention has a more pronounced effect in preventing a new CEO from making big performance gains than in preventing big drops.

We introduce a new explanation for one of the most pronounced phenomena on the American business landscape in recent decades: a dramatic increase in attributions of CEO significance. Specifically, we test the possibility that America's CEOs became seen as increasingly significant because they were, in fact, increasingly significant. Employing variance partitioning methodologies on data spanning 60 years and more than 18,000 firm-years, we find that the proportion of variance in performance explained by individual CEOs, or "the CEO effect," increased substantially over the decades of study. We discuss the theoretical and practical implications of this finding.

We introduce multiple refinements to the standard method for assessing CEO effects on performance, variance partitioning methodology, more accurately contextualizing CEOs' contributions. Based on a large 20-year sample, our new 'CEO in Context' technique points to a much larger aggregate CEO effect than is obtained from typical approaches. As a validation test, we show that our technique yields estimates of CEO effects more in line with what would be expected from accepted theory about CEO influence on performance. We do this by examining the CEO effects in subsamples of low-, medium-, and high-discretion industries. Finally, we show that our technique generates substantially different-and we argue more logical-estimates of the effects of many individual CEOs than are obtained through customary analyses. a The variance explained by CEOs is calculated by multiplying the unexplained variance from Model 3 (79.8%) by the marginal R-squared from Model 4 (48.7%).

A recent study by Fitza argued that the prior estimates of the Chief Executive Officer (CEO) effect are conflated 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 findings, 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.

Research Summary We introduce an open‐source dataset documenting the reasons for CEO departure in S&P 1500 firms from 2000 through 2018. In our dataset, we code for various forms of voluntary and involuntary departure. We compare our dataset to three published datasets in the CEO succession literature to assess both the qualitative and quantitative differences among them and to explore how these differences impact empirical findings associated with the performance‐CEO dismissal relationship. The dataset includes eight different classifications for CEO turnover, a narrative description of each departure event, and links to sources used in constructing the narrative so that future researchers can validate or adapt the coding. The resulting data are available at (https://doi.org/10.5281/zenodo.4543893). Managerial Summary This article describes the development of an open‐source database of all CEO dismissals and departures in the S&P 1500 between 2000 and 2018. Prior research on CEO turnover either does not capture the cause of departure or has coded the event independently, leading to inconsistencies and a lack of transparency in coding schemes. This has made it difficult to generate knowledge on the causes and consequences of CEO dismissal. We describe how we developed the database, and we explore how our dataset compares to prior CEO dismissal research. The resulting data are available at (https://doi.org/10.5281/zenodo.4543893).