Time Inconsistency: The Temptation to Organise Surprise Inflation

Starting from this latter insight, the stage was prepared for an analysis of the misuse of the policymaker’s discretion. Following Kydland and Prescott (1977), Barro and Gordon (1983) see the policymaker minimise a loss function of the form:

where π is inflation, y is log output and k some positive desired deviation from market equilibrium y*.

The form of this supply function, with quantities following price signals, reveals the Walrasian roots of NCM; this is a further difference compared to the Marshallian tradi­tion of Friedman’s monetarism (Hoover 1984).

This simple game-theoretic exercise nicely demonstrates NCM’s simplistic attitude towards the relationship between politics and markets: policymakers appear as inclined to fool private agents, in order to aim for a policy target that does not conform to a direct aggregation of individual preferences. This justifies the recommendation to abolish the policymaker’s discretionary powers (that is, let him optimise a per-period loss function) and to impose an obligation to obey predictable rules of decision making.

The list of critical points that can be raised against the surprise-inflation theory is long (Forder 2001). The arguments given for the k > 0 target are hardly convincing: if income taxes produce a distortion in individual labour supply so that private and social benefits of work do not conform, one might consider a tax reform, but not expansive monetary policy. Blinder (1997) claims never to have met central bankers who succumb to the temptation of initiating a surprise inflation because they can easily anticipate the inevitable damage in their intertemporal loss function.

Game theoretic models of time inconsistency have been so popular, because we have wanted to believe them, despite these models being unrealistic in several respects... [They] have ignored the fact that there are long lags between monetary policy adjustments and their effect on the real economy, and that both inflation and output have persistence. But so long as wages and prices are fully flexible, such monetary policy lags would imply that the policy would be trans­parently observed before it affects the economy; consequently the Central Bank could not fool anybody... If monetary instruments operate with a lag, then a rational public would observe them and adjust their expectations accordingly if they have not bound themselves into a con­tract longer than that lag. Hence the public would not be fooled, and the time inconsistency problem would vanish.