“Too much money chasing too few ideas” has been the death knell for investment fad after fad. This will never cease to be so. – Howard Marks
With the recent announcement at Point72 of a $250MM investment in quantitative trading platform Quantopian and a recent FT article, there has been a surge in interest in Do-It-Yourself quant strategies. Here’s one from WSJ.
There are some significant challenges for these startups in my opinion*:
- Has already been done before – The most obvious criticism is that this has been done before with sites like Collective2. The website boasts $35MM in “linked” money, which couldn’t come close to supporting the initial investments in some of these platforms. While Collective acts as more of a marketplace model, the barrier to DIY was never very high to begin with. Interactive Brokers offers a free API and plenty of intraday or outside leverage sources for those willing to look.
- Is quant a skill, luck or talent? – These firms employ an “American Idol” approach of finding new managers in that everyone gets a chance to audition with the hope that a handful of people will find long term profitable trading strategies. This may be true in a practice like singing, but what about quant trading? Is it more like a skill based sport like golf? If you give everyone golf clubs, are they ready for the PGA? I would think it’s very likely that the worst person on the PGA could beat a random sample of the population a majority of the time. Professional baseball attempted something similar by hosting a reality tv show in India looking for pitchers. One would be forgiven if they don’t recognize the names of the winners, Rinku Singh and Dinesh Patel.
- Intellectual Property ownership – The ownership of the IP can be uncertain or at risk as it may be shared with potential investors and principals of the firm, if not publicly available. Any manager should be worried about a team of PhDs at Point72 actively trying to reverse engineer their strategy. It is not that difficult in most circumstances given enough experience in development and a few details about the strategy. In the past, some more unscrupulous firms would initiate fake job interviews with quant managers to get a snif of the IP. Big firms fight tooth and nail to protect their IP for good reason. Ultimately if a manager truly finds something proprietary** it would be in their best interest NOT to share.
- Individual strategies exhibit high elements of randomness – It has been my experience that the most successful approach is not via individual strategy but as a portfolio of strategies that are either uncorrelated or that implement a regime changing logic of some sort. There are really only 2 trades possible, mean reversion or trend following (or if you are a derivatives trader, long and short vol) The only thing that changes are the amplitude and frequency of the price series inherent in the regime. Strategies based on price action or transformation of price action (Kalman filtering, RSI, correlations) that exploit some artifact of the recent past price action are usually unstable, i.e. past performance not indicative of future returns. Publishing a portfolio of trading strategies may or may not be viable.
- Scalability – Finding alpha is hard, finding it at scale is even harder. If a manager found something that can generate a few 100k in profit in the personal account but can’t scale further, that likely wouldn’t interest most hedge fund. Scaling also typical requires greater infrastructure and focus on execution ability which would be difficult for a home gamer.
- Revenue model of firms – A lack of scalability puts into question the viability of the platform to generate meaningful earnings. The Jet.com model seems like the most attractive exit strategy because there isn’t enough juice in small scale strategies to spread among equity owners, VC, and managers.
Much of the interest in strategies outside the Wall Street domain is understandable. Likely due to frustration over performance of their human trading talent, Tudor recently announced a layoff of almost 15% of its workforce and brought on board a number of quants. Platforms such as these could serve as a new pipeline of future talent as current hiring tactics are notoriously inefficient and place an emphasis on past experience or pedigree rather than merit or potential. This can lead to rampant groupthink and herding behavior as opined by many people including Howard Marks and Steve Cohen.
However, I think some of the poor performance by humans is partially attributable to the management of the funds. There is a fundamental misalignment on what firms say they want (process, risk-adjusted returns) and what is valued (outsized and quick returns) Altering the hiring process won’t necessarily change this fundamental psychosis.
A lot of the blame however falls on the asset owners themselves whose expectations and investments horizon are misaligned. In fact I think ‘disruption’ in the current practices of asset owners represent a ripe and potentially lucrative opportunity.
*Disclaimer: I am an advisor and performing consulting work for a startup that proposes some solutions to the challenges above. Feel free to email me if you’d like to know more.
**Proprietary involves strategies based on more than price & volume data. It might best be served by some examples: Front-running rebalances of indices or new products, finding available short shares in hard to borrow situations, pushing FX markets in exotic pairs during illiquid market hours, momentum strategies on illiquid stocks, regulatory/operational loopholes. These are some anecdotal examples of strategies I’ve come across that worked for people in the past.