Thursday, August 28, 2008

Evaluating short extension funds (130/30)

The current issue of the FAJ has a great piece about How variation in signal quality affects performance. On the face of it, the article deals with the generic perception that any relaxation of investment restrictions (read: long only) will invariably result in better performance. This is the selling proposition of currently fashionable 130/30 funds. But it doesn't - or only in the special case of an information coefficient which is stable over time. Managers with unstable ICs will see their performance deteriorate with increasing short positions. 

This approach delineates a way to help evaluating 130/30 funds which do not have a long enough track record in the product (i.e. the majority of vendors), but a more significant one in long only mandates. If that track record were to exhibit an unstable IC in long only portfolios already, then there might be a problem in the short extension.

Tuesday, August 26, 2008

Prediction markets

Let me be explicit from the start: I have a problem with the judgmental connotations of the term speculation. In my view, there is no meaningful distinction between investing and speculating. Investing as well as speculating is about the assumption of risk in return for the uncertain possibility of reward. There is only inappropriate investment or speculation, but not bad speculation per se. 

The cause for these rather philosophical considerations is my interest for prediction markets such as intrade, where you can effectively trade in all sorts of event probabilities. Literature has it that predictions produced using a bid/ask mechanism with real money will be of significantly better quality than survey forecasts for instance. Naturally, they cannot tell the future, but they will be more efficient at processing all currently available information than any one expert. Therefore, these markets serve two important purposes: If reasonably liquid, they offer high quality, quantified predictions of event probabilities of presidential elections or hurricane landfall severities, and they offer a hedging possibility, although the liquidity is definitely a limiting factor there at this point.

It is very timely that the CFA Institute has just published the results of its last Monthly Question survey about prediction markets. 65% of respondents think that such markets offer valuable information, and a majority of 54% think that they are fit for investment purposes (mostly hedging).

Tuesday, August 12, 2008

Challenges in Quantitative Management

At last, I got round to finishing Challenges in Quantitative Equity Management, a freely available CFA Institute Research Foundation monograph by Fabozzi, Focardi and Jonas. The book is based on surveys and conversations with practitioners of quant management, capturing the collapse in industry outperformance following the summer of 2007. I was particularly interested to find out whether there would be any attention paid to XBRL, which I suspect should be quite relevant to quant management, but it wasn't even mentioned throughout the book. That's not to say it is considered to be irrelevant: Too many people using similar models and the same data was considered to be one of the major challenges to the quant approach going forward. 

Nevertheless, it is a worth while read for a number of thoughts and discussions, such as the distinction between quantitative and judgmental investment processes, the style correlation of quant management (which tend to be value-driven), the negative correlation between fund capacity and alpha-generation and, last but not least, the adaptive markets hypothesis as opposed to the efficient markets hypothesis.