For all their troubles going about alpha generation, algorithmic traders get remarkable little respect these days. Last week while I was attending The Trading Show in New York, a select group of quant-oriented, buy-side participants and fintech enablers debated the limits of traditional computers in financial research and, conversely, of quantum computing’s promise. Many of these investors chase new paradigms for investing. For this, they look for high-quality unstructured data, they invest in quants from various disciplines, they give machines a greater role spotting and acting upon trading opportunities, and they manage a variety of risks. Some of these fund managers—echoing Nassim Taleb’s arguments on the study of Black Swan events—complain that models based on linear algebra and calculus are too simplistic and look instead for greater predictive accuracy in stochastic math and machine learning.
What’s driving the pursuit of ever-more esoteric predictive models? A very, very complex investing world, a mess of financial interconnections we call capital markets walking erratically along a “single path”—borrowing from the statistical expression that looks at causality sifting through multiple factors. Think of the events leading up to the Flash Crash of May 6, 2010, or the post-Brexit market reaction. What caused those events? Could those factors have been predicted and modeled independently, much like scientific experiments in biology or physics? The answer is “no,” and while some of these events may repeat, they do so with a different intensity or outcome because markets adapt to new data. Models feed off data. But complicating matters is that, as new data emerges, it alters predictive models in ways that have yet to be experienced, such as how will financial checker pieces rearrange should Trump become the new U.S. president?
Additionally, G7 regulators haven’t made the buy-side’s investing decision-making process much easier. Increasingly commonplace central bank interventions cause price distortions in fixed income and securities markets. The notion of risk premium for discounting future cash flows to value securities sadly doesn’t work well in a deflationary, negative interest rate environment. Greed and fear now coexist—it used to be that these antagonistic emotions would cause equities and fixed income markets to move in inversely correlated patterns. But now those fearing a broad market correction drive fixed income prices higher (fixed income yields move opposite into negative territory), while a large part of those not in the fear camp reluctantly flock to low-cost passive vehicles (ETFs, index mutual funds) and cause equities to rise despite the high risk of a strong market correction.
As further proof that this is not a great time to be a fund manager, investors have fund management styles under close scrutiny. Active fund managers are under increasing pressure to deliver yield to justify the standard 2%/20% fee structure that many of them charge. The speed with which these firms trade has become a liability in the eyes of Main Street investors. Lost to HFT critics is that being a fast market participant today is a necessity and also a cost of doing business. Having strategies based on speed hasn’t been a priority for the e-buy-side for three to five years, according to several people we spoke to with knowledge of the matter. With larger HFT firms increasingly playing a role as market-makers and some (like Virtu) acting as trading infrastructure partners to banks and brokers for select markets, it would seem that sophisticated buy-side firms are using scale rather than pure speed as the new way of making money.