Methods and Problems in Business Cycle Theory

Robert E. Lucas, Jr.

Journal of Money, Credit and Banking, Vol. 12, No. 4, Part 2: Rational Expectations. (Nov.,1980), pp. 696-715.

Robert Lucas begins his critique with an observation; the intention behind Business Cycle Theory is to create “models” of reality, rather than actually model reality to mimic it. It is the purposes behind each pursuit that distinguish the usefulness of RBC models, from any discussion derived from actual economic behavior. The purpose of models is to help us understand the mechanism that exist in reality; to confuse this with a purposed aimed at “realism” is to denigrate or ignore the inherent usefulness of RBC models. Lucas complains that “the theory is not being effectively used to help us to see which opinions about the behavior of actual economies are accurate and which are not.” Its not so important to be a “real” interpretation, as it is to be a “better imitation.” The question arises: “what makes a better imitation than mimicking what is real?” and this is the question that Lucas is setting out to answer.

Lucas points to a few key developments that have allowed RBC models to make a contribution. First, is what Lucas calls a “purely technical” development that enlarges our “abilities to construct analogue economies.” He is of course, talking about formalized mathematical models that allow us to investigate situations or cause-effect relations that would otherwise be to costly to observe in “real economies.” Second, Lucas points us to the fact that RBC models are built in response to different question. Unlike the journalist for whom every new story is a phenomenon based on a new theory, RBC models approach the world with unbending, unchanging formalism. This is probably an overstatement, but it illuminates the need and use of formal models in RBC theory. Third, as more economist and mathematicians specialize in the field, it is no wonder that the field requires more specialization to understand and utilize it in itself. This has caused a misunderstanding amongst outside observers, such that historians have attempted to “understand developments of monetary economics in terms entirely internal to the sub-discipline.”

The author spends some time criticizing Keynes, only in the end to show how much “Keynesianism” in all its advancements are indebted to the neo-classical formalized approach to economics. In this context, Lucas continues with the history of RBC theory in the context of general equilibrium theory, a mode of thought that begins, so he contends, with Alfred Marshall. Paul Samuelson advanced the course of finding a neo-classical synthesis with Keynesianism, by supplying the “main ingredients” for a mathematically explicitly theory of general equilibrium; “an artificial system in which households and firms jointly solve explicit ‘static’ maximum problems.” But, as Lucas points out, business cycles are not static. To correct for this, Samuelson included additional “free” parameters that still had an “equilibrium” point to appease the neo-classicals, while maintaining enough parameters to show a wide variety of possible paths coming as a result of an exogenous shock. There was equilibrium with dis-equilibrium; a synthesis.

Real business cycle theory developed out of the new classical tradition and attempts to explain economic activity fluctuations in a world without coordination failures, “price stickiness, waves of optimism or pessimism, monetary policy, or government policy in general.” Therefore, in contrast to the Keynesian approach, RBC theory describes all fluctuations as real fluctuations in supply, and hence every fluctuation is a new equilibrium rather than a point of dis-equilibrium.

An important historical moment for RBC models came with the publications of Ragmar Frisch (1933) and Eugen Slutzky (1937), who distinguished between impulse and propogation mechanisms. Impulse mechanisms are what initially cause a “variable to deviate from its steady state value, whereas propagation mechanisms: “causes deviations from the steady state to persist for some time.” All deviations from the steady state arise from exogenous productivity shocks, in contrast to the Keynesian theory in which demand-side fluctuations account for output fluctuations. This is why RBC theory designates little importance to other impulse mechanisms like tax rates, monetary policy, and preferences, all of which are very important in Keynesian theory.

The growth of output in Robert Solow’s model (1956, 1957) is dependent upon the capital labor ratio and the “solow residual”, which embodies the “technical possibilities” that augment the productivity of labor. This residual is stochastic, exogenously determined, and accounts for almost half of the growth in output. RBC theory is then founded in a neo-classical growth model augmented by stochastic shocks to productivity. Stadler points out that “once one incorporates stochastic fluctuations in the rate of technical progress into the neoclassical growth model, it is capable of displaying business cycle phenomena that match reasonably closely those historically observed in the USA.” (pg 1753).

Stadler then points out a few common features of RBC models. 1) they have their foundations in micro-economic behavior. The behaviors of representative agents are aggregated to determine the overall impact on the economy. Furthermore, all agents have rational expectations and all markets are instantly clearing with no information asymmetries. For this reason, all RBC theories are built within the framework of competitive general equilibrium models. 2) All firms and households are maximizing agents. 3) Cycles are dependent upon the shocks to technology that impact the production function—therefore, rather than deviations, these productions functions have been redefined so that all cyclical behavior is still an equilibrium condition. 4) There are propagation mechanisms that magnify or mitigate the exogenous shocks to labor-augmenting technology.

Kydland and Prescott (1982) present a classic example of an RBC model in which all volatility in output can be accounted for by stochastic changes in productivity. Like most RBC models, it has the problem that it overpredicts the correlation between productivity and output. Because these models are developed and tested around the same data, the standard deviation of output predicted in the model is identical with the data.