2. Patchwork ‘evolutionary economics’

Research conducted over the past decades under the label of ‘evolutionary economics’ has undoubtedly improved the understanding of the ongoing transformations of the economy. However, a coherent theory of economic evolution, able to integrate the diverse results, has not yet emerged. Evolutionary economics is rather a patchwork of different approaches. The presently most influential among them is the Neo-Schumpeterian approach going back to the seminal work by Nelson and Winter (Reference Nelson and Winter1982).Footnote ¹ It revives Schumpeter's ideas on the restless, innovative nature of capitalism. Accordingly, focus resides mainly on innovation competition, R&D, technical progress, and the growth of firms, industries, and economies. The corresponding works do not all embrace evolutionary theorizing. Indeed, a good part of the success of the Neo-Schumpeterian school is due to its close affinity to the agenda of contemporary, non-evolutionary innovation research.Footnote ²

Nelson and Winter used the attribute ‘evolutionary’ economics to point to the crucial role, which they assign to a ‘loose analogy’ between natural selection theory in biology and their theory of economic adaptation (Nelson and Winter, Reference Nelson and Winter2002). Their aim was to construct an alternative to the dominant optimization-cum-equilibrium approach in economics. In that approach, the transformation of the economy is interpreted as a sequence of simultaneous, optimal adaptations of economic agents to a state of general equilibrium. Adaptations are necessary because of exogenous, i.e., unexplained, shocks that disrupt the equilibrium state over and again. Drawing on Simon's theory of bounded rationality (e.g., Simon, Reference Simon, McGuire and Radner1972), Nelson and Winter rejected the idea that individual adaptations are always optimal. Particularly at the organizational level, the plasticity of behavior is not that perfect. Instead, the prevailing bounded rationality results in routinized and, hence, inert behavior (Cyert and March, Reference Cyert and March1963; March and Simon, Reference March and Simon1958). If routine behavior at the individual level is inert but adaptations are necessary, selection processes will effectuate them by driving out the less well-adapted among the behavioral routines. The level of adaptation is thus shifted to the ‘population’ of routines (defined for a firm or an industry), and the adaptation process is modeled analogously to how natural selection winnows out less fit genetically coded traits in the population of a species. In this way, the canonical metaphor of an auctioneer who coordinates individual adjustment efforts is replaced by the metaphor of natural selection forces effecting changes in the composition of firm routines (see Joosten, Reference Joosten2006).

Furthermore, Nelson and Winter criticized the mismatch between the incessant transformation of the capitalist economy that one can observe and the insistence of canonical economics on the significance of the general equilibrium of the economy, disturbed only by exogenous shocks.Footnote ³ Nelson and Winter followed Schumpeter (Reference Schumpeter1934)[1912] here who argued that innovative activities keep the economy in a state of unrest that needs to, and can, be explained by economic theory rather than being treated as exogenous shocks. In line with their loose evolutionary analogy, but in contrast to Schumpeter (Reference Schumpeter2002), Nelson and Winter interpret innovations as variations of routines analogously to genetic variation. Like mutations in the gene code of organisms can change the distribution of traits in the population of a species, so can innovations affect the composition of firm routines in an industry. For instance, a variation in an organizational routine may lower a firm's unit production costs. If, to stick to the example, the resulting lower average unit costs in the industry lead to a decreasing price level, non-innovating firms with the highest unit costs will sooner or later make losses and be driven out of the market (Metcalfe, Reference Metcalfe2008).

As the analogy constructions show, a direct application of the Darwinian theory to the economic domain is not intended (Nelson, Reference Nelson2006).Footnote ⁴ This is in stark contrast to the historical precedent of ‘evolutionary economics’ proclaimed by Veblen (Reference Veblen1898) with his naturalistic approach. Veblen indeed advocated an extension of the Darwinian theory of evolution to economic phenomena. He was no less critical of the mechanical, rational adaptation hypotheses underlying the economic theory of his time. Like Nelson and Winter after him, he also called for a more ‘realistic’ view of economic behavior. But his critique of the marginal calculus did not induce him to put a selection algorithm in its place. Instead, he abandoned calculus completely and speculated about the role of instinct (genetic influences) and habit (the influence of cultural learning) on economic institutions. Accordingly, he was not much, if at all, interested in technical change and industrial dynamics, but rather in how economic, political, and cultural habits and institutions shape the development of the economy (Veblen, Reference Veblen1899, Reference Veblen1914).

As mentioned, Veblen's successors soon abandoned his naturalistic, evolutionary approach. Apparently, they did not find it useful for explaining the various political, industrial, and macroeconomic topics they were interested in. However, some of Veblen's naturalistic intuitions re-emerged later in works of diverse authors such as Georgescu-Roegen (Reference Georgescu-Roegen1971), Hayek (Reference Hayek1979; Epilogue), North (Reference North1997), North et al. (Reference North, Wallis and Weingast2009), and Ostrom (Reference Ostrom2014). The subjects, hypotheses, and political messages of these contributions differ, and yet all share three basic assumptions that separate them from canonical economics. First, they recognize that the human economy and its institutions are embedded in, and contingent on, nature, not least of which human nature (i.e., innate human proclivities). Second, they portray the institutions prevailing at a given time as the result of an ongoing evolutionary process. Finally, they emphasize the necessity of explaining in terms of a historical analysis why some institutional forms prevail while others decline. These commonalities expound why these contributions can be said to represent an evolutionary branch of a new, naturalistic institutionalism. Their common evolutionary perspective notwithstanding, Neo-institutionalists (with the exception, perhaps, of North and Ostrom) and the Neo-Schumpeterians hardly take notice of each other, not to speak of the missed chance of a cross-fertilization.

A similar mutual silence characterizes the relationships between evolutionary economics and the new field of evolutionary game theory.Footnote ⁵ In contrast to rational game theory, players in evolutionary games are assumed to represent, rather than to choose, mixed or pure strategies as part of their behavioral repertoire. This assumption is due to the explanatory requirements of evolutionary biology, particularly sociobiology (Maynard Smith, Reference Maynard Smith1982; Trivers, Reference Trivers1971), from which evolutionary game theory originally emerged. In that literature, the behavioral repertoire of an organism is understood to be innate and the competing strategies subject to natural selection. Economists resort to that interpretation to explain apparently ‘irrational’ features of human economic behavior such as altruism, cooperativeness, fairness, moral aggression, etc. in terms of a possible reproductive advantage (Binmore, Reference Binmore1998; Bowles and Gintis, Reference Bowles and Gintis2011). The association with the extended Darwinian theory of evolution is straightforward (Henrich et al., Reference Henrich, Ensminger, McElreath, Barr, Barrett and Bolyanatz2010).

Beyond these fundamental features of human sociality, it is not clear, though, how an interpretation of strategies as genetically fixed traits can contribute to explanations in the domain of economics. For that reason, attempts have been made to re-interpret the strategies as part of a behavioral repertoire that is not innate, but emerges from reinforcement learning (see Brenner, Reference Brenner1998). The time scale on which behavior evolves would in that case differ dramatically from the case of innate behavior. Yet, the adaptation algorithm – usually some form of replicator dynamics – and the equilibrium points of the games remain unchanged and different from those of rational game theory (see, e.g., Friedman, Reference Friedman1998). In non-strategic settings (i.e., in ‘playing against nature’ as game-theorists put it), reinforcement and conditioning learning are significant forms of adapting behavior with little or no cognitive participation (Herrnstein and Prelec, Reference Herrnstein and Prelec1991). Yet in strategic settings, their relevance is less clear and the explanatory value of evolutionary game theory, therefore still uncertain.

As these newer developments have largely been ignored so far by the Neo-Schumpeterians, there was no need to change their attitude toward individual economic behavior. Perhaps owing to the dominance of the population perspective on selection and variation processes (Metcalfe, Reference Metcalfe2008), individual behavior has never been high on the agenda of evolutionary economics. The only aspect where a reference to individual behavior figures prominently is the bounded rationality hypothesis going back to Simon (Reference Simon, McGuire and Radner1972). However, the rejection of the fiction of full rationality does not yet imply an elaborate theory of human economic behavior and its underlying motivations. Such a theory can draw on insights from behavioral economics and evolutionary psychology.Footnote ⁶ For understanding the role of the evolved parts of the human behavioral repertoire, human behavioral ecology and sociobiology, evolutionary anthropology, and evolutionary psychology are also relevant (Brown and Richerson, Reference Brown and Richerson2014). Thoroughly anchored in the extended Darwinian theory of evolution, all these human science disciplines point out the direction in which ‘evolutionary’ economics may move when developing its behavioral foundations further.

3. Modalities of an evolutionary analysis

With a variety of hypotheses that is only matched by the variety of topics addressed, the diverse approaches to evolutionary theorizing in economics do not seem to have much, if anything, in common. A behavioral foundation that could provide a common ground by tapping the potential of the extended Darwinian theory of evolution has much been neglected so far. A conceptual integration on the basis of the abstract notions of variation, selection, and replication, as proposed by Hodgson and Knudsen (Reference Hodgson and Knudsen2010), is subject to controversy (Cordes, Reference Cordes2006; Levit et al., Reference Levit, Hossfeld and Witt2011; Nelson, Reference Nelson2006; Vromen, Reference Vromen2007). Instead of discussing possible integration strategies at the conceptual level, the focus will therefore be shifted here to the methodological level. As will turn out, hitherto unnoticed common ground can be identified underlying the diversity of theoretical approaches. It rests in the methodology, more specifically the modalities of causal explanations, germane to the Darwinian paradigm. The modalities of evolutionary explanations are independent of the particular (controversial) hypotheses used for explaining economic phenomena and a characteristic feature by which evolutionary and non-evolutionary research in economics can be distinguished.

The modalities of evolutionary explanations have been outlined exemplarily for evolutionary ethology by Tinbergen (Reference Tinbergen1963), but have been generally accepted in biology. Drawing on distinctions that go back to Aristotle, Tinbergen singled out four types of causal explanations as follows. Suppose the object of research is a particular trait of an organism, the capacity to process acoustic signals, say. Focus may then be on

1. (i) proximate explanations of how the trait works, i.e., of how acoustic signals are processed;

2. (ii) ultimate explanations of the function of the trait and its adaptive value, i.e., of why the capacity to process acoustic signals implies an advantage in terms of survival and reproductive fitness that is sufficient to fix the trait in the gene pool of a species;

3. (iii) explanations of the descent of the trait, i.e., of how the capacity to process acoustic signals has historically evolved in one or several species’ phylogeny;

4. (iv) explanations of how the trait develops over an organism's life cycle, i.e., of how the capacity to hear unfolds during the ontogeny of the organism's body.

The broader relevance of Tinbergen's scheme for evolutionary research based on the extended Darwinian theory of evolution has recently been emphasized by Wilson et al. (Reference Wilson, Hayes, Biglan and Embry2013). It is also pertinent to economic research.Footnote ⁷ With slight modifications, the scheme will therefore be used here to characterize the distinct common ground in evolutionary economics. By the same token, an answer can be derived regarding two further questions. The first concerns the main difference between evolutionary economics and canonical economics. The second question is whether the evolutionary paradigm should be considered replacing, complementing, or including the canonical approach to economics.

To begin with the latter two questions, it is worthwhile first to note that both evolutionary and canonical economics have a common interest in proximate type (i) explanations. As a matter of fact, this type of explanation is center stage in canonical economics when analyzing all sort of economic effects and the mechanisms that cause them. Familiar examples abound: How is the daily market clearing exchange rate between two currencies fixed in the money market? How does an increase in the tax rate on cigarettes affect the quantity of tobacco consumed per unit of time? By what mechanisms does a central bank influence the price level in an economy? How does the profit rate influence the growth rate of a closed economy? Whether concerning the workings of individual response patterns, of markets and other exchange mechanisms, of organizations and institutions, or of technological constraints, to answer questions like these, a proximate explanation is required.

The specific explanation given can, of course, be based on very different hypotheses, and the hypotheses themselves can be satisfactory or unsatisfactory. The fact that evolutionary economics and canonical economics share an interest in proximate type (i) explanations does not mean that they need rely on the same explanatory hypotheses. When economics is envisioned to be the science of optimization and coordination, as in the canonical approach, the rational (or optimal) choice hypothesis supplies the foundation for all ensuing explanations. This hypothesis is, however, rejected almost unisono in evolutionary economics (and not only evolutionary economics) as empirically inadequate idealization. Laboratory experiments designed to test the rationality hypothesis demonstrate that the hypothesized behavior is often a far cry from the actually observed behavior (Kahneman, Reference Kahneman2003).Footnote ⁸ In evolutionary economics, it is usually argued that instead of full rationality, the economic agents exhibit only a more bounded version of rationality. What precisely the behavior is that bounded rationality induces in a particular setting, i.e., how the states of the environment map into individual choices, is rarely expounded, though.

The rational choice explanations of canonical economics thus represent a subclass of proximate explanations. The distinguishing characteristic of this subclass is the generic use of the optimization hypothesis. If economics were to only provide proximate explanations (and if the optimization hypothesis were less problematic), canonical economics could claim to be sufficient. However, proximate explanations are not enough. They take as given the mechanisms whose workings or effects they are supposed to explain. Whenever the current properties of these mechanisms depend on the mechanisms’ pre-history, i.e., on how they have evolved, it is important to go beyond proximate explanations. Different types of causal explanations are required then to answer the questions of why and how the mechanisms have evolved in the first place. (Correspondingly, such explanations are also necessary to explain how it is possible that, at different times and different places, different mechanisms have evolved that serve similar purposes.) It is here that the explanatory interest and, hence, the agenda of evolutionary economics goes beyond that of canonical economics.

In Tinbergen's scheme, why something has evolved requires an ultimate explanation (type ii). In contrast, how something has evolved requires an explanation of its history (type iii). These two types of explanation are unique to evolutionary research and shared by all attempts to apply encompassing theories of evolution.Footnote ⁹ In biology, both types of explanations draw on the Darwinian theory of evolution, but highlight different aspects of descent with variation. Their power and broad significance in the biological domain is due to the dominant effect that the shaping force of natural selection has on the evolution of the species in the long run. This is no longer the case in the domains where the Darwinian theory evolution is extended to account for cultural influences. Natural selection plays less of a role in those domains precisely because of the increasing influences of the evolved human faculties of observational learning, inference, inventiveness, and intentionality. The latter imply adaptation processes of their own, which are shaped by human conditioning learning and insight.Footnote ¹⁰

Often occurring within just one generation, these culturally contingent adaptation processes operate on much shorter time scales than natural selection. They also produce more variety as the particular culture emerges from the specific social interactions by which humans are brought up, act, and transmit behavior patterns. In fact, the weaker the selection pressure at the species level, the more can culturally acquired preferences and cognitively constructed intentions lead to behavior that has no biological adaptive value, i.e., does not enhance individual fitness in terms of survival and reproduction or number of surviving offspring. The fact that there are two influences on individual behavior in modern economies – the effects of the dual, biological and cultural, human inheritance as Richerson and Boyd (Reference Richerson and Boyd2005) put it – makes ultimate type (ii) explanations of economic phenomena and processes more complicated. Unlike in biology, observable phenomena are governed not only by the criterion of reproductive fitness but also by various other, culturally contingent criteria as well, particularly in the short run. No one who has tried to explain the ongoing transformation of modern economies and their driving forces will be surprised by this finding.

Biology and economics also differ with respect to type (iii) explanations. Again, the more the influence of the additional, culturally conditioned, adaptation criteria increases, the more complex the explanations get. Where the criterion of reproductive fitness plays little part (except via the inherited dispositions of human behavior), the learned preferences and more or less idiosyncratic intentions and ideas of agents take over as criteria of adaption. In the limiting case, type (iii) explanations may then resemble what Menger (Reference Menger1985)[1883] long ago had in mind with his ‘causal-genetic’ explanations, explicitly accounting for a multiplicity of causal factors not including reproductive fitness. Economic change is then still explicable in terms of an adaptation process. But the explanatory hypotheses now refer to rather specific criteria such as culturally contingent goals or even idiosyncratic motivations to expound how certain adaptations occur.

Despite these reservations, type (ii) and type (iii) explanations add substance and scope to economic theorizing. The reconstruction of how presently observable mechanisms have evolved often uncovers significant, contingent features of these mechanisms by which their functioning (or reasons for their mismatch) can better be understood. This can be shown for each of the examples just mentioned in relation to type (i) explanations. For instance, daily fixing of the exchange rate between two currencies in the money market has evolved from quotation rules originally well adapted by trial and error to the rather slow process of market-maker-mediated trade. When computerized electronic currency trade, still following the logic of market-maker-mediated trading rules, was introduced, information processing and market transparency were first much improved. However, with the new information technology, opportunities for exploiting the overcome logic emerged, too. High-frequency currency traders now employ computer programs that can use millisecond time differentials between orders and actual transactions to change electronic quotes and manipulate the terms of transactions to their own advantage. As a result, the currency trade mechanism is in these days less transparent. Since the majority of market participants still follows the overcome arbitrage rules, high-frequency trade can also develop feed forward effects that can produce erratic outlayers in the exchange rate time series.

Turning now to the question of a common ground beneath the present heterogeneity of theoretical approaches to evolutionary economics, the preceding reflections on the modalities of explanation suggest a succinct answer. Irrespective of what topics are considered and how evolution is conceptualized, all theoretical approaches share one condition. They can unfold their potential only by means of type (ii) and/or type (iii) explanations. Even if not explicitly recognized, they therefore have in common that they pose questions and develop hypotheses required by the modalities of these two types of explanations. In doing so, work in evolutionary economics goes beyond canonical economics typically posing question relating to proximate type (i) explanations. By distinguishing between type (ii) explanations (why something has evolved) and type (iii) explanations (how something has evolved), evolutionary theorizing can be structured in similar ways, and hypotheses can be assessed with respect to whether they are adequate for ‘why’ or ‘how’ questions, respectively. Some confusion and unnecessary controversy at the theoretical level may on this basis be overcome.

4. The Case of analyzing institutional evolution

The analysis of institutions has a century-long evolutionary tradition in economics from Veblen (Reference Veblen1899) to North (Reference North1997) and Ostrom (Reference Ostrom2014). It is therefore a straight forward case for illustrating what can be gained by focusing on the modalities of explanation. Before, however, a problem needs to be addressed that confronts all research on institutions, namely the enormous variety and heterogeneity of phenomena associated with the term ‘institutions’. Many efforts have been undertaken to come to grips with it by identifying some generic features that are shared by all the different institutions and on which a general theory could be built.

With the rise of non-cooperative game theory and its method of abstraction, a solution to the problem has gained prominence that interprets institutions as an outcome of strategic interactions between individuals. Accordingly, the generic features of institutions are identified with the characteristics of games supposed to reflect the particular interaction context.Footnote ¹¹ The analysis of institutions then becomes a matter of applying general game theory, and specific insights are derived by determining the properties of the equilibria of the relevant games. For many institutions, particularly the formal ones (see Schubert and von Wangenheim, Reference Schubert and von Wangenheim2006), it will not be possible to capture the underlying strategic interactions by games simple enough for their equilibria to be determined. In that case, the method of abstraction runs idle and can at best result in another, perhaps more abstract, description of institutions in game-theoretic terms. However, where the strategic interactions underlying an institution can be represented by a sufficiently simple game, interesting insights into the logic and the working of that institution can be obtained.

A simple, frequently used example is that of coordination games epitomizing the logic of a particular institution, namely conventions. It can serve here too for the intended illustration of the role of the modalities of explanation. A convention can be represented by one of at least two co-existing, locally stable equilibrium points in a many-person coordination game (see, e.g., Schlicht, Reference Schlicht1998; Young, Reference Young1998). Its working is then explained (in the sense of type (i) explanations) by the condition that, in equilibrium, the involved players are coordinated on a compatible interaction behavior. This is in their common interest. (At least in the pure form of coordination games, there is no conflict between the players regarding their preferences over the alternative equilibrium points of the game.) For that reason, any of the possible conventions corresponding to one of the alternative, locally stable equilibria is self-enforcing once it has spontaneously emerged.

The details of the corresponding game-theoretic models of convention-institutions need not interest us here. The point to be made is that all of them explain in the form of generic, proximate type (i) explanation how the coordination function of this institution works. Almost all game-theoretic models of other institutions also represent this type of explanations.Footnote ¹² But, valuable as the insights which they offer are, in an evolutionary analysis, it is required to explain in addition why and how a particular convention has emerged. This is rarely done. Indeed, it is usually stated that the process of equilibrium selection (i.e., the process of emergence of a particular convention that would require a type (iii) explanation) is ‘underdetermined’.Footnote ¹³

To answer the why question one can try to construct a type (ii) explanation based on a hypothesis about some adaptation advantage, which the involved players gain by coordinating on the convention. The context-dependent task then is to find the relevant criterion of adaptation in the first place. Is it reproductive fitness? (In biology, the ultimate criterion is the beneficiaries’ reproductive fitness in a given environment.) Is it the interest of political and economic agents in power? Or is it individual (phenotypic) utility? Game-theoretic models usually remain vague on this point. Their criterion of adaptation is expressed by the payoffs associated with the strategies. When, as usually, the payoff values are interpreted in utility terms, game theory ends up with the finding that the advantage of convention-institutions is, well, efficiency – whatever adaptation criterion the placeholder ‘utility’ stands for.

Such an (unspecific) efficiency explanation is often considered congenial with economic theorizing. However, lacking a reason for why efficiency has adaptive value it is not sufficient for a type (ii) explanation in an evolutionary analysis. For the small bands that characterized early hominid groups it can be hypothesized, for instance, that conventions facilitated the coordination of group members on joint hunting tactics, on a shared mode of signaling the approach of enemies, or on ritualized inter-personal conflict resolution. This most likely meant an advantage in terms of survival chances and reproductive fitness.Footnote ¹⁴ Due to the tremendous technological change and concomitant growth of society since ancestral times, large, anonymous groups have formed, and entirely new conventions have emerged. The adaption advantage of many of these conventions seems to have shifted from reproductive fitness to specific (phenotypic) sources of utility.

Consider a particularly simple example of modern conventions: traffic rules such as driving either on the left- or the right-hand side of the road (see, e.g., Hodgson and Knudsen, Reference Hodgson and Knudsen2004). In so far as this convention prevents accidents, it could be argued to still mean an adaptation advantage in terms of reproductive fitness. Yet, the willingness to adopt, and abide by, the conventional driving on a specific side of the road is not an innate behavior trait and, hence, not differentially genetically reproduced. Nor is it a culturally fostered trait that spontaneously emerges from group-based communication and observational and conditioning learning. Instead, the implementation and enforcement of the traffic convention is the result of orders and sanctions invoked by authorities who, by the preceding evolution of other institutions, have the power to do so.

Accordingly, the criterion of adaption could be the interests pursued by these authorities. They may, perhaps, have a vicarious interest in saving traffic participants from suffering the disutility of injuries and pains (rather than raising their reproductive success), as well as an interest in an efficient traffic flow that saves time and resources. However this may be, the example shows that the advantage which the coordination on a particular convention entails – and, consequently, the explanation for why the convention evolved – is likely to depend on specific conditions of time and place. This means that type (ii) explanations cannot, like in biology, rely on just the one powerful, general principle of reproductive fitness. They must unfortunately be expected to be more complex than their counterparts in biology.

Type (iii) explanations are significant in the evolutionary analysis when it comes to understanding how evolution actually takes place. For the present case, this requires reconstructing how functions of institutions change over time and, accordingly, how that change raises or lowers the advantage institutions entail in terms of the adaptation criteria prevailing in the particular context. In biology, type (iii) explanations are particularly challenging in situations in which a genetically coded trait is not functional, but nonetheless survives in the population. This is closely related to the problem of evolutionary mismatch, i.e., a situation in which something initially functional when it first emerged, becomes dysfunctional later and nonetheless survives. Such situations are relevant also for economic evolution. But, due to the potential multiplicity of the criteria of adaptation in an economic context, yet another problem can emerge here. The problem is related to the fact that in economic evolution and cultural evolution more generally multiple, potentially incompatible, criteria of adaptation can be effective. It may then be difficult to assess what the adaptation value of something that has evolved, e.g., a convention-institution, is. Consequently, deciding whether it is still functional or already dysfunctional may become a problem. It can be inferred, thus, that type (iii) explanations too are substantially more complex in the domain of economics (and other cultural domains). Type (iii) explanations of how institutions evolve are a case in point.

With their specific demands on historical information, type (iii) explanations represent a research program that transcends the abstract game-theoretic analysis of institutions having its strength in the context of type (i) explanations. The broader research program does not exclude type (i) explanations, but cannot be reduced to them. (It has much in common with Veblen's institutionalism and the evolutionary branch of New Institutionalism. Both develop type (iii) explanations without explicitly acknowledging this.)

The level of abstraction that may be adequate for proximate type (i) explanations of institutions is simply no longer adequate for explanations of the historical evolution of conventions. A type (iii) explanation of how conventions evolve can serve as example. If a game-theoretic analysis was to be used for that purpose, it would have to keep track of innovations in the rules of the underlying coordination game, of changing constellation of motives and players. The capacity of game theory to handle changes of this complexity would quickly be exceeded. Moreover, it would be necessary to step outside game-theoretic reasoning in order to explain what criteria of adaptation govern the implementation, enforcement, and change of conventions. A more or less plausible ad hoc specification of the payoffs and the order relation between them would not suffice.

From the point of view of how economic phenomena evolve historically, it is important that type (iii) account for the fact that there are usually causal factors in the pre-history. They can have a strong impact on the later path of evolution. Among them are genetic pre-adaptations that do not only constrain the behavioral repertoire of human agents (their strategy set) but also the way in which they perceive the consequences of their action (their payoffs). Such pre-adaptations can also influence the agents’ capacity to coordinate more or less spontaneously and, hence, the way in which conventions emerge. An example is the unique human capacity of seeing the self in others and perceiving them as intentional agents (Tomasello, Reference Tomasello1999).Footnote ¹⁵ Likewise, there are cultural pre-adaptations that can develop an effect for the evolution of conventions. For example, under certain conditions, previous learning to behave in a manner that supports coordination in one context may strongly facilitate the spontaneous emergence of conventions in entirely different contexts.

Obviously, the analysis of institutions holds different meaning for evolutionary economics and canonical economics. The latter basically confines itself to type (i) explanations of the abstract generic features of institutions developed from an elaborate game-theoretic formalism. Evolutionary economics, in contrast, acknowledges the truism that human institutions did evolve and continue to do so. For that reason, the theoretical understanding of institutions cannot be complete without searching answers for the why and how of institutional evolution, i.e., without integrating type (ii) and (iii) explanations. A different question and matter of controversy between canonical and evolutionary economics remains: are the proximate type (i) explanations sufficient for practical purposes? Or can it be shown that policy advice and what is called ‘mechanism design’ in canonical economics are indeed incomplete or even misleading without paying attention to how existing institutions have evolved in the first place? If so, a broad adoption of the modalities of evolutionary explanations is likely to enrich research into institutional economics in the future in a similarly dramatic way as these modalities have improved biological research.