Jackson Hole Economics

When Following Rules Leads to Chaos

In recent decades, a rising fraction of financial markets trading has been guided by algorithms. An algorithm is simply a mathematical rule that determines when a computer, not a human, should buy or sell a financial asset.

Some rules are simple. Passive investing via exchange traded funds (ETFs) is one example. The investor simply purchases shares in the ETF, which is then automatically rebalanced each day in order to track its underlying benchmark, such as the S&P500. No human intervention is required to make the corresponding buy and sell decisions, which are based on rules, not economic fundamentals.

Other rules may be more complicated and hidden from view. Various firms have invested billions of dollars to implement algorithms that purport to recognize signals in price patterns that will give them an edge—often measured in nanoseconds—about where to profitably buy and sell. Such active quantitative strategies (‘active quant’) increasingly rely on artificial intelligence and machine learning to update their trading algorithms, further removing them from a human interface.

It has been estimated that algorithmic trading, whether passive or active now comprises well over half of all trading on global equity markets. According to research at Deutsche Bank, on a typical day over 80% of daily transactions in equity futures markets and stocks are executed according to algorithms. Bond, currency and commodity markets are witnessing increasing mechanized trading as well.

Of itself, algorithmic trading is not ‘wrong’. These approaches have served investors well. Passive investing via ETFs enables investors to participate in markets at much lower cost (lower fees). Some active quant strategies have delivered above average returns.


But there is a darker side to investing without reference to the fundamentals that determine asset values, such as growth, earnings, inflation or interest rates. And that darker side extends beyond the scope for unhinged prices—i.e., asset prices bubbles—or the potential for capital misallocation. 

The more pernicious risk mechanical investing interferes with the very mind-matter dynamics that are essential for a functioning, healthy and symbiotic relationship between financial markets and the real economy. This is a central tenant of the application of cognitive science to economics and finance, as I have pointed out in these pages before.

Cognitive science suggests that the economy and markets embed a form of dualism. Human decision-making in both markets and economic activity creates invaluable information, above all price discovery, that makes for better individual and collective decision-making. Collective market psychology (‘the market mind’) shapes and reshapes outcomes in the real economy and vice versa. In a proper functioning market comprised of atomistic individuals, intrinsically valuable information is relayed to and from Wall Street and Main Street, which, while not ensuring instantaneous nor continuous equilibrium, nevertheless is the sole sustainable process for delivering, on average, the least bad outcomes for each.

Introducing an increasingly dominant form of investing, derived either from passive re-balancing or to exploit price patterns (active quant), places that symbiosis at risk. Neither form of algorithmic trading reflects changes in underlying fundamentals, rather only in market data itself. In extremis, the natural equilibrium between ‘Mr. Market’ and ‘Mr. Smith’ (aka between Wall Street and Main Street) breaks down.

The reason is simple. When algorithms decide, price discovery is suppressed. No less an authority than Jack Bogle—the father of passive investing—noted that the result of pure index trading would be chaos and catastrophe.

The critical point is as follows. Any approach, whether derived from economic, statistical or financial models, that considers interaction between conscious humans as mechanical is bound to fail and, in doing so, may cause damage beyond the realm of its inventor. Human interaction in all markets, financial and real, reflects fundamental factors as well as human perceptions, including awareness, of those factors. Mr. Market’s mind, warts and all, mirrors the wisdom and sometimes also the foolishness of crowds. 

Put differently, in econometric terms, any mechanistic model that ignores individual and crowd mentality, and the feedbacks between them, will be mis-specified. In other words, any rule based on mechanistic approaches inherently ignores relevant variables, including market mood. That means the estimated coefficients of such mechanical models will never be stable. 

In finance terms, the application of the mistaken notion that prices are derived mechanistically means that unstable asset price patterns will defy attempts to predict how they may change. 

But it is the application of algorithmic approaches in finance, economics and public policy, treating all aspects of the economy and markets mechanically, that has the potential to do the greatest damage because it ignores or suppresses the basis on which decision-making occurs.

What is a better way forward? To begin, we must recognize the consequences of what is going on around us. Passive investing is a classic fallacy of composition. What is good for the individual investor (cheap access to capital markets) makes everyone worse off if it is universally adopted, as Jack Bogle reminded us. 

Rather, more research and effort must be devoted to understanding the interplay between economics, finance, and cognitive science. If we are to devise tools to enable investors to more sustainably and fruitfully turn their savings into investment via capital markets, then those tools must incorporate the principles from each of these disciplines.

Human interaction resides at the core of value creation. But humans acting alone and among each other are not automatons. Models, statistics and algorithms help us to detect patterns in human decision-making, but they are not a substitute for it. Worse still, if we assume away consciousness, individually and collectively, we risk destroying the very foundation upon which value creation resides.