Stagflation and the Phillips Curve, Part II

The aftermath of the stagflation decade featured a battle between the Early Keynesian (EK) and Rational Expectations (RE) versions of the Phillips curve: w_j(t)=a_o+a₁(U^N(t)–U(t))+p_k(t) versus w(t)=a_o+a₁(U^N(t)–U(t))+E_tp(t+1). (The variables were defined last week.) Robert Lucas’s RE equation won that battle but is now on the edge of losing the war.

The game fundamentally changed when the GEM Project demonstrated that employers and employees rationally reject the expectations approach to periodically adjusting labor prices for inflation. Agents rely instead on catch-up to inflation that has already occurred. Generalized-exchange analysis has reestablished the EK Phillips version as the superior macro estimating equation. That is good news. Lucas’s version was never any good at explaining the data and, as a result, gravely misled policymakers. The widespread acceptance of the centrality of rational inflation expectations in the Phillips curve can be understood as one of the worst mistakes in the history of macroeconomics.

Accounting for Nonstationary Inflation Regimes 

There is one detail to clean up. Recall that last week’s wage-arrangements analysis supporting the superiority of the EK wage equation assumed stationary mean inflation. SMI is a reasonable condition for a lot of modeling but not for the stagflation decade. In the 1970s the Fed’s regime was clearly not stationary. Fortunately, it is not difficult to enrich GEM modeling to accommodate instability in the central bank’s inflation regime (p^N). The more complete analysis increases its explanatory power without damaging the conclusions derived.

For ease of presentation, assume that the wage-adjustment period (k) remains constant. Also posit that the monetary authority has at time t a target trend inflation rate (p^N), which it can implement immediately: p^T(t) = p^N(t). Adjusting wages for inflation is now analytically separable into two parts: (a) the influence of price change given the inflation regime and (b) the influence of changes in the that regime.

For the first, the use of expectations has been shown to be suboptimal. It is relatively costly, with no corresponding benefits, to produce and implement inflation forecasts in support of periodic compensation adjustments. Catch-up to past inflation (p_k(t)) is the superior strategy. The second  is also a realatively straight-forward affair yielding the same conclusion. It features the efficient recalibration of rational wage-setting arrangements emdodied in the compensating wage premium: W^P(t)=W^m(t)(1+p^N).

The optimal adjustment of W^P little affects the GEM Project’s revival of the EK Phillips curve. The product-price consequences from ΔW^P, induced by the central bank’s altered inflation regime, is simply absorbed by the powerful catch-up term, p_k(t). The EK wage estimating equation is unchanged:

w_j(t)=a_o+a₁(U^N(t)–U(t))+p_k(t)+ε(t),

where ε is an error term.

Summing Up

Given sufficient firm-specific human capital to make worker turnover costly, optimizing labor-pricing arrangements mandate that catch-up to inflation that has already occurred must always be used in periodic wage adjustments, replacing Lucas’s irrational “rational expectations”. The new form will take some getting used to. The nature of catch-up is badly understood by New Keynesians. Carlin and Soskice (2006, p.160) are illustrative: “The key point to highlight is that although the inertial or backward-looking Phillips curve matches the empirical evidence concerning inflation persistence, it has a major shortcoming. Because it rests on ad hoc assumptions – in particular about the inflation process – rather than being derived from an optimizing micro model of wage or price setters’ behaviour, it does not allow a role for ‘credibility’ in the way monetary policy affects outcomes.” It is dismaying how easily NK theorists ignore powerful clues provided by evidence.

Once important evidence is no longer pushed aside, it is easily seen that, given a shift in the central bank’s nominal regime, rational catch-up incorporates any price-inflation consequences from recalibrating the wage premium (W^P). Throughout, the GEM analysis remains rooted in the formal method of rational mediated exchange organized by general decision-rule equilibrium rightly prized by modern theorists.

Most significant, however, is another point that New Keynesians ignore. Coherent mainstream (friction-augmented general-market-equilibrium) modeling, with its inherent wage recontracting, itself demonstrates the necessity of meaningful labor-price rigidity (MWR) for any Phillips curve to produce useful results. Catch-up by itself helps explain wage-price inertia, but only MWR rationally suppresses labor-price recontracting. Such suppression is a necessary condition for adverse disturbances in nominal demand to induce evidence-consistent losses in employment and output. Absent suppression, macro theory cannot be stabilization-relevant and can only mislead policymakers.

Blog Type: Wonkish Saint Joseph, Michigan