At the latest since 1978, international shifts in the demand for money have widely been recognized as a major problem for monetary-policy making under flexible exchange rates. To the extent that such shifts are unpredictable, the case for monetary preannouncement is weakened (Giersch, 1977; Girton, Roper, 1978; Sachverständigenrat, 1978/9; Vaubel, 1980). To the extent that they are unidentifiable once they occur, the case for flexible exchange rates is weakened (Giersch, 1977; King et al., 1977; Sachverständigenrat, 1978/9; Vaubel, 1980). Empirical identification of international shifts in the demand for money (or, what is the same, of currency substitution in demand) has been attempted though estimation of money-demand functions and exchange-rate functions. Concentrating on international stock shifts in the demand for money rather than on portfolio-growth effects (which are likely to be far less volatile), most researchers used the forward premium/discount or, what is approximately the same, the international interest- rate differential as the most promising determinant of currency substitution; for these variables reflect the expected difference in the opportunity cost of holding the monies concerned if interest is not paid on money balances or on required reserves.