Notes on Rational and Irrational Expectations

This file is a single-author scholarly essay. Its scope is compact but wide: Haberler moves through psychological theories of the business cycle, the monetarist-Keynesian “post-Keynesian consensus,” the rational-expectations challenge, and William Fellner’s credibility approach to disinflation.

Haberler’s main thesis is that rational expectations sharpened macroeconomic thinking but overstated its claims. Expectations matter, and policy cannot permanently exploit inflation-unemployment tradeoffs; yet real-world cycles persist because agents do not share one model, government behavior is not cleanly predictable, wages and prices are institutionally rigid, and money illusion has not vanished.

He begins with Pigou, Keynes, and Jöhr on “errors of optimism and pessimism.” Haberler accepts the rational-expectations objection that people should not repeat identical forecasting errors indefinitely, but he denies that this abolishes cyclical misjudgment. Forecasting institutions and cycle-conscious firms did not make cycles disappear:

Yet the business cycle did not go away; it is still with us.

Psychological waves are therefore not independent first causes but amplifiers of cumulative expansion and contraction. Haberler then connects this to monetary causation, especially the Great Depression, where Friedman and Schwartz’s monetary interpretation is largely endorsed without excluding nonmonetary triggers. Keynesian deficit spending and monetarist analysis can be reconciled once depression becomes cumulative collapse:

The Great Depression surely was a landslide of major proportion.

The second section reconstructs a broad consensus after Keynes and monetarism. Fisher’s distinction between nominal and real interest, and between anticipated and unanticipated inflation, becomes central. Haberler criticizes Keynesian policy discussion for neglecting changing inflation expectations, then credits monetarism with restoring money to macroeconomics. The consensus he describes is anti-fine-tuning but not anti-stabilization: money matters, inflation requires monetary expansion, and no permanent Phillips curve exists.

Rejection of the Phillips curve as a long-run proposition is an essential feature of the consensus which I tried to describe.

Rational expectations enters as the “radical wing” of monetarism. Haberler treats its strongest claim—that systematic monetary and fiscal policy affects only nominal variables—as both intellectually important and empirically implausible. Its validity depends on a sharp distinction between predictable rules and unpredictable shocks, but actual policy does not come so neatly packaged:

In reality, government policies are spread out over the whole range between the two extremes.

His core conceptual move is to replace the rational-expectations image of homogeneous, model-consistent agents with a plural and institutionally embedded economy. People do form expectations about policy, but they do not all perceive policy identically or infer the same “correct” monetarist conclusion. More damagingly, the theory assumes competitive market clearing and neglects wage rigidity, labor unions, contracts, and other constraints.

The neglect of institutional rigidities greatly impairs the relevance of the rational expectations theory for the real world.

The oil shock of 1973–74 is Haberler’s concrete test case. Against Barro’s claim that monetary policy had no role in offsetting a perceived real supply shock, Haberler argues that downward money-wage rigidity changes the policy problem. If real wages must fall, some rise in the price level may be less destructive than unemployment caused by forcing nominal wages downward. This defense of limited accommodation relies on residual money illusion, which he refuses to treat as extinct:

But money illusion is a fairly hardy plant.

The conclusion turns to Fellner’s credibility hypothesis as a better synthesis. Anti-inflation policy must shape expectations through consistent, believable action, not assume that the public has already decoded a stable policy rule. Credibility matters because disinflation succeeds only if wage and price setters believe the authorities will persist.

Market participants must be persuaded that the government will firmly stick to a policy of reducing money GNP growth to a noninflationary level within a reasonable period of time.

The essay’s relevance lies in its effort to defend a moderate monetarist-Keynesian position at the moment rational expectations was reshaping macroeconomics. Haberler grants the new school its insight—anticipated inflation erodes stimulus—but rejects instantaneous neutrality. The final judgment is deliberately balanced:

In summary, the post-Keynesian consensus still stands, fuzzy though it is on the edges like most concepts in the social sciences.