After a bit of shadenfruede shared with a colleague and occasional quant manager earlier this week, m* finally got around to reading the investor letters of several quant funds, as neatly collected and summarized in numerous articles of late. Collectively, they sound like no one so much as the homebuilder ceo’s, blaming other companies for building too many homes and saying things will improve just as soon as the other guy leaves the business. m* thinks these guys are missing a factor or two.
Their strategy, also called statistical arbitrage, exploits various market inefficiencies well known in academic circles. Returns to small size, value, the January effect, momentum, and mean reversion form the theoretical basis for these strategies. However it was only with the onset of electronic trading systems and large commercial databases of regressions of stock returns on various risk factors that these strategies could be implemented on an efficient scale. The low cost of electronic trading made possible vast numbers of transactions, most earning mere pennies or fractions of pennies. The risk factor regressions made it possible to create long-short portfolios positioned to exploit the inefficiencies while remaining statistically neutral to a myriad of macroeconomic, industry, and company characteristic factors. The product has expanded into global equity markets and even into markets for currencies and interest rates.
Pre-2002 let’s say, before the serious explosion in hedge fund assets, and the no-volatility-four-year-without-so-much-as-one-10%-pullback-bull-market (have you been feeling very smart in your investing until recently?), quant funds were on the whole rather small, an under the radar presence, and far from being dominant participants in equity markets relative to large mutual funds and institutional investors. However they were very successful at generating consistent returns and this, along with the intrigue stemming from the intense secrecy that shrouded their intellectual property, made them must-have investments for asset allocators.
Of course success breeds imitators and with the basic principles in the public domain, there was little to stop two guys in a garage from programming up their own systems, proving it in backtesting, and raising investor capital, often from an existing hedge fund eager to expand into the space. Moreover, as the original operations matured, more and more employees with the requisite knowledge left to start their own firms which also found ready investors. The amount of capital employed in quant strategies mushroomed from a few billion to likely in excess of $100bn in equity capital today, levered 3x to 6x, and with their high frequency transactions magnifying their influence, a daily volume estimated to be more than half the overall equity market.
The natural question to ask of course is how much alpha is available to these strategies? At what point do they go from exploiting the over trading noise of other traders in the market to being the market themselves?
That is exactly what happened. When you have the CFO of Goldman Sachs talking about multiple 25 standard-deviation events and hedgefund managers publicly dissecting their daily performance data like FAA crash investigators while saying things like “a perfect storm of negative events” m* is pretty certain they lack (publicly anyway) an understanding of the true distributions or their true risk, not to mention any semblance of personal accountability or embarrassment.
They seem to have mistaken the collective reinforcing effect of their own behavior on the trend for continued success of the strategy. Over time, as capital flowed in the funds poured more money into the same long and short value strategies while actively dampening volatility with the mean reversion strategies, in effect self-validating their models from the past even while the real situation had changed.
In short this most sophisticated group of investors became just another group of levered momentum traders, creating their own temporary low volatility trend. Their alpha capture became beta follow, and they were none the wiser until the infinitely creative market disrupted their somnolence with levered losses that appeared in a flash accompanied by the insight “Oh, we’re trend followers and we’re short vol!”
Except of course that not one of the letters from these managers mentions that insight, nor any explanation that demonstrates an understanding and acceptance of the true nature of the fund's risk. m* finds this amusing because in all the writhing gymnastics blaming the other guy, the market, the universe, or those darn frustrating statistics, (anything to avoid taking responsibility), a simple declarative “Yes, well, the strategy is inherently short volatility, a bit of a trend formed, and these event do occur from time to time,” would be the most exculpatory and truthful explanation.
What now for these funds and the prognosis for the strategy? Beyond this week’s easily foreseeable rebound (as they lever up again and GS puts $3bn more to work on the small cap buy list) a double whammy of diminished single digit returns and higher returns volatility is in store. How so? Lower leverage. The true alpha here is vanishingly small hence the intense focus on transaction cost and the need for considerable leverage to earn returns. Higher exogenous volatility in the marketplace, from the turmoil in credit for example, necessitates running lower leverage. In addition, as the mean reversion strategies were essentially providing liquidity to the market, acting as market makers in effect, a reduction of leverage there will lead to still higher volatility. Finally, the recent sharp losses are the clearest and loudest signal possible to investors that the strategies may not work as expected from here on out. Expect inflows and self reinforcing trends to diminish and volatility of returns to increase accordingly.
No comments:
Post a Comment