PIIE Blog | RealTime Economic Issues Watch
The Peterson Institute for International Economics is a private, nonprofit, nonpartisan
research institution devoted to the study of international economic policy. More › ›
Subscribe to RealTime Economic Issues Watch Search
RealTime Economic Issues Watch

Improving the National Accounts of the United States

by | July 10th, 2013 | 09:21 am

Every five years, the Commerce Department’s Bureau of Economic Analysis (BEA) undertakes a comprehensive revision of what are known as the National Income and Product Accounts (NIPA) in an endeavor to provide a more complete and accurate picture of the US economy. This year the revision due on July 30 is awaited with some anticipation because it will include an ambitious effort to expand the concepts of “investment” and “capital” to include spending on such activities as research and development, which should improve considerably our ability to track an increasingly knowledge-based economy. One by-product of this new analysis is likely to be that the nation’s GDP will be shown to have been bigger than it was under previous measures. There will also be an improved accounting for pension benefits and liabilities, furnishing a clearer picture of the current and evolving financial conditions of households, businesses, and governments. By themselves, the innovations to be introduced by the BEA are unlikely to change substantially the contours of the Great Recession and the subsequent pattern of slow recovery. But they should shed a brighter light on important longer-term developments in the US economy.

It has long been the custom for the BEA to fold in data from annual economic surveys every year. But the comprehensive revisions, which occur once every five years, incorporate changing data definitions, improved methodologies, and the results of the five-year benchmark input-output accounts. A key consequence of the innovations being introduced in the 2013 comprehensive revision will be to align the US national accounts more closely with the international standards embedded in the System of National Accounts (SNA)—standards that, for the most part, have been advocated and advanced by the United States. This international effort has been motivated by a goal of providing countries with a common conceptual framework for economic measurement that reflects both current economic theory and the changing nature of modern economies. Bringing national accounts into greater congruence with the SNA also facilitates international comparisons of economic performance that will be of help to researchers and policymakers both here and abroad. The United States will be taking a meaningful step forward with the implementation of these innovations.

Measuring Investment in Research and Development

Accounting for investment in intangible capital has long been supported by economists on analytical grounds. One prominent form of investment in intangible capital is spending on research and development, which is currently treated as an intermediate input in the production process for purposes of our national accounts. Intermediate inputs are those that produce current output. Examples would be the use of labor, energy, and materials in production. By contrast, research and development (R&D) spending is undertaken specifically for the purpose of raising future output, much like investment made in physical capital. Moreover, R&D spending has other attributes that make it similar to investment in fixed assets: ownership rights can be established and transferred, the results of R&D spending are long lasting, and they can be used repeatedly in the production process. The same logic led the BEA in 1999 to reclassify spending on software as investment rather than as an intermediate input. As the conceptual content of output becomes increasingly important, acknowledging and accounting for R&D spending as a form of investment, and its accumulation as part of the nation’s capital stock, will be critical in accurately tracking the health of the national economy.

Of course, in doing so the BEA has confronted the daunting task of translating the analytical concept of R&D investment into practical measurement. In this effort, the BEA will largely employ surveys conducted by the National Science Foundation of R&D spending by businesses, universities, other nonprofit research institutions (like the Peterson Institute), and governments.1 In addition, like any form of capital, R&D capital can be expected to depreciate over time, as the R&D assets become less valuable or obsolete. For example, we can assume that the R&D conducted by the Remington Company to improve its typewriters is fully depreciated at this point. The BEA is using establishment- and firm-level studies of the relationship between investment in R&D and future profits to estimate depreciation rates. Based on these studies, R&D spending in any given period is assumed to contribute to future profits at a geometrically declining rate.

Based on preliminary calculations by the BEA, classifying R&D spending as investment rather than intermediate expense boosted the level of nominal GDP by roughly $300 billion in 2007—or a bit more than 2 percent. Because the accounts are being revised back over history, the level of measured GDP will be shifted up by nearly equal percentage amounts across time so that the pattern of estimated quarterly growth rates will be little affected by the addition of R&D investment to GDP.

Measuring Intangible Investment in Artistic Endeavors

While less widely discussed, spending on entertainment, literary, and other artistic originals has many of the same characteristics as R&D spending—most importantly, today’s effort is designed to produce a flow of services into the future. (Who would dispute that Beethoven’s efforts from 1822 to 1824 to compose his Ninth Symphony should be represented as an investment in enduring “capital” that has rendered services over nearly two centuries, rather than as an intermediate input that contributed to output only from 1822 to 1824?)

 As with R&D spending, the measurement challenges are considerable. For each type of artistic “original,” the BEA must estimate capital values from the net present value of future royalties and revenues. Estimates are developed from economic censuses, trade sources, and databases maintained by websites such as the popular IMBd.com. Depreciation rates for these “capital investments” are constructed based on the pattern of declining revenues observed from the originals. (I have been amazed to turn on a TV in some far-flung location in the world only to find some long-cancelled US television show still providing “services” to the local population or to lonely hotel occupants.) Like R&D expenditures, spending on entertainment, literary, and other artistic originals will be treated as investment spending and added directly to measures of GDP. According to preliminary estimates by the BEA, this investment amounted to about $70 billion in 2007, or about .5 percentage point of nominal GDP. Again, the expectation is that incorporation of this added investment spending will shift up the level of GDP but not affect appreciably the pattern of growth that is currently recorded in the accounts.

Improved Accounting for Defined Benefit Plans

A second major conceptual advance in the US national accounts to be implemented in the upcoming comprehensive revision will be a shift in the accounting of defined benefit pension plans from a cash basis to an accrual basis. In a cash accounting system, compensation and costs are measured when payments are made and received. But under an accrual accounting system, compensation and costs are measured when benefits are earned and liabilities are incurred regardless of whether a cash payment is made. For example, at present, the BEA records employers’ cash contributions to pension plans as employee compensation and the interest earned on pension plan assets as personal interest and dividends. Underfunding or overfunding of these plans is ignored. From an accrual perspective, this presents a misleading picture. The pension benefits being accrued by employees are better thought of as compensation for the employees for this period’s labor services, whether or not the employer is actually making a cash contribution that funds the future benefit.

Likewise, failure to account for an unfunded liability incurred by an employer overstates the economic profits of the firm in that period. The distortions would run the opposite direction for overfunded pension plans. Under the new accrual accounting approach, if cash contributions fall short of the estimated changes in the liabilities of defined benefit plans, the shortfalls will be added to compensation and deducted from profits.2 The BEA will take a similar approach to under- or overfunded public pension plans. For example, shifting from cash to accrual accounting in the state and local sector, where pension plans are, on net, underfunded, will raise measured compensation of state and local workers and reduce measured state and local government surpluses.3

These accounting changes also have important corresponding implications for personal income and saving. Underfunded pension plans forgo interest income. Under the new accounting treatment, an imputed interest cost on the unfunded actuarial liability will be “paid” for purposes of the accounts by the employer to the pension plan and, through the pension plan, onto persons in the form of imputed interest and dividends. Again, the BEA has published preliminary estimates that point to some sizable upward revisions to personal income and saving, reflecting the aggregate underfunding of pension liabilities in the United States. For the years between 2000 and 2007, the estimated upward revision to the personal saving rate averages about 3 percentage points. Because the upward shift in the personal saving rate will be evident through much of the historical data, it is unlikely to have large implications for the near-term outlook for household consumption.

I should hasten to note that the shift to accrual accounting will not affect the level of national saving. Rather it will affect the composition of national saving; the increase in estimated personal saving will be offset by corresponding downward adjustments to business and government saving. But the revisions should result in a more accurate picture of trends in personal saving and wealth, allowing for better assessments of the financial wherewithal of US households. Similarly, the accounting will yield a more accurate assessment of the financial conditions of businesses and governments.

The BEA will introduce other methodological changes as part of the comprehensive revision, and, as it does every year, the BEA will fold in data that have become available since its last annual revision. All told, these additional changes could either amplify or moderate some of the revisions to GDP, incomes and saving that will result from the broader inclusion of intangible investment in the accounts and from the shift to accrual accounting for defined benefit pension plans. But whatever the finer points of these revisions, the picture painted of the US economy is likely to more comprehensive and more accurate because of these important innovations.


1. There are many thorny issues of classifying the ownership of R&D capital. For example, federal spending on R&D benefits private industry as well as the federal government. For the most part, the BEA is classifying the funder as the owner. For more details, see the Preview of the 2013 Comprehensive Revision of the National Income and Product Accounts [pdf].

2. For true aficionados of national accounts, it should be noted that the BEA will be identifying a separate pension plan subsector in the NIPAs to track the corresponding current receipts, expenditures, and cash flows of this sector. These changes will also bring the US NIPA into closer alignment with the standards set by the SNA.

3. Munnell, Alcia H., Jean-Pierre Aubry, Josh Hurwitz, Madeline Medenica, and Laura Quinby. 2012. “The Funding of State and Local Pensions: 2011–2015,” Issues in Brief, Center for Retirement Research, revised May 2012.