2. Background

The year 1959 was a critical one in terms of my personal engagement with evolutionary thinking, and also for other aspects of my career. As I have described elsewhere (Winter, Reference Winter, Hitt and Smith2005), it was in 1959 that a largely fortuitous circumstance directed my attention to Armen Alchian's Reference Alchian1950 paper, Uncertainty, Evolution and Economic Theory (Alchian, Reference Alchian1950). Those ideas initially appealed to me as a possible basis for the theoretical component of the dissertation I had in progress, which dealt with the determinants of corporate research and development (R&D) spending. I was in need of such a basis because I had become convinced that my initial framing of the problem, which featured far-sighted firms attempting to pursue R&D plans that maximized present value, was too far-fetched at the behavioral level to be acceptable as a foundation. After reading Alchian, I turned back to Milton Friedman's paper on methodology (Friedman, Reference Friedman1953). While I was already familiar with that essay, I had not focused specifically on Friedman's use of the ‘natural selection’ argument. Where Alchian had proposed that evolutionary ideas offered a way for economists to avoid reliance on maximization assumptions, Friedman's similar ideas were advanced as a basis for such reliance.

Over the space of a few months, the empirical project on corporate R&D disappeared from the dissertation plan and was replaced by a focus on the evolutionary arguments of Alchian and Friedman – particularly on Friedman's invocation of natural selection as a defense of the standard profit maximization assumption. Unfortunately, this major shift of focus left little time in the career schedule for actual execution of the plan. I had accepted a job at The RAND Corporation, which at the time offered an exciting research environment across a broad intellectual front.Footnote ¹ Initially, I was attracted particularly by the interest and obvious importance of the national security work going on in the Economics Department, which was then under the leadership of Charles J. Hitch.

I joined RAND's Washington office in September; in December I relocated to Santa Monica. There I found Dick Nelson, who quickly became a close friend and also a sounding board for the ideas emerging from the new direction of my dissertation research. These conversations began, also, the merger of the two major strands of what became the Nelson –Winter collaboration – the firm behavior strand and the economic growth/technological change strand.Footnote ² It took quite a while for that merger to come to full fruition (actually, we may hope that it still has some distance to go). Back in 1959–60, I continued to devote part of my time to the development of the Friedman-related dissertation ideas, exploiting RAND's permissive environment. Early formulations of those ideas appeared in a draft shown to Dick and a few others in the middle of 1960. Most of my time, however, went to more ‘mission-related’ work.

Around the same time, a third major perspective on the conception of economics as an evolutionary science came into focus. This reflected the influence of Thomas Kuhn and his book, The Structure of Scientific Revolutions (Kuhn, Reference Kuhn1970). It is my recollection that I first encountered the book, or at least its main ideas, in 1959 or early 1960. The first edition of the book is dated 1962; I am not entirely clear as to how my first encounter with the ideas happened. In the preface to the first edition, Kuhn indicates that the thesis of the book fell into place during his 1958–59 stay at the Center for Advanced Study in the Behavioral Sciences (CASBS) at Stanford. Interestingly, he attributes that consolidation partly to his exposure to the conversations of his social scientist colleagues at CASBS – and his reflection on the fact that they talked past each other in ways that natural scientists usually did not.

In any case, I certainly was already familiar with the ideas when, in the spring of 1960, I had the privilege of encountering Kuhn in person. Both of us attended the Nelson-organized conference in Minneapolis that was ultimately reflected in the celebrated Nelson-edited volume, The Rate and Direction of Inventive Activity (Nelson, Reference Nelson1962). In a conversation at the conference, I asked Kuhn who else had a general perspective on science that was similar to his own. He pointed me to Michael Polanyi's book Personal Knowledge (Polanyi, Reference Polanyi1962 [1958]). That book became another major influence on the evolutionary project, though not so much because of its perspective on science generally. It was the skills chapter, and the general development of the concept of tacit knowledge, that was most important. In an account of the evolution of ‘methods of doing things,’ a key point is that such methods are not fully represented as codified knowledge (Winter, Reference Winter2013). Economic evolution would be a very different thing if they were.

Most relevant to the present paper is the fact that in the early discussions between Nelson and me, Kuhn's then-recent contribution loomed imposingly in the background. It was then natural to see the central role of technological change in economic progress, which was increasingly acknowledged, as an anomalous result for the neoclassical paradigm – perhaps constituting an anomaly in the Kuhnian sense. It was natural to see this as simply the most prominent of many visible indications of the deep clash between the neoclassical assumption that actors maximize over known opportunity sets and the commonplace realities of economic choice – and hence that the anomaly was of even broader significance than might at first appear. It was further natural to speculate that this broad anomaly might conceivably produce ‘crisis’ in the neoclassical paradigm, and such a crisis might lead to a ‘revolution’ in which a new, evolutionary paradigm might emerge.

A half century later, there has been a lot of progress in evolutionary thinking generally and in its application to economics in particular. Meanwhile, the theoretical core of the neoclassical paradigm has matured in ways that make its facial implausibility even greater. Much more productive adjustments and extensions have taken place around the edges, and an econometric revolution has transformed the empirical study of economics. At the same time, much potentially relevant progress has been made in adjacent disciplines, including psychology and technology studies, but with negligible impact on the neoclassical mainstream. The foreseen crisis in economic theory did not develop – though a crisis in the economy ultimately did. The evolutionary revolution has only managed to secure a limited beachhead, but the basic case for it is stronger than ever.

3. A tour of the beachhead and some neighboring positions

The ‘beachhead’ of my title is the large piece of intellectual territory in which scholars with an evolutionary orientation have developed a substantial body of research. A wide-ranging literature lies within that beachhead, and it is not my purpose to review it here, nor would that be feasible given the length constraints of this article and the competence limitations of its author. Instead, I point out some general features of the beachhead, features that seem particularly relevant to the future of evolutionary economics and to its potential claim to paradigm status. A few references are noted for the sake of concreteness, with emphasis on works that offer substantial bibliographies and useful orientations toward literatures that lie close to evolutionary economics but are not strictly part of it – the ‘neighboring positions’.

For a first perspective, we can visualize in imagination how evolutionary contributions are distributed within the field addressed by the remainder of the economics discipline, or by neoclassical economics specifically. Plainly, the scatter is far from random. Much of the pattern is readily understood as a reflection of core commitments of the evolutionary viewpoint. The most obvious and powerful case of that kind is the study of technological change.Footnote ³ The fact that evolutionary economists have directed much of their attention to that field is arguably an ‘over-determined’ result. It is useful to review the several causes that have promoted this attention to technological change.

Addressing the list of causes from the top down, there is first the long-term orientation of evolutionary studies in general and the derived concern with understanding the mechanisms of long-term economic change. Then there is the Schumpeter-derived emphasis on the key role of innovation and technological change in producing the rising living standards that generally accompany modern economic development – which evolutionary economists regard as the quintessential example of a set of facts that the economics discipline ought to address convincingly. Then there is the fact that close examination of the histories of particular technologies typically reveals a clear pattern of ‘cumulative causation’ (Veblen, Reference Veblen1898), so that real explanation of the present and the options for the future requires attention to origins in the past. This observation in turn opens a field of fertile inquiry into the mechanisms of technological evolution, a field with which the formal apparatus of neoclassical production theory can make no contact. It is a field in which evolutionary economists have been particularly active, but there have also been major contributions from other sources, particularly from students of technology per se and from economic historians who have emphasized the role of technology.Footnote ⁴

Such a setting is also one that severely challenges any attempt to explain key decisions in rational choice terms. This is partly a reflection of mundane points about limited attention and computational capacity, partly a matter of Knightian uncertainty and partly a matter of the complex sequential structure that typically characterizes the decision options. Obstacles to rational choice formulations that are still more fundamental arise from ontological uncertainty (Lane and Maxfield, Reference Lane and Maxfield2005) – the difficulty of seeing in advance the nature of the entities that will shape the evolution of the system in the future, and with which future choices will be concerned. Serendipitous discovery of unanticipated entities is a common happening in scientific and technological history, with Roentgen's discovery of X-rays being perhaps the classic example.Footnote ⁵ Such examples cannot be mapped convincingly into the ‘states of the world’ framework of contemporary rational choice theory; they illustrate that history is genuinely an unfolding process that brings forth wholly unanticipated phenomena from the indeterminately large range of possibilities.Footnote ⁶ Finally, realistic studies of technology and its changes make clear the inadequacy of the standard economic accounts of productive knowledge, represented analytically by production functions and production sets but based fundamentally on a ‘recipe’ metaphor. These accounts encourage a systematic overemphasis on the role of abstract information and its representations, and correspondingly understate the role of various forms of tacit knowledge, especially the forms that are the fruit of learning from sustained experience (Dosi and Nelson, Reference Dosi, Nelson, Hall and Rosenberg2010). Also neglected by the standard framework is the role of firms and other organizations in performing the related social functions of storing, applying and extending the more complex forms of productive knowledge.

These are not small matters. Each consideration has powerful implications for the way economic reality is perceived and analyzed. Each thus contributes to a sort of fence across the intellectual terrain, tending to impede realistic and probing engagement from the neoclassical side in significant domains like trade policy, antitrust, intellectual property, and public support of science and education – to name only a few of the areas where evolutionary and mainstream instincts tend to diverge. Meanwhile, work goes forward on the evolutionary side of the fence – the territory where some evolutionary economists grew up, and to which others migrated long ago, having found some sort of gate through the fence. They all find the territory congenial and have been highly productive there.

The study of firm behavior is a second broad area where evolutionary theorizing has shaped a substantial body of research. The reasons for this include those just adduced in the case of technological change, but there are differences in emphasis and in specific scope. Putting the matter rather simply, there are many scholars who cannot swallow the neoclassical representation of organizations as optimizers, and they are alert to alternatives. Although that basic neoclassical commitment is particularly hard to accept when studying the uncertainty-ridden domain of technological change, its limitations are very evident in other spheres as well – including some that involve problems that are much less challenging from a decision-theoretic viewpoint. Conceptions of firm behavior developed in the evolutionary literature, emphasizing organizational routines and capabilities, have long appealed to many of these scholars. The appeal has been particularly strong for those whose main interest is in the great Schumpeterian themes – important phenomena of change in the real economy, including the behavior and strategies of firms in industries where innovation is important, the distinctive roles of large and small firms, and the phenomena of industrial dynamics – particularly the characteristic patterns associated with the emergence and growth of new industries.Footnote ⁷

A third area of significant concentration of effort involves the role of institutions. In its origins, evolutionary economics was institutional economics. That thread remains important in the modern literature. As in the case of firm behavior, there is a cluster of work near the key topic of technological change. Diverse institutions support the technical progressiveness of the system in diverse ways. Many of these institutions stand at a considerable distance from those that express the incentives and coordinating principles of the ‘market economy’, assuming the latter is understood as a field contested by for-profit firms. These areas are addressed in the segments of the evolutionary literature concerned with business history, national innovation systems, intellectual property policy, and with industry's connections to university and government-supported research.Footnote ⁸

More generally, evolutionary economists see the institutional environment as much richer and more consequential than the term ‘market economy’ suggests. Cultural and institutional factors shape the development of consumer preferences across the full course of life. Business practices are profoundly institutional, in the ‘rules of the game’ sense of what ‘institutions’ are. They reflect prevailing norms and cultural attitudes at multiple levels of analysis, from the broad society down to the specific norms governing contract in an online or telephone market. Sector by sector, the political, regulatory, and economic aspects of the marketplace co-evolve, partly as a consequence of the activity of self-interested actors, and partly as a reflection of the ‘law of unintended consequences’.

While these three areas have been the locus of a large proportion of their total activity, evolutionary economists have certainly ventured more widely. For example, consumer behavior is a topic that has suffered relative neglect compared to firm behavior, but evolutionary economists have not ignored it (Witt, Reference Witt2001; Nelson and Consoli, Reference Nelson and Consoli2010). Similarly, the evolutionary literature has given inadequate attention to macroeconomics, but there have been significant efforts to rebalance the allocation.Footnote ⁹ Analysis that aims at direct policy relevance has, similarly, been in relatively short supply – but not entirely absent.

In the three key areas, in the regions where evolutionary economists have ventured more tentatively, and in the broader field of influence of evolutionary ideas, the cause of evolutionary economics has been advanced by the efforts of a broad range of intellectual allies. As is discussed below, these allies occupy much of the adjacent intellectual territory. Collectively, they not only add weight to the case for evolutionary economics, but contribute to its coherence by filling significant gaps. Some offer critiques of neoclassicism that complement the evolutionary one; others apply distinctive tools to illuminate questions that are of common concern. In general, the theoretical perspectives of the allies are congenial from the evolutionary point of view and diverge significantly from the neoclassical view. While there certainly are important differences among these various perspectives, the realities of competition and specialization in academe often promote exaggeration of the differences, and an under-exploitation of the opportunities inherent in the commonalities. As we attempt to envision the future of evolutionary economics, it is important to recognize that it is likely to resemble the past at least in the sense that it will be profoundly influenced by the collective strength and contributions of the allies.

A dense network of relationships connects the allied fields to each other and to evolutionary economics. This density testifies to the coherence of the total picture, but also enhances the challenge of providing an interpretive structure for the picture. It seems useful to identify two broad categories of allied contributions, though the lines of demarcation between them are not sharp.

In the first category, the overall scope is narrower than the evolutionary perspective, but the questions addressed are both of interest to evolutionary economists and potential candidates for their research attention. Scholars studying technological change have long been particularly prominent among the allies, and it is apparent that the evolutionary economics of today would be unrecognizable if the shaping influence of those allies were absent. Indeed, there the fields clearly overlap (Verspagen and Werker, Reference Verspagen and Werker2004). Many of the allies concerned with technology and innovation were, or are today, associated with one of the several European centers of excellence in this field of inquiry. They sometimes have (or had) colleagues who self-identified as evolutionary economists – and sometimes the lines of demarcation are decidedly unclear. Historically, the leadership provided early on by Christopher Freeman, and subsequently by Keith Pavitt, at the Science Policy Research Unit of the University of Sussex, was particularly important in shaping modern evolutionary economics. There are multiple lines of descent from that origin to the European centers of excellence today, not to speak of its earlier shaping influence on the Nelson and Winter version of an evolutionary theory. Meanwhile, far to the west, other lines of long-term influence derived from the work of Nathan Rosenberg, Paul David, and others.Footnote ¹⁰

Other types of allied scholarship lie at a slightly greater distance from the core concern. Decision making is studied from various directions, with a variety of methods and diverse questions in view. There is much work on decision biases at the individual level, with empiricism that is mostly experimental; see, for example, Camerer et al. (Reference Camerer, Lowenstein and Prelec2005), Kahneman (Reference Kahneman2011), and Tversky and Kahneman (Reference Tversky and Kahneman1974). A quite different decision-related literature is concerned with leadership and internal politics in large organizations, as in March (Reference March1962), Cyert and March (Reference Cyert and March1963), and Bower and Gilbert (Reference Bower and Gilbert2005), in some cases with an evolutionary framing, e.g., Burgelman (Reference Burgelman1994). It is typically case-based or historical in method, for obvious reasons. In quite a different direction, there is the field of the population ecology of organizations, a field where the theoretical perspective joins evolutionary thinking to sociology and the subject matter overlaps the study of industry evolution and the statistical distributions of firm growth and size (Hannan and Freeman, Reference Hannan and Freeman1984). We also owe to sociologically oriented scholars some particularly well-documented accounts of how economic theorizing has influenced business practice (MacKenzie, Reference MacKenzie2006; Zbaracki and Bergen, Reference Zbaracki and Bergen2010). Transaction cost economics in the Coase–Williamson tradition (Coase, Reference Coase1937; Williamson, Reference Williamson1985) shares the evolutionary emphases on bounded rationality and the importance of institutions, and that literature offers a wealth of details about how business transactions are actually done. Still other perspectives are found in literatures devoted to realistic treatment of business practice in particular functional areas (finance, accounting, marketing, operations). None of these approaches adopts the neoclassical perspective on choice, and each complements evolutionary economics in a significant way.

The second category of allied contributions includes lines of inquiry that help to paint the large-scale picture of the economic system, located in space and time, that evolutionary thinkers aspire to create. For example, studies of socio-cultural evolution address even larger ‘systems’ and larger time scales than those in evolutionary economics, and relate to key elements of economic behavior. A question of particularly broad significance is this: to what degree has biological evolution produced human beings with a natural inclination toward ‘altruistic’ or ‘cooperative’ behavior? On this topic and cultural evolution more generally, see, for example, Bowles and Gintis (Reference Bowles and Gintis2011), Hodgson and Knudsen (Reference Hodgson and Knudsen2010), and Richerson and Boyd (Reference Richerson and Boyd2005). Scholars of business and economic history form another substantial category, offering contributions at all levels of analysis, from the details of business routines to the historically intertwined developments of capitalism and technology. In some of these contributions, an evolutionary interpretation is explicit (Mokyr, Reference Mokyr1990; Murmann, Reference Murmann2003). In other cases there is abundance of material for an evolutionary story but not much of the story line itself.Footnote ¹¹ While the more self-consciously evolutionary part of the historical literature is particularly welcomed by evolutionary scholars, the rest is certainly helpful. More important, it signals a wealth of opportunity that could be seized if the links of evolutionary economics to the historical fields were strengthened.

A final field in the second category is organization studies. Particularly in the part of that broad field shaped by the influence of the Carnegie School there are long-standing connections to evolutionary economics (Cyert and March, Reference Cyert and March1963). Those connections and re-combinations persist today in the work of the many intellectual descendants, and there are major overlaps with the field of strategic management (Gavetti and Levinthal Reference Gavetti and Levinthal2004; Gavetti et al. Reference Gavetti, Levinthal and Ocasio2007). Even at the telegraphic level of this review of the allies, it is not realistic to attempt even enumeration of the points of contact between evolutionary economics and contemporary organization studies. One point that seems to stand out is the way computational modeling has become prominent as a technique of theoretical research in both fields, particularly where there is a focus on the dynamics of organizational learning. (See the papers collected in Lomi and Larsen (Reference Lomi and Larsen2001), and the references there, and the rapidly expanding literature exploiting the ‘NK’ modeling technique.)

Although the scholars in question might not take pride in the title of ‘allies of evolutionary economics’; evolutionary economics has always drawn heavily on the mainstream tradition (Mazzoleni and Nelson, Reference Mazzoleni and Nelson2013) and continues to do so today. These contributions from the mainstream discipline overlap the two categories of the discussion above. Avoiding a long digression in the full history of economic thought, I note that Kenneth Arrow not only figured importantly in the afore-mentioned Stanford–Yale–Sussex synthesis on technology (Arrow, Reference Arrow and Nelson1962), but has provided much relevant insight on organizations; see, e.g., Arrow (Reference Arrow1974). A number of empirically-oriented economists have helped to establish the factual context that evolutionary economists largely take for granted. A prominent and entirely contemporary example is the work of Nicholas Bloom and John Van Reenen and their numerous co-authors and colleagues. They have produced a remarkable body of significant evidence and analysis about the management of business firms (see Bloom and Reenen (Reference Bloom and Reenen2010), and the many works cited there . . . but recognize that this overview is now out of date). This work confirms again the fact that there is a large amount of heterogeneity in firm behavior and results (cf. the valuable review by Syverson (Reference Syverson2011)), and shows that the selection environment is an important factor shaping that heterogeneity. In particular, a weak selection environment is reflected in poor management. This result matches the basic evolutionary stance: it is not the case that individual economic actors are reliably effective ab initio in addressing the problems that confront them, and it is not the case that selection decisively eliminates the low performers, but it is the case that the selection environment powerfully influences the realized distribution. This work and much related work on firm heterogeneity (see the discussion in Jacobides and Winter (Reference Jacobides and Winter2012)) speak quite resoundingly in support of the evolutionary view, though it is not reported that way.

Finally, I would like to point to two particularly important allies. They come to mind because of their contributions, and also because of the sad circumstance that these colleagues were lost to us in 2013, passing away at ages when they were still extremely active scholars and when there were strong intimations of even greater contributions in the future.Footnote ¹² Michael D. Cohen (1945–2013) was an organization theorist in the Carnegie School tradition (thus, not an economist of any stripe), and one with particularly strong interests in the psychological foundations of individual behavior in organizations. His 1994 paper with Paul Bacdayan (Cohen and Bacdayan, Reference Cohen and Bacdayan1994) pointed convincingly to the possibility that the concept of organizational routines could be firmly grounded in individual psychology, identifying that issue as a key intellectual cause and moving the cause forward with an ingenious experiment. That work underscored the fundamental character of the distinction between habit-based behavior and deliberate choice, arguing persuasively that the distinction rests on a corresponding distinction at the level of mechanisms in the brain.Footnote ¹³

The second key ally lost last year was Steven Klepper (1949–2013). Klepper was an economist with strong empirical interests and econometric skills. Beginning in his graduate student years, he produced a large body of work on industry evolution (Klepper, Reference Klepper1996; Klepper and Graddy Reference Klepper and Graddy1990; Klepper and Simons, Reference Klepper and Simons2000). That work and its influence developed roughly in parallel with the efforts to explore a dynamic theory of industry structure in which technological change is central – efforts inspired by Nelson and Winter (Reference Nelson and Winter1982) and pursued by many others. The contributions of Klepper (et al.) filled an important gap in that program. Nelson and Winter had discussed and modeled industry evolution, but they did not attend significantly to the historical phenomenology of the subject. That phenomenology, the stories of particular technologies and industries as they come and go in historical time, constitutes a key part of the causal linkage that relates the firm-level, behavioral aspects of the evolutionary view to the aspects that feature the aggregate level and the long term. Klepper's work offered much insight and extremely helpful documentation of the working of that linkage. Although Klepper did not, as far as I know, ever self-identify as an evolutionary economist, he did become increasingly aware that he was not a neoclassical economist.Footnote ¹⁴ In his theoretical modeling, he favored a restrained version of neoclassicism, in which firms are represented as conventional optimizers on some matters, but do not attempt implausible inter-temporal optimizing. He did not adopt the ‘routines and capabilities’ perspective on those questions. This approach served him well, but a more behavioral perspective on the allegedly optimized decisions might, I think, have served him even better.

This tour of the beachhead provides some perspective on the findings of Hodgson et al. (Reference Hodgson, Jarvinen and Lamberg2013) in their bibliometric analysis of the evolutionary literature. They remark that ‘One of the key problems (in the field of ‘evolutionary’ studies in economics and business) is not a lack of diversity but its fragmentation and specialization’ (p. 25). This is in part a reflection of the roles played by the numerous and diverse allies, who are aware of the evolutionary project in economics – and often cite it favorably – but are not actually focused on advancing that project. Because of the importance of the allies, the project is wide in apparent scope but lacking in coherence, as Hodgson et al. suggest. At the same time, in the part of the beachhead occupied by more committed adherents of the evolutionary program, there is relatively little interest in engaging many of the traditional areas of interest of the economics discipline. Hodgson et al. confirm the centrality of Nelson–Winter (Reference Nelson and Winter1982) in the evolutionary literature. That book contained, half-deliberately, quite a large number of ‘loose ends’ – sketches of starting points for inquiries that might have explored matters of broad and traditional interest in the economics discipline. But the loose ends on those traditional topics were not picked up; it was the students of innovation and of business behavior who paid most of the attention that was actually paid. This again seems consistent with the Hodgson et al. findings. Nelson and Winter were hoping to reform economics; what they actually achieved was something smaller and different, and this fact left visible traces in the bibliometric record over a span of several decades.

Among these many contributions, there are numerous examples where the scholars that I have designated as ‘allies’ offer the most persuasive evidence on theoretically significant hypotheses that evolutionary economists generally espouse. Against that background, it is easy to imagine a claim that the evolutionary economics program is superseded by the diverse accomplishments of the allies. In response to that suggestion, evolutionary economists might point out that their focus is on economics, understood in Marshal's sense . . . the ‘study of mankind in the ordinary business of life’ (Marshall, Reference Marshall1920: 1). Valuable as the allies’ contributions are, most of them do not consider that to be the subject matter.Footnote ¹⁵ Further, the evolutionary view places the ordinary business of life in its large spatio-temporal context – large enough, at least, to see the importance of the long-term interrelations of economic, technological, and institutional progress, and appreciate their great significance for the human race. Neither the allies nor the mainstream economists fully embrace that combination, nor, or course, do they embrace the basic explanatory toolkit shared by the evolutionary sciences generally.

Therefore, important as they are, the contributions of the allies do not offer a substitute for the past accomplishments or the continuing development of evolutionary economics. In particular, they do not exploit the value of its dual contact with, first, the day-to-day operational realities that we all interact with in today's economy, and, second, its contact with the long-term and large-scale phenomena of economic growth. Walk into a coffee shop and order a latte, take out your phone and make a call to secure a flight reservation, arrive at the boarding area to board the flight that you reserved on your phone, get on a very large airplane of a design that first flew in 2005. Those are four direct encounters with organizational routines/capabilities/‘methods of doing things’. There are deep evolutionary histories lying behind the details of each of those four encounters, involving interacting developments in technology, business behavior and institutions. Assuming that the ‘ordinary business of life’ is indeed the subject matter, careful examination of any of them would do much to further understanding of it.

As Veblen said,

That was the program; that is the program, and evolutionary economists have done a great deal to pursue it. In that endeavor, they have had many invaluable allies, but no rivals who are similarly committed to that same cause.

4. Core strengths of the evolutionary view

In evolutionary economics, the fundamental theoretical commitments are at the system level and are very general. It is fundamental that economic evolution, like biological evolution, is a time-consuming dynamic process involving sub-processes of variation, inheritance/retention and selection. In such a context, evolutionary explanations necessarily attend to antecedents: answering a question about the ‘origins’ of an observed phenomenon requires attention to the historical/evolutionary process from which it emerged.Footnote ¹⁶ It is likewise fundamental that populations of entities are marked by competition for scarce resources. By contrast, commitments regarding the behavior of the entities are tentative, instrumental and, at least in principle, empirically grounded. Thus, the account of actor behavior invoked in evolutionary theory remains open to the possibility of empirical correction from any and all perspectives, which is a fundamental strength that is emphatically not shared by the mainstream paradigm.

Further, the quest for fruitful (though tentative) commitments is strongly shaped by the consideration that evolutionary theory emphasizes the elements of continuity in actor behavior (the broad rubric being ‘inheritance’). Thus, the extent of such continuity, and its character and dynamic laws, is a recurring issue in both theoretical and empirical inquiry. Illumination about those ‘dynamic laws’ is derived from psychology, particularly the psychology of habit, institutional economics, organization studies, and other sources. The explanatory stance emphasizing the shaping role of the past is complemented by enthusiasm for empirical study of what has actually happened – whether in economic history, or in current domains of technological and organizational change. For policy analysis or for short-term prediction purposes, these evolutionary orientations encourage close study of the ways of doing things that are recently or currently prevalent – on the premise that those ‘ways’ are likely to persist in the short term, and that a stronger basis for short-term prediction is hard to find.

Admittedly, this approach is more effective in some cases than in others. Evolution sometimes moves fast – challenging, for example, the ability of the creators of flu vaccines to guess which strain will be dangerous next year. Regulators of the financial sector are similarly challenged by the task of guessing which innovative financial product will be dangerous next year. These phenomena arise in evolutionary processes in which ‘inheritance’ is weak and ‘mutation’ is rapid, but they are still evolutionary processes; they are ones that invite enhanced attention to the mechanisms of variation relative to those of inheritance and selection. Beyond that (evolutionists would claim), there is not much that can actually be done to improve the quality of the guesses. Evolution is fundamentally creative, and that inevitably presents a challenge for prediction efforts.

In these respects, there are fundamental contrasts between the evolutionary view of behavior and the optimization approach of the neoclassical mainstream. The latter looks to the conjectured future, not the observable past, to understand the actor's behavior today. It sees the elements of continuity in terms of ‘frictions’ that impede an otherwise perfect equilibration process. Its preferred methods provide an impulse toward over-simplification that tends to persist even when the high stakes would warrant greater care – as in the domain of financial regulation. And it appraises institutional arrangements largely by reference to a timeless, market-centered ideal. As has been widely remarked, these commitments take a toll in terms of relevance, leaving analysis floating in an abstract, fictional world that is disengaged from the realities of behavior.

Yet, in spite of embarrassing shortfalls in performance, such as the failure to anticipate the Financial Crisis and the Great Recession, the mainstream economics discipline today gets a lot of respect from society. Individual economists rise to positions of power, sometimes prove to be able administrators, and collectively constitute an influential demographic within the larger population of ‘talking heads’. What is the basis of this success? It does not seem likely that it is attributable to the scientific value of the discipline's accumulated wisdom on individual choice, since so much of that wisdom is reflected in discourse only at an abstract level. There are other intellectual camps that are engaged with understanding choice behavior, notably including the behavioral economists, the psychologists and the operations research scholars (who actually practice and extend the use of optimization techniques), and they suffer much less from excess abstraction.

Rather, it seems that the perceived competence of economics lies in its fund of insight at higher levels of analysis – broadly speaking, ‘the system level’. From Adam Smith's day to the present, economists have sought to explain and improve the functioning of whole economies (trade policy, general equilibrium), and smaller social units down to the level of dyads (bargaining, prisoner's dilemma). The strong thread of continuity in that effort is reliance on a basic view of individual choice that emphasizes self-interested motivation and careful economic calculation. The manifestations of that basic view have, however, changed dramatically from Smith's day to the present, to the point that the kinship between Smith's view and a contemporary theoretical model can take quite a while to explain – perhaps several semesters.

The significant question posed is whether the system-level insights are valued because of, or in spite of, the characteristic set of commitments about individual-level choice. Here, there is something of a paradox. That many mainstream economists now take the ‘because of’ view is suggested by their concern with the soundness of ‘microfoundations’ as a desideratum in judging a theory. They understand ‘sound microfoundations’ in terms of conformity to the rational choice paradigm. Mainstream economists have long since abandoned any serious engagement with the actual truth of these foundational propositions about choice; what appeals about them is their received status as highly evolved conventions of the neoclassical paradigm – which accounts for their ascribed legitimacy in the foundational role.

To anyone who lacks commitment to those conventions, this approach seems peculiar. After all, since the warrant for conclusions arrived at by deductive logic depends crucially on the truth of the premises, there is no logical warrant for reliance on deduction from mere disciplinary conventions as such; so what is the scientific payoff to obsessive concern with the fine points of such arguments? In any case, observers from outside the economics mainstream (and its scattered colonies) are rarely enamored of rational choice, and often respond with enhanced skepticism rather than assent when reliance on those particular ‘microfoundations’ is emphasized. The more esoteric the model expressing that reliance, the more the eyebrows go up in the lay audience. If the system-level conclusions are taken seriously, it is in spite of those cherished microfoundations.

Yet, the respect for the system-level insights does remain. As suggested above, I believe that the respect for these insights (and the belief that they are the accepted update of Adam Smith) is the main explanation for the substantial prestige that the economics discipline retains. This can be explained partly on the ground that the insights can be largely justified on the basis of much less restrictive claims about individual behavior than those formally invoked – a view that many economists share (and one closely related to the Friedman-generated ‘realism of assumptions’ controversy).

A more important consideration is that system-level insights are particularly valuable, and the competition on the supply side of that intellectual marketplace is mostly unimpressive in both quality and quantity, especially as it relates to the economic system specifically. I believe that this is, in fact, the crucial consideration. Theories that establish a good frame for systematic discussion contribute value that is partly independent of the accuracy of the picture presented in the frame.Footnote ¹⁷ For this reason and others, informed observers from outside of economics may attend to the conversation and respect the system-level insights of economists, or the policy conclusions that are derived from such insights, and the more so if they stand at a safe distance from the underlying logic.

Evolutionary economics is oriented to the system level (or the ‘population’ level) from the start, and is not encumbered by commitments to fiction at lower levels of analysis (organism, individual, organization). Its actual commitments encourage precisely the engagement with reality that many observers find lacking in the mainstream. Why then, does the evolutionary approach struggle to gain attention while the visible ‘talking heads’ offer policy conclusions based on theoretical arguments that nobody actually believes? That is a large question of intellectual evolution that deserves attention in its own right, but it is too big for the present essay.

5. Evolution and intentionality

According to the argument just advanced, a core strength of the evolutionary approach is that it is not constrained by commitments to a specific and unrealistic view of individual actor behavior. It can be argued in reply, however, that this liberation is purchased at too high a price. At the individual level, the theory is left underspecified. Further, it fails to incorporate the strength of the basic ideas about motivation and calculation that have powered economic analysis since the time of Adam Smith. These deficiencies, the critic says, are not adequately compensated by the theory's embrace of considerations of learning, habit and organizational behavior.

A critique along this line provided the stimulus for the recent flurry of efforts in the management literature that explore the ‘microfoundations’ of theories of organizational routines and capabilities. The issues raised in that discussion have much bearing on the question addressed in the following section of the present essay, which considers the prospects of the evolutionary paradigm in future conflict with the mainstream.Footnote ¹⁸ They involve, fundamentally, the question of how an evolutionary theory should address the intentionality of human action – a question important enough to deserve separate treatment.

A persistent source of confusion in the discussion of evolutionary economics derives from the apparently sharp contrast between the calculated behavior featured in mainstream economics and the understanding that evolution is ‘blind’. It is easy to identify the root of this confusion as inappropriate reliance on the example of evolutionary theory presented by biology, where much philosophical significance is assigned to the claim that evolutionary processes are entirely non-teleological. No such sweeping claim is made in the context of social and economic evolution. There is, nevertheless, a major difference between mainstream and evolutionary views of intentionality, and the ‘easy’ resolution of the problem leaves open the question of where intentionality fits in the evolutionary picture.

Some participants have emphasized that the routines concept in the evolutionary economics framework responds to the theoretical need for an inheritance mechanism in an evolutionary theory, and this is certainly correct. Some have gone beyond that point and argued that routines must be conceived as simple and/or highly inertial and rigid for them to play that theoretical role, and this is certainly not correct.Footnote ¹⁹ The response is at three levels, corresponding to the observations that routines may be broad in scope, flexible, and intentionally designed – though no particular routine is necessarily any of those things.

First, as to scope, many routines enable complex, coordinated responses to information arising from a rapidly changing environment – as is true also of many genetically-determined traits of animals. It is possible for very complex behavior to be highly routinized, repetitive, and ‘rigid’ – but only in the sense that the same stimuli reliably produced the same responses, and not at all in the sense of the manifest constancy of a simple behavior. Instead, variation in the environment translates systematically in to variation in behavior. In such cases the empirical parsing of a particular behavioral pattern between its more inertial ‘genotypic’ aspects and its reactive ‘phenotypic’ elements can be a tricky business, as is vividly illustrated by the long-running controversy about language acquisition by infants.

Second, as to flexibility, the practiced pattern of the routine may not control all relevant details of required behavior, so some sort of intentional adjustment necessarily occurs as behavior unfolds. Alternatively, adjustment may be forced by implementation difficulties. Over time, some of these adjustments may become habitual, leading to change at the level of the routine itself. Further, a complex habit or routine may have occasions for deliberation built in; the immediate response that is automatically triggered by some stimulus may be ‘think carefully about what to do next’. Or, a relatively rigid routine may govern the preparation phase for a subsequent effort at creative problem-solving.Footnote ²⁰ In extreme cases, the ‘routine’ offers such limited specific guidance that it might better be called a ‘heuristic’ or a ‘strategy’ or a ‘principle’ (Cohen et al. Reference Cohen, Burkhart, Dosi, Egidi and Winter1996).

Finally, as Nelson and I pointed out long ago, the repetitive behavioral patterns of the routine may themselves have been intentionally selected in some antecedent process – perhaps by an optimization technique, or through the work of a project team (Obstfeld, Reference Obstfeld2012). Indeed, the term ‘standard operating procedure’ is sometimes (incorrectly) taken to be roughly synonymous with ‘routine’, and such procedures are typically the product of intentional design efforts of some kind. A large part of the world's efforts to improve organizational performance takes the form of intentional change in organizational routines. This provenance does not make the resulting patterns less routine, once the design process is in the past and specific responses are appearing automatically.

For evolutionary theory, the important point is that patterns are established and then subjected to real-world feedback that strongly affects what happens next. Because that feedback is often difficult to anticipate, especially at some distance into the future, there is still a sense in which evolution is ‘blind’: intentionality is expressed situationally, and with imperfect understanding of its implications in the future or in spatially remote locations.

These diverse possibilities are very real, and easily illustrated. We should thus be very wary of letting any particular example drive us toward a general conclusion that ‘this is what routines are really like’. When these possibilities are viewed in relation to the full scope of habitual or routinized behavior, it becomes apparent that the opposite possibilities are also real and easily illustrated: some habits and routines are minimally shaped by deliberative processes of any kind; they may be acquired inadvertently and exercised subconsciously. Others may have become rigid over time as a result of processes that are quite independent of functionality considerations. Still others are deliberately designed to be rigid as a matter of operational necessity; the system interactions and sensitivities are such that giving free rein to local innovation presents excessive hazards to overall coordination. The picture is challenging in its complexity.

In Winter (Reference Winter2013), I advanced a proposal for sorting out this complexity. It involves a schematic representation of John Dewey's psychology, as succinctly presented in Dewey (Reference Dewey2002 [1922]). Dewey saw behavior as shaped by the joint action of habit, impulse (emotion), and intelligence (deliberation). This tripartite scheme offers an orderly framework for sorting out the complexity of behavior. Near the habit vertex of the ‘Dewey triangle’ we have behaviors that are highly automatic. At the deliberation vertex we have behaviors that are carefully considered and thus reflect intentionality in the most specific sense. At the ‘impulse’ vertex we have behaviors that are substantially ‘hard-wired’ though not habitual; they manifest what evolution has determined to be useful responses to extreme motivation – ‘fight or flight’, for example. Actual behavioral patterns may hover close to a vertex, or maintain a stable intermediate position, or wander in the triangle. This is a complex picture, but it is not quantum mechanics – or even ‘rocket science’. It is complex, but manageable. It is complex by the standards of single-vertex theories like mainstream economics, which is a deliberation-based theory to the limited extent that it is willing to reference any psychology at all. By the standards of what existing knowledge can tell us about behaviors near any single vertex, it is extremely simple – but those vertices tend to correspond to the domains of disciplines that rarely talk to each other, so the relationships tend to be obscured.Footnote ²¹

The appropriate scientific response to such complexity is to sort the cases carefully, examine them individually, and then illuminate the unifying generalizations and, crucially, the role of the shared origins behind the current diversity. This general approach has been followed with great success in developing evolutionary accounts of, for example, chemical elements, biological species, and (perhaps surprisingly) mineral species.Footnote ²² It is also what evolutionary economists propose and pursue in the quest for understanding of economic behavior. In the social sciences, unfortunately, a scholar who commits to this effective and cumulative but incremental program may have trouble getting published in the prestigious journals where narrower conceptions hold sway.

In summary, the stance of evolutionary economics is that complex mixtures of habit and deliberation are involved in the shaping of behavior. In the processes of design or modification of routines, there is often intelligent deliberation about means–ends relationships, operating alongside other factors. Such deliberation is a cousin to the rational calculation featured in mainstream theory, but (with rare exceptions) it is only a distant cousin. The distance between the concepts is accounted for by a list of considerations that have been examined extensively in the evolutionary literature, technology studies, organization theory, and other fields.Footnote ²³ Further, evolutionary theorists have produced a wealth of models in which goal-directed mechanisms are a feature of the innovation process that modifies routines.

6. Paradigm change and scientific progress

It is quite plain that there are fundamental differences in view between evolutionary economics and the neoclassical core of mainstream economics – differences that can be called differences in paradigm, with at most a modest stretch of Kuhn's terminology. It is clear also that the neoclassical paradigm is currently dominant. That dominance might be permanent, but it is worth considering the fact that even very established paradigms sometimes give way.

Among the several examples of paradigm change analyzed by Kuhn, the twentieth century revolution in physics stands out in several respects. The scope and prestige of the accomplishments of Newtonian physics were so great that the overthrow of such a great system remains an astonishing fact in retrospect, even after we have had a century to absorb it. As Kuhn remarked, the example shows that ‘even the most striking past success provides no guarantee that crisis can be indefinitely postponed’ (Kuhn, Reference Kuhn1970: 121). For this reason and others the revolution in physics provides a useful reference model for major scientific change – useful in part because it is in some ways so extreme. The example evokes and illustrates important questions about the character and consequences of a paradigm shift. It can serve as a guide for imaginative investigation of how the paradigm conflicts in economics might play out. I use it for that purpose here, noting some parallels between the real example in physics and the one in economics, which I think of as set in a hypothetical future. For conciseness in drawing these parallels, Newtonian physics and neoclassical economics will sometimes be referred to as ‘the old paradigm’, while relativity and evolutionary economics will be ‘the new paradigm’. That there are hazards in even suggesting such parallels is obvious, but the potential gain in insight seems worth the risks of the exploration.Footnote ²⁴

In any case of paradigm change, the question arises of whether the truths featured in the young paradigm are essentially complementary to the truths featured by the old (a possibility inviting the claim that both sets are valid), or whether the old truths are undercut at a fundamental level. That broad question can be framed in different ways. For the physics case, Kuhn explored a strong form of the complementarity thesis, considering the question of whether Newtonian physics can appropriately be considered a logical special case of relativistic physics, though subject to a domain restriction – the domain of phenomena unfolding at far below light speed. In short, can Newton's physics appropriately be considered the same as Einstein's physics, once that domain restriction is taken into account? Kuhn's analysis arrives at a negative conclusion. He notes that while it would be possible to produce in this way a formal justification for an accurate Newtonian prediction, such a justification would use Einstein's language and conceptual apparatus. Thus, the fact that some terms are common to the two theories (e.g., ‘time’, ‘mass’) is superficial and potentially misleading; the two theories are saying different things.

The issue here is the distinction between scientific truth and the pragmatic realities of everyday action or sound engineering. As the foregoing suggests, Kuhn took the view that the old paradigm of physics was scientific error:

To say that ‘Newton was wrong’ is not to deny the massive achievements of Newtonian physics, or to deny its continuing utility today. But utility is not truth. Regarding his analysis of the status of Newtonian physics as a special case, Kuhn observes, ‘Our argument has of course explained why Newton's Laws ever seemed to work. In doing so it has justified, say, an automobile driver behaving as though he lived in a Newtonian universe’ (p. 102). Further, ‘Though an out-of-date theory can always be viewed as a special case of its up-to-date successor, it must be transformed for the purpose. And the transformation is one that can be accomplished only with the advantages of hindsight, the explicit guidance of the more recent theory’. (pp. 102–103). Here, Kuhn's ‘always’ pertains to the task of explaining why the old theory ‘ever seemed to work’, and certainly not to the domain where the old theory is plainly inadequate. An engineer can design a good bridge without being alert to relativistic physics, but relativistic physics can be invoked to explain why that is the case. The bridge still manifests the underlying truths of relativistic physics, regardless of whether the difference between those truths and the Newtonian approximations to them needs to be taken into account by engineers. By contrast, the engineer designing a supercomputer needs such alertness; the task definitely requires attention to what can safely be disregarded in the bridge design task – and Newtonian physics cannot explain this.

In economics, the ‘complementarity’ view could be summarized as saying that the accomplishments in the evolutionary beachhead are valid in the sense that relativity theory is valid for phenomena at relativistic speeds – setting aside the question of whether the new paradigm is a special case of the old, or vice versa, or neither. (In the old paradigm view, it is the new paradigm that is the special case, as it seeks to deal with a domain of exotic phenomena that were not much encountered (previously).) As is further discussed below, I believe that in the vast body of conclusions derived by the methods of neoclassical economics there are many pragmatically valid propositions. It is an important task for evolutionary economists to identify such propositions and to explain in evolutionary terms why the neoclassical theory ‘ever seemed to work’ in these cases. When it seems to work, it does so in spite of the systematic underlying error at the level of actor behavior. There are multiple reasons why that might be the case, including the ‘natural selection’ reasons proposed by Friedman, and which I critiqued and explicated long ago (Hodgson, Reference Hodgson1994; Winter Reference Winter1964, Reference Winter1971). Nelson and I analyzed the multiple reasons why the textbook comparative statics of a demand shift seems to work (Nelson and Winter, Reference Nelson and Winter1982: chapter 7).

In the great majority of the cases, however, the real problem is not that there are actual causal mechanisms creating the appearance of underlying optimization, but that the bar for ‘seems to work’ is set extremely low. Very generally, evidence indicating that economic motivation matters is touted by neoclassical economists as if it were supportive of neoclassical optimization-based predictions, with zero attention to the problem of distinguishing optimization from the many other ways in which economic motivation can be manifested (Winter, Reference Winter2014). Those other ways include, for example, the susceptibility to greed-induced blindness that underpins the widespread propensity to be victimized by confidence schemes. Another form of motivated but non-optimizing behavior, much more central to the paradigm dispute, is the propensity to act in a habitual manner that seems to be well adapted to the specific environment in which it occurs. In neoclassical practice, the ‘optimization’ designation is freely conferred on any apparently effective habit that is shaped by any combination of repetition, learning, deliberation or design. Thus, for neoclassicals, the theory of optimizing behavior seems to work whenever the behavior theorized seems to work.Footnote ²⁵

Evolutionary economists note that such an approach diverts attention from the actual origins of behavioral patterns, and from the important possibility that the likely responsiveness and robustness of those patterns in future environments may differ from what is seen in the observed situation. More fundamentally, the approach shrinks from offering credible support for the optimization claim. Credible support would go beyond evidence on motivation, internal consistency, and apparent worldly success, and look into the possibility that the availability of more effective behaviors had been overlooked in the alleged optimization.

Both of the young paradigms are more concerned with the ‘microfoundations’ of their theoretical propositions than are their old-paradigm counterparts. The young ones draw upon, feature, and build upon micro-level observations that were neither pursued in the old paradigm nor fully admitted when they surfaced by other mechanisms. In the economics case, the neglected micro-level evidence refers to the actual decision processes of individuals and organizations. In the evolutionary view, respect for the scientific truth about that has to underpin any claim for attention at higher levels of analysis. By contrast, the neoclassical camp largely admits to the falsity of its theoretical representations in that domain, but believes that reliance on these fictive devices does not present a serious hazard for deduced conclusions relating to higher levels of analysis. More important, this supposition is all that exists to justify the dedication of high-quality intellectual resources to the rigorous development of elaborate, logically persuasive accounts of what things would be like if the conventionally agreed premises of neoclassical economics were true. But they are not true, and this allocation of effort is questionable not only because of its cost, but because it contributes to unrealistic policy instincts across a broad front – often providing ill-founded support for vague arguments about the virtues of laissez-faire capitalism.Footnote ²⁶

Strikingly, each of the two younger paradigms differs from the corresponding old one by offering very different accounts of the role of time. Indeed, the word ‘role’ fails to suggest the extent of the reconceptualization involved in the paradigm shifts.Footnote ²⁷ One might say that both of the older paradigms feature quite a tame version of time, a kind that does not evoke probing questions about the causal relations among other entities in the theory. In the physics case, Einstein probed deeply into the question of how fast a physical cause can travel. His official conclusion was expressed with remarkable though deceptive simplicity in his 1905 theory of special relativity. On the economics side, critics of neoclassical optimization assumptions have long made reference to the fact that the computational costs of true optimization could not legitimately be dismissed – to ignore them is to ignore one of the world's most fundamental examples of the resource scarcity with which economics is purportedly concerned (Radner, Reference Radner, Smelser and Baites2001; Winter, Reference Winter, Day and Groves1975). Even in the computer age, computation takes time – as is underscored by the recent attention to high frequency trading in the stock exchanges. Also vividly illustrated there is the point that it is the software and the institutional context that determine where the benefits of cheap computation accrue (Lewis, Reference Lewis2014).

While the evolutionary and neoclassical paradigms divide on many specific issues about economic behavior, systems, and policy, what is centrally at stake in the paradigm clash is the matter of scientific truth. The truths at issue relate particularly to the basis of individual actor behavior, but certainly include several others. In my view, neoclassical economic theory is ‘wrong’ in much the sense that Newton was ‘wrong’, per Kuhn. That does not mean that Newtonian engineers build failing bridges, or will do so in the future – but they might build pitifully poor supercomputers. If you seek to understand how and why the old paradigm can lead you astray, you cannot begin by accepting its flawed premises. You need the wider view that the new paradigm provides.Footnote ²⁸

7. Breakout?

There are numerous realistic reasons why the evolutionary economists of the future are likely to make their contributions within or near the beachhead topics – certainly there is no shortage of opportunity there. I will not bother to enumerate specific reasons, but invite the reader to imagine such specifics within broad categories like ‘the sociology of knowledge’, ‘tenure and promotion’, ‘teaching assignments’, and ‘resource constraints’. Alternatively, we could imagine convening a council of distinguished evolutionary economists and asking that council for strategic guidance on how the field might be made to prosper in the future. The question is how ambitious the council's recommendations would turn out to be. Would they favor a vigorous pursuit of paradigm change in the economics discipline as a whole? Or would they, being sensible economists and realists above all, give great weight to some of the same considerations just suggested? They might also take note of the potential difficulty of getting to critical mass in a wider range of research areas, given that the aggregate mass of evolutionary influence is now so limited.

There is indeed a sensible case for caution and limited ambition, and the corresponding prediction has a lot to recommend it. As I argued in the beachhead tour above, the need and promise of evolutionary thinking in the topic areas of the beachhead are particularly apparent; it is not surprising that the beachhead is where it is. Nevertheless, there is doubtless an element of historical accident involved there as well, and the persistence of the pattern may partly reflect inertia – or the influence of prevailing routines – as much as the inner logic of the subject matter. Many evolutionary economists have ventured the sly observation that the persistence of neoclassicism can be understood as an example of the persistence of routines. I now make the sly observation that this point can be turned around: perhaps the stability of the topic areas addressed by evolutionary economists has now become a part of their routines. And perhaps they ought to think about this, and consider whether realistic opportunities for significant innovation are being ignored.

Also, contra the argument invoking the ‘critical mass’ problem against an expansion of the beachhead, there is the corresponding argument about economics as a whole: Evolutionary economics has certainly not achieved critical mass there, and is much weaker even in its areas of relative strength than it would be if the movement were broader. The probability of survival, let alone success, might actually be enhanced by a more broadly expansionary policy.

The cautious approach would leave unexploited the intellectual assets mentioned in the ‘core strengths’ discussion. First and foremost, there is the importance of offering a system-level interpretation of the economy and hence of the discipline's explanatory problem. Economics is widely perceived as adrift today, and the problem can reasonably be ascribed to the fact that the exploration of the great unifying theme promoted by Adam Smith was brought to a kind of culmination in the middle of the last century, in competitive general equilibrium theory (Debreu, Reference Debreu1959). That was a remarkable achievement, which continues to have a profound influence on how most economists (even heterodox ones) view the system. But there has been no successor in the role of providing a system view; instead the mainstream discipline is largely defined today by a broad portfolio of research techniques. As I suggested above, a compelling system view is what the economics discipline is best qualified to contribute, by virtue of its intellectual heritage. Because there is no compelling system view on offer, there is politically tainted scientific cacophony – or as Kuhn would say, a re-emergence of pre-paradigmatic discussion. Evolutionary theory offers the quintessential system view. In the one direction are the credentials it shares with the natural science that reach all the way back to the Big Bang, and in the other it is focally concerned with the processes of change that will create the economy of the future. We know where we, and the economy, stand in space-time, and think about the economy and its possible futures in that perspective. As Veblen suggested long ago, an evolutionary approach offers economics the opportunity to join the sciences. To me, that seems so powerful an appeal that it is hard to believe that it will forever be ignored.

Complementing the ‘system view’ aspect of evolutionary economics is its commitment to open borders – to free trade in the domain of social science theory. Nelson and I commented long ago on the problem of the ‘intellectual autarky of economics’ (Nelson and Winter, Reference Nelson and Winter1982: 405–407). The problem is much worse today because other fields of social science inquiry have been rushing forward in the interim, a situation that economics has largely ignored.Footnote ²⁹ That is particularly true in the case of psychology. Although the broad fund of experimental work in social psychology is not organized around the task of proving neoclassical economists wrong about individual behavior, it achieves that nevertheless.Footnote ³⁰ The commitment to optimization/rational decision forms the sturdy core of the neoclassical barriers to free trade in ideas. Making that commitment (as a matter of descriptive theory) entails affirming that the actors generally get the answers to their problems right. Qualifying that affirmation seriously creates complexities that mainstream reviewers probably will not like. But nobody else in the social sciences, apart from those in the rational choice colonies, actually believes that – and one wonders how many neoclassical economists would likely affirm it as a valid generalization about the outcomes of significant meetings in their own academic departments.

This lack of a system view is a major problem, not just for the academic discipline, but for society. A coherent view of the economy as a system is an indispensable ingredient of sound economic policy, but the main supplier of significant system views for the economics discipline suffers from a disabling addiction. An intervention is called for.

Evolutionary economists could find much useful work outside the usual beachhead. I conclude this essay by offering a single, but very important example in this large category. Generally, it is essential to see business practices in their institutional context, and to recognize the power of that context to determine the shape and limits of plausible action. The specific case in point is the US mortgage market in the years that preceded the Financial Crisis of 2008. There were at least three major differences between the institutional context prevailing then and the context that prevailed earlier, in a period in which no such dramatic problems emerged: (1) it used to be the case that mortgage originators participated significantly in the risk that the mortgage would not be repaid, but the securitization of mortgages largely eliminated that risk; (2) it used to be the case that investment banks in the US were partnerships, but they were converted to public corporations; this change eliminated much of the personal financial risks of the management, leaving them primarily with a stake in the upside of risky bets; (3) it used to be the case that the US rating agencies, which are governmentally empowered to define the limits of safe investment for other actors, were remunerated by the potential buyers of risky securities – but that changed, so they were remunerated by the sellers instead. In the new context, a variety of new and risky business practices flourished – temporarily. If any one of these three institutional changes had not prevailed, the crisis and its enormous social costs might have been avoided. (Of course, it would be hard to prove that.) The theoretical point, however, is that all such changes in institutions and business practices are beneath the threshold of awareness of neoclassical theorists (see Winter (Reference Winter2010) and Jacobides (forthcoming) for further discussion). They still constitute naturally inviting territory for evolutionary and institutional economists – and still are important for future policy, though major social costs have already accrued from their neglect.