Central bank cooperation in foreign exchange markets: an empirical test

Atlantic Economic Journal, Sept, 1996 by Thomas J. Carter, Paul R. De Lancey

The study looks at pairs of central banks and their reactions to each other's policies. The equations below are estimated for several pairs of countries.

[R.sub.it] = [[Alpha].sub.i] [[Beta].sub.1i] [R.sub.jt] [[Beta].sub.2i] [X.sub.it],

[R.sub.jt] = [[Alpha].sub.j] [[Beta].sub.1j] [R.sub.it] [[Beta].sub.2j] [X.sub.jt],

where [R.sub.it] is the foreign exchange reserve of country I at time t. It is a function of reserves in country j and a vector of other variables, [X.sub.it]. [X.sub.it], includes GNP, imports as a proportion of GNP, previous period reserves, interest rates, and the variability of exchange rates. These equations are estimated using iterative three-stage least squares after correcting for serial correlation. The monthly data are all from the floating exchange rate period.

Cooperative intervention generates negative reaction coefficients, [[Beta].sub.1i] and [[Beta].sub.1j], for both countries if the purpose of the intervention is bilateral, that is, designed to realign the exchange rate between the two countries. For example, Germany purchases dollars with marks and France sells dollars for francs to depreciate the mark versus the franc. The central banks cooperate, but France loses reserves while Germany gains reserves. Such cooperation is to be expected for countries within the European Monetary System.

A positive [[Beta].sub.1i] implies that France sells francs when Germany sells marks. Since French policy raises the mark relative to the franc while German policy does the opposite, France is competing with Germany.

This study looks at the German-Italian, German-French, and French-British pairs within Europe. Also, the equations are estimated for the German-U.S. relationship. The estimates of the intra-Europe [[Beta].sub.1i]'s are almost all negative but statistically insignificant at the 95 percent level with t-ratios ranging from 0.61 to 1.55. This gives some evidence that European countries tend to cooperate in their bilateral exchange rates. The only reaction coefficient that is significantly positive is that showing Germany's reaction to changes in U.S. reserves (t-ratio = 2.89), meaning that Germany tends to lean against U.S. interventions.

International policy cooperation may be valuable, but does it happen? This study suggests that cooperation in exchange rate intervention exists within Europe but not between the leading European central bank, the Bundesbank, and the U.S.