Here’s a short essay on how nudge theory would be interpreted in an ecologically rational frame. This essay is a bit of a teaser for a paper I’m working on about nudge theory and its parallels to the socialist calculation debate. Enjoy!
In 2008, Chicago economist Richard Thaler and Harvard Law School Professor Cass Sunstein published Nudge: Improving Decisions about Health, Wealth, and Happiness. The authors contrast so-called Econs, “rational” individuals who maximize utility using some kind of optimization method, against Humans, who sometimes optimize but sometimes make “mindless” decisions that display empirical biases not present in “rational” decision-making. A “nudge,” in the words of Thaler and Sunstein, is “any factor that significantly alters the behavior of Humans, even though it would be ignored by Econs.” Humans respond to incentives and nudges, which, when properly deployed, “improve people’s lives, and help solve many of society’s major problems” (Thaler & Sunstein 2008: 8).
Nudges change in the context in which choices are made, not the content of choices. That is, how an agent makes her choice depends on the set of choices that she perceives and the methods by which she can transform those choices into the outcome she desires. Suppose we imagine a buffet table with salad on one side and cake on the other. An agent whose firm end is to cut her calories will choose the salad; an agent whose primary end is being polite and secondary end is cutting calories may take cake only if no people are watching her; and the agent who values cake over her self-respect may eat cake now and beat herself up over it later.
Nudges of the type proposed by Thaler and Sunstein require a belief about how a particular agent may make a better choice in the estimation of the agent if the agent were facing a difference choice context. Thaler and Sunstein argue that since individuals always make choices in a particular context anyway—and will face a variety of pre-existing influences—nudging is liberty-preserving. The authors model the effectiveness of nudges in a rational choice framework, whereby successful nudges assist agents in becoming more “rational” and thereby result in welfare gains. Hansen & Jesperson see pro-nudging arguments as splitting into two distinct camps: the “libertarian” camp, which constrains nudge policy to be preferences- and choices-preserving, and the “democratic” camp that would constrain nudge policy no more than traditional interventions are constrained, that is, subject to final review by voters in a representative democracy (Hansen & Jespersen 2013: 12).
It is in the best judgment of this author that agents cannot act rationally in the manner delineated in traditional optimization-based equilibrium economics due to computational and decidability issues (see discussions in Koppl, Velupillai, Wagner, Simon, Axtell for instance), and rather exhibit what Gerd Gigerenzer has called “ecological rationality,” (Gigerenzer & Todd 1999) or what Herbert Simon famously called “bounded rationality” (Simon 1996). Simon’s studies of behavior, in his words, were mostly of the “residual categories” of human behavior, as the reigning theory of choice at the time, and still currently, took into account utility gains only with respect to a very few parameters (Simon 1977).
As we’ve learned since, the “residual categories” of behavior are no less relevant to the decisions people make than the few parameters taken into account by rational choice models. Given the bounded rationality frame made necessary by expanding the analysis of human behavior to an analytically intractable set, it makes sense to analyze the effect of nudge theory in a boundedly rational frame replete with Simon-esque “residual categories.” Candelo & Wagner (hereafter, C & W) take a similar tack in an upcoming article on behavioral economics in an ecologically rational frame, and I see my analysis as an extension or refinement of theirs.
2. The cake-slice nudge
If we reject neoclassical theory as in C & W, we immediately see that nudges proclaiming to move a social system closer to a Pareto social welfare optimum are out of the question, analytically. Without the objective measurement device provided by rational choice theory, we can say little about social welfare, as defined. However, that doesn’t mean we can’t investigate the effects of nudges on the efficiency of agent computation, how well agents are able to learn a solution given a problem to solve, and whether nudges advantage the preferences and ends of policy makers.
An explicatory device used in C & W’s analysis is to outline how behavioral economics ignores Simon’s “residual categories” of behavior. They reference an experiment conducted that shows intransitive preference orderings arising from a choice between differently sized slices of cake (C & W: 7; Rizzo 2015). First an agent is offered three differently sized sliced of cake, and takes the medium sized slice. The experimenter then removes the largest slice and asks the agent to choose again; this time, the agent chooses the smallest slice of the three. C & W explain how what looks like intransitive preferences to a rational choice theorist could be explained by invoking the residual categories of choice ignored by the rational choice theorist: the agent may value being seen as polite, and translate that into never taking the largest slice offered. Therefore, in our analysis, a nudge could not be based on making an agent more rational, as there is no sense in which a choice made isn’t rational in the sense of acting according to one’s values.
An example of a nudge would be to explicitly offer people a choice between two sized slices of cake at first, perhaps written as such on the menu, but only offer the large size on request. This is an example of Thaler and Sunstein’s so-called choice architecture, where no choices are removed but it becomes more costly to make a choice deemed “bad” by policy makers. The policy implications of such a nudge are obvious, and touted with much gusto in hand-wringing documentaries like Morgan Spurlock’s “Super Size Me,” where Spurlock suggests that putting a cap on portion sizes offered to customers will have positive effects on health while having no negative effects on the utility of the customers. The Nudge-esque version of Spurlock’s suggestion would be to remove the super size category from the menu, but still make it available upon request. In effect, the policy maker sees herself not as manipulating market actors but as nullifying existing manipulations.
3. Transferring utility from “bad” decision-makers to “good” decision-makers
Nudge theorists claim they are giving agents avenues to make better choices that agents truly wish to make. How nudge theorists are supposed to know agents wish to make the choices nudge theorists wish them to make is typically explained using anecdotes. Regardless, the implication of nudge theory is that the agent learning mechanism is broken or needs a helping hand. Nudge theorists recognize on some level that instead of optimizing, agents rely on heuristics in order to make choices. Altering the context in which the heuristics of choice operate can short-circuit choice to favor behaviors preferred by policy makers. People can be manipulated, an ancient truth perhaps best explicated by George Orwell in his “Politics and the English Language.” But can a mass of people be manipulated in a general way with no negative consequences for anyone, at the time of the intervention or thereafter?
In an ecologically rational frame, our objects of analysis are no longer an agent’s budget constraint, the set of goods available on the market, and the price vector of those goods. By deviating from traditional optimization in creative ways, agents can lower the cost of making the combinatorially large number of decisions facing them at any particular moment. Agents have different methods of making choices from which to choose. Decisions-making methods are as combinatorially diverse as combinations of goods, and there is as fierce competition between methods as there is goods on the market, with a far smaller multitude of “winning” decision-making methods emerging from the process. Some of these methods are for sale in the form of books, courses, and paid expertise, or granted in exchange for supporting an idea and its leaders, but many more can be gleaned from pure observation and trial-and-error. Analogous to choosing which goods to consume, choosing a decision-making method depends strongly on the characteristics and subjective preferences of the agent.
Making it cheaper to engage in some methods of decision-making is analogous to subsidizing some goods. Agents who would have preferred “bad” decision-making methods will find their preferred methods costlier, or choose less preferred methods by default at a loss of utility unbeknownst to them. There’s no analogy to neoclassical social welfare in an ecologically rational regime, but we can say that a transfer of utility has taken place from agents who would have made the choices policy makers deem as “bad” to agents policy makers deem as “good.” Therefore nudges, as implemented in the real, ecologically rational world, have real, negative effects on some agents in order to benefit others.
Candelo, R., Wagner, R. (upcoming). Pareto’s Theory of Action and Behavioral Economics.
Gigerenzer, G., & Todd, P. M. (1999). Simple heuristics that make us smart. Oxford University Press, USA.
Hansen, P. G., & Jespersen, A. M. (2013). Nudge and the manipulation of choice: A framework for the responsible use of the nudge approach to behaviour change in public policy. Eur. J. Risk Reg., 3.
Orwell, G. (2013) (originally, 1946). Politics and the English language. Penguin UK.
Simon, H. A. (1977). ‘The logic of heuristic decision-making,’ in R. S. Cohen and M.W. Wartofsky, eds., Models of Discovery. Boston: D. Reidel.
Simon, H. A. (1996). The sciences of the artificial. MIT press.
Thaler, Richard, Sunstein, Cass. (2008) Nudge: Improving decisions about health, wealth, and happiness. Yale University Press.