The extent to which deviations from covered interest parity can be attributed to transactions costs has been exaggerated in the economic literature because the swap market in foreign exchange has been ignored. It is shown that such deviations should be no greater than the lowest of the transactions costs in one of three markets: the swap market or either of the two relevant securities markets. This reconciles the theory with the data, which show spreads of no more than a few basis points. However, the empirical results have no direct bearing on the conventional market efficiency hypothesis.

The extent to which deviations from covered interest parity can be attributed to transactions costs has been exa ggerated in the economic literature because the swap market in foreig n exchange has been ignored. It is shown that such deviations should be no greater than the lowest of the transactions costs in one of thr ee markets: the swap market or either of the two relevant securities markets. This reconciles the theory with the data, which show spreads of no more than a few basis points. However, the empirical results h ave no direct bearing on the conventional market efficiency hypothesi s. Copyright 1988 by University of Chicago Press.