The purpose of the thesis has been to explore the use of hedge fund styles when constructing portfolios of hedge funds (i.e. funds of hedge funds). The central question is if the use of hedge fund styles can significantly explain and improve risk adjusted returns (characterized by Sharpe ratios). The study has been done in collaboration with Optimized Portfolio Management AB who desire further knowledge and evaluation of hedge fund styles for their fund of hedge funds.To be able to create successful ex ante portfolios we have explored various prediction models for both risk and return. Our findings indicate that return prediction is problematic using simple models such as regression since the risk exposure of the indices appear to change significantly over time. One can however using exponentially weighted moving averages (EWMA) achieve relatively promising estimations of future returns. Covariance matrix estimation seems to be more straightforward. We have achieved promising results using both traditional EWMA models as well as improved estimators using principal component analysis.Covariance prediction models were evaluated separately using a minimum-variance portfolio optimization technique and provided a significant risk reduction compared to the aggregated hedge fund universe (represented by a naively diversified portfolio). Combinations of risk and return prediction models were evaluated using traditional mean-variance portfolio construction methods, which were optimized for Sharpe ratios. These provided a significant increase in risk adjusted returns relative to the aggregated hedge fund universe. The allocation is however discouraging due to serious instability over time.Our findings indicate that there indeed is an advantage of taking hedge fund styles into consideration when constructing funds of hedge funds in a risk adjusted perspective. However, further research into return prediction needs to be done in order to stabilize portfolio allocation. An alternative seems to be tactical style allocation on a more fundamental analysis basis.
| Identifer | oai:union.ndltd.org:UPSALLA1/oai:DiVA.org:liu-2759 |
| Date | January 2005 |
| Creators | Håkansson, Matilda |
| Publisher | Linköpings universitet, Ekonomiska institutionen, Ekonomiska institutionen |
| Source Sets | DiVA Archive at Upsalla University |
| Language | Swedish |
| Detected Language | English |
| Type | Student thesis, info:eu-repo/semantics/bachelorThesis, text |
| Format | application/pdf |
| Rights | info:eu-repo/semantics/openAccess |
Page generated in 0.0022 seconds