A flurry of recent activity indicates that evolution is beginning to occupy center stage in economic debates—and not a moment too soon.
Recently published books include Robert Frank’s The Darwin Economy: Liberty, Competition, and the Common Good (in which he predicts that Darwin will eventually be regarded as the father of economics), Yochai Benkler’s The Penguin and the Leviathan: How Cooperation Triumphs Over Self-Interest, Geoffrey Hodgson and Thorbjorn Knudsen’s Darwin’s Conjecture: The Search for General Principles of Social and Economic Evolution, and my own The Neighborhood Project: Using Evolution to Improve My City, One Block at a Time.
Behind the trade books is a growing academic movement, including a recently concluded conference organized by Denise Dollimore and Geoffrey Hodgson titled “Evolutionary Thinking and Its Policy Implications for Modern Capitalism”. As president of the Evolution Institute, I am privileged to function as a coordinator in addition to my own contribution, including a collaboration with NSF’s National Evolutionary Synthesis Center (NESCent) on integrating economics and evolution. One result has been a white paper submitted to NSF titled “The Relevance of Evolution for Economic Theory and Policy”, co-authored with economist John Gowdy and with 64 signatories, including luminaries such as Pulitzer prizewinner E.O. Wilson and Nobel laureate Elinor Ostrom.
Economic policies informed by evolution do not fall cleanly into any current political camp, presenting an opportunity to formulate a new Middle Path. Another exciting prospect is that the dialogue can be based on scientific norms of accountability, which, while not entirely genteel, are vastly more constructive than the fight to the death and disregard for the facts that current political discourse has become. In this spirit, I will focus on a review of Frank’s book in Slate magazine by science writer John Whitfield, whose own book, People Will Talk: The Surprising Science of Reputation, will be published in November.
Frank does not aspire to entirely rethink economics in The Darwin Economy. Instead, he focuses on a single foundational difference between evolutionary and economic theory. Evolutionary theory is based on relative fitness—which individuals survive and ultimately reproduce better than other individuals. Economists sometimes think in terms of “positional goods”, in their own parlance, but the main edifice of economic theory is based on absolute utility maximization—as much as possible for me, regardless of what anyone else gets.
Thinking about human economic behavior in more relative terms explains some of the excesses that are on display everywhere we look, from the $1000 suit that provides a competitive edge over a $500 suit for a job interview, to the 4000 square foot house that’s worth buying in a neighborhood of 2000 square foot houses but not in a neighborhood of 8000 square foot houses, to the millions of dollars worth spending for your daughter’s sweet sixteen party, if you happen to have billions of dollars at your disposal. The fact that these excesses are obvious does not mean that they have been incorporated into economic theory. Frank therefore provides an important service by pointing out the difference between absolute vs. relative utility thinking and exploring some of the implications—including a very feasible tax policy that should appeal to thinking libertarians in addition to liberals. Frank is a highly respected economist, an insider rather than an onlooker such as myself. His proposed tax policy richly deserves to occupy center stage in political and economic debates. If it is adopted, it could literally alter the fate of the nation and the world.
Whitfield agrees with Frank’s tax plan but disagrees with his evolutionary rationale. According to Whitfield, Frank pays insufficient attention to another foundational difference between evolutionary and economic theory. For an economist, the pursuit of self-interest typically results in a well-functioning economy. That’s what the metaphor of the invisible hand (and a lot of neoclassical economic theory) is all about. For an evolutionist, functioning well as a group requires cooperation, which is often undermined by individual self-interest. Explaining how cooperation, altruism, and anything else that appears “for the good of the group” is one of the central problems of evolutionary theory. Many attempts have been made to provide an explanation, which sail by names such as group selection, kin selection, inclusive fitness theory, and evolutionary game theory. Even so-called selfish gene theory treats cooperation as a central problem with a (partial) solution, using its own vocabulary. The relationships among all these theories are famously confusing, even to evolutionary insiders, not to speak of onlookers such as Frank and Whitfield.
Whitfield is right that Frank pays scant attention to the problem of cooperation in The Darwin Economy. You’d think that I would have beaten Whitfield to the punch in my own videocast with Frank, but I was content to focus on Frank’s positive contribution. Now I welcome the opportunity to reconcile the points made by Frank and Whitfield. Intriguingly, if we follow Frank by thinking assiduously in relative terms, we can resolve most of the confusion that famously shrouds evolutionary theorizing about cooperation.
To begin, consider a tinker toy evolutionary model in which a single group consists of two types, solid citizens and slackers, that breed true. One of the joys of being an evolutionist is that these could be bacterial types just as easily as human types. The two types can exist in any initial proportion. The solid citizens provide a benefit to the whole group, including themselves, at no personal cost. You might think that such a no-cost public good is unrealistic, but at least it is imaginable. Even though the solid citizens benefit themselves in absolute terms, the proportion of solid citizens in the group has not changed. No evolution has occurred.
What is required for solid citizens to be favored by natural selection in this model? Imagine that there are several groups that vary in the proportion of solid citizens. Within each group, no evolution takes place. The proportions remain the same or vary only by chance. But the groups with more solid citizens fare better than groups with fewer solid citizens. The addition of several groups to the model creates a relative fitness difference that was lacking in the single group model—and the fitness difference is between groups, not within groups.
This is a multilevel selection model for the evolution of solid citizen behavior. In keeping with Frank’s emphasis, it assiduously keeps track of relative fitness differences within and among groups. It is easy to understand, but it has important consequences for studying economic behavior from an evolutionary perspective.
First, notice that between-group selection is required to explain the evolution of a no-cost public good, even before we consider examples in which benefitting others requires a personal cost. There are no fitness differences in a win-win situation, at least among the winners. Thinking assiduously in terms of relative fitness within and among groups changes the way that we classify individual- and group-level benefits. A no-cost public good qualifies as selfish in absolute fitness terms, and need not require caring about others psychologically, but still requires a multi-group population structure and between-group selection to evolve in an evolutionary model.
Second, since a no-cost public good is neutral with respect to within-group selection, any amount of variation among groups provides a sufficient differential for it to evolve. When we consider traits that benefit others or the whole group at a personal cost, then between-group selection favoring the trait must be strong enough to overcome within-group selection operating against the trait. The stronger the selection differential within groups, the more variation among groups is required for group selection to carry the day.
Third, models framed in terms of kin selection, inclusive fitness theory, evolutionary game theory, selfish gene theory, or any other formalism, are not exceptions to the rules outlined above. All of them assume that social behaviors are expressed in groups that are small compared to the total population (e.g., the N in N-person game theory), the behaviors labeled “cooperative” and “altruistic” are typically selectively disadvantageous within groups, and evolve only by virtue of the differential contribution of groups to the total population. These models often seem as if they don’t invoke between-group selection because they track what evolves in the total population without tracking selection differentials within and between groups. As soon as the appropriate partitioning is made, the roles of within- and between-group selection are there for all to see. The apparent differences among the models are artifacts of different accounting methods.
Fourth, once we become accustomed to translating among the various accounting methods, the importance of between-group selection for the evolution of solid-citizen traits becomes uncontroversial. This is already regarded as “obvious” among the cognoscenti, including some that prefer the inclusive fitness accounting method to the multilevel selection accounting method, but this quibble should not be confused with the earlier and much more substantive claim that group-level selection is invariably weak compared to within-group selection. That claim has been authoritatively rejected long ago.
Fifth, humans employ a panoply of mechanisms to function well as groups, including rewards, punishments, reputations, and all that. Some of these mechanisms appear self-interested in psychological terms. For example, I might behave well purely to enhance my reputation without caring at all about anyone else’s welfare. If we want to know how these mechanisms evolved by genetic or cultural evolution, we need to model them as traits that evolve in competition with other traits. If we think assiduously in relative terms, then we’ll want to know if these traits evolve on the strength of fitness differentials within vs. between groups. Based on an extensive modeling literature, mechanisms that sail by names such as indirect reciprocity, reputation, and policing are higher-order public goods that provide benefits to whole groups by causing others to provide lower-order public goods. The cost of providing a higher-order public good can be very low (e.g., gossiping about the misbehavior of others), verging on a no-cost public good. But even a no-cost public good requires between-group differentials to evolve, as we have seen.
Sixth, an important type of multilevel selection called equilibrium selection takes place when social interactions result in multiple stable local equilibria. Although every equilibrium is locally stable, they can differ greatly in their group-level properties (e.g., all defect vs. tit-for-tat in an iterated prisoner’s dilemma model). Between-group selection takes the form of selection among equilibria, rather than the selection of a trait that is locally disadvantageous within all groups. Since the panoply of human social control mechanisms can stabilize just about any norm, equilibrium selection is an especially powerful form of between-group selection in our species.
Against this background, I can return to Frank’s thesis in The Darwin Economy and Whitfield’s critique in his review. If we read Frank’s book with multilevel selection in mind, between-group selection makes many appearances. For example, he notes that countries that don’t raise sufficient revenue to function well are replaced by countries that do. At least there is a replacement of social organizations, which is what multilevel cultural evolution is all about. This is not just an idle conjecture, but has been documented in considerable detail by Peter Turchin in his book War and Peace and War: The Rise and Fall of Empires.
Or consider Frank’s example of hockey players deciding whether to wear helmets. Those without helmets beat those with helmets in any given game, but everyone is worse off without helmets. It’s a classic multilevel selection scenario. The solution of creating a helmet-wearing convention enforced by punishment is a classic example of equilibrium selection. The work required to establish and maintain the convention is a classic example of a higher-order public good. The actual process of establishing the convention might involve deliberative planning rather than a raw process of between-group selection, but that only pushes back the inquiry to why we deliberatively plan that way. In short, there is plenty of multilevel selection in Frank’s thesis. If it doesn’t seem that way, it is because of how he describes it, not because of what is actually taking place.
Whitfield is right to call attention to “the toughest question in evolutionary biology: how has natural selection, which drives individuals to compete with their own kind, nevertheless produced so many examples of cooperation and altruism?” However, he’s out of touch with the current literature when he states that “Most evolutionary biologists, however, think that examples of group selection are somewhere between rare and nonexistent, because the benefits of individual selfishness almost always outweigh those of group solidarity.” He is confusing the earlier consensus about group selection, which collapsed a long time ago, with current quibbles about which accounting method one happens to prefer. If he assiduously divides relative fitness differentials into within- and between-group components, he will discover that not only is Frank’s economics implicitly group-selectionist, but also his own.
Let’s follow some of Whitfield’s thinking that he displays in his review. Commenting upon Frank’s tax plan, he states: “This is not bad economics: Government is an exercise in trying to make group selection work, doing the best for everyone within the borders while competing with other nations.” Here, he seems to be explicitly invoking group selection. As an aside, governments do not invariably strive to compete with other nations. Frequently they strive to cooperate with other social entities in a multi-tier population structure. One of the beauties of multilevel selection theory is that the same rules apply to all levels of the multi-tier hierarchy. If cooperation can evolve within groups, then it can also evolve within groups of groups.
Whitfield continues: “The best way to do this though, is not just to work out what’s best for everyone, but to align the public interest with self-interest.” It is unclear how he relates this statement to group selection in his own mind, but we can see that, at best, the alignment results in a win-win situation with an absence of fitness differences within the group. And work is required to create and maintain the alignment.
Whitfield continues: “…what evolutionary biology teaches us is that it’s not enough to assume, as Frank does, that everyone just wants to create the biggest economic pie. That’s like saying a gazelle cares more about the average speed of its herd than whether it can outrun a hungry cheetah.” Here, Whitfield is introducing psychological terms such as “wanting” and “caring” into the conversation. As we have seen with our tinker toy model, it doesn’t matter how the solid citizen who provides a no-cost public good thinks or feels about his actions. Perhaps he genuinely cares about the welfare of others, or perhaps only about his own share. Either way, there are no fitness differences within the group and between-group selection is required for the solid citizen type to increase in frequency in the total population. Whitfield is confusing proximate causation (psychological mechanisms) with ultimate causation (the consequences of behaviors for reproductive success, regardless of the psychological mechanism).
Whitfield concludes his review this way:
In fact, those leading and funding opposition to progressive taxation are rational enough—they’re the ones who do best if society becomes an arms race won by those with the biggest antlers and the priciest suitcases, with the lions getting anyone who can’t keep up. What those opposing them need to show is not just how the common good can be maximized, but how it can be reconciled with the self-interest of enough people to vote it into being.
I agree with Frank’s tax plan not so much because it maximizes the common good but because, as someone who relies on public services, I think it would be good for me. I also think that my country contains more people like me than it does plutocrats who have no need of a public realm and no desire to fund it. Evolution shows us how cooperation can come about despite the inescapable logic of selfishness.
According to Whitfield, opponents of progressive taxation are pursuing a strategy favored by within-group selection. If they succeed, they will be the last ones standing aboard the sinking ship. Whitfield’s use of the word “rational” implies that the opponents are consciously aware of their strategy and its consequences. I have my doubts. It’s equally likely that in their own minds, opponents of progressive taxation are convinced that they’re saving the ship. Once again, it is important to distinguish between proximate and ultimate causation. There are many ways to skin a cat, and there are many different psychological motivations that can result in the same political strategy. However the opponents of progressive taxation think about their actions, I agree with Whitfield about the ship-sinking consequences.
Whitfield then asserts the need to justify Frank’s tax plan in terms of individual self-interest and reports his own motivation. He’s the type of solid citizen (at least in this context) who cares only about his own share. He wrongly concludes that everyone must think that way because of evolution’s “inescapable logic of selfishness”. Evidently, those who were psychologically less self-centered weren’t among our ancestors. I feel sorry for Whitfield if he thinks that his own motivation is species-typical. In any case, evolutionary theory provides no comfort for his position. There is compelling theoretical and empirical evidence that other-oriented and society-oriented psychological mechanisms can fare well in a multilevel Darwinian world. And whatever Whitfield means by evolution’s “inescapable logic of selfishness” it is not that everything evolves by within-group selection.
What about the other authors who are beginning to move evolution onto center stage in economic debates? In The Penguin and the Leviathan, Yochai Benkler focuses his attention on cooperation, like Whitfield, but does a much better job reporting on the current state of knowledge, including multilevel selection and how it relates to other accounting methods. In Darwin’s Conjecture, Geoff Hodgson and Thorbjorn Knudsen are true scholars of evolutionary theory in relation to economic theory, from the beginning to the present. They do aspire to rethink economics, in contrast to Frank’s more narrow objectives. And my own Neighborhood Project shows how economics can be seen as a mere star in a universe of subjects that can all be viewed through the lens of evolutionary theory.
By calling Frank’s objectives “more narrow”, and reformulating some of his ideas in terms of multilevel selection, I am not detracting from the importance of his thesis in any way. On the contrary, by focusing on the foundational distinction between absolute and relative utility, Frank has come up with workable tax policy that deserves our most serious consideration. He could have been more explicit about multilevel selection, but there is nothing wrong with his evolutionary rationale.
Frank’s tax plan provides a concrete example of a Middle Path that can be followed by thinking people of all political stripes. And this commentary provides an example of political discourse bound by the norms of accountability associated with science and good scholarship. As evolution begins to occupy center stage in economic debates, there will be a Darwinian contest of views, but one that is carefully refereed to lead to productive outcomes.