Parallel exchange rates in developing countries : lessons from eight case studies

In parallel (dual) foreign-exchange markets - extremely common in developing countries - a market-determined exchange rate coexists with one or more pegged exchange rates. The authors report the main lessons from a World Bank research project on how these systems work, based mainly on case studies in Argentina, Ghana, Mexico, Sudan, Tanzania, Turkey, Venezuela, and Zambia. On the whole, the experiences were disappointing. Most countriestolerated high premiums for long periods, which harmed the allocation of resources and growth. The studies indicate no clear gains from prolonging a dual system. The case for a dual foreign exchange system is stronger when the system is adopted as a temporary option to deal with a severe balance of payments crisis. Argentina, Mexico, and Venezuela resorted to a dual system at the time of the debt crisis, to smooth out the devaluation in the exchange rate to achieve the needed real depreciation. This helped to maintain limited control over domestic inflation, and avoided a sharp drop in real wages while protecting the balance of payments. In the longer term, not much was gained. In the cases studied, the dual system was misused more often than not: it was used too long and the premium was higher than is should have been. Venezuela, for example, used the system for six years with an average 120 percent premium, Mexico for five years (average 30 percent), and Argentina for eight years (average 44 percent). In Argentina and Venezuela, the dual system was used to avoid macroeconomic adjustment while protecting international reserves. It is doubtful the macroeconomic gains (in terms of keeping equilibrium in the balance of payments and lower inflation) were greater than the costs in terms of misallocation of resources. In Ghana and Tanzania, the dual exchange rate system was prolonged to maintain overvalued real exchange rates and expansionary macroeconomic policies. The large premium in those countries (at times more than 1,000 percent) shows the dramatic inconsistency between exchange rate policy and monetary and fiscal policies. On determinants of the parallel exchange rate, the evidence indicates that macroeconomic fundaments (such as fiscal deficit, credit policies, and so on) matter most. In the short run the premium is driven by expectation about the evolution of these macroeconomic factors. Overall, in the countries examined in the project, the existence of a parallel foreign exchange market generated fiscal losses. These losses resulted because the public sector was a net seller of foreign exchange rate. This means that unification has some pleasant fiscal arithmetic. The experience with unification indicates that it usually takes place at the parallel exchange rate. Most countries unified to a crawling peg system, though some opted for floating exchange rates. Successful unification to a fixed exchange rate was less frequent, and it required strong adjustment in fiscal and monetary policies. Regarding speed, unification was quick in countries where the parallel system was used temporarily, and gradual in those where the system existed for long periods and with a tradition of widespread foreign exchange controls.