Refine
Document Type
- Working Paper (3)
- Article (1)
Language
- English (4)
Has Fulltext
- yes (4)
Is part of the Bibliography
- no (4)
Keywords
- Portfolio optimization (4) (remove)
Empirical evidence suggests that asset returns correlate more strongly in bear markets than conventional correlation estimates imply. We propose a method for determining complete tail correlation matrices based on Value-at-Risk (VaR) estimates. We demonstrate how to obtain more efficient tail-correlation estimates by use of overidentification strategies and how to guarantee positive semidefiniteness, a property required for valid risk aggregation and Markowitz{type portfolio optimization. An empirical application to a 30-asset universe illustrates the practical applicability and relevance of the approach in portfolio management.
Tail-correlation matrices are an important tool for aggregating risk measurements across risk categories, asset classes and/or business segments. This paper demonstrates that traditional tail-correlation matrices—which are conventionally assumed to have ones on the diagonal—can lead to substantial biases of the aggregate risk measurement’s sensitivities with respect to risk exposures. Due to these biases, decision-makers receive an odd view of the effects of portfolio changes and may be unable to identify the optimal portfolio from a risk-return perspective. To overcome these issues, we introduce the “sensitivity-implied tail-correlation matrix”. The proposed tail-correlation matrix allows for a simple deterministic risk aggregation approach which reasonably approximates the true aggregate risk measurement according to the complete multivariate risk distribution. Numerical examples demonstrate that our approach is a better basis for portfolio optimization than the Value-at-Risk implied tail-correlation matrix, especially if the calibration portfolio (or current portfolio) deviates from the optimal portfolio.
Tail-correlation matrices are an important tool for aggregating risk measurements across risk categories, asset classes and/or business segments. This paper demonstrates that traditional tail-correlation matrices—which are conventionally assumed to have ones on the diagonal—can lead to substantial biases of the aggregate risk measurement’s sensitivities with respect to risk exposures. Due to these biases, decision-makers receive an odd view of the effects of portfolio changes and may be unable to identify the optimal portfolio from a risk-return perspective. To overcome these issues, we introduce the “sensitivity-implied tail-correlation matrix”. The proposed tail-correlation matrix allows for a simple deterministic risk aggregation approach which reasonably approximates the true aggregate risk measurement according to the complete multivariate risk distribution. Numerical examples demonstrate that our approach is a better basis for portfolio optimization than the Value-at-Risk implied tail-correlation matrix, especially if the calibration portfolio (or current portfolio) deviates from the optimal portfolio.
Socially responsible investing (SRI) continues to gain momentum in the financial market space for various reasons, starting with the looming effect of climate change and the drive toward a net-zero economy. Existing SRI approaches have included environmental, social, and governance (ESG) criteria as a further dimension to portfolio selection, but these approaches focus on classical investors and do not account for specific aspects of insurance companies. In this paper, we consider the stock selection problem of life insurance companies. In addition to stock risk, our model set-up includes other important market risk categories of insurers, namely interest rate risk and credit risk. In line with common standards in insurance solvency regulation, such as Solvency II, we measure risk using the solvency ratio, i.e. the ratio of the insurer’s market-based equity capital to the Value-at-Risk of all modeled risk categories. As a consequence, we employ a modification of Markowitz’s Portfolio Selection Theory by choosing the “solvency ratio” as a downside risk measure to obtain a feasible set of optimal portfolios in a three-dimensional (risk, return, and ESG) capital allocation plane. We find that for a given solvency ratio, stock portfolios with a moderate ESG level can lead to a higher expected return than those with a low ESG level. A highly ambitious ESG level, however, reduces the expected return. Because of the specific nature of a life insurer’s business model, the impact of the ESG level on the expected return of life insurers can substantially differ from the corresponding impact for classical investors.