On the Random Walk Characteristics of Short- and Long-Term Interest Rates In an Efficient Market

As IS NOW WELL KNOWN, tests of market efficiency that asset prices rapidly and fully reflect all relevant information are inevitably tests of joint hypotheses. In each test, a particular model of market equilibrium examined simultaneously with the question of market efficiency. Early studies of the behavior of stock prices, however, often failed to make explicit the underlying model of market equilibrium. In particular, those who studied the random walk hypothesis typically failed to make explicit their assumption that equilibrium returns are constant over time. Only under this assumption does evidence on the autocorrelation of successive one-period returns bear directly on the question of market efficiency. Recently, Phillips and Pippenger [9] have called attention to a body of evidence that suggests that interest rates may follow a random walk. They then note that this evidence is consistent with the hypothesis that capital markets are [9, p. 1 1] . By implication, they suggest that if interest rates do not follow a random walk, then the bond market not efficient. Poole [11, p. 476], too, suggests that random walk behavior of interest rates to be expected in an efficient market. The premise of this paper that, if the shortcomings of the early work on stock prices are to be avoided, researchers must exercise great caution in linking random