Bounded rationality and the need for low-cost heuristics

6 H. Hayakawa J. of Economic Behavior Org. 43 2000 1–34 abstracted by way of interdependence via reference groups, in which social status ranking of such groups gives direction to the emulation–avoidance motives. In Section 4, holding that physical and social wants are not commensurate generally, we consider a sequential satisficing decision rule over physical wants as an alternative to the usual utility theory. In Section 5, we discuss properties of a preference relational system that follows from this sequential rule and propose a two-step procedural model of consumer choice where social want appears in the second step as an instrument of indeterminacy reduction. In Section 6, we present a formal model of interdependence via reference groups and define what social want constitutes. We show that a reaction function over relevant social groups and information on the whereabouts of social norms together make it possible to quantify the social want-satisfying property; this quantification eliminates or substantially reduces the indeterminacy of choice that remains in the first step. We show that a norm-oriented choice behavior can be rationalized by a norm-guided ordering of choice objects. In Section 7, we relate our model of interdependence to the work of Duesenberry 1949 and Leibenstein 1950, and Section 8 concludes the paper.

2. Bounded rationality and the need for low-cost heuristics

It is now widely recognized that the decision-making environment including the internal psychology of decision makers is short of being perfect. Various elements can account for such imperfection. 1 Decision makers’ cognitive and computational capacities are signif- icantly bounded Simon, 1955, 1959. 2 The severity of this limitation is compounded by the fact that the time endowment is fixed, so that all activities including cognitive ones compete for the use of this endowment Becker, 1965; Linder, 1970. 3 Decision makers seldom have perfect information about choice alternatives, but information gathering and processing, like any other activity, is costly in time and other resources Stigler, 1961. 4 Many decision-making situations involve elements of risks, so that the anticipated con- sequences of decisions can only be assessed in probability terms, subjective or objective. 5 If decision-making situations are imbued with uncertainty so that the means–end re- lationships that are necessary for economic calculations break down, the consequences of an action cannot be assessed even in probabilistic terms Keynes, 1921; Knight, 1921. The total failure of the objective rationality in the face of uncertainty leaves a gap that is beyond the cognitive capacity of decision makers Beckert, 1996. While various models or apparatuses have been developed to cope with some of these limitations, we focus here on decision makers’ motives to save the resources that would otherwise be required to solve complex choice problems. In our view, the key to the success of a decision maker in coping with the limitations of the decision-making environment is the availability of cost-absorbing choice mechanisms or devices that are rooted in social capital and order. For sure, such mechanisms have to be simple enough to afford a painless adaptation to the decision-making environment; at the same time they should reflect the social structure of this environment. Specifically, what is needed is a socially meaningful choice mechanism that absorbs risk and uncertainty, reflects costs of acquiring and processing information, reduces the pressure of time constraints, and eases the computational pain of problem solving. If one’s cognitive capacity is limited and if H. Hayakawa J. of Economic Behavior Org. 43 2000 1–34 7 there are limits to decision making costs that can be borne, a resulting decision mechanism will be of a simple kind. Assume, for a moment, that there are neuron locations in human brains to process in- formation and that two kinds of information occupy such locations: a decision rules and transformation, and b data about choice objects and internal states of a decision maker. Naturally, an economizing problem arises over such locations. The more complex are the problem-solving algorithms, the less room is available for data, and the more complete are the data, the less room is available for complex algorithms. The total computational capacity has an upper limit that is short of global optimization. Given costs of obtaining data and given time constraints on how long is permitted to solve a problem, further limits are placed on the complexity of problem solving. After all, the costs of decision making should not exceed the net value of choices made. Once costs of reaching a decision are taken into account, the need for a simple-choice mechanism that relies on low-cost heuristics cannot be ignored. Given social capital that has been accumulated in the form of life styles and given social and cultural order that has transformed the environment into a well-directed social field so that socially desirable ends and means can be identified, it is only natural for individuals to search for low-cost heuristics in the life styles of their relevant social groups. Viewing the consumer choice process as one of utilizing low-cost heuristics is not an escape from the conventional rationality hypothesis. Rather, it is best viewed as an extension of this hypothesis when the decision-making environment is imperfect. The central issue is still one of cutting costs of problem solving as we focus on the use of socially desirable, cost-saving means see Vriend, 1996 for a much broader interpretation of rationality as pursuance of self-interest. What is novel of this view is that it interprets the life styles of social groups as social capital that has been accumulated through error-learning processes and relates this capital to human behavior of bounded rationality.

3. Social capital as sources of low-cost heuristics