In the 1970s, Robert Lucas famously argued that adaptive expectations left information on the table and, as a result, could not be consistent with optimizing behavior. Lucas’ target was then-mainstream Early Keynesians who used adaptive expectations to motivate catch-up adjustments for inflation in their reduced-form wage equation (the Phillips Curve). In his alternative wage equation, Lucas substituted rational expectations (RE), which reasonably requires the cost-effective use of available information. Lucas’ neoclassical Phillips Curve was the vanguard of the macro-theory insurgency that eventually banished EK thinking from mainstream discussion and debate. The advent of RE was widely accepted as a very big deal, a conclusion that contrasts to the relatively minor role the GEM Project has more recently showed them to play in stabilization-relevant theory.
Here’s why RE got so much attention. The neoclassical RE-augmented market-centric general-equilibrium model produces a Phillips Curve that is vertical at the natural rate of unemployment. That implies discretionary monetary policy had no real-side (employment and output) effects, an outcome that came to be known as the “Policy Ineffectiveness Proposition”. At the time, a surprising number of macro theorists believed that the introduction of Lucas’ impeccably rational expectations was the cause of stabilization-policy ineffectiveness. If true, RE would have been the coup de grace to the stabilization-relevance of EK modeling.
Keynesians, of course, fought back. (It’s called the macro wars for a reason.) First out of the gate in the concerted attempt to weaken the power of rational expectations was Stanley Fischer (1977), who was to become the Governor of the Bank of Israel and later the Vice Chairman of the Federal Reserve. He restored the real effects of monetary policy by imposing arbitrary staggered nominal wage setting on the neoclassical RE-augmented macro model. Fischer demonstrated what macroeconomists who read and understood Patinkin (1965) already knew; market-flexible prices (especially wages), not rational expectations, are the necessary condition for the ineffectiveness proposition. That conclusion certainly came as no surprise to economists at the Fed who were tasked to model wages and Keynesian nominal-to-real causality. I know; I headed that talented research and policy unit. But, busy with the stagflation crisis, we welcomed Fischer’s help in deflecting some of the heat from the not-fully-moored pestering academics who somehow believed that rational expectations were a dagger to the heart of discretionary demand-management.
With a decades-later perspective, I believe that the true import of Fischer’s analysis was not illustrating the criticality of wage rigidity in stabilization-relevant macro modeling. To reiterate, the Early Keynesians had already accomplished that. Moreover, EK theorists knew that his arbitrarily staggered wages necessarily leave a lot of information on the table, implying that Fischer’s model did little to advance the development of rigorous macro theory. The insurgent neoclassicists were unimpressed.
Fischer’s work is, instead, most significant as early indicator of how New Keynesians would go badly wrong during and after the macro wars, dooming their macroeconomics to ineffectiveness in the central quest to understand the rational suppression of wage recontracting as well as the fundamental nexus between rigorous micro- and microeconomics. Fischer surely knew that his analysis of policy effectiveness was constructed on arbitrary restrictions on wage flexibility; he surely knew that he had not rationally suppressed wage recontracting. Indeed, New Keynesians generally knew that. Yet the practice of assuming obviously irrational wage rigidity and asserting it as part of settled macro theory became the norm in NK research.
Lazy model-building has consequences. Most consequently, it deprived NK theorists of the crucial insight on how modern economies work that comes from careful rational-behavior modeling. It cannot surprise that Fischer’s irrational staggered-wage analysis mangles the essential nominal-to-real causation that is at the core of stabilization-relevant macro theory. Most consequently, had he adequately modeled Keynesian causation, i.e., adequately modeled meaningful wage rigidity, he would have been on his way to insightfully explaining the economics of the stagflation decade, its optimal remedial policy design, and anticipate the other shoe to drop – the downsizing crisis of the 1980’s that generated millions of permanently lost good (rent-paying) jobs. The neoclassical macro insurgents, using the stagflation crisis to discredit EK modeling despite themselves having little grasp of what was actually going on, would have been stopped in their tracks.
NK theorists, given their quiet reliance on arbitrarily staggered wages (Calvo or Taylor) to inform stabilization dynamics, have never produced an adequate explanation of the how and why of the stagflation and downsizing decades. New Keynesians remain innocent of the close relation between the two crises. Had Fischer actually modeled rational labor pricing in his rejection of policy ineffectiveness, he could have addressed the most important questions of the day. (Those questions unsurprisingly involved involuntary unemployment.) He could have changed the course of the macro wars, avoiding the great damage to policymaking and the reputation of macroeconomics that they caused. The next two weeks’ posts elaborate on the stagflation and downsizing decades, demonstrating again and then again the power of generalized-exchange modeling to properly explain misunderstood, overlooked, important macro phenomena.
Blog Type: Wonkish San Miguel de Allende, Mexico