Representative Agent

views updated

Representative Agent

The representative agent is a device used by economists to model the macroeconomy. The general idea is to solve a well-specified microeconomic problem, and then use the relationships between the variables in that model as a description of the macroeconomy. For example, we can model the decisions that would be made by a person who would like to consume goods and enjoy leisure, but can only purchase goods after earning wages through working. Additionally, we can posit a specific production technology that determines how many goods are produced when the person works. From this problem, we can solve for the relationship between the state of the economy and the person’s decision on how much to work and consume. A representative agent model would then take these relationships, which were derived for an individual, and use them as an exact specification of the macroeconomy. Thus, if the representative agent would increase consumption by 10 percent in response to a 5 percent increase in the level of technology, it is assumed that aggregate consumption will also rise by 10 percent in response to the same change.

The earliest form of a representative agent was in Alfred Marshall’s Principles of Economics (1890). Marshall constructed a representative firm as a means of specifying an industry supply curve. This representative firm was vigorously criticized by economists at the time, and the notion vanished from economics.

Representative agent models were resurrected by macroeconomists in the 1970s and became the predominant means of studying the macroeconomy in the 1980s. Surprisingly, nobody has ever written an explicit defense of this technique. The underlying rationale seems to be a desire to provide microfoundations for macroeconomic research. If we acknowledge that macroeconomic relationships are the result of a vast number of decisions made by individuals, then it might be desirable to have our macroeconomic models derived from relationships found in microeconomic problems. Because it is impossible to model the decisions of every individual in an economy separately, some shortcut is needed. Hence the representative agent is a stand-in for an average person, and the macroeconomy is assumed to behave, on average, like the average person.

Representative agent models have been roundly criticized as being an improper means of studying the macroeconomy for several reasons. It is simply not the case, even in very simple economies, that the aggregate of many decisions is exactly the same as the decision made by the representative agent. Representative agent models are incapable of capturing the effects of heterogeneity and interaction among agents, nor can they capture the effects of problems related to asymmetric information or strategic interactions. Fundamentally, representative agent models suffer from the fallacy of composition (that what is true of the part is true of the whole). The extensive criticisms demonstrating that representative agent models are neither good nor useful means of studying the macro-economy have never been addressed by those who use such models. Yet, the use of the representative agent model continues apace.

SEE ALSO Macroeconomics; Marshall, Alfred; Microfoundations