One of our primary objectives at AllAboutAlpha.com is to encourage debate and discussion on emerging topics in investment management. That is why we cover new academic studies, surveys, counter-intuitive viewpoints and controversial opinions. Today, we bring you the latest in an ongoing debate between two well-known and highly regarded figures in the hedge fund industry, Professor Harry Kat of the Cass Business School and Dr. Lars Jaeger of fund manager Partners Group (previous postings: Jaeger…Kat…Jaeger…). Although Kat and Jaeger differ on many issues, they share a common interest in furthering the field of finance through frank, collegial and mutually-respectful debate. And judging from our traffic, so do you the reader.
Today, Kat responds to Jaeger’s rebuttal…
Special to AllAboutAlpha.com by: Professor Harry M. Kat, Cass Business School, London
Before I respond to Lars Jaeger’s comments in more detail, it is probably good to backtrack a bit. In my note of February 20, 2008, I made the following 3 points:
First, if you want to replicate a diversified hedge fund index, you don’t need alternative betas since such an index is almost fully driven by traditional risk factors.
Second, the (traditional) factor exposures of diversified hedge fund indices do not seem to change quickly enough over time to completely invalidate the factor model approach. The performance (backtested or live) of the various factor model based replication products supports this. I showed the evolution of the Goldman Sachs ART index because the Bloomberg data go back until 1996, but I could well have picked another comparable product.
Third, thinking about investment risks in terms of risk factors is very useful, but implementing this framework in practice is not as easy as it may seem. I used the live performance of the Partners Group ABS fund to supports this hypothesis because it is the only fund of this nature with data readily available.
In his comment, Lars claims I have treated him unfairly as I have shown the ABS fund performance net of fees. This is not true! I applied a flat fee of 100bps, comparable to what factor model based replication products tend to charge, despite the fact that the ABS fund itself charges a much higher fee. The graph below shows the ABS performance gross of fees (Bloomberg PGABS), net of 100bps and net of the actual fees charged. From the graph it is clear that the figures I have used are much closer to the fund’s gross results than they are to the fund’s actual net results. The graph also makes it clear what an enormous drag the 1.25% plus 15% fee structure is on the performance of the fund. Over its three and a half year life, the fund has turned an initial investment of $100 into $111.53. If it had only charged 100bps, however, it would now stand at $120.52.
Lars says I should have included the ABS fund’s backtested returns as well and that things would have looked better if I had done so. That is true. However, I chose not to do so for two reasons. First, comparing backtested and live performance, the ABS fund’s performance seems to have gone sour the moment it went live. That could be an unfortunate coincidence of course, but it could equally well be the result of data mining. Second, the fund strategy appears to have changed over time. In February 2007, Lars told a conference in London that the ABS fund combined 18 different strategies. In New York in October, the number of strategies had grown to 38. Obviously, this makes the earlier backtest results less relevant.
But let’s skip the nitpicking and move on to the one billion dollar question: how useful is modern finance theory really? Lars says: I cannot see how rigorous economic arguments can be taken ‘just one step too far'”. This statement reflects the devout belief in the power of economic reasoning one would expect from a novice. One will not often hear experienced economists make such claims. They know and understand the limitations of what Lars refers to as rigorous economic arguments and sometimes, when amongst themselves, jokingly refer to economics as the art of telling plausible stories.
Looking at what rigorous arguments in finance have brought us so far, one cannot help but feel utterly disappointed. CAPM and APT don’t fit empirical data, mean-variance portfolio theory is largely unworkable as expected returns are impossible to forecast, the efficient markets theory is clouding the judgment of millions, etc. etc. The only good thing to come out of modern finance so far has been option pricing theory, although one could argue that since its foundations were laid as early as 1900, this success should not be attributed to modern finance either.
Given the above, I think Lars grossly overestimates what rigorous economic arguments can bring us. This leads him to overcomplicate matters to such an extent that his creations start to work against him. This is supported by the facts. The ABS fund has underperformed significantly, as Lars admits himself. He states that “we ultimately rely on the actual performance of our models to judge their quality”. Well, if that is indeed the case, he will have to agree with me that there is something not right with his model.
What is the methodological alternative to Lars’ approach? Well, it’s exactly the imperfect statistical analysis that he dislikes so much. Instead of fantasizing about how the world would be if everybody behaved fully rational all of the time and studying abstract theoretical worlds where everything that makes life complicated is assumed away, we should pay more attention to the world around us and let the data speak out more often. This is the approach underlying FundCreator, our risk management system, and it has paid off tremendous dividends so far. Yes, FundCreator is complicated, but it is not overly complicated. All complication in FundCreator serves a purpose. In short, models should be as complicated as necessary, but not more than that.
– Harry M. Kat, March 3, 2008
The opinions expressed in this guest posting are those of the author and not necessarily those of AllAboutAlpha.com.