Research & the Role of Big Data
Traditional fundamental research shops became popular prior to the financial crisis and boomed as a cottage industry shortly thereafter, as analysts sought third-party expertise. However the research industry is not a revenue generator; it provides research on subscriptions paid by broker revenues via commissions. And currently, with commissions down and soft-dollar spending under attack, the business model is suffering.
Industry followers have watched boutique and full-service brokers massively cut back and consolidate in the last five to seven years, and coverage has shrunk significantly across remaining providers. Very little attention is given to names outside the S&P 500, and it's estimated the average coverage of the top full-service brokers is no more than 250 companies per organization. Not that it matters much: In recent years stock picking strategies have lost popularity in favor of basket trading, or simply moving in and out of asset classes, which has further drained commissions.
"To an extent, individual stock picks have come back into play in this last year so we've seen a bit of an upsurge, but we haven't seen a commission flow for research go up," says Indy Sarker, chief executive at Analec, a financial technology company offering platforms and solutions to the investment research industry. "At best it's still flat, if not declining." He says for the first time the industry has realized this is not a cyclical market, but is a structural problem. "There's a massive difference between waiting for a rebound and realizing there will be not be one if you don't reengineer the business model."
So the question is: How can research firms reposition themselves to provide valuable insights to buy-side clients and remain profitable?
Alternatives are to hire quants that use big-data technologies, hire analysts who can work with big data to streamline their workflows, or replace the analysts completely with big data. After all, years of effort have brought forth technologies capable of aggregating multiple data sources that analysts are already looking at on a piecemeal basis, and software can apply fundamental research rules that a typical stock analyst would be doing anyway to find meaning and insight.
But most brokers are not headed in this direction. They are largely stuck in the traditional methods, using the same old model of research by individuals. Understandably, many analysts have spent decades in the business, during which they've become pros at manually aggregating streams of data and running proprietary analysis, pocketing 6+ figures a year to do so. Entire boutiques have been founded around the celebrity of their star analyst, who likely covers no more than 20 firms at a time. Providing big-data research to clients would be a huge transformation, but to ask those seasoned analysts to train on new tools, build statistical models to come up with directional trading signals, and streamline workflow so they can cover more names... that's unlikely to be well received.
As for research shops bringing in analysts with these new technical and analytical skill sets, Sarker says he just isn't seeing that happen very much, a trend that makes him skeptical of the industry's future.
Meanwhile, those that are making attempts at leveraging data tools are struggling to incorporate it into the traditional flow, because the human analyst is still a bottleneck. Chaith Kondragunta, CEO of AnalytixInsight, a research company that produces automated fundamental stock research reports, notes that if an analyst traditionally had access to 100 gigabytes worth of data on the companies he covered, and there are now 100 terabytes on those companies, the person consuming it is still only a single analyst, so his throughput is still never more than 20 firms. "For them, just because there are more data available doesn't mean there's more research available."
Opportunities in the black hole of market coverage
Still, conversation is brewing on the role of technology to help research shops more efficiently deliver commercially viable research to the buy-side. One glaring target is mid-cap and small-cap firms that are thinly, if at all, covered by the remaining analysts.
Although automated fundamental research reports may not be as good as one constructed by an analyst who has covered a firm her whole professional life, one could argue that today's models can get her pretty close at great speed. Skilled and seasoned analysts are also continually improving upon those models.
Kondragunta argues, "Investors today are getting better served with immediate automated reports than having analysts say, 'Give me a few days.' We cover 50,000 stocks around the globe; there's a massive long tail in the companies that are searched. Previously those companies just didn't have any research available, or it was hard to access if there was any available. Now people are coming in and able to access it -- it's right at their fingertips."
AnalytixInsight's software platform, CapitalCube, completes more than 100 billion computations on stocks each day, automatically updating the research reports that are written in plain English. "Our firms duplicate a lot of the work of analysts, and that's where the game changes." Kondragunta estimates this kind of automation can complete 80 to 90 percent of the workflow, the rest is the analysts adding color and their unique insights.
Mass automation of fundamental stock research also lends itself well to the rising popularity of exchange traded funds (ETFs), which can track hundreds of stocks, big and small. "If somebody wants to provide fundamental research on an S&P500 ETF, they better know the research on all 500 stocks under that ETF, but because there's no coverage in bank brokerage on all institutions that make up ETF, they can't currently provide good fundamental research on it," explains Kondragunta, who says he is seeing exponential demand for ETF reports. He adds that institutional investors are frequently using ETFs to hedge positions. "If you want to use ETFs as hedge you must know what you're looking for to counter your positions. You'll want to do fundamental risk matching on speed and scale."
Becca Lipman is Senior Editor for Wall Street & Technology. She writes in-depth news articles with a focus on big data and compliance in the capital markets. She regularly meets with information technology leaders and innovators and writes about cloud computing, datacenters, ... View Full Bio
Tweet This