Hedge fund replication: Difference between revisions
John E Smith (talk | contribs) Added small piece regarding current investment level. Hedge Fund Replication is a valid method of investment that is gaining traction, and I see no reason why this page should be deleted. |
spelling |
||
| Line 1: | Line 1: | ||
'''Hedge Fund Replication''' is the collective name given to a number of different |
'''Hedge Fund Replication''' is the collective name given to a number of different methods that attempt to replicate hedge fund returns. As the hedge fund industry has boomed over recent years, so more participants{{Who}} have begun to question the ability of hedge fund managers in aggregate to produce alpha.{{What}} Replication removes the illiquidity, transparency and fraud risk associated with direct investment in hedge funds. With the belief that alpha is a zero-sum gain, more investors are looking to simply add "Hedge Fund Beta" to their portfolio. So far, around $5 billion is estimated to have been allocated to replication strategies, including around $400 m by the USS Pension Scheme. These early investors have been rewarded as the replicators outperformed their direct investment cousins in 2008 due to their greater liquidity and lower use of leverage. |
||
Consensus has not yet been reached as to the best methodology to replicate hedge fund returns. Currently, there are three schools of thought, as follows<ref>http://www.hedgefundreplication.com</ref>: |
|||
| ⚫ | |||
| ⚫ | |||
The most intuitive replication methodology essentially looks at each hedge fund strategy in isolation, and qualitatively asking the question: What is the manager trying to do to generate returns? For each hedge fund strategy a replication strategy is put in place, that attempts to mimic what the hedge fund manager is doing in a mechanical fashion. Consequently, one would expect to generate the beta, but not the alpha, of that particular strategy. This process is carried out for each of the hedge fund strategies, and then these strategies are combined to produce a product that attempts to replicate the entire hedge fund universe. |
The most intuitive replication methodology essentially looks at each hedge fund strategy in isolation, and qualitatively asking the question: What is the manager trying to do to generate returns? For each hedge fund strategy a replication strategy is put in place, that attempts to mimic what the hedge fund manager is doing in a mechanical fashion. Consequently, one would expect to generate the beta, but not the alpha, of that particular strategy. This process is carried out for each of the hedge fund strategies, and then these strategies are combined to produce a product that attempts to replicate the entire hedge fund universe. |
||
== Factor |
== Factor replication == |
||
The factor replication method attempts to replicate the return stream of the hedge fund universe. The simplest way of doing this is by carrying out a linear regression, using one of the headline hedge fund indices and a number of factors. The model selects the factors that have the highest explanatory powers for the index returns, and then each month the model will perform the regression analysis over a rolling time frame, and select the weightings for each of these factors. This method has been criticised for being backward looking and taking too long to adjust to changes in hedge fund allocations. Improvements to this method have been suggested, such as using Kalman or Particle Filters, which improve the speed at which the model reacts to change. |
The factor replication method attempts to replicate the return stream of the hedge fund universe. The simplest way of doing this is by carrying out a linear regression, using one of the headline hedge fund indices and a number of factors. The model selects the factors that have the highest explanatory powers for the index returns, and then each month the model will perform the regression analysis over a rolling time frame, and select the weightings for each of these factors. This method has been criticised{{By whom}} for being backward looking and taking too long to adjust to changes in hedge fund allocations. Improvements to this method have been suggested, such as using Kalman or Particle Filters, which improve the speed at which the model reacts to change. |
||
== Distribution |
== Distribution replication == |
||
The distribution replication method does not actually attempt to replicate hedge funds, but rather the distribution payoff that one associates with hedge funds. The original proponent of this methodology, Harry Kat, has in recent times changed tack and no longer pushes this as a "replication" method, but rather one of "creation". Essentially, the desired distribution properties are selected (risk, skew, kurtosis and correlation) together with a reserve asset, and the model will attempt to give returns that cause the distribution model to tend towards the required distribution over time. |
The distribution replication method does not actually attempt to replicate hedge funds, but rather the distribution payoff that one associates with hedge funds. The original proponent of this methodology, Harry Kat, has in recent times changed tack and no longer pushes this as a "replication" method, but rather one of "creation". Essentially, the desired distribution properties are selected (risk, skew, kurtosis and correlation) together with a reserve asset, and the model will attempt to give returns that cause the distribution model to tend towards the required distribution over time. |
||
| Line 18: | Line 17: | ||
== Hedge Fund Replication Index == |
== Hedge Fund Replication Index == |
||
A recent advancement of the replication community is the introduction of a benchmark index, called the Hedge Fund Replication Index, published by hedgefundreplication.com. This index represents all the main product providers, and analysis of this index shows that around 85% of all hedge fund returns can be replicated. Further research shows that the fees associated in investing via a fund of hedge funds are actually greater than the alpha that the individual hedge funds produce, therefore the average fund of hedge funds will underperform the average replicator. The |
A recent advancement of the replication community is the introduction of a benchmark index, called the Hedge Fund Replication Index, published by hedgefundreplication.com. This index represents all the main product providers, and analysis of this index shows that around 85% of all hedge fund returns can be replicated. Further research{{By whom}} shows that the fees associated in investing via a fund of hedge funds are actually greater than the alpha that the individual hedge funds produce, therefore the average fund of hedge funds will underperform the average replicator. The index has a correlation of 0.93 to the broader hedge fund indices. |
||
| ⚫ | |||
| ⚫ | In time, many industry experts believe that hedge fund replication will become a core holding of an investor's hedge fund allocation, with investments in |
||
| ⚫ | |||
| ⚫ | In time, many industry experts{{Who}} believe that hedge fund replication will become a core holding of an investor's hedge fund allocation, with investments in individual hedge funds being seen as satellite holdings. This will mirror the growth in index-tracking mutual funds and ETFs in the long only world. |
||
== References == |
== References == |
||
Revision as of 06:11, 6 December 2009
Hedge Fund Replication is the collective name given to a number of different methods that attempt to replicate hedge fund returns. As the hedge fund industry has boomed over recent years, so more participants[who?] have begun to question the ability of hedge fund managers in aggregate to produce alpha.[clarification needed] Replication removes the illiquidity, transparency and fraud risk associated with direct investment in hedge funds. With the belief that alpha is a zero-sum gain, more investors are looking to simply add "Hedge Fund Beta" to their portfolio. So far, around $5 billion is estimated to have been allocated to replication strategies, including around $400 m by the USS Pension Scheme. These early investors have been rewarded as the replicators outperformed their direct investment cousins in 2008 due to their greater liquidity and lower use of leverage.
Consensus has not yet been reached as to the best methodology to replicate hedge fund returns. Currently, there are three schools of thought, as follows[1]:
Mechanical trading strategies
The most intuitive replication methodology essentially looks at each hedge fund strategy in isolation, and qualitatively asking the question: What is the manager trying to do to generate returns? For each hedge fund strategy a replication strategy is put in place, that attempts to mimic what the hedge fund manager is doing in a mechanical fashion. Consequently, one would expect to generate the beta, but not the alpha, of that particular strategy. This process is carried out for each of the hedge fund strategies, and then these strategies are combined to produce a product that attempts to replicate the entire hedge fund universe.
Factor replication
The factor replication method attempts to replicate the return stream of the hedge fund universe. The simplest way of doing this is by carrying out a linear regression, using one of the headline hedge fund indices and a number of factors. The model selects the factors that have the highest explanatory powers for the index returns, and then each month the model will perform the regression analysis over a rolling time frame, and select the weightings for each of these factors. This method has been criticised[by whom?] for being backward looking and taking too long to adjust to changes in hedge fund allocations. Improvements to this method have been suggested, such as using Kalman or Particle Filters, which improve the speed at which the model reacts to change.
Distribution replication
The distribution replication method does not actually attempt to replicate hedge funds, but rather the distribution payoff that one associates with hedge funds. The original proponent of this methodology, Harry Kat, has in recent times changed tack and no longer pushes this as a "replication" method, but rather one of "creation". Essentially, the desired distribution properties are selected (risk, skew, kurtosis and correlation) together with a reserve asset, and the model will attempt to give returns that cause the distribution model to tend towards the required distribution over time.
Hedge Fund Replication Index
A recent advancement of the replication community is the introduction of a benchmark index, called the Hedge Fund Replication Index, published by hedgefundreplication.com. This index represents all the main product providers, and analysis of this index shows that around 85% of all hedge fund returns can be replicated. Further research[by whom?] shows that the fees associated in investing via a fund of hedge funds are actually greater than the alpha that the individual hedge funds produce, therefore the average fund of hedge funds will underperform the average replicator. The index has a correlation of 0.93 to the broader hedge fund indices.
The future of hedge funds
In time, many industry experts[who?] believe that hedge fund replication will become a core holding of an investor's hedge fund allocation, with investments in individual hedge funds being seen as satellite holdings. This will mirror the growth in index-tracking mutual funds and ETFs in the long only world.