The stagflationary experience of the 1970s has shaken the confidence of both economists and policy-makers in activist macroeconomic management. In the light of this experience, the authorities in many countries turned to simple macroeconomic policy rules such as fixed targets for the money supply, first advocated by Friedman (1960). The experience with unconditional money supply targets has, however, proven rather disappointing. Although inflation has indeed fallen since the early 1980s, disinflation was accompanied by a large increase in unemployment: in many European economies this unemployment persists. Attention has therefore recently focused on another simple policy rule, James Meade's 1978 proposal for targeting the growth of nominal GDP. Nominal income targets seem preferable to money supply targets because they prevent shifts in the velocity of money from affecting real output and employment.
The properties of nominal income targets have since been examined both empirically and theoretically. The theoretical analyses of nominal income targets are implicitly based on a social welfare function that penalizes only deviations of output (or employment) from equilibrium output (or employment). In addition, most of the analyses concentrate solely on monetary policy. In this paper, I extend these earlier analyses to consider a government objective function which depends on both employment and monetary stability. In addition, I examine the role of supply-side fiscal policy, and its implications for optimal stabilization policy and nominal income targets.
I use a theoretical model of a small open economy that produces both internationally traded and non-traded goods and services. Perfect capital mobility and perfect commodity arbitrage are assumed for simplicity. In addition, all domestic and external disturbances are assumed transitory, and the focus is on the deviations of actual variables from their long-run equilibrium values. The paper presents a careful derivation of the welfare cost of aggregate fluctuations, which, it is assumed, the authorities attempt to minimize. I assume, following Gray (1976), Fischer (1977) and Bean (1983), that wages are negotiated before the realization of the current disturbances, and that they are imperfectly indexed. Although wages react to both inflation and unemployment, this ex post response is sub-optimal and causes a reduction in welfare, which one could call the welfare cost of unemployment. In the context of my model this is nothing more than a Harberger (1971) triangle in the labour market, and was first suggested in a similar context by Aizenman and Frenkel (1985). The second welfare cost, that of monetary instability, arises because money is assumed to pay no interest. This cost can be measured by the familiar Bailey (1956) triangle. The authorities are assumed to minimize a weighted average of these two welfare losses.
The results can be summarized as follows: if the authorities can use both monetary and fiscal policies in the short-run, then the welfare losses caused by domestic and external disturbances can be fully neutralized. The optimal policy assignment is straightforward: monetary policy is devoted to the monetary stability objective, in order to eliminate deviations of domestic interest rates from their long-run equilibrium values, while fiscal policy concentrates on the full employment objective. Alternatively, incomes policy could be used in place of fiscal policy in pursuing the employment objective. If the authorities cannot use either fiscal or incomes policy, or if they do not wish to use them for allocative, distributional or ideological reasons, then only a second-best policy is available: monetary policy should be devoted to managing the exchange rate, so as to achieve a balance between the the welfare costs of a marginal increase in unemployment and a marginal increase in monetary instability. The optimal trade-off between these two costs depends on all the parameters of the model, including the relative weights given to the full employment and monetary stability objectives in the policy-makers' social welfare function.
The analysis suggests that nominal income targets are in general suboptimal, unless there are no supply or world interest rate shocks. Alternatively, such targets are optimal if monetary stability carries no weight in the objective function of the authorities and if supply-side fiscal policies could be directed to counteracting supply shocks. If this is not possible, then the optimal nominal income target must accommodate supply shocks to the extent that they are disequilibrating due to the short-run sluggishness in the labour market.