Department of Economics and Business Economics

Modelling asset correlations during the recent financial crisis: A semiparametric approach

Research output: Working paperResearch


  • Rp10 71

    Final published version, 1.91 MB, PDF document

  • Nektarios Aslanidis, University Rovira Virgili, Spain
  • Isabel Casas, Denmark
  • School of Economics and Management
This article proposes alternatives to the Dynamic Conditional Correlation (DCC) model to study assets' correlations during the recent financial crisis. In particular, we adopt a semiparametric and nonparametric approach to estimate the conditional correlations for two interesting portfolios. The first portfolio consists of equity sectors SPDRs and the S&P 500 composite, while the second one contains major currencies that are actively traded in the foreign exchange market. Methodologically, our contribution is two fold. First, we propose the Local Linear (LL) estimator for the correlations instead of the standard Nadaraya Watson (NW) estimator used in Hafner et al. (2006) and Long et al. (2010). Second, we perform an extensive set of Monte Carlo experiments to compare the semiparametric and nonparametric models with the DCC specification. Unlike the aforementioned papers we also perform multivariate simulations in addition to the bivariate ones. Our simulation results show that the semiparametric and nonparametric models are best in DGPs with gradual changes or structural breaks in correlations. However, in DGPs with rapid changes or constancy in correlations the DCC delivers the best outcome. Finally, portfolio evaluation results show that the nonparametric model dominates its competitors, particularly in minimum variance weighted portfolios.
Original languageEnglish
Place of publicationAarhus
PublisherInstitut for Økonomi, Aarhus Universitet
Number of pages32
Publication statusPublished - 2010

    Research areas

  • Semiparametric Conditional Correlation Model, Nonparametric Correlations, DCC, Local Linear Estimator, Portfolio Evaluation

See relations at Aarhus University Citationformats

Download statistics

No data available

ID: 22380489