Secret weapon to beating the market
After being stationary for weeks, an oil rig leased by energy company Apache Corp. off the northeastern Atlantic coast of South America was on the move again last spring.
Satellite and radio signals tracked the rig moving northwest towards Aruba on April 11, an area where Apache didn’t have drilling rights, suggesting its exploratory drilling project off the coast had failed.
Quandl Inc., a Toronto-based startup that specializes in finding, buying and re-selling digital records that predict how a stock or sector will perform, purchased a dataset of those signals and processed it for sale to investors.
On April 21, Apache confirmed its exploration had failed. Investors who had purchased Quandl’s data were able to figure that out 10 days before everyone else, allowing them to sell or short the stock ahead of the pack. Apache’s stock price underperformed the S&P Commodity Producers Oil & Gas Exploration & Production Index by 4.3 per cent from April 12 to 24.
Legal experts say investors may be risking more than their capital when using such alternative data since case law hasn’t yet determined what crosses the line into privacy violations or insider trading, but it’s a risk a growing number of financial institutions are willing to take, especially since in Apache’s case, and many others, it has paid off.
Hedge funds are willing to pay handsomely for Quandl’s alternative data because it helps them achieve alpha, or market-beating returns, that are harder and harder to come by as information becomes easier to find and analyze.
To some, that may sound like an unfair advantage. But to Tammer Kamel, Quandl’s founder and chief executive, alternative data analysis is just a high-tech, high-quality version of the research investors have been doing for decades.
“That is the original alpha source, knowing something the market doesn’t know. It’s beautiful,” he said. “If you can come to them with a genuine information advantage, where they can know something their peers in the market do not know that’s tradable, that’s hugely valuable.”
Alternative data has grown explosively over the past two years and Kamel said his company is the world’s top provider of it, with annual revenue on track to double for the second year in a row. The space is heating up, however, with startups such as RS Metrics and Goldman Sachs Group Inc.backed Kensho providing competition.
But Quandl, which also sells traditional financial data in analysis-friendly formats such as CSV (comma-separated values) files via an API, now counts almost all the top hedge funds, asset managers and investment banks in North America as clients.
Kamel said the idea for Quandl came while working at a hedge fund as a quant — a computer expert who uses data to build analytical models that look for predictive signals — and became frustrated with financial data providers such as Bloomberg and Reuters, which tend to bundle their data with proprietary hardware and software that make high-level analysis difficult.
Quandl launched in early 2014 and has had a good business selling traditional financial data, which continues to be the company’s main source of revenue.
But about a year and a half ago, Kamel took note of all the new data being produced by satellites, drones, mobile phones and other sources and realized there was an opportunity to provide investors with an additional edge.
“This is the opportunity to go into the wilderness and find data that Wall Street has never seen before,” he said.
Kamel said he encourages his employees to think like their hedge-fund clients in identifying data sources that predict market-moving information before the market has a chance to move. Quandl is particularly interested in companies that produce what it calls “exhaust” data, or data collected as part of a company’s normal operations without intending to turn it into a revenue source.
For example, insurance companies keep records of how many new car insurance policies they sell, as well as which vehicle manufacturer’s model is being insured, which happens to be a great predictor of new car sales before the automakers release the data themselves.
But, Quandl faces a dilemma after convincing suppliers to sell their data: the more clients the company sells the data to, the less of an investing edge it provides, making it less valuable.
To solve that problem, Quandl uses the data to build a predictive model to make an educated guess about how much money could be invested before the data loses its advantage and then sells it to a limited number of clients accordingly.
Kevin McPartland, head of research for market structure and technology at research firm Greenwich Associates, said this technique will become more and more challenging as competition heats up in the alternative data sector. For now, there are still plenty of financial institutions that haven’t caught onto the trend, but he said that won’t be the case for long.
“The more you sell, the less valuable it becomes,” McPartland said. “Finding that middle ground is a real business challenge.”
Another business challenge Quandl faces is in navigating uncharted legal territory. Legal issues raised by alternative data fall into two categories: privacy and insider trading.
Kamel said Quandl refuses to purchase any data with personally identifiable information and vets each dataset to make sure it’s scrubbed of all identifying traces. As for insider trading, he said he prefers to buy data produced by one company that tells an investor information about a different company to head off allegations that the information came unfairly from the “inside.” For example, the insurance data noted above tells investors how the automakers are doing, not the insurance company.
“They’re not telling us anything insider about themselves. They’re telling me an interesting fact about Ford and General Motors,” Kamel said. “It can look like insider information, but it’s how you get it that matters.”