Crude Oil Inventory Build a "correction" by EIA

[dan_s]

As I have posted here many times, I have little faith in what EIA reports each week. Their production and inventory numbers are primarily based on formulas and over the years have been poor indications of what is really happening. The speculators who set the near-term oil price over-react to their weekly reports. - Dan

From FuelFix:

Yesterday’s weekly EIA inventory report showed a 14.4 million-barrel build to oil inventories, the largest since weekly records began in 1982.

Here at ClipperData, we use U.S. Customs bills to report oil imports in a timely fashion. Imports play a key role in the weekly inventory data, making a major contribution to supply, and unsurprisingly, have a big influence on inventory numbers.

Comparing our numbers to those reported by the EIA shows their import data lagging behind the totals reported for the weeks ending October 14 and October 21, and then playing catch-up in the last week, the one ending October 28.

In the two weeks prior, the EIA reported import numbers into the U.S. Gulf Coast that were significantly below the Customs totals. For the week ending October 14, EIA imports into PADD 3 were 550,000 barrels per day below the Customs number, and 116,000 bpd below for the week ending October 21.

For the week ending October 28, there was a massive correction in the data, and the EIA reported imports that were 522,000 bpd higher than Customs data.

Not only do we believe that the EIA missed import volume in the weeks ending October 14 and October 21, but the massive stock build in the week ending October 28 suggests that there was some pent-up stock building that was not reported for at least one week if not two, and that led to the high October 28 number.

Taking a look at the stock change over a three-week period yields a similar result to that of EIA’s PADD3 imports: they are in line with our numbers. The EIA’s total stock change over the last three weeks is an 8.6mn bbl build, in line with our projection of 8.5mn bbls.

This is an unusual time shift in the EIA data. Usually the agency is much more in line with the numbers reported by Customs, and this looks like a one-off occurrence.