The role of data in investing…

In the early 1990’s Eugene Fama and Kenneth French introduced the Fama-French three-factor model. They posited that asset returns could be explained through the use of three factors; a size variable, a value variable and a systematic risk variable.

At the time the model was considered a breakthrough, certainly a marked improvement on the ubiquitous Capital Asset Pricing Model. It was also deemed, in some circles, an overly quantitative, data-driven model.

We might look back on this today and say that just three factors might not even qualify as a quantitative, data-driven model. In fact, though the model remains useful for systematic explanations of asset returns, most experts would observe it as being outdated or, at least, underspecified.

The point here is that Fama and French published their paper in a time were heavy data processing was still in its nascent stage. The concept of data driving analysis was only just beginning to take charge and financial analysts were still firmly entrenched in the bottom-up, qualitative mindset.

But the advent of prolific data generation has allowed for the creation of quantitative models that are more accurate, have better fit and greater predictive power. We can look at data in financial analysis through the three V’s of big data; volume, variety and velocity. In every one of these aspects there’s been a rapid increase in data availability. The sheer volume of security data available now is greater as is the broad nature in which is captured. By virtue of this, the velocity at which data is generated and the speed at which it can be crunched have both drastically increased over the years.

However, the use of data in finance isn’t just limited to traditional methods being improved upon. Analysis has extended to the incorporation of Google Trends data to determine the popularity of a company’s brand or products. This data is then correlated to the sales recorded in the company’s accounts to test whether a lead-lag effect exists. Other novel macro variables, previously uncaptured or too infrequent as a data series, are also being used predictively for more traditional financial and economic metrics.

However, the use of data isn’t just limited to moving towards a more quantitative method of investing. Robo-advisors are providing a whole new realm of financial advice altogether. Providing portfolio recommendations based on algorithms crunching vast amounts of data might have been unseemly 10 years ago, but with passive investing becoming more popular it makes sense for retail investors to forego high management fees and invest with a robot.

There is plenty of room yet for data to take up a greater role in investing. As all aspects of society and business become recordable as data points, more of our understanding will shift from expert qualitative knowledge to skilful quantitative analysis. Of course, this data and its results still have to be interpreted. Whether financial analysts are comfortable letting a machine do that is still to be seen.

 

[1] https://www.infoq.com/articles/big-data-in-finance

[2] https://www.wallstreetmojo.com/quantitative-financial-analyst/

[3] https://www.huffingtonpost.com/entry/what-can-you-do-with-big-data-in-finance_us_58bd70f1e4b02eac8876d116

[4] https://portfoliosolutions.com/latest-learnings/fama-french-three-factor-model

[5] https://www.lynda.com/Excel-tutorials/Financial-Forecasting-Big-Data/569337-2.html