Lucas’s assessment of the General Theory
To Lucas, Keynes ought to be honoured for the role his ideas have played in the expansion of socialism rather than for his theoretical contribution. The latter, Lucas wrote, “is not Einstein-level theory, new paradigm, all this” (2004: 21).
In Lucas’s opinion macroeconomics started off on the wrong foot by being Keynesian. He regretted that Keynes did not try to make Walras’s static model dynamic, as Hayek had suggested (before changing his mind). In Lucas’s view, Keynes abandoned this line of research to tackle the easier task of demonstrating the existence of unemployment at one point in time, that is, in a static framework.Lucas also criticised Keynes for having discarded what he called the “equilibrium discipline”, a basic premise by which, he felt, economists should abide when constructing theories. It consists of two postulates: (1) that agents act in their own self-interest and (2) that markets clear (Lucas and Sargent 1979 [1994]: 15). These postulates are deemed to constitute a universal requirement, rather than being linked to the specific purposes of particular models. That is, they are viewed as constituent parts of neoclassical theory, which in turn is equated simply with economic theory. The counterpart of the equilibrium discipline is the rejection of the disequilibrium notion. According to Lucas, by betraying this equilibrium discipline, Keynes gave an example of “bad social science: an attempt to explain important aspects of human behaviour without reference either to what people like or what they are capable of doing” (1981: 4). Lucas accepted that Keynes’s lapse from the equilibrium discipline was understandable in view of the apparent contradiction between cyclical phenomena and economic equilibrium, but it remains true, he claimed, that in retrospect it prompted a long detour in the progress of economic theory.
Turning now to Lucas’s assessment of Keynesian economics, as distinct from the economics of Keynes, the following points should be brought out. First, Lucas praised Keynesian macroeconomics for having engaged in econometric modelling and empirical testing, in contrast to Keynes’s reasoning in prose. The Keynesian macroeconomic models were the first to attain this level of explicitness and empirical accuracy; by doing so, they altered the meaning of the term “theory” to such an extent that the older business cycle theories could not really be viewed as “theories” at all (Lucas 1977 [1981]: 219).
Second, Lucas took a strong stance on the Phillips-curve controversy. The latter opposed Keynesians and monetarists a la Friedman: Keynesians defended the stable Phillips curve allowing for a trade-off between unemployment and inflation, while monetarists argued for the natural rate of unemployment hypothesis. Lucas’s distinct contribution to the debate was to provide stronger foundations for Friedman’s insight in his path-breaking article, “Expectations and the Neutrality of Money” (Lucas 1972 [1981]). The 1970s stagflation episode, Lucas claimed, demonstrated the failure of Keynesian activation policy, while confirming Friedman’s predictions.
The most influential of Lucas’s judgments about Keynesian theory is the famous “Lucas critique” (Lucas 1976 [1981]). It asserts that the econometric models of the time, all derivatives of the Klein-Goldberger model, could not serve their avowed purpose of comparing alternative economic policies because the coefficients of the models were estimated by econometric methods (rather than being derived from theory), and their numerical values were independent of any changes in institutional regime that might occur. Therefore, the model-builder will miss the fact that agents would change their decisions when faced with a policy change. As a result, a model of the economy estimated at a period during which a particular institutional regime held sway could not but provide inadequate information for assessing what might occur under a different regime. According to Lucas, only deeper, “structural models”, that is, derived from the fundamentals of the economy, agents’ preferences, and technological constraints, were able to provide a robust grounding for the evaluation of alternative policies.
Lucas’s critique was part and parcel of the rational-behaviour hypothesis introduced by Muth (1961). It was meant to capture the idea that economic agents ought to be ascribed the ability of guessing (on average) the outcome of the market in which they are participating, conditional on the information available. That is, their subjective expectations about any coming event should coincide (on average) with the model-builder’s objective expectations. The change involved is radical, a move away from a backward looking towards a forward-looking depiction of economic agents.