After yesterday’s story on EDHEC’s new hedge fund replication research, we were curious about Harry Kat’s take. With some cajoling by us, Professor Kat responds below
By: Prof. Harry Kat, Cass Business School, City University (London)
In a recent interview published at All About Alpha and a research paper presented at the Edhec Asset Management Days in Geneva on March 13, Edhec researchers make some statements with respect to the workings of and results that can be obtained from our FundCreator technology that could be misconstrued. In this note I hope to clarify a few important issues.
But first, I’d like to start with a short story that will clearly illustrate the point that I’d like to make.
Leo and Chris are walking down the High Street and suddenly notice this beautiful shiny new car. They stop to take a closer look. What is it? says Leo.
I don’t know, says Chris. never seen anything like it.
Hey, wait a minute, Leo shouts, It says ‘Rolls Royce’ over here.
That must be it then. says Chris. It’s the latest Roller. I think I read something about it in a car magazine in the supermarket last week. After taking another good look, both wander off fantasizing about what it would be like to have a car like that themselves.
Closer to home, Leo suggests they build a Rolls Royce for themselves. Chris agrees and offers his tools and his garage as a workshop. The two go to work and by the evening their car is finished. Indeed, it looks very much like the real thing.
They both jump in and hit the road. Driving around the neighbourhood, however, it quickly becomes evident that the car they built looks a lot better than it drives. It’s bumpy, takes 20 seconds from 0 to 60, doesn’t steer properly and the brakes squeak. Of course, that’s not really surprising, as they only got to see the outside of the Rolls they were trying to replicate. And Rolls Royce has been refining the inside of the car for nearly 100 years. Disappointed, Leo and Chris return home to evaluate their efforts.
What are Leo and Chris to conclude? Will they conclude from today’s experience that Rolls Royce doesn’t make good cars? No. Ultimately, Leo and Chris draw the only rational conclusion: that their attempt to build their own Rolls Royce was unsuccessful.
Now turning our attention back to FundCreator and the Edhec research, one could tell a very similar story. FundCreator’s roots go back more than 7 years. Over that period, more than 10 man-years have been invested in developing the various procedures and the system itself. Packed with state-of-the-art estimation, optimisation and stochastic control routines, the FundCreator C++ code currently spans over 17,000 lines. Nobody can realistically expect to recreate this in the comfort of their own garage. Indeed, with the help of our November 2005 research paper one could build something that looks like FundCreator on the outside, but since it is difficult to surmise what really goes on inside the machine, one should not expect the same results. In other words, and in my view, EDHEC researchers simply did not accurately replicate FundCreator.
Apart from not accurately replicating FundCreator, the Edhec researchers may have misinterpreted some of the published results. In the aforementioned interview, for example, EDHEC says, Using the S&P 500 index as the reserve asset (as Kat does), our replications under-perform systematically and significantly, by approximately 500 bps per year.
However, in FundCreator, the reserve asset is not simply the S&P 500 index. Since the reserve asset is at the core of every FundCreator strategy, it should contain as little uncompensated risk as possible. We therefore advise everybody using the system to use a well-diversified basket as the reserve asset, not just one single index. In our own research we use a portfolio of Eurodollar, 5-year and 10-year note, Russell 2000, S&P 500 and GSCI or crude oil futures. Apart from leaving a lot of diversifiable risk, using only the S&P 500 as the reserve asset will pull in the full 2000-2003 equity bear market. It is therefore not at all surprising that the Edhec study finds the strategy underperforms significantly – it is significantly under-diversified.
So how can one evaluate something of which one doesn’t know what is going on inside? I think that’s simple – try it out and examine the results themselves rather than the methodology. We have never held back information about the results achievable with FundCreator. On the contrary, we have published a whole variety of back tests, on the FundCreator website (and elsewhere) going back to 1995. In my view, these back tests show one thing very clearly: FundCreator is efficient and remarkably robust with respect to transaction costs, the rebalancing frequency, the choice of reserve asset, etc.
The proof of the pudding is in the eating Don Quixote once said, and he was 100% right.
– Harry M. Kat March 14, 2007
Professor Harry Kat is an occasional contributor to AllAboutAlpha.com. The opinions expressed herein are those of Professor Kat and not necessarily those of Alpha Male or AllAboutAlpha.com.