Financial Analysts vs. Data Scientists

Financial analysts could’ve been the first data scientists…

Earlier this year, the Chartered Financial Analyst (CFA) Institute announced a decision to include artificial intelligence, automated investment services, and mining unstructured data in its 2019 exams (Reuters). Stephen Horan, the Institute’s managing director for credentialing, says that the expanded coverage will help future analysts “…distinguish between structured and unstructured data analytic methods, as well as identify characteristics of robust investment algorithms.”

It’s been a long time coming.

Using data for business is nothing new. Financial analysts have been at it for decades, long before the first data scientists popped onto the scene, and, unfortunately, long before data became cool. Like hipsters before hipster things hit mainstream.

Like data scientists, financial analysts have good analytical skills and a nose for business problem solving: What data should we measure? What types of analysis can we perform? How do we use data to drive action? The only big difference between the two is technology and statistics.

But, clearly, data scientists are getting the better end of the deal.

A quick search on GlassDoor shows that, on average, data scientists earn nearly twice as much as financial analysts. What is more, they’re taking jobs from financial analysts—apparently, data science is now the hottest job in Wall Street.

Why is this the case?

Financial Analyst Average Salary (Glass Door)
Financial Analyst Average Salary (Glass Door)
Data Scientist Average Salary (Glass Door)
Data Scientist Average Salary (Glass Door)

Mo data, mo problems

Back in the day, data was simple. We didn’t have too much of it, and what we did have could easily fit into neatly arranged rows and columns in spreadsheets. Excel was more than enough in most cases.

Then the Internet, relational databases, and their concomitant technologies happened. All of a sudden, even the most esoteric phenomena were turned into data points. We started collecting data on everything and anything we could get our hands on—online activity, transportation patterns, food consumption, exercise habits, even cats in movieswhich animals fart, and abandoned trolleys in Bristol river.

A study by IBM showed that over 90-percent of the world’s data was created post-2010, much of it unstructured data from sensors, websites, and the ubiquitous Internet of Things (IOT). We have reached a point where data has grown so vast in terms of volume, velocity, and variety that the old rules no longer apply. We have effectively outgrown the spreadsheet.

This posed a conundrum for financial analysts. Companies were no longer content with financial reports and balance sheets; they wanted to extract and analyze all kinds of data from all kinds of sources, and get feedback in real-time. This called a new set of skills altogether.

The most prepared to meet this demand were software engineers, computer scientists, and statisticians. Not necessarily people who understood business but who, at least, knew how to source, publish, and analyze data. Over time, these people picked up on the business side of things, and data science was born.

We should ask ourselves why the inverse did not occur—that is, why didn’t financial analysts pick up data skills to meet the new demand? Some attribute this to the general slowness of financial professions, others, to the lack of the right incentives (and disincentives) to encourage upskilling.

Whichever the case, it was an opportunity missed. Had financial analysts learned to manage new types of data, they would’ve easily been more relevant today, especially that companies now demand faster reporting, smarter metrics, and, of course, more data.

Unfortunately, much of that is already beyond the purview of the average financial analyst.

A word on computers

It can be hard to imagine, but computers used to be people. In the early 1940s, firms like NASA employed human computers (mostly women) to solve complex math problems—such as the number of engines needed to keep a plane airborne, or the optimal trajectory of a rocket.

Human computers - the women of NASA
The lady on the right reminds me of my mother… board
Good Housekeeping

Quo vadis?

Our data-supercharged future isn’t all doom and gloom for financial analysts. If they keep the status quo, the pay is still pretty decent; and, if they want to adapt to today’s data, the gap isn’t as wide as many believe.

Many data skills can be learned fairly quickly, and there are a lot of free data courses to take advantage of. A PhD in computational statistics or machine learning might be nice, but it isn’t mandatory. Starting with something simple, say SQL or Python, can already open a lot of doors; the important part is to combine those new things with some good old-fashioned critical thinking and problem solving skills.

All that said, the recent CFA decision is a veritable sign that a major shift is in order. Old tools and techniques in financial analysis can’t possibly handle the volume, velocity, and variety of today’s data.

The jury has spoken: financial analysts need to adapt to the changing data landscape if they want to remain relevant…or at least get paid more.