Historically, market data tells us where the markets are now and where they’ve been, with traders leveraging latency-reducing technologies to get as close to “now” as possible. But the next step for data is to use predictive analytics to tell us where the markets will be in future.
Firms already use historical data to analyze how markets respond to events. By examining their reaction in the past, one can take a qualified guess at how they will react to similar situations now. The better you can predict the market in a second, minute or hour‘s time, the greater advantage you have over those basing their analysis on where it stands now.
Predictive analytics are also being used to manage risk, such as in the case of a new suite of tools developed by Dun & Bradstreet to assess the credit risk of borrowers and predict not just whether but when they will default. Yes, those same algorithms that recommend with uncanny accuracy what you might want to buy next on Amazon are already being put to use in the financial markets.
It is in part the capabilities of predictive analytics—and the sensitivity to others being able to predict trading behavior—that led to such a scandal over Bloomberg journalists’ access to high-level information on how clients used their terminals. After all, if a reporter could ascertain your strategy based on your activity on the terminal, what they do with that knowledge could render your strategy useless.
So it’s understandable that banks would be uncomfortable at the suggestion that anyone—reporter, rival, or anyone else—could see any detail of their research, since any indication of what data they are looking at could reveal something about their strategy. For example, Tick Data Corp, which last week executed a management buyout from new owner Disnat (the online broker of Desjardins Group, which acquired parts of Tick Data’s now-bankrupt former parent, Penson), realizes that the data its clients request can indicate their next trading moves.
Tom Myers, VP of business development at Tick Data, says that since all its developments—including the addition of historical data from new exchanges—are driven by client needs, “We are a leading indicator of sorts—because if a firm is running a strategy in the US and wants to try it out in Europe or Asia, their first call might be to me… to acquire historical data to test their models.”
Indeed, one of the main reasons why trading strategies have a much shorter shelf-life today—a matter of days or weeks before they are no longer competitive, compared to months at a time in the past—is that firms put as much effort into trying to figure out their rivals’ strategies as they put into creating their own. In some cases, the two are part of the same effort: to build an effective strategy of your own, you have to understand how your competitors will behave against it.
“Firms like Goldman Sachs and JP Morgan are in the business of finding out this kind of information, so it’s ironic that they are whining to Bloomberg for doing to them exactly what they do to everyone else,” one source quipped.
As if to prove this point, when I visited JP Morgan’s website seeking confirmation of the latest twist in the saga—that the bank apparently demanded five years’ worth of Bloomberg’s data on its activity—my antivirus software blocked access to its page of press releases, warning that it behaved like a phishing site.
Industry observers say the data industry already operates with a high degree of integrity, but one source says the current furore could be the tip of the iceberg if any suggestion arises that anyone was able to profit by trading on knowledge resulting from monitoring clients’ activity. And—if only because in the world of premium data terminals, it can really pay to complain—I predict that the vendor’s clients will have a lot more to say.
Of course, the prediction that everyone wants to make is who’s won this year’s Inside Market Data Awards, but for that, you’ll have to wait until our awards dinner following the North American Financial Information Summit on Tuesday.
Should regulators take a more active role when it comes to AI oversight, or leave it to the professionals? What will M&A look like in 2018?Subscribe to Weekly Wrap emails