In the large macroeconomic literature on rules versus discretion, many argue that `punishing' governments that seek to effect `surprise' inflation while retaining flexibility in face of shocks is impossible. In Discussion Paper No. 767, Research Fellow Patrick Minford argues that this is not merely theoretically possible but actually occurs in modern democracies. Models that stress maximizing behaviour by individuals lack the mechanisms required to initiate coordinated action and thus enforce punishment strategies, which leads to multiple equilibria. But democratic institutions can determine a unique equilibrium if every five years or so elections take place in which all voters are expected to take part.

Minford assumes that the public are informed about some current events and can monitor whether the government reacts in accordance with its preannounced rules. This discourages the government from reneging on promises. It seeks to maximize its re-election chances, while voters withdraw votes both relatively in response to poor performance and absolutely in sufficient numbers to ensure its defeat if it relaxes monetary policy except in response to `news'.

In the run-up to the election, the government has full information about current shocks, but voters have such information only for past shocks and observe current inflation and output, although they observe current shocks once the election is over. Even though the government has unfettered discretion, it will follow a policy of tying its hands and responding only to shocks. If the public has less information before the election, this framework increasingly resembles the `pure rule' case, and the trade-off with discretion re-emerges; but giving it access to all publishable macroeconomic data enables the government to pursue the optimal policy of setting prior targets but allowing contingent responses. Minford maintains in conclusion that such a regime with contingent rules and floating exchange rates will outperform either the ERM or EMU.

Time-inconsistency, Democracy and Optimal Contingent Rules
Patrick Minford

Discussion Paper No. 767, February 1993 (IM)