As the old joke says, “math is what mathematicians do.” Somehow this simple tautology is lost in the dishonest world of finance

Quantitative investing: A crisis waiting to happen

In a recent WSJ article, Jason Zweig brilliantly summarizes the unbearable hype and hubris exhibited by some self-titled “quants”:

BlackRock, the giant asset manager, recently announced it will rely more heavily on computers to pick stocks. Rob Arnott, a leading advocate of mechanical investing approaches, said this past week that it’s “actually relatively easy to beat the market” if you get the math right.

Mr. Zweig is of course a great professional and a polite journalist. Some of us would have mentioned, in passing, that Rob Arnott’s funds have been divested by PIMCO after years of underperformance. So it seems beating the market is not “actually relatively easy” after all. Many other “factor investing” and “smart beta” zero-humility funds will follow.

Some call it “factor investing”, others call it “smart beta”. Whatever this “dumb alpha” is, it does not perform like real quant investments

What is a quant?

Like any buzzword, the term quant has been abused to mean everything. Historically, a quant has been a PhD in a STEM fields who applies her or his scientific knowledge to finance. Perhaps the best known example is Emanuel Derman, a PhD in physics who in the early 2000s wrote a bestseller titled “My Life as a Quant.” Some of the most successful funds in history are led by quants, with no background in academic finance: RenTec, Two Sigma, DE Shaw, CFM, etc. Moreover, these quant funds do not hire candidates with a PhD in finance. Clearly, the most successful quant funds do not consider academic finance as part of the quant world. Let us understand why.

RenTec was founded by the eminent mathematician James H. Simons, who had no prior financial background. The firm has managed to produce average returns of 35% over more than two decades, without ever employing a PhD in finance as quant researcher

Finance is not a science

Given the success of these firms, and the term quant, soon financial academics saw profitable to call themselves “quants” too. They must have thought: “Forget about real math, financial models use numbers and econometrics, right?” In the uncertain world of finance, it is good marketing to attach a scientific aura to a financial product. Here is a little problem: Finance is not a science according to accepted epistemological definitions. In order to be considered a science, finance must overcome three hurdles:

  1. It must discover immutable positive laws.
  2. Experiments must be independently reproducible.
  3. Predictions must be accurate.

Problem #1: Nothing works forever in finance

Finance is the result of changing human institutions, agents, laws. Unlike physical phenomena, financial markets are an adaptive system. There are no positive laws in finance. Competition means that every edge is doomed to disappear. At best academics may find something that used to work, as the opportunity will be arbitraged away following its publication.

Problem #2: The scientific method cannot be applied to finance

The only financial laboratory is the market, and that is too busy solving real-life problems. Academics cannot repeat experiments, hence their discoveries cannot be independently validated. We cannot go back to 6 May 2010 and repeat the events of the flash crash by removing some actors, in order to derive a precise cause-effect mechanism. Instead, academics pretend that historical simulations (backtests) can replace true experiments. Even if that were true, virtually no study controls for multiple testing, a practice the American Statistical Association considers highly misleading. As the current President of the American Finance Association has acknowledged, this means that the great majority of the discoveries published in journals are likely false, due to selection bias. The reason these pseudo-scientific financial theories are taught in classrooms is because they are unfalsifiable in a Popperian sense.

The scientific method requires experimentation, not historical simulation. Finance is an art, and there is nothing wrong with that

Problem #3: All evidence indicates that academic finance has failed miserably

Where are the billionaire financial academics? There are no billionaire Nobel Prize winners in Economics. If you could make accurate financial predictions, would you not act on them, if for no other reason to prove a point? No academic financial theory has generated substantial investment returns for their authors. Of course, some academics have enriched themselves by charging outrageous management fees to their investors, but that has not made their investors wealthier. Useless predictions means finance is not a science in a Lakatosian sense.

Left: The original Nobel medal, given for Physics, Chemistry and Medicine. Right: The medal funded by the Bank of Sweden since 1968, against the will of the Nobel family. Yes, they look very similar, like quants vs. pseudo-quants, but do not be fooled by appearances: What would you call a “science” that cannot predict anything?

The gullibility crisis

Many investors have been misled to believe that financial products originated in academic finance are scientific. Pension allocators have poured hundreds of billions of dollars in so-called “factor investments” and “smart beta funds”, not because they perform well, but because they have a good academic pedigree. Remember, these asset allocators are not rewarded for success, and their decision is informed primarily by risk management. That is, managing their risk of being fired! When fund ABC fails, they will point out that ABC invested in an idea by Nobel Prize XYZ, and that they are not going to second-guess science. The term con-art seems more appropriate than science, as these funds abuse people’s trust in science.

Many of these products, known as “factor investments” or “smart beta funds”, take the form of ETFs that offer immediate liquidity. That may have been alright when the industry was $100 billion, but assets keep growing steadily approaching $1 trillion, propelled by the “Fed’s put”. Because they rely on the same principles, these funds are highly correlated. Sooner or later the Fed will cease to suppress volatility, and let market forces dictate prices. When customers realize that factor investments underperform the market (especially given the hefty fees!), they will pull their money. One possibility is that instant liquidity may enable scenarios like the flash crash.

A more likely scenario is that investors will steadily lose money… bled by fees and arbitrageurs. Here is how:

  1. Companies will adjust their accounting to become attractive to these publicly-known factor models.
  2. A small number of securities favored by the same academic models will receive large inflows.
  3. Prices will rise beyond their fair value, not because these theories are right but because the academic marketing machine keeps inflating their prices in a predictable manner.
  4. Quant funds will prey on pseudo-quant funds.

And that’s a key flaw of academic financial models, that they are public: Humans will always find a way to turn a purported financial “law” on its head, and profit from those gullible enough to believe in it.

Comments are closed.