Analysis of optimal policy design has usually assumed that decision-making is centralized, i.e. that the political institutions charged with policy-making always act in harmony. This is not the case in many countries where the Treasury and Central Bank manage fiscal and monetary policies separately. If they do not agree on targets for policy values or priorities to be assigned to those targets, their policies may conflict. In Discussion Paper No. 251, Maria Luisa Petit and Research Fellow Andrew Hughes Hallett illustrate the effects of noncooperative policies on the trade-off between output and inflation in Italy during the 1970s and 1980s. In the 1970s, the Bank of Italy was obliged to buy all the public bonds which the Italian Treasury was unable to sell elsewhere. Faced with very substantial budget deficits in the 1970s, this requirement meant that monetary policy was directed entirely at deficit financing with stable interest rates, and it was not considered worthwhile publishing any figures on the growth in money supply.
The first policy change was to offer Treasury bills to the public in 1976. The Bank of Italy could try to control the growth in the monetary base by operating in the secondary debt market, but monetary policy remained subordinate since the Bank of Italy was still obliged to buy any unsold bonds. The Treasury created rapid monetary growth by issuing debt, which the Central Bank attempted to destroy by openmarket operations in order to protect its targets for inflation and external balance. The 1977-81 period represents a classic case of non-cooperative policy-making, according to Hughes Hallett and Petit. This changed in 1981, when the Bank of Italy was freed from its obligation to buy unsold Treasury bonds. The Bank of Italy and the Treasury could then, in principle, choose between following a fully non-cooperative approach or they could agree to cooperate across the range of targets pursued by both institutions.
Hughes Hallett and Petit use an empirical model of the Italian economy to calculate the performance of optimal non-cooperative, `enforced' cooperative, and `cooperative-by-agreement', policies over the period from 1977 to 1981. The analysis assumes that the Italian Treasury and the Bank of Italy act as two different policy-makers, who can decide whether or not to cooperate. The government's targets are first output growth and then low inflation, while the targets of the central bank are first inflation and then the balance of payments (in particular, the level of international reserves).

The authors find that the policy options under non-cooperative decision-making are completely different and inferior to the cooperative case. Reduced inflation is obtained in the cooperative case through more restrictive monetary policy. Public expenditure is the main instrument for expanding output in the model, together with the extra internal and external demand generated by reductions in the rate of inflation. But public expenditures alone are not able to reduce inflation and so the stimulating effects of lower inflation on output and aggregate demand do not materialize. A subservient monetary policy (such as the Bank of Italy was obliged to follow in the 1970s) therefore results in higher inflation and loss of that extra output stimulus. There are also clear losses from non-cooperative policy-making where the central bank feels obliged to block unacceptable fiscal expansion by the government, as the Bank of Italy was doing at the time of the `divorce' in 1981.

Decentralised Policies and Efficient Tradeoffs: An Essay on the Costs of Uncoordinated Fiscal and Monetary Policies Andrew Hughes Hallett and Maria Luisa Petit