Printer Friendly

Satellite accounts in a modernized and extended system of economic accounts.

Reflecting growing interest in moving the U. S. economic accounts toward the United Nations System of National Accounts (SNA), the working group on improving economic statistics headed by the chairman of the President's Council of Economic Advisers recently recommended such a move. This article introduces the SNA and describes satellite accounts, which promise to be an important and powerful addition to that system. A satellite account for research and development is presented to illustrate the general features of satellite accounts and the kinds of problems faced in implementing them statistically.

THE IDEA OF A comprehensive, consistent set of macroeconomic accounts that embraces nonfinancial and financial flows and stocks, adjusted for price change over time where appropriate, is not new. For example, in the mid-1950s a report written by a group of leading economists for the Budget Bureau, the predecessor of the Office of Management and Budget, recommended organizational and statistical changes that would bring the United States closer to such a system. Several private researchers - among them Richard and Nancy D. Ruggles, John W. Kendrick, and Robert Eisner - have developed comprehensive systems. With new developments in economic theory, increases in the ability to manipulate data, and the emergence of new economic problems, interest in the idea has picked up. When interest picked up, it reflected an additional concern: the international comparability of the measures of economic growth, inflation, and saving rates that are part of the economic accounts.

The concern for the international comparability of policy variables leads directly to the United Nations System of National Accounts (SNA). The SNA is the set of guidelines in economic accounting followed by most other countries, which adapt it to fit their economies and statistical systems. [1]

The SNA is comprehensive in coverage and it is integrated. The SNA includes information that is now in the U. S. national income and product accounts (NIPAs), the input-output accounts, the flow-of-funds accounts (prepared by the Board of Governors of the Federal Reserve), and balance sheets. As well, it also includes information in addition to that in the existing U. S. economic accounting systems; it includes explicit revaluation accounts, which show changes in value of financial and nonfinancial assets, and a fuller set of accounts for business, households, and the other sectors of the economy. The SNA is an integrated framework in that it uses consistent definitions, classifications, and accounting conventions throughout.

The original SNA took shape in the 1940s and 1950s, when it was the product of the same theoretical developments and economic conditions as the NIPAs. A revision of the SNA in 1968 substantially expanded its framework, which is now being updated. The revised system, targeted for completion in 1993, will include satellite accounts, which promise to be important and powerful new tools for measurement and analysis.


The inclusion of satellite accounts in the revised SNA represents the maturing of a tool that has been found to add substantial flexibility to a nation's system of accounts. The first satellite accounts - which for the moment will be defined loosely as supplementary information tied to the main set of national economic accounts - were developed in France beginning in the 1960s. One of the earliest was a satellite account for housing. Since then, primarily in Europe, Canada, and Japan, satellite accounts have been built or are being built for such fields as agriculture, health, research and development, transportation, trade, education, the environment, natural resources, and social protection.

Over time, satellite accounts for particular fields have come to be associated with the following characteristics:

1. They feature data for a whole field of economic

activity and provide a framework for arraying more

comprehensive information about the field than can

be shown in the main accounts.

2. They are purpose-oriented in that the criterion for

a transactor's or transaction's inclusion is its linkage

to the field.

3. They are articulated with the main accounts and

contain at least one measure that is also in the main


4. They present information in ways that are different

from the main accounts; definitions, classifications,

and accounting conventions may differ from those

used in the main accounts in order to provide the

most useful presentation of information about the

field. What is counted as current or capital in the

main accounts may be changed, or the boundary

of production may be moved. The definitions, classifications,

and accounting conventions must be

consistent within the account, however.

5. They often contain tables that answer several questions:

Who is producing, and what are the means

of production? Who is financing? What is the result

of the expense, and who is benefiting or using the


6. They often encompass monetary and physical data

in an integrated fashion. Physical data may relate

to production, for example, the number of persons

employed in the field or the stocks of equipment.

Physical data may also relate to beneficiaries, for

example, the number of persons being affected by

activities in the field.

Satellite accounts thus go beyond BEA's tables, such as for gross auto product, that supplement the main U.S. accounts. BEA's tables typically provide further detail but without changing definitions, classifications, and accounting concepts, and they are single- rather than multi-dimensional.

When definitions, classifications, or accounting conventions are changed, the principal aggregates, such as GNP, might be altered in the satellite account. The advantage of satellite accounts is that such an alternative view of the economy can be obtained without disturbing the main accounts.


Satellite accounts will be a prominent feature of the revised and expanded U.S. national economic accounts. Because they can be constructed independently of the much larger effort needed to rework the accounts, some satellite accounts will be made available before completion of the full set of revised and expanded accounts. BEA has begun exploring two satellite accounts: (1) for research and development (R&D), and (2) more tentatively, for natural resources.

R&D was chosen as the field to explore first for several reasons. Data to support an R&D account are already largely available; in fact, Kendrick and Eisner have constructed economic accounts for the United States that show R&D in some of the ways that a satellite account would. The data include the results of National Science Foundation surveys of industrial and academic R&D and tabulations of government R&D expenditures. [2] The theoretical and technical challenges to the construction of an R&D account may be addressed by quantitative techniques that have been pioneered by academic and federal government researchers. Also, BEA has some experience in estimating constant-dollar R&D expenditures series for defense and biomedical research.

The satellite account will treat as capital formation the expenditures on R&D that are treated in different ways in the NIPAs. For example, R&D performed by the business sector is treated in the NIPAs as intermediate products consumed if it is financed by business or as government purchases if it is financed by the government; the former is netted out, but the latter is counted, in deriving GNP. In the satellite account, R&D expenditures will be treated as formation of intangible capital. Thus, estimates will be needed of the consumption of this capital and of the value of the services it provides.

Table I uses hypothetical (but roughly representative of actual) values to illustrate the effect on GNP if R&D expenditures were capitalized. The top panel shows the product and income components that sum to GNP and charges against GNP in the NIPAs. The bottom panel shows the components modified to provide R&D investment by the several sectors that sum to $130 billion and R&D capital consumption allowances that sum to $95 billion. The services rendered by R&D capital are assumed equal to the capital consumption allowances. The services are imputed to personal consumption expenditures, business intermediate products consumed, and government purchases. These assumptions would alter personal, business, and government expenditures. For example, personal consumption expenditures are lowered by the amount of nonprofit institutions' expenditures on R&D, which are reassigned to investment, and increased by the services of the R&D capital, which are imputed.

The effects on the national income and product account are sizable: In all, GNP is increased more than 2 percent. The resulting estimates represent an increase in imputations and corresponding large changes in major NIPA aggregates. The location of the estimates in a satellite account, however, provides an alternative way of looking at the economy that does not require the sacrifice of the GNP measure in the NIPAs.

An R&D satellite account should be a valuable addition to the nation's economic accounts. The most important innovations of the R&D satellite account are expected to be constant-dollar flow and stock series. These series should be very useful in analyzing the sources of productivity growth, for example. Several economists have used a neoclassical production-function approach to examine the quantitative linkage between research and productivity and, in spite of data limitations, have found that research has significant effects on productivity (Eisner 1989, Griliches 1980). R&D capital is very unlike most conventional capital stocks; it is intangible, it is much like a public good in that multiple users can make use of it without diminishing the stock (Pakes and Schankerman 1984), and R&D in one industry or sector can affect productivity in other industries or sectors (Scherer 1982). It is possible to use marginal conditions to relate R&D flows to productivity increases (Mansfield 1984). In a more straightforward manner, however, constant-dollar R&D capital stocks can be used to estimate directly production functions that show the relationships between R&D and productivity (Eisner 1989).


The R&D satellite account, as it is envisaged, will have most of the characteristics of satellite accounts listed earlier. It will provide an overall picture of R&D activity as well as estimates of the R&D capital stock. The account will array activities and stocks by sector and disaggregate R&D by type of research or development. The account will be articulated with the main accounts through major components of R&D spending, such as federal nondefense purchases of space R&D. The classification between current account and capital account used in the satellite account will differ from that of the NIPAs. R&D classified as current-account expenditures in the NIPAs will be counted as investment in the satellite account.

The R&D account will include tables that show the production and financing of R&D, numbers of persons engaged in R&D, and the capital stocks resulting from R&D; these will answer the questions typical of satellite accounts. Where data are available, the account will also feature joint presentations by category (industry or discipline) of R&D expenditures and numbers of scientists and engineers engaged in R&D activities.

The first versions of the R&D satellite account will be based on information that is available to BEA from the National Science Foundation and other sources. (Later versions may elaborate the detail shown or expand the areas of coverage.) BEA will reconcile the available information with the underlying detail of its currently published accounts. The reconciled expenditures in the

R&D account will match the corresponding items in the NIPAs and the underlying input-output accounts. For example, federal government expenditures for intramural R&D and for R&D carried out by business and nonprofits will match the corresponding components of federal purchases in the NIPAs.

The most comprehensive current-dollar table will show R&D expenditures disaggregated by the sector performing the R&D and the sector financing the R&D. This two-way disaggregation gives a comprehensive overview of the R&D process.

Several additional current-dollar tables will provide sector detail. An industry R&D table will show performance and financing of R&D disaggregated by industry. This table will be designed to detail the R&D process by industry and to enable comparison of the relative amounts of the various types of R&D in different industries. (The level of disaggregation by industry will attempt to match that published by BEA in the NIPAs; in the future, emphasis will be placed on increasing the level of detail shown for nonmanufacturing.) Another industry table will show R&D performed by industry and the numbers of full-time equivalent scientists and engineers engaged in R&D activities in the industries. An academic R&D table will provide an overview of R&D performed by colleges and universities. It will show various types of R&D expenditures disaggregated by discipline. A government R&D table will provide an overview of federal funding for R&D. The funding will be disaggregated by discipline and by department funding the R&D. The level of disaggregation by discipline will be the same as that of the academic R&D table.

A table for constant-dollar expenditures for R&D, disaggregated by sector, will be a major feature of the account. Corresponding deflators will be shown in an additional table. The deflators will be cost-type indexes that are composed of weighted sums of the compensation of persons engaged in R&D and the supplies, materials, and indirect costs used to support the R&D. BEA already calculates deflators for defense R&D and for federal biomedical R&D. The methodology for deflation of other types of R&D will reflect this existing methodology as well as the work of other economists on developing deflators for other types of R&D (Mansfield 1987, Jankowski 1990).

Additional tables will show R&D capital stocks and services of R&D capital. Once time series have been created for current- and constant-dollar R&D flows, current- and constant-dollar capital stocks can be calculated. The stocks, on both a net and gross basis, will be calculated using a perpetual inventory methodology similar to that currently used by BEA to calculate capital stocks for business structures and equipment. Capital consumption allowances will be calculated using the same methodology; these allowances will be used as measures of the services provided by R&D capital to the household, business, and government sectors. [3]


The measure of R&D capital is intended to capture the value of the stock of information created by R&D efforts. Accordingly, it is impossible to measure directly, and indirect measures will be constructed.

R&D output will be measured by the expenditures on it. Measuring R&D in this way has parallels in other national accounting practices. Expenditures for R&D by the government and nonprofit institutions are made because they value the information gained from R&D enough to make the expenditure. Similarly, business R&D expenditures are made because the value created by the expenditures is sufficient to justify them. Because business tangible investment is measured by the expenditures on it, an analogous treatment for intangible R&D investment seems justified. Further, this treatment is consistent with deflators that are cost-type indexes based on weights and prices specific to R&D (Kendrick 1976, Mansfield 1984, Jankowski 1990). BEA currently uses this approach in calculating deflators for defense and biomedical R&D (Holloway and Reeb 1987).

With regard to the capitalization of R&D expenditures, several issues arise that may be dealt with as what?" who?" and when?"

What types of R&D should be capitalized? At one extreme, only business expenditures for development that are directly linked to planned commercial applications could be capitalized. At the other extreme, all R&D expenditures could be capitalized.

The focus is often on basic research: Should it be capitalized, or should it continue to be classified as consumption? By the National Science Foundation's definition, basic research includes the cost of projects that "represent original investigation for the advancement of scientific knowledge and ... do not have specific immediate commercial objectives." Basic research is undertaken for the purpose of improving man's understanding of the world and not for the purpose of increasing productivity or adding to production. Although basic research increases knowledge that forms the background for applied research and development, it has no immediate linkage to commercial applications.

The results of basic research are typically published and become public goods that are available to all, not to just the organization performing or financing the research. In addition, basic research is highly mobile internationally; unless the results of the research are published only very obscurely, the knowledge gained by research in one country becomes rapidly available in other countries. The issues raised by international mobility apply to a lesser extent for all R&D; one analyst has found indications of considerable international technology transfers within multinational companies (Mansfield 1984).

The "who" issues refer to the sectoring of R&D capital. To begin with, what sectors' R&D will be capitalized? Current NIPA practice would suggest capitalizing only business R&D; in the NIPAs, only the business sector carries out investment. However, this treatment seems inappropriate for R&D capital because of its similarity to public goods, i.e., government or nonprofit R&D may provide benefits to business. Then, to what sectors does the capital accrue? The capital could accrue to the sector financing the R&D or the sector performing the R&D. Alternatively, some analysts have assumed that most or all applied research and development capital accrues to the business sector (Kendrick 1976, Eisner 1989). Others have implicitly assumed that only an industry's own R&D would accrue to it and have found significant productivity effects for industrial R&D (Mansfield 1980).

One of the "when" issues is about the point in time at which R&D is capitalized. R&D could be capitalized as it is performed. Because much R&D involves multiyear efforts, the analogous capitalization of large-scale structures as they are built could serve as a guide. Alternatively, R&D could be capitalized when the products or services based on it come to market (Kendrick 1976). This approach requires the estimation of two lags: The first is the gestation lag (that is, the time between the beginning of the R&D project and its completion), and the second is the application lag (that is, the time between completion of the R&D project and the time its results find commercial application). Although some academic research has addressed these lags, additional work will be needed to estimate the lags' lengths.

Another "when" issue concerns the length of the period of time over which the R&D capital is depreciated. Some researchers have assumed that basic research has an infinite lifespan, but this leads to problems in calculating initial capital stock levels and requires that basic research be viewed as something very different from applied research. Another approach would be to depreciate R&D capital over the "patent lives" of commercial products, but some analysts have found that the value of patents declines rapidly following their issue, suggesting that the true economic lifespan of R&D capital is shorter than patent lives (Pakes and Schankerman 1984). An alternative approach would base the lifespan of R&D capital on the average lifespans of fixed capital assets. This approach, however, fails to recognize that the market value of R&D capital declines as, among other things, new innovations displace the original innovation. Another approach would be to base the lifespan of R&D capital on the periods over which companies expect to product goods and services using the technology developed by an R&D project (Blades 1988). These periods would have to be obtained by surveying companies. Yet another approach would be to assume plausible, but fairly arbitrary, lifespans for R&D capital. This approach has been used in some recent estimates of R&D capital stocks (CBO 1988, Eisner 1989).

* Carol S. Carson is the Deputy Director of the Bureau of Economic Analysis, U.S. Department of Commerce, Washington, DC. She is participating in the revision of the United Nations System of National Accounts. Bruce T. Grimm is Assistant to the Deputy Director, Bureau of Economic Analysis. This article is a shorter version of one presented at the 32nd Annual Meeting of the National Association of Business Economists, September 23-27, 1990, Washington, DC.

When this article was being written, a request for funding to plan and carry out a move to the United Nations system was before Congress. Although the outcome was uncertain, the article describes the work as though the funding were in place.


[1] For an overview of the SNA and a brief comparison of it with the U. S. national income and product accounts, see Carol S. Carson and Jeanette Honsa, "The Unite Nations System of National Accounts: An Introduction,' Survey of Current Business 70 (June 1990), pp. 20-30.

[2] An internationally accepted definition of R&D is "creative work undertaken on a systematic basis in order to increase the stock of knowledge . . . and the use of this stock of knowledge to devise new applications" (OECD 1981).

[3] There are other measures of capital services, including the user cost of capital measure, which is the sum of the capital consumption allowance and the real own return to capital.

[4] Much of this section is based on earlier work by BEA's Thae Park.

REFERENCES Blades, Derek. 1989. "Treatment of R&D Expenditures

in the National Accounts." Paper prepared for the

SNA coordinating group meeting, Luxembourg,

January 23-27, 1989. Congressional Budget Office. 1987. Trends in Public

Investment. Washington, DC: U.S. Government

Printing Office. Eisner, Robert. 1989. The Total Incomes System Of

Accounts. Chicago: University of Chicago Press. Griliches, Zvi. 1980. "Returns to Research and Development

Expenditures in the Private Sector." In

New Developments in Productivity Measurement

and Analysis. ed. John Kendrick and Beatrice

Vaccara, pp. 419-54. Chicago: University of Chicago

Press for the National Bureau of Economic


, ed. 1984, R&D, Patents, and Productivity.

Chicago: University of Chicago Press. Holloway, Thomas and Jane Reeb. 1987. "The Biomedical

Research and Development Price Index: Revised

and Updated Estimates." Bureau of

Economic Analysis Discussion Paper No. 20.

Washington, DC. Jankowski, John. 1990. "Construction of a Price Index

for Industrial R&D Inputs." National Science

Foundation. Draft. Kendrick, John. 1976. The Formation and Stocks of Total

Capital. New York: Columbia University Press

for the National Bureau of Economic Research. Mansfield, Edwin. 1980. "Basic Research and Productivity

Increase in Manufacturing." American Economic

Review 70 (December).

1984. "R&D and Innovation: Some Empirical

Findings." In Griliches 1984, pp. 127-48.

1987. Price Indexes for R and D Inputs,

1969-1983." Management Science 33 (January). Organization for Economic Cooperation and Development

(OECD). 1981. The Measurement of Scientific

and Technical Activities. Paris. Pakes, Ariel and Mark Schankerman. 1984. "The Rate

of Obsolescence of Patents, Research Gestation

Lags, and the Private Rate of Return to Research

Resources." In Griliches 1984, pp. 73-88. Scherer, F.M. 1982. "Inter-industry Technology Flows

and Productivity Growth." Review of Economics

and Statistics 64 (November).
COPYRIGHT 1991 The National Association for Business Economists
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1991 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Carson, Carol S.; Grimm, Bruce T.
Publication:Business Economics
Date:Jan 1, 1991
Previous Article:The statistics corner.
Next Article:The Age of Diminished Expectations: U.S. Economic Policy in the 1990's.

Terms of use | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters