Chairman PROXMIRE. Mr. Hodges, you have a somewhat longer prepared statement than the other witnesses. I would appreciate it if you could limit the oral presentation to about 10 minutes and the entire prepared statement, which is a very useful and helpful statement, will be printed in full in the record, including the exhibits, at the end of your oral statement.

Mr. HODGES. And there may be other materials, Mr. Chairman, you might like to have that we will provide.

Chairman PROXMIRE. Very good.

Mr. HODGES. With your permission, my prepared statement, I assume, has been accepted for the record?

Chairman PROXMIRE. Yes, sir.

STATEMENT OF JOHN E. HODGES, DIRECTOR OF STATISTICS, AMERICAN PETROLEUM INSTITUTE

Mr. HODGES. I am director of statistics at API and the record shows I have had extensive experience both on the academic side and in industry as a collector of primary data and as a user of data pertaining to the petroleum industry.

Indeed, I have been involved with these matters since 1944, long before I came to the API.

Because of the current interest in this problem, I am pleased to review what the API is doing in this area. But before describing the program, I would like to emphasize that the API does not publish forecasts or make projections of any kind and we do not collect statistics on prices.

The first major element of our program is the weekly statistical bulletin.

This bulletin provides data pertaining to petroleum supply and refinery operations. In recognition of the importance of the time factor, the bulletin is usually published each Wednesday on the basis of information reported directly to us by the companies as of the previous Friday.

In order to put the bulletin in proper perspective it should be noted that the U.S. Bureau of Mines publishes monthly reports on the domestic supply system that are much more comprehensive than the API weekly reports.

However, their monthly reports run at least 2 months behind in publication. Nevertheless, the industry and I believe many Federal Government agencies consider the Bureau of Mines to be the definitive source of information of the types reported in their monthly petroleum statement.

It should be emphasized that the primary purpose of the weekly statistical bulletin is to provide the most timely information that can be collected covering selected operating series pending release of the definitive data by the Bureau of Mines.

Now we do not get as good a coverage as the Bureau of Mines and it has been customary for the API to introduce a factor to take our 90-odd-percent coverage and make an estimate for the total industry. Exhibits A and B attached to my prepared statement show the percentages of volumes reported to the API against the Bureau of

Mines, and also shows a comparison of the accuracy of the data published by the API in relation to the Bureau of Mines.

We believe that this record shows that we have done an extraordi

narily good job in this area.

The second report that we concentrate on is that pertaining to reserves. The API has been involved in this in an organized fashion since 1935.

In 1966, in response to a petroleum statistics report sponsored by the Bureau of the Budget, the work of this committee that prepares the reserves estimates was greatly expanded and also expanded to include estimates of productive capacity.

Now, there is a common misconception about API reserve estimates and how they are derived. I believe many people think that we take information submitted directly by the individual companies. In fact, estimates published by the API are prepared by highly qualified professional personnel who prepare individual and independent estimates of what proved reserves may be in their respective areas. The details of the organization of the API committees that actually do the work, and the procedures and definitions are set forth in exhibit C attached to my prepared statement.

The third major project in the division pertains to drilling costs and expenditures. In cooperation with the Mid-Continent Oil & Gas Association and the Independent Petroleum Association of America, API compiles and publishes data pertaining to the estimated cost of drilling oil wells, gas wells, and dry holes, by depth ranges for various areas of the United States.

The second part consists of an estimate of expenditures for oil and gas exploration, development, and production in the United States.

These data are published in what we call the joint association survey.

Now, there are no annual sources of cost and expenditure information comparable to that published in the joint association survey. However, the "Census of Mineral Industries" is now available every 5 years previously at 4-year intervals. This provides benchmark data against which the results in the JAS can be compared for accu

racy.

We have made a very careful study of the two documents for the years 1963 and 1967 and we are satisfied that we have come up with results that in our opinion are better than what Census has done.

The Federal Power Commission has also collected voluminous data of a similar nature. The Chase Manhattan Bank collects data and prepares estimates of total expenditures. The point is there are a lot of cross-checks to the data we are publishing.

The fourth area pertains to drilling statistics. Prior to 1966, the basic sources of statistics pertaining to the drilling of oil and gas wells were the trade journals. The States published drilling statistics, the American Association of Petroleum Geologists published drilling statistics. There were also commercial scouting associations. However, none of these agencies were covering the United States uniformly and they were not using standard definitions.

According to the Kruger report, the study I referred to a moment ago, there was no single, timely, uniform, accurate, and complete set of data. In response to that report the API worked out an arrangement with the American Association of Petroleum Geologists whereby we are getting a ticket filled out on each well drilled and completed in the United States. A copy comes to the API for tabulation. In the meanwhile a copy passes to the American Association of Petroleum Geologists Committee on Statistics of Drilling. They review this in the light of their professional judgment as to how the well is classified and their judgment prevails when the final tabulations are made.

Again there are sources of data against which these API publications can be cross-checked. The "Census of Mineral Industries" every 5 years, the trade journals still publish data. There are also scouting associations and commercial companies that collect a great deal of drilling information for the United States.

Now beyond these basic elements of the program, a major function of the API for the past 7 years has been to establish standard definitions which are crucially important to the development of useful and compatible statistical systems. I can say for certain that the Bureau of Mines definitions and API definitions are uniform. The definitions used by Census are uniform with respect to what the API uses. We also publish petroleum facts and figures in recognition of need for a compendium of all data pertaining to the petroleum industry. This is simply an encyclopedia.

On an annual basis we publish the annual statistical review, which is primarily a summarization of data published by the Bureau of Mines. The purpose is to put it in a more convenient reference form. Very soon we anticipate expanding the annual statistical review to include other series. In effect it will be come an annual supplement to "Petroleum Facts and Figures."

Mr. Chairman, this completes my statement but I would like to make several observations which I believe are related to these hearings.

First, contrary to certain opinions, there is a voluminous amount of data available with respect to the petroleum industry. The API statistical program provides only a small fraction of the total data available. I am fully in agreement with the panelist a moment ago who said we have so dang much of it it is hard to correlate; it needs organization; definitions are different. We have a problem.

Second, I am convinced that much of the current controversy concerning the availability and credibility of petroleum statistics stems from confusion as to the difference between statistics on the one hand and projections on the other. This is reflected in the confusion of the public as to the degree of the energy shortage. What the public really wants to know is what the situation will be next February, next March, what the situation will be in August, with respect to gasoline supplies. The real measure of the current adequacy of stocks of gasoline and fuel oil is not so much how they compare a year ago, but how they relate to an expected summer travel demand, whether in February and March, the level of refinery runs that will be possible to support demand in view of the embargo.

Statements about the future must be based on assumptions involving numerous variables such as these. When forecasters differ in their assumptions they may arrive at different conclusions.

Finally, the allegation has been made that the API gathers statistics solely for the benefit of its larger members: This is not the case. All of the information the API collects is published in summary form. Our publications are distributed on a complimentary basis to all members of the press and government agencies without restrictions of any kind. Our publications are available to all others on a subscription basis regardless of whether the individual or the agency is a member of API.

Thank you.

Chairman PROXMIRE. Thank you, Mr. Hodges.

[The prepared statement, with exhibits, of Mr. Hodges follows:]

PREPARED STATEMENT OF JOHN E. HODGES

Mr. Chairman, I am John Hodges, Director of Statistics at the American Petroleum Institute. I joined the staff of the API in 1966. Prior to that time, I taught statistics and economics at three universities for a period of approximately twenty-five years. Concurrently, I accumulated twenty-three years of full-time industrial experience as statistician, economist, and vice president of a large oil well equipment manufacturer.

Since 1944, I have been continuously involved with problems pertaining to petroleum industry statistics. My work in private industry involved collection of primary petroleum data and the use of petroleum statistics for forecasting. My academic career (and personal interests) led to my involvement with the collection, analysis, and use of petroleum statistics, long before I became a member of the API staff.

Because of current interest in data pertaining to the petroleum industry, I welcome this opportunity to review the basic elements of the API statistical program.

The API was organized in 1919. From its inception, the API has been profoundly aware of the public's interest in the adequacy of petroleum supply.

In 1920, the API established a statistical department charged with responsibility for the collection and prompt publication of vital statistics pertaining to the petroleum industry.

The first efforts of the new department were directed toward the preparation of estimates of daily average crude oil production. From that beginning, the work of the department has been gradually expanded in well-defined but limited areas.

Before describing API statistical publications I wish to emphasize the API does not make or publish forecasts or projections of any kind, and it does not collect statistical information on prices.

The Weekly Statistical Bulletin.-The Weekly Statistical Bulletin in its present format is a direct outgrowth of the work that was started in 1920. This Bulletin provides data pertaining to the domestic petroleum industry such as weekly crude runs to crude oil distillation units, the utilization of operable refining capacity, weekly refinery production of gasoline, jet fuels, distillate fuel oil, and residual fuel oil. In addition, the Bulletin shows weekly stocks of crude oil and refined products, weekly estimates of domestic crude oil production, and weekly information pertaining to the imports of crude oil and refined products. Stocks (i.e., inventories) reported by the API and the Bureau of Mines are referred to as "primary stocks." These stocks are located at refineries, bulk terminals, and in pipelines. They do not include inventories in the hands of distributors and consumers, i.e., secondary stocks.

In recognition of the importance of the time factor, the Bulletin is usually published each Wednesday on the basis of information reported directly to the API by refiners and pipeline companies involved with supply and refinery operations. The information reported to the API each week reflects operations for a seven day period ending on the Friday preceding publication.

In order to put the Weekly Statistical Bulletin in proper perspective, it should be noted that the U.S. Bureau of Mines publishes monthly reports on the domestic supply system that are much more comprehensive than the API weekly reports. However, their monthly reports dealing with various aspects of supply, demands, movements of crude oil and products, refinery operations, and inventories, run at least two months behind in publication; e.g., as of today, the latest figures available are for the month of October. (This lag represents about the normal time required to collect and compile the voluminous statistics in the 24 tables in their "Monthly Petroleum Statement.") Nevertheless, the industry as well as the federal government, considers the Bureau of Mines to be the definitive source of information of the types reported in the Monthly Petroleum Statement.1

It should be emphasized, therefore, that the primary purpose of the API Weekly Statistical Bulletin is to provide the most timely information that can be collected covering selected operating series pending release of the definitive data from the Bureau of Mines.

The data published in the Bulletin are made possible by the voluntary cooperation of those members of the API willing to provide information pertaining to their operations. A comparison of volumes reported to the API with volumes reported to the Bureau of Mines for the first nine months of 1973 is shown in Exhibits A and B on a percentage basis.2 For example, the volumes of crude runs reported to the API during the first nine months of 1973 averaged approximately 97 percent of volumes of crude runs reported to (and published by) the Bureau of Mines.

In order to publish estimates of total industry volumes reported in the Bulletin, the API has traditionally introduced an estimate of unreported volumes. These estimates are based on a comparison of the latest figures available from the Bureau of Mines with data actually reported to the API for the same period. In other words, we are currently using adjustment factors based on a comparison of Bureau of Mines data for the month os September, 1973 with data actually reported to the API for weeks ending in September 1973. API weekly estimates are audited regularly against Bureau of Mines reports as they become available. This enables us to detect deficiencies in our estimates and to make necessary adjustments in our estimating procedures. Comparisons of the reliability of data published by the API as related to data published by the Bureau of Mines are shown in Exhibits A and B. For example, with respect to crude runs, API data published for the first nine months of 1973 were within four tenths of one per cent of crude runs published by the Bureau of Mines.

Our agreement with the companies that cooperate in this statistical program is that we will use the data supplied to arrive at and publish estimates for the industry, and that the API will not release information pertaining to individual companies operations. This understanding reflects a basic view that release of data on the operations of a company is the proper function of that company and not of a trade association.

The accuracy of weekly API data can be verified for any given month by comparison with monthly results published by the Bureau of Mines. As shown in Exhibits A and B, the weekly averages have compared exceptionally well with final data published by the Department of the Interior. Furthermore, the major companies in the industry are called upon periodically to testify before the Texas Railroad Commission as to their inventories of crude oil and major products. Those statistics are a matter of public record available to everyone, and are frequently published in trade journals. They have not been considered of general public interest in the past when supplies were adequate to meet demands and only industry totals were of concern to business publications and newspapers.

In the present period of shortages, additional interest has developed in current statistics other than those published by the API. For example, questions are being raised as to the accumulation of secondary inventories outside the hands of major refiners and pipeline companies, and trends in the actual cur

1 Definitions of all terms used by the API are the same as those used by the Bureau of Mines.

2 October results from the Bureau of Mines were not received in sufficient time to update Exhibits A and B.