商品期货投资组合与市场收益的尾部相依研究——基于混频分位点相协模型

Abstract

Diversified allocation of asset portfolios is a core topic in investment management, and risk factor portfolios are important references for investors' asset allocation choices. Combining the idea of mixed data sampling models, this paper extends and constructs the qpr-MIDAS (quantile-specific probability ratio-Mixed Data Sampling) quantile association model. Based on the Chinese commodity futures market, the paper uses daily high-frequency commodity indices and stock indices to characterize the tail risk dependence between weekly risk factor portfolios and the market factor. From the model perspective: the qpr model combined with MIDAS can effectively characterize tail dependence with better fitting performance; in economic applications, the model can more effectively achieve diversified allocation of investment portfolios. The empirical results show that: (1) the risk spillover relationship between risk factor portfolios and market factor returns is relatively independent; (2) the risk spillover effect among factor portfolios has an obvious negative co-movement with commodity index and stock index returns; (3) the risk transmission effect weakens with the relaxation of market access policies. This research provides theoretical reference for investors' commodity futures asset allocation and portfolio design considering extreme risk conditions.