Why EIA's Weekly Petroleum Product Demand Trends Are Flawed

|
Includes: BNO, DBO, DNO, DTO, DWTI, OIL, OLEM, OLO, SCO, SZO, UCO, UGA, USL, USO, UWTI
by: Robert Boslego

Summary

Demand is a function of product exports.

Product exports are estimated based on an incorrect method.

Lagged data provide evidence of the flaw.

Monthly data show demand declined in November instead of rising.

It is well known that the Energy Information Administration's (NYSEMKT:EIA) Weekly Petroleum Status Report's (WPSR) data are just estimates. Yet, they move the market because they are the best recent data available.

Seasoned observers do not put too much weight on a single week's data, but rather look at the trends in the data to gauge supply and demand fundamentals. I have written previously about why even the trends in weekly crude production and crude oil export data are flawed. Now, I conclude that the trends in weekly petroleum product demand data, reported as "product supplied" (line 26 of the WPSR), are flawed.

Taking a recent example, the demand trend based on data from the WPSR for November 2015 implied a year-over-year gain of 2.7%. With release of the EIA's more accurate Petroleum Supply Monthly data, the data show a decline of 0.9% from November 2014.

This article is on the technical side, but the conclusions are important, in my mind, in thinking about how to interpret the weekly data. For those interested in the technicalities, here they are.

Calculating Demand

The EIA reports "product supplied" as its proxy for demand. It is important to keep in mind that this is not equivalent to an estimate of end-user consumption for the week. Rather, is a "disappearance" from primary storage. This does not include inventories at gasoline stations, for example.

The equation the EIA uses (with reference to the lines in the WPSR) is:

Total Product Supplied = Crude Oil Input to Refineries (line 14) + Other Supply Production (line 15) + Net Product Imports (line 21) - Stock Change (line 24) + Adjustment (line 25).

Net product imports (line 21) includes finished petroleum products, unfinished oils, gasoline blending components, fuel ethanol, and NGPLs and LRGs.

My reason for signaling out net product imports is due to the rise in net product exports, combined with the highly flawed EIA product export estimation procedure. Although the crude oil export ban has just recently been lifted, the oil industry has been able to export petroleum products. In recent years, these exports have become sizable, averaging around four million barrels per day (mmbd), which is roughly equal to 20% of U.S. consumption.

Official U.S. exports statistics for petroleum products are compiled by the U.S. Bureau of the Census and are published monthly. The EIA obtains this data approximately six weeks after the close of the reporting month. Weekly estimates of exports for petroleum products are forecast using an autoregressive, integrated, moving average (ARIMA) procedure. The weekly estimate is updated when a new monthly estimate is calculated.

The ARIMA procedure models a value as a linear combination of its own past values and present and past values of other related time series. The most recent five years of past data are used to obtain the exports forecast.

This type of model will not produce reliable results because exports depend on price differentials among destinations. Although the U.S. exports a large share of its petroleum products to Mexico and Canada, products are exported worldwide. An ARIMA model, without price differentials, cannot capture the dynamics of world oil trade.

Example

When the EIA initially reported its weekly statistics for November 2015, total product demand averaged 19.714 million barrels per day (mmbd), based in part on petroleum product exports of 2.279 mmbd. More recently, when the EIA released its PSM, it reported product exports of 2.743 mmbd, 0.464 mmbd higher than the WPSR figure. As a result, it revised down its demand estimate by 0.525 mmbd to 19.189 mmbd. The upward revision in product exports accounted for 88% of the downward revision in demand.

Petroleum Products WPSR PSM
(mmbd) Nov. 2015 Nov. 2015 Difference
Demand 19.714 19.189 .525
Exports -2.279 -2.743 .464
Click to enlarge

Because November 2014 demand had been reported at 19.37 mmbd, the WPSR estimate indicate a demand gain of 2.7%. Instead, when the demand estimate was reduced in the PSM, the new estimate indicated a decline in demand of 0.9%.

A look back at demand estimates in earlier months reveals that the demand trend in 2015 had actually begun to decline in September. That trend was not evident in the WPSR numbers.

Conclusions

Exports of petroleum products have become sizable in relation to domestic demand. Because they impact the derivation of domestic demand, the accuracy of their estimate is important in assessing demand where a few percentage points matter.

The EIA ARIMA procedure to estimate exports is inherently flawed. It lead to a conclusion that demand was still rising from September through November, when demand had actually begun declining year over year.

Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.