Rationality as a Process

Individual decision-making is not adequately portrayed by focusing on static rationality properties. The static approach can mistake rationality-in-process for bounded rationality or irrationality. We consider a sampling of intellectual frameworks that address decisionmaking rationality as a process, including intrapersonal arbitrage, Wicksteed’s principle of price, dialectical reasoning, and errordriven learning. We conclude that the approach to normative analysis shared by both neoclassical and behavioral economists is not the only possible one and that, in fact, it misses an important aspect of human decision-making. Evaluations based on the static approach are at best incomplete and likely to be misleading. The rationality norms accepted by neoclassical economists and applied by behavioral economists are static. The individual is either “rational” or he is not. His preferences are either completely ordered, consistent through time, and transitive, or they are not. In the neoclassical version of rational choice theory, actual behavior conforms to the static axioms of preference,1 whereas in the behavioral version it falls short. In each conception, the individual’s rationality is described in terms of preferences and beliefs at a given point in time. For some types of analysis – like simply predicting the direction of change in an endogenous variable – this may not be a problem. On the other hand, in welfare and prescriptive analysis it can be inappropriate. If we only take a snapshot of the individual at a moment in time, we will fail to see how the individual’s decision-making evolves. We may simply have of picture of it in an inchoate state. This is especially relevant for policies that aim to intervene in decision-making. If individuals are in the midst of a decision-making process, then the evaluation of their behavior even at a 1In brief, the most important rationality properties are completeness and transitivity of the preference ordering. In addition, actual behavior must be consistent with the preference ordering. ISSN 2326-6198; DOI 10.1561/105.00000098 ©2018 M. J. Rizzo and G. Whitman 202 Mario J. Rizzo and Glen Whitman point in time should be made through the lens of that process. To put matters another way, the unit of evaluation should not be a part of the process but the process as a whole. The purpose of this article is to argue that individual decision-making is not adequately portrayed by focusing on static rationality properties. The static approach can mistake rationality-in-process for bounded rationality or irrationality. In what follows we consider a sampling of intellectual frameworks in increasing order of their “radicalness” in addressing decision-making rationality as a process. Some may be familiar to the reader; others less so. Our fundamental purpose is to show that the approach to normative analysis shared by both neoclassical and behavioral economists is not the only possible one and that, in fact, it misses an important aspect of human decision-making. Evaluations based on the static approach are at best incomplete and likely to be misleading. 1 Intrapersonal Equilibrium There are many manifestations of equilibrium in economics.2 Among the least explored is intrapersonal equilibrium. What is sometimes called the “equilibrium of the individual” generally refers to the adjustment of an individual’s various actions into a coherent whole or, viewed intertemporally, parts of a single plan. This equilibrium is said to be attained immediately. However, in another sense, it is not really attained at all but is “simply” a consistency property, and hence equivalent to a definition of abstract rationality. Accordingly, for standardly rational agents, in either interpretation, (a) every decision is made without the elapse of time and (b) all of these decisions are always mutually consistent. Neoclassical economists accept this both normatively and descriptively, while behavioral economists accept it only normatively. There is good reason to believe that this state of affairs is not likely ex ante. To assume otherwise would be to assume that an individual would have settled, in effect, on a life plan from the beginning of their consciousness. Such an intrapersonal state of equilibrium is analogous to the Arrow-HahnDebreu model of general equilibrium that incorporates demand functions for all commodities in every contingent state of the world, at every time and in every location. The equilibrium of the individual excludes a trial-and-error process by which individuals discover in particular cases what they want and how to get it (Dold, 2018; Plott, 1996). Furthermore, the notion of a fully consistent equilibrium ignores the costs of discovering inconsistencies and then resolving them. There is no reason that an individual can or should try to achieve such consistency, as the cognitive cost of doing so likely exceeds the benefits. “The expected marginal benefit of discovering and/or forming these preferences presumably declines as the 2For a discussion of different types of equilibrium concepts see Tieben (2012). Rationality as a Process 203 compared options get further from one’s likely future experience” (Whitman and Rizzo, 2015, p. 419). At some point, then, achievement of additional degrees of consistency will fail the cost-benefit test. 1.1 Intrapersonal Equilibrium: Austrian Background Precise adherence to formal features of economic rationality was considered unrealistic in the Austrian economics tradition, even before the formalization of the axioms of rational preference and choice. One of the early criticisms of marginal utility theory was that it presupposed deliberations about the value across goods on various margins that are too complicated for real-life individuals to perform. In the late nineteenth century, Böhm-Bawerk (1959 [1889]) replied to this criticism by saying that the individual will, in effect, satisfice, that is, “his performance will do well enough for his purposes” (202). Generally, precision is not to be expected because this can only be achieved at the cost of significant scarce mental effort.3 Therefore, even in a state of complete individual equilibrium, there is not likely be precise determination of marginal valuations. In his criticism of indifference analysis, the interwar Austrian Mayer (1994 [1932]) pointed out that Pareto’s characterization of the individual agent was based on an equilibrium construct in which all of the relevant “magnitudes operative in genetic-causal sequence [are portrayed] as if these existed together at the same time. A state of affairs is synchronized in the ‘static’ approach, whereas in reality we are dealing with a process” (92). While Mayer did not carefully distinguish between the equilibrium of the individual and the determination of equilibrium prices, his criticism is quite general. For Mayer the idea of an individual with a preference or indifference ordering that satisfies all the requirements of consistency ex ante would confuse the possible result of a process with a collection of synchronic characteristics.4 Robbins (1935) characterized one of the central consistency criteria of neoclassical rationality, the transitivity of preferences, as a claim “that in a state a perfect equilibrium the possibility of advantage from further ‘internal 3Böhm-Bawerk suggests a useful maxim: “In really important things, be really exact; in moderately important things be moderately exact; in the myriad trifles of everyday economic life, just make the roughest sort of valuation” (202). This maxim captures the idea that the value of precision may be different in different circumstances and so the degree of precision in decision-making will be contextand content-dependent. 4In fairness to Pareto he sometimes did recognize the existence of a completed process. “A man who buys a certain food for the first time may buy more of it than is necessary to satisfy his tastes, price taken into account. But in the second purchase he will correct his error, in part at least, and thus, little by little, will end up procuring exactly what he needs. We will examine this action at the time when he has reached this state.” (Pareto, 1971 [1906], 103, emphasis added). Pareto extended this idea to reasoning itself: “Similarly, if at first he makes a mistake in his reasoning about what he desires. He will rectify it in repeating the reasoning and will end up making it completely logical” (103). Pareto’s analysis is of the completely logical state. 204 Mario J. Rizzo and Glen Whitman arbitrage operations’ is excluded” (92). This internal equilibrium is analogous to the perfectly competitive market equilibrium insofar as all problems due to lack of foresight, costs of decision-making, or insufficient awareness of opportunities for gain are absent. For the individual as for the market, at any point in time, the process of (internal) arbitrage may not be complete. Nevertheless, economics, in Robbins’s view, is not “limited to the explanation of situations in which action [is] perfectly consistent” (92). Indeed Robbins thought that “it is only in terms of irrational choice, that many of the more complex situations which economics has to study can be explained” (1934, 101). Here, Robbins uses the word “irrational” to mean intransitive, yet he is arguing for adopting a more inclusive model that does not require transitivity. Modern-day economists working in the Austrian tradition have often expressed dissatisfaction with neoclassical models built on the assumption of equilibrium. The equilibrium “solution” to such a model is a state of affairs that satisfies all defining assumptions of the model at once. It is a state of rest with no internal tendency to change. The classic supply-and-demand model is a simple example: at the equilibrium price and quantity, all expectations are fulfilled. The situation is self-reinforcing and therefore, absent any exogenous changes