RDP 8705: The Economics of International Policy Coordination 3. Empirical Analyses of Policy Co-ordination

Attempts to measure the gains to coordination are rare. The issues of interdependence have been studied in the large scale models such as project LINK^[11] or the linked U.S./Canada model in Helliwell and McRae (1977). These studies attempt to measure the size and sign of the transmission of economic disturbances between countries rather than applying a game theoretic approach. The first attempt at an empirical implementation of the strategic game theoretic approach was undertaken by Oudiz and Sachs (1984). In this study, the authors used policy multipliers from several large scale models of the world economy and applied static game theory to determine the gains from coordination. They assumed that the estimated results were generated from a Nash-Cournot equilibrium and then estimated the gains from coordination. The results showed a very small gain.

Hughes-Hallet (1987) used dynamic game theory in a three-region empirical model (consisting of the U.S., EEC and the rest of the world) to assess the difference between non-cooperative and cooperative policy responses to the experience of the world economy since 1974. He found larger gains to the U.S. from cooperation but very little difference for the ESC.

McKibbin and Sachs (1985) have applied the results from dynamic game theory to a five-region simulation model of the world economy to measure the gains to coordination. The model was calibrated using 1983 data. In this paper, the authors found small gains to coordinating the macroeconomie policies of the major industrial regions in the face of an inflationary shock, such as occurred in the late 1970's. However, they do show that the gains to the developing countries are potentially large. In the face of a global inflationary shock, a non-coordinated disinflation results in monetary contraction and fiscal expansion in the industrial region. This leads to high world interest rates and a strong U.S. dollar. When the authors simulate the same shock assuming a coordinated response, the policy mix is less extreme and world interest rates are much lower. In these studies the comparison is the outcome of a Nash-Cournot game versus a cooperative equilibrium. In a further study ishii, McKibbin and Sachs (1985) found that if the model is used to generate a future baseline (assuming OECD projections for macroeconomle policies rather than optimizing policymakers) and then policymakers optimised in a cooperative manner, the result was a substantial gain from coordination.

Taylor (1985) examined the issue of the gains from coordination using an estimated reduced form model consisting of the seven largest OECD countries. Comparing the outcome of Nash-Cournot equilibria and cooperative equilibria, he also found small gains from coordination depending on the weights in the policymakers' objective function.

Footnotes

See Fair (1978) for an early survey of the multipliers from the major models. Also see Bryant, Henderson et al (1986) for a recent Brookings Conference which considered the policy multipliers of the major world models and Frankel (1986) for an application of the results of this conference to policy coordination issues. [11]