Hundreds of studies regress the one-period-ahead change in the spot exchange rate on the forward premium – henceforth called the Bilson (1981)-Fama (1984) (BF) regression – without making any allowance whatsoever for the possibility that the process underpinning exchange-rate movements may have changed, at least intermittently, over the modern period’s four decades of floating rates. Researchers report estimates of the slope coefficient that are not only less than unity, but less than zero. A slope coefficient that is less than unity implies that the forward premium is a negatively biased predictor of future changes in the spot exchange rate. A slope coefficient that is negative implies that spot-rate changes tend to be in the opposite direction of that predicted by the forward premium.2Interpreting their results as implying a stable relationship in the data, international macroeconomists conclude that “one can make predictable profits by betting against the forward rate” (Obstfeld and Rogoff, 1996, p. 589).