Approaches and Limits in Replicating Hedge Fund Returns
Summary
The note surveys ways to approximate a hedge fund’s return stream using a different, potentially passive strategy. It groups prior research into linear factor models, nonlinear or less transparent models, and work emphasizing the difficulty of reliable replication. Linear approaches include returns-based style analysis, which can estimate exposure to observable assets or factors. The answer also points to research on hedge fund returns and the possibility that some reported alpha may reflect exposure to systematic risk factors.
The discussion presents replication as a question whose feasibility depends on what is meant by “replicate.” Matching the beta component may be more achievable than reproducing the full return stream, while hidden strategies or nonlinear exposures can make a close clone difficult. It offers research references and broad schools of thought, but no empirical test, implementation steps, or evidence that a particular correlation target can be reached. The guidance is therefore a conceptual map rather than a validated replication recipe.
Key ideas
- Linear factor models can estimate a hedge fund’s exposures to observable return sources.
- Returns-based style analysis is one established framework for linear replication.
- Nonlinear and undisclosed strategies create additional challenges for passive replication.
- Some apparent hedge fund alpha may be explained by systematic beta exposures.
- Whether replication is feasible depends on whether the goal is matching beta or the full return stream.
Tags
Full text
# Cloning Return Streams # Cloning Return Streams How would one go about cloning/replicating returns of a hedge fund or a strategy. That is given a return series of the object to be clone, is it possible to decompose return and reconstruct another different passive strategy that generates return that are correlated above 75%+? Perfect replication is definitely not possible, but I wanted to understand the process and possibly the success people have had in this area. My initial thought on such subject would be to decompose its return using factor models... Some pictures (source: bwater): Edit: 2012-12-14 ## Answer by Vijay (score 5, accepted) https://quant.stackexchange.com/a/4744 You are treading controversial waters. It's hard to summarize, but at the risk of oversimplifying, there are three broad schools of thought: - "Linear Models": Classic Examples are a string of papers from Jasmina Hasanhodzic and Andy Lo at MIT (scholar.google.com should give you plenty). For similar work related to Mutual Funds that you may be able to repurpose you should look at the classic "Sharpe Returns Based Style Analysis (aka RBSA)" upon which most linear approaches are based. - "Non Linear and 'Mystery' models': (i.e. details undisclosed) models. Classic example is from the infinitely entertaining Harry Kat et al (examples include "Tell Me What You Want, What You Really, Really Want!" but if you google it, you'll need to add some sort of hedge fund replication tag to avoid getting nothing but Spice Girls references) - "It's not easy": Classic examples here are hard to find, but the most eloquent (and imho unbiased) are from Amenc et al at EDHEC. Examples include "Performance of Passive Hedge Fund Replication Strategies" The short version, sadly, is that the general feeling (amongst the majority of academics, at least imho) is that what you are asking about is not easy to truly "clone" :-( but check out the references above anyway, perhaps you'll spot something new and interesting. Having said that, many academics (and, apparently, some of your commenters) feel that most hedge fund "alpha" is really beta disguised as alpha. In that case, depending on what you mean by "replicating returns of a hedge fund" you may or may not have an plausible task on your hands. Replicating the "beta" portion of a hedge fund may indeed be possible. The classic reference here is probably any of the string of papers by Fung and Hsieh (again, scholar.google.com is your best friend). Bottom Line: it's a matter of opinion and if you had a more precisely stated question, you might get a more precise answer. I hope that helps at least a little :-)
Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.