Case for Short-Term Quants
Investors have become more interested in short-term trading strategies. Rishi Narang, founder of Telesis Capital LLC and author of Inside The Black Box—The Simple Truth About Quantitative Trading, explains the advantages of short-term approaches.
Mr. Narang manages a fund of managed accounts specializing in trading strategies. He has been involved in the hedge fund industry, with a focus on quantitative trading strategies, since 1996. He was managing director and co-portfolio manager at Santa Barbara Alpha Strategies from 2002 to 2005 and previous to that was president and vice chairman of Tradeworx Inc., a quantitative hedge fund manager.
Opalesque Futures Intelligence: What’s your investment approach?
Rishi Narang: I focus on quantitative strategies in equities, futures and options. Our core approach has four components. One is that we want pure alpha—I stay away from beta exposure. Even in my personal portfolio, I don’t own stocks, bonds or credit. Our second requirement is liquid instruments and the third is mostly quantitative strategies. Non-systematic, discretionary strategies are a very small part of the portfolio. Our fourth goal is to be mostly in short-term trading. I generally do not invest in long- or medium-term trend-following commodity trading advisors.
OFI: How do you define short term?
RN: In our portfolio, the longest term quant trader holds positions on average for 10 days. The shortest term quant turns over positions many times a day. Overall, the average time for turning over positions is slightly less than two days.
OFI: Do you prefer short-term trading because it is more liquid?
RN: No, that’s not the reason. Long-term trend followers also use liquid instruments.
OFI: Why do you like short-term strategies?
RN: Here’s an interesting analogue. Take weather forecasts. You can predict the weather really well over the next few hours. The further out the date, the less accurate the prediction. You can’t really predict the weather four days from now. Weather systems at this point are not that well understood and the data is so noisy that building a model for long-term forecasting is extremely difficult. Similarly it is extremely difficult to build long-term trading models. Models look into the past to predict the future. The greater the distance from the starting point in the past to the prediction in the future, the greater the odds that the future will change and cause the strategy to fail. A shorter span gives you better odds.
OFI: What’s the catch?
RN: Transaction cost is the key problem for short- term strategies. If you can solve the problem of transaction costs, then the risk-adjusted expected return from short-term trading is better than the expected risk-adjusted return from long-term strategies.
OFI: Can short-term trading accommodate large amounts of capital?
RN: To invest like we do, you can’t manage much money. We’re in a small corner of the quant niche. We do not aspire to become a big, multi-billion dollar asset manager. Some managers can take more money—I know one very good short-term futures trader that now manages $4.5 billion. We could increase our capacity by skewing our weighting to traders who can take more money. But our goal is to make the best investments, rather than maximize capacity. Almost all of my own money is in the fund, so I mainly care about making the fund as good as possible. By contrast, most people in asset management look to maximize capacity.
OFI: Is it possible to overcome the capacity constraint?
RN: Anyone who thinks they’re managing tens of billions of dollars in short-term strategies is either managing Renaissance Technologies’ Medallion fund or is likely deluded. It is really hard to manage large money with short-term strategies. The problem of market impact is simply too severe and it is very difficult to find people who have solved that problem. Big shops may have short-term trading elements, but on the whole their average holding period is measured in weeks or months, not hours or days.
OFI: Haven’t quantitative strategies performed poorly in recent years?
RN: There have been some major headwinds for quantitative trading, but more so for long-term strategies. There were lots of traps for fundamentally oriented longer-term trading. Last year stocks that looked cheap kept getting cheaper; this year stocks that looked like junk kept getting more expensive. Shorter term quant strategies were less affected
OFI: What’s tripping up the models?
RN: Quantitative models are based on fairly stable behavioral patterns from the past, so dislocations that break with the past can be painful. We’ve had a number of such dislocations in the past several years.
OFI: Can’t quantitative methods account for such developments?
RN: Sharp breaks from the past – like 9/11, the Iraq war or Lehman Brothers going under – are almost impossible to factor into models. A model based on rare events will be wrong most of the time—not a good way to play any game.
OFI: What about the quant losses in 2007?
RN: The massive sell-off by quants in August 2007 was caused by the beginnings of the credit crisis. Big multi-strategy funds that had credit and quant trading got stuck with too much risk in their credit allocation and had to liquidate something. They reduced their quant allocations, since that was easier than selling credit. Asset prices buckled under the tidal wave of liquidations.
OFI: Did these events affect long- and short-term traders differently?
RN: In the summer of 2007, longer term CTAs had big drawdowns, but shorter-term traders were more successful. Then, 2008 was great for trend followers while short-term traders had mixed results and on average did only half as well as longer term traders. By contrast, this year short-term quant strategies are doing better. Considering this three-year period, it has been less challenging for shorter term strategies. In general quant strategies have done very well, despite the common perception that they have not.
OFI: Why is this year difficult for technical trading?
RN: Regime changes are painful for technical trading. There were two kinds of regime change this year; a transition from bear market to bull market and a steep decline in volatility.
OFI: What do you seek in a quant?
RN: I look for a number of characteristics, but a very important one is humility. Yes, they need to be confident enough to believe that they can make money. But at same time they need to have respect for the market and recognize that taking risk is a very serious matter. Some quants believe they’re so smart that they’re smarter than the market. But even if they’re right in the long run, that does not mean that they’ll actually win. Long-Term Capital Management’s trades were right in the long run, and yet the firm did not survive.
OFI: What else defines the ideal quant?
RN: The second hallmark of a good quant is a lucid understanding of what they’re good at and what they’re not good at, coupled with the ability to take exposures in the former and avoid the latter. Say someone thinks they have an edge in predicting convergence between stock prices. If they don’t control for sector risk, they could take unintentional bets on sectors as they buy under-performing equities and sell out-performers. The good quant would avoid sector exposure—where there’s no edge. A third hallmark is mild pessimism. Most models that predict the future don’t work. It is a bad flaw to expect a given idea to work and look to support it. A quant should take the approach that his job is to falsify his theory and only when he can’t falsify the theory will he implement it. You can think of this as Type I errors in science – which in finance causes blow-ups – versus Type II errors – which mean not making as much profit as you could. Better quants make Type II errors more than Type I errors.
OFI: What’s the future for quant strategies?
RN: Quantitative trading is a blend of human intelligence with computer capability. It combines the strengths of the human with those of the computer. How this hybridization will develop further is a very interesting issue. Technology is becoming faster at a faster pace. How does a trader make use of such advances? The object should be not only to do the same thing as other people only faster, but also to do it better and furthermore to do new things altogether. In the meantime, there is more interest from investors in short-term quant trading, because that’s what’s done well over the past few years. This aspect of human behavior I don’t expect to change at all: investors tend to chase whatever has done well recently.