Performance of the Trading Strategies with Risk Constraints in Finnish Market
Salkanovic, Ibro (2005)
Kuvaus
Kokotekstiversiota ei ole saatavissa.
Tiivistelmä
The main aim of this master’s thesis was to compare two risk constraint methodologies Value-at-Risk and Expected Shortfall on the sample of Finnish stocks from the Helsinki Stock Exchange in order to determine a more suitable model for Finnish stock market.
Theoretical part presented main theories explaining the problems with which these two risk constraining methodologies are faced. The term coherence was introduced and explained. Risk limit approach was also presented.
The data comprised of daily returns in the period of 01.01.1997-31.12.2004 and it was divided into portfolios using Sharpe ratio. Mean-variance optimization and equal weights were used as asset allocation methodologies. As proxy for riskless asset Finnish Government T-Bills were used. Data was tested with Jarque-Bera Test, Jensen’s Alpha and Chow Test. To determine significance t-Test and p-value were used.
Results showed that Value-at-Risk risk constraining model yields better returns than Expected Shortfall risk constraint, given the assumptions made. Assumed behavior mirrored Finnish investment fund. There was no evidence that risk constraining methodologies can improve profitability from choosing a passive strategy of investing into the HEX-Market indices. It was also shown that dynamical adjusting of portfolios yields better returns than buy-and-hold strategy if the portfolio allocation methodology is mean-variance optimization.
The results are in most part consistent with previous literature. Surprising result was that Expected Shortfall constraining methodology was inferior strategy to Value-at-Risk regardless of the allocation strategy.
Theoretical part presented main theories explaining the problems with which these two risk constraining methodologies are faced. The term coherence was introduced and explained. Risk limit approach was also presented.
The data comprised of daily returns in the period of 01.01.1997-31.12.2004 and it was divided into portfolios using Sharpe ratio. Mean-variance optimization and equal weights were used as asset allocation methodologies. As proxy for riskless asset Finnish Government T-Bills were used. Data was tested with Jarque-Bera Test, Jensen’s Alpha and Chow Test. To determine significance t-Test and p-value were used.
Results showed that Value-at-Risk risk constraining model yields better returns than Expected Shortfall risk constraint, given the assumptions made. Assumed behavior mirrored Finnish investment fund. There was no evidence that risk constraining methodologies can improve profitability from choosing a passive strategy of investing into the HEX-Market indices. It was also shown that dynamical adjusting of portfolios yields better returns than buy-and-hold strategy if the portfolio allocation methodology is mean-variance optimization.
The results are in most part consistent with previous literature. Surprising result was that Expected Shortfall constraining methodology was inferior strategy to Value-at-Risk regardless of the allocation strategy.