In the world of modern finance, there has always been the search for the Holy Grail. Ever since the advent of computers, practitioners have looked to harness the power of computing and direct it towards the goal of producing endless profits. Today the buzz words being used across industries include, “AI – Artificial Intelligence,” “Machine Learning,” “Neural Networks,” and “Deep Learning.” Regrettably, nobody has found a silver bullet, but that hasn’t slowed down people from trying. Wall Street has an innate desire to try to turn the ultra-complex field of finance into a science, just as they do in the field of physics. Even banking stalwart JPMorgan Chase (JPM) and its renowned CEO/Chairman Jamie Dimon suffered billions in losses in the quest for infinite income, due in large part to their over-reliance on pseudo-science trading models.
Preceding JPM’s losses, James Montier of Grantham Mayo van Otterloo’s asset allocation team gave a keynote speech at a CFA Institute Annual Conference in Chicago, where he gave a prescient talk explaining why bad models were the root cause of the financial crisis. Montier noted these computer algorithms essentially underappreciate the number and severity of Black Swan events (low probability negative outcomes) and the models’ inability to accurately identify predictable surprises.
What are predictable surprises? Here’s what Montier had to say on the topic:
“Predictable surprises are really about situations where some people are aware of the problem. The problem gets worse over time and eventually explodes into crisis.”
When Dimon was made aware of the 2012 rogue trading activities, he strenuously denied the problem before reversing course and admitting to the dilemma. Unfortunately, many of these Wall Street firms and financial institutions use value-at-risk (VaR) models that are falsely based on the belief that past results will repeat themselves, and financial market returns are normally distributed. Those suppositions are not always true.
Another perfect example of a Black Swan created by a bad financial model is Long Term Capital Management (LTCM) – see also When Genius Failed. Robert Merton and Myron Scholes were world renowned Nobel Prize winners who single-handedly brought the global financial market to its knees in 1998 when LTCM lost $500 million in one day and required a $3.6 billion bailout from a consortium of banks. Their mathematical models worked for a while but did not fully account for trading environments with low liquidity (i.e., traders fleeing in panic) and outcomes that defied the historical correlations embedded in their computer algorithms. The “Flash Crash” of 2010, in which liquidity evaporated due to high-frequency traders temporarily jumping ship, is another illustration of computers wreaking havoc on the financial markets.
The problem with many of these models, even for the ones that work in the short-run, is that behavior and correlations are constantly changing. Therefore any strategy successfully reaping outsized profits in the near-term will eventually be discovered by other financial vultures and exploited away.
Another pundit with a firm hold on Wall Street financial models is David Leinweber, author of Nerds on Wall Street. As Leinweber points out, financial models become meaningless if the data is sliced and diced to form manipulated and nonsensical relationships. The data coming out can only be as good as the data going in – “garbage in, garbage out.”
In searching for the most absurd data possible to explain the returns of the S&P 500 index, Leinweiber discovered that butter production in Bangladesh was an excellent predictor of stock market returns, explaining 75% of the variation of historical returns. By tossing in U.S. cheese production and the total population of sheep in Bangladesh, Leinweber was able to mathematically “predict” past U.S. stock returns with 99% accuracy. To read more about other financial modeling absurdities, check out a previous Investing Caffeine article, Butter in Bangladesh.
Generally, investors want precision through math, but as famed investor Benjamin Graham noted more than 50 years ago, “Mathematics is ordinarily considered as producing precise, dependable results. But in the stock market, the more elaborate and obtuse the mathematics, the more uncertain and speculative the conclusions we draw therefrom. Whenever calculus is brought in, or higher algebra, you can take it as a warning signal that the operator is trying to substitute theory for experience.”
If these models are so bad, then why do so many people use them? Montier points to “intentional blindness,” the tendency to see what one expects to see, and “distorted incentives” (i.e., compensation structures rewarding improper or risky behavior).
Montier’s solution to dealing with these models is not to completely eradicate them, but rather recognize the numerous shortcomings of them and instead focus on the robustness of these models. Or in other words, be skeptical, know the limits of the models, and build portfolios to survive multiple different environments.
Investors seem to be discovering more financial Black Swans over the last few years in the form of events like the Lehman Brothers bankruptcy, Flash Crash, and Greek sovereign debt default. Rather than putting too much faith or dependence on bad financial models to identify or exploit Black Swan events, the over-reliance on these models may turn this rare breed of swans into a large bevy.
See Full Article on Montier: Failures of Modern Finance