As the world economy becomes more closely integrated, any assessment of the effects of national economic policies must take into account their international effects - both direct and indirect. This suggests the use of multi-country models for policy simulations. But most available models (such as Project Link) are constructed along very traditional Keynesian lines.

Research Fellow Patrick Minford reports preliminary results of an alternative 'new classical' approach to multi-country modelling in a recent CEPR Discussion Paper. He models the world economy by linking 9 small models, each representing a major OECD country. Each country model is identical in structure, and resembles the Liverpool model of the UK economy. The models thus feature both rational expectations and market clearing. There are both union and non-union sectors in the labour market. In the preliminary work he reports, Minford takes the equilibrium or 'natural' values of real variables such as output and real interest rates to be exogenous - there is as yet no 'supply side' in the models.

A problem much stressed recently in model-building has been the 'Lucas critique': that model parameters, which depend on expectations may vary in response to changes in government policies or other exogenous features of the environment. Minford notes that in principle, we can avoid this problem by specifying all expectations explicitly. In practice, however, so many expectations enter a multi-country model that parsimony enforces some choice of critical expectations to be modelled explicitly, leaving others implicit and so vulnerable to the critique. He argues that only empirical trials will tell us how well such choices have been made and whether it would be profitable to widen the choice.

To date Minford's model has been based on preliminary econometric estimates using limited information methods. When satisfactory estimates could not be obtained in this manner, some parameter values based on earlier work were imposed. Minford views the model as still tentative and essentially untested. But he argues that in so far as its structure reflects a major strand of modern macroeconomic thinking and its parameters are related to available empirical work, its simulation properties are of interest.

The main simulations of the model to date have dealt with US policy 'surprises' - i.e., policy changes which were not expected by the market at the time they were announced. The model predicts that a US budget deficit financed by bonds (i.e. holding money supply constant) will not raise US or world GNP but will raise world real interest rates quite substantially. This result - '100% crowding out' - is of some interest in view of recent US budget deficits. It is not consistent with the view that these deficits have stimulated the world economy in a Keynesian manner. It does lend support to the complaint that these deficits have created substantial problems for countries borrowing in world markets - notably LDCs whose 'debt crisis' can be laid squarely at the door of US fiscal policy. The model also says that US monetary policy is very powerful - a 1% once-for-all rise in money supply raises world GNP by 0.8% in year one. Are these model simulations consistent with recent world experience? Minford argues that they are: that large US deficits have not been 'stimulatory' but have raised world real interest rates substantially; and that the US monetary contraction in 1980-1981 and expansion in late 1982 have been the major cause of the latest world business cycle.

What implications do these simulations have for international economic policy? Minford's model indicates that even though theoretically possible, international fine tuning is unnecessary, as the model stabilises itself fairly rapidly. Stimulative ('locomotive') policies - i.e. rises in budget deficits and money growth - will have their principal effect on inflationary expectations and will not speed up recovery. Minford argues that predictability of policy is clearly desirable - shocks and surprises disturb agents plans and in particular increase the variance of output. At the 'micro' country level, it would pay EEC governments to borrow less when US deficits have pushed world interest rates up; this would ease pressure on the world capital market and reduce the disturbance to world output.