Portfolio optimization has long been used in asset management to mitigate risks of fluctuating asset prices. In this study, we use copula models and portfolio optimization to investigate how inventory financing providers (IFP) can utilize the timely market information of collaterals to optimize their portfolios of collaterals to mitigate default risks. Through comparing the predictive performance of copula strategies with that of the multivariate normal distribution (MVN) strategy, we find that the general canonical vine copula can characterize the dependence structure among collateral return series and has superior predictive performance over the MVN and other copulas. Our findings suggest that the general canonical vine copula can be constructed into portfolio strategies that can be adopted by the IFP to mitigate default risks and improve their risk profile.

投资组合优化长期以来一直用于资产管理，以减轻资产价格波动的风险。在本研究中，我们使用 copula 模型和投资组合优化来研究库存融资提供商 (IFP) 如何利用抵押品的及时市场信息来优化其抵押品组合以降低违约风险。通过比较copula策略与多元正态分布（MVN）策略的预测性能，我们发现一般规范藤copula可以表征抵押收益序列之间的依赖结构，并且比MVN和其他copula具有更好的预测性能。