Big Data’s Big Role in Finance and Financial Regulation
August 01, 2018
Picture this: you’re tired of the spam in your inbox, so you download a new app for your browser that blocks it. While downloading, the Terms of Agreement pop up, and you click ‘Agree’ – because why wouldn’t you? Unbeknownst to you, while you are now enjoying your spam-free email, the Slice Technologies app is analyzing your emails for purchase receipts and selling this anonymized data to hedge funds. Is that an invasion of privacy? Not quite, as you agreed to the terms. But why would hedge funds, and other investment advisers, want this information? Well, with this kind of alternative data, investment firms can make much more accurate predictions about a company’s sales revenue and its health. This new world of alternative data poses incredible alpha-creating potential for investment advisers, as well as new legal concerns for the courts and regulators.
As the old adage goes, knowledge is power. In today’s information age, data is being created incessantly by nearly everything we do: the websites we visit, the places we go, the purchases we make, and much more. This alternative data, paired with powerful computational analysis, has the ability to inform financial decisions and identify economic trends well before traditional analyses ever could – and investors are catching on. Alternative data are data sources outside of the typical traditional sources used by investment advisers, like government reports, SEC filings, and quarterly earnings reports. Alternative data can be satellite imagery of crops used to predict yields or anonymized reports of what people are purchasing with their credit and debit cards. This data is typically collected and cleaned by data vendors that package and sell it to Wall Street for millions. According to AlternativeData.org, in 2016, the alternative data market was worth $200 million, and by 2020, spending on alternative datasets and associated infrastructure is expected to reach $1.7 billion. Some larger firms like Blackrock and Goldman Sachs already use internal programs to interpret news stories, analyst reports, and social media by using natural language processing to make informed inferences about capital markets.
Alternative data provide algorithms with the fuel to make financial decisions and predictions more accurately than ever before, and investment advisers are increasingly incorporating quantitative trading methods into their investment strategies. According to the Tabb Group, quantitative hedge funds account for 27% of all U.S. stock trades by investors, and these firms are constantly searching for more innovative and effective sources of data. For instance, data tracking corporate jets is used to predict mergers and/or acquisitions. In 2017, hedge fund managers anticipated the acquisition of Actelion, a Swiss pharmaceutical company, by Johnson & Johnson after tracking the company’s flights to Switzerland.
While alternative data is extremely valuable, the point at which alternative data crosses the line legally is not yet clear, and regulators know it. When does having this type of granular insight cross the line into say, privacy violations or insider trading? In terms of privacy, data vendors typically try to avoid any liability and remove personally identifiable information (PII) from data that is sold. The Investment Data Standards Organization (IDSO), a 501(c) nonprofit founded in January 2018, creates standards for the new alternative data industry. One of their products, a current work-in-progress, is a list of best practices when dealing with data sets including PII. Fortunately for us, granular level data is not that important to those purchasing alternative datasets; what is important is aggregate level trends for entire companies or industries that can be inferred from the sum of many individuals’ purchases or other behavior. However, some states are taking a more proactive approach to the data protection problem. On June 28, 2018, California passed a bill that requires companies to disclose what kind of data is being collected and to what third parties it is being sold. It also requires that consumers have an ability to opt out of having their data sold.
In terms of insider trading, there are still questions regarding alternative datasets. The point at which alternative data becomes material nonpublic information (MNPI) is not yet clearly defined. According to Bloomberg Law, much of the risk that a data purchaser takes on lies in ensuring the data vendor has the legal right to sell the data at all. Investment firms can minimize this risk by asking data vendors for warranty that the data sets do not include MNPI, and that the data was collected in a way that did not breach the vendor’s legal duties to the source. Tammer Kamel, CEO of an alternative data provider Quandl, prefers to purchase data from companies that provide investor information on different companies to avoid the idea that the information came from an insider. However, much of the discussion on the role of alternative data in insider trading will hinge on what is considered nonpublic information. For example, while alternative data on new iPhone sales may not be widely disseminated, nothing is stopping an investment adviser from surveying everyone that walks out of an Apple store to ask if they bought the new iPhone.
However, clear cases of insider trading using alternative data do exist. In SEC vs. Huang, two former Capital One employees were prosecuted for misappropriating a confidential database of credit card purchases and making a $2.8 million profit on a $147,300 investment. The employees made searches on the database of transactions to better predict the sales of public companies ahead of their quarterly earnings reports. In one instance, after analyzing sales of outdoor gear retailer Cabela’s, the employee purchased $51,890 in put option contracts. The next day, after Cabela’s announced a 10% decrease in sales, he sold the contracts and made a 108% return. Here, the SEC successfully argued that this case met the criteria for insider trading: material nonpublic information was misappropriated in a breach of duty, and was then traded on. While this fits the bill for misappropriated insider trading, this example illustrates how data can and will be misused to outperform the market going forward and why regulatory agencies must be able to address this new problem.
As big data and alternative data continue to play a larger role in investment decisions, the U.S. Securities and Exchange Commission (SEC) and other regulatory bodies will have to keep up to ensure fair and efficient markets in the changing financial landscape. With the power of big data, regulatory agencies like the SEC and FINRA can move away from reactionary enforcement policies – like only starting investigations after suspicious activity has been detected – and toward a more proactive approach. Using a “trader-based” method of detection allows agencies to identify suspicious trading patterns among networks of traders and potential sources of material nonpublic information using analytics. The SEC has already started with the creation of the Market Abuse Unit (MAU). In one 2017 case, the MAU was credited with detecting “patterns of insider trading” that led to the sentencing of an investment banker and plumber on insider trading charges., The banker had sold tips on the status of 10 mergers and acquisitions, allowing the other to earn over $76,000 in profits. With new techniques and technology, the SEC and FINRA, among others, will be able to more effectively identify illicit activity.
While innovative uses of big data propel investment firms forward in terms of performance, investment advisers must be careful to avoid legal risks that include privacy violations and insider trading. This unprecedented era of data collection and analysis has changed the financial landscape, but it will also change the regulatory landscape: enforcement agencies must adapt and use the same analytical techniques to detect illegal activity more efficiently.
Student Blog Disclaimer
The views expressed on the Student Blog are the author’s opinions and don’t necessarily represent the Penn Wharton Public Policy Initiative’s strategies, recommendations, or opinions.