Buy the market when everyone is in despair, sell when everyone is exuberant. But is this as simple as it sounds? Analysis and strategy are certainly required. Thus, a Swiss asset manager has developed a unique and successful investment strategy based on neuroscience.
By Michael Kometer, Founder and CIO, and Wolfram Klingler, Partner and Investor at Neuronomics AG (an asset manager specializing in Computational Neuroscience and Quantitative Finance)
The full quote from Warren Buffet goes: “Be fearful when others are greedy and greedy when others are fearful”. In other words, buy the market when everyone is in despair, sell when everyone is exuberant. The rule is so basic, it makes sense, it is empirically validated, Warren Buffet has become one of the richest persons on earth applying it. Yet why is it difficult to impossible to follow his advice?
What neuroscience says about economic decision making
Neuroscience has highly relevant and interesting answers to this question. The answers come from experiments and research where scientists use advanced imaging methods to track and understand what happens in the brain while we take economic decisions. Such experiments have gone as far as analyzing the neuronal processes of traders while they trade real money in live environments, tracking their neuronal processes with EEG’s and fMRI’s.
But before we discuss insights neuroscience can give us regarding this question, let’s have a look at how the prevailing understanding of the market looks like today. One of the most used hypotheses, both in academia and in business, is the efficient market hypothesis. According to this theory, all available information is already priced into each asset and the market participants are rational actors, trying to maximize their utility (the infamous “Homo Oeconomicus”). This hypothesis assumes a linear information transformation, in which information is processed by rational participants that take rational decisions.
Surviving in the wild vs navigating financial markets
Neuroscience shows that this is far from reality and that there are hypotheses that do a lot better explaining what effectively happens in financial markets. Our brain has been built to survive in the wild, not to navigate financial markets. And the time span since financial markets exist is much too short for our brains to adapt. Experimental research shows that we weight losses much stronger than gains and that we are challenged when it comes to taking low probability events into account. It is safe to say that neuroscientific research proves our processing of information to be highly non-linear.
Understanding how information is processed in the brains of financial market participants opens new horizons. The evolutionary development of our brain has led to two basic reactions: approach and avoidance. We are designed to avoid and shun away from situations that are perceived to be risky and uncertain, as such situations reduce our chances of survival in the wild. We are neurologically designed to look for situations in which rewards seem to be available. Our learning is strongly shaped by both aversive and pleasant situations. Against the background of this very basic observation, the asymmetric assessment of gains and losses becomes quite obvious and makes sense from a biological perspective.
The illustrations show the avoidance and approach brain network which are activated as financial market participants process gains and losses. These networks influence the decision making of financial market participants and may lead to overreactions and market inefficiencies.
In other words: while many financial market decisions may be suboptimal from an economic point of view, they make sense and are optimized for biological survival. Equally, research shows that the general risk aversion increases massively after a crash, although the statistical probability for a further crash is significantly lower after a crash. We are biologically designed to avoid aversive situations.
How our brain creates market inefficiencies
Neuroscientific research shows how neuronal processes of market participants lead to inefficiencies. In other words, such processes lead to market movements that go beyond what is fundamentally justified. This is easy to observe in bear markets, but also in bull markets, where the continued reaping of rewards leads to an increase in confidence levels beyond reason, which often is the basis for the reverse market movements that inevitably follow. Ultimately, we can observe such patterns at a smaller scale on a daily basis in all markets, whether it's equities, fixed income, commodities, or cryptocurrencies. This must be the case if the patterns are a result of human neuronal processing and not of market information. The more efficient and liquid markets are, the weaker and less frequent such patterns appear, but even in the most efficient markets, for instance, the US stocks with the largest market capitalization, we can still observe these patterns regularly.
While the typical quant-based asset management strategies might find such patterns in their statistical analysis of historical market price data, it is difficult to impossible to distinguish stable patterns induced by neuronal processes of market participants from the indefinite number of patterns that can be found in any given set of historical market data.
Applying neuroscience in quantitative finance
To solve this problem, Neuronomics has developed a novel approach. We use the latest research from computational neuroscience to define patterns that are induced by neuronal processes and then scan the markets for such patterns, rather than trying to identify patterns from the markets themselves. This is crucial in avoiding the biggest challenge in quantitative asset management: Overfitting of the models. Overfitting is the process by which we adapt our algorithm too much to a given market data set, which leads to unreliable results as soon as we apply it to live markets.
With a deep understanding of the neurological processes which underlie economic decision-making, we can fundamentally improve our market understanding. As these patterns are stable, their identification allows for a high-quality short-term prediction of specific market situations across different asset classes. Ultimately, these market inefficiencies can be instrumental to generate systematic alpha. There might be something like a Neuro-Factor, just as we can observe factors such as liquidity, size, or momentum.
A novel approach to investing in Crypto markets
After years of intense research and development, Neuronomics AG is trading live since July 2020 with its quantitative approach based on computational neuroscience. The result is a strategy with little correlation to any market and consistent high risk-adjusted returns. Initially, we have applied our approach to the Crypto markets, as this is the market with the least fundamental data available, a market that is relatively inefficient and strongly driven by investors' emotions. Therefore, this strategy also offers the highest return potential amongst all asset classes, even if it is limited in size due to the relatively limited liquidity in Crypto markets. A quantitative strategy based on computational neuroscience for the equity market will follow in 2021.
Return of the Neurofin AMC (green) vs CCI 30 Index (violet) during market reversals in September and November 2020. The net performance of the strategy since its inception at the end of July 2020 is 12.55% per November 30th, with an annualized Sharpe ratio of 2.5. The strategy offers a superior risk-adjusted return with low market correlation.
To make the strategy investable, Neuronomics has partnered with GENTWO, a Zurich-based innovative securitization specialist offering a new generation of financial products. Through GENTWO’s securitization platform, the strategy is available to professional investors, and despite an underlying Crypto portfolio, fully bankable with a Swiss ISIN.