Is the world awash in crude, or is supply tight enough to prop up high oil prices? During the past few months, even OPEC can't seem to decide if it should increase production. These questions hinge on knowing exactly how many barrels are being produced and exported, or are in storage. But confusion reigns about true global oil production, deliveries and supplies. During the bear market in oil a few years ago, the financial community and industry itself virtually ceased attempts at original analysis, says George Littell, principal with Groppe, Long & Littell, a Houston energy consulting firm, in a recent report. The void was filled by the International Energy Agency (IEA) in Paris and its Oil Market Report (OMR). But within reports on oil supply and demand from IEA-or from any other entity, for that matter-lurk many variables that analysts would be unwise to ignore. Processing gains at the refinery often contribute to over- or understatements of volumes. People confuse demand with physical deliveries. Finally, differences in the way countries account for their oil supply/demand (trade) balances should be noted. Because of data problems, both the U.S. Energy Information Administration (EIA) and the IEA supplement their monthly or quarterly reports with an annual report that "cleans up" the data to correct errors of classification and omission. EIA distinguishes between crude oil and refinery feedstocks in all its imports. IEA does so in its annual "Oil Information Report," but not quarterly. Accuracy requires figuring a trade balance-tracing the physical flow of oil. It is straightforward for statisticians of importing countries, who have no ax to grind, to determine where a cargo of crude came from. The origin of refined products or gas liquids is only slightly more complex. After a vessel docks, any shipbroker can find out where its voyage began. Confusing demand with deliveries is the first sign of trouble, says Littell. Demand is the relationship between consumption and price. It can be determined by how much oil is consumed at different prices. After the fact, the statistical system measures deliveries because statistics are collected from refiners and marketers rather than consumers. Oil consumption is not measured by any statistical system. If consumers are not doing anything unusual with their tanks-they are a little over half full on average-deliveries and consumption should be about the same. Processing gains Refineries convert crude oil, condensate and gas liquids into gasoline, jet fuel, diesel, residual fuel oil, lubricants, wax, asphalt and coke. These processes do not affect the weight of the hydrocarbons refined, however, these products are less dense than the crude oil input. As a result, the volume of output is 1% to 6% greater than the volume of input, depending on the processes used. This difference is called processing gain. A signature of the IEA is the low processing gain it reports of 1.7- to 1.8 million barrels per day. With an incorrect processing gain, it is not possible to achieve an accurate oil balance by volume. To compensate, IEA overstates production, understates deliveries or confesses error with a "balancing item." In practice it does all three. Further confounding the issue: production is made up of oil, condensate and natural gas liquids. Thus, it is easy to overstate production unless attention is paid to these categories. IEA makes little effort to do so. Industry sources tend to blur the distinctions. The IEA always finds more oil being produced than anyone else, so readers who accept its data anticipated an imminent price collapse throughout 2003. During 2003, it said, the oil industry accumulated stocks of about 290 million barrels. The price of West Texas Intermediate (WTI) was $31.25 per barrel when 2003 began and $32.46 when it ended. It is hard to imagine how that could be so while 290 million barrels were added to stocks. A better story is that there was almost no change in total stocks in 2003. U.S., Canada and Mexico Domestic oil statistics are collected by the EIA. They are reported by volume (barrels) rather than by weight. EIA includes several line items in its data under the category "other." "Other deliveries" are unfinished oils and gasoline blending components. Imbalances result from trade between refineries and are offset by errors in deliveries of products (one refinery's production of residual fuel is another refinery's input of feedstock.) EIA's "other production" is a mixture of ethanol, methanol and MTBE. Including them in the oil balance creates confusion. EIA's "unaccounted for" results from the difference between crude as produced and as received by refineries, which contains butanes, natural gasoline and condensate transported in pipelines. Normally this is a positive number. The negative numbers reported in 2003 are a clue that EIA's estimates of oil production were overstated. Oil statistics in Canada, like those of the U.S., are reported by volume. But Canada is the world's largest producer of bitumen and synthetic crude, confounding the data Mexico reports statistics monthly to the IEA in Paris by metric tons rather than volume. Pemex routinely overstates its production. Understating product imports and overstating exports disguises this. Nearly 90% of crude oil exports are "Maya" heavy crude (22° API gravity) but Pemex's refineries cannot use much Maya. As production of other crude oil declines, there will be either an increase in importing products, or Mexico will have to begin importing crude. The latter will likely force an end to its overstatements of production. OECD and OPEC statistics Reported refinery input for European members of the Organization for Economic Cooperation and Development (OECD) consistently exceeds refinery output. Statistics for much of the area continue to be suspect. Statistics from the Organization of Petroleum Exporting Countries (OPEC) are also unreliable. The motives are either to conceal fraud or preserve the option to commit it. The problem is not confined to data for production and consumption of oil. It extends to reserve estimates and even population -the other numbers that might affect the members' quotas for crude oil. This was adapted from "Oil Statistics, April 2004," by George S. Littell of analytical and forecasting firm Groppe, Long & Littell in Houston. See groppelong.com.