Beating the Foreign Exchange Market

摘要: Filter rule profits found in foreign exchange markets the early days of current managed float persist later periods, as shown by statistical tests developed and implemented here. The test is consistent with, but independent of, a wide variety asset pricing models. cannot be explained risk if premia are constant over time. Inclusion home-foreign interest rate differential computing has little effect on comparison filter returns to those buy-and-hold. IN THE EARLY YEARS generalized floating that began March 1973, excess buy-and-hold were for many countries (Logue, Sweeney Willett [18], Dooley Shafer [6], Cornell Dietrich [5]). It was unclear, however, whether such indicated inefficiencies. First, it not clear adequately handled tests, often ignored. Second, there no evidence would available postsample. And third, significance these profits. This paper uses logic risk/return tradeoffs analyze role illustrates discussion using Sharpe-Lintner Capital Asset Pricing Model (CAPM), although Breeden consumption-based CAPM, Merton intertemporal or an Arbitrage (APM) could easily used. develops appropriate (Section I), analyzing return speculation less foreign-domestic differential. explicitly assumes sample. rates observable filters going from risk-free dollar Deutsche Mark (DM) into 1980's II), this true even when taking account transactions costs. turns out results $/DM case do depend very much interestrate differential, primarily behavior. useful know because quite difficult find matching, interest-rate data high quality. For sample nine other currencies, only rates, Section III shows made first 610 next 1,220 days, 1980's. (Dooley [7] similar persistence experiments use have tests.) When been spot markets, argued due risk. CAPM both strategies. * Claremont McKenna College Graduate School. Arthur D. Warga, Douglas Joines, Thomas offered helpful comments, Ning-Ning Koo provided research assistance.