Normalizing Risk Measures in Risk-based Portfolios Through Covariance Misspecification Error Analysis
Year: 2024
Type: article
Abstract: The core focus of this article is to examine and test the allocation strategies utilized in portfolio management for weight determination. Our investigation centers around the covariance matrix, a crucial input of the risk-based portfolios, including minimum volatility, inverse volatility, equal risk contribution and maximum diversification portfolios.The principal aim of this thesis is to analyze the robustness and sensitivity of different portfolios when confronted with potential misspecifications or errors in the covariance matrix.To accomplish this, we employ a Dynamic Conditional Correlation model and utilize a Monte Carlo approach to generate a substantial number of covariance matrices. When we use risk-based strategies on the simulated matrices, we can assess the robustness of each allocation methodology by calculating the distance between the true and simulated allocations.Following this, we estimate the probability of model accuracy to adjust two ’conventional’ risk measures, which represent the performance of risk-based portfolios, by incorporating the risk of covariance misspecification in their value. The resulting enhanced risk measures are more likely to be reliable, as they are normalized based on the probability of obtaining them in the absence of misspecification.
DOI: 10.2139/ssrn.4758251
Authors: Enrico Sergi, Claudio Conversano
Institution: University of Cagliari
Language: English
Field: Decision Sciences
Domain: Social Sciences
Open Access status: diamond
OpenAlex ID: W4414995035