Quandl Uses Algos to Predict a Company's Future Earnings

Alternative data provider Quandl, which was acquired by Nasdaq last year, has launched a new dataset that tracks and ranks companies based on their earnings quality and provides additional insight for equities trading.

The service, called Earnings Quality Rankings, or EQR, looks at a company’s earnings report and determines how likely its income will grow, remain profitable or fall.

It uses algorithms to scrape financial statements to connect patterns and combines that with human expertise to figure out which patterns are important. Quandl then takes the results and assigns each company a rank. The ranking goes from “1” being the lowest quality to “5” as the highest. The rankings—released and updated weekly as companies release their income—point to firms whose earnings reflect true figures and predict which ones will be the steadiest to bet on.

We find that if earnings quality is high today, it’s likely to stay high next quarter and maybe the quarter after that; but if you’re looking for earnings quality, say, three years from now then there’s less correlation.

Abraham Thomas, Quandl

Earnings quality refers to the ability of current earnings to predict future earnings. This discounts any accounting changes that may have boosted the bottom line for that quarter, says Abraham Thomas, chief data officer at Quandl.

“It’s presenting a view of earnings quality that we think is unique in its methodology and its predictive ability. Earnings quality is not a new concept at all; academic papers have been written about it, and everybody thinks about the quarterly earnings of a company when investing anyway,” Thomas says. “But when you talk about calibrating 20 years of data in thousands of fields, it’s probably something that was computationally quite hard to do until maybe five or 10 years ago. The size and scope of the computation and the availability of data were not that easy to get a hold of a decade ago.”

He adds that it’s clear that a company generating profitable, repeatable sales have high earnings quality because it can be inferred this strong performance will continue. However, if a company makes money by changing its accounting or by selling off its assets, continued strong performance is less of a guarantee.

The dataset covers 9,235 US equities mainly listed on the New York Stock Exchange and Nasdaq. Quandl ranks firms that meet a market capitalization of at least $150 million. Each company is classified by sector to make company comparisons more useful, Thomas says. Baseline financial statement data is collected from Thomson Reuters.

Thomas says ranking companies by earnings quality by sector gives an “additional sanity check” for investors seeking to filter out companies when working out its short- or long-term investment strategies. These high earnings quality firms, he says, are less volatile and outperform those of a lower ranking.

While EQR is able to predict earnings quality for a few quarters, Thomas notes the predictive quality is not meant for estimates made on a longer time horizon.  

“We find that if earnings quality is high today, it’s likely to stay high next quarter and maybe the quarter after that; but if you’re looking for earnings quality, say, three years from now, then there’s less correlation,” he says.