Abstract:
The present study investigated weak-form market information efficiency in Tehran security exchange (TSE) as an emerging market and in Dow Jones United States security exchange (DJUS) as a developed market based on random walk model. In each market, the random walk model was examined using daily and monthly returns of a set of indices. The results of the parametric and non-parametric tests indicated that the daily returns are not independent and identically distributed in TSE. Moreover, according to the results of the variance ratio test, a trending behavior in daily returns and mean-reversion behavior in monthly returns were observed. In DJUS, however, the daily returns were found to be independent and identically distributed and the results of variance ratio test did not confirm that the returns follow a particular pattern in this market.
Machine summary:
"In another study, Fadainejad (1974) presented some pieces of evidence for the rejection of both weak and semi-strong efficient markets hypotheses using serial correlation and normal distribution test.
(2012) provided evidence for the rejection of random walk model and weak-from market efficiency in TSE using a variety of variance ratio tests.
Perhaps the simplest version of the random walk hypothesis is the independently and identically distributed (IID) increments case in which the dynamics of /is obtained by the following equation: / (5) Where /represents stock returns, / is the expected price change or drift, and /shows that / is independently and identically distributed with mean 0 and variance /.
5. The Research Tests In this study, Ljung-Box test, runs test, Lo-MacKinlay variance ratio test, and Wright`s variance ratio test were employed to test the independent and identically distribution of the returns according to random walk hypotheses (Campbell et al.
The null hypothesis of Wright`s (2000) ranked-based variance ratio test states that the logarithm of prices follows the random walk model and / is independent and identically distributed.
In effect, the null hypothesis of Wright`s sign-based variance ratio test states that the logarithm of the price follows random walk model.
5. Accordingly, Wright (2000) presented the sign-based test statistics under null hypothesis assuming homoscedastic error term and as follows: and (23) (24) Table 6: Wright Variance Ratio Test Results Tehran Security Exchange به تصویر صفحه مراجعه شود)) به تصویر صفحه مراجعه شود) ) Note: In each lag, the first row represents the test statistic under the assumption of homoscedastic error term and the second row represents the test statistics under the assumption of heteroscedastic error term."