Printer Friendly

The comprehensive revision of the U.S. national income and product accounts: a review of revisions and major statistical changes.

In early December, the Bureau of Economic Analysis (BEA) released results of a comprehensive--that is, benchmark--revision of the national income and product accounts (NIPA's). This comprehensive revision is the ninth of its kind; the last such revision was released in December 1985. In this revision, most current-dollar series are revised back to 1978, and many are revised back to earlier years; all constant-dollar series are revised back to 1929. (For information on the availability of the revised estimates, see the box on page 26.)

Comprehensive revisions differ from annual NIPA revisions because of the scope of the changes incorporated and because of the number of years subject to revision. (Annual NIPA revisions are typically carried back 3 years to incorporate new and revised source data from a variety of annual government surveys and administrative records and reports.) Comprehensive revisions incorporate three kinds of changes: Definitional and classificational changes, statistical changes, and new and redesigned tables. The definitional and classificational changes incorporated in this comprehensive revision were described in an article in the September 1991 SURVEY OF CURRENT BUSINESS, the major statistical changes are reviewed in a section of this article beginning on page 34, and the new and redesigned tables were presented in an article in the October 1991 SURVEY.

The incorporation of definitional and classificational changes (hereafter referred to as "definitional changes") and of statistical changes leads to revisions in current-dollar estimates and in estimates of prices. In turn, these revisions lead to revisions in constant-dollar estimates. The first three sections of this article describe the impact of the definitional and statistical changes on the current-dollar annual estimates for 1959-90, the constant-dollar annual estimates for 1977-90, and the quarterly estimates for 1988:I-1991:III. The final section reviews the major statistical changes incorporated in this comprehensive revision.

Annual Revisions in Current

Dollars

As shown in table 1, the current-dollar revisions in level for major aggregates are generally upward. Prior to 1977, the revisions are mostly small, and they are largely traceable to definitional changes. The largest revisions result from the change that redefines government sales and nontaxes. This change lowers both State and local government purchases and personal tax and nontax payments and raises personal consumption expenditures for services; it has little effect on gross domestic product (GDP) and gross national product (GNP) because the revisions are largely offsetting. Two other definitional changes result in noticeable revisions--amounting to $2.0 billion or more--prior to 1977. The reclassification of interest paid by the Federal Government to foreigners raises payments of factor income, thus lowering GNP but not affecting GDP, and lowers net interest, thus lowering
 Table 1.--Revisions in Gross Domestic Product,
 Gross National Product, National Income, and
 Personal Income
 [Billions of dollars]
 1959 1977 1982 1987 1990
Gross domestic product:
 Total revision 1.5 9.0 34.8 53.2 90.4
 Definitional 1.3 .9 6.0 10.0 5.6
 Statistical .2 8.1 28.8 43.2 84.8
Gross national product:
 Total revision 1.2 4.1 13.8 28.9 59.4
 Definitional 1.0 -4.0 -12.7 -14.9 -31.9
 Statistical .2 8.1 26.5 43.8 91.3
National income:
 Total revision .9 -8.7 4.1 32.0 41.2
 Definitional .8 -4.1 -41.6 -16.5 -43.5
 Statistical .1 -4.6 25.7 48.5 84.7
Personal income:
 Total revision .5 -6.2 20.1 35.6 34.3
 Definitional .3 -3.3 -2.1 -4.5 -19.3
 Statistical .2 -2.9 22.2 40.1 53.6


national income. The reclassification of bad debt losses as financial transactions raises corporate profits and lowers proprietors' income, rental income of persons, and business transfer payments. This change has little effect on GDP and GNP; it raises national income and lowers personal income. One statistical change, the introduction of a new procedure for estimating construction of private nonresidential structures and of State and local government nonresidential structures, results in upward revisions back to 1970, raising both GDP and GNP. Another statistical change, the introduction of a new methodology for estimating imputed rental income of farm dwellings, lowers personal consumption expenditures (PCE) and farm proprietors' income back to 1973; this change lowers GDP, GNP, national income, and personal income.

The remainder of the discussion of the revisions in the NIPA estimates focuses on 1977-90, the period in which the revisions are concentrated. The dollar revisions for 1977-90 in the components of the five summary NIPA's are shown in appendix A. Revisions in the major components of these accounts are discussed in some detail in the following sections.

National income and product

GDP in current dollars is revised up for all years from 1977 through 1990. The percent revision in the level of GDP ranges from 0.5 percent for 1977 to 1.7 percent for 1990. Statistical changes account for most of the revision; the change that has the largest impact is the introduction of the new procedure for estimating nonresidential structures. Other statistical changes that contribute to the upward revisions are the incorporation of the results of the 1982 benchmark input-output (I-O) accounts (largely into the estimates of producers' durable equipment), the annual Survey of Manufactures for 1988 and for 1989 (change in business inventories), and the Annual Retail Trade Survey for 1988 and for 1989 (PCE). Definitional changes have their greatest impact in 1986, when they account for about one-fifth of the total revision. the changes that contribute most to the upward revisions in GDP are (1) the change in imputed brokerage charges (PCE and State and local government purchases), (2) the reclassification of nine Federal Government agencies (Federal nondefense purchases), (3) the reclassification of capital gains distributions (PCE), and (4) the capitalization of monetary interest on own-account construction (nonresidential structures).

GNP is revised up for all years. The percent revision in the level of GNP is less than that for GDP, ranging from 0.2 percent for 1977 to 1.1 percent for 1990. (In the last comprehensive revision, the largest percent revision in the level of GNP was 4.0 percent.) In general, the impacts of the definitional and statistical changes on GNP are similar to those on GDP. Two exceptions are definitional changes that affect receipts and payments of factor income: The reclassification of interest paid by the Federal Government to foreigners as a payment of factor income (which lowers GNP); and the recording of nonresident taxes as transfer payments, which raises both receipts and payments of factor income (because receipts are raised more than payments, GNP is also raised).

National income is revised down for 1977--81 and up for 1982--90. The largest percent revision in the level of national income is 0.9 percent in 1990. (In the last comprehensive revision, the largest percent revision in the level of national income was 4.4 percent.) The downward revisions for 1977--81 largely result from the definitional change that reclassifies interest paid by Federal Government to foreigners. Beginning with 1982, substantial upward statistical revisions more than offset downward definitional revisions. For all or part of 1982-87, the following are the major statistical changes that result in upward revisions: The improved adjustments for misreporting on tax returns that are used in estimating nonfarm proprietors' income and wages and salaries, information from the 1982 I-O accounts that lowers rental expenses and thus raises rental income of persons, information from a trade source that increases the employer's share of health insurance premiums, and a variety of improvements in the procedures for estimating net interest. For 1988-90, the upward revisions also reflect the incorporation of tax return sources for the estimates of corporate profits and net interest and of Bureau of Labor Statistics (BLS) tabulations of wages and salaries covered by State unemployment insurance.

Product-side components

Personal consumption expenditures (PCE) for goods.--PCE for goods is revised down for 1977-88 and up for 1989 and 1990 (table 2). These revisions are largely accounted for by a number of statistical changes.

PCE for durable goods is revised down for all years; the revisions are primarily in motor vehicles, mainly in purchases of used autos and of "tires, tubes, accessories, and other parts." The downward revisions in used auto purchases by consumers reflect an updating of the depreciation schedule used to determine the constant-dollar value and, prior to 1984, the carrying
 Table 2.--Revisions in Personal Consumption
 Expenditures
 [Billions of dollars]
 1977 1982 1987 1990
Personal consumption expenditures:
 Total revision 14.3 8.5 42.8 85.3
 By source of revision
 Definitional 17.4 37.8 57.8 74.6
 Redefine government sales and
 personal nontaxes 17.2 34.8 53.0 71.0
 Reclassify bad debt losses as
 financial transactions -.2 2.1 -1.5 .3
 Reclassify capital gains
 distributions .2 .1 2.1 .3
 Remove capital gains from
 brokerage charges .2 .9 4.1 2.9
 Statistical -3.1 -29.3 -15.0 10.7
 By component
 Durable goods -1.9 -16.2 -19.7 -14.4
 Nondurable goods 0 1.3 9.8 24.0
 Services 16.1 23.4 52.7 75.7


back of an adjustment for quality change in the valuation of used autos in current dollars. Beginning with 1981, expenditures for "tires, tubes, accessories, and other parts" is revised down substantially; the revisions result from information from the 1982 I-O accounts on the consumer share of purchases of these items.

PCE for nondurable goods is revised up beginning with 1980; the revisions for recent years are sizable. Much of the revision is traceable to upward revisions in expenditures for gasoline and oil, reflecting newly available information from the 1982 I-O accounts, the 1987 Census of Retail Trade, and the 1987 Truck Inventory and Use Survey. Within nondurables, new information from the 1982 I-O accounts and the 1987 Census of Retail Trade on the product composition of sales by retail stores leads to substantial downward revisions in expenditures for food and substantial upward revisions in expenditures for "other nondurables." Sizable upward revisions in "other nondurables" also reflect the incorporation of data from the 1988 and the 1989 Annual Retail Trade Survey. Beginning with 1978, expenditures for fuel oil and coal are revised down, reflecting information from the 1982 I-O accounts on the consumer share of purchases.

PCE for services.--PCE for services is revised up substantially for all years. These revisions are largely accounted for by the definitional change that redefines government sales and nontaxes; this change adds expenditures for certain services provided by government--primarily tuition payments for higher education and charges for medical care--to PCE for services. Statistical changes result in sizable upward and downward revisions in a number of services components. Data from the 1982 I-O accounts result in upward revisions for most years--especially since 1982--in transportation services, services furnished without payment by financial intermediaries, expense of handling life insurance, and religious and welfare activities. Housing service is revised down for all years for both nonfarm and farm dwelings: The revisions in nonfarm housing services mainly reflect newly available data on units from the 1990 Census of Housing, and the revisions in farm housing services reflect the new methodology for estimating imputed rental value of farm dwellings. Among other services, professional medical services is revised down to reflect the elimination of a double-counting of physicians' fees included in hospital charges, and personal care services is revised down to reflect data from the 1982 I-O accounts; net foreign travel is revised down to reflect data from the 1982 I-O accounts and, for years prior to 1986, to carry back several balance-of-payments-accounts revisions that had been introduced in the 1987 and 1989 annual NIPA revisions.

Nonresidential structures.--Nonresidential structures is revised up substantially for all years (table 3). These revisions largely reflect the introduction of the new procedure for estimating construction. Within structures, the upward revisions are concentrated in industrial and commercial buildings.

Nonresidential producers' durable equipment (PDE).--Nonresidential PDE is revised up for all years. The revisions are largely in information processing and related equipment, mainly in computers and in communication equipment, and reflect information from the 1982 I-O accounts. Within transportation and related equipment, net sales of used autos to consumers by business--a negative entry in PDE--is
 Table 3.--Revisions in Gross Private Domestic
 Investment
 [Billions of dollars]
 1977 1982 1987 1990
Gross private domestic investment:
 Total revision 14.2 56.1 49.8 61.6
 By source of revision
 Definitional 4.0 9.0 -2.2 -.1
 Reclassify CCC commodity loans 3.5 7.1 -3.2 -.8
 Capitalize monetary interest on
 own-account construction .6 1.9 1.0 .7
 Statistical 10.2 47.1 52.0 61.7
 By component
 Fixed investment 10.7 47.5 51.8 56.6
 Nonresidential 9.0 47.0 52.9 56.6
 Nonresidential 9.0 47.0 52.9 62.9
 Structures 7.9 35.5 37.6 51.7
 Producers' durable equipment 1.3 11.5 15.3 11.2
 Residential 1.7 .6 -1.1 -6.3
 Change in business inventories 3.4 8.6 -2.0 5.0
 Nonfarm 0 1.4 .4 5.4
 Farm 3.4 7.2 -2.5 -.4
 CCC Commodity Credit Corporation


revised up (see the entry on used car purchases in PCE goods), and aircraft is revised down, reflecting information from the 1982 I-O accounts and the Annual Survey of Manufactures for 1988 and for 1989.

Residential investment.--Except for 1990, the revisions in residential investment are small. A downward revision for 1990, which is primarily in single-family construction, reflects the incorporation of revised data from the Census Bureau on the value of new construction put in place.

Change in business inventories.--Revisions in change in business inventories vary from year to year in terms of direction and size. For all years, revisions in the change in farm inventories largely reflect the reclassification of commodity loans made by the Commodity Credit Corporation (CCC). Within nonfarm inventories, the change in manufacturing inventories is revised up for 1988-90, reflecting newly available information from the Annual Survey of Manufactures for 1988 and for 1989; the change in "other" inventories is revised down for 1988 and 1989, reflecting the incorporation of Internal Revenue Service (IRS) tabulations of inventory book values from corporate tax returns for 1988.

Net exports of goods and services.--As part of this comprehensive revision, GDP replaces GNP as the primary production aggregate; this replacement leads to a redefinition of net exports of goods and services to exclude net receipts of factor income. (1) (Exports

[TABULAR DATA OMITTED]

is redefined to exclude receipts of factor income, and imports is redefined to exclude payments of factor income.)

Revisions in net exports of goods and services (as redefined) are generally small (table 4). For all years, both exports and imports are revised up; the revisions largely result from the definitional change that records exports and imports of services on a gross basis and from the statistical change related to territorial adjustment, which affects both goods and services. Statistical revisions also reflect the incorporation of the latest balance of payments accounts; these revisions are mainly in travel and passenger fares, beginning with 1984, and in foreign students' expenditures, beginning with 1981.

The addenda to table 4 show the revisions in net receipts of factor income. The downward revisions for all years reflect larger upward revisions in payments of factor income than in receipts of factor income. The upward revisions in payments of factor income largely result from two definitional changes: The reclassification of interest paid by the Federal Government to foreigners, and the recording of payments of interest
 Table 5.--Revisions in Government Purchases
 [Billions of dollars]
 1977 1982 1987 1990
Government purchases:
 Total revision -19.3 -34.1 -39.9 -55.2
 By source of revision
 Definitional -20.5 -40.8 -45.6 -68.9
 Reclassify nine Federal
 Government agencies .2 1.6 4.2 2.0
 Redefine government sales and
 personal nontaxes -17.4 -35.7 -54.6 -73.2
 Reclassify CCC commodity loans -3.5 -7.1 3.2 .8
 Remove capital gains from
 brokerage charges .1 .4 1.6 1.5
 Statistical 1.2 6.7 5.7 13.7
 By component
 Federal -3.2 -6.1 3.6 .9
 National defense 0 0 -2.5 -.2
 Nondefense -3.3 -6.2 6.2 1.1
 State and local -16.1 -27.9 -43.6 -56.1
 CCC Commodity Credit Corporation


on a gross basis. The upward revisions in receipts of factor income largely result from the recording of interest on a gross basis.

Government purchases.--Government purchases are revised down substantially for all years (table 5). The downward revisions are accounted for by the redefinition of government sales and personal nontaxes, which substantially lowers State and local government purchases of services for all years. These downward revisions are partly offset by upward statistical revisions in State and local government purchases of structures, reflecting the new procedure for estimating construction, and compensation of employees. Prior to 1990, the revisions in employee compensation largely reflect the inclusion of the wages and salaries of students working for public educational institutions and of other State and local government employees not covered by unemployment insurance; for 1990, they largely reflect BLS tabulations of wages and salaries covered by State unemployment insurance.

Revisions in Federal Government purchases are generally small except for years when the reclassification of CCC commodity loans results in sizable upward or downward revisions in Federal nondefense purchases.

Income-side components

Compensation of employees.--Compensation of employees is revised up for all years, substantially for 1990 (table 6). Upward revisions in wages and salaries are mainly attributable to the following statistical changes: The improved adjustments for misreporting on tax returns; the inclusion of wages and salaries of students working for
 Table 6.--Revisions in Compensation of Employees
 [Billions of dollars]
 1977 1982 1987 1990
Compensation of employees:
 Total revision 0.7 9.0 12.3 46.1
 By source of revision
 Definitional 0 0 .2 .6
 Reclassify nine Federal
 Government agencies 0 0 .2 .6
 Record nonresident taxes as
 transfer payments 0 0 0 0
 Statistical .7 9.0 12.1 45.5
 By component
 Wage and salary dsbursements 0 7.2 11.5 33.6
 Supplements to wages and salaries .7 1.8 .8 12.5
 Employer contributions for social
 Insurance .1 0 -.1 -3.5
 Other labor income .6 1.8 1.0 15.9


public educational institutions and of other State and local government employees not covered by unemployment insurance; and, for 1990, to BLS tabulations of wages and salaries covered by State unemployment insurance.

Upward revisions in supplements to wages and salaries are more than accounted for by upward statistical revisions in other labor income. Within other labor income, revisions in health insurance are attributable to revised estimates of the employer's share of these contributions derived from information from the U.S. Chamber of Commerce and BLS; revisions in pensions are attributable to IRS tabulations of data on contributions from corporate tax returns for 1988.

Proprietors' income with inventory valuation adjustment (IVA) and capital consumption adjustment (CCAdj).--Proprietors' income with IVA and CCAdj is revised down for all years except 1984-86 (table 7). Farm proprietors' income is revised down for all years; the revisions reflect both definitional and statistical changes. The definitional changes are the reclassification of CCC commodity loans and the elimination of the defaulters' gain associated with the reclassification of bad debt losses as financial transactions. The statistical changes are the introduction of a new procedure for estimating imputed rental value of farm dwellings and the incorporation of revised estimates of production expenses from the Department of Agriculture based on the 1987 Census of Agriculture and on the 1988 Agricultural Economics and Land Ownership Survey.

Nonfarm proprietors' income is revised down for 1977-79, up for 1980-86, and down for 1987-90. The elimination of the defaulters' gain associated with bad debt losses results in
 Table 7.--Revisions in Proprietors' Income With
 Inventory Valuation and Capital Consumption
 Adjustments
 [Billions of dollars]
 1977 1982 1987 1990
Proprietors' income with inventory
 valuation and capital
 consumption adjustments:
 Total revision -6.5 -4.8 -13.2 -29.3
 By source of revision
 Definitional -1.3 -4.4 -13.5 -12.6
 Reclassify CCC commodity loans -.1 -1.2 0 -.1
 Reclassify bad debt losses as
 financial transactions -1.3 -3.5 -13.4 -12.4
 Capitalize monetary interest on
 own-account construction .1 .3 -.1 -.1
 Statistical -5.2 -.4 .3 -16.7
 By component
 Farm -3.4 -11.1 -11.5 -7.4
 Proprietors' income with
 inventory valuation adjustment -3.3 -10.8 -11.2 -7.2
 Capital consumption adjustment -.2 -.4 -.3 -.3
 Nonfarm -3.0 6.4 -1.6 -21.9
 Proprietors' income -2.6 9.3 5.3 -15.4
 Inventory valuation adjustment 0 -.1 .2 .9
 Capital consumption adjustment -.4 -2.9 -7.1 -7.3
 CCC Commodity Credit Corporation


downward revisions for all years, substantially beginning with 1987. Statistical revisions are upward through 1987, as the improved adjustments for misreporting on tax returns more than offset downward revisions from two sources. The first of these sources is revisions in the CCAdj, which reflect the incorporation of the revised estimates of fixed investment into the estimates of the consumption of fixed capital (see consumption of fixed capital for additional detail); the second source is the carrying back and correction of an adjustment, introduced in the July 1988 annual NIPA revision, for interest paid by partnerships that is passed through to partners, who, in turn, treat it as an expense on their own tax returns. Beginning in 1988, statistical revisions turn negative; the upward revisions due to the misreporting adjustments become smaller, and they are more than offset by downward revisions stemming from the other two sources. The incorporation of IRS tabulations of noncorporate business tax returns for 1989 also contributes to the downward revisions.

Rental income of persons with CCAdj.--Rental income of persons with CCAdj is revised up for 1979-85 and down for 1986-90 (table 8). The revisions are largely statistical. The upward revisions for 1979-85 reflect information from the 1982 I-O accounts that lowers expenses and thus raises rental income of persons. The downward revisions for 1986-90 reflect both lower estimates of rental receipts and higher estimates of interest expenses. The
 Table 8.--Revisions in Rental Income of Persons
 With Capital Consumption Adjustment
 [Billions of dollars]
 1977 1982 1987 1990
Rental income of persons with
 capital consumption adjustment:
 Total revision 0.8 8.3 -10.5 -19.8
 By source of revision
 Definitional -.7 -.5 -2.0 -3.2
 Reclassify bad debts losses as
 financial transactions -.7 -.5 -2.0 -3.2
 Statistical 1.5 8.8 -8.0 -16.6
 By component
 Rental income of persons 1.2 10.0 -11.6 -21.0
 Capital consumption adjustment -.4 -1.6 1.1 1.3
 Table 9.--Revisions in Cororate Profits With
 Inventory Valuation and Capital Consumption
 Adjustments
 [Billions of dollars]
 1977 1982 1987 1990
Corporate profits with inventory
 valuation and capital
 consumption adjustments:
 Total revision 0.9 1.5 11.5 20.7
 By source of revision
 Definitional 3.5 -.2 21.7 9.0
 Record nonresident taxes as
 transfer payments -.8 -3.1 -2.9 -4.3
 Reclassify bad debt losses as
 financial transactions 4.6 4.9 27.4 19.0
 Remove capital gains from
 brokerage charges -.8 -3.2 -2.5 -5.0
 Capitalize monetary interest on
 own-account construction .5 1.3 -.4 -.5
 Recognize capital consumption
 for abandoned nuclear power
 plants 0 -.2 0 0
 Recognize personal-injury trusts
 as corporate businesses 0 0 0 -.2
 Statistical -4.4 1.7 -10.2 11.7
 By component
 Profits before tax 2.9 6.7 12.6 27.6
 Inventory valuation adjustment 0 .5 4.9 -2.8
 Capital consumption adjustment -2.0 -5.7 -6.0 -4.1


lower receipts are based on data on the number of units from the 1990 Census of Housing, and the higher expenses reflect a new procedure for handling refinanced mortgages.

Corporate profits with IVA and CCAdj.--Revisions in corporate profits with IVA and CCAdj are small except for downward revisions for 1981 and 1986 and substantial upward revisions for 1987-90 (table 9). The downward revisions are largely attributable to statistical revisions in the CCAdj, which is revised down for all years (see consumption of fixed capital for a description of the major sources of these revisions). The upward revision for 1987 is more than accounted for by the reclassification of bad debt losses as financial transactions. The upward revisions for 1988-90 are accounted for by this reclassification and by statistical revisions that largely reflect the incorporation of IRS tabulations of corporate tax return data for
 Table 10.--Revisions in Net Interest
 [Billions of dollars]
 1977 1982 1987 1990
Net interest: Total revision -4.6 -9.9 31.8 23.4
 By source of revision
 Definitional -5.5 -16.5 -22.8 -37.2
 Reclassify interest paid by
 Federal Government to
 foreigners -5.5 -19.3 -25.3 -37.9
 Record nonresident taxes as
 transfer payments 0 .1 .1 .3
 Reclassify bad debt losses as
 financial transactions -.2 2.4 -.3 -.1
 Reclassify capital gains
 distributions .2 .2 2.3 .4
 Remove capital gains from
 brokerage charges .1 .1 .3 .2
 Statistical .9 6.6 54.6 60.6
 By component
 Net monetary interest -4.5 -4.3 11.9 -9.6
 Net imputed interest -5.6 20.0 33.0


1988 into the estimates of profits before tax. By industry, the upward revisions for 1987-90 are largely in profits of financial institutions, reflecting the reclassification of bad debt losses.

Net interest.--Net interest is revised down for 1977-83 and up for 1984-90 (table 10). The downward revisions are attributable to the reclassification of interest paid by Federal Government to foreigners. Downward definitional revisions continue through 1990; although these revisions become large, they are more than offset by even larger upward statistical revisions. For 1982-87, the statistical revisions are largely accounted for by the following changes: (1) Newly available IRS tabulations of interest paid by partnerships for residential rental property, (2) improvements in the adjustment to interest reported on partnership tax return data to account for interest passed through to the individual partners (see proprietors' income), (3) improvements in the procedure used to convert data from federally insured banks to represent all commercial banks in estimating imputed interest paid by them, (4) improved adjustments for interest misreported as business receipts by credit agencies and captive finance companies, (5) revised estimates for 1987 of income earned by regulated investment companies, (6) newly available IRS tabulations of interest paid by several types of nonprofit institutions, and (7) revised estimates of mortgage interest (see rental income of persons). Beginning with 1988, upward revisions also reflect the incorporation of IRS tabulations of corporate tax returns for 1988 and the incorporation of data from several other annual sources for 1989 and 1990.

Table 11.--Revisions in Indirect Business Tax and Nontax Liability [Billions of dollars]
 1977 1982 1987 1990
Indirect busines tax and montax
 liability: Total revision -0.2 -2.4 -2.8 -1.1
 By source of revision
 Definitional .6 .6 3.8 4.6
 Reclassify nine Federal
 Government agencies .5 1.6 3.8 4.6
 Recognize court-mandated
 escrow accounts .2 0 0 0
 Statistical -.8 -4.0 -6.6 -5.7


Table 12.--Revisions in Business Transfer Payments [Billions of dollars]
 1977 1982 1987 1990
Business transfer payments:
 Total revision -1.3 1.1 -4.5 -7.3
 By source of revision
 Definitional -1.4 -.8 -5.1 -5.4
 Reclassify nine Federal
 Government agencies 0 0 -.2 -.2
 Record nonresident taxes as
 transfer payments 1.4 3.6 3.2 4.5
 Reclassify bad debt losses as
 financial transactions -2.8 -4.4 -8.0 -10.0
 Recognize personal-injury trusts
 as corporate businesses 0 0 0 .2
 Statistical 1 1.9 .6 -1.9
 By component
 Transfers to persons -2.7 -2.5 -7.7 -11.8
 Transfers to rest of the world 1.4 3.6 3.2 4.5


Indirect business tax and nontax liability.--Indirect business tax and nontax liability is revised down for all years (table 11). The downward revisions are more than accounted for by a statistical change that transfers interest earned by State and local governments on industrial development bonds from rental receipts, which is included in nontaxes, to interest received. These downward statistical revisions are partly offset by upward revisions for all years that result from the definitional changes that involve the reclassification of nine Federal Government agencies.

Business transfer payments.--Revisions in business transfer payments are small for 1977-86; business transfer payments are revised down for 1987-90 (table 12). The downward revisions are more than accounted for by downward revisions in transfer payments to persons, which largely reflect the elimination of the defaulters' gain associated with bad debt losses. The down ward revisions in transfers to persons are partly offset by upward revisions in transfers to the rest of the world, which result from another definitional change, the recording

Table 13.--Revisions in Subsidies Less Current Surplus of Government Enterprises [Billions of dollars]
 1977 1982 1987 1990
Subsidies less current surplus of
 government enterprises:
 Total revision 0.3 -2.5 -3.3 2.3
 By source of revision
 Definitional .3 -2.7 -2.8 3.4
 Reclassify nine Federal
 Government agencies .3 -1.5 -2.8 3.4
 Reclassify CCC commodity loans -.1 -1.2 0 -.1
 Statistical 0 .2 -.5 -1.1
 By component
 Subsidies 0 0 .7 .2
 Current surplus -.3 2.4 3.9 -2.1
CCC Commodity Credit Corporation


of nonresident taxes as transfer payments.

Subsidies less the current surplus of government enterprises.--Revisions in subsidies less the current surplus of government enterprises are generally small (table 13). They largely reflect the definitional changes involving the reclassification of nine Federal Government agencies.

Consumption of fixed capital.--Consumption of fixed capital--that is, economic depreciation and allowances for accidental damage--is revised up for all years (table 14). The upward revisions largely result from the higher estimates of fixed investment.

Capital consumption allowances (CCA) is also revised up for all years, but by less than is consumption of fixed capital. Upward revisions in the CCA of corporations and of nonfarm sole proprietorships and partnerships--which are tax-return-based depreciation measures--reflect a number of statistical changes. These changes include the incorporation of corrections to IRS tabulations of corporate tax return amortization for 1982-85 and the incorporation of IRS tabulations of corporate tax return data for 1988 and of noncorporate business tax return data for 1989. Upward revisions in the CCA of nonprofit institutions serving individuals reflect a new estimating procedure that results in higher estimates of fixed investment by these institutions.

The CCAdj--which is derived as the difference between the CCA and the economic measure of depreciation--is revised down for all years. The revision is almost entirely in the adjustments for corporate profits and for nonfarm proprietors' income.

Table 14.--Revisions in Consumption of Fixed Capital [Billions of dollars]
 1977 1982 1987 1990
Consumption of fixed capital:
 Total revision 3.7 15.9 15.2 19.2
 By source of revision
 Definitional 0 1.0 .5 .6
 Capitalize monetary interest on
 own-account construction 0 .2 .5 .6
 Recognize capital consumption
 for abandoned nuclear power
 plants 0 .7 0 0
 Statistical 3.7 14.9 14.7 18.6
 By component
 Capital consumption allowances .7 5.1 2.4 8.0
 Corporate 0 3.1 -1.4 -1.7
 Noncorporate .7 2.0 3.8 9.7
 Less: Capital consumption
 adjustment -3.0 -10.8 -12.8 -11.1
 Corporate -2.0 -5.7 -6.0 -4.1
 Noncorporate -1.0 -5.1 -6.8 -7.0
Addenda:
 Consumption of fixed capital:
 Corporate 2.0 8.7 4.5 2.5
 Noncorporate 1.7 7.2 10.7 16.7


Personal income and outlays

Personal income is revised down for 1977-79, up for 1980-88, down for 1989, and up for 1990 (table 15). These revisions reflect a combination of substantial upward and downward revisions in several components. The revisions in wages and salaries, other labor income, proprietors' income, and rental income of persons were described previously. Among other components, revisions in personal interest income are particularly large.

Personal interest income is revised up for all years, substantially beginning with 1984. The upward revisions largely stem from the previously described statistical revisions in net interest. (2) These revisions are partly offset by sizable upward revisions in interest received by State and local governments, which reflect the inclusion of interest paid by businesses for industrial bonds (see indirect business taxes), and in interest received by the Federal Government, which reflect higher estimates of the interest paid by persons and by business because of late filings of income tax returns.

Personal dividend income is revised up for all years. The revisions largely result from the definitional change that recognizes capital gains distributions of regulated investment companies as dividends.

Transfer payments to persons are revised down for all years. These revisions

Table 15.--Revisions in Personal Income [Billions of Dollars]
 1977 1982 1987 1990
Personal income: Total revision -6.2 20.1 35.6 34.3
 Definitional -3.3 -2.1 -4.5 -19.3
 Statistical -2.9 22.2 40.1 53.6
 By component
 Wage and salaries, other labor
 income, proprietors' income with
 IVA and CCAdj, and rental
 income of persons with CCAdj -5.1 12.5 -11.2 .4
 Personal dividend income 1.2 3.2 8.6 1.0
 Personal interest income .3 7.1 46.8 40.9
 Transfer payments -2.8 -2.5 -7.7 -9.9
 Less: Personal contributions for
 social insurance 0 0 .7 -1.9
Less: Personal tax and montax
 payments: Total revision -18.1 -37.9 -59.1 -78.4
 Definitional -18.1 -37.1 -56.7 -75.9
 Statistical -.8 -2.4 -2.5
Equals: Disposable personal income:
 Total revision 12.0 58.2 94.8 112.7
 Definitional 14.9 35.2 52.3 56.6
 Statistical -2.9 23.0 42.5 56.1
Less: Personal outlays:
 Total revision 14.5 12.1 44.7 86.2
 Definitional 17.8 39.3 61.3 78.2
 Statistical -3.3 -27.2 -16.6 8.0
Equals: Personal saving:
 Total revision -2.6 46.1 50.1 26.5
 Definitional -3.0 -4.4 -9.1 -21.7
 Statistical .4 50.5 59.2 48.2
Addendum:
 Personal saving as a percent of
 disposable personal income -.3 1.8 1.4 .5
IVA Inventory valuation adjustment
CCAdj Capital consumption adjustment


largely result from the elimination of the defaulters' gain associated with bad debt losses.

Revisions in personal contributions for social insurance are small.

Personal tax and nontax payments.--Personal tax and nontax payments is revised down substantially for all years. These revisions are largely traceable to the redefinition of government sales and personal nontaxes, which removes payments for certain services--primarily tuition payments for higher education and charges for medical care--provided by government from personal nontaxes and adds them to PCE. Personal taxes is also lowered by the statistical change related to taxes and interest paid to the Federal Government because of late filings of tax returns.

Disposable personal income.--Disposable personal income (DPI) is revised up substantially for all years. The upward revisions in DPI reflect both the largely statistical revisions in personal income, upward except for 1977-79 and 1989, and the largely definitional revisions in personal tax and nontax payments, downward for all years.

Personal outlays.--Personal outlays is revised up substantially for all years. These revisions largely reflect the previously mentioned revisions in PCE. In addition, they also reflect a definitional change that recognizes interest payments by persons to government.

Personal saving.--Personal saving is revised down for 1977 and 1978, up for 1979-88, down for 1989, and up for 1990. The pattern of revision in personal saving reflects the revisions in both personal income and personal outlays. Beginning with 1982, the principal statistical revisions in personal income that raise personal saving are the improved adjustments for misreporting on tax returns, the revisions in personal interest income that stem from the revisions in net interest, and the revisions in employer contributions for health insurance. For 1977-87, personal saving is also raised by the downward statistical revisions in personal outlays (which is subtracted in deriving personal saving), mainly stemming from revisions in PCE for durable goods and for services. For 1988-90, personal saving is lowered by upward statistical revisions in personal outlays, mainly in PCE for nondurable goods. For all years, personal saving is lowered by the definitional change that eliminates the defaulters' gain associated with bad debt losses. Most of the other major definitional and statistical changes that affect personal income, personal taxes, DPI, or personal outlays have little effect on personal saving. For example, the redefinition of government sales and personal nontaxes raises DPI and the PCE component of personal outlays by virtually the same amount. Similarly, several of the statistical changes that affect personal income also affect the PCE component of personal outlays; the most important of these offsetting changes are the downward revisions in imputed rental receipts, which lower farm proprietors' income and rental income of persons as well as PCE for housing services, and the upward revisions in imputed interest paid by commercial banks, which raise both personal interest income and PCE for services.

The personal saving rate--personal saving as a percentage of DPI--is revised up for all years except 1977, 1978, and 1989. The largest upward revisions are for 1980-87, when the personal saving rate is raised 1-to-2 percentage points each year. In 1990, the rate is raised 0.5 percentage point to 5.1 percent.

Table 16.--Revisions in Government Receipts and Expenditures [Billions of dollars]
 1977 1982 1987 1990
 Federal
Receipts -0.1 0.1 0 -4.9
Expenditures -3.7 -10.3 -6.3 -5.6
Surplus or deficit (-), national
 income and product accounts 3.6 10.4 6.4 .7
 State and local
Receipts -18.2 -40.4 -60.9 -76.0
 Of which: Redefine government
 sales and personal nontaxes -17.4 -35.7 -54.3 -72.8
Expenditures -16.9 -32.2 -49.9 -66.3
 Of which: Redefine government
 sales and personal nontaxes -17.4 -35.7 -54.3 -72.8
Surplus or deficit (-), national
 income and product accounts -1.3 -8.2 -10.9 -9.7


Government receipts and expenditures

Federal.--Revisions in Federal Government receipts are small for all years except 1990 (table 16). For most years, personal income tax receipts are revised down, reflecting the definitional change that records nonresident taxes as transfer payments and the statistical change related to taxes and interest paid by persons because of late filings of tax returns. (In the NIPA's, interest received by government is recorded on the expenditures side of the government account as an offset to interest paid.) The downward revisions in personal taxes tend to be offset by upward revisions in indirect business nontax accruals, reflecting the definitional changes related to the reclassification of nine Federal Government agencies and a statistical change related to the valuation of purchases of used assets. For 1989 and 1990, corporate profits tax accruals are revised up, reflecting information from regular annual sources on taxes paid by liability year. For 1990, information from regular annual sources adds to the downward revisions in personal income tax receipts.

Federal Government expenditures is revised down for all years except 1983. For most years, the downward revisions reflect revisions in net interest paid. Net interest paid is revised down for all years, as interest received by the Federal Government (recorded as an offset to interest paid) is revised up, reflecting the statistical change related to interest paid by persons and business because of late filings of income tax returns. The reclassification of CCC commodity loans results in sizable revisions in nondefense purchases; these purchases are revised down considerably for 1982 and 1985 and up considerably for 1983 and 1988.

The Federal Government deficit on a NIPA basis is reduced for all years. Prior to 1984, the upward revisions are traceable to definitional changes, primarily the reclassification of CCC commodity loans and the reclassification of nine Federal Government agencies. For 1984-90, the upward revisions reflect a number of statistical changes, primarily the change related to higher interest payments by business (the higher interest payments by persons are offset by lower personal taxes) because of late filings of tax returns, the change in the revaluation of used assets, and for 1989 and 1990, the upward revision in corporate profits tax accruals. For particular years, the reclassification of CCC commodity loans also resulted in sizable revisions in the Federal deficit. (The effects of the CCC loan reclassification are largely those of timing; over long periods, revisions in the deficit due to this change will net to zero.)

State and local.--State and local government receipts are revised down substantially for all years. The revisions largely result from the redefinition of government sales and personal nontaxes, which lowers personal nontax receipts. Indirect business tax and nontax accruals is also revised down for all years, reflecting the transfer of interest earned on industrial development bonds from nontaxes to interest received (revised estimates of earnings results in adding more to interest received than is deducted from nontaxes). For 1989 and 1990, these revisions are partly offset by upward revisions in personal tax receipts and in indirect business tax and nontax liability; these upward revisions reflect Census Bureau Government Finances tabulations for fiscal year (FY) 1989 and preliminary tabulations for FY 1990 and Census Bureau quarterly tax revenue data for 1990.

State and local government expenditures are revised down substantially for all years. These revisions largely result from the redefinition of government sales and personal nontaxes, which raises government sales (a negative entry in government purchases). Net interest paid is revised down for all years, as interest received is revised up, reflecting the inclusion of interest paid by business for industrial development bonds. For 1983-89, these revisions are partly offset by the carrying back of a statistical change, introduced in the July 1988 annual NIPA revision, that excludes capital gains from interest received by State and local government retirement systems. The downward revisions are partly offset by upward revisions in purchases of structures, reflecting the introduction of the new procedure for estimating construction, and in purchases of employee compensation, reflecting the inclusion of wages and salaries of students working for public educational institutions and of other State and local government employees not covered by unemployment insurance.

The State and local government surplus on a NIPA basis is reduced for all years. The downward revisions are largely traceable to the higher estimates of purchases of structures and of employee compensation. (The reclassification of government sales and personal nontaxes results in equal, and offsetting, revisions in receipts and expenditures; the transfer of interest earned on industrial development bonds results in largely offsetting revisions.) For 1989 and 1990, the incorporation of the data from regular annual sources holds down the reduction in the surplus.

Rest-of-the-world transactions

Receipts from the rest of the world is revised up for all years (table 17). Both exports of goods and services and receipts of factor income are revised up. The upward revisions in exports largely reflect the recording of exports of services on a gross basis; the upward revisions in factor income receipts largely reflect the recording of receipts of interest on a gross basis.

Payments to the rest of the world is revised up for all years. Both imports of goods and services and payments of factor income are revised up. The upward revisions in imports largely reflect the recording of imports of services on a gross basis; the upward revisions in factor income payments reflect the reclassification of interest paid by the Federal Government to foreigners and the recording of payments of interest on a gross basis. Net transfer payments is revised up; the revisions are more than accounted for by the recording of nonresident taxes as business transfer payments to the rest of the world. The downward revisions in interest paid by the Federal Government to foreigners result from the deletion of this interest item as a separate entry.

Table 17.--Revisions in Rest-of-the-World Transactions [Billions of dollars]
 1977 1982 1987 1990
Receipts from the rest of the world 4.9 18.0 19.4 25.4
 Exports of good and services 3.3 12.3 11.0 15.0
 Receipts of factor income 1.6 5.7 8.5 10.3
 Capital grants received by the
 United States (net) 0 0 0 0
Payments to the rest of the world 4.9 18.0 19.4 25.4
 Imports of goods and services 3.4 8.1 10.4 16.4
 Payments of factor income 6.5 26.6 32.9 41.3
 Transfer payments (net) .7 2.6 1.4 3.7
 Interest paid by government to
 foreigners (1) -5.5 18.3 -25.3 -38.8
 Net foreign investment -.3 -.9 .1 2.7
 (1) In this comprehensive revision, the reclassification of
interest paid
by government to foreigners as a payment of factor income
results in
the deletion of this line.


Net foreign investment--the difference between receipts from the rest of the world and payments to the rest of the world--is revised down for 1977-86 and up for 1987-90. The revisions are small for most years. For 1986-90, the revisions are largely traceable to statistical revisions from the balance of payments accounts that were not previously incorporated into the NIPA's; these revisions are primarily in net receipts of factor income.

Gross saving and investment

Gross saving is revised up substantially for all years (table 18). The upward revisions are largely traceable to the previously described statistical revisions in personal saving and in undistributed corporate profits with IVA and CCAdj. (The upward revisions in consumption of fixed capital are largely offset by the revisions in the CCAdj for corporate profits and for proprietors' income.) The government surplus or deficit on a NIPA basis is revised down for most years, as reductions in the State and local government surplus exceed reductions in the

Table 18.--Revisions in Gross Saving and Investment [Billions of dollars]
 1977 1982 1987 1990
Gross saving 3.0 62.5 64.1 54.5
 Gross private saving .7 60.4 68.7 63.4
 Personal saving -2.6 46.1 50.1 26.5
 Undistributed corporate profits
 with IVA and CCAdj -.4 -1.6 3.3 17.7
 Consumption of fixed capital 3.6 15.8 15.1 19.2
 Government surplus or deficit(-),
 national income and product
 accounts 2.3 2.2 -4.6 -8.9
Gross Investment 14.0 55.2 49.9 64.3
 Gross private domestic investment 14.2 56.1 49.8 61.6
 Net foreign investment -.3 -.9 55.1 2.7
Statistical discrepancy 10.9 -7.3 -14.2 9.8
IVA Inventory valuation adjustment
CCAdj Capital consumption adjustment


Federal Government deficit. Most of the definitional revisions in the components of gross saving are offset in total gross saving; for example, the reclassification of bad debt losses as financial transactions lowers personal saving but raises undistributed corporate profits with IVA and CCAdj.

Gross investment is also revised up substantially for all years. The upward revisions largely result from the following statistical changes: The introduction of the new procedure for estimating nonresidential structures; the incorporation of information on purchases of new equipment from the 1982 I-O accounts; the revised estimates of net purchases of used autos in PDE; and, for 1989-90, the incorporation of regular annual source data for the change in business inventories. The definitional change in the classification of CCC commodity loans also affects the change in business inventories; it reduces gross investment for most years but contributes to the upward revisions for 1977, 1981, 1982, and 1985.

Annual Revisions in

Constant Dollars

The preceding section examined the revisions in the current-dollar estimates in terms of the differences between the revised and the previously published levels. This approach is not useful for the current-dollar estimates, because the shift in the base period changes the prices in which these series are measured. The differences in the levels of the revised and the previously published series primarily reflect this change; the effect can be seen by comparing the previously published levels of real GNP for 1990 in 1982 dollars ($4,157.3) with that in 1987 dollars ($4,846.9). (3) This section will examine the revisions in the constant-dollar estimates in terms of the differences between the revised and the previously published rates of change.

In general, revisions in aggregate constant-dollar--or real--estimates arise from three main sources. First, they reflect the revisions in the current-dollar estimates; the contribution of the current-dollar revisions is measured as the revision in percent change of the current-dollar estimates. Second, they reflect the shift in the base period from 1982 to 1987; the contribution of the base-period shift is measured as the difference between the change in the previously published estimates in 1987 dollars and the change in the previously published estimates in 1982 dollars. Third, they reflect "other" revisions that result from such sources as revisions in prices and the interaction of redistributions of current-dollar levels within components, which, in effect, reweight the constant-dollar component detail; because the contribution of the "other" revisions is obtained as a residual, the measure of this contribution depends upon the way the contributions of the first two sources are measured.

During 1977-90, the revised estimates of real GDP show an average annual growth rate of 2.5 percent, 0.2 percentage point less than was shown in the previously published estimates (table 19). The growth rate of real GNP, shown in the addendum to the table, is also revised down 0.2 percentage point to 2.5 percent. As indicated in the columns of the table showing the sources of the constant-dollar revisions,

[TABULAR DATA OMITTED]

the downward revisions in the growth rates of both real GDP and real GNP are mainly accounted for by the shift in the base period. (As discussed in the section on the major statistical changes, a shift to a more recent base period tends to reduce the measured change in real aggregates.) By major component, the downward revision in the increase in real GDP is largely accounted for by downward revisions in fixed investment and in exports of goods and services.

Revisions in components of real GDP.--Largely as a result of the rebasing, the average annual growth rate of real fixed investment during 1977-90 is revised down 0.6 percentage point to 2.1 percent. The growth rate of producers' durable equipment is revised down 1.3 percentage points to 3.5 percent; the downward revision is more than accounted for by rebasing, mainly reflecting the importance of computers in this component (the volume of computer purchases jumped from 1982 to 1987, while their price fell sharply). The growth rate of residential investment is also revised down, 0.6 percentage point to -0.2 percent; the downward revision is attributable to current-dollar revisions and to "other" revisions that largely result from methodological changes in the price estimates for single-family

[TABULAR DATA OMITTED]

houses and for multifamily houses. In contrast, the growth rate of nonresidential structures is revised up 1.8 percentage points to 2.7 percent, as the sharp upward revision in current dollars is augmented by an upward revision due to the rebasing.

As a result of the rebasing, the growth rate of exports of goods and services is revised down 1.0 percentage point to 5.7 percent. The growth rate of merchandise exports is revised down 1.4 percentage points to 5.5 percent; the downward revision is largely accounted for by the rebasing, mainly reflecting the importance of computers. The growth rate of exports of services is revised up 0.2 percentage point to 6.1 percent, mainly reflecting the current-dollar revisions.

Revisions in the growth rates of the other major components of GDP are small. Within PCE, durable goods is revised down, largely reflecting the rebasing. Within government purchases, Federal national defense purchases is revised down, largely reflecting the rebasing, and Federal nondefense purchases is revised up, reflecting current-dollar revisions. Within imports, merchandise imports is revised up, reflecting the rebasing.

Annual changes.--For 1977-90, the annual, or year-to-year, percent change in real GDP is revised down for 8 years, is revised up for 4 years, and is unrevised for 2 years (chart 1). The revisions for the 14 annual percent changes average 0.3 percentage point (without regard to sign). Changes for 3 years are revised by 0.5 percentage point or more. For 1984, the growth of real GDP is revised down 0.8 percentage point to 6.2 percent. This revision is largely in nonresidential fixed investment, reflecting the rebasing, and in Federal Government purchases, reflecting current-dollar revisions. For 1987, the growth of real GDP is revised down 0.5 percentage point to 3.1 percent. This revision is largely in nonresidential fixed investment, reflecting current-dollar revisions and the rebasing, and in exports, reflecting the rebasing and "other" revisions. For 1988, the growth of real GDP is revised down 0.5 percentage point to 3.9 percent. This revision is largely in nonresidential fixed investment and in exports, both mainly reflecting the rebasing.

Quarterly Revisions

Revisions in quarterly estimates come about in three major ways: (1) Adjustment of the estimates to reflect the annual revisions, (2) incorporation of new and revised source data (including the updating of seasonal factors) that are used to indicate quarterly patterns, and (3) changes in the methodology used to prepare the quarterly estimates. The following discussion of quarterly revisions focuses on the estimates for 1988:I-1991:III, the period that would normally have been covered by an annual NIPA reviion.

For this period, the quarter-to-quarter percent change in real GDP is revised down for nine quarters, is revised up for five quarters, and is unrevised for one quarter (chart 2 and table 20). The preponderance of downward revisions is largely attributable to the just-described revisions in the constant-dollar annual estimates. The revisions for the 15 quarterly percent changes (at annual rates) average 0.8 percentage point (without regard to sign), somewhat larger than the average for the past five annual NIPA revisions. Changes for three quarters are revised by 1.5 percentage points or more. For the first quarter of 1988, the increase in real GDP is revised down 2.0 percentage points to 2.6 percent. The downward revision is accounted for by nonresidential fixed investment, change in business inventories, exports, and State and local government purchases. (The quarterly estimates of exports--and of imports as well--are revised back to 1978, reflecting the application of seasonal adjustment factors developed jointly by BEA and the Census Bureau.) For the fourth quarter of 1988, the increase in real GDP is revised up 1.5 percentage points to 3.9 percent. The upward revision is largely accounted for by change in business inventories; PCE, nonresidential and residential fixed investment, and State and local government purchases also contribute. For the fourth quarter of 1990, the decrease in real GDP is revised down 1.8 percentage points to 3.9 percent. The downward revision is accounted for by nonresidential fixed investment, change in business inventories, and State and local government purchases.

Major Statistical Changes

This section of the article reviews the major sources of the statistical revisions in the revised estimates for 1959-90. In this comprehensive revision, the following statistical changes are incorporated:

* Shift of the base period by the substitution of the prices of 1987 for those of 1982 in calculating constant-dollar series and the substitution of the composition of output in 1987 for that in 1982 in calculating fixed-weighted price indexes;

* New and revised data from sources that are usually incorporated at the time of the annual July revisions of the NIPA's, referred to as "regular sources for 1988-90";

* New and revised data from sources that are usually incorporated at the time of comprehensive revisions because they become available less often than annually, referred to as "regular benchmark sources"; and

* Changes in methodologies--that is, new source data and/or new estimating procedures--including the use in earlier years of methodological changes introduced in the annual NIPA revisions that have taken place since the 1985 comprehensive revision.

Shift of the base period

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 1985 comprehensive revision, when the base period was shifted from 1972 to 1982. The year 1987 was selected as the new base period because it is the latest year for which the economic censuses were conducted and for which the NIPA estimates will not be revised until the next comprehensive revision. The effects of the shift on the revised constant-dollar estimates are described in a section of this article beginning on page 32; a hypothetical example that shows these effects is described in a note on page 35.

Newly available regular source data

Regular data sources for 1988-90.--Major regular data sources for 1988-90 are shown in table 21. These data would normally have been used to prepare an annual revision in July 1991; the annual revision was combined with this comprehensive revision. The effect of incorporating these sources reflects the newly available or revised data for 1988-90 as well as changes in procedures related to them.

Regular benchmark source data.--The single most important regular benchmark source is BEA's 1982 benchmark I-O tables. I-O tables provide the benchmarks for the estimates of the expenditure components of GDP and for parts of several income components. (4) In turn, the I-O tables are based primarily on the detailed industry statistics collected by the Census Bureau in the 1982 Economic Censuses and in the 1982 censuses of agriculture and governments. The tables also are based on 1982 data from the following regular sources: BEA's balance of payments accounts and perpetual-inventory-method estimates of private capital stock, U.S. Department of Agriculture (USDA) farm statistics reports, and Health Care Financing Administration reports on health insurance.

The detailed industry statistics from the economic censuses provide additional data items and more detailed information than the annual surveys that cover the same industries, and they also cover additional industries. For example, sales and receipts of retail and service establishments are reported for the types of products sold, and manufacturers' shipments are reported by more detailed type of product. The additional retail and service data provide for a more accurate distribution of products purchased by persons; the additional shipments data provide for a more accurate identification of purchases of business equipment and of related parts.

The censuses also provide additional information, such as goods and services purchased by business, for both the industries covered by annual surveys and for most of the industries--mining, construction, and nonmerchant wholesale trade--covered only in the censuses. This information--in combination with the more detailed sales, receipts, and shipments data--provides more accurate estimates of purchases of products by persons. (5) The additional industry coverage also provides more accurate estimates of inventories and, as explained later in this section, the basis for improved estimates of investment in structures.

In addition to the incorporation of estimates from the 1982 I-O tables, other regular benchmark sources incorporated into the comprehensive revision include the following. The 1987 Economic Censuses and the 1987 censuses of agriculture and governments (to the extent data were available in time); the 1990 Census of Housing; periodic audits conducted by the IRS; and periodic IRS tabulations on private pension plans and nonprofit institutions. The 1987 Economic Censuses data that are incorporated in the revision are limited to the following: Sales, receipts, and shipments for retail rtrade, service industries, and manufacturing; inventory change for mining, manufacturing, and trade; mileague and other

[TABULAR DATA OMITTED] characteristics of truck use from the truck inventory and use survey; and drilling expenses for mineral industry establishments.

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 the major methodological changes, including those needed to incorporate the definitional and classificational changes, is shown in table 22. For major changes incorporated in the annual NIPA revisions since the 1985 comprehensive revision, the table identifies the July issue of the SURVEY in which a summary description appears. Several of the most important changes made in this comprehensive revision are described in the following paragraphs.

Inputed rental value of farm dwellings.--Beginning with 1973, the methodology used to prepare current-dollar estimates of the imputed rental values of farm dwellings is changed; the change is based largely on a similar change recently introduced by USDA in their farm statistics reports. [6] The new methodology, which reduces both PCE services and farm proprietors' income by $7.9 billion for 1982, is based on the rental equivalency method of valuing the imputed rent. This method, the same as that used for nonfarm dwellings, is based on the convention that assumes that the imputed rent paid by owner occupants is the same as that paid by tenants for the rent of a similarly valued dwelling. Both the BEA and USDA estimates for 1987 are based on the total value of farm dwellings derived from the 1987 Census of Agriculture and the related 1988 Agricultural Economics and Land Ownership Survey and on the rent-to-value ratio for nonfarm housing from BEA's estimates. The two estimates differ for other periods. BEA's estimates are extrapolated backwards and forwards by the net value of the stock of farm housing at current replacement cost from BEA's capital stock series; the USDA's estimates are extrapolated by the current market value of the farm housing stock including the value of the farm land. Previously, for all periods, BEA used USDA's estimates, which had been derived as the sum of expenses related to farm housing--such as depreciation and maintenance and repair--and an estimated return to the owner equal to the market value of farm housing times the mortgage interest rate.

The methodology for the constant-dollar imputed rental values is not affected by the changes in the current-dollar estimates. The constant-dollar estimates continue to be calculated by extrapolating the base-period value by the constant-dollar net value of the farm housing stock.

Imputed commissions of domestic security dealers.--The definitional change related to imputed commissions of security dealers required the development of a procedure for eliminating capital gains (or losses) from the previously published estimates of these commissions, which were based largely on reports filed with the Securities and Exchange Commission (SEC). In the revised estimates, the commission is defined as the "spread," or the difference, between the cost of acquiring a security or an equity and its sale value, based on the acquisition cost on the date of sale. For Federal Government securities, the revised commissions are estimated using the dollar volume of trading as reported by the Federal Reserve Bank of New York and on bid and ask prices as published in the Wall Street Journal (WSJ). For equities, the revised commissions are estimated in two parts: For stocks sold over the counter, volume data are from the National Association of Securities Dealers, and bid and ask prices are from the WSJ; and for exchange-traded stocks, volume data are based on transactions on the New York and other stock exchanges and on related bid and ask prices. Imputed commissions paid by various purchasers--such as persons, businesses, and State and local government pension plans--are estimated for securities using data on holdings of these purchasers from the flow of funds accounts published by the Federal Reserve Board and for equities using data on transactions from New York Stock Exchange reports.

The definitional revision also results in the elimination of the capital gains element of the commissions from the profits of the dealers. For dealers in securities, the capital gains were derived from SEC trading accounts data by subtracting the revised commissions estimates and estimates of interest included in the SEC data. The capital gains were allocted between security dealers and bank dealers using data from the SEEC and the Federal Deposit Insurance Corporation.

Nonresidential structures.--Estimates of nonresidential structures, both privately owned and owned by State and local governments, are revised up substantially; the revisions begin with 1970 for private structures and with 1975 for State and local structures. For 1982, the revised estimate for private nonresidential structures is 24.8 percent, or $35.5 billion, higgher than the previously published estimate; the revised estimate for State and local structures is 12.1 percent, or $5.3 billion, higher. These revisions reflect the introduction of a new procedure to benchmark the estimates from the Census Bureau's monthly construction-put-in-place survey, the source of the previously published estimates of nonresidential structures. The new procedure was introduced in an effort to resolve a longstanding problem with the put-in-place survey--the need for benchmark estimates for most types of nonresidential construction.

The benchmark procedure, which provided estimates for 1977 and 1982, is based primarily on information from the census of construction industries on the value of contract construction receipts for new construction, from the census of service industries on architectural and engineering fees, and from Federal regulatory agency reports on the value of own-account new construction of utility companies. (Work is in progress on the development of similar benchmark estimates for 1987.) The procedure provided estimates of total new construction except for investment in mining exploration, for which estimates for all periods were taken from regular benchmark sources. The 1977 and 1982 estimates for the types of construction covered by the new procedure were prepared as follows. (1) The latest estimates from the put-in-place survey, with minor adjustments for definitional differences, were used for all of new residential construction and for new nonresidential construction by the Federal Government, public utilities, and agriculture. (2) The benchmark procedure estimates were used for State and local government new construction. (3) For the remainder of private new nonresidential construction, estimates by type of construction

[TABULAR DATA OMITTED]

were judgmentally prepared using both the corresponding benchmark and put-in-place estimates.

For the years between 1977 and 1982, estimates of State and local government new construction were interpolated using estimates from annual Census Bureau surveys of State and local government expenditures. The same survey data were used to extrapolate backwards from 1977 and forwards from 1982. For each type of private new construction based on the new procedure, estimates for the years between 1977 and 1982 were interpolated using the latest corresponding put-in-place estimates. The same put-in-place estimates were used to extrapolate backwards from 1977 and forwards from 1982. For the estimates after 1982, the put-in-place extrapolators were adjusted for the downward bias in growth in earlier periods. (7 )

Tax return misreporting adjustments.--Beginning with 1977, the adjustments for misreporting on tax returns are updated to reflect regularly available audit information from the 1982 and the 1985 Taxpayer Compliance Measurement Program (TCMP) and the following additional information: The results of two special audit studies--one on partnership returns for 1981 and the other on employment tax returns for 1984; the results of new research by the IRS on the extent to which the audits detect all unreported income; and a BEA assumption that recent changes in tax law on compliance by noncorporate businesses have reduced misreporting beginning with 1987. (8) For 1982, the revised misreporting adjustments added $23.3 billion to both national and personal income: Proprietors' income is raised by $17.0 billion; wages and salaries, by $5.9 billion; and rental and interest income combined, by $0.4 billion.

The special partnership audit, conducted as part of the TCMP effort, was the first comprehensive audit of the detailed partnership return Form 1065. Previously, misreporting by partnerships was assumed to be proportionately the same as that by sole proprietorships. The new study, whose results are incorporated beginning with 1977, generally showed more misreporting by partnerships. The 1984 employment tax study updates the previous adjustments, which were based on information from the 1979 TCMP. The new study provides moree accurate information because it is based on reporting of wages and salaries by all employers covered by Federal Unemployment Insurance; the TCMP study was based only on reporting of wages paid by sole proprietorships and small corporations. In the revised wage and salary estimates, the misreporting adjustments are interpolated between 1978 and 1984; for 1984 forward, the total 1984 adjustment is extrapolated using total private wages and salaries.

The new IRS research incorporated into the revised estimaters relates to that part of the adjustment that covers incomes undetected by the auditors. (9) Previously, IRS assumed that the auditors detected only $1 out of every $3-1/2 of unreported net income. The new research shows that this adjustment should be made only to underreported gross income and not to that part of underreported net income caused by the overreporting of expenses. In the revised BEA estimates, the auditor's adjustment is reduced, beginning with 1977.

The other change in procedure is a reduction, beginning with 1987, in the estimates of the amount of misreporting undetected by the auditors. This reduction is based on a BEA assumption that the compliance provisions and provisions related to the use of passive losses of the Tax Reform Act of 1986 will reduce misreporting.

Interest paid by real estate rental partnerships.--Beginning with 1981, the estimates of interest paid are revised to reflect the incorporation of newly available tabulations of interest paid b partnerships primarily engaged in the rental of real estate. Prior to 1981, interest paid by these businesses was included in the published IRS tabulations of partnership tax returns. Beginning in 1981, the partnership tax return was redesigned with regard to the reporting of rental income. As a result, the related interest payments were not tabulated for 1981; for 1982-86, they were not tabulated when the partnership returns were tabulated. Because these tabulations were not available, BEA judgmentally estimated the missing interest paid. Their incorporation in the comprehensive revision has the largest effect for 1984, when it adds $10.4 billion to net interest.

Computer prices.--Several improvements are incorporated in the price indexes for computers (processors and peripheral equipment). The most important of these improvements is the preparation of a separate index for imports, which is used in the deflation of imported computers in PDE and in merchandise imports. The new index uses import weights to combine separate indexes for imported processors, imported personal computers, imported printers, and for domestic and imported disk drives and displays. Previously, these series were deflated using an aggregate deflator for domestic and imported computers. (10) Other improvements include the separate deflation of computers and peripheral equipment held in manufacturing and trade inventories and the extension from 1969 back to 1958 of separate deflation of computers based on indexes developed in several independent studies. (11)

Price index for multifamily houses.--Beginning with 1977, a newly developed BEA price index for multifamily houses is used for this component of residential fixed investment. Previously, the Census Bureau's price index for single-family houses was used. The multifamily index is an hedonic index derived from Census Bureau data on the prices of multifamily construction projects and on the following characteristics: Units per project, square feet per unit, bathrooms per unit, bedrooms per unit, percent of units with central air conditioning, percent of units with parking in the structure, and percent of units having more than three stories. In addition, two characteristics are included to account for the effects of the geographic location--construction wage rates and average cooling-degree days for the State where the structure is located. The index is calculated only annually; for quarterly estimates, the Census Bureau single-family index is used as both the interpolator and the extrapolator.

[TABULAR DATA OMITTED]

[TABULAR DATA OMITTED]

[TABULAR DATA OMITTED]

[TABULAR DATA OMITTED]

[TABULAR DATA OMITTED]

(1) For more information, see "Gross Domestic Product as a Measure of U.S. Production" in the August 1991 SURVEY OF CURRENT BUSSINESS.

(2) Personal interest income is derived as net interest plus interest paid by persons and by government less interest received by government.

(3) The levels of real GNP are from the August 1991 SURVEY, table 1.2 of the "Selected NIPA Tables" and table 4 of the "Reconciliation and Other Special Tables."

(4) The 1982 I-O tables were presented in "Benchmark Input-Output Accounts for the U.S. Economy, 1982," SURVEY 71 (July 1991) 30-71. For several GDP expenditure components, these estimates were subsequently revised to reflect the definitional, classification, and statistical changes incorporated into this NIPA comprehensive revision. Revised 1982 I-O tables reflecting these changes will be prepared and made available later in 1992.

(5) For a description of the estimation of PCE in the I-O accounts, see U.S. Department of Commerce, Bureau of Economic Analysis, Personal Consumption Expenditures, Methodology Paper Series MP-6 (Washington, DC: U.S. Government Printing Office, June 1990): 31-40.

(6) The USDA change is described in Economic Indicators of the Farm Sector: National Financial Summary, 1989, U.S. Department of Agriculture, ECIFS 9-2 (Washington DC: U.S. Government Printing Office, January 1991): 8-9.

(7) Since BEA began work on preparing benchmark estimates, the Census Bureau has incorporated several major revisions inthe put-in-place estimates. For State and local nonresidential construction, the revised put-in-place and BEA benchmark estimates for 1982 are only slightly different. For private new nonresidential construction, the put-in-place estimate remains substantially below the benchmark estimate for 1982, but the difference has been narrowed. This remaining difference appears to be due mainly to inadequate coverage of construction and reconstruction work performed at industrial or commercial sites in the put-in-place survey. For a discussion of the Census Bureau revisions and of the major coverage problem in the survey, see U.S. Bureau or the Census, Current Construction Reports: Value of New Construction Put in Place (Washington, DC: U.S. Government Printing Office, May 1991): 2-3.

(8) For a description of the previous adjustments, see "Improved Adjustments for Misreporting of Tax Return Information Used to Estimate the National Income and Product Accounts, 1977," SURVEY 64 (June 1984): 17-25.

(9) The IRS research is described in Income Tax Compliance Research, IRS: Supporting Appendices to Publication 7285, IRS production No. 1415 (July 1988): A46-A48.

(10) For a description of the previously used prices for computers, see "Improved Deflation of Purchases of Computers," SURVEY 66 (March 1986): 7-9.

(11) The indexes used to construct these deflators are shown in Jack E. Triplett, "Price and Technological Change in a Capital Good: A Survey of Research on Computers" in Technology and Capital Formation, ed. Dale W. Jorgenson and Ralph Landau (Cambridge, MA: The MIT Press, 1989), Table 4.6A on p. 176 and Table 4.11 on p. 190.
COPYRIGHT 1991 U.S. Government Printing Office
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
Title Annotation:includes a note on the effect of shifting the base period
Publication:Survey of Current Business
Date:Dec 1, 1991
Words:12526
Previous Article:National income and product accounts.
Next Article:Gross state product by industry, 1977-89.
Topics:


Related Articles
Revised estimates of the national income and product accounts of the United States, 1929-85: an introduction.
Selected national income and product estimates, 1929-85.
Gross product by industry: revised and updated estimates, 1947-85.
A preview of the comprehensive revision of the national income and product accounts: new and redesigned tables.
Comprehensive revision of the national income and product accounts.
Fixed reproducible tangible wealth in the United States, revised estimates.
Mid-Decade Strategic Plan.
Note on the upcoming comprehensive revision of the National Income and Product Accounts.
Subject guide: volume 83 (2003) January-December.
BEA current and historical data: appendixes.

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