The Response of Interest Rates to Unexpected Weekly Money: Are Policy ChangesImportant?* R. W. HAFER SouthernIllinois University at Edwardsville Edwardsville, Illinois RICHARD G. SHEEHAN Notre Dame University Notre Dame, Indiana I. Introduction The evidence presented in prior studies of the interest rate-weekly money relationship generally indicates that financial marketsreact only to unexpectedmoney changes. The data also support the hypothesis thatthe response of interest ratesto the money announcement significantly changed in October 1979 and in October 1982. On 6 October 1979 the Federal Reserve announcedthat it would henceforth place more policy emphasis on the behaviorof the money stock and reduce the weight given to fluctuations in the federal funds rate. In October 1982 the Fed again changed its operating procedure,shifting away from targeting the monetary aggregates to emphasizing the importance of discountwindow borrowing. The evidence to date from almost every studyrejects the hypothesis of a stable money announcement interestrate relationship in favorof the alternative thatboth October policy regime changes represent structural breaksin the estimated regressions. Reading through this growing literature, we were struck by the noticeable absence of tests for breaks in the regressions other than the above-mentioned regime shifts.' Previousresearchers generally have ignored the potentially destabilizing impacts stemming fromthe numerous financial innovationsthat occurred after 1979 such as the redefinition of the monetary aggregates in 1980 and the nationwide legalization of NOW accounts in 1981. The effect of policy events like the special credit control program of 1980 and shifts in the weight attached to M1 as a policy target also have not been adequately treated in tests of structural stability. Given the numerous financial changes occurring between 1980 and 1982, it shouldnot be surprising that selecting a break point *We would like to thank Jerry Dwyer, Barry Falk, Gikas Hardouvelis, Scott Hein, Doug Pearce, participants in seminarsat the 1986 Southern Economics Association and at Washington University-St. Louis and an anonymous referee for comments on an earlier version. Nancy Juen provided excellent researchassistance. An earlierversion of this paper was completed while Hafer was at the FederalReserveBank of St. Louis. The usual disclaimer applies. 1. This statementshould be tempered somewhat. As we will show in section III, Roley [21; 22] tested for a break in February 1980 when the money announcement was moved from Thursday to Friday and when the expected money survey was changed from twice a week (Tuesday and Thursday) to once a week (Thursday). In addition,severalothers [4; 14; 17] use approaches similar to that employed here. 577