Kumesh Aroomoogan has come a long way from his days at Wall Street bank Citigroup, where his job was to stay on top of breaking news. He remembers getting yelled at by a trader because he stopped watching news feeds to go the bathroom and missed a market-moving headline. Now, he’s the co-founder and CEO of Accern, a company that automates his old job.
Tedious office roles are increasingly done by computers, freeing up workers to do things that are (hopefully) more useful. To some extent, even the trader who once yelled at Aroomoogan has been automated: One of the hottest areas of finance these days is quantitative investing, which uses artificial intelligence to sift through massive troves of data to identify signals that humans can’t see. Quant funds are Accern’s biggest customers.
Aroomoogan’s bathroom-break schedule is less fraught now, because instead of supplying a human trader with tips about breaking news, Accern’s technology sweeps up data from 300 million websites, 150 million Twitter feeds, as well as analyst presentations and FactSet reports for traders—either humans or algorithms—to analyze. It uses natural-language processing to find keywords like company names, and measures when a story is rising up the media food chain, such as from blogs to newswires, to indicate that it may be important enough to act on.
“Quant hedge funds are buying as much data as they can,” Aroomoogan says. Top banks like Goldman Sachs are hosting events for their clients to meet with data vendors because their customers want to know everything about the latest available data. Investment banks want their own quants to be on top of things, too.
The so-called “alternative data” market was worth about $200 million in the US last year and is expected to double in four years, according to research and consulting firm Tabb Group. In addition to public websites, hedge funds are collecting and crunching data generated by credit card transactions, satellite images of parking lots, and customers reviews.
The amount of stored electronic data is growing exponentially, but the trick is getting it into a useful state for algorithms to devise profitable trading strategies based on it. To this end, firms that specialize in refining the information (paywall) are springing up, providing actionable data for investors to shovel into their computers. Tammer Kamel, CEO of alternative-data platform Quandl, says companies unknowingly sit piles of profitable data of great interest to Wall Street.
Traditional banks and old-school buy-and-hold investors are using alternative data, too. Some of it is marketing hype—adding the words “artificial intelligence” or “big data” to any investment offering can lure some customers. But analytics can also save a traditional money manager time by sifting through news and data on their behalf.
Hedge funds need better marketing, as investors are increasingly skeptical of them charging a 2% management fee as well as a 20% performance fee for results that have been lackluster in recent years. Billionaire investor Warren Buffett expects to win a bet (for charity) that hedge funds would underperform the S&P 500 index over a 10-year period.
The enthusiasm for quant funds is encouraging for hedgies. Family offices that manage assets for the wealthy are backing away from hedge funds generally, but appear to be boosting their investments in quant funds.
Sadly for them, however, the scope for making money from these strategies could be fleeting. One way to maintain an edge is to get exclusive access to data so competitors can’t benefit from it, but the legalities quickly become murky (paywall).
In the meantime, sweeping up as much data as possible into high-powered machines running whizzy algorithms is an increasingly common practice among hedge funds. News and data companies like Bloomberg and Thomson Reuters now include alternative data in their offerings, and about 75% of hedge funds already use social media and social-driven news feeds to inform investing decisions, according to Greenwich. Alternative data, it seems, is fast becoming mainstream.