演講摘要 : In this paper we propose a new class of tests for the martingale difference hypothesis. In contrast with the autocorrelation-based tests and the spectrum-based tests, the proposed test requires a weaker moment condition and has power against a much larger class of non-martingale difference alternatives that may be serially correlated or uncorrelated. Moreover, this test does not rely on the assumption of conditional homoskedasticity. As compared with many existing consistent tests, this test is easy to implement and has a standard limiting distribution. Our simulations confirm that the proposed test is powerful against various linear and nonlinear alternatives and is quite robust to the failure of higher-order moments. Our empirical study on exchange rate returns also shows that the conclusion resulted from the proposed test is different from that of the conventional tests.
