X. Vidal-Llana, F. Morillas
In the asset pricing literature, a primary challenge is to achieve more accurate estimations of a firm's future returns than competing models, particularly through the analysis of extreme values. However, such estimations are inherently complex. For investors and regulators, estimating extreme values offers a novel perspective, revealing behavioral patterns and asset dependencies that are not observable in traditional expectation-based regressions. In this study, we extend the conventional framework of quantile regression to quantile-on-quantile regression and further to multivariate quantile-on-quantile regression. This approach allows for a more comprehensive estimation of market dependence structures and facilitates the differentiation of asset behaviors under varying market conditions. An application that compares market indexes using macroeconomic factors demonstrates the behavioral differences across each conditional distribution.
Keywords: Extreme values, Asset dependencies, Market dependence structures, Conditional distribution
Scheduled
AR1 Risk analysis II
June 12, 2025 3:30 PM
MR 1