Mining data on cab rides to show how business information flows

AS COMPUTING power has grown, it has become easier to uncover information hidden inside datasets that seem totally unconnected. Some recent studies have used this approach to reveal business-related information flows. One linked the movements of 18th-century share prices with the arrival of ships bringing news. Another looked at the relationship between business activity and the movements of corporate jets. A third mined White House visitor logs for the names of executives and examined their companies’ subsequent stockmarket returns.

A paper in this vein published on March 5th pores over a dataset released by New York City’s government covering more than 1bn cab rides between 2009 and 2014. David Finer, a graduate student at University of Chicago’s Booth School of Business, analysed trips connecting the headquarters of big banks and the Federal Reserve Bank of New York. He extracted trips starting at commercial banks and at the New York Fed that converged on the same destination around lunchtime, and those directly from banks to the New York Fed late in the evening.

The number of such journeys rose sharply around the dates of meetings when interest rates were determined by the Federal Reserve’s monetary-policy committee in Washington. (The New York Fed plays an outsized role in setting and administering American monetary policy. Its president is a permanent member of the Fed’s rate-setting committee and sits over the trading desk that puts policy into effect.) The jump in journeys was especially marked in 2012, when the committee decided to extend quantitative easing, the purchase of securities with newly-created money. The policy had a profound impact on financial markets. There was even a noticeable change in the data that matched the timing of an office relocation by Goldman Sachs.

Mr Finer builds on a provocative paper by Anna Cieslak of Duke University, and Adair Morse and Annette Vissing-Jorgensen of the University of California at Berkeley. They asserted that information on monetary policy could be used to profit from stockmarket movements, and that such information had leaked from the Fed. Mr Finer’s “assumptions are flawed and misleading”, the New York Fed responded. “It is simply not credible to imply that an increase of a few taxi rides by unknown passengers between densely populated areas of the city—business, transportation and hospitality hubs—increased the risk of inappropriate communication.”

The data do have obvious shortcomings. They do not show who was in the taxis, and departure and arrival points are accurate only to within 100 feet. Even if private bankers and New York Fed staff did meet and discuss policy, they may have broken no law. Many of the journeys were outside the “blackout” periods during which communication between Fed officials and bankers is strictly forbidden.

But the tortuous way the Fed’s policymakers release information, through an initial announcement, then weeks later the release of minutes and years later transcripts, means that a direct meeting with its officials can be extremely useful. The Fed has acknowledged that merely having a discussion can lead to accidental disclosure. And even lawful private discussions that transfer government information of value raise questions about fairness.