Sydney Finkelstein, Donald C. Hambrick
5 min
This study investigates the "upper-echelons" theory, which posits that organizational outcomes—such as strategy and performance—are reflections of the characteristics of a firm's top management team. Specifically, the authors examine how the tenure of top management teams affects strategic persistence (the stability of a firm's strategy over time) and strategic/performance conformity (the degree to which a firm aligns with industry averages). Furthermore, the authors introduce the concept of "managerial discretion" as a moderator, arguing that the influence of top managers is constrained or enabled by their environment and organizational context.
Using a sample of 100 large organizations across three industries—computers (high discretion), chemicals (moderate discretion), and natural-gas distribution (low discretion)—the researchers analyzed data from 1978 to 1982. They employed generalized least squares (GLS) regressions to test the relationship between team tenure and organizational outcomes, while controlling for firm size, slack resources, and performance.
The study finds that as top management team tenure increases, firms exhibit greater strategic persistence and a stronger tendency to conform to industry norms in both strategy and performance. Short-tenured teams, by contrast, are more likely to pursue novel, deviant strategies that can lead to performance outcomes significantly different from industry averages.
Crucially, the strength of these relationships depends on the level of managerial discretion. In high-discretion environments (like the computer industry) or organizations with high slack and smaller size, the influence of management tenure on organizational outcomes is significantly stronger. In low-discretion environments, such as the highly regulated natural-gas industry, the link between management tenure and organizational outcomes is much weaker, suggesting that environmental constraints can override the influence of executive characteristics.
This research bridges the gap between two competing views in management theory: the "upper-echelons" perspective, which emphasizes executive influence, and the "environmental determinism" perspective, which suggests that external forces dictate firm outcomes. By demonstrating that managerial discretion acts as a critical moderator, the authors provide a more nuanced framework for understanding when and why top managers matter. The findings suggest that boards and stakeholders should consider the industry context when evaluating the impact of leadership changes, as the potential for executive-driven strategic change varies significantly across different business environments.
Alex: How did they operationalize discretion without it becoming circular? That seems like a genuine measurement challenge.
Sam: They used environmental indicators — product differentiability, regulatory freedom, demand instability — measured at the industry level. A natural gas utility scores low on all of these; a computer firm scores high. It's not a perfect solution, but it keeps the discretion construct separable from the outcome variables, which matters for the regression logic. [[RP_SECTION:empirical-methodology-and-results|Empirical methodology and results]]
Alex: And the empirical test itself?
Sam: They ran GLS regressions across a hundred organizations, with strategic persistence as the dependent variable and team tenure as the key predictor, interacted with the discretion measure. The tenure-persistence relationship was significantly stronger in the computer industry than in natural gas — which is the load-bearing result. They also controlled for return on equity, and the tenure effect held after that adjustment, which rules out the simplest performance-based confound. [[RP_SECTION:limitations-and-identification-challenge|Limitations and identification challenges]]
Alex: Where would a careful referee push back?
Sam: Two places. The first is reverse causality — the classic identification problem. Firms following stable strategies may simply choose not to turn over their management teams, which would inflate the observed tenure effect without any causal story running from tenure to persistence. The authors acknowledge this, but longitudinal design alone doesn't resolve it. You'd need something closer to an exogenous shock to tenure to really nail the direction.
Alex: And the second?
Sam: The proxy for the dominant coalition. They use board membership, which is objective and measurable, but it likely misses influential executives who operate outside formal board structures. And for some industry subgroups, the sample is small enough that the industry-specific estimates carry real uncertainty — the pattern is consistent, but the precision isn't always there to support strong inference. [[RP_SECTION:theoretical-contributions-and-implicatio|Theoretical contributions and implications]]
Alex: So what's the right way to read this? Strong framework, provisional estimates?
Sam: I'd put it that way, yes. The conceptual contribution — integrating managerial discretion into upper-echelons theory — resolves a genuine tension in the literature between the view that leaders are all-powerful and the view that they're captives of their environment. The answer is neither: managerial influence is a variable, and discretion is what sets its range. That's a durable insight. The specific effect sizes are less durable, partly because of the identification issues and partly because the organizational contexts have shifted considerably since 1990.
Alex: Which raises an obvious extension — the discretion available to a CEO in an AI-native firm versus a legacy utility is probably wider than anything in this sample.
Sam: Exactly. And that's where the framework still has traction. The implication for anyone studying executive effects today is that you can't treat tenure as a demographic variable in isolation. You have to account for the latitude of the role. Strip that out and you're likely misattributing variance — either overcrediting executives in constrained environments or undercrediting them in open ones. The moderating structure is the contribution, and it holds up even when the specific estimates don't.
Alex: A conditional theory that's more honest about when it applies. That's a reasonable place to land. Thanks for listening to ResearchPod.