Multiscale Investment Portfolios’ Nonsystematic Bias Analysis And Its Calibration

While using sample data to calculate the return and variance of portfolio in different investment horizons, the traditional finance research has usually chosen a random sample scale to measure multiscale portfolio performance. Actually, This simple processing has implied an assumption that the expected return and variance of portfolio is independent of the timescale. However, because of the complexity of investment environment in the real investment practice, investors have to constantly adjust the asset investment period, which has caused the investors to ignore the timescale effects while measuring the intertemporal investment portfolio performance, arbitrarily select the benchmark timescale to calculate the portfolio performance to make the corresponding investment decisions, and finally results in a nonsystematic bias in dec is ion-making. Therefore, how to reasonably measure the multiscale portfolio performance is one of the high practically significant problems in the fields of financial economics.This paper mainly analyzes the nonsystematic bias of investment portfolio performance measure under the condition of multiscale and discusses its calibration empirically. Firstly, the paper has proposed the research value and significance, as well as put forward the corresponding research ideas and methods from the theoretical divergence and investment practice confusion. From the three perspectives of intertemporal portfolio performance measure, financial market multiscale research framwork and relations between investment horizon and portfolio performance, we have logically reviewed the relations among investment horizon, portfolio performance and its measurement. The third part is about the measure of multiscale portfolio performance and nonsystematic bias analysis. Based on the improvement of the calculation of Sharpe ratio(SR), we’ve divided into two cases(Tk> T0and Tk<T0) to analyze the nonsystematic bias by the comparison of theoretical Sharpe ratio(TSR) and real Sharpe ratio(RSR), and the underlying rules are recognized. Through the design of bias calibration functions on the timescale T, the paper has compared the actual calibration effects towards different functions exemplified by CSI300index, and has also acquired the single-scale and multiscale efficient frontiers. The conclusion of empirical research is that on the condition of Tk>T0, GTSR> RSR and while Tk<T0, RSR> GTSR and The calibration function based on Taylor expansion is more ideal than that of binomial fittings, and the multiscale efficient frontier present a nonlinear characteristics along with the timescale.Through the empirical study of the multiscale portfolio performance measurementnonsystematic bias using CSI300index, the aim of this paper is to put forward thecorresponding bias calibration ideas and methods to help investors calibrate thenonsystematic bias in the process of multiscale portfolio performance measure, andthereby accurately guiding investors’ decision-making and risk management.