The MAFFIA group (Mathematicians Against Fraudulent Financial and Investment Advice) offers a convincing alternative view in an insightful blog post, pointing out the difference between pseudo-quants and actual quants. Specifically, there is an important distinction between academic finance--and theories popular within academic finance--and actual mathematical finance. The gap between the returns of such math-based firms as Renaissance Technologies and Two Sigma and strategies based on academic finance theories reflects the differences in approaches to investing.
Because I am intimately involved in the recruitment processes of trading firms, I see first hand the rise in pseudo-quant practitioners: those using math in casual ways and marketing their approaches as quantitative. An extreme example occurred in a job interview with a junior candidate who asserted his quant background and skill. I mentioned to him my development of an ensemble model for the ES futures and asked him how he deals with large data sets to avoid overfitting. The candidate looked distinctly uncomfortable and said that he had not developed any models. Instead, he said, he plots market price changes on a graph and looks for patterns. Needless to say, our conversation about quant came to a crashing halt!
Less egregious but still highly problematic was the trader who came to the interview with a regression model developed over the past few years of market data. The model had a very high fit with the data, relying on a variety of rate of change measures. He confidently asserted that his model was valid because he had tested it "out of sample". Unfortunately, research suggests, if the search space is sufficiently large, it is not difficult to find a strategy that "works" in and out of sample merely by chance. What looks like "smart beta" is all too easily mined with large data sets, resulting in inferior forward performance. Such "quant" approaches can easily crash if they are implemented by the trading and investing herd.
True quant is the application of mathematics to the world of finance. For those interested in mathematical finance, a wide-ranging collection of papers can be found on the MAFFIA site. You'll see there insights into everything from the Sharpe Ratio to fresh strategies for hedging risks and what to look for in legitimate backtests. The answer to the limitations of pseudo-quant strategies is not to abandon mathematics altogether, but rather to employ rigor in the application of mathematics. Just as medicine has evolved from a discipline dominated by village doctors to more of an evidence-based science, finance is doing the same.
Resources:
Posts
