The analysis of exchange rate currency bands received a major set-back from the many empirical studies which found that EMS data from the 1980s appeared to violate the main empirical implication of Krugman's simple target zone model: that exchange rates will usually lie close to the edges of their bands. Extensions to consider official intramarginal intervention and the risk of realignments have helped to explain why they are usually close to the centre of the band, but a puzzle remains. In Discussion Paper No. 845, Research Fellow Patrick Honohan notes that the empirical evidence has been based entirely on the bilateral exchange rates between the Deutschmark and other EMS member currencies. Since member countries in fact intervene against all other participating currencies, simply relating a currency's position to its bilateral intervention limits vis-à-vis the Deutschmark does not adequately specify its position within its EMS band.

Honohan presents calculations to show that the widely-reported pattern of clustering around the centre of the band is a direct consequence of using these bilateral rates. The exchange rates of four of the six currencies in the ERM narrow band during 1979-88 were in fact distributed almost uniformly across the full bands within which the EMS allowed them to fluctuate; the Belgian franc was usually near a band-edge, while only the Irish punt was usually close to the centre. In practice, currencies often run up against a limit with another currency before reaching the limit against the Deutschmark, which accounts for the clustering of their bilateral rates in the middle of the permitted range. Honohan notes that this corrected methodology finds little clustering at the bands' edges, as Krugman's original model predicts, but it considerably weakens the case against it. Empirical tests of the theory's extensions and its predictions concerning correlations between interest rates and exchange rates should adopt this improved methodology which takes proper account of the multi-currency nature of the EMS.

The European Monetary System and the Theory of Target Zones
Patrick Honohan

Discussion Paper No. 845, September 1993 (IM)