The word ‘panic’ has an interesting etymology. It derives from the Greek god Pan, who during times of crisis could unleash an unearthly cry that instilled terror into humans and gods alike, causing them to descend into utter chaos. You’d be hard pressed to find a more apt term to describe the carnage we’ve witnessed in the markets over the past couple of weeks. The rapid spread of the Covid-19 epidemic, better known as Coronavirus, has seen financial markets around the world enter freefall, with nearly $10 trillion in market value being wiped out in the US alone. While a market crash is an understandable reaction to the prospect of the global economy grinding to a halt, what’s more curious is the sudden rally in the markets on 24 March in response to a $2 trillion stimulus package announced by the American government, despite no visible signs of respite in the spread of the pandemic.
We ought to take this moment to ponder the very existence of the phenomenon of market panics. The one we’re going through may be unique in the scale and breadth of its causal factors but it is hardly unprecedented, historically speaking. Market panics have existed since the foundation of capital markets, and we’ve seen several happen in living memory, from the 1929 stock market crash to the 1997 Asian financial crisis to the 2008 financial crisis. The question really should be why? For years, the dominant paradigm in financial economics was that the main impediment to efficient markets is access to information. Yet even today, with a plethora of statistical models and computing tools at our disposal that can instantly access trillions of gigabytes of data from every corner of the world, we still experience events like bubbles, crashes and panic selloffs with breathtaking regularity. To understand why this happens, and indeed how we can solve this problem, we need to delve into one of the most important cornerstones of contemporary financial thought: The Efficient Markets Hypothesis, or EMH.
The myth of efficient markets
EMH essentially states that the price of an asset on the open market reflects all available information related to the asset, including probable future events that might occur. This theory has been around in various forms since the mid-19th century but was brought to the forefront of financial economics by Eugene Fama in the 1970s. The implication of this theory is that securities are always valued at a ‘fair price’ and it is consequently impossible to buy undervalued or overvalued assets as all available information would already have been priced in. It suggests that the only way an investor can hope to generate above market returns is to assume more risk, thereby precluding the possibility of any investment strategy outside of sheer luck in generating sustained alpha.
This, to put it mildly, has aroused considerable debate in the world of finance. Critics point out that this theory can’t explain the ability of celebrated investors, like Warren Buffet, managing to consistently outperform the markets over several decades. Supporters, on the other hand, shrug these off as statistical anomalies, and instead point to studies showing that actively managed funds rarely managed to outperform markets in the long run, with one showing that only 1 in 4 active managers actually managed to beat the market. The highlight of this debate was the famous Wall Street Journal Dartboard Contest in 1988, where professional investors attempted to beat a ‘dummy’ investor in generating higher returns from the market, the ‘dummy’ investor in question being the staff of the WSJ throwing darts at a dartboard to pick securities at random. Though the professional investors won, it was a pyrrhic victory; they only managed to beat the market 51 times out of 100 and only beat the dummy investors 3 times out of 5. The implication that the majority of highly paid investment professionals could barely outperform a chimpanzee with a marker was devastating to the active-management approach to investment, leading to the rise in popularity of passively-managed, index-linked funds.
This theory, however neat, would raise a few eyebrows in the seasoned finance veteran community. Firstly, it assumes that all the available information can be gathered and processed by all investors in a more or less equal manner. However, it is clear that the volume of information out there is far too much for even the most enterprising investor to fully and efficiently monitor. More damningly, EMH cannot convincingly explain why market panics happen. All those bubbles and crashes, bull and bear runs that characterize modern markets simply should not exist, according to EMH. The fact is that the EMH model omits one crucial component of the financial markets: human beings. The implications of human emotions, sentiment and fallibility in this system is addressed by another approach to financial markets; Behavioural Finance.
Human bias affects investment decisions
Behavioural Finance attempts to explain how the market functions in practice, rather than in theory, by studying how human psychology affects investment behaviour. These include Mental Accounting, where investors often allocate money sub-optimally to conform to irrational mental biases; Anchoring, where investors lock on to an arbitrary reference point for the value of their investments; Self-attribution, where investors attribute success to their own efforts and ingenuity but failures are shrugged off as bad luck; and Herd Behaviour, the scourge of modern capital markets. The theory also identifies several biases and tendencies that lead to counter-productive investment activities, such as confirmation bias, where investors tend to accept information when it conforms to their predisposed sentiments, familiarity bias, where investors prefer to invest in assets they are knowledgeable about, ignoring better opportunities in other areas, and loss aversion, where investors are vastly more concerned about avoiding losses than they are seeking opportunities for growth. Behavioural Finance goes a long way in explaining the holes and inaccuracies present in EMH, and also answers the question as to why market panics still occur despite massive leaps in analytical technologies like high frequency trading algorithms. These solutions are programmed to mirror the fallacies in human reasoning and merely do so on a much larger and faster basis, leading to ‘flash crashes’ like the one in the Dow Jones index on 6th May 2010, where the market dropped precipitously and recovered almost as quickly in the span of one afternoon, for no discernible reason.
All that being said, is the market forever enthralled to the fallibility of human emotions? Turns out, there may be a nascent solution to this dilemma in the form of Artificial Intelligence (AI). The term ‘AI’ encompasses a wide body of simulation techniques aimed at replicating human intelligence, reasoning and logic to analyze and solve problems. You can find AI systems in nearly every facet of modern finance; it helps banks identify fraud and money laundering, analyzes news sources to gauge market sentiment and powers trade algorithms that search for patterns and opportunities in the equity markets. AI is even being deployed in the fight against coronavirus, to diagnose afflicted patients as well as to identify molecules that can potentially combat this pandemic. In short, it is an incredibly useful and versatile tool.
AI can help reduce bias – and fear
One of the most promising fields within AI today is machine learning, which uses mathematical models to process vast amounts of data, identifies patterns and correlations within it and makes predictions with little to no human intervention. In the process, it effectively ‘learns’ from the data and uses this ‘experience’ to improve its performance and predictions, much like a human being would. The implications for the financial world are substantial; AI systems are capable of processing vast quantities of data at a speed and precision far exceeding human capabilities. Moreover, while AI systems could potentially exhibit unintentional biases (like the infamous case where a noted Youtube creator applied for an Apple credit card along with her husband and was assigned a substantially lower credit limit than him, despite having a better credit score), they can largely do so without falling prey to the psychological conjectures and biases that humans do. The result is that investors incorporating AI systems into their decision-making processes can benefit from actionable insights and predictions derived from constantly self-improving systems that can analyze more data every second than a human being could do over their entire lifetimes. AI is already revolutionizing the world of finance, with a booming industry of new startups aimed at developing AI solutions to maximize efficiency, cut costs and deliver actionable insights for investors of all stripes. Our company, for instance, uses AI to assess credit risk for corporate debt by analyzing over 350 different variables including, rather interestingly, reading textual news sources, and can predict the likelihood of a credit rating downgrade up to 12 months before it actually happens.
Would these solutions completely eradicate market panics in the future? It seems unlikely, as long as human beings are at all involved in the process of investing. Yet the proliferation of AI can help reduce the impact of human biases and fears in setting the pulse of the market, as well as empower investors by giving them access to better information and insights. At a time of crisis like this, it is ever more important that investors look past the tumult and chaos, and AI can go a long way towards doing that.