Structural Breaks, Long Memory, or Unit Roots in Stock Prices: Evidence from Emerging Markets

This paper investigates whether daily stock price indices from fourteen emerging markets are random walk (unit root) or mean reverting long memory processes. We use an efficient statistical framework that tests for random walks in the presence of multiple structural breaks at unknown dates. This approach allows us to investigate a broader range of persistence than that allowed by the I(0)/I(1) paradigm about the order of integration, which is usually implemented for testing the random walk hypothesis in stock market indices. Our approach extends Robinson's (1994) efficient test of unit root against fractional integration to allow for multiple endogenously determined structural breaks. For almost all countries, we find support for the random walk hypothesis, with the exception of four stock markets, where weak evidence of mean reverting long memory exist. Structural breaks have impact on the unit root behavior only for Mexico; for all other 11 markets unit roots exist even when structural breaks are not taken into account. In order to check the robustness of our results, we use the two-step feasible exact local Whittle (FELW2ST) estimator of Shimotsu (2010), which allows for polynomial trends, non-normal distributions, and non-stationarity. The results from the semi-parametric FELW2ST approach shows that, except for Mexico, stock price indices of 13 emerging markets are not mean reverting.