The ex post failure of uncovered interest parity (i.e. the forward premium bias) in the post-Bretton Woods era is well documented. Recently, explanations have been offered for this failure which have centred upon the unusual monetary experience over this period. We test these explanations using data from earlier periods as well as subsequent to the adoption of an inflation target. Canada operated a flexible exchange rate regime during the Bretton Woods era, providing a unique opportunity to examine ex post deviations from uncovered interest parity. Canada is also unusual in that it has pursued an explicit inflation rate target since February 1991. We find no forward premium bias over the flexible rate period during Bretton Woods, as well as prior to Canada's adoption of inflation rate targeting (when learning would be expected to have taken place), while a forward premium bias does exist during the inflationary/disinflationary period or subsequent to the new monetary regime.