Nonlinear properties of the Korea fund market

Abstract

This paper attempts to analyze the return time series of both equity and balanced funds by using a statistical physics approach to estimate the mechanism of fund performance. Our empirical results suggest that for equity and balanced fund data, the Hurst exponent of the return time series is approximately 0.5 whereas the volatility data yield a long term correlation with the Hurst exponent of 0.9. After the temporal correlation and nonlinearity has been eliminated by using the shuffle and surrogate methods, the Hurst exponent is approximately 0.5. In addition, we measure the complexity by using the multifractality, which can help us to understand the mechanism of the fund performance. We find the presence of multifractality regardless of the data used, which is attributed to extreme events and temporal correlation. In particular, the degree of multifractality of equity funds is very similar to that of the balanced fund even though the funds are characterized by differences in the degree of risk.