|
| Portfolio Construction |
 |
| Introduction |
| To create a portfolio with hedge funds,
mutual funds, ETF, stocks or bonds, qualitative and quantitative analysis should be performed. The qualitative analysis is the most important part. Quantitative analysis
has more and more importance in the Portfolio construction
process, because academic research papers find new ways to analyze hedge fund risks. |
| |
| Qualitative Analysis |
| For each hedge fund strategy, the qualitative analysts select the hedge funds which best fulfill the criteria as: |
| • |
Who are the manager and his traders? |
| • |
Is the strategy able to generate the promised returns? |
| • |
How do they react to changing conditions? |
| • |
Are they still open to new investment? |
| • |
What is the hedge fund leverage? |
| • |
Do they use options to hedge? |
| • |
Do they use leverage? |
| • |
Do they have a risk management system in place? |
| • |
Do the returns reflect the strategy? |
| • |
What is the hedge fund extreme risk hedge? |
| • |
What is the general degree of transparency of the hedge fund? |
| • |
What are the investment strategies? |
| • |
Does the hedge fund follow consistently its style? |
| • |
Who is the broker and administrator? |
| • |
And more ... |
|
| |
| The qualitative analyst should answer these questions. Each bank investing in hedge funds has a team of analysts. To their qualitative views, they will add quantitative criteria like past returns, past standard deviation, past semi-standard deviation, past linear correlation, drawdown, maximum monthly loss, percent of positive monthly returns, behavior of the hedge funds during the equity/bond market
crashes or maximum time to recover from a loss. This job can be performed with the software platform from AlternativeSoft. |
| |
 |
Source: AlternativeSoft's platform. |
|
| |
| Quantitative Analysis |
| The problems with the quantitative method of selecting hedge funds is that the hedge fund manager strategy can change or there are not enough monthly data to compute statistics. We propose four steps to construct portfolios with hedge funds, namely: |
| • |
Portfolio extreme risk analysis |
| • |
Portfolio and hedge fund style sensitivity |
| • |
Hedge fund portfolio optimization |
| • |
Hedge fund portfolio simulation |
|
 |
| Extreme Risks Analysis |
| To analyze the hedge fund risk, the volatility is not enough as autocorrelation, liquidity risk, and non-normality are common features in those assets. We propose the use of 5 new risk measures: |
| • |
The Modified Value-at-Risk, which accounts for volatility, skewness and kurtosis. |
| • |
The Conditional Value-at-Risk, which accounts for the extreme negative portfolio returns. |
| • |
The bear correlation, which accounts for equity or bond crashes. |
| • |
The non-linear regression technique between the hedge fund and a benchmark to measure the hedge fund exposure to option strategies. |
|
| |
| To analyze these
four extreme risks, we recommend to use the AlternativeSoft's software platform. |
|
Source: AlternativeSoft's platform. |
|
| |
| Style Analysis |
| In this section, the aim is to determine in which strategy the hedge fund is invested. There are two ways to do that: |
| • |
A principal component analysis. According to Fung and Hsieh, Performance Attribution and Style Analysis, 1996, p.18, they extract 5 orthogonal factors, which explains 50% of the return's standard deviation: distressed, macro, opportunistic, value and market timing. |
| • |
A style analysis versus an equity index, a bond index, a commodity index, a money market index, out-of-the money put option, out-of-the money call option as independent variables and a stepwise regression technique. Other factors such as interest rates, exchange rates, a credit
index or the VIX index can be used. In hedge funds, the risks are concentrated in equity market and liquidity. |
|
| |
| To analyze hedge funds or portfolio style exposure, AlternativeSoft's software platform is available. |
|
|
Source: AlternativeSoft's platform. |
|
| |
| Portfolio Optimization |
| In the previous section, we measured the hedge fund's downside risks and verified their sensitivities to economic factors. The next step is to construct an optimal portfolio with an optimization. |
| |
| To do that, we select the hedge fund candidates that pass the qualitative and quantitative screening processes described above. We insert some minimum and maximum weights' constraints, beta constraints, hedge funds' expected returns, hedge funds' expected volatility. Then, we optimize with a technique that accounts for extreme risks. Several techniques are available to optimize with extreme risks: |
| • |
Bear correlation minimization |
| • |
Conditional Value-at-Risk minimization |
| • |
Modified Value-at-Risk minimization |
| • |
Maximum Drawdown minimization |
| • |
Omega optimization |
|
| |
| We recommend using optimization techniques which account at least for mean, variance, skewness and kurtosis. It has been shown in academic papers that the investor is rewarded for purchasing negative skewness. |
| |
| To optimize a hedge fund portfolio, AlternativeSoft's software platform is available. |
 |
Source: AlternativeSoft's platform. |
|
| |
| Portfolio Simulation |
| The aim of the paragraph is to simulate the optimal portfolio computed above, in the future. With hedge funds, past returns are not always available or the time series have short history. We propose to take the risk from the historical time series, to use the hedge fund expected returns and to simulate the hedge funds' portfolio, in the future. To perform this task, we assume: |
| • |
The past correlation between the hedge funds will be the same, in the future |
| • |
The risk is stable (i.e. stationary assumption), in the future |
| • |
The hedge fund expected returns are a good approximation of what the future will be |
| • |
The investor is risk averse against extreme negative hedge fund returns |
|
| |
| A portfolio simulation gives the investor a view of the potential large future drawdowns, his returns' probabilities and probabilities to be above a certain return. |
| |
| To simulate a hedge fund portfolio, AlternativeSoft's software platform is available. |
 |
Source: AlternativeSoft's platform. |
|
| |
| Conclusion |
| We have shown that qualitative analysis is the main part in portfolio construction with hedge funds. Then, we have shown that traditional risk instruments have to be more accurate, especially to measure downside risks. We proposed a sensitivity approach to determine hedge funds' sensitivity to economic factors. We explained that a portfolio optimization should be performed in order to find the optimal hedge fund weights. Finally, we said that a simulation is required to whether the hedge fund portfolio answered the investor's needs. |
| |
 |
