Industry Evolution

Contributed by Brian Silverman




This reading list explores industry evolution, with a particular focus on its relationship to innovation. Industry evolution refers to cumulative change in industry characteristics, notably the processes of firm entry, exit, and growth. Models that explore such processes generally exhibit three features: 1) an emphasis on industry change as a dynamic process, 2) an embrace of path-dependence, and 3) a search for empirical regularities and irregularities across industries. Industry evolution is of particular interest to students of firm strategy because anticipating and exploiting environmental change is one of the most fundamental tasks that managers face, and the literature on industry evolution provides insights into the interdependencies among industry change, a firm’s strategic choices, and changes in the basis of competitive advantage.

Early research on industry evolution focused on developing theory that could explain key empirical regularities noted in life spans of industries, notably related to entry rates, exit rates, firm growth and survival, market demand, and the nature of technological innovation. Recent research has increasingly emphasized heterogeneity among firms (that is, which firms survive and generate certain types of innovation) and heterogeneity within and among industries (that is, what industry characteristics, such as the existence of industry segments, change the nature of competition within an industry). Recent research has also used industry evolution models as theoretical and empirical baselines for exploring related issues such as diffusion of knowledge developed by failed firms, innovation-based competition between incumbents and entrants, and the effect of institutional change on industry development.

The seven core readings listed above provide a solid introduction to theoretical underpinnings and recent developments in the literature on industry evolution and firm strategy.1 This commentary traces the development of this literature. The optional readings in the annotated list following the commentary provide additional background and further development of core ideas for the interested reader. Articles referred to in the commentary that appear in the supplementary readings list are marked with an asterisk (*).

Much of the modern field of industry evolution stems from the concept of the industry life cycle, in which industries are conceived as progressing through several stages: emergence (with the introduction of the first relevant product or service), growth (during which the industry experiences a wave of entry and sharply increasing demand), maturity/shakeout (characterized by more stable demand and a wave of net exit), and decline (characterized by stagnant or declining demand and a stable set of firms). Although this model seems to have first been articulated to help stock market investors make investment decisions (Grodinger, 1953), it quickly became a staple of strategy scholarship and managerial prescription (e.g., Levitt, 1965). Much of the research in industry evolution has focused on the development of theoretical explanations for the stylized facts associated with the industry life cycle. One impressive feature of this effort has been the progression from models in which industry change is exogenously imposed to those that grapple with endogenously driven change.

Core reading #1, Gort & Klepper (1982), provides a clear and accessible review of five competing theories of industrial development and their implications for net entry and exit rates, output growth, and technological change in an industry over time. Using data on 46 industries, Gort & Klepper develop several stylized facts regarding entry, exit, growth, and innovation, and run a “horse race” among these theories to determine which can best explain the stylized facts. The article is a valuable demonstration of how to derive clear, testable implications from theory, how to creatively and painstakingly collect data that will support a clear empirical test, and how to match theory to real-world stylized facts. With 25 years of hindsight, however, one striking feature of the model is its reliance on exogenous events to change key parameters. For example, although the specified entry function relies heavily on the nature of innovation in an industry, the model is silent on what triggers a change from “major” to “minor” innovation. Hence, although there is some path-dependence in the model, not all key parameters evolve in a historically dependent fashion.

Several scholars tackled the challenge of developing a fully endogenous model of industry evolution. In core reading #2, Klepper (1996) develops a formal model that has been widely influential in strategy research. In this model, firms with stochastically different innovative ability enter an industry at different times. Those that develop appealing product or process innovations grow, subject to adjustment costs. Firms become more productive over time thanks to process innovations. As each firm makes decisions about R&D investment and output, three key events occur: 1) entry slows as productivity improvements among incumbents raise the bar for profitable entry; 2) exit increases as increased competition ultimately drives out less competitive firms – i.e., those that are smaller (and hence lack scale economies) or that are less capable at R&D; and 3) increasingly large incumbent firms increasingly shift R&D investment from product innovation to process innovation, as the returns to process innovation are increasing in firm output. The resulting model explains six stylized facts about industry evolution and offers a parsimonious story explaining how the interaction of profit-maximizing firms leads systematically to shakeouts, an emphasis on process innovation, the long-term survival of early entrants, and other features.

Other approaches to the development of such models of industry evolution include Jovanovic (1982*) who develops a model in which firms learn their efficiency only after entering an industry: those firms that are particularly efficient grow quickly, while those that are inefficient exit. This model can explain the simultaneous high likelihood of failure and of rapid growth for small firms vs. large firms, as well as explain the relationship between industry concentration and industry-wide rate of return. Alternatively, Jovanovic & MacDonald (1994*) explain shakeouts as a consequence of a major innovation that increases the minimum efficient scale of production; early adopters grow rapidly, forcing out laggards. Related natural-language models that emphasize the shift in type of innovation within an industry include Abernathy & Utterback (1978*) and Anderson & Tushman (1990*).

Although much of the early work on industry evolution was rooted in economics, sociological approaches to the topic flourished beginning in the late 1970s. Building on the work of Hannan & Freeman (1977; 1984), whose theory of structural inertia emphasized the obstacles to change encountered by organizations, organizational ecologists emphasized the influence of selection (that is, changes in the composition of firms in an industry) over adaptation (that is, changes in existing firms’ activities) in changing the nature of an industry. The primary distinction of the organizational ecology approach centered on the role of “legitimation,” the development of a taken-for-grantedness for a product or organizational form. The addition of a firm to a population or industry increases both competitive intensity and the industry’s legitimacy. When population “density” (i.e., number of firms) is low, the legitimation effect swamps the competitive effect. However, the second derivative of legitimation with respect to population density is negative while that of competition is positive; thus, above some population level, competitive effects swamp legitimation effects, leading to net negative entry. Thus, population ecology offers an alternate explanation for certain stylized facts of industry evolution. Core reading #3, Hannan et al. (1995), provides a clear articulation of density dependence and extends the original theory to consider whether it works at different geographic levels.

Alternate sociological approaches to industry evolution include resource partitioning (Carroll, 1985*; for refinements see Dobrev, Kim & Carroll 2002*) and niche-overlap (Baum & Singh 1994a), both of which focus on competition among organizations that rely on different sets of resources; as well as institutional ecology, which looks at the role of institutional pressures on founding and failure rates (Baum & Oliver 1996*).

Over the last decade, strategy scholars have devoted increasing attention to firm heterogeneity – which firms are likely to enter, survive, and thrive in an industry, which are at greatest risk of exit, and what contingent characteristics affect this? One stream of research has investigated the role of pre-entry experience as a predictor of post-entry success. Several studies have found that, compared to de novo firms, diversifying entrants have greater success rates in terms of survival (Klepper & Simons 2000*), even if their products are technically inferior to those of new entrants (Khessina & Carroll 2008*). Diversifying entrants also play a larger role in legitimizing the industry (McKendrick et al. 2003). Core reading #4, Thompson (2007), distinguishes between different types of pre-entry experience to identify which offers the most valuable benefits in a new industry. Of particular interest, Thompson finds that two key stylized facts – that older and larger firms have superior survival rates than younger and smaller firms – disappear after controlling for pre-entry experience. 2

Relatedly, core reading #5, Tripsas (1997), explores the competition between industry incumbents and entrants after a major innovation upsets an industry. Integrating the literature on competence-destroying innovation (Tushman & Anderson 1986) with that on appropriating the returns to innovation (Teece 1986), Tripsas proposes that the post-technological-shock balance of power between incumbent and entrant turns on the distribution of ownership over scarce complementary assets. If incumbents continue to own key assets that retain their importance post-innovation, then the incumbents will continue to dominate the industry. Conversely, if the assets are devalued, then the incumbents will be overwhelmed by the entrants. Tripsas tests her predictions in a study of the typesetting industry, which experienced three major innovations during the century that she studies. Not only does this article demonstrate a valuable theoretical integration across literatures, but it also demonstrates the use of qualitative and quantitative data to convincingly demonstrate results.

Also over the last decade, strategy scholars have devoted increasing attention to heterogeneity of the landscape within an industry. Whereas most of the early economic research assumed that all firms compete equally within an industry – that is, an industry is a homogenous-good market – a stream of literature has emerged that focuses on the importance of segments or niches within an industry. Core reading #6, de Figueiredo & Silverman (2007), develops a natural language model of competitive dynamics with both firm heterogeneity and industry segmentation. This model makes several predictions about the pattern of expansion by (dominant) incumbent firms across segments and about the way that competition plays out at the segment level. For studies exhibiting a more sociological approach – or, depending on your level of cynicism, reflecting less self-serving bias by the producer of this commentary – see Dobrev et al’s (2002*) study of niche evolution or Mitchell’s (1989) study of entry by incumbents into industry subfields. Related articles include King & Tucci’s (2002*) study of entry into new generations of a product, and Klepper & Thompson’s (2006*) model of industry dynamics in “subsectors” that appear and expire randomly.

The above literature frequently discusses exit rates from an industry, but what actually happens to a firm’s assets after it exits? An exiting firm leaves behind various resources that might live on in an industry – former employees, equipment, industry-specific knowledge, and the like. Economists have generally been silent on this issue. Organizational ecologists have studied whether prior exits influence current entry, presumably by releasing useful resources, with mixed results. Core reading #7, Hoetker & Agarwal (2007), extends this by focusing on the diffusion of a particular asset – technological knowledge – following the exit of firms from the disk drive industry. They find evidence that a living, functioning firm enhances the ability of other firms to access knowledge. 3 More important, this article demonstrates how to use industry evolution concepts to explore strategic questions in related fields such as diffusion of technological knowledge.


Bibliography




1 These readings do not do justice to the broader literature on evolutionary theories of economic and organizational phenomena. The supplemental list appended to this commentary includes a handful of readings that would likely be considered core readings for that literature, most notably Nelson & Winter (1982).
2 The exploration of pre-entry experience links the industry evolution literature to that of diversification (see Villalonga in this Reader for a commentary on that subject).
3 This is also relevant to the literature that explores how institutions influence the flow of scientific and technological knowledge (e.g., Furman & Stern 2006).