Randomness of stock market movement: a nonparametric approach

Abstract

In the weak form of the efficient market hypothesis (EMH), it is assumed that stock prices move in a random fashion. In this study, the focus is on the national stock indices rather than on the stock prices, and to determine whether the stock indices behave in a manner consistent with the random walk. Due to the problem of nonnormality with the data under study, this study employed various nonparametic tests as opposed to many previous studies which employed parametric tests. The data for this study consist of daily, weekly and monthly closing indices of the KLSE Industrial (Malaysia), Hang Seng (Hong Kong), Nikkel Dow Jones (Tokyo), Dow Jones Industrial Average (New York), Australian All Ordinaries (Sydney), and Financial Times Industrlal Ordinaries (London). The results of thIs study indicate that daily changes do deViate from the random walk for smaller markets of Malaysia and Australia. In addition, the results are not quite consistent or stable over time for these smaller markets. For weekly data, only the market of MalaySia shows nonrandomness for the period 1984-1988. The Cox-Stuart test for trend usmg monthly data failed to detect a sIgnificant trend In the movements of these markets (except for the market of MalaYSia). Overall, this study shows that the random walk model is still valid for the national stock indices just like in the case of stock, especially in the more established and active world's markets. However, smaller and "not-so-actlve" stock markets are not quite consistent in terms of their conformity to the random walk.