Smartphone data can highlight when firms are under SEC investigation—and could expose insider trading
Investigations into companies by the Securities and Exchange Commission (SEC) can be market-moving events. And for those looking for an early tip as to the agency’s movement, it’s possible to gain some insight into which companies might soon fall under investigation by tracking the data of smartphones that spend a lot of time in SEC offices during the day. That’s the finding of a new paper by academics at four U.S. universities, which tracks geolocation data across 26 major U.S. metro areas. The academics identify what they believe could be SEC devices based on repeated visits to SEC offices during working hours, then monitor the devices as they move around—including visiting the headquarters of companies during working hours. This method allows the researchers to capture interactions between the SEC and firms that may not be observable through traditional datasets. “We cannot observe any one particular individual,” says Steven Irlbeck, assistant professor of finance at the University of New Hampshire. “We definitely don’t want to do that. What we can do is kind of aggregate. We look at devices that are very commonly in the different SEC offices, the regional offices during the workday.” Irlbeck and his colleagues discovered from the data that many visits made by whom they believe to be SEC employees occur outside formal investigations. In all, 84% of visits are made to firms that never fall under investigation. But they also found that firms that have a history of actions against them by the SEC are visited more frequently by the commission. Of those who were under SEC investigation during the period the paper studied, three-quarters saw an SEC visit before the investigation was formally announced. Around 14% of visits are cross-regional, meaning SEC devices from one office visit firms in other regions. Any knock on the door is not good news—whether an investigation is opened or not. “On average, we show that these visits do predict future stock declines,” says Irlbeck. Within three months, firms visited by the SEC see stock price drops of between 1.4 and 1.94%, which the researchers posit could be because traders pick up on informal signals that a firm is under the regulatory microscope, whether through tip-offs, following similar geolocation data, or because the company changes its behavior to show it’s complying with investigatory demands. Despite the price drop, Irlbeck and his colleagues find the temptation of insider trading is outweighed by the potential threat of criminal enforcement action against such behavior. Insiders are 16% less likely to sell stock in the week before and week after an SEC visit than they are in periods with no visits. However, those that take the risk and get away with it are rewarded: Insiders who sell up following a visit, but before a formal investigation, avoid losses of around 5%. “It’s very difficult for us to know in any one given circumstance if anything nefarious is actually going on,” says Irlbeck. “But on average, what we found was quite surprising. When we do see trading following these visits, the figures are worth 1,000 words.”
Business Economic Syndicated News 
Investigations into companies by the Securities and Exchange Commission (SEC) can be market-moving events. And for those looking for an early tip as to the agency’s movement, it’s possible to gain some insight into which companies might soon fall under investigation by tracking the data of smartphones that spend a lot of time in SEC offices during the day.
That’s the finding of a new paper by academics at four U.S. universities, which tracks geolocation data across 26 major U.S. metro areas. The academics identify what they believe could be SEC devices based on repeated visits to SEC offices during working hours, then monitor the devices as they move around—including visiting the headquarters of companies during working hours. This method allows the researchers to capture interactions between the SEC and firms that may not be observable through traditional datasets.
“We cannot observe any one particular individual,” says Steven Irlbeck, assistant professor of finance at the University of New Hampshire. “We definitely don’t want to do that. What we can do is kind of aggregate. We look at devices that are very commonly in the different SEC offices, the regional offices during the workday.”
Irlbeck and his colleagues discovered from the data that many visits made by whom they believe to be SEC employees occur outside formal investigations. In all, 84% of visits are made to firms that never fall under investigation.
But they also found that firms that have a history of actions against them by the SEC are visited more frequently by the commission. Of those who were under SEC investigation during the period the paper studied, three-quarters saw an SEC visit before the investigation was formally announced. Around 14% of visits are cross-regional, meaning SEC devices from one office visit firms in other regions.
Any knock on the door is not good news—whether an investigation is opened or not. “On average, we show that these visits do predict future stock declines,” says Irlbeck. Within three months, firms visited by the SEC see stock price drops of between 1.4 and 1.94%, which the researchers posit could be because traders pick up on informal signals that a firm is under the regulatory microscope, whether through tip-offs, following similar geolocation data, or because the company changes its behavior to show it’s complying with investigatory demands.
Despite the price drop, Irlbeck and his colleagues find the temptation of insider trading is outweighed by the potential threat of criminal enforcement action against such behavior. Insiders are 16% less likely to sell stock in the week before and week after an SEC visit than they are in periods with no visits. However, those that take the risk and get away with it are rewarded: Insiders who sell up following a visit, but before a formal investigation, avoid losses of around 5%.
“It’s very difficult for us to know in any one given circumstance if anything nefarious is actually going on,” says Irlbeck. “But on average, what we found was quite surprising. When we do see trading following these visits, the figures are worth 1,000 words.”
