Sources and Techniques Used in the Construction of Annual GDP, 1790-Present

by

Samuel H. Williamson
Professor of Economics Emeritus
Miami University
and
President of MeasuringWorth

The computation of annual GDP in the U.S. has two parts, before 1929 and since that year. The earlier data are not as reliable as those for the last 83 years. The reason is simple: GDP data were not collected or even defined before the 1930s and thus any measures for years before 1932 rely on sources that were not collected for the purpose of constructing national income and product accounts.

BACKGROUND

In 1932, the U.S. Senate directed the Department of Commerce to "report estimates of the total national income of the United States." The work was assigned to the Bureau of Foreign and Domestic Commerce; but, by the end of 1932, arrangements were made with the National Bureau of Economic Research (NBER) for Simon Kuznets to assume direction of the study and he and his team completed the report in late 1933 and it was published in U.S. Congress (1934). This report was where the concept was perfected into its current form and Kuznets received the Nobel Prize in Economic Science in 1971 in part for this achievement.

The Commerce Department starting in 1942 began using a methodology that differed from Kuznets in its treatment of government expenditures. Kuznets felt that the activities of the government are an intermediate service and should not be included in final output. Commerce argued that all final production and spending must be counted; and that by subsuming government spending as an intermediate good, one missed an important sector of economic activity in a modern economy.

According the their count, the Bureau of Economic Analysis (BEA), which publishes these numbers for the Department of Commerce, has revised its annual series 14 times. For a good discussion of the history of the development of GDP in the period after Kuznets report, see U.S. National Income and Product Statistics Born of the Great Depression and World War II by Rosemary D. Marcuss and Richard E. Kane. At the international level a System of National Accounts (SNA) was first adopted in 1953 and has been revised five times since then, with the latest taking place in 2008.

BEFORE 1929

There have been several studies that have made estimates of US GDP in the pre 1929 period. Kuznets and his student Robert Gallman were the most important researchers. Others include John W. Kendrick, Stanley Lebergott, Thomas S. Berry, Christina D. Romer, Nathan S. Balke and Robert J. Gordon, Thomas Weiss, Paul David, John J. McCusker, and Richard Sutch.

In 2006, we published our series for US GDP from 1790 to 1929 linked to the BEA series to the current year. It was first on EH.Net and then MeasuringWorth when the EH.Net data sets were moved there. Since then, that series was updated whenever the Bureau of Economic Analysis (BEA) has revised its annual series. This series is the basis the current observation up to 1929. See: Sources and Techniques Used in the Construction of Annual GDP, 1790-1928.

LATEST REVISION

In July of 2013, the new revision of annual GDP from 1929 to 2012 data publishedby the BEA was announced. Many of the earlier revisions did not changed the nominal GDP for the pre WWII period. This time is different, with the 1929 observation being one percent greater and increasing to more than three percent by the end of the period. In all the revisions, real GDP has changed because of a new base year, but when corrected for that, often the earlier years have stayed the same. Again that is not true this time, as real GDP for 1929 is about .6% less than the old series and ends at 4.7% higher in 2012. This means the annualized growth rate for the 83 years will be slightly higher at 3.29% instead of 3.22%.

In July 2018, the BEA published a new version of the 1929 to present series. There were minor changes and the base year was changed to 2012.

There are many technical that caused these revisions in the GDP series. One change that BEA made that impacted all the years' observations is the capitalization of research and development expenditures. Before this revision, expenditures for purchased R&D are classified as intermediate inputs, and were included with the other costs of production and were not identified as contributing to output separately. Now, these expenditures are counted as fixed investment and the depreciation of these assets in consumption of fixed capital. This includes a new asset category of "intellectual property products" for entertainment, literary, and artistic originals. The interested reader can learn about by consulting this article in the Survey of Current Business.

The film Gone with the Wind was made in 1939 for a cost of about four million dollars. Under the old rules, the value of this movie was counted in GDP as its revenue in that year. Under the new rules, that cost is thought of as an investment, the same as if it were spent on a new factory. Now that four million was an investment that over the years has created revenue of 50 times its cost. Then there is Heaven's Gate (1980), which cost $44 million to make and had revenue of about $3.5 million, a very poor investment.

2014 DATA UPDATE

In March 2015, the BEA issued its annual revision of the GDP series. These updates are not the same as a revsion, as there are no, as they say, "major conceptual changes."

LINKING

The new series link the latest BEA series to the Williamson series using the 1929 values of both. Besides new nominal and real GDP, the current series incorporates small changed in population since 2000.

IMPACT ON GROWTH RATES

Users of the Annualized Growth Rates and Relative Values -- US $ calculators will have the same answers they had before if they confine their request to the period between 1790 and 1929; the revised BEA series has no effect. For any period beginning prior to 1929 and ending after that year, and requestsinitiating after 1929,will have different results.

Please read our Note on Data Revisions.