IN THE EARLY 1970s most economists were confident that received theory, in either a fiscalist or monetarist version, could adequately explain and forecast the demand for money, Goldfeld [14]. This belief was important because the money demand function is among the foundations of both the classical and Keynesian macro models in their traditional presentations, and it is a cornerstone of econometric forecasting. During the past two years, however, evidence of the money demand function's failure has been accumulating, Enzler, Johnson, and Paulus [8], Garcia and Pak [12], Goldfeld [15], and Sinai [22].1 At the same time it has become apparent that the ratio of currency to demand deposits began rising in 1961 and was by 1978 approaching levels unprecedented in its history since 1892. Further, as the ratio in the U.S. appears to rise during times of crisis-wars and severe recessions-its recent acceleration might be interpreted as a portent of doom. This paper demonstrates, however, that the cause of the money demand function's failure contributes pari passu to explaining the rise in the currency ratio, for which there is no need, therefore, to resort to inimical forces.2 Section I demonstrates that the demand for currency has remained stable since 1952, while demand deposit demand has shifted downward. The cash ratio's behavior results, therefore, not as sometimes has been mistakenly claimed from an unexpected increase in the currency numerator, but from a decrease in the demand deposit denominator. Causes of the demand deposit reduction are discussed in Section II, where the growth of federal funds transactions are singled out for particular consideration. Section III shows that including net federal funds purchases in the dependent variable restores the validity of the demand deposit demand function and substantially improves the forecasting performance for both deposits and the currency ratio. (See Figure 1.)