This paper uses new survey data on foreign exchange expectations for Argentina, Brazil and Uruguay to test the\nhypothesis of unbiasedness. The pattern emerging is revealing: only Argentinean forecasts are unbiased predictors\nof exchange rate movements, while agents err systematically for Brazil and Uruguay. We argue that the systematic\nintervention of the Argentinean Central Bank in the foreign exchange market is likely to explain this result, as it\nsimplifies the forecast exercise in that market. As long as the requirements to predict well are simple, agents perform\nwell. If instead the exchange rate determination model is intricate, expectation failures arise.
Loading....