Persistence and efficiency of OECD stock markets: linear and nonlinear fractional integration approaches

The analysis of the dynamics of the share price series using the fractional integration technique that accommodates nonlinear deterministic trend based on the Chebyshev polynomials in time is considered in this study. It is important to account for the nonlinear behavior and autoregressive structure of the error disturbance term of the share price indices if correct estimates of the integration order are to be obtained. The findings of this study show that it is important to put the autoregressive order into consideration following the significant effect it imposes on the fractional integration estimates, and there are evidences of nonlinearities in nine countries. The nonlinear effect significantly imparts on the degree of persistence of the nine countries by either inducing more or less persistence. However, it is more profound for Belgium, Japan and Hungary whose share market statuses change from efficient to inefficient, as their d estimates under the linearity case, respectively, fall from 0.9715, 0.9998 and 0.9344 to 0.3230, 0.3112 and 0.8579 under the nonlinearity case. For the sake of policy, these findings have important implications for portfolio management and diversification, investment choices, and policy making and forecast.