Printer Friendly

Revised estimates of the national income and product accounts of the United States, 1929-85: an introduction.

Revised Estimates of the National Income and Product Accounts of the United States, 1929-85: An Introduction

THE Bureau of Economic Analysis has released the results of a comprehensive--that is, benchmark--revision of the national income and product accounts (NIPA's). This revision is the eighth of its kind; the last such revision was released in December 1980. All series in the NIPA's are revised beginning in 1973, and many of them also are revised for earlier years to provide continuous time series. All constant-dollar series and price indexes are revised back to 1929 to reflect a shift in the base period from 1972 to 1982.

This article focuses on the estimates for 1972-84, the period in which the revisions are concentrated. On the whole, the revised estimates do not appear to have substantially altered the view of the economy provided in the previously published estimates (chart 1).

* For current-dollar GNP, the average annual rate of increase from 1972 to 1984 is the same--9.9 percent--in the revised as in the previously published estimates.

* For real GNP, the average annual rate of increase from 1972 to 1984 is less in the revised estimates--2.5 percent compared with 2.7 percent. A major source of the revision is the shift in the base period from 1972 to 1982.

* For prices, the average annual rate of increase in the GNP fixed-weighted price index from 1972 to 1984 is less in the revised estimates--6.6 percent compared with 7.3 percent. Major sources of the revision are the shift in the base period and the introduction of a new price index for computers.

* The pattern of short-term, or cyclical, fluctuations in real GNP in the revised estimates is similar to that shown in previous ones. The revisions do not alter the peak and trough quarters of the business cycles, and thus do not change their duration. In the most recent contraction, 1981-82, the average quarterly decline is somewhat more in the revised estimates than in the previously published ones; through the third quarter of 1985, the subsequent expansion is somewhat less vigorous.

The "National Income and Product Accounts Tables" in this issue of the SURVEY OF CURRENT BUSINESS contains the revised estimates for the series in the regular quarterly presentation for the years 1983 and 1984 and for quarters beginning with the second of 1984. Revised estimates for all NIPA series will be available in the spring in two separate volumes: for 1929-82, years for which estimates will not be revised until the next comprehensive revision, and for 1982-84. Information on obtaining the revised estimates prior to the publication of these volumes is described in the box below.

This SURVEY article is one of several about the comprehensive revision. Earlier, preliminary revised current-dollar estimates for 1977 appeared in the May 1984 SURVEY, a description of the improved adjustments for misreporting on tax returns appeared in the June 1984 SURVEY, and explanations of the changes incorporated in the revision appeared in the October 1985 SURVEY. Forthcoming articles will describe the new price index for computers, one of the major statistical changes, and will review aspects of the economy as seen through the revised estimates.

Part I of this article briefly reviews the definitional and classificational changes and the statistical changes presented in the October SURVEY and provides some additional information. Part II describes the size of the revisions in selected NIPA series and traces them to their sources. Part III describes the redesign of NIPA tables.

Part I. SOURCES OF REVISIONS

This part of the article describes the sources of the revisions in terms of the two kinds of changes--definitional and classificational, and statistical--incorporated in the revised estimates. In general, definitional and classificational changes are carried back to the initial year of impact, which, in many cases, is 1929. Statistical changes, on the other hand, affect various time periods; only a few affect the estimates back to 1929. Several begin in the 1950's, but most begin in 1973.

Statistical Changes

Statistical changes, which are the outstanding feature of this revision, have several origins:

* Shift of the base period by the substitution of the prices of 1982 for those of 1972 in calculating constant-dollar series and the substitution of the composition of output in 1982 for that in 1972 in calculating fixed-weighted price indexes;

*

* Incorporation of new and revised data from regularly used sources that become available less often than annually, referred to as "benchmark sources;"

* Incorporation of new and revised data from regularly used sources that are usually incorporated at the time of the annual July revisions of the NIPA's, referred to as "regular sources for 1982-84;"

* Use of source data not previously available; and

* Use of new estimating procedures.

Shift of the base period and regular

data sources

A shift to a more recent base period is a standard procedure designed to provide measures of real output and prices that are more relevant for many purposes. The NIPA's were last "rebased" in the comprehensive revision released in 1976, when the base period was shifted from 1958 to 1972. The year 1982 is selected as the new base period because it is the latest year for which the NIPA estimates will not be revised until the next comprehensive revision.

Major regular data sources for 1982-84 are shown in table 2 (reproduced from the October article). Data from these sources are incorporated because the revisions of the estimates for 1982-84 that would have customarily been published in July 1985 are combined with the longer term revisions. Of the benchmark sources, the single most important is BEA's input-output (I-O) table for 1977. Summary or, in some cases, preliminary statistics from the 1982 economic censuses are also used.

Changes in methodology

Because the availability of new source data and the development of new estimating procedures often go hand in hand, these changes are described as "changes in methodology." A list of such changes is shown in table 3. Most of them--the ones marked with asterisks--were described in the October article. The others are described next: Four affect current-dollar estimates, and four affect deflation procedures. In addition, for the improved adjustments for misreporting on tax returns, the change that has the single largest effect, table 4 shows estimates for more years than were presented in the October article.

Improved estimates of imputed interest paid by life insurance carriers and by savings and loan associations.--The estimates of imputed interest paid by life insurance carriers and by savings and loan associations are revised beginning in 1973 to reflect the recognition of the conversion of a growing number of these organizations from mutual associations to stock corporations. Imputed interest paid differs for the two types of organizations; for stock corporations, it is defined as property income received less interest paid to depositors or beneficiaries, while for mutual associations, profits before tax also are subtracted. The effect of the change is to raise the imputed interest components of net interest, personal interest income, and personal consumption expenditures for most years. In the previously published estimates, as well as in the revised estimates for the pre-1973 period, all carriers and associations are treated as mutual associations for the estimation of imputed interest paid.

Improved estimates of consumer expenditures for goods.--The revised estimates incorporate commodity-flow procedures not only for 1977--the year of the I-O table--but also for 1980. The use of commodity-flow procedures for 1980 marks the first use of these procedures in a comprehensive revision in estimating personal consumption expenditures for an intercensal year. The 1980 estimates are derived from (1) shipments and inventory data from the Census Bureau's 1980 Annual Survey of Manufactures; (2) detailed merchandise exports and imports data from the Census Bureau; and (3) the margin rates for trade and transportation and the consumer share of industry output determined to be final purchases from the 1977 I-O table. For years when commodity-flow-based estimates are not available, consumer purchases of goods are prepared by interpolation and extrapolation using retail store sales. The previously published estimates for 1973 forward were extrapolations of the commodity-flow estimates from the 1972 I-O tables.

Improved estimates of producers' durable equipment.--The revised estimates of producers' durable equipment incorporate new detail in the commodity-flow procedures for 1982. The revised estimates are derived from (1) detailed industry shipments and inventory data from the 1982 Census of Manufactures; (2) merchandise exports and imports from the Census Bureau; and (3) margin rates for trade and transportation and the business share of industry output determined to be final purchases from the 1977 I-O table. For years when commodity-flow-based estimates derived from census of manufactures shipments data are not available, estimates of producers' durable equipment are prepared by interpolating and extrapolating them using estimates from an abbreviated commodity-flow procedure.

Improved estimates of consumer purchases of energy.--The revised estimates of consumer purchases of electricity, natural gas, and gasoline, beginning in 1968, reflect new information from the Energy Information Administration (EIA) of the U.S. Department of Energy. For electricity and natural gas, the effect of the introduction of EIA survey data on residential energy use for 1978-82 is to reduce the estimates. The new levels are extrapolated backward and forward based on trade association data. For gasoline, the effect of the introduction of data for 1972-82 on the price per gallon developed by EIA and by the Census Bureau is to reduce the estimates. The new 1977 level is interpolated back to 1968 and the new 1982 level is extrapolated forward based on retail sales of gasoline service stations.

Deflation procedures.--Procedures used for deflation are improved for several components. In most cases, the new procedures had already been introduced for recent years, and the revisions--which affect estimates of both constant dollars and prices--extended the procedures back in time.

The estimates of residential additions and alterations for 1959-79 are revised to reflect the use of a different price series in the deflation procedure. This change, which was to substitute data from the Consumer Price Index for the Census Bureau price index for single-family houses sold, was noted in the July 1982 SURVEY.

The estimates of the Commodity Credit Corporation (CCC) inventory change component of government purchases are revised to incorporate commodity detail on acquisitions and dispositions of farm crops by the CCC. Separate deflation is introduced for annual estimates for 1947-77 and for quarterly estimates for 1972-77.

The estimates of national defense purchases are revised to reflect an expanded level of detail at which deflation is carried out. The new detail mainly affects the estimates for 1972-77.

The estimates of net exports are revised to reflect (1) the introduction of a new price index for computers and (2) the expansion of the level of detail at which deflation is carried out for merchandise imports and for exports and imports of services. A new price index for exports and imports of business machines, which incorporates the new price index for computers, is introduced beginning in 1964. The level of detail for deflation for merchandise imports, beginning in 1929, is expanded to provide for the separate deflation of petroleum; in the previously published estimates, separate deflation began in 1967. For exports and imports of services other than factor income, the level of detail for quarterly deflation, beginning in 1959, is expanded to several subcomponents; in the previously published estimates, the use of this level of detail began in 1980.

Definitional and

Classificational Changes

Definitional and classificational changes are made so that the accounts depict in an analytically useful way the evolving institutions in the economy and the economic activities in which they engage. These changes usually are introduced at the time of a comprehensive revision so that they can be carried back historically on a consistent basis.

Ten definitional and classificational changes (hereafter referred to as definitional changes) are incorporated in this comprehensive revision.

* Two relate to the classification of certain business expenditures as investment: Capitalize replacement railroad track, and capitalize major replacements to residential structures.

* Three relate to Federal Government employment benefit programs: Impute a social insurance fund for military retirement, impute a social insurance fund for Federal unemployment insurance for military personnel, and treat the Civilian Health and Medical Plan of the Uniformed Services as an employer-paid health insurance plan.

* Three relate to foreign transactions: Impute bank service charges to foreigners, reclassify military shipments financed by "forgiven" loans, and reclassify membership contributions to international organizations.

* Two relate to government assistance programs: Reclassify medical vendor payments, and reclassify payments under two State and local government assistance programs.

All of the definitional changes except the reclassification of membership contributions to international organizations were described in the October article. For this change, back to 1946, U.S. membership contributions to these organizations are now treated as transfer payments to foreigners. Federal Government nondefense purchases and also imports of services other than factor income are reduced by the change, so that GNP is not affected. The change reflects that the membership contributions are being used primarily for humanitarian and developmental activities, not for administrative expenses, and that data to identify the latter amounts are not available. This revision, which will be made in the U.S. international transactions accounts in June 1986, brings the U.S. treatment into conformity with international guidelines.

Part II. THE SIZE OF THE REVISIONS

This part of the article considers the size of the revisions of the previously published estimates. The first section covers the revisions of annual current-dollar estimates, tracing them to the definitional and statistical changes. The second section covers the revisions of annual constant-dollar and price estimates. The revisions of the constant-dollar estimates are traced to revisions of the current-dollar estimates, the shift in the base period, and other revisions, including improved deflation. The revisions of the price estimates are traced to the shift in the base period and other revisions. For the most part, these discussions will focus on the period from 1972 to 1984. The third section covers the revisions of quarterly constant-dollar GNP over the business cycles since 1960, with particular attention to the revisions during the present expansion.

Annual Current-Dollar

Estimates

GNP in current dollars is revised up in all years, 1972-84 (table 5). The percent revision is 2.3 percent in 1972 and 3.1 percent in 1984. Definitional changes account for almost one-third of the revision in GNP in 1972 and for about one-fourth of the revision in 1984. Among the definitional changes, the capitalization of major replacements to residential structures and the imputation of a social insurance fund for military retirement have the largest effect in raising GNP. Among the statistical changes, the improved adjustments for misreporting on tax returns and the improved methodology and new data for residential investment have the largest effect. Of the major product-side components of GNP, personal consumption expenditures and residential investment are revised up substantially in all years from 1972 to 1984. Nonresidential investment and net exports are also revised up in all years; government purchases are revised down in all years. Change in business inventories is revised up in some years and down in others.

National income is revised up in all years, as definitional and statistical changes, particularly the imputation of a social insurance fund for military retirement and the improved adjustments for misreporting on tax returns, lead to large upward revisions in compensation of employees and proprietors' income. Rental income of persons is revised down substantially in all years; net interest is revised up beginning in 1976; and corporate profits is revised up through 1980 and down thereafter.

Personal income is revised up, reflecting, for the most part, the statistical changes that affect national income. In addition, personal income is raised by the definitional change that reclassified medical vendor payments; in personal income that change adds to transfer payments.

Revisions in level--in dollars and in percent--of the NIPA components of the five-account summary system are shown for selected years in table 6. Revisions in the four major product-side and seven major income-side components of the national income and product account--the first accoun in the summary system--are discussed in some detail in the following subsections. For each major component, a table gives the total revision for 1972, 1977, 1982, and 1984. The total revision is broken down both by source of revision--definitional and statistical--and by subcomponent.

Product-side components

Personal consumption expenditures (PCE).--The revisions in PCE are shown in table 7. The upward revision from definitional changes is accounted for by the reclassification of medical vendor payments, which raises purchases of medical care in services and also of drugs in nondurables. The upward revision from statistical changes is largely accounted for by the improved adjustments for misreporting on tax returns, which raises all three components of PCE, particularly nondurables and services. Another part of the substantial upward revision in services is from the incorporation, through the 1977 I-O table, of data from the 1977 Census of Service Industries, which covered non profit organizations for the first time. Two other statistical changes have a noticeable impact on recent years: The incorporation of data from the 1983 American Housing Survey lowers housing services in 1982-84; the incorporation of registration data, the basis of the consumer-business allocation of motor vehicles, raises consumer purchases (and lowers business purchases) of both new and used autos in 1984.

Gross private domestic investment.--Revisions in gross private domestic investment, shown in table 8, are substantial. The large upward revisions through 1982 are mainly from definitional and statistical changes in residential investment and from statistical changes in producers' durable equipment (PDE).

Residential investment is revised up in all years. About one-half of the revision is from the capitalization of major replacements to structures; statistical revisions are primarily from the incorporation of a new Census Bureau series on construction value and a revised series on additions and alterations.

The upward revision in PDE through 1982 is from the incorporation of information from the 1977 I-O tables and from preliminary 1982 commodity-flow estimates. In 1983 and 1984, PDE is revised up much less. The revision for those years reflects the incorporation in an abbreviated commodity-flow procedure of shipments data from the 1983 Annual Survey of Manufactures and of more detailed information on exports and imports. The detail in these sources improves the allocation between complete equipment and parts; only the former is included in PDE. In addition, for 1984, there is new information for the consumer-business allocation of autos. For 1983 and 1984, as well as earlier years, there are substantial downward revisions in the office, computing, and accounting machinery category, which includes computers, and upward revisions in a number of other PDE categories.

Revisions in nonresidential structures are small: upward for railroads, from the definitional change and, except for 1983, upward for petroleum and natural gas exploration, from the incorporation of information from the 1977 Census of Mineral Industries and the 1983 Joint Association Survey of Drilling Costs; downward for electric utilities, reflecting the new estimating methodology and regular data sources for 1983 and 1984.

Revisions in change in business inventories reflect the incorporation of regular source data. The revisions are small, except for nonfarm inventories in 1983 and 1984.

Net exports.--As shown in table 9, net exports of goods and services are revised up. Definitional changes have only a small impact on net exports. The definitional change that imputes bank service charges to foreigners does not affect net exports because it raises exports and imports by equal amounts: Exports of services other than factor income and imports of factor income services are each raised $0.6 billion in 1972, $1.5 billion in 1977, $2.8 billion in 1982, and $5.0 billion in 1984. Through 1982, the revision in net exports is largely from the exclusion of the Commonwealth of Puerto Rico and U.S. territories from exports and imports in the NIPA's. Beginning in 1983, merchandise trade, particularly imports, is affected by the Census Bureau's timing adjustments to trade data. Statistical changes also reflect the incorporation of other information from BEA's international transactions accounts.

Government purchases.--Revisions in government purchases of goods and services, shown in table 10, are downward. The largest revisions are from definitional changes: The upward revision in Federal national defense purchases from the imputation of a social insurance fund for military retirement is more than offset by the downward revision in State and local purchases from the reclassification of medical vendor payments. Revisions in Federal Government purchases from statistical changes are small. In State and local purchases, structures are revised up, beginning in 1975, due to the improved estimating methodology, which incorporates data from the Federal Highway Administration.

Income-side components

Compensation of employees.--Revisions in employee compensation, shown in table 11, are upward for both wages and salaries and for supplements to wages and salaries. In supplements to wages and salaries, most of the revision in employer contributions for social insurance is accounted for by the definitional change that imputes a contribution by the Federal Government for military retirement. In other labor income, the revisions are progressively larger through 1981 And then taper off. This pattern reflects upward revisions in employer contributions for health insurance and, beginning in 1982, downward revisions in employer contributions to private pension and profit-sharing plans. The former is due to the incorporation of a revised series from the Health Care Financing Administration. The latter is due to the incorporation of corporate tax return tabulations for 1982 and fragmentary information that indicates reduced levels of contributions due to over-funding of many companies' defined benefit pension plans.

In wages and salaries, the revision largely reflects the improved adjustments for misreporting on tax returns. The revision in 1984 is also from the incorporation of unemployment insurance information on private wages and salaries.

Proprietors' income.--Revisions in proprietors' income with inventory valuation and capital consumption adjustments are shown in table 12. The revision raises the level of proprietors' income by substantial amounts throughout the 1970's and 1980's. The revised estimate of proprietors' income is $233.7 billion in 1984, about 1-1/2 times larger than previously published.

Nonfarm proprietors' income is revised up substantially, largely due to the improved adjustments for misreporting on tax returns. Other statistical changes have small and offsetting effects except in 1983 and 1984, when the nonfarm capital consumption adjustment--the difference between tax-return-based capital consumption allowances and capital consumption allowances with capital consumption adjustment ("economic depreciation")--is revised up. This revision reflects the estimates of economic depreciation, which are the result of revised perpetual inventory calculations. Revisions in farm income are largely from the U.S. Department of Agriculture's introduction into its income estimates of benchmarks based on the 1978 and 1982 Census of Agriculture.

Rental income.--Revisions in rental income of persons with capital consumption adjustment are shown in table 13. The revision lowers the level of rental income of persons by substantial amounts throughout the 1970's and 1980's. The revised estimate of rental income is $10.8 billion in 1984, only about one-fifth as large as previously published.

The substantial downward revisions in rental income stem primarily from two statistical changes. The first is the methodological change that incorporates an expanded list of homeownership expenses in the calculation of net rental income of owner- and tenant-occupied nonfarm dwellings. The second is the incorporation of data from both the 1980 Census of Housing and its supplementary Survey of Residential Finance and from the 1983 American Housing Survey. These data lower the estimate of space rent (the "receipts" item in the calculation of net rental income of owner- and tenant-occupied non-farm dwellings) beginning in 1973. The revisions are particularly large in 1982-84. The revisions in the capital consumption adjustment primarily reflect the definitional change for resisdential replacements.

Corporate profits.--Revisions in corporate profits with inventory valuation adjustment and capital consumption adjustment (CCAdj) are shown in table 14. Statistical changes, mostly in the CCAdj, account for practically all of the revision in profits from current production. Except in 1984, revisions in the CCadj reflect the revised estimates of economic depreciation from the revised perpetual inventory calculations; the tax-return-based measure is revised very little.

Profits before tax are revised up somewhat in all years, reflecting revisions from a variety of sources, including the improved estimates of the defaulters' gain adjustment, the exclusion of the Commonwealth of Puerto Rico and U.S. territories from NIPA estimates, and revised estimates of the oil and gas exploration expenditures adjustment. Incorporation of newly available Internal Revenue Service profits data for 1982 leads to smaller upward revisions in 1982-84.

Net interest.--Revisions in net interest, shown in table 15, are small until the 1980's, when this component is revised up by sizable amounts. After the mid-1970's, the downward revision in net interest due to the definitional change that imputes bank services to foreigners is more than offset by the effect of statistical changes. The improved estimates of investment income of private noninsured pension plans raise net interest, beginning in 1973, because net imputed interest of these plans is revised up more than net monetary interest is revised down. The methodological change that relates to interest received by life insurance carriers and by savings and loan associations raises net interest in 1973-81. Net monetary interest, and thus net interest, is boosted by the incorporation of newly available data for 1982 from the Internal Revenue Service.

Capital consumption allowances with capital consumption adjustment.--Revisions in capital consumption allowances with capital consumption adjustment are shown in table 16. Definitional changes in investment account for the upward revisions until 1981, when revisions that are primarily due to revised perpetual inventory calculations begin to dominate. Revisions of the investment flows used in the perpetual inventory calculation raise economic depreciation; revised service lives and revised deflators reduce the estimates.

Capital consumption allowances are revised up in all years, primarily due to definitional changes. Corporate capital consumption allowances also reflect a sharp downward revision in 1983 and 1984 due to new information on the effect of the accelerated cost recovery system under the Economic Recovery Tax Act of 1981. Such information is used in estimating corporate depreciation for years for which information from tabulations of tax returns is not available. Noncorporate capital consumption allowances reflect an upward revision in all years in capital consumption allowances for residences. The estimates for residences, which affect rental income of persons, are based on perpetual inventory calculations. These perpetual inventory calculations, in turn, incorporate revised investment flows.

Other charges against GNP.--Revisions in the remaining components of charges against GNP--that is, GNP measured as the sum of income-side components--are shown in table 17. The revisions, all of which are statistical, are small except in 1984. The 1984 revisions reflect the incorporation of regular source data. For indirect business taxes, they are concentrated in State and local taxes. For subsidies less the current surplus of government enterprises, they are mainly in the current surplus of State and local government enterprises.

Annual Constant-Dollar and

Price Estimates

The previous section examined revisions in terms of differences between the revised and previously published levels. This approach is useful for current-dollar estimates, but not for constant-dollar estimates or for price indexes. The shift in the base period changes the units in which these series are measured, and the difference in level between the previously published and revised series primarily reflects the difference in the unit of measurement. For example, restating from 1972 dollars to 1982 dollars approximately doubles the level for real GNP, reflecting the increase in the level of prices from 1972 to 1982; it does not indicate a doubling of output. Therefore, this section, rather than examining revisions in terms of difderences in levels, examines revisions in terms of rates of change.

Table 18 provides an overview, for selected periods, of the revisions in rates of change in constant-dollar GNP and in the GNP fixed-weighted price index, as well as in current-dollar GNP and in the GNP implicit price deflator. For current-dollar GNP, the average annula rate of increase from 1972 to 1984 in the revised estimates, at 9.9 percent, is the same as in the previously published estimates. Constant-dollar GNP on the revised basis increases from 1972 to 1984 at an average annula rate of 2.5 percent--0.2 percentage point less than in the previously published estimates. The GNP fixed-weighted price index on the revised basis increases at an average annual rate of 6.6 percent--0.7 percentage point less than in the previously published estimates. The GNP implicit price deflator on the revised basis increases at an average annual rate of 7.3 percent--0.4 percentage point more than in the previously published estimates.

Revisions in year-over-year changes in real GNP within the period from 1972 to 1984 are small (chart 2). Reflecting the downward revision in the average annual rate of increase over the period, year-over-year changes are revised down in 9 of the 12 years. These changes are reduced by 1 percentage point in 1977 And 1/2 percentage point in 1973, 1976, 1979, 1981, and 1982. Only in 1978 is the change revised up by as much as 1/2 percentage point.

Reflecting the downward revision in the average annual rate of increase, year-over-year changes in the GNP fixed-weighted price index for the period from 1972 to 1984 are revised down in 11 of the 12 years (chart 3). The largest downward revisions, ranging from 1/2 to 1-1/2 percentage points, are in the 1970's. In 1980 and 1981, the change in GNP prices is revised down about 1/2 percentage point, but remains above 9 percent. The revised estimates still show a sharp deceleration in prices from 1981 to 1983. In 1983 and 1984, the change in GNP prices is virtually unrevised from about 4 percent.

Constant-dollar estimates

Table 19 provides, in the last three columns, a decomposition of the revision in the average annual rate of change in real GNP and in its major components for 1972-84 into three parts: (1) the part that is due to the current-dollar revision; (2) the part that is due to shifting the base period from 1972 to 1982; and (3) the part that is due to other revisions, including revisions in price indexes used in deflating current-dollar estimates and revisions in (shifts among) current-dollar component detail that, in effect, reweight the constant-dollar component detail. The part due to the current-dollar revision is the revision in percent change in current-dollar GNP and in each of the components. The part due to the shift in base period is measured as the difference between the change in the previously published estimates restated in 1982 dollars and the change in the previously published estimates in 1972 dollars. The part due to other revisions is obtained as a residual.

For real GNP, the downward revision of 0.2 percentage point in the average annual rate of increase for 1972-84 is more than accounted for by the shift in the base period, which has a -0.4 percentage point impact. As discussed in the October article, a shift to a more recent base period tends to reduce measured change in real GNP. The note on page 14 demonstrates with a simple example how this effect occurs. Because the current-dollar change in GNP is unrevised, the current-dollar revision does not contribute, and other revisions contribute a small offset.

Within the major components of real GNP, most of the average annual rates of change over 1972-84 are revised down. The revisions reflect, to varying extents, the current-dollar revisions, most of which raise the rate of change; the impacts of rebasing, most of which reduce the rate of change; and the effects of other revisions, which are mixed.

The change in real PCE is revised down 0.4 percentage point, from 3.1 to 2.7 percent, largely due to the impact of rebasing. Nondurables and services are revised down about the same as total PCE. Durables is revised down only slightly, as the upward revision in current dollars nearly offsets the impact of rebassing.

The downward revision in the change in real gross private domestic investment is more than accounted for by a sharp revision in PDE. The change in PDE is revised down from 5.6 to 4.4 percent, the largest revision among the major components of GNP. This revision is more than accounted for by rebasing; the current-dollar change is unrevised, so it does not contribute, and other revisions--including the new computer price index--contribute a small offset. (See the note on page 16 for a more detailed description of the revision in PDE, particularly with respect to business expenditures on computers.) The change in structures, the only component for which rebasing has a positive impact, is revised up 0.5 percentage point. The rebasing impact reflects the petroleum mining exploration, shafts, and wells component, in which the changes in prices and quantities have a positive relationship (both increase more than the average of prices and quantities of all items included in structures), rather than the negative one usually observed, over much of this period. In addition, other revisions, largely the improved deflation of nonresidential buildings, which reduces prices in most years during this period, contribute to a higher change. The change in residential investment is revised up 0.6 percentage point due to current-dollar revisions.

In real net exports, the change in exports is revised up slightly, and that in imports is revised down substantially. For exports, the negative impact of rebasing is more than offset by an upward revision due to the incorporation of the new price index for computers. The change in imports is revised down 1.1 percentage points due to the effects of rebasing. The large rebasing impact reflects the petroleum and petroleum products component. (As as result of these revisions, the average annual rate of change in real exports is slightly more than that in real imports over the 1972-84 period; it was 1 percentage point less than the change in imports in the previously published estimates.)

Revisions in the change in real government purchases are small. The change in Federal defense purchases is revised down due to the negative impact of rebasing. The change in Federal nondefense purchases is revised up due to current-dollar revisions and to the introduction of several new price indexes, including that for computers. The change in State and local purchases is not revised.

Price estimates

Table 20 provides, in the last two columns, a decomposition of the revision in the average annual rate of change in the GNP fixed-weighted price index and in indexes for GNP components into two parts: (1) the part that is due to shifting the base period, and (2) the part that is due to other revisions, including revisions in price indexes and revisions in current-dollar component detail that, in effect, reweight the detailed price indexes.

About one-half of the downward revision of 0.7 percentage point in the average annual rate of increase over the 1972-84 period in the GNP fixed-weighted price index is attributable to the shift in the base period; the remainder is due to other revisions, mainly the new price index for computers. The average annual rates of increase for most of the major GNP components over 1972-84 are revised down. In general, shifting the base period and the other revisions both reduced the changes.

The change in PCE prices is revised down 0.2 percentage point, from 7.2 to 7.0 percent. The changes in durables, nondurables, and services are also revised down. These revisions are largely due to the shift in the base period.

The downward revision in the change in fixed investment prices is more than accounted for by a sharp downward revision in the change in PDE prices from 6.9 percent to 1.7 percent. Some of the downward revision of 5.2 percentage points is attributable to the shift in the base period; however, most of it is due to the in-corporation of the new price index for computers. The new index declined at an average annual rate of 14 percent from 1972 to 1984; in the previously published estimates, computer prices had been assumed constant. Nonresidential structures is the only major component of GNP for which the change in prices is revised up substantially. The revision is due to the atypical relationship between changes in prices and quantities for petroleum exploration.

In net exports, the change in both export and import prices is revised down substantially. For both, the other revision is largely due to the new price index for computers, which is used to deflate the business machines categories. In addition, for imports the impact of rebasing contributes substantially.

Revisions in the change in the prices paid by government largely refect the incorporation in Federal purchases of the new price index for computers. Revisions in the prices paid by State and local governments are small.

Business Cycles

Changes in real GNP during periods of contraction and expansion since 1960 are shown in table 21. Although the revisions do not alter the peak and trough quarters for these business cycles, and thus do not change their duration, they do affect the size of the declines and increases in each period. The shift in the base period is a major factor in the revisions for almost every cycle.

For the five periods of contraction, the revised estimates show steeper declines for three. The change in real GNP that is revised down the most is the contraction of 1981-82, where the previous decline of 3.0 percent per quarter at an annual rate is revised to 3.4 percent. The change that is revised up the most is the 1973-75 contraction, where the decline previously was 3.9 percent per quarter at an annual rate and is now 3.5 percent. Thus, in the revised estimates, the declines in both the 1973-75 and 1981-82 contractions are of about the same steepness, whereas in the previously published estimates, the 1973-75 contraction was considerably steeper.

As shown in table 22, the shift in the base period is the primary source of the revisions in both contractions; it reduces the decline in the 1973-75 contraction 0.4 percentage point per quarter at an annual rate and increases the decline in the 1981-82 contraction 0.7 percentage point per quarter. Because the constant-dollar estimates for all periods are measured using the same base period, comparisons of the recent cycles with earlier ones using the revised estimates should be made with caution. For many purposes, an up-to-date base period is considered to be most appropriate to the measurement of constant-dollar output in the most recent periods. Thus, the 1982 base period may be more appropriate in measuring the steepness of the 1981-82 contraction, and for a similar reason, the 1972 base period may be more appropriate in measuring the 1973-75 contraction. Removing the impact of the shift in the base period from the 1973-75 contraction by subtracting the 0.4 percentage point per quarter referred to eariler results in a decline of 3.9 percent per quarter--measured in 1972 dollars--compared with the 3.4-percent decline of the 1981-82 contraction--measured in 1982 dollars. This comparison indicates that the 1973-75 contraction is steeper than the 1981-82 contraction.

Table 22 also shows the importance of the revisions in the current-dollar estimates in each cyclical fluctuation. With the exception of the contraction that consists only of the one-quarter drop in the second quarter of 1980, the impact of the current-dollar revisions is upward but small. The other sources of revisions--e.g., revisions in prices--are available only for the last three contractions and also are small.

For the five periods of expansion, the revised estimates show less vigorous increases, with the expansions of 1970-73 and 1980-81 showing the largest revisions. The shift in base period again is a major source of the downward revisions for expansions after 1972 and is the most likely source for the revisions in the two earlier expansions. In addition, in the 1980-81 and present expansions, downward revisions in the current-dollar estimates also are a major source. Finally, for the present expansion, the effects of rebasing and of current-dollar revisions are partly offset by revised price estimates.

During the present expansion, the revisions in the quarterly rates of change in real GNP are unusually large (chart 4). From the third quarter of 1983 to the second quarter of 1985, much of these revisions is due to the correction of the time of recording U.S. merchandise exports and imports. In the previously published estimates, merchandise exports and imports were recorded in the month the documents were processed, which often lagged the actual month of departure or entry. Over longer time-spans--such as several quarters--the revisions to correct monthly timing tend to even out. The two largest quarterly revisions in the annual rate of increase in real GNP are in the fourth quarter of 1984 and the first quarter of 1985. In the fourth quarter, the increase is revised down 3.7 percentage points, from 4.3 to 0.6 percent. About two-thirds of the downward revision is accounted for by revisions in current-dollar GNP, which, in turn, are more than accounted for by net exports. In the first quarter, the increase is revised up 3.4 percentage points, from 0.3 to 3.7 percent. More than one-third of the upward revision is accounted for by revisions in current-dollar GNP, which are far more than accounted for by net exports.

To reflect the definitional changes and to make the set of NIPA tables more informative, several tables and a number of new items are added. The new tables and new items are summarized in table 23 of this article. The tables continue to be arranged in nine sections:

1. National Product and Income

2. Peronsal Income and Outlays

3. Government Receipts and Expenditures

4. Foreign Transactions

5. Saving and Investment

6. Product, Income, and Employment by Industry

7. Fixed-Weighted Price Indexes and Implicit Price Deflators

8. Supplementary Tables

9. Seasonally Unadjusted Estimates

Section 7 was substantially redesigned to include more tables of fixed-weighted price indexes and fewer of implicit price deflators. Changes in fixed-weighted price indexes reflect only price change, whereas changes in implicit price deflators reflect not only price changes, but also changes in the composition of GNP and, as a result, can give misleading signals of price change. (For an explanation of why deflators can give these signals, see the note in the May 1985 SURVEY, p. 6.) In most cases, fixed-weighted price indexes for goods, services, and structures have replaced implicit price deflators. (In all cases, the deflators can be calculated from published current- and constant-dollar estimates.) For GNP and several major components, the initial year of the fixed-weighted indexes is 1959, because information for the components for prior years is not available in detail consistent with that for 1982. In addition, there are three new tables: fixed-weighted price indexes for GNP by major type of product (table 7.2); fixed-weighted price indexes for the relation of GNP, gross domestic purchases, and final sales to domestic purchasers (table 7.3); and implicit price deflators for command-basis GNP (table 7.8).

The last two tables are associated with new tables in section 1. The relation of GNP, gross domestic purchases, and final sales to domestic purchasers (table 1.5--current dollars; table 1.6--constant dollars) is designed to bring out the relationship between GNP, a measure of production, and two aggregates that can be viewed as measures of aggregate demand: gross domestic purchases and final sales to domestic purchasers. The increased role of exports and imports in the economy has led to more interest in these measures of demand. Command-basis GNP (table 1.11) is a measure adjusted for changes in the terms of trade. It had been introduced in the May 1981 SURVEY and had since been shown quarterly as a special table.
COPYRIGHT 1985 U.S. Government Printing Office
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1985 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:includes methodological discussion, tables of revised estimates, sources of revisions, note on shifting base period and producers' durable equipment revisions, and NIPA tables
Publication:Survey of Current Business
Date:Dec 1, 1985
Words:7314
Previous Article:U.S. affiliates of foreign companies: operations in 1983.
Next Article:Plant and equipment expenditures.
Topics:


Related Articles
An advance overview of the comprehensive revision of the national income and product accounts.
Selected national income and product estimates, 1929-85.
The cyclically adjusted federal budget and federal debt: revised and updated estimates.
National income and products accounts tables.
Gross product by industry: revised and updated estimates, 1947-85.
The U.S. national income and product accounts: revised estimates.
GNP: an overview of source data and estimating methods.
The U.S. national income and product accounts: revised estimates; annual 1986-88 and quarterly 1986:I - 1989:I.
Annual input-output accounts of the U.S. economy, 1986.
The comprehensive revision of the U.S. national income and product accounts: a review of revisions and major statistical changes.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters