To Be or Not to Be: The Entrepreneur in Neo-Schumpeterian Growth Theory

Abstract

Based on a review of 700+ peer-reviewed articles since 1990, identified using text mining methodology and supervised machine learning, we analyze how neo-Schumpeterian growth theorists relate to the entrepreneur-centered view of Schumpeter Mark I and the entrepreneurless framework of Schumpeter Mark II. The literature leans heavily toward Schumpeter Mark II; innovation returns are modeled as following an ex ante known probability distribution. By assuming that innovation outcomes are (probabilistically) deterministic, the entrepreneur becomes redundant. Abstracting from genuine uncertainty, implies that central issues regarding the economic function of the entrepreneur are overlooked such as the roles of proprietary resources, skills, and profits.

Introduction

What is the problem we wish to solve when we try to construct a rational economic order? [ . . . ] If we possess all the relevant information, [ . . . ] and if we command complete knowledge of available means, the problem which remains is purely one of logic. This, however, is emphatically not the economic problem which society faces. And the economic calculus which we have developed to solve this logical problem [ . . . ] does not yet provide an answer to it. (Hayek, 1945)

The advent of sustained economic growth that began some two centuries ago has been crucial for the dramatic increase in human welfare compared to earlier periods in the history of our species. Understanding the causes of economic growth is therefore of enormous value. The first generation of modern mainstream growth models showed that the accumulation of factors of production could explain only a small part of growth (Solow, 1957). This triggered the development of endogenous growth models in the 1980s, which added human capital accumulation and knowledge capital to the models. By assuming that some knowledge was nonrival and nonexcludable, the explanatory power was greatly increased. ¹ Nevertheless, the models lacked an agent that combined and applied the new knowledge with other factors of production to generate growth. In short, the models were entrepreneurless. ²

The first major step to introduce the entrepreneur into mainstream models of aggregate economic growth was neo-Schumpeterian growth theory, which first appeared in the early 1990s. ³ This line of economic inquiry is called “Schumpeterian” because it incorporates the process that Schumpeter (1942) termed “creative destruction,” that is, the process by which new innovations challenge and—if successful—replace existing economic structures, into a new breed of endogenous growth models. ⁴

However, Schumpeter took two divergent views on the entrepreneur during his career. In earlier conceptualizations, Schumpeter (1934) considered the entrepreneur to be the primus motor of innovation-based growth by carrying out discontinuous innovations. ⁵ Later in life, Schumpeter (1942) predicted that entrepreneurs would become redundant as innovations became routinized and carried out by large corporations. Following Phillips (1971), these two opposing views of innovative activities are customarily referred to as Schumpeter Mark I and Mark II, respectively (e.g., Malerba & Orsenigo, 1996; Nelson & Winter, 1982).

Hence, creative destruction, according to Schumpeter, can be modeled either with the entrepreneur at the center of analysis or with the individual entrepreneur having a marginal role or being altogether absent. These two conceptualizations have vastly different implications for theory and policy. However, although comparative discussions on entrepreneurial definitions and their implications for research and policy have a long history in related fields, such discussions have been largely absent among neo-Schumpeterian growth theorists.

The aim of this article is to analyze how entrepreneurship is represented and conceptualized in neo-Schumpeterian growth theory. We do this by analyzing the content of 714 peer-reviewed articles on neo-Schumpeterian growth published over the period 1990 to 2020. The articles are identified through text mining of relevant articles and supervised machine learning. The analysis distinguishes between highly influential articles in the field (“core articles”), reviews, and other articles. In addition, two textbooks by seminal authors are analyzed. ⁶ First, the use of key terminology and references to formative literature, for example, Schumpeter (1934), are examined quantitatively. Then, the use of the term “entrepreneur” is studied and categorized relative to Schumpeter Mark I and II.

Previous reviews of the literature on neo-Schumpeterian growth (Block et al., 2017; Bogliacino, 2014; Dinopoulos & Şener, 2007; Shabnam, 2014) have taken entrepreneurial concepts as given and have not addressed the fundamentals of the underlying theory. Other studies have challenged the theoretical foundation of neo-Schumpeterian growth theory (Acs & Sanders, 2013; Bianchi & Henrekson, 2005; Johansson & Malm, 2017; Nelson, 1997, 1998). However, these studies have been limited in their empirical scope; they have not provided exhaustive evidence on what the literature does include.

Research in the Schumpeterian tradition since its inception has been prolific; there are several strands of literature inspired by Schumpeter’s work such as the literature on national innovation systems (e.g., Edquist, 2013; Lundvall, 2012) and entrepreneurial ecosystems (e.g., O’Connor et al., 2018; Wurth et al., 2022). Meanwhile, although highly regarded and enriching on their own merits, each of these literatures lies outside the core of mainstream economics. Within the mainstream, Schumpeterian arguments are predominantly framed in the neo-Schumpeterian tradition highlighted in our paper. ⁷ In turn, this means that the neo-Schumpeterian tradition constitutes one of the dominant channels for disseminating ideas on entrepreneurship from economics more generally to policy makers, graduate students, and the general public (Aghion et al., 2015b). In our view, this provides a strong motivation for conducting an in-depth analysis of this literature. Moreover, in contrast to the rival Schumpeterian literatures, all mainstream endogenous growth models are based on the existence of an economic equilibrium and optimization. Hence, based on the political and scholarly impact of the literature as well as for the sake of logical clarity, the current article is strictly limited to the domain of neo-Schumpeterian growth. ⁸

We contribute to the literature in at least two ways. First, we provide the first large-scale empirical study connecting the description of the entrepreneurial function in neo-Schumpeterian growth theory to its Schumpeterian antecedents. Second, we position the literature relative to Schumpeter’s work, thus highlighting the limitations of current discussions and pointing to potential areas of future development.

We find that the literature leans more toward Schumpeter Mark II than toward Schumpeter Mark I. For instance, the literature refers to Schumpeter (1942) more often than Schumpeter (1934) and it does not relate to the key Schumpeter Mark I concept of “new combinations.” A critical aspect is that innovative ventures are modeled as processes whose return is determined by an ex ante known probability distribution, which greatly curtails the literature’s potential for providing causal guidance regarding the economic processes that precede innovation and the processes that influence its outcome. Moreover, the literature is shown to heavily emphasize routine over disruptive innovation, which is inferred to be a direct consequence of its conceptualization of the innovation process. Hence, the neo-Schumpeterian entrepreneur may be described as a routine decision-maker who pursues business opportunities based on exogenous and ex ante given parameters. This modeling of innovative activities renders the entrepreneur—as presented in Schumpeter Mark I—redundant in discussions of neo-Schumpeterian growth. The literature thus abstracts from the role of nonroutinized entrepreneurial decision-making in ventures that contain elements of genuine uncertainty, that is, non-calculable risk. Genuine uncertainty is central to understanding the economic role of the entrepreneur, particularly his or her key role in the introduction of disruptive innovations. Thereby, the literature is also lacking in its ability to explain about economic conditions that promote radical technological shifts and, as a corollary, in identifying policy measures that foster radical innovation. ⁹

Our findings highlight what is arguably a fundamental limitation of the current discourse. Given that researchers subscribe to the notion that innovations are, at least partly, associated with genuine uncertainty, this implies that extant neo-Schumpeterian growth models run the risk of providing misleading guidance to policy makers who would like to stimulate economic growth. In particular, in cases where it emanates from the introduction of disruptive innovations.

A potential objection to our examination is that neo-Schumpeterian growth models seek to explain and predict the historical macroevolution of the economy, and at the aggregate level, it may be fair to abstract from the genuine uncertainty of innovative activity at the micro level. ¹⁰ However, we argue that a causal understanding is necessary to inform us of how to stimulate future economic growth. This is likely to be particularly important for economies at the technological frontier, where the relationship between R&D output and economic growth is far from unequivocal and where there are only minor opportunities for improvements to basic institutional quality. The development of modern growth theory may itself serve as an illustrative example: Departing from modeling growth as exogenous, knowledge and creative destruction were endogenized to strengthen the models’ explanatory power and to identify additional economic policy tools. The current analysis can be regarded as a further step in that process.

To extend and enrich the discussion beyond routinized innovation, our study suggests that the literature could draw on insights from extant discussions of genuine uncertainty and its key role in entrepreneurial activity. We believe that the discussion could benefit from incorporating ideas from the entrepreneurship literature that has so far developed parallel to neo-Schumpeterian growth theory, notably Knight (1921) and the subsequent literature emphasizing the role of uncertainty-bearing and judgmental decision-making (e.g., Foss & Klein, 2012). A fundamental idea in this tradition is the experimental character of the economy caused by uncertainty. The need for policy to provide favorable incentives for novelty and adaptability is therefore stressed to a greater extent than in the mainstream literature. Among other things, this includes a focus on diversity in skills, different forms of finance, and free access and flow of knowledge (e.g., Elert et al., 2019; Henrekson & Johansson, 2009; Sanders et al., 2020). By incorporating aspects like these, we may gain a deeper understanding of entrepreneurship, innovation, and, ultimately, economic growth as an endogenous process.

The rest of the paper is organized as follows. The second section discusses Schumpeter’s two concepts of entrepreneurship. The third section presents the study’s data and method, and the fourth section presents the results. The fifth section offers a concluding discussion. Appendix A describes the process and search terminology to identify the neo-Schumpeterian growth literature, and Appendix B reports the core article that the identification process is based on. A complete list of identified articles is presented in the Online Appendix.

The Schumpeterian Entrepreneur

Schumpeter first laid out his theory of the economic function of the entrepreneur in German in 1911, but it took until 1934 before the work was available in English. In Schumpeter (1934), he sets out to identify the causal mechanisms that connect innovative activity to economic growth. He posits that economic growth cannot be adequately explained by increases in factors of production; in his view, long-run growth also involves change in the sense that the factors of production are repurposed in new and more valuable ways, that is, what he refers to as economic development. Such repurposing of existing resources—the creation of “new combinations”—is carried out by the entrepreneur, who, consequently, is conceptualized as the primus motor of innovation-based growth. Since economic change is seen as an endogenous process driven by the creativity and actions of individual actors, the ability of the individual entrepreneur becomes central. ¹¹ Since entrepreneurial skill is a scarce resource, it was identified as a bottleneck in innovative processes (Acs & Sanders, 2013).

Schumpeter (1934) reasoned that new ideas are only economically relevant if they are put to economic use, and the entrepreneur is seen as the primary link between new ideas and their market introduction in the form of valuable commodities. The entrepreneur identifies the potential economic uses of new ideas and realizes their economic value through commercialization, and new firm entry provides an important channel for entrants to introduce radically new ideas and to challenge existing economic structures.

New combinations translate into economic development through a three-step process. The first step involves the conception of a novel idea, a new combination, which Schumpeter referred to as an invention. Once a novel and potentially profitable invention has been identified, the second step consists of identifying its potential economic uses and realizing its economic value by introducing it to the market, which is referred to as carrying out an innovation. When an economic use of a new combination has been identified, the third step of the process consists of spreading the innovation in the economy, which Schumpeter referred to as dissemination. If successful, the new innovations and their dissemination give rise to creative destruction manifested in structural change that alters the composition of the economy.

In describing this process, Schumpeter (1934) was careful to distinguish the role of inventors—actors who conceive new inventions—from those who identify and realize their economic value—entrepreneurs. This distinction is essential because it illustrates the assertion that economic change requires, in addition to novel ideas, the ability to commercialize them. This view is commonly referred to as Schumpeter Mark I.

Schumpeter defined innovation more broadly than what is typically referred to by the term in everyday language as well as in economic analysis, where innovation is most commonly thought of as emanating from R&D. However, Schumpeter maintained that this definition was too narrow and argued that innovations did not have to be of scientific origin. Rather, he envisioned innovations as taking five principal forms: the introduction of new products, the introduction of new methods of production, the opening of new markets, the conquest of new sources of supply, and new methods of organizing a firm or industry.

Later in life (Schumpeter, 1942), he argued that the innovative activity of individual entrepreneurs would be gradually phased out and replaced by routinized R&D processes in large corporations—a view customarily referred to as Schumpeter Mark II. ¹² A notable aspect of the thesis is that its primary intent was not to account for the process of creative destruction but rather to provide detail on the workings of socialism. Schumpeter predicted that increased routinization of innovation would lead to the disappearance of the entrepreneurial class, which, in turn, would pave the way for a structural shift toward socialism in the West.

Innovative activity may thus be modeled either as having the individual entrepreneur at the center of analysis or as a process in which the entrepreneur is marginalized or even absent. The choice of conceptualization has far-reaching implications for how one understands the workings of the economic system and the impact of economic policy. Schumpeter Mark II, with its emphasis on large corporations and central planning, lends support to the idea of interventionism and active industrial policy to stimulate economic growth. ¹³ In contrast, Schumpeter Mark I’s focus on individual entrepreneurs and nonroutinized innovation speaks in favor of a decentralized market economy. ¹⁴

Although Schumpeter’s work has influenced subsequent economic thought, a shortcoming is that it largely abstracts from the roles of risk and uncertainty in economic growth. Therefore, researchers have recently begun to show increasing interest in the work of Knight (1921), who likewise argued that entrepreneurial profit is a product of innovative entrepreneurship. He thereby extended our understanding of profit and, by extension, our understanding of the nature and economic role of entrepreneurship. ¹⁵ Importantly, Knight made a distinction between risk and uncertainty: risk is probabilistically quantifiable, while uncertainty is not. Thus, Knight refers to uncertainty as events about which we know so little that we are unable to assign any probabilities. This is customarily termed Knightian uncertainty or genuine uncertainty. ¹⁶

Building on the concepts of risk and uncertainty, Knight stipulated that entrepreneurial actions are inherently uncertain because they involve the creation of new combinations. Therefore, the outcomes of these actions cannot be known ex ante. Based on this premise, he argued that entrepreneurial profit cannot persist in competitive markets unless the expected value of innovative activity is also, at least in part, subjective because it would otherwise be absorbed through price adjustments of inputs in the innovation process. Consequently, the Knightian entrepreneurs differ in skill, which is expressed in terms of differences in their ability to make subjective assessments of the viability of innovative ventures, which Knight referred to as judgment.

Given the conception of risk as the outcome of calculable events, Knight was also of the opinion that risk should be seen as an ordinary cost, not as a residual of the returns on innovative activity. Therefore, it should not be understood as part of entrepreneurial profit; entrepreneurial profit should only be seen as the residual returns of innovative activity given risk, which he denoted “pure profit.” Hence, in the Knightian tradition, entrepreneurial profit refers to bearing the uncertainty associated with the introduction of new ideas and where he saw the pursuit of such profit opportunities as one of the key mechanisms in explaining long-run economic growth.

In contrast to Schumpeter Mark I, who asserted that employees could also fulfill the entrepreneurial function, Knight argued that entrepreneurship was inextricably linked to ownership. He based his view on three arguments. First, given that owners hold the ultimate decision-making rights, Knight inferred that owners ultimately decide whether to pursue innovation activities, including any decision to delegate this task. Second, owners are the residual claimants of the return on innovative activity; as their resources are invested, they are the ultimate bearers of uncertainty. ¹⁷ Third, given that entrepreneurial activity is inherently uncertain, the value of entrepreneurship is also uncertain; hence, the role of ownership becomes central to understanding entrepreneurial incentives. By virtue of these three arguments, Knight suggested that unlike other factor inputs, remuneration for entrepreneurial activities cannot be determined ex ante, not even in a probabilistic sense, due to the inherent uncertainty associated with entrepreneurial activity. Therefore, to foster innovation and establish a “price” on entrepreneurship services in the face of uncertainty, the entrepreneur must hold a residual claim on profits, that is, be made an owner. Hence, in the Knightian conceptualization, entrepreneurial profit and ownership serve the role of both providing incentives for entrepreneurship and a contractual solution for the pricing of the entrepreneurial function.

Another framework complementary to Schumpeter’s view is provided by Kirzner (1973). In contrast to Schumpeter, Kirzner envisioned the role of the entrepreneur as the actor who restores equilibrium by identifying existing arbitrage opportunities. ¹⁸ Moreover, he described entrepreneurship as a process of alertness and discovery, where entrepreneurs pursue objectively known arbitrage opportunities under competition.

Related to the Knightian and Kirznerian discussions on the nature of entrepreneurial activity, a strand of contemporary research studies the epistemological underpinnings of entrepreneurship by distinguishing between discovered and created business opportunities (e.g., Alvarez & Barney, 2010; Leyden & Link, 2015; Venkataraman, 2003). Discovered opportunities are exogenously existing opportunities whose intrinsic value can be objectively assessed by actors ex ante. In contrast, created opportunities are endogenously created by entrepreneurs based on their subjective valuations and cognitive abilities, and the market value of these opportunities is continuously realized by the entrepreneurs through a process of trial-and-error whereby their intrinsic value only becomes manifest ex post. In practice, it is probably not a question of either or; arguably, innovations typically include elements both of discovery and creation.

The distinction between discovered and created opportunities provides a framework for understanding both the nature of business opportunities and the entrepreneurial skills needed to pursue them. By applying the concepts of discovered and created business opportunities, it is possible to gain insight regarding the position of the neo-Schumpeterian entrepreneur relative to the frameworks of Schumpeter Mark I and II. Specifically, when prospective innovative ventures are assumed to be completely or mostly based on discovered opportunities, that is, when ventures are generally modeled as taking calculable risks, theory inadvertently assigns a central role to routinized investments and calculated risk preferences in firms for determining innovation and economic growth, which is in line with Schumpeter Mark II. In contrast, when innovative ventures are assumed to be completely or mostly based on created opportunities, that is, when opportunities are generally modeled as genuinely uncertain, theory assigns a central role to the nonroutinized decision-making of individual entrepreneurs, which is in line with Schumpeter Mark I. ¹⁹

Method and Data

Identifying the Population

We follow Aghion and Howitt (2009) and Acemoglu (2009) and date the conception of neo-Schumpeterian growth theory to 1990 based on the publication of Segerstrom et al. (1990) and Aghion and Howitt (1990). As a result, our investigation is confined to peer-reviewed articles published between 1990 and 2020.

To identify the field, a number of influential—or core—articles were reviewed to capture relevant terminology. The selection was based on the reviews by Acemoglu (2009), Aghion and Howitt (2009), Aghion et al. (2015a), and Akcigit and Nicholas (2019). ²⁰ This yielded an initial dataset of 44 publications (listed in Appendix B). Next, the content of these articles was analyzed to capture pervasive terminology across the articles by using text mining tools; see Appendix A. ²¹ As shown in Figure 1, the most common terms and phrases across the identified core articles are, for example, competition, productivity, and technological change. A striking feature of Figure 1 is the absence of the terms “entrepreneur” and/or “entrepreneurship.”

Figure 1.

Co-occurrence of the most common terminology across core articles on neo-Schumpeterian growth, 1990 to 2020.

Once the core terminology across articles was identified, combinations of key terms and auxiliary terminology were selected based on within-article co-occurrences. The resulting search strings were then inserted into Google Scholar, Scopus, and the Web of Science. The initial search process yielded a total of 40,388, unique results. ²² By means of a stepwise iterative procedure detailed in Appendix A, we arrive at a final population of 714 peer-reviewed articles featuring neo-Schumpeterian growth models. A complete list of identified articles is presented in the Online Appendix.

Text Analysis

Once the literature was identified, all peer-reviewed articles were manually surveyed to review their conceptualizations of the entrepreneur. Moreover, to characterize the literature, all articles were subjected to a word search for terms related to the work of Schumpeter as well as the complementary work of Knight (1921) and Kirzner (1973). To capture terminology related to all of the above works, articles were searched for the occurrence of the terms “entrepreneur” and “innovation.” Next, to capture terminology related to Schumpeterian discussions, articles were also searched for the terms “creative destruction,”“new combinations,”“invention,”“inventor,” and “innovator.” Moreover, to capture discussions by Knight (1921) and Kirzner (1973), articles were searched for the terms “alertness,”“genuine uncertainty,” and “judgment.” ²³ Finally, articles were searched for direct references to Schumpeter (1934, 1942), Knight (1921), and Kirzner (1973). To manage inconsistencies in formulations across texts, all search strings were applied using n-gram approximate or “fuzzy” string matching (e.g., Pfeifer et al., 1996).

Qualitative Analysis

To capture the conceptual nature of the literature and how it is positioned relative to Schumpeter Mark I and Mark II, a qualitative review was undertaken. This process departed from two key differences in the Schumpeterian characterizations of the value-creating process, namely the role of the entrepreneur in capturing the value of new ideas—active in Mark I, versus passive or superfluous in Mark II—and the type of innovative activity that is emphasized—nonroutinized in Mark I versus routinized in Mark II. Accordingly, the scope and focus of the review were limited to these areas of inquiry. ²⁴ This required, in turn, operationalization of the involved constructs and their domains. To capture discussions on the nature of innovation relative to the Schumpeterian debate, two factors were considered: (a) the conceptualization of the investment decisions that precede innovation and (b) the representation of new technologies relative to existing ones, that is, the composition of creative destruction. Similarly, to capture relevant discussions on the economic role of the entrepreneur, the literature was surveyed for statements and assumptions that explicitly connect the actions of entrepreneurs to innovative outcomes. ²⁵

The qualitative analysis was thereafter conducted in three steps. First, all three topics were surveyed using open coding (e.g., Blair, 2015). As such, this initial coding procedure sought to identify distinct subsets of theoretical and methodological traditions against which the literature’s orientation could be understood. Next, to more closely relate these findings to the Schumpeterian discussion, the respective outcomes were categorized in terms of: (a) Whether or not innovative outcomes were represented as having an ex ante objective value, that is, as being routinized, versus nonroutinized; (b) whether potential innovation outcomes were characterized as being incremental or non-incremental, that is, “step-by-step,” versus radical innovation; ²⁶ (c) whether the entrepreneur was conceptualized as playing an active role throughout the value-creating process, or whether parts of this role could be performed by other agents, that is, the entrepreneur as the primus motor, versus the entrepreneur as a partially or wholly substitutable actor.

Results

In this section, the use of entrepreneurship constructs in neo-Schumpeterian growth research is analyzed and positioned relative to Schumpeter Mark I and Mark II. First, article contents are surveyed for terminology use and literature references related to Schumpeter (1934, 1942) and the complementary frameworks of Knight (1921) and Kirzner (1973). Then, theoretical conceptualizations of the entrepreneur and his/her roles in innovative activity are qualitatively reviewed across articles.

Descriptive Results

Table 1 presents the use of Schumpeterian terminology related to Schumpeter (1934, 1942) as well as the complementary frameworks of Knight (1921) and Kirzner (1973). In addition, the table reports the number and share of articles that include direct citations to Schumpeter (1934, 1942), Knight (1921), and Kirzner (1973). In an effort to capture seminal discussions in the literature, the results are presented across the categories “core articles,”“review articles and textbooks,” and “other articles.”

Table 1.

The Number and Share (%) of Peer-Reviewed Articles and Textbooks That Include Direct Citations and Terminology Related to Schumpeter Mark I and II, Knight (1921), and Kirzner (1973), 1990 to 2020.

	(1)Core articles		(2) Review articles and textbooks		(3) Other articles		(4) Total
	Number	Share	Number	Share	Number	Share	Number	Share
Key terminology
Creative destruction	26	59	4	100	291	44	321	45
Entrepreneur	15	34	3	75	205	31	223	31
Innovation	44	100	4	100	626	94	674	94
Innovator	28	64	3	75	324	49	355	50
Invention	17	39	3	75	226	34	246	34
Inventor	17	39	3	75	169	25	189	26
New combination	1	2	0	0	6	1	7	1
Uncertainty	17	39	2	50	173	26	192	27
Risk	22	50	3	75	350	53	375	53
Genuine uncertainty	0	0	0	0	8	1	8	1
Judgment	0	0	0	0	3	0.5	3	0.4
Alertness	0	0	0	0	0	0	0	0
Literature references
Schumpeter (1934)	2	4	2	50	21	3	23	3
Schumpeter (1942)	7	16	3	75	87	13	96	13
Knight (1921)	0	0	0	0	2	0.3	2	0.3
Kirzner (1973)	0	0	0	0	1	0.2	1	0.1
Total number of included works	44	6	4	1	666	93	714	100

Note. The category “genuine uncertainty” also encompasses the terms “Knightian uncertainty,”“true uncertainty,” and “radical uncertainty.” See Appendix A for “Core articles.”“Review articles” are Aghion et al. (2015a) and Akcigit and Nicholas (2019). “Textbooks” are Acemoglu (2009) and Aghion and Howitt (2009). “Other articles” are listed in the Online Appendix. The terms “entrepreneur” and “innovator” are weakly complementary: approximately 20% of articles use both terms.

The table reports that 31% of all articles mention the term “entrepreneur,” whereas almost all include the term “innovation.” ²⁷ Rather than using the term “entrepreneur,” the literature is found to use the term “innovator,” which appears in half of all articles. ²⁸ This usage is likely to result from the terminology used in early papers, such as Aghion and Howitt (1992), to denote actors that pursue innovative activity. ²⁹ Notably, the term “entrepreneur” does not appear in either of the seminal articles of Segerstrom et al. (1990) or Aghion and Howitt (1992). Grossman and Helpman (1991) include the term but use it as a synonym to innovation-based firms. It was not until later that the term became pervasive in the literature. Other early contributions are represented around the genesis of the literature, such as Boyer (1991) or Cheng and Dinopoulos (1992). ³⁰ However, it would take until the mid-2000s for the first emergence of this term in an article that spurred a significant number of subsequent studies (i.e., Aghion et al., 2005).

The concepts of “invention” and “inventor,” which are central to the Schumpeterian discussion, are found in approximately one-third and one-fourth, respectively, of all articles. In this case too, it could be noted that the early articles do not contain these terms, while only one review article does. Moreover, the few references made to “new combinations” use the term only to position the presented discussions relative to Schumpeterian terminology, whereas none apply the concept to the analysis (Albaladejo & Martínez-García, 2015; Murakami, 2017; Olsson, 2000, 2005; Stein, 1997). ³¹

Next, by studying references to Schumpeter (1934) and (1942) (Schumpeter Mark I and Mark II), the results in Table 1 again suggest that the literature primarily relies on the work of Schumpeter Mark II. The two works are cited in 3 and 13% of articles, respectively. ³² Moreover, most core articles make no reference to Schumpeter (1934), including the early articles by Segerstrom et al. (1990) and Grossman and Helpman (1991). This strengthens the conjecture that neo-Schumpeterian growth theory is primarily oriented toward Schumpeter Mark II. ³³

There are three observations that stand out in Table 1. First, given the large difference in the share of articles that include the term “innovation” compared to the terms “entrepreneur” and “innovator,” the focus of neo-Schumpeterian analyses is primarily innovation per se and not the actor(s) who conduct(s) it. This implies reliance on Schumpeter Mark II rather than Schumpeter Mark I. Since the latter views the innovator-entrepreneur as the persona causa of innovation-based growth, the exclusion of these terms indicates that Schumpeter Mark I is not applied.

Second, turning to the frameworks of Knight (1921) and Kirzner (1973), the literature frequently refers to the concepts of “risk” and “uncertainty.” However, a qualitative analysis suggests that these terms are used interchangeably rather than denoting two separate constructs. Similarly, only approximately 1% of all articles include discussions using the key Knightian concepts “genuine uncertainty,”“Knightian uncertainty,”“radical uncertainty,”“true uncertainty,” and “judgment,” and no articles include the Kirznerian concept of “alertness.” Finally, Knight (1921) is only cited in two articles. ³⁴ Kirzner (1973) is cited in one single article, Sanders and Weitzel (2012), who also apply Kirzner’s framework in their modeling. Given the small number of occurrences, these observations strongly suggest that the overall neo-Schumpeterian literature to date has not incorporated insights from Knight (1921) or Kirzner (1973). ³⁵

Conceptual Analysis

By examining the prevalence of key terms and references related to the formative literature—Schumpeter (1934), Schumpeter (1942), Knight (1921), and Kirzner (1973)—on entrepreneurship, the analysis in section “Descriptive results” offers a preliminary understanding of the orientation of neo-Schumpeterian analyses relative to Schumpeter Mark I and Mark II. We will now proceed to a qualitative assessment of the literature by reviewing the boundaries of entrepreneurial conceptualizations across articles.

In reviewing the literature’s orientation relative to Schumpeter Mark I and Mark II, we depart from their different characterizations of the value-creating process through the representation of innovative activity and the economic function of the entrepreneur. A prominent finding made during the initial open coding procedure is the high degree of homogeneity in the field’s conceptual foundations. ³⁶ First, the entrepreneurial function in neo-Schumpeterian growth theory is found to be exclusively modeled as an agent that pursues R&D investments in search of ex ante calculable monopoly rents (e.g., Aghion et al., 2015b). The entrepreneur is thus conceptualized as a decision-making agent in an intermediate sector firm that is responsible for allocating firm resources between two activities: production and R&D. As such, this conceptualization is silent about the role of the entrepreneur within the firm, that is, whether the role can be fulfilled by a manager or whether it refers to the owner(s) of a firm. A second characteristic of the literature is the conceptualization of innovation outcomes. Throughout the literature, it is assumed that returns on innovation investments follow an ex ante and objectively known probability distribution. The expected costs and returns of innovations are thus objectively calculable, and the value and economic uses of innovations are known once a new product or technology has been developed. Hence, the innovation concept is reminiscent of the concept of discovered opportunities. ³⁷

This assumption may be motivated by a quest for theoretical parsimony, which admittedly is both a common and necessary practice in economic modeling. ³⁸ However, this particular simplification comes at a significant cost to causal interpretability. Specifically, by assuming that entrepreneurial decision-making follows an implicit distribution, the literature is effectively applying a form of backward induction, where behavior is assumed to follow a deterministic pattern that is observable ex post. Although this may be sufficient to explain historical growth rates, it hampers our understanding of the growth process. This makes it less likely to be helpful in advising us on how to induce innovation and growth through a proper design of policy instruments. A richer conceptualization of the innovation process would benefit the debate by highlighting additional relevant points of inquiry, such as the macroeconomic role of ownership compositions (Andersson et al., 2018).

By depicting the entrepreneur as an actor whose economic function is to invest in calculable outcomes, the role of the neo-Schumpeterian entrepreneur is relegated to the role of a routine decision-maker in pursuit of discoverable business opportunities. ³⁹ This implies that neo-Schumpeterian economic modeling closely resembles the entrepreneurless growth process of Schumpeter Mark II rather than the entrepreneur-centered view of Schumpeter Mark I. In fact, despite being referred to as “Schumpeterian,” the current literature is arguably more reminiscent of models on variety-expansion (e.g., Romer, 1990), which do not include an entrepreneur. Owing to this theoretical affinity, the two literatures are often cited in tandem, and there have been theoretical efforts to bridge the gap between them (e.g., Bondarev & Greiner, 2019; Futagami & Ohkusa, 2003; Madsen, 2008). ⁴⁰ As such, the current framework can be understood as primarily following a tradition of describing incremental quality improvements of established products or services where the potential payoffs on investments are partly or wholly calculable, that is, what neo-Schumpeterians commonly refer to as “quality ladders.” ⁴¹

Moreover, when the profitability of R&D investments is modeled as being probabilistically deterministic, the innovation process becomes of subordinate interest, which explains why the literature does not elaborate on the different stages of the innovation process: invention, innovation, and dissemination, that is, the processes that connect the conception of a new idea to its subsequent market introduction and dissemination in the economy. ⁴² It also explains why the literature delves less than related research into the importance of “agent heterogeneity” for successfully carrying out the different phases of the innovation process, that is, the need for economic actors such as entrepreneurs and firms with different skills. ⁴³ Likewise, discussions of different types of innovations and their relative importance are largely absent. This constitutes yet another departure from Schumpeter.

Given the above discussion, it is inferred that the neo-Schumpeterian framework in its current state is likely to be ineffective in providing causal insights regarding the process of innovation and entrepreneurship, and in particular in cases where innovative outcomes are nonroutinized and non-calculable, that is, in cases where they contain elements of Knightian uncertainty. In turn, Knightian uncertainty is likely to be particularly central to forming an understanding of the antecedents of disruptive innovations and entrepreneurship (Assink, 2006; Naar et al., 2019). To the extent that researchers subscribe to the notion that both incremental and disruptive innovations are important for providing causal explanations to modern economic development, this implies a need for a different conceptualization of innovation-driven growth. Growth results both from radical and incremental innovation and their relative contribution varies across countries, sectors, and years; a model that focuses solely on either incremental or radical innovation will only capture part of the story.

A causally informative model of endogenous growth under genuine uncertainty must incorporate the fact that many—perhaps most—innovations are undertaken without full information on their potential value, meaning that they lack strictly objective benefits against which their costs can be weighed. Instead, innovation can be expected to be wholly or partly pursued based on the subjective valuations and judgment-based decisions of individual entrepreneurs (e.g., Bylund & Packard, 2021). Hence, given elements of genuine uncertainty, entrepreneurs cannot solely rely on objective knowledge regarding the final economic uses of ideas to determine their expected economic value (cf. Boettke, 1997). Instead, they must retain an active role in identifying the economic uses of innovations and appropriating their economic value, as the entrepreneur of Schumpeter Mark I. This implies that the focus of analysis is directed toward the process of invention, innovation, and dissemination. In contrast, neo-Schumpeterian growth models—like their variety-expanding kins—make no distinction between the invention and innovation stages and then assume instant diffusion.

At the same time, introducing incalculability and subjectivity into the economic models does not imply that innovation outcomes are driven solely by chance and subjectivity. On the contrary, several determinants of innovation success can likely be incorporated to increase both the causal interpretability and predictive power of existing frameworks. Notably, Knight (1921) stresses the central role of the knowledge, experience, and innate abilities of entrepreneurs in the selection and outcome of disruptive innovations, that is, what he refers to as “judgment.” ⁴⁴ For example, it is likely that the tacit knowledge gained from past experiences of creating and exploiting innovations is a core element of entrepreneurial acumen. In this area, initial steps have been taken to include the concept in theoretical models (Aghion & Howitt, 1998; Haruyama, 2009; Mukoyama, 2003; Thoenig & Verdier, 2003) and empirical operationalizations of tacit knowledge have been presented in the microeconomic literature (e.g., Balconi et al., 2007). However, the concept has seemingly failed to gain a wider traction in the neo-Schumpeterian growth literature. This absence is, in turn, notable given the significant role attributed to tacit knowledge in the wider literature on entrepreneurship and innovation (Gertler, 2003; Pérez-Luño et al., 2019). ⁴⁵

Similarly, a superior ability to act, adapt, and learn may explain why some entrepreneurs consistently maintain a competitive advantage over time (Alvarez & Barney, 2010; Alvarez & Busenitz, 2001). Along these lines, a handful of neo-Schumpeterian contributions have sought to capture heterogeneity of innate abilities across entrepreneurs (Acemoglu et al., 2006; Dohse & Ott, 2014; Lloyd-Ellis & Bernhardt, 2000; Michelacci, 2003). However, the discussion has not gained significant traction over time. ⁴⁶ Again, this stands in contrast to the attention that entrepreneurial skill and adaptive abilities have received in the wider entrepreneurship field (e.g., Baker & Nelson, 2005; Eshima & Anderson, 2017).

Moreover, performance and profits may derive from the ability of founding entrepreneurs to build efficient organizational structures that are capable of sustaining competitive advantages through continuous innovation and adaption to changed circumstances (cf. Alvarez & Busenitz, 2001; Penrose, 1959). This involves forming a team of competent coworkers with complementary skills and providing them with incentives to work toward a common goal (e.g., Elert & Henrekson, 2021; Wurth et al., 2022). One contractual solution to accomplish this is to offer stock options to key personnel, thus giving them future ownership stakes in the firms (e.g., Bengtsson & Hand, 2013; Gompers & Lerner, 2001; Henrekson & Sanandaji, 2018). The concepts of managerial and organizational innovations are briefly alluded to in the literature (e.g., Francois & Lloyd-Ellis, 2003; Martimont & Verdier, 2000), but the subjects have not been thoroughly explored in theory. Instead, organizational innovations are typically housed under the wider umbrella of process innovations, where the literature has an explicit focus on patentable processes, while core non-patentable concepts such as managerial structures and their intangible components remain elusive.

In line with Knight (1921), entrepreneurship under uncertainty also emphasizes that ownership is intertwined with entrepreneurship and that remuneration to entrepreneurs—pure profit—emerges from bearing uncertainty as a residual claimant. In contrast, in the absence of uncertainty, ownership itself is unnecessary because any actor can simply contractually achieve the required control over assets and obtain the foreseen returns (Foss et al., 2021). As such, the incorporation of uncertainty may also help to motivate why there is a need for a private sector in neo-Schumpeterian models, which is not necessary in extant models where innovation follows directly from R&D investments. In fact, the ideas of standardized innovation and economies of scale were what led Schumpeter (1942) to his prediction that the government (or rather what he called “laborism”) would eventually replace the entrepreneur as the primus motor of the economy. ⁴⁷ Finally, genuine uncertainty may enlighten our causal understanding of business failure among previously successful entrepreneurs because they can never fully anticipate the value of a novel idea.

Despite criticism from prominent economists (Acs & Sanders, 2013; Bianchi & Henrekson, 2005; Nelson, 1997, 1998), another notable finding is that the conceptual limitations of the neo-Schumpeterian entrepreneur have not been addressed so far within the literature. In fact, during the literature review process, no instances of critical reflections concerning alternative entrepreneurial constructs were identified. At the same time, recent neo-Schumpeterian work explicitly acknowledges the disparity between core measures of R&D (patent output) and economic growth (Aghion et al., 2019). This may signal an increasing awareness in the literature that its workhorse models are currently lacking key elements. Another dimension is the fact that all endogenous growth models after Romer (1986) are supply driven. Gries and Naudé (2021) offer an endogenous growth model specification that is demand driven—illustrating that demand constraints can create fundamental doubt as to how much of potential production (supply) can be sold in the market, leading in turn to reduced entrepreneurial effort and investment. ⁴⁸ This is yet another dimension that may be relevant for further exploration by neo-Schumpeterian growth theory.

A potential objection to our examination is that neo-Schumpeterian growth models seek to explain and predict the macroevolution of the economy, and at the aggregate level, it may be fair to abstract from the genuine uncertainty of innovative outcomes at the micro level. Although the validity of this assertion is debatable per se (Frydman et al., 2019), this line of reasoning is also likely to be debatable in this specific context in at least two respects. First, given that economics seeks to explain the causes of economic growth, a deeper causal understanding is required. Second, economists aspire to provide reliable policy advice and the adequacy and precision of policy proposals hinge on a good causal understanding of the growth process and its microeconomic foundations. Furthermore, given that endogenous growth models—both neo-Schumpeterian and variety-expansion models—make an explicit point of being grounded in microeconomic fundamentals, the argument that it is fair to abstract from the micro level becomes contradictory. In fact, this is at odds with the literature’s own perception and goal of capturing fundamental causes of growth (Acemoglu, 2009, p. 19).

The above points are likely to be particularly relevant for economies at the technological frontier, such as the United States and Western Europe, where the causal effect of R&D on economic growth is weak (Aghion et al., 2019; Bloom et al., 2020) and where there are only marginal opportunities for improvements to basic institutional quality such as enabling free entry, securing property rights, or increasing accessibility to higher education.

A broader understanding of the growth process may lead research onto previously unexplored paths that will increase its explanatory power. For instance, the existence of substantial information problems caused by genuine uncertainty may help explain the global predominance of family ownership of firms (Andersson et al., 2018). This is likely to have significant macroeconomic implications as family firms have been found to systematically deviate from the standard assumptions of firm behavior. Hence, by better understanding the interplay between entrepreneurship, ownership, firm organization, and innovation, growth theory may be augmented in ways that enhances both its predictive power and usefulness for policy analysis and guidance.

So, why do current neo-Schumpeterian models not incorporate genuine uncertainty? The explanation is likely linked to the theoretical and methodological approach of the literature and, specifically, to the prevalence of equilibrium modeling in the field (Hébert & Link, 2007). Equilibrium is fundamentally incompatible with genuine uncertainty; this suggests a need for a more pluralistic methodological approach (Hébert & Link, 2007). In the presence of uncertainty, an equilibrium or “optimum” output of innovative activities cannot be objectively defined. Hence, to yield a richer and more inclusive theory of entrepreneurship and economic growth, a more inclusive approach to economic theory and methodology is needed.

Finally, despite the presented criticism, it is worth emphasizing that in comparison to its neoclassical predecessors, the contributions of neo-Schumpeterian growth models are one step toward a more realistic conceptualization of the economic growth process. In effect, neo-Schumpeterian growth models have reintroduced the notion of an entrepreneur to the core of mainstream economics, and the neo-Schumpeterian literature has contributed to an increased focus on economic history to further our understanding of how institutions and policy enable or impede economic growth.

Concluding Remarks

This study explores the position of the neo-Schumpeterian entrepreneur relative to Schumpeter Mark I and II, that is, Schumpeter (1934), where the entrepreneur is the persona causa of innovation and economic growth, and Schumpeter (1942), where the entrepreneur becomes superfluous. This is accomplished by quantitatively reviewing the terminology applied in neo-Schumpeterian growth theory and by qualitatively reviewing neo-Schumpeterian conceptualizations of entrepreneurship and innovation. The analysis is based on more than 700 peer-reviewed articles on neo-Schumpeterian growth published from 1990 to 2020.

By quantitatively analyzing the literature, we find that less than one-third of all articles include the term “entrepreneur.” Moreover, a mere 1% mention Schumpeter’s key concept “new combinations,” and then only to position their discussion relative to Schumpeterian terminology. Our analysis of the reference lists shows that less than one-twentieth of articles include references to Schumpeter (1934) and about one-tenth to Schumpeter (1942). Rather, the literature adheres closely to its roots in endogenous growth theory, which abstracts from the Schumpeterian entrepreneur. Similarly, Frank Knight and Israel Kirzner, who, together with Schumpeter, are the most important scholarly pioneers in terms of influence on contemporary entrepreneurship research, are absent from the examined literature. ⁴⁹ Only 3 out of 714 articles mention either Knight or Kirzner, and only 2 of these discuss implications for entrepreneurial decision-making. As a corollary, the related concepts of judgment, genuine uncertainty, and alertness are not discussed.

Our qualitative analysis of the literature’s conceptualizations of entrepreneurship and innovation reveals two common themes. First, the neo-Schumpeterian entrepreneur is defined based on his/her role as an undertaker of innovative investments, notably in terms of R&D. Second, the outcome of innovative activity is assumed to follow an implicit probability distribution that is observable ex post, while simultaneously modeling it as if it was objectively known ex ante. Hence, the expected costs and benefits of innovative ventures are assumed to be ex ante calculable.

By assuming that the expected value of innovative activity is fully calculable, the economic role of the neo-Schumpeterian entrepreneur is reduced to that of a routine decision-maker. As a result, the disruptive role of the Schumpeter Mark I entrepreneur becomes redundant, that is, in congruence with the prediction of Schumpeter Mark II. This conclusion is strengthened by the fact that references to Schumpeter (1934) are rare and that the articles do not use the terminology associated with Schumpeter’s early work, including the concept of “new combinations.”

When innovations are modeled as discovered opportunities whose expected value is exogenously given, analysis of the value-creating process becomes unnecessary. This includes the different stages of the innovation process, that is, the processes that connect the conception of a new idea to its subsequent introduction and market dissemination. This abstracts from the distinction between the inventive and entrepreneurial functions, which is a further digression from Schumpeter Mark I. Relatedly, the assumption that the value of a given innovation is objectively and ex ante calculable may partially explain the seeming lack of diversity in terms of how innovative ventures are currently modeled. As a result, other types of innovations discussed in Schumpeter (1934), such as the creation of new or improved organizational structures with the ability to generate and exploit innovations through time, become less relevant. This also explains why discussions of ownership and the nature of entrepreneurial skills are so scarce.

In contrast to the neo-Schumpeterian conceptualization of the innovation process, a key component of the innovation process under uncertainty consists of the value generation process undertaken by entrepreneurs in the absence of calculable outcomes. In this process, the value of a new idea is endogenously imputed based on the subjective valuation of the entrepreneur and, over time, through its dissemination in the marketplace.

By assuming that returns on innovative activity are ex ante calculable, the neo-Schumpeterian conceptualization of economic growth is effectively disregarding key economic antecedents that are a pervasive characteristic of economic life. Hence, by reintroducing the entrepreneur into mainstream growth models, neo-Schumpeterian growth theorists also need to address theoretical and methodological issues that have hitherto remained underexplored, such as the entrepreneurial function as bearing uncertainty and the appropriateness of equilibrium modeling for analyzing economic growth as an innovation process.

Supplemental Material

sj-pdf-1-etp-10.1177_10422587221141679 – Supplemental material for To Be or Not to Be: The Entrepreneur in Neo-Schumpeterian Growth Theory

Supplemental material, sj-pdf-1-etp-10.1177_10422587221141679 for To Be or Not to Be: The Entrepreneur in Neo-Schumpeterian Growth Theory by Magnus Henrekson, Dan Johansson and Johan Karlsson in Entrepreneurship Theory and Practice