Abstract thinking matters

A week ago, I wrote about how strategic thinking (the ability to think several steps ahead) is a crucial success factor for traders. Meanwhile, companies that hire traders typically focus on analytical skills rather than the ability to think strategically as a key factor for successful traders. Analytical thinking is indeed important, but less so for traders and more for financial analysts. In particular, the ability to think in abstract terms and scenarios helps analysts make better recommendations.

Financial analysts are obsessed with data. They try to collect as many detailed data points as they can, in an effort to understand a company and its market as deeply as possible. Yet, having more data does not make your forecasts better – a fact that has been well documented for more than 50 years. In 1965, Stuart Oskamp published a study where he asked clinical psychiatrists, psychology post-grad students and undergrad students to assess the personality of a person and forecast his likely behaviour in certain situations. He provided information in four rounds, each round disclosing more detailed information about the person’s past behaviour and circumstances. The chart below shows the accuracy of the forecasts of clinical psychiatrists together with their confidence, in how accurate their forecasts were going to be.

More data leads to higher confidence but not higher accuracy

Source: Oskamp (1965)

This effect has been observed in all kinds of fields including with investors in the decades since. The key driver seems to be that more data helps make better predictions in an environment where the relationship between cause and effect is stable (e.g. the natural sciences or medicine) and where there is a relatively low level of uncertainty. But if uncertainty is high, no matter how stable the relationship between cause and effect, forecasts are not going to be accurate (think of playing the lottery or forecasting a roulette wheel). Similarly, if uncertainty is low but the relationship between cause and effect is unstable forecast accuracy is again very low (think of the screeching noise a microphone can make if it is subject to a feedback loop, you cannot forecast what tone will come next in that feedback loop).

But financial markets just like individual human beings are impossible to forecast. Uncertainty around outcomes is high all the time and the relationship between cause and effect changes constantly. Hence, more data may not lead to better forecasts.

Instead, using more abstract ideas and scenarios seems to be the way to go. A study from researchers at Singapore University analysed the type of questions financial analysts asked at company earnings calls as well as the comments they made about the answers from company executives.

The study created what they called an Abstract Thinking Index (ATI) for each analyst that indicated how much the analyst thought in abstract or concrete terms. Analysts that asked more about future events rather than past developments were considered to think more in abstract terms. Analysts that focused more on the ‘why’ rather than the ‘how’ of certain developments were considered to think more in abstract terms. Analysts that talked more about broad topics rather than narrow, concrete topics were considered to be more abstract thinkers. And finally, analysts who used more abstract words and concepts in their questions were considered to think more in abstract terms (duh!). In essence, analysts who think more in abstract terms were less concerned about gathering more and more factual data about a company, trying to model every detail, but rather focused on the big picture and the key drivers of performance.  

In the end, the study found that analysts who thought more in abstract terms than in concrete terms or examples gave better forecasts and were given more attention by investors. When these analysts changed their recommendations on a stock, share prices moved more than for analysts who think more in concrete terms. And on average, following the recommendations of analysts who think more in abstract terms when they change their recommendation on stocks is creating outperformance. In particular, it seems listening to abstract thinking analysts when they downgrade their recommendation seems to be a good way of avoiding losing stocks.

Abnormal returns of stocks up- and downgraded by analysts

Source: Jin et al. (2022)