Robert Lucas led the macro academy’s rejection of both the keystone status of meaningful wage rigidity and the consequent centrality of nominal demand in stabilization analysis. His influence has been enduring despite a remarkable fact. Lucas’s seminal contribution, i.e., the application of rational expectations to nominal-wage determination, is for the largest and most relevant class of labor pricing simply wrong. It is a familiar story to readers of the GEM Blog. Lucas was badly misled by his core, unexamined assumption that confinines rational price-mediated exchange to the marketplace.
The Irrational Rational-Expectation Phillips Curve
The reason that much of Lucas’s anti-Keynesian thinking is wrong is not because of his strong, and often criticized, commitment to rationality in economic modeling. Such objections are badly off-base. The GEM Project’s staunchness in defense of rational decision-making is at least as strong as Lucas’s.
Instead, his famous analysis breaks down because of his failure work through the implications of a critical assumption. He assumes, for convenience, that labor pricing in modern economies is conducted in a spot market. Everybody knows that most employees are more permanently attached to their employers, with tenures spanning multiple annual wage adjustments. It turns out that, given the absence of spot-market labor pricing, the rational-expectations Phillips curve is not compatible with rational behavior.
In the 1970s, Lucas capped his interest in writing down micro-coherent labor-supply equations by imposing John Muth’s rational expectations on the then-consensus Early Keynesian “adaptive expectations” Phillips curve. The EK version is:
w(t)=ao+a1(UN-U(t))+a2pL(t)+e(t),
where w is the rate of change of nominal wages, UN is the natural rate of unemployment, U is the actual rate, pL is lagged consumer price inflation, and e is an error term. The Lucas version is:
w(t)=ao+a1(UN-U(t))+Etp(t+1)+ε(t),
where E represents expectations rooted in the cost-effective use of available information. The innovation is of course the inflation term (pL(t)→Etp(t+1)). Lucas confidently attacked the EK catch-up version of adjusting nominal wages for inflation as irrationally leaving information on the table. He believed his expectations version to be self-evidently consistent with optimizing behavior. His forward-looking take inflation adjustment quickly became mainstream.
The GEM Project easily shows that Lucas’s inflation-term innovation is inconsistent with rational employee-employer behavior and badly misleads on how wages are actually determined. In the real world, the rational payment a premium relative to market-opportunity wages compensates workers for losses implicit in any lagged adjustment for inflation, allowing both employers and employees to costlessly avoid the significant costs of informing workable rational expectations of future inflation. Both sides are better off using the more efficient catch-up approach. The GEM argument is supported by the actual behavior of practitioners, who always use catch-up. The best-practices literature that demonstrates the disuse of expectations in annual wage adjustments is characteristically ignored by Lucas and other mainstream theorists.
Chapter 4 of the Website’s e-book carefully works through the rationality of catch-up, and the irrationality of expectations, in the periodic adjustment nominal labor pricing for price inflation. It further demonstrates that the forward-looking Phillips curve has been a badly inaccurate guide to understanding the stagflation decade – the massive market failure that rational expectations was supposed to have explained.
More generally, Lucas’s irrational Phillips curve played a crucial role in upending then-mainstream Early Keynesian macro thinking, relegating Samuelson’s Neoclassical Synthesis to the dustbin. The resulting vacuum in stabilization-relevant research motivated the doomed multi-decade quest to figure out how to restore policy-relevance to macro theory now organized by market-centric general-market-equilibrium. The GEM Project helps solve that fundamental problem by constructing a truly microfounded Phillips curve:
w(t)=bo+b1(UN−U(t))+b2pŁ(t)+b3(EtpN(t+1)−Et-1pN(t))+ε(t),
where pN denotes the central bank’s inflation objective. (For elaboration, see Chapter 4 in the Website’s e-book.)
Notably overlooked in the mainstream analysis here is that the decades-long, often bitter debate over how to adjust wages for inflation has always been wrong-headed. The issue is not particularly important. The much more critical issue is making sense out of the relationship between unemployment and labor pricing. What matters, if for example involuntary job loss is to result from nominal demand disturbances, is how to rationally suppress wage recontracting. That crucial question, in its rational-behavior context, is uniquely answered in the GEM Project’s generalized-exchange macroeconomics.
Quite a Man
Lucas’s fundamental contribution turns out to be wrong. And, worse, the mistake helped pushed the profession into a prolonged period of ever-increasing damage to the credibility of mainstream macro modeling, denying mainstream theory of any serious claim to stabilization relevance. Anyone who thinks otherwise should recall the bitter denunciations of mainstream thinking by policymakers during and after the 2007-09 Great Recession. All that, however, does not mean Lucas should be dismissed. His intellect is formidable, and his skeptical perspective on the market-centric general-equilibrium macroeconomics that he help entrench in the academy is insightful. I also can’t help believing there is something to the Robert Oppenheimer quote: “Any man whose errors take ten years to correct is quite a man.” Lucas’s error, at least partly because of his towering persona, took more than a generation to correct. There are important messages in that delay.
Blog Type: Wonkish Saint Joseph, Michigan
Write a Comment