Changes in persistence of performance over time

• DOI: http://doi.org/10.1002/smj.3185
• Published date: 01 October 2020
• Author: Victor Manuel Bennett
• Date: 01 October 2020

RESEARCH ARTICLE

Changes in persistence of performance

over time

Victor Manuel Bennett

Fuqua School of Business, Duke

University, Durham, North Carolina

Correspondence

Victor Manuel Bennett, Fuqua School of

Business, Duke University, Durham, NC.

Email: vmb10@duke.edu

Abstract

Research Summary: I revisit a research tradition laying

out trends in persistence of performance, to create shared

facts to inform theory development. I build on prior work

in Strategy by: (a) bringing together two distinct measures

of persistence of performance and explaining how they

complement each other and help distinguish theories, (b)

extending the time series from those prior studies and

applying modern statistical improvements to demonstrate

new patterns. Specifically, while I find ordinal persistence

increased monotonically since the 1980s, autocorrelation

of performance was not monotonic and is now

approaching lows from the mid-1990s. These patterns are

inconsistent with mechanisms raised in the debate about

declining competition. I suggest that these facts can

inspire Strategy theory, which would contribute to that

debate.

Managerial Summary: A central concern of managers

is how precarious is their position, or for new ventures,

whether they will be able to displace current leading

firms. Even within the same year, management writing

and the media might claim that it is both: a) a time of

faster-than-ever turnover, b) or that large firms have

become unassailable. This manuscript tracks two mea-

sures of performance persistence over time. I find that

This manuscript has benefited tremendously from the efforts of the Editor, four anonymous reviewers, Ashish Arora,

Sharique Hasan, Jim Bessen, and seminar participants at SDABocconi, the Strategy Research Forum, LMU Munich

Institute for Strategy, Technology, and Organization (ISTO), and the White House Council of Economic Advisers (CEA)

Research Series. It would not have been possible without the ideas and efforts of Claudine Gartenberg. John Elder

provided excellent copyediting, but all errors remain my own.

Received: 25 June 2016 Revised: 12 February 2020 Accepted: 14 April 2020 Published on: 28 June 2020

DOI: 10.1002/smj.3185

Strat Mgmt J. 2020;41:1745–1769. wileyonlinelibrary.com/journal/smj © 2020 Strategic Management Society 1745

persistence of rank has increased roughly steadily since

the 1980s. How associated are performance measures

year-to-year was actually at its peak in the mid 1980s,

potentially explaining the debate. The final patterns are

consistent with top firms separating from followers in

performance so much that increasing volatility in their

performance does not lead to changes in rank.

KEYWORDS

competition, persistence of performance, rank, strategy, trends

1|INTRODUCTION

A frequently reoccurring political issue that is currently prominent (Economist, 2016; Fleming

& Fox, 2018; Scaggs, 2018), is persistence of business' performance—the tendency for firms with

high performance last period to also have high performance this period. Scholars have long

been central in this argument, with some arguing that the performance of firms is becoming

ever more entrenched, making dominant firms more dominant and potentially leading to

increased inequality (Autor, Dorn, Katz, Patterson, & Reenen, 2020) and decreased innovative

investment (Gutiérrez & Philippon, 2017). Others, however, suggest that product lifecycles are

becoming more rapid (Bayus, 1998) and that it is becoming harder for firms to sustain an

advantage (D'Aveni, 1994; D'Aveni, Dagnino, & Smith, 2010; McGrath, 2013; Thomas &

D'Aveni, 2009). Both the Strategic Management Journal (D'Aveni et al., 2010) and Organization

Science (Ilinitch, D'Aveni, & Lewin, 1996) have had special issues dedicated to how firms must

respond to the asserted decline in persistence.

1

Arguments differ, in part, because the facts themselves are disputed. For example, much of

the work in economics asserting increased entrenchment does not actually measure entrench-

ment, but rather measures either concentration (e.g., Autor, Dorn, Katz, Patterson, &

Reenen, 2017; Gutiérrez & Philippon, 2017) or structurally estimated markups (e.g., De

Loecker, Eeckhout, & Unger, 2020). These measures have been criticized by those saying that

(a) concentration should not be measured at the national level and, when measured at the mar-

ket level, it is actually decreasing (e.g., Rossi-Hansberg, Sarte, & Trachter, 2020; Rinz, 2018;

Handwerker and Dey, 2019) (b) even correctly measured concentration does not represent com-

petition (Shapiro, 2018), and (c) estimated trends in markups are inconsistent with trends in

inflation (Syverson, 2019). Shedding light on what is actually happening in the American mar-

ketplace therefore calls for new measures.

Fortunately, Strategy scholars have a long history of studying the persistence of perfor-

mance. In fact, Rumelt, Schendel, and Teece (1991: 12) argue that this was one of the origins of

the field:

Some firms simply do better than others, and they do so consistently. Indeed, it is the fact of

these differences that was the origin of the strategy [field]. In standard neoclassical economics,

1

This assumption was not necessarily a dominant view, as work in the Strategic Management Journal has suggested that

persistence did not decline (e.g., Jacobsen, 1988; McNamara, Vaaler, & Devers, 2003).

1746 BENNETT