quantities fluctuate; large shocks can have large effects on prices, employment, and exchange rates. (2) The best we can do currently is design institutions or arrangements that minimize the cost of adjusting to shocks. (3) Generally, central banks do not have superior information about real and nominal, permanent and transitory effects. Unless they have superior information about shocks and their effects through time, they cannot expect to stabilize exchange rates. (4) Agents of the central bank often mistake the temporary effect of central bank intervention, arising from the reluctance of speculators to bet against the momentary effects of intervention, for longer-term effects on the value of a currency.
Central bankers and governments have contributed to, and even created, high and variable inflation, low growth, and high unemployment by inappropriate policies during the past two decades. Their past mistakes do not give us reason to believe that they have information about prices, economic activity, interest rates, and other variables that permits them to stabilize exchange rates by means of coordinated efforts in the exchange market.
The time passed long ago when we should have shifted our focus from attempts to fine-tune prices, output, interest rates, and exchange rates to the more important problem of finding arrangements that reduce risks arising from shocks inherent in nature and resulting from social arrangements. The list of such arrangements begins with a monetary rule, adds a fiscal rule setting the maximum ratio of government spending to output or setting the anticipated growth of taxes and spending, and includes removal of remaining impediments to capital and to trade.
Policies of this kind establish procedures to which markets adapt. They eliminate the risks arising from destabilizing shifts in policy. Known procedures and credible preannounced policies permit the type of coordination, through the action of informed traders and speculators, that reduces variability in markets. Variability of exchange rates will be "low," however, under constant policies only if real shocks are small.
The perceived permanent values of the variables are obtained from the equations for the output market (A.1), the money market (A.2), the production function (A.3), and the belief that purchasing power parity holds for permanent values (A.4). The fact that the model is dichotomized removes permanent effects A unanticipated inflation on real variables.