London 26 August 2011: Reuters

The weakness of the gasoline market in the United States has been hidden, at least in the short term, by mounting exports, which have distorted weekly demand estimates. Between January and May, the Energy Information Administration's weekly "product supplied" figure -- a proxy for wholesale demand -- has overstated U.S. gasoline consumption by an average of 279,000 bpd when compared to the more detailed monthly data that EIA produces.

As a result, the agency moved this month to adjust its methodology for estimating exports in a bid to reduce the error rate in its weekly consumption estimates. But if the weekly data for recent months is off by an amount similar to that seen in the first five months of the year, U.S. gasoline consumption may be heading for its weakest summer in a decade.

Demand will be below the 2010 level even with no downward revisions for June, July and August, which must be the most optimistic scenario.

In the past, exports have proved problematic for the EIA, which does not collect its own data. But market watchers should be used to the problem, which has been around for years. This time around, weekly estimates have miscounted rising gasoline exports to Mexico and elsewhere in Latin America as domestic consumption.

EIA officials acknowledge the difficulty they have producing weekly consumption estimates due to a lack of hard data on oil exports. For now, the agency relies on other, less speedy sources, such as the U.S. Census, for export data, which is used to produce its more definitive monthly oil data sets, but these lag considerably.

June monthly data will not be available until the end of August. By the time we can look at August gasoline demand the focus of the market will be on winter. Collecting weekly export data would help, EIA analysts acknowledge, but may be beyond its means as spending cutbacks loom in Washington.

For now, market participants should be aware of the risks of accepting at face value the EIA's estimate of demand on a week-by-week basis.

DEMAND DESTRUCTION

The more interesting question is to look at what is going on in the U.S. gasoline market this year. Certainly it seems that if it is a U.S. economic slowdown depressing gasoline demand, it started earlier and is more pronounced than optimistic commentators have assumed.

Over the first five months of the year, only February showed higher gasoline demand in 2011 than 2010 based on the monthly EIA data. Moreover, both April and May showed substantial shortfalls.

Adjusting the four-week moving average on the weekly data from the end of June and July and the most recent reading for August down by the error rate seen this year suggests the U.S. may have gone through the bulk of the summer burning less than 9 million barrels per day of gasoline.

Even if the weekly overestimate is not as high as it was earlier in the month, U.S. gasoline demand will almost certainly come in below 2010 levels for June, July and August, given the weakness of the weekly readings.

This demand weakness seems to reflect the wider struggles of the U.S. economy. U.S. GDP growth in the first quarter was revised down sharply to a mere 0.4 percent, in part due to a revision in the way petroleum imports are accounted for.

Early indications for second quarter growth are also anemic. So gasoline demand destruction is well and truly underway in the United States. Whether it is permanent or temporary seems an academic dispute.

What is clear is that demand for gasoline in the world's largest consumer of the fuel is presently going down and may well weaken substantially more if a full blown recession occurs.

Ends --

By Robert Campbel, Reuters market analyst – for Commodities Now.

The views expressed here are his own.