ADVERTISEMENT

Divergent Data Can Make for a Whiplash World for Oil Traders

Divergent Data Can Make for a Whiplash World for Oil Traders

(Bloomberg) -- About 1.5 million barrels of oil. That was the difference this week between the change in U.S. stockpiles suggested by an industry report and data from the federal government.

Think that’s a lot? It’s actually the closest the American Petroleum Institute and the U.S. Energy Information Administration have been in their reports in almost two months. On Dec. 27, the API data implied a 4.46 million barrel decline while the EIA showed a 7,000-barrel build the very next day.

Welcome to the whiplash world inhabited by America’s oil traders, who depend on the reports that arrive just a day apart to determine the price they pay. On Dec. 27, the price of oil immediately dropped 1.3 percent in New York after the API data hit. It quickly jumped 2.6 percent after the next day’s EIA report.

“The EIA is from Mars and API is from Venus," lamented Phil Flynn, a senior market analyst at Price Futures Group in Chicago. Nonetheless, he said, “you can’t ignore the data from API. It can move the market."

Divergent Data Can Make for a Whiplash World for Oil Traders

For attentive oil watchers, it’s no surprise. Over the past year, the weekly change implied by the two influential reports has differed by 3.4 million barrels on average, according to a Bloomberg analysis. About a third of the time, the two estimates moved in opposite directions.

The industry-sponsored API releases its data at 4:30 p.m. New York time on Tuesday. EIA figures are released to the general public on the agency’s website at 10:30 a.m. on Wednesday.

Judging the API based on weekly changes isn’t fair as its report doesn’t estimate that number, said Hazem Arafa, the group’s director of statistics. Instead, the API, like the U.S. EIA, measures the total stockpile number each week. By that measure, its figure is typically within a percentage point of the government’s, Afara said.

Small differences in methodology, as well as “random statistical noise," can make the estimates different, he added.

To contact the reporter on this story: Alex Nussbaum in New York at anussbaum1@bloomberg.net

To contact the editors responsible for this story: Simon Casey at scasey4@bloomberg.net, Reg Gale, Will Wade

©2019 Bloomberg L.P.