Beyond the Hype

I. Introduction

The idea of bringing evolutionary theory into economics is becoming increasingly popular (see, e.g., Hodgson and Knudsen 2010; Witt 2008; Nelson and Winter 1982). From consumer choice to the behavior of entire economies, there is hardly an area of economics that has not been subjected to some kind of evolutionary analysis. However, this kind of evolutionary approach to economics also is very controversial—in particular, critics from both inside and outside the discipline are concerned that, for various theoretical reasons, the approach has merely confused the issues being discussed, rather than enlightened them (see, e.g., Rosenberg 1994; Gould 1996; see also Vromen 2001). In this paper, I seek to assess this dispute in more detail.

In particular, I argue that, in principle, there are two ways in which evolutionary theory might be relevant to economics—evidentially and heuristically—and that neither of them actually shows much promise at the moment. The evidential approach suffers from the fact that we are lacking crucial pieces of economic or evolutionary biological data, and the heuristic approach is made problematic by the fact that it has not yet suggested a major novel idea for economists to investigate. In doing this, though, I also try to make clear that while the critics of evolutionary economics are right in their negative assessment of the current value of the approach, their reasoning is often flawed. Specifically, I try to show that the main issues with evolutionary economics are empirical—not theoretical—in nature and that the future prospects of evolutionary economics are far less dim than is often supposed by its critics.

The paper is structured as follows. In section II, I sketch the main ideas behind evolutionary economics and delineate the two main ways in which evolutionary theory could be seen to matter to economics. In sections III to V, I assess whether evolutionary theory can be useful in either of these two ways in three key debates in economics. I conclude in section VI.

II. Evolutionary Economics

There are two different ways to understand the term “evolutionary economics,” a broader and a narrower one. ¹ On the broader understanding, “evolutionary economics” simply refers to the application of evolutionary theory to economic issues. For example, much of the recent work on cultural evolution can be applied to various economic questions (after all, economies are parts of human culture), and the same goes for work on cognitive and neural evolution (after all, economics has much to do with thought and decision making) (for more on the former, see, e.g., Boyd and Richerson 2005; Henrich et al. 2004; Skyrms 1996; for more on the latter, see, e.g., Camerer 2007; Glimcher, Dorris, and Bayer 2005). On this understanding of “evolutionary economics,” no further theoretical commitments are made by the program: in particular, what, exactly, the upshot is of this application of evolutionary theory to economics—what hypotheses it yields or supports—is being left open.

Precisely this is different on the narrower understanding of the term, however: there, a specific set of economic doctrines is said to be supported by evolutionary considerations. ² Most important among these doctrines are the following three (see also Vromen 2001; Hodgson 1999):

Since I will say more about these doctrines below, a quick sketch of their content is sufficient here. Doctrine 1 endorses the idea that economic decision making cannot be analyzed well with models that presume agents are optimizing, computationally unbounded, risk-neutral reasoners. Doctrine 2 has it that firms are economic agents in their own right, not merely alternative transactional spaces within which economic agents interact. Finally, doctrine 3 states that economies should be studied with the tools of chaos theory (or the like) and not with the kinds of equilibrium analyses that are commonly employed in economics.

Most of the controversy surrounding evolutionary economics has centered on this narrow interpretation—which is unsurprising, as this interpretation does not just make a commitment to a specific methodology of economics but also to a specific set of economic viewpoints (see also Witt 2008; Vromen 2001; Rosenberg 1994). For this reason, it is also this understanding that will be at the forefront here.

With this in mind, it is important to distinguish two ways in which evolutionary theory can be seen to enter into economics: evidentially and heuristically (see also Schulz 2011b, 2011c). ³ These two ways are not mutually exclusive, in that some set of evolutionary considerations can be both evidentially and heuristically useful to a particular economic hypothesis; however, the two ways can come apart and therefore need to be distinguished from each other. Consider them in turn.

If evolutionary theory is used evidentially, the aim is to use it to provide evidence for the truth of some economic hypothesis: it is meant to support that hypothesis further. ⁴ Of course, this evolutionary evidence need not be taken for the final word on the matter in question: the point is just that these evolutionary considerations confirm the truth of the economic hypothesis—they raise the probability of that hypothesis being true. For this reason, whether or not the evolutionary considerations end up swamped by other considerations, they at least need to be taken into account when assessing the hypothesis in question (see also Schulz 2011c). ⁵

By contrast, if evolutionary theory is used heuristically, the aim is to suggest novel economic hypotheses or ideas that it would be interesting to explore further or which would allow one to better test other, existing economic hypotheses or ideas. ⁶ Note that, on this way of applying evolutionary theory, actually confirming some particular economic hypothesis is not the goal; what is meant to be achieved is merely the discovery of a new way of thinking about the economic world that can then be further explored. For this to be a plausible use of evolutionary theory, it must therefore be shown that (1) this use actually does suggest novel insights: the mere possibility of suggesting novel insights is not enough, as most theories can do this much; (2) these hypotheses must in fact be novel: employing evolutionary theory in economics cannot plausibly be defended from a heuristic point of view if the results of this employment have already been known about beforehand; and (3) the hypotheses suggested must be genuinely interesting: pointing to hypotheses that are minor variations on existing ones is not enough to justify an application of evolutionary theory as genuinely heuristic. In respect to (2) and (3), note further that it is not fully clear exactly what it takes for some result to be genuinely “novel” or “interesting.” However, no sophisticated account of this is needed here—in particular, for present purposes, it is enough to note that some particular result cannot plausibly be seen to be novel if it is already being investigated and discussed in the literature; a particular account of what is interesting will not be appealed to.

The goal in the rest of this paper is to see whether either or both of these two uses of evolutionary theory can be applied to the above three doctrines of the narrowly understood evolutionary economics. Consider, then, these doctrines in turn.

III. Evolution and Rational Choice

The appropriate analysis of rational decision making is a key issue of contemporary economic debate—rational choice arguably is right at the heart of the discipline and has implications for virtually all of the questions the latter investigates. ⁷ It is therefore unsurprising that one of the major ways in which evolutionary theory is thought to have entered economics is located right here.

In this context, what evolutionary economists have argued is that the dominant account of rational decision making in the economic literature—rational choice theory (RCT)—is drastically flawed. ⁸ To understand their attack on this theory, it is best to begin by noting that, in broad outline, RCT sees economic agency to be based on (expected) utility maximization—that is, on the choice of actions whose expected consequences rank highest in the agent’s preference ordering. More specifically, RCT sees economically rational decisions as stemming from a mechanism with the following four key features: ⁹

Importantly, evolutionary economists contend that, for one reason or another, evolutionary considerations can be seen to undermine all of these core tenets of RCT. ¹¹ While there is much to be said about the particular details of this evolutionary attack on RCT, for present purposes, a brief survey of its different components is sufficient to bring out its overall structure. ¹²

In the first place (and perhaps most famously), Gerd Gigerenzer and his collaborators have argued that an evolutionary informed perspective on (human) decision making reveals economic agency to be the result of “simple heuristics”: decision rules that yield quick results that are good enough for the case at hand (though not necessarily optimal and universally applicable) and do not require many informational or computational resources (see, e.g., Gigerenzer and Selten 2001). The reasons why evolutionary theory is said to support this picture are twofold.

On the one hand, relying on an optimizing mechanism such as RCT to make decisions can be extremely costly in terms of concentration, attention, energy, and time spent. By contrast, “satisficing” mechanisms like simple heuristics are designed to make humans act in a (biologically) appropriate manner without having to incur the costs of making these decisions in an optimizing way. On the other hand, it is claimed that decision making based on simple heuristics is a natural consequence of the fact that humans are biological organisms and thus subject to the way biological evolution works. In particular, natural selection is a process that typically builds specialist solutions to specialist problems; as far as natural selection is concerned, general problems do not exist and thus do not require (general) solutions. ¹³ Hence, it is implausible to suppose that humans employ an optimizing, domain-general reasoning mechanism such as RCT, rather than a set of satisficing decision-making tools for specific situations. Therefore, claims a and b are false.

Next, Brian Skyrms and other evolutionary game theorists have noted that when analyzing decision making from the point of view of evolutionary game theory, it becomes clear that there are reasons to expect agents to sometimes choose dominated strategies (see, e.g., Skyrms 1996, 2004; Alexander 2007). ¹⁴ The general gist of these arguments is as follows (see, e.g., Skyrms 1996, chap. 3).

Assume that an agent can choose between two different strategies, S₁ and S₂; assume also that S₁ dominates S₂. Next, assume that agents pair up and that the payoff an agent receives upon choosing S₁ or S₂ depends on what her partner chooses. Finally, assume that agents change their strategy choices in light of the payoffs they receive: agents who get higher payoffs stick to their chosen strategy; others change it. ¹⁵ Given all of this, it can turn out that, in the evolutionarily stable state—that is, in the situation that is such that, if small numbers of players chose different strategies from the ones in the stable state, they would do worse and thus the system would return to the stable state—S₂ may be chosen by some or all members of the population; moreover, these states can be reached from many different starting strategy distributions. What this shows, therefore, is that evolutionary theory again suggests that RCT should not be seen to be a plausible hypothesis about economic agency: it can be fitter—and thus, presumably, more economically successful—to choose dominated rather than dominant strategies. Hence, claim c above is false as well.

Finally, a number of researchers have suggested that evolutionary theory underwrites the idea that RCT gives a bad account of the attitudes toward risk of economic agents (see, e.g., Robson 2001; Orr 2007; Okasha 2007). ¹⁶ In particular, they argue that evolutionary theory would seem to lead one to expect economic agents to be risk averse: agents will be fitter—and thus, presumably, more economically successful—to the extent that they care about how risky their actions are, not just what their average payoffs are.

To see this, note that it can be shown that the fittest organisms are those whose decision making is sensitive not just to the expected number of offspring these decisions yield but also to their variance. While it is controversial exactly how the variance should enter into an organism’s fitness calculation, it is not controversial that it should (see, e.g., Gillespie 1977; Sober 2001). This comes out clearly from comparing two strategies A and B, where A leads an organism to consistently leave 1 offspring every period and B leads an organism to either leave 2 or 0 offspring, with equal probability. Note that, in absolute terms, the expected number of A and B offspring are the same in both cases—1—however, in relative terms, the expected number of offspring of the two strategies will be different: the expected relative number of A offspring is 1/2 [1/(2+1)] + 1/2 (1/1) = 1/6 + 3/6 = 4/6 = 2/3. Since 2/3 > 1/2, we would expect that, over time, there will be more and more A choosers in the population. In short: risk aversion would sometimes seem to be adaptive. Thus claim d above should be seen to be disconfirmed from an evolutionary point of view, too.

In a nutshell: if considered from an evolutionary point of view, it seems to become clear that RCT cannot be supported as a plausible account of rational decision making: some of its key tenets—claims a-d above—are shown to be false from this perspective. Hence, some other account—such as the satisficing, heuristics-based one of Gigerenzer et al.—ought to be adopted instead. However, when analyzed in more detail, it turns out that the evolutionary considerations do not in fact support this conclusion. This is for two reasons.

On the one hand, these considerations are not well seen as heuristic devices that have pushed the discussion surrounding the plausibility of RCT forward greatly. Primarily, this is because the above evolutionary arguments have few new, major implications that can lead (or have led) to important discoveries being made. To see this, note firstly that the general plausibility of RCT—and especially its commitment to maximization—has been a topic for discussion for a long time (arguably since Bernoulli’s time) and secondly that heuristics-based alternatives to the theory have also been discussed—for nonevolutionary reasons—for decades (see, e.g., Simon 1957). The same goes for the reasonableness of the requirement to avoid dominated strategies: this, too, has been discussed frequently and for a long time in a nonevolutionary context (see, e.g., Hargreaves Heap et al. 1992, 110-11; Cantwell 2006). Finally, the plausibility of RCT’s depiction of an economic agent’s attitudes toward risk is one of the oldest discussion points surrounding the literature and did not need evolutionary theory to become noticed (for discussion, see, e.g., Jeffrey 1983, chap. 10). ¹⁷ In short: it is not plausible to see evolutionary theory as having made a useful heuristic contribution to this debate.

On the other hand, it is not clear that it can be seen as having made an evidentially useful contribution to the denial of claims a-d above either. This is due to the fact that, while the above evolutionary arguments differ in numerous particulars, there is a common element in all of them that, as they stand, makes their evidential status problematic. This common element concerns the fact that the above arguments all are parameter specific: they require certain conditions to be in place for RCT-based decision-making mechanisms to be maladaptive. ¹⁸

More specifically, none of the above arguments show that it is always—or at least most of the time—adaptive to make decisions in a non-RCT like way. At best, all that these arguments can show is that, under certain circumstances, making decisions in a non-RCT-like way is adaptive. This can be seen from the fact that all of these arguments are hedged by “sometimes” and “typically” clauses at crucial places. In particular, it is not generally true that RCT-like decision making must be slower and more costly than simple heuristics-based decision making—this will depend on the kind of decisions one has to make (see also Sadrieh et al. 2001; Sterelny 2003; Richerson and Boyd 2005; Schulz 2011a). Equally, whether evolution by natural selection leads to specific solutions or more general ones depends on the nature of the adaptive problem to be solved and the traits the organism in question already has (see also Schulz 2008a). Next, whether dominated strategies can be expected to be stable states in an evolutionary dynamic depends (among other things) on the degree to which players choosing the various strategies are correlated with one another: if players pair up randomly, dominated strategies will never be part of the stable state (see, e.g., Skyrms 1996, chap. 3). Finally, while it is not controversial that some kind of risk aversion is adaptive, there is considerable controversy over the extent to which risk aversion matters: in particular, how strongly risk neutrality is selected against depends very much on the details of the case and can range from hardly at all to very strongly (see, e.g., Gillespie 1977; Cooper 1987, 1989; Sober 2001).

This parameter specificity of the above arguments matters, as it means that these arguments will only be evidence against claims a-d above to the extent that the conditions in which humans actually evolved were in the relevant range of these parameter values. However, to say the least, it is not clear that this is true.

Firstly, few scholars have even tried to make a serious attempt at connecting the known evolutionary biological facts with the conditions for when non-RCT-based decision making is adaptive: evolutionary economists typically leave it at rather broad claims about the fact that non-RCT-based decision making is “plausibly” more adaptive than its opposite. That, though, is not enough to make evolutionary theory evidentially useful in the debate surrounding the cogency of RCT: an explicit, well-grounded connection between these two fields needs to be established.

Secondly, there is still not all that much known about human evolutionary history in the first place, so that it is by no means easy to fill the gap left by the evolutionary economists (see, e.g., Mithen 1990; Sterelny 2003, 2012, for some careful attempts to do so)—and what is known does not always support their case. In particular, given the importance of the social environment to human cognitive evolution (see, e.g., Byrne and Whiten 1988; Sterelny 2003, 2012), the adaptiveness of simple heuristics is often thought to be quite implausible as a completely general account of decision making: relying on simple behavioral rules would seem to be highly detrimental if the environment (in the form of predators or competing conspecifics) actively seeks to exploit fixed action patterns like this (see also Sterelny 2003; Schulz 2011a). Further, how often conditions were right to make choosing dominated strategies adaptive is also unclear: while it seems true that many human interactions have been based around nonrandom associations (e.g., centering on kin), many human interactions seem to involve virtually random associations (see Cosmides and Tooby 1992b; Sterelny 2003). Finally, something similar is true when it comes to risk aversion: it does seem true that a number of our decision problems might have rewarded considerable risk aversion; however, it also seems to be true that many others did not (e.g., many foraging or hunting decisions: see, e.g., Mithen 1990).

In short: a closer look at the evolutionary accounts of decision making presented suggests that, as matters stand, they cannot be considered either as heuristically useful for debating the plausibility of RCT or as providing evidence against this plausibility. Note that this point has nothing to do either with the fact that these accounts are typically based just on natural selection (and thus ignore other evolutionary possibilities, such as drift or phylogenetic constraints) or with the fact that establishing what is really adaptive is prohibitively complex (as has been claimed by Rosenberg 1994). Rather, the claim is just that, as matters stand, we neither learn something new and interesting from these arguments, nor do we have the data to support the assumptions that underlie them. As will become clearer momentarily, something similar holds true for the second economic hypothesis that the narrow evolutionary economics is based on.

IV. Firms, Economic Agents, and Evolution

Ever since the inception of economics as a science, the two paradigm economic entities have been taken to be individual agents and firms (see, e.g., Arrow and Hahn 1971; Hausman 1992). ¹⁹ However, this picture of the central importance of firms within economic theory belies the controversy surrounding the nature of these entities: in fact, there is much debate over what, exactly, firms are, why they exist, and why there is more than one of them in an economy (see, e.g., Coase 1937; Schmalensee and Willig 1989; Hodgson 1999, chap. 11). ²⁰

When it comes to answers to these questions, two broad approaches can be distinguished in the literature (see, e.g., Hodgson 1999, 247-49; see also Fama 1980; Foss and Klein 2008). ²¹ Firstly, there are what may be called the “contract-based” approaches. According to these approaches, firms are a special kind of “transactional space” (see, e.g., Coase 1937; Williamson 1971; Klein, Crawford, and Alchian 1978; Hart 2008; Aghion and Holden 2011). ²² In particular, on this kind of account, the key thing to be noted about firms is that they take some economic interactions out of the market—instead of buying and selling their inputs and outputs on the market, parts of a firm simply take their inputs from some other parts of the firm or give their outputs to these other parts. This internalization of economic interactions is noteworthy, since, in principle, economic agents should be able to buy and sell any factor of production on the market. ²³ To explain this flight from the market, this view of the firm therefore appeals to the existence of “transaction costs.” ²⁴

Transaction costs are the costs that an agent incurs from taking part in a market. There are two main types of these kinds of costs: on the one hand, there is the cost that comes from acquiring the relevant information—where which goods and services are bought and sold, at what price, what the differences are between the various goods and services, and so on—and on the other hand, there is the cost of drawing up explicit contracts about what is to be bought and sold. Of course, this does not mean that producing the goods and services in house will be costless: the relevant skills need to be learned, further capital and other inputs might need to be acquired, and so on. However, the important point to realize is just that taking part in a market does not come free.

This matters, as it provides a reason for taking some economic transactions out of the market. In particular, firms can be seen to exist to minimize these kinds of transaction costs: they will grow up to the point at which, for the last transaction in question, the costs that come from taking part in a market are matched by the costs that come from producing the relevant good or service in house—thus implying that the economic agent minimizes overall production costs. In short: on the contract-based approach, firms are a certain kind of economic environment within which economic agents act—they make it possible for economic agents to transact more cheaply than what would be possible on the market.

By contrast, on the other main approach to the theory of the firm—the so-called competence-based approach—firms embody a certain kind of quasi-mental competence about how to produce various goods and services—they are collective agents (see, e.g., Knight 1921; Penrose 1959; Clark 1997, chap. 9; Hodgson 1999, chap. 11). In particular, here, firms are seen as having accumulated, on the one hand, a certain set of beliefs about how to produce some particular good or service and, on the other, a set of motivational structures with which to see through this production. Importantly, these beliefs and motivations can differ from those of any of the employees of the firm: what the firm believes might be different from what any of its particular employees believes (e.g., it might know more than what any one employee knows), and what the firm wants to do (e.g., maximize profits) might be different from what any of its employees wants to do (e.g., work fewer hours) (see also Hodgson 1999, 251-58; Simon 1957; Pettit 2003).

This matters, as it seems to imply that firms must be seen as economic agents in their own right: after all, they have quasi-beliefs and desires and generally appear to have all the necessary ingredients for genuine agency (see also List and Pettit 2006). In short: on the competence-based approach, firms are seen as more than mere transactional spaces—they are taken to be economic actors in their own right.

Given this dispute between the contract-based and competence-based approaches toward the firm, evolutionary economists have argued that it is possible to use evolutionary theory to defend the latter view (see, e.g., Hodgson and Knudsen 2010; Hodgson 1999; Nelson and Winter 1982; see also Vromen 2001). This defense takes off from two key claims. Firstly, evolutionary economists argue that markets can select firms—for example, for their profitability (see, e.g., Hodgson and Knudsen 2010, chap. 7; see also Alchian 1950; Enke 1951). Much like natural selection selects fitter organisms, markets are said to select more profitable firms: these are more likely to persist and create other firms as spinoffs. Secondly, evolutionary economists note that evolutionary biologists have increasingly come to accept groups of organisms as possible evolutionary biological entities in their own right. In particular, much recent work on the “units of selection” problem (i.e., the problem of which entities—genes, organisms, groups, etc.—are the targets of natural selection) adopts “multilevel selection theory”: ²⁵ according to this theory, natural selection can act on many different levels and target groups as well as organisms or genes. ²⁶ In the present context, these two claims are important, as they can be seen to make two kinds of arguments for the competence-based theory.

Firstly, in a heuristic way, they can be taken for a “how possibly” story that shows that it is possible for collectives to be individuals of their own. ²⁷ In particular, the idea here is that the adoption of multilevel selection theory in evolutionary biology shows that it is possible for (closely interacting) groups of individuals to be the targets of a selective mechanism—and thus (or so it may appear) to be biological individuals in their own right. If this insight is combined with the (alleged) fact that markets can be seen as just such a selective mechanism, then it becomes clear that it ought to be seen as something worth exploring whether a similar conclusion might not also hold true in economics: maybe (closely interacting) groups of individual agents—that is, firms—should also be seen to be economic entities of their own (see also Hodgson and Knudsen 2010; Vromen 2001; Nelson and Winter 1982).

Secondly, the evolutionary biological work on multilevel selection theory can also be taken to be evidentially important for the debate surrounding the theory of the firm. This is so, as it might be thought that the fact that markets can select among different firms implies that one needs to recognize firms as individual agents—for after all, for something to be the target of a selective mechanism, it has to be a bona fide entity of the relevant kind. In particular, one might argue that if it is a basic economic fact that markets select between different firms, and if it is a basic evolutionary biological fact that the targets of a selective mechanism must be autonomous evolutionary biological entities themselves, then firms simply have to be seen as economic agents of their own (see also Witt 1999; Vromen 2001; Hodgson and Knudsen 2010). In this way, one might see the adoption of multilevel selection theory in economics to be evidence for the competence-based view of the firm.

However, when considered in more detail, it turns out that, as they stand, these evolutionary arguments are unconvincing as well. When it comes to the heuristic argument, the reason for this is again relatively simple: it was already known that it was possible for firms to be economic agents of their own (in some sense at least). Indeed, the formulation and defense of the competence-based approach toward the firm derives to a large extent from nonevolutionary considerations: in particular, it is typically argued for on sociological and organizational grounds (see, e.g., Nickerson & Zenger 2004; Foss and Klein 2008; see also Hodgson 1999, chap. 11) or even on the basis of problems of coordination failure in standard game theory (see, e.g., Bacherach 1999). In a nutshell: the heuristic version of the above evolutionary argument can only be seen to be weak, as it does not suggest something that is genuinely novel.

Unfortunately, the evidential version is not much better off. This is so for two reasons. ²⁸ Firstly, as presented, the above evolutionary argument is not actually sound. As far as biological evolution is concerned, groups can be the targets of selection without being evolutionary biological entities in their own right (see, e.g., Okasha 2006; Godfrey-Smith 2009). In fact, the emergence and persistence of novel evolutionary biological entities—“Darwinian individuals”—requires special conditions, the exact nature of which is still controversial (see, e.g., Buss 1987; Maynard Smith and Szethmary 1995; Michod 1999; Okasha 2006; Godfrey-Smith 2009, 121-28). ²⁹ This matters, as the pure fact that firms, qua groups of interacting economic agents, are being selected on the market should then not be seen to make them economic entities of their own right—indeed, it is possible to accept that firms are selected by the market and that evolutionary biology allows the selection of groups and still defend a contract-based view of firms. This is possible, as group selection alone is not sufficient for the existence of high level Darwinian individuals: indeed, the typical models of group selection (what is often called “MLS1”—see, e.g., Damuth and Heisler 1988; Sober and Wilson 1998; Okasha 2006) have it that groups are not Darwinian individuals of their own—the focus of these analyses always remains the individual members of the group (see also Godfrey-Smith 2009; Okasha 2006). ³⁰

Moreover (and this is the second problem with the evidential version of the evolutionary argument here), the situation does not improve if the argument is reformulated to say that, under some special conditions, evolution by natural selection can lead to the emergence and persistence of new Darwinian individuals and that, therefore, we ought to expect the same to be true in the economic realm when structurally similar conditions obtain. This reformulation does not save the argument, for we now run into the same parameter-specificity problem we encountered in the context of the evolutionary arguments against RCT—and that in an especially egregious version.

This is so, as, on the one hand, there is no clear theory yet as to which conditions are conducive to the formation and persistence of novel Darwinian individuals: while there are suggestions about both the abstract nature of these conditions (such as the need for very close cooperation among existing Darwinian individuals—see, e.g., Godfrey-Smith 2009; 121-28; Okasha 2006, chap. 8) and the details of the generation and persistence of particular novel Darwinian individuals (see, e.g., Buss 1987; Maynard Smith and Szethmary 1995), there is currently no general theory as to which conditions are such that they typically lead to the generation and persistence of new Darwinian individuals. On the other hand, even if we had such a general theory, it is clear that there is no reason to think, as yet, that the conditions it identifies match those prevailing in any—not to mention every—actual economy. At the very least, much more work is needed to show this—and no one has even attempted to do this work yet. In short: even a coherent version of the evolutionary argument for competence-based theories of the firm is far from being able to make for genuine evidence in favor of these theories.

Overall, therefore, it becomes clear that the argument in favor of the second tenet of the narrow understanding of evolutionary economics is unconvincing as well. Consider then the third component to see if the situation is different there.

V. Economies, Equilibria, and Ecology

The last of the major theses of evolutionary economics concerns the claim that the emphasis on equilibrium analysis that characterizes contemporary mainstream economics is misguided: instead, market economies should be seen as dynamic systems that constantly change and are never in any kind of rest or equilibrium state. To represent them, therefore, a different kind of analysis is needed: one that is focused on the constantly changing dynamics of the system, rather than its equilibrium states (see, e.g., Vromen 2001; Hodgson 1999, 258-60; Young 1998).

In a little bit more detail, the idea here is the following. Note firstly that contemporary economic theory is based on equilibrium analysis: its goal is to determine the steady, stable state of the economy and the dynamics leading up to that stable state (see, e.g., Mas-Colell, Whinston, and Green 1995; Hausman 1992). For example, we might consider a model with (linked) equations for the interest rate, price level, stock market index, and various other economic variables to determine the steady state growth rate of the economy and the dynamics of the system: from which initial positions will the economy converge to the steady state, and at what rate? ³¹ Given the identification of this equilibrium, we might then go on to consider changes in the basic parameters of the system to see what the new stable state is, how it compares to the old one, and how it can be reached. So, for example, we might consider changing the interest rate in the above model to determine what the new steady state growth rate will be (e.g., whether higher or lower than the old one) and how a system in the old steady state will adjust to the new one (e.g., will some variables overshoot and then correct back to the normal level?). In short: standard economic theory seems to see the crucial goal and value of economic analysis as identifying the equilibrium states of an economy, together with the transition dynamics leading to these equilibrium states.

However, it may be thought that this sort of analysis is based on some drastically wrong presuppositions (see, e.g., Lawson 2006). In particular, it would seem obvious that economies are highly complex systems: a large variety of constantly changing causal factors (both in nature and in quantity) is impinging on them at any given moment (see also Hausman 1992). Given this, it would seem that it should be expected that real-life economies are (a) hardly ever in equilibrium and (b) constantly changing in their fundamentals. In turn, this would seem to make equilibrium analysis highly implausible as a methodology for studying these sorts of systems: after all, due to point (a), the equilibrium of the system will not describe any actual state it is ever likely to be in, and due to point (b), knowing the transition dynamics will not help in this either, as real economies are not even likely to ever really approach any particular equilibrium. Hence, a better methodology for studying these economies would be based on some kind of chaos theory: just like in the study of meteorological systems, economic models should apply the tools used for understanding complex, chaotic systems. ³²

To support this viewpoint, evolutionary economists could appeal to the supposed fact that evolutionary biology and related disciplines have given up on equilibrium analysis as well (see, e.g., Young 1998; Groenewegen and Vromen 1999; Vromen 2001). ³³ In particular, they might note that it now seems standardly assumed that (evolving) ecosystems, for one, are hardly ever in equilibrium and that, therefore, the best kind of analysis of these sorts of systems is a chaos theoretic approach (see, e.g., Hastings et al. 1993; Cushing et al. 2003; see also Sterelny 2001). This matters, as it also seems to be the case that ecosystems are structurally quite similar to real-life economies. In particular, both cases feature a multitude of causes with many feedback cycles and other complex interdependencies. Given these structural relations between the two cases, one might therefore think that the same kinds of methodological tools should be applied to both cases as well. Since it is widely accepted (or so it might claimed) that a plausible science of evolutionary ecology does not rely on equilibrium analysis, the same should be taken to be true when it comes to economics.

As before, this argument can be read in two ways, heuristically and evidentially; and as before, neither is particularly compelling. The heuristic reading of the argument struggles once again with the novelty requirement of plausible heuristic uses of evolutionary theory. In particular, the appeal to evolutionary theory does not seem to have been instrumental in suggesting the possible need for nonequilibrium analyses in economics—in fact, this possible need has been widely discussed in economics for quite some time (see, e.g., Smith 1776; Schumpeter 1934). To be sure, it may be that, while known, this idea was not very strongly appreciated before the evolutionary argument was put forward: this argument may have brought it more to the forefront in (methodological) debates in economics. However, it should be clear that this can only be considered a quite minor role for evolutionary theory in economics and hardly the foundation of a drastically novel approach toward the subject.

Turning then to the evidential reading of the argument, it turns out that this is somewhat implausible, too. The reason for this is that it is neither clear that its evolutionary presuppositions are true nor that the situations in (evolutionary) ecology and economics are as parallel as made out to be by the above argument. To see this, note firstly that it is by no means so obvious that evolutionary ecologists have abandoned equilibrium analysis (broadly understood) completely. In fact, in some cases, they seem to stick to it quite closely, having merely made it more sophisticated: for example, many versions of the Lotka-Volterra approach toward modeling predator-prey relationships are inherently equilibrium focused (see, e.g., Odenbaugh 2005); much the same is true for many other areas of ecological research (for an overview, see, e.g., MacArthur and Wilson 1967; Roughgarden 1979; see also Justus 2008). For this reason, it is not obvious that it is indeed true that ecologists have concluded that equilibrium analysis is untenable in their discipline: in fact, to what extent equilibrium analysis is tenable in ecology is still a hotly contested subject (see, e.g., Hastings et al. 1993; Sterelny 2001).

Secondly, it is far from clear that ecology should be taken for the model of economics here in the first place. It is true that both ecology and economics center on complex systems that are affected by many different causes in a myriad of complex ways. This, though, does not mean that these kinds of systems cannot make up a very heterogeneous bunch, with some requiring a genuine dynamic analysis and some being amendable to equilibrium approaches. In particular, it may be true that (many or most) ecological systems are structurally like meteorological systems and therefore require a more chaos theoretic methodology to be understood; however, it may also be true that (many or most) economies are not like this and therefore are easily understandable in terms of equilibrium models. The reason for this is that, while both systems are based on complex causal interrelations, the way they are based on these causal interrelations might be different in the two cases: in the relevant ecological cases, no genuine approach to an equilibrium might ever occur (due to the frequency and strength of external and internal effects on the system). By contrast, it may well be the case that economies typically are subject to fewer and weaker shocks—in particular, the shocks might leave the underlying equilibria untouched—and spend most of their time approaching some equilibrium or other. If that is so, though, then we have not been given a reason to take the alleged problems of equilibrium analysis in ecology to similarly be problems for equilibrium analysis in economics. Importantly, moreover, we have, as yet, no reason to think this is not a very real possibility about the relationship between the two cases: at this point at least, whether economies are like ecological systems in these respects is an entirely open question—and one that evolutionary economists have not even begun to answer.

Overall, therefore, it becomes clear that the third major tenet of evolutionary economics—the dynamic nature of real economies—also lacks the needed evolutionary support (either heuristically or evidentially). What remains to be done is to put the conclusions of the past three sections together, to see what overall picture of the state of evolutionary economics emerges.

VI. Conclusion

I have tried to argue that the appeal to evolutionary theory in economics, if understood narrowly as defending three key theoretical hypotheses in the subject, cannot yet be seen to be fully plausible or useful. In particular, I hope to have shown that, on the one hand, the relevant evolutionary considerations are of little heuristic value: the major ideas they suggested have virtually all been known and debated for some time. On other hand, these evolutionary considerations also do not add much of evidential value to these debates: the argumentative presuppositions of these considerations are not yet clear enough to permit them to ground any further economic hypotheses. Hence, I conclude that evolutionary economics, as it stands, does not add much to traditional economic theorizing.

However, there are also two (related) more positive lessons that can be taken away from the present discussion. Firstly, it is in particular the evidential uses of evolutionary theory that future work in evolutionary economics should focus on. This matters, as it goes contrary to what is sometimes being suggested in this context (see, e.g., Rosenberg 1994; Witt 1999, 2008; Hodgson 1999; Hodgson and Knudsen 2010): it is commonly thought that heuristic defenses of evolutionary economics can be gotten off the ground easily and that it is evidential applications of the program that are hard to achieve. However, what the above makes clear is that, if there is any promise in evolutionary economics, then it is precisely on the evidential side: if the necessary information could be obtained, then evolutionary economics would indeed be able to significantly push forward various discussions in the discipline. Making evolutionary theory heuristically useful in economics, by contrast, looks to be much harder: achieving the necessary novelty just is no mean feat.

Secondly, the present criticism of evolutionary economics differs in important ways from others in the literature. In particular, unlike Rosenberg (1994) or Gould (1996), I have not tried to show that the idea of using evolutionary theory in economics is somehow generally misguided or could never be made to work. Rather, the idea has been merely to show that, as matters stand, applications of evolutionary theory in economics have not been useful; this, though, is consistent with there being future suchlike applications that do not fall prey to the above worries. In other words, my point here was not that there are timeless structural limitations to work in evolutionary economics but merely that there are current epistemological limitations to it. For this reason, it can also be seen as an upshot of the paper that more, not less, work in evolutionary economics is desirable.