Next Generation Models for Portfolio Risk Management: An Approach Using Financial Big Data

TL;DR

This paper introduces a novel dynamic risk measurement model leveraging financial big data to enhance portfolio risk analysis, addressing high-dimensional challenges and outperforming benchmark methods in empirical tests.

Contribution

It is the first to utilize financial big data for improving the accuracy of portfolio risk measurement, overcoming the curse of dimensionality.

Findings

Financial big data improves small portfolio risk analysis.

The proposed model outperforms benchmark approaches.

Benefits are most significant when incorporating extensive asset information.

Abstract

This paper proposes a dynamic process of portfolio risk measurement to address potential information loss. The proposed model takes advantage of financial big data to incorporate out-of-target-portfolio information that may be missed when one considers the Value at Risk (VaR) measures only from certain assets of the portfolio. We investigate how the curse of dimensionality can be overcome in the use of financial big data and discuss where and when benefits occur from a large number of assets. In this regard, the proposed approach is the first to suggest the use of financial big data to improve the accuracy of risk analysis. We compare the proposed model with benchmark approaches and empirically show that the use of financial big data improves small portfolio risk analysis. Our findings are useful for portfolio managers and financial regulators, who may seek for an innovation to improve the accuracy of portfolio risk estimation.