A frequently asked questions in the hedge fund literature is 'What are the systematic risk factors in hedge fund returns?". Existing efforts can be classified as 'bottom-up' or 'top-down', i.e. analysing specific styles of funds or taking a portfolio approach, respectively. In my first essay, I take a 'bottom-up' approach and analyse convertible arbitrage (CA) returns. The magnitudes of their reported returns suggest that there are severe inefficiencies in the pricing of convertible bonds. Alternatively, CA excess returns may be compensation for exposure to extraordinary events, i.e. market crashes, and may be related to systematic risk in a nonlinear way. To overcome database biases, statistical issues, the dynamic use of leverage, etc., I replicate three core CA strategies and show that these adequately represent real investor experiences. Panel regressions show that the replicated strategies have risk exposures that can be related to their construction. To overcome the limitations of a linear framework, I show that once crash risk is hedged with options while maintaining the same ex-ante exposure to crashes, excess returns are statistically indistinguishable from zero. CA therefore partly or fully represents compensation for systematic risk. In the second essay we take a 'top-down' approach and analyse whether accounting for nonlinear relationships between hedge funds and the market in the construction of a portfolio is beneficial for a myopic investor. We expand the classic mean-variance framework and dynamically optimise a portfolio based on time varying moments. Nonlinear relationships enter by allowing for different correlations conditional on market movement. We show that an investor is indeed better off in terms of risk and return if he accounts for correlation asymmetries. His portfolio returns are less negatively skewed, less kurtosed and have a lower turnover. The driving factor appears to be that less capital is allocated to non-directional, arbitrage-style hedge funds.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:551331 |
| Date | January 2011 |
| Creators | Siepman, Marvin |
| Publisher | London School of Economics and Political Science (University of London) |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
Page generated in 0.0024 seconds