A Comparison of Selected
EIA-782 Data With Other Data Sources
EIA Home > Petroleum > Petroleum Feature Articles
A Comparison of EIA-782 Data
With Other Data Sources
by Jacob Bournazian
The EIA-782 survey series collects data on petroleum markets to fulfill legislative mandates from Congress and to provide comprehensive information for evaluating market behavior. It includes three surveys: Form EIA-782A, "Refiners'/Gas Plant Operators' Monthly Petroleum Product Sales Report;" Form EIA-782B, "Resellers'/Retailers' Monthly Petroleum Product Sales Report;" and Form EIA-782C, "Monthly Report of Prime Supplier Sales of Petroleum Products Sold for Local Consumption." This article compares the data from the EIA-782 survey series with other sources to assess the quality of the EIA-782 data. Significant differences and trends among data series may indicate the need for changes in data collection and processing, the reporting population, survey or sample design, or may simply reflect conceptual differences across surveys.
The data sources used to compare with the EIA-782 series include:
This article discusses the differences among the data sources and the reasons for variation among the data series. Some differences are irreconcilable and exist among the data sources because of different reporting populations, point in time measurements of market activity, survey design, methodology, and metadata issues relating to product and energy-use sector definitions. Other factors that contribute to differences between data sources include differences in geographic and/or market coverage, product definitions, and units of measure. A more detailed description of each data source is contained in the Notes section at the end of this article.
Generally, Tables FE1 - FE3 show that EIA-782 national prices are lower than the BLS, EIA-878, and EIA-888 retail price data series. Differences in the survey methodology across the surveys explain some of the price differences.
Residential No. 2 Fuel Oil
Table FE1 shows BLS prices are 3 to 5 percent higher than EIA-782 prices from 1995 through 2001. The difference between the two series has gradually grown each year since 1999. BLS prices are obtained from urban areas only and do not reflect complete geographic coverage for this product. EIA-782 prices are volume weighted price estimates.
|Table FE1. U.S. Residential No. 2 Distillate Prices, 1995-2001|
|(Cents per gallon)||Percentage|
On-Highway Diesel Fuel
Table FE2 shows the annual estimates for EIA-782 and EIA-888 prices from 1995-2001 and for BLS prices from 1998-2001. BLS began publishing retail diesel fuel prices beginning in 1998. EIA-782 and EIA-888 prices track closely, however, the EIA-888 prices range from 9 to 23 percent below the BLS prices. No adjustment was made to the BLS and EIA-888 prices for additional State and/or local taxes relating to environmental regulations and transportation use, so both price series were expected to be higher than the EIA-782 prices.
|Table FE2. U.S. Retail On-Highway Diesel Fuel Prices, 1995-2001|
|n/a = not available|
BLS prices are obtained from urban areas and do not reflect complete geographic coverage for this product. Diesel prices are one of five types of fuel prices that are collected for the motor fuels item strata in calculating the Consumer Price Index. Outlets are selected based upon responses to the BLS Telephone Point of Purchaser Survey (TPOPS) on where the consumer purchased any motor fuels during the survey period. The companies reporting on the EIA-888 survey were selected from the EIA-782 surveys.
Table FE3 shows the annual estimates for EIA-782, EIA-878, and BLS prices from 1990-2001. The price differences were similar across each grade of gasoline so this paper only discusses the prices for regular grade gasoline. For the past 11 years, BLS prices vary between 5 and 9 percent above the EIA-782 prices. EIA-878 prices are between 1 and 5 percent higher than EIA-782 prices during the same time period. Since both BLS and EIA-878 prices contain additional taxes relating to sales taxes, highway use taxes, and other local taxes that could not be removed for this analysis, the EIA-782 prices should be lower than the other series.
|Table FE3. U.S. Retail Motor Gasoline Prices, Regular Grade, 1990-2001|
|n/a = not available|
Since BLS does not calculate an annual price, a simple average of monthly prices was calculated to obtain the annual average price. The BLS monthly prices are calculated based on approximately 900 price quotes. Approximately 25 to 35 prices are collected from each published geographic area. EIA-782 prices represent sales transactions in all 50 States. There are other limitations in comparing a simple average data series with a volume weighted average price series because of the effect of volume changes throughout the year on the annual price estimate.
The EIA-878 also uses fixed volume weights for calculating prices based on approximately 900 price quotes each week. A simple average of weekly prices was calculated to obtain the annual average price. At the national level, EIA-878 regional prices are weighted based on fixed weights for each region. The EIA-878 sample used from 1998 through 2001 is based on EIA-782 sales volumes from October 1996 through October 1997. The EIA-782 prices utilize all sales transactions throughout the reference month whereas the EIA-878 are point in time estimates.
EIA-782C volumes were compared with volumes reported in the EIA-821, the PSA, and the FHWA. Product supplied in the PSA is an estimate of the demand for petroleum products because it measures the production and adjustments from primary sources of supply for a given time period. It is calculated by adding and subtracting volumes as they are moved into and out of the primary distribution stream. Sales volume data from the EIA-782C and EIA-821 reflect the transfer of product title from a seller to a buyer into those States where the products are locally marketed and consumed. FHWA doesn't collect actual sales data on gasoline and diesel fuel. States report their fuel volumes to FHWA based on the beginning inventory at the terminal facility minus exports plus shipments to the terminal during the reporting cycle. This difference in survey concepts and methodology underlies some of the differences that exist between the data series.
Distillate Fuel Oil
Table FE4 shows volumes of distillate fuel oil from the EIA-782C, EIA-821, and PSA series from 1990 through 2001. From 1990 through 1992, the EIA-782C volumes are higher than the EIA-821 and PSA volumes. The difference between the EIA-782C series and the PSA series narrows after 1993. An important reason for this change is the addition in 1993 of several importers and traders to the exclusionary list used by respondents when reporting sales on the EIA-782C. For a description of the changes made in 1993 see Changes to Form EIA-782C, "Monthly Report of Petroleum Products Sold into States for Consumption", by Kenneth I. Platto, Petroleum Marketing Monthly, May 1993. Respondents to the EIA-782C should exclude sales to any company on the exclusionary list that is not a local distributor, local retailer, or end user. These changes improve the EIA-782C's market coverage and eliminate double counting of volumes sold.
|Table FE4. U.S. Distillate Fuel Oil Volumes, 1990-2001|
Since 1997, EIA-782C volumes have been below both EIA-821 volumes and PSA volumes. This suggests that the EIA-782C may still not have complete coverage on distillate sales. Table FE4 also shows that the difference between the EIA-782C and EIA-821 volumes is greater than the difference between the EIA-782C and the PSA volumes.
One possible source for the widening gap between EIA-821 and EIA-782C volumes is the sales coverage for on-highway use. The EIA-821 on-highway energy use sector shows an increase of approximately 16% since 1997 and is the largest component contributing to the increase in distillate volumes for that survey during the past 5 years. The EIA-821 volumes for on-highway use are obtained from the Federal Highway Administration and used in place of data reported for this category. However, some EIA-821 respondents may report sales to commercial and institutional fleet vehicles in the commercial use category. If the commercial category contains some misreported transportation volumes, and exogenous data is used to replace the data for on-highway use, then some double counting of distillate volumes for transportation use on the EIA-821 may occur. If sales to fleet vehicles, which are reported in the commercial category, are increasing at approximately the same rate as other distillate sales for other on-highway use, then the amount of double counting of distillate sales in the EIA-821 survey may also be increasing over the past five years, and may contribute to any differences between the EIA-782C and EIA-821 data series.
Table FE5 shows volumes of motor gasoline from the EIA-782C, FHWA and PSA series from 1990 through 2001.
|Table FE5. U.S. Motor Gasoline Volumes, 1990-2001|
|n/a = not available|
EIA-782C and PSA
Table FE5 shows that during 1990-1992 PSA motor gasoline volumes are between 8 - 10 percent lower than the EIA-782C volumes. One reason PSA volumes are significantly lower than EIA-782C volumes prior to 1993 is that double counting occurred on the EIA-782C because some respondents were not excluding sales to companies that should have listed on the exclusionary list. A second reason EIA-782C volumes are greater than PSA volumes prior to 1993 was that PSA did not have complete coverage of downstream blending of finished motor gasoline. Blending of fuel ethanol, methanol, methyl tertiary butyl ether (MTBE), and other blend stock with gasoline often occurs downstream from the refineries. Prior to 1993, this is included in the EIA-782C volumes but not in the PSA volumes. Since 1993, the published PSA motor gasoline volumes include downstream blending at bulk terminals. As a result, the difference between the EIA-782C and the PSA volumes narrows after 1993.
EIA-782C and FHWA
Table FE5 also shows that EIA-782C volumes generally track the FHWA motor gasoline volumes beginning in 1993. Prior to 1993, FHWA volumes were reported by wholesale distributors to State motor fuel tax agencies that compile data on gasoline taxes and these data were reported by the State agencies to FHWA. In 1993, the point of Federal tax collection was moved upstream from the last wholesale sale to the terminal operators. This change in the reporting volumes results in more accurate reporting of the FHWA volumes.
Residual Fuel Oil
Table FE6 shows volumes of residual fuel oil from the EIA-782C, EIA-821, and PSA from 1990 through 2001.
|Table FE6. U.S. Residual Fuel Oil Volumes, 1990-2001|
EIA-782C vs. PSA and EIA-821
Table FE6 shows that the difference between the EIA-782C and the other two series widen after 1993 and reach its highest level in 2000 with PSA volumes exceeding the EIA-782C volumes by 43% and the EIA-821 exceeding the EIA-782C volumes by 35%. The large and continuous divergence between the EIA-782C and the other two series suggests that the EIA-782C may be missing some coverage of residual fuel oil sales. In addition, there may have been some misreporting by respondents in the PSA surveys during this time period.
During 1999 through 2001, some importers misreported their imports of residual fuel oil on the EIA-814, "Monthly Imports Report." Their imports of unfinished oils were combined with residual fuels and both products were reported as residual fuel oil. This may be due to the misconception by respondents that imports reported to EIA should match the same import volumes reported to the U.S. Customs office. When a shipment reaches the United States, the importer of record will complete Customs Form 7501, "Entry Summary" (CF-7501). CF-7501 is used to validate data reported on Form EIA-814. CF-7501 does not have a category for unfinished oils so imports of unfinished oils are mistakenly reported as residual fuel oil. It is unclear how long the reporting of unfinished oils as residual fuel oil has been occurring during the past 10 years. The reported imports of unfinished oil as residual fuel oil by some PSA respondents contributes to the higher PSA volumes for residual fuel oil. This is notable in the past 2 years as imports of residual fuel oil reached their highest levels and account for 39 percent of product supplied for residual fuel oil in 2000 and 47 percent in 2001. A data correction for 2001 PSA residual fuel oil volumes moved some volumes from imports of residual fuel oil to imports of unfinished oils. This resulted in reducing the gap between the EIA-782C and PSA residual fuel volumes. No other data corrections were made to the 1999 and 2000 PSA volumes.
Another reason EIA-782C volumes are below the other series is that some firms may be missing from the reporting population of the EIA-782C survey. Missing firms from the EIA-782C will undercount sales and contribute to the difference between the series as the percentage of imports that comprise product supplied of residual fuel oil increases. The EIA-782C has been below both other data series for the past nine years and the gradual widening of this disparity, even after the revision of the 2001 PSA volumes, indicates that the EIA-782C is missing some coverage of residual fuel oil.
One method of evaluating the quality of petroleum market prices and volumes is to compare EIA-782 series data with other sources. Differences among data sources could lead to a review and possible update of the reporting populations for the EIA-782 surveys and research on alternatives for adjusting the data. Other differences across data sources indicate differences in survey methodology and conceptual differences with data collection.
The Form EIA-782A, "Refiners'/Gas Plant Operators' Monthly Petroleum Product Sales Report," collects monthly price and volume data at the State level for 14 petroleum products for various retail and wholesale marketing categories. It is a census of refiners and gas plant operators. The frame is updated on an ongoing basis using respondent lists from surveys such as the Form EIA-810, "Monthly Refinery Report;" the Form EIA-816, "Monthly Natural Gas Liquids Report;" and industry trade publications. Currently, 110 companies respond to the EIA-782A survey.
The Form EIA-782B, "Resellers'/Retailers' Monthly Petroleum Product Sales Report," is sent to a sample of resellers and retailers of motor gasoline, distillate, propane, and residual fuel oil. Respondents to Form EIA-863, "Petroleum Product Sales Identification Survey," are used as the sampling frame of resellers and retailers for the EIA-782B. Firms having 5 percent or more of sales in a State are selected with certainty. The remaining companies on the frame are sampled by geographic area, product, type of sale, and by probability proportional to size. The EIA-782B sample includes approximately 2,000 companies.
The Form EIA-782C, "Monthly Report of Prime Supplier Sales of Petroleum Products Sold for Local Consumption," collects volumes of prime supplier sales of selected petroleum products into States where they are locally sold and consumed. A prime supplier is a firm that produces, imports, or transports any of the selected petroleum products across State boundaries and local marketing areas and sells the product to local distributors, local retailers, or end users. This survey provides a measure of consumption in most States. Currently, 183 firms respond to the EIA-782C survey.
Data collected on the Forms EIA-782A, EIA-782B, and EIA-782C are published in the Petroleum Marketing Monthly (PMM) and the Petroleum Marketing Annual (PMA).
In addition, production, import, and export data collected by EIA's Petroleum Division are published in the Petroleum Supply Monthly (PSM) and the Petroleum Supply Annual (PSA). The Petroleum Division uses the Petroleum Supply Reporting System (PSRS) for data collection. The PSRS is composed of a family of data collection survey forms, data processing systems, and publications systems. Detailed data on refinery and natural gas plant operations, bulk terminal and pipeline stocks, petroleum products imports, and movements of petroleum products among Petroleum Administration for Defense (PAD) districts are collected monthly. Figures for product supply originate from Forms EIA-810, "Monthly Refinery Report;" EIA-811, "Monthly Bulk Terminal Report;" EIA-812, "Monthly Product Pipeline Report;" EIA-813, "Monthly Crude Oil Report;" EIA-814, "Monthly Imports Report;" EIA-816, "Monthly Natural Gas Liquids Report;" and EIA-817, "Monthly Tanker and Barge Movement Report." Aggregate export data obtained from the Bureau of the Census are also included in computations for the PSM and PSA. The PMA and the PSA may contain revisions of the data published in the PMM and the PSM, respectively, due to late submissions or revisions to the monthly data.
The Form EIA-821, "Annual Fuel Oil and Kerosene Sales Report," collects data on the sales to end users of distillate fuel oil, residual fuel oil, and kerosene. The data are used to determine the level of sales by energy-use category and product at the State, regional, and national levels. The sample size is approximately 4,000. The sampling frame for the EIA-821 is also derived from the respondents to Form EIA-863. The EIA-863 is a quadrennial census used to collect information on size, type, and geographic location of firms selling petroleum products. Data from the Federal Highway Administration (FHWA) of the U.S. Department of Transportation replace EIA-821 data reported as on-highway diesel sales.
The Highway Statistics Division of the FHWA collects information related to highway transportation. Sales volumes of motor gasoline are published on a calendar year basis and are a cumulative tabulation of gross gallons of gasoline reported by wholesale distributors to State motor fuel tax agencies. The FHWA collects information on finished motor gasoline, with no distinction made among motor gasoline grades. The data include gasoline for both highway and non-highway use. The FHWA includes gasohol but excludes exports, fuels for military use, and dealer transfers.
The Bureau of Labor Statistics (BLS) publishes the aggregate index for household fuels and its component indexes for electricity and natural gas, as well as the motor fuels. The component index for fuel oil and diesel fuel are only published at the national level. These retail prices are collected monthly by BLS representatives in the urban areas, and support the estimation of the Consumer Price Index (CPI). The CPI uses fixed volume weights to measure the change in price over time for a defined market basket of goods and services bought by urban consumers. It measures the percent change in consumers' expenditures on a fixed list of items whose values and qualities do not change over time. The base period weight of the fuels indexes for the time period evaluated represents the out-of-pocket expenditures on household fuels as reported in the Consumer Expenditure Survey (1993-1995). Approximately 2,400 prices are collected for all three grades of gasoline in approximately 87 urban areas across the country and include all taxes directly associated with the purchase and the use of the items. The 87 areas cover 39 States and the District of Columbia.