ABSTRACT
Tests of market efficiency using technical trading rules generally compare the returns obtained from the trading rules to the returns from a simple buy-and-hold strategy. Most of the tests in the literature are informal, simple comparisons of the returns from the two strategies and involve no statistical tests of hypotheses. To obtain a clear-cut result from comparing the returns from the trading rule and from buy-and-hold, however, requires an appropriate statistical test. The literature has provided few statistical tests to evaluate trading rules relative to buy-and-hold strategies. Moreover, some of these tests have important limitations and deficiencies. The most useful, but still simple, test is probably the X-test discussed below. The X-test can be extended in a variety of ways to cover single-asset speculation or simultaneous speculation on several assets or a portfolio of assets. It gives a measure of risk-adjusted profits from speculation. Further, these are real risk-adjusted profits. Later chapters report X-statistics of from 3 percent per year to over 5 percent per year; these are impressive in a world where estimates of real rates of return on riskless assets are essentially zero and of the real return on the market are on the order of 8 percent to 10 percent.