A Study Of Portfolio Scale-variable Beta Coefficient: Based On DMA Model

Capital Asset Pricing Model(CAPM),the cornerstone of modern finance, which is undergoing linear to nonlinear expansion。However,Nonlinear CAPM system has not yet formed. Based on previous study, this paper tries to improve Scale Variable CAPM(SVCAPM) and propose DMA-based Variable CAPM. When using SVCAPM to estimate the Beta coefficient of portfolio, noise of price volatility of linear low-frequency will be filtered off and the Beta coefficients of the model can only consider the nonlinear systemic risk, for the reason that they chose Detrended Fluctuation Analysis(DFA) and Detrended Cross-correlation Analysis(DCCA). If we use SVCAPM to estimate Beta coefficient of market prices or price index when there exists a marked linear trend portfolio fluctuation, the results will be underestimated or overestimated. With that in mind, we use Detrended Moving Average Analysis(DMA) to replace DFA, and Detrended Moving Average Cross-correlation Analysis(XDMA) to replace DCCA, to define a new SVCAPM. By introducing a new parameter, this article improves the filter of DMA and XDMA, which called double filters, to define DMA-based dual scale capital asset pricing model(DMA-TSCAPM). Empirical test shows that this model presents a higher degree of recognition of the systemic and also a better hedging effect of Beta coefficient.