Foreign Exchange Markets
Hysteresis effects

Recent empirical research on foreign exchange markets has rejected the hypothesis that the difference between the forward rate and the realized future spot rate in an efficient market is white noise. This is most commonly interpreted as evidence of a risk premium in such markets, although no satisfactory theoretical model of this risk premium that performs well empirically has yet been found.
In Discussion Paper No. 407, Research Fellow Richard Baldwin argues that even very small transaction costs may help to account both for the rejection of these market efficiency tests, and for the failure to find a good empirical model of the risk premium. Baldwin draws on recent research in industrial economics which indicates there is a range of inactivity in firms' dynamic entry-exit strategies, and that this band permits hysteresis. Since transaction costs in asset markets are sunk costs, the prediction errors used in market efficiency tests are in fact a combination of the true prediction errors and a wedge consisting of the `option value' of being in foreign currency, plus or minus the transaction cost. Baldwin suggests that the failure of measured prediction errors to be white noise may be due to this wedge.
Baldwin investigates the consequences of relaxing the assumption of uncovered interest parity and shows that the presence of tiny transaction costs plus uncertainty implies that a central bank may not face pressure on its reserves, as long as the interest differential is less than about one percentage point. This allows a moderate degree of monetary independence in a fixed exchange rate regime and may explain why there has been so little arbitrage across EMS currencies. By extension, removing all uncertainty on cross-interest arbitrage (e.g. by maintaining zero margins around central parities in the EMS) may lead to greatly increased arbitrage, and there may therefore be increasing difficulties as the margins on interest rate differentials within the EMS are reduced to zero. Baldwin's analysis applies equally to asset prices in the domestic market, and it suggests that even a small Tobin tax on asset trade could allow rates of return on various assets to get fairly far out of line, with clear deleterious effects on economic efficiency.
This analysis also points to a real difference between irrevocably fixed exchange rates and a common currency: the elimination of small transaction costs in moving to a single European currency would lead to efficiency gains above and beyond those coming from the elimination of exchange rate uncertainty.

Richard E Baldwin
Re-Interpreting the Failure of Foreign Exchange Market Efficiency Tests: Small Transaction Costs, Big Hysteresis Bands

Discussion Paper No. 407, April 1990 (AM/IM)