US Production May Be Significantly Less Than EIA Estimate: Bloomberg, Reader -- June 4, 2020

Wow, this is really, really cool.

After the EIA weekly petroleum report was released yesterday, a reader who follows this very, very closely, noted:

I haven't even looked at the rest of the report, but the first thing I noticed is that a million barrels of oil per day went missing for the third week in a row, ie, production + imports + storage withdrawal has been 1 mbpd greater than refinery use + exports + the SPR addition... 

Best guess is that their production number has been wrong...

You can see my reply at this post.

Overnight, Bloomberg posted an article saying the very same thing: oil traders are asking why US inventory math doesn't add up. 

Oil traders and analysts scrutinizing U.S. inventory data for signs of a market recovery are being confronted by an odd situation: the math just doesn’t add up.

Various government data sets including stockpiles, production, imports and exports are signaling that current official figures on at least some supplies are excessive.

The excess is showing up in the U.S. Energy Information Administration’s so-called crude supply adjustment factor -- the difference between stockpile numbers and those implied by production, refinery demand, imports and exports.

That has averaged negative 980,000 barrels daily over the past four weeks -- the largest in records going back to 2001, and equivalent to more than 27 million barrels.

That's nearly exactly what the reader reported: a one-million-bopd discrepancy. The Bloomberg article continues:

The adjustment factor tends to swing back and forth ...

Some investors lay the blame for the current discrepancy on U.S. oil production numbers. While daily output fell 700,000 barrels to 11.2 million in May, they believe oil’s plunge into negative territory in April should have led to a steeper decline.

Just last month, consultancy IHS Markit said that U.S. oil producers are in the process of curtailing 1.75 million barrels a day of existing output by early June...

The EIA pushes back:

While output may be a factor, it’s unlikely the full answer. According to the EIA’s Robert Merriam, the accuracy of its production modeling compared to subsequent monthly data has been good, often within 1-2% in most months.

... the EIA’s weekly data has recorded a production decline of 1.9 million barrels, which he said was substantial compared with historical numbers.

He said that there have also some large weekly swings in reported inventory levels and refinery runs, so all the elements within the crude data are moving around far more than they usually do, adding that “the timing of reporting those could also be driving the adjustment lately, as they always do.”

A year ago, the crude adjustment factor caught the attention of energy enthusiasts when the figure was more than 800,000 barrels a day for four weeks, implying the reverse -- that something in the data was undercounted.

The EIA, at the time, had suspected that besides understating oil production, one of the reasons was plant condensate that was associated with natural gas output, but blended into the crude oil stream.

Bottom line: there is evidence that US production has declined further than the official estimates. That's interesting because it seems to me that production numbers from Russia and Saudi Arabia are coming in much lower than I had anticipated. [I'm probably the only one that holds that opinion but the production cuts in those two countries seem to be quite remarkable.]

Related: over at twitter after the weekly EIA petroleum report was released --