ORMAT Cognitive Economics
February, 2007
Miles Kimball and Robert Willis
I. Human Capital Theory
The development of economics in many topical areas often follows a progression from treating a model element as a black box viewed from the outside to peering into the mechanism inside the black box. In black box treatments, the cognitive and informational nature of human capital and technology is pushed into the background. For human capital and technology, looking deeply into the black box involves looking at what is going on inside human minds. Thus, we see Cognitive Economics as a logical extension and broadening of human capital theory—including both those types of human capital that are acquired and those types that one is born with.
II. Cognitive Economics
Cognitive Economics is defined as the economics of what is in people’s minds. In practical terms, this means that cognitive economics is characterized by its use of a distinctive kind of data. This includes data on expectations, hypothetical choices, cognitive ability, and expressed attitudes. Categorizing a field of economics by the type of data used and theorized about makes some sense from a practical point of view because working with a particular type of data often requires some specific human capital. A simple typology of areas of economic research by type of data addressed would be something like the following. (Note that each of these areas of economics is concerned with data on naturally occurring market choices and time allocations as well as its distinctive type of data.)
· Traditional Economics: data on naturally occurring market choices and time allocations only.
· Experimental Economics: data on choices in artificial situations with real stakes.
o field experiments
o lab experiments
· Neuroeconomics: data from brain imaging and other ways of measuring brain activity, data on eyeball orientation, and data on other physiological measures such as skin conductance, muscle activation, or hormone levels.
· Genetic Economics: data on genes.
· Cognitive Economics: data on hypothetical choices, psychometric data, and self-report data on mental contents.
o survey measures of expectations
o survey measures of preference parameters
o direct measures of intelligence
o direct measures of decision-making skill
o self-reported emotions, including self-reported happiness
o survey measures of beliefs about how the world works
We return shortly to a discussion of the different branches of Cognitive Economics, each defined by use of a particular subtype of data.
The name “Cognitive Economics” is coined by analogy to “Cognitive Psychology,” the area of psychology that examines internal mental processes such as problem solving, memory and language. Historically, Cognitive Psychology was a departure from Behaviorism--the insistence that only outward behavior was a legitimate subject of study.[1]
The name “Cognitive Economics” might initially sound as if it might be yet another synonym for Behavioral or Psychological Economics. But although there will be overlap,[2] we mean something different. The most obvious difference is that Cognitive Economics is narrower. Psychological Economics addresses a huge range of issues and cuts across all of the data types listed above, while Cognitive Economics focuses primarily on innovative kinds of survey data, along with lab data of the same basic type.
Second, important pieces of cognitive economics are inspired by the internal dynamic of economics rather than by psychology. As examples, in addition to the interest in intelligence measures that arose out of human capital theory the importance of expectations and preference parameters in macroeconomics has spurred a desire for direct measurement of expectations and preference parameters. Finally, we think it is fair to say that Psychological or Behavioral Economics has been to an important degree a school of thought as well as an area of study. In coming up with a definition of Cognitive Economics, we set as a requirement for ourselves that we only indicate an area of study, not a particular viewpoint. To make this point clear, a research agenda arguing that, in fact, data on mental contents and hypothetical choices was unreliable would be part of Cognitive Economics.
Having defined the field of Cognitive Economics in as neutral a way as we know how, we will give our own opinion on existing research and future directions in Cognitive Economics, organized around three themes: using data on hypothetical choices and mental contents (1) to identify individual heterogeneity, (2) to revisit welfare economics and (3) to study bounded cognition. Data on hypothetical choices and what is in people’s minds has obvious relevance to these three themes. Bounded cognition also raises some important theoretical issues that we will discuss.
III. Identifying Individual Heterogeneity
To the extent one is interested in heterogeneity of what is in people’s minds, trying to get “direct” measures of what is in people’s minds is one obvious strategy. Relatively direct measurement of what is in people’s minds encompasses what has become quite a robust literature. The common use of IQ measures in labor economics is one of the simplest examples. Survey measurement of expectations at the individual level has a long and distinguished history. There has been a great deal of recent interest in the survey measurement of preference parameters, using hypothetical choices or self-assessments. Happiness research, with its use of self-assessed mood, accounts for scores, if not hundreds, of economics papers. Direct measures of the implicit theories people have of how the economy works are harder to come by, but these, too, exist.
IV. Revisiting Welfare Economics
Concern with policy and overall welfare motivates some of the concern with measuring preference parameters that we discussed above in the context of identifying individual heterogeneity. In particular, the population distributions of the elasticity of intertemporal substitution, labor supply elasticities, and interpersonal dependencies in preferences have important implications for the welfare effects of capital and labor taxation. In addition to data on preferences based on hypothetical choices, there has been considerable interest in using data on self-reported happiness to study welfare issues.
The use of self-reported happiness to study welfare issues illustrates a key methodological issue in Cognitive Economics. Whenever a new measure is used, its relationship to standard concepts of economic theory is at issue. For example, since welfare economics is based on preferences, with the objective of getting people as much as possible of what they want, in order to use self-reported happiness to address welfare issues, it is crucial to establish the relationship between self-reported happiness and preferences. The most common assumption in the economic literature using self-reported happiness has been that self-reported happiness is equal to some version of utility. If self-reported happiness were, in fact, tightly linked to preferences in this way, its importance for welfare economics would be of enormous importance.[3] In Kimball and Willis (2007), we argue at length that self-reported happiness does not behave like utility, but has a more complex relationship to utility. In particular, to explain the data, we suggest that a large component of self-reported happiness depends on recent innovations in lifetime utility. Whenever people receive good news about lifetime utility, self-reported happiness temporarily spikes up; whenever people receive bad news about lifetime utility, self-reported happiness temporarily dips down. This means that while it is questionable to use the level of happiness to infer preferences, the dynamics of happiness are informative about preferences and so can be used to inform welfare economics. This is particularly valuable in areas where market choices are not fully informative about preferences. For example, the dynamics of happiness can be used to study interpersonal dependencies in the utility function, preferences over events largely outside of one’s control such as the death of one’s spouse, and preferences over nonfinancial aspects of public policy.
V. Studying Bounded Cognition
Taking a simplified view of information as recorded data and data summaries, for the purposes of this paper we will call all of the other operations of the human mind besides the bare recording and accessing of information “cognition,” without the finer distinctions that psychologists often focus on. Moreover, to avoid the judgment Herbert Simon’s phrase “bounded rationality” can inadvertently suggest, we will refer instead to “bounded cognition.”[4] Bounded cognition means something more than just imperfect information, it means finite intelligence, imperfect information processing, and decision-making that is costly. Bounded cognition is the second key theme we see for Cognitive Economics.
If true, explanations based on bounded cognition have enormous practical consequences and policy implications. In particular, well understood bounded cognition would imply that even in the absence of externalities, welfare can often be improved by economic education, setting up appropriate default choices for people, or providing disinterested, credible advice. By contrast, if true, explanations of puzzling behavior on the basis of individuals maximizing exotic preferences imply that welfare improvements must come in the standard way from addressing externalities, or in the case of inconsistent preferences, by taking sides in an internal conflict. Once puzzling behavior that is difficult to explain on the basis of standard economic theory is identified, it is hard to think of a more important question than whether people behave that way because they want to, or simply because they are confused.
Our perspective on bounded cognition is close to that of the excellent discussion by Conlisk (1996). So we will limit ourselves to highlighting a few of what we consider the most salient points, with our own spin.
A. The Reality of Bounded Cognition. The first key point is the reality of bounded cognition. Although the inadequacies of our current tools can make it hard to study bounded cognition theoretically, the claim that human intelligence is finite and that finite intelligence matters for economic life is not really controversial.[5] Even those economists whose opinion of their own intelligence is unreasonably exalted are regularly reminded by their students and coworkers that not everyone has unlimited intelligence. Many people pay substantial sums for financial advice even aside from commissions on transactions. Even people who have low wage rates, so that their time is less expensive, often pay others to do their tax returns. Even if one considers talking in a courtroom a special skill that is not just a matter of intelligence, people pay a lot of money to other lawyers who only read law books and extract the relevant information. If everyone had infinite intelligence one could understand the law books on one’s own, and this would only make sense if one’s wage rate were higher than the lawyer’s wage rate. With everyone having infinite intelligence, even finite reading speeds would not give trained lawyers enough of an edge for them to charge what they do.
One of the most important economic manifestations of finite intelligence is the expensive and time-consuming acquisition of human capital. Most obviously, the large amount of resources devoted to mathematical education and research would make little economic sense in a world in which everyone had unbounded cognition. Mathematicians spend their entire careers discovering and teaching things with very little informational content about the external world—things that would be easily deducible by anyone with infinite intelligence. In other subject areas, education may involve a significant amount of straightforward information transfer. But in most areas, the acquisition of useful habits of thought is at least as important.[6] Teaching students to “think like an economist” is itself somewhere between information transfer and the inculcation of one of those useful habits of thought. Below we present a model of the effects of being taught a standard model of portfolio choice. This is an area where we think many individuals are confused and where making the right choices is important. The model is relatively simple, but breaks some of the normal theoretical taboos. The remainder of this section makes the case for why it is sometimes necessary to break these taboos.
B. Difficulties in Studying Bounded Cognition with Standard Theoretical Tools. One key reason it is not easy using our standard theoretical tools to model bounded cognition is the “infinite regress” problem emphasized by Conlisk (1996). The infinite regress problem afflicts models that assume a cost of computation or other decision-making cost. The problem is that figuring out how much time to spend in making a decision is almost always a strictly harder decision than the original decision. In particular, one would typically need to know the right choice to the original decision in order to calculate the value of making additional computations in order to make the right choice instead of another choice. Given costly decision-making, the agent faces a serious issue of figuring out whether it is worth thinking carefully about the original problem, which leads to the issue of figuring out whether it is worth thinking carefully about thinking carefully about the original problem, and so on.
Costs to decision-making are a natural enough assumption for economists that a substantial percentage of all applied economic theory papers might include them, if it were not for the infinite regress problem. Finessing the infinite regress problem somehow is essential if economists are to develop effective theoretical tools for studying bounded cognition. We see several feasible strategies for getting around the infinite regress problem—every one of which requires breaking at least one inhibition shared by many economists. Least transgressive are models in which an agent sits down once in a long while to think very carefully about how carefully to think about decisions of a frequently encountered type. For example, it is not impossible that someone might spend one afternoon considering how much time to spend on each of many grocery-shopping trips in comparison shopping. In this type of modeling, the infrequent computations of how carefully to think about repeated types of decisions could be approximated as if there were no computational cost, even though the context of the problem implies that those computational costs are strictly positive.
A second strategy is to give up on modeling bounded cognition directly and use models of limited information transmission capacity as a way of getting agents to make more imperfect decisions. In other words, one can accept the fact that our standard tools require constrained optimization with its implication of infinite intelligence somewhere in the model, but handicap agents in the model by giving them a “thick skull” that is very inefficient at transmitting information to the infinitely intelligent decision-maker within (that is, the perfect constrained optimizer within). This is a way to interpret the program of Sims (2002) program that disconnects the implied transmission bit-rates from anything in the external world, since low bit-rates would only be a metaphor for bounded cognition.[7]