The U.S. national income and product accounts: revised estimates: 1981-83: first and second quarter 1984.
The revised estimates for the period including the first quarter of 1981 through the second quarter of 1984 are shown following this article. (The box below describes the other forms in which the revised estimates are available.) Estimates for 1980 can be found in the July 1983 SURVEY OF CURRENT BUSINESS, for 1977-79, in the July 1982 SURVEY, and for earlier years, in The National Income and Product Accounts of the United States, 1929-76: Statistical Tables.
In the new estimates--
* The cylical fluctuations during the period are similar to those in the previously published estimates, although in the new estimates the expansion in real GNP in 1981 is interrupted in the second quarter and the trough 1982 is the third quarter rather than the fourth (chart 5).
* The course of productivity change, largely mirroring the cyclical fluctuations, is also similar.
* The pattern of slowing inflation in 1981 and 1982 is the same as in the previously published estimates. No revision in percent change from the preceding quarter in the GNP fixed-weighted price index exceeded 0.4 percentage point and the level of the index (1972=100) by the fourth quarter of 1983 was only 0.1 index point lower than in the previously published estimates (chart 6).
* The personal saving rate is higher, especially in the quarters of 1982, when it averaged about one-half a percentage point higher (chart 7).
* The Federal deficit on a NIPA basis differs little from that shown in the previously published estimates, except in the last two quarters of 1983, when it is somewhat lower. The State and local surplus (excluding social insurance funds) is substantially lower in 1983.
The first section of this article provides a summary of the revisions for 1981-83 in terms of annual estimates of current-dollar GNP, constant-dollar GNP and associated price indexes, charges against GNP, and sector and saving-investment transactions. The next section provides methodological notes on several aspects of the revisions.
Summary of Revisions in Annual Estimates
Table 1 shows the most important new and revised source data for current-dollar GNP and charges against GNP. The years shown are the years of the estimates into which these data are directly incorporated. Current-dollar GNP
Current-dollar GNP was revised up $3.7 billion in 1981, down $3.7 billion in 1982, and down $5.7 billion in 1983 (table 2). The revisions in the percent changes from the preceding year were up 0.2 and down 0.2 percentage points, respectively, in 1981 and 1982; the percentage change was not revised in 1983. The personal consumption expenditures component (PCE) was revised down substantially in both 1981 and 1982 and each of the other final sales categories was revised up by a small amount. Thus, the contrasting direction of the revision in GNP in the 2 years was due to the change in business inventories, which was revised up sharply in 1981 and down by a small amount in 1982. In 1983, PCE was again revised down, but only $2.1 billion. Government purchases and the change in business inventories were also revised down, the former $4.0 billion and the latter $7.0 billion. Fixed investment was revised up $6.7 billion, mainly reflecting a large revision in producers' durable equipment, and net exports was revised up $0.7 billion.
Revisions in PCE goods were downward, by small to moderate amounts each year. In durable goods, the revisions were negligible. In 1982 and 1983, small to moderate downward revisions in motor vehicles were offset by upward revisions in other durables. Nondurable goods were revised down each year, and included large downward revisions in food purchases. In services, the revisions were downward by moderate amounts in 1981 and 1982 and negligible in 1983. Medical care was revised down by moderate amounts each year. Several other components--including religious and welfare services, net foreign travel in 1982; and personal business services in 1983--were also revised down.
Revisions in nonresidential fixed investment were upward each year, by small amounts in 1981 and 1982 and $4.5 billion in 1983. In structures, the revisions were negligible in 1982 and small in 1981 and 1983. Public utilities were revised up each year; these revisions were offset by downward revisions in petroleum and natural gas mining and exploration and in farm structures in 1982 and more than offset by downward revisions in these components in 1983. Producers' durable equipment was revised negligibly in 1981 and up in 1982 and 1983. The 1983 revision was concentrated in motor vehicles and also included upward revisions in computers. Residential investment was revised negligibly in 1981 and 1982. In 1983, more than one-half of a $2.2 billion upward revision was accounted for by the additions and alterations component.
Revisions in the change in business inventories were up $7.6 billion in 1981, down $1.6 billion in 1982, and down $7.0 billion in 1983. The 1981 revision was in the change in the book value of inventories held by industries other than manufacturing and trade. The small 1982 revision was due to downward revisions in farm inventories and in the inventory valuation adjustment for nonfarm inventories. The large 1983 revision was mainly in farm inventories, a small upward revision in the change in book values was offset by a downward revision in the inventory valuation adjustment.
For net exports, the revisions were small and upward each year. Exports were revised up each year, due to small upward revisions in the services components. Imports were revised down in 1981 and 1982 and negligibly in 1983; as in exports, the revisions were in services (see note 3).
Government purchases were revised up by small amounts in 1981 and 1982 and down $4.0 billion in 1983. In State and local purchases, the total and most components were revised up each year. In Federal purchases, revisions were negligible in national defense purchases each year and in non-defense purchases in 1981 and 1982. In 1983, a $5.3 billion downward revision was concentrated in purchases of the Commodity Credit Corporation (see note 2). Constant-dollar GNP and associated price indexes
In constant (1972) dollars, GNP was revised down each year--$1.7 billion, $5.4 billion, and $0.6 billion in 1981, 1982, and 1983, respectively (table 2). The revisions in the percent changes from the preceding year were down 0.1, down 0.2, and up 0.3 percentage point, respectively.
The differences between the revisions in percent change in current-dollar GNP and in constant-dollar GNP, and also in the components, appear as revisions in percent change in the implicit price deflators (table 3). The revisions in percent change in the implicit price deflactors can be decomposed into two parts: one due to revisions in the composition of goods and services, and one due to revisions in prices. The former can be seen as the difference between revisions in percent change in implicit price deflators and in fixed-weighted price indexes; the latter can be seen as revisions in percent change in fixed-weighted price indexes. In the fixed-weighted price indexes, the revisions in percent change were generally small. Revisions of as much as 0.3 percentage point occurred in more than 1 year only in residential investment, imports, and Federal purchases. Charges against GNP
Revisions in total charges against GNP, as shown in table 4, were similar to those in GNP except in 1981. In that year, when GNP was revised up but charges against GNP was revised down, the revision in the statistical discrepancy was $10.6 billion. In 1982 and 1983, revisions in the statistical discrepancy were small.
Each year, the revision in charges against GNP reflected moderate to large downward revisions in components that were only partly offset by upward revisions, some of which were large. The largest revisions--not all in the same direction--were in net interest and proprietors' income. Two components-compensation of employees and corporate profits--were revised down each year by small to moderate amounts. Revisions in other components were small.
Compensation of employees was revised down $3.8 billion in 1981, $1.4 billion in 1982, and $5.3 billion in 1983. In 1981, the revision was concentrated in other labor income; in 1982, in employer contributions for social insurance; and in 1983, in wages and salaries.
Proprietors' income, which was revised up in 1981 and 1982 by $5.0 and $2.1 billion, respectively, reflected small to moderate upward revisions in both the farm and nonfarm components. In 1983, when the revision was down $6.8 billion, the farm income component was revised down $7.1 billion.
Rental income of persons was revised up each year, by small amounts in 1981 and 1982 and by $3.5 billion in 1983.
Corporate profits were revised down by a small amount in 1981, by $5.7 billion in 1982, and by $4.0 billion in 1983. Profits before tax--that is, book profits--were revised down even more. Domestic profits in most industries except public utilities were revised down. The inventory valuation adjustment was revised down by small amounts in 1982 and 1983, and the capital consumption adjustment was revised up by moderate amounts each year.
Net interest was revised down $8.9 billion in 1981, negligibly in 1982, and up $9.1 billion in 1983 (see note 4).
REvisions in capital consumption allowances with capital consumption adjustment, indirect business taxes, and subsidies less the current surplus of government enterprises were negligible with one exception. Indirect business taxes were revised down $5.5 billion in 1983, largely due to State and local property taxes. Sector and saving-investment transactions
The current-dollar revisions discussed so far are for GNP, charges against GNP, and their components, which make up the two sides of the national income and product account in the NIPA five-account summary system (see table A on page 18). This summary system also includes three sector accounts--for persons, for government, and for foreigners in their transactions with the United States--and a gross saving and investment account. Many of the transactions in these accounts are the counterentries to transactions for which the revisions have already been mentioned. The following references are limited to revisions in other transactions and aggregates.
In the personal income and outlay account, personal income was revised down $5.5 billion in 1981, up $6.0 billion in 1982, and up $2.1 billion in 1983 (table 5). Most of the revisions were accounted for by components that have already been referred to: wages and salaries, proprietors' income, rental income of persons, and personal interest income. Revisions in components of personal income not yet referred to--personal dividend income and transfer payments--were negligible to small. Revisions in personal tax payments were also negligible to small. Accordingly, revisions in disposable personal income were similar to those in personal income. The revisions in personal outlays, which were dominated by those in PCE, were downward, by large amounts in 1981 and 1982; outlays were not revised in 1983. The revisions in personal saving were up each year, by $2.1 billion in 1981, $10.6 billion in 1982, and $4.4
billion in 1983.
In the government account, revisions in Federal receipts and expenditures except those in nondefense in 1983, were small (table 6). The deficit on a NIPA basis was revised up by small amounts in 1981 and 1982 and down $3.0 billion in 1983. Revisions in State and local receipts were small and generally upward in 1981 and 1982; the 1983 revision, which included a moderate downward revision in indirect business taxes, was downward $5.2 billion. Revisions in expenditures were small and upward in 1982 and 1983. The surplus on a national income and product accounts basis was revised up by small amounts in 1981 and 1982, but down $7.3 billion in 1983.
In the foreign transactions account, the revisions in components other than exports and imports, which were discussed earlier, were negligible (table 7). Accordingly, the revisions in net foreign investment were similar to those in net exports.
Revisions in the components of gross saving and investment were all referred to earlier (table 8). In summary, in 1981, the revision in gross saving was negligible and the revision in gross investment was upward $11.1 billion. In 1982, gross saving and investment were revised up, both by small amounts. In 1983, gross saving was revised down by $2.4 billion, and gross investment was revised negligibly.
1. Seasonal adjustment.--Seasonal adjustment factors, generally calculated with the Census Bureau's X-11 program, were updated by bringing in the year 1983. The updating was routine; several special situations were examined, but they did not warrant modification of the standard procedures. This updating accounted for roughly one-half of the revision in the quarterly change in GNP. As a result of the updating, the first and fourth quarters tended to be higher and the second and third quarters lower, largely reflecting the updated seasonal factors for the change in business inventories.
For imports of crude petroleum, seasonal adjustment factors were not applied to the quarters of 1982 and 1983 in the revised estimates. Highly erratic movements in these 2 years made it difficult to estimate seasonals. Crude petroleum imports for 1981, and distilled products for all years, were seasonally adjusted in the usual way.
2. Payment-In-Kind (PIK) transactions.--The PIK program was one of the 1983 Federal farm programs designed, in part, to reduce production and thereby reduce the large stocks of grains and cotton accumulated in 1981 and 1982. Under the PIK program, farmers could idle acreage and receive crops from Commodity Credit Corporation (CCC) inventories in return. On the basis of revised estimates, farmers took title to $5 billion of crops previously held by the CCC in 1983. This amount reflects an upward revision of $1.5 billion. This revision reflected better information about the dates on which the transfers were made; it was required because the unexpectedly large participation in the program caused a delay in the processing of reports on which the estimates used in the NIPA's are based. In the NIPA's, these transfers do not affect gross farm product, GNP, or Federal Government expenditures. However, they do affect components of these totals. Thus, the revision in PIK transactions, although it did not affect the totals, did affect many components. Table 9 brings the most important of these components together.
In gross farm product (and also output), the reduction in CCC stocks due to PIK is a reduction in net CCC loans, which is shown in the table combined with cash receipts from marketings. This reduction is offset by an increase in farm inventories. (As shown by the table, these components of farm product (and output) were affected as well by other large revisions. These revisions reflected lower farm production--due to the midyear drought--than estimated earlier.)
In GNP, the increase in farm inventories due to PIK is offset by a reduction in CCC inventories. The latter is a reduction in Federal net purchases. Correspondingly, PIK transfers do not affect GNP seen as the sum of income components. GNP is valued at market prices--that is, exclusive of subsidies. PIK transfers appear as subsidies to farmers and are part of farm proprietors' income (or, for corporate farms, corporate profits); they are subtracted along with other subsidies as a separate item in the derivation of GNP.
In Federal expenditures, the PIK subsidy payments offset the reduction of CCC inventories due to PIK, so that PIK transactions have no effect on total Federal expenditures.
3. International services transactions.--Exports and imports of services were affected by revisions in the foreign travel and direct investment income components.
Revisions of travel and passenger fare receipts reflected the resumption of tabulations of forms filed by nonresident aliens upon entry into the United States and the introduction of data from an expenditure survey, by the Bank of Mexico, of Mexicans in the border area as they returned from the United States.
The remitted and reinvested components of earnings of unincorporated affiliates of U.S. residents and of foreign residents have been separated beginning in 1982 and 1981, respectively. Previously, when the two components were inseparable statistically, both had been treated as remitted. Remitted earnings continue to be treated like dividends of incorporated affiliates. Remitted earnings of U.S. residents from their unincorporated foreign affiliates are treated like dividends received by domestic corporate business and like dividends paid by the the rest of the world. Remitted earnings of foreign residents from their unincorporated U.S. affiliates are treated like dividends paid by domestic corporate business and ad dividends received by the rest of the world. Reinvested earnings are now treated as undistributed corporate profits. U.S. residents' share of reinvested earnings of their unincorporated foreign affiliates is treated as receipts of undistributed profits from the rest of the world. Foreign residents' share of reinvested earnings of their unincorporated U.S. affiliates is treated as payments of undistributed profits to the rest of the world. The amounts of the reinvested earnings are shown in footnotes to the National Income and Product Accounts Tables in tables 6.24B, 6.25B, 8.6, and 8.12.
Further information is available in the discussion of the annual balance of payments revisions, which appeared in the June 1984 SURVEY OF CURRENT BUSINESS.
4. Interest -- The revisions in interest were the largest among the major components of charges againt GNP and personal income. The methodologies fro these components, which are highlighted in what follows, make them particularly subject to revision.
Net interest is estimated as part of an accounting for interest flows by sector and by legal form of organization, separately for monetary and imputed interest. The last year for which complete source data are available is 1981; for later periods, the estimates are based on less reliable source data and thus are subject to increasing margins of error. Table 10 is an arrangement of these interest estimates designed to show the sources of revision.
For years for which source data are complete, estimates of net monetary interest paid by business are derived mainly from tax return tabulations, prepared by the Internal Revenue Service, of interest paid and received by corporations, sole proprietors, and partnerships. Tabulations for corporations become available with a 3-year lag, and those for sole proprietorships and partnerships with a 2-year lag. For years when these tabulations are not available, estimates are prepared using a different breakdown. For financial corporations, data from regulatory agencies (for example, the Federal Deposit Insurance Corporation, and Federal Home Loan Bank Board, and the National Credit Union Administration) and trade sources are used. For nonfinancial corporations, estimates are prepared by extrapolating interest paid by the product of debt outstanding for several kinds of debt (largely from the Federal Reserve Board's flow-of-funds accounts) and appropriate interest rates, and by extrapolating interest received using a similar procedure. Interest paid and received by partnerships and proprietorships are extrapolated by past trends. For all years, interest payments by homeowners on mortgage and home improvement loans are estimated using data on mortgages outstanding from the flow-of-funds accounts along with appropriate interest rates. For current quarters, estimates of net business interest are derived as a residual, as will be explained in connection with personal interest income.
Imputed interest is paid by two kinds of financial intermediaries: banks and other depository institutions, and life insurance carriers and private noninsured pension funds. For the former, imputed interest is estimated as the property income earned on investment of deposits less monetary interest paid on deposits (less profits, in the case of mutual depositories). For the latter, it is estimated as the property income earned less profits of mutual life insurance carriers. The sources and methods used in estimating imputed interest are similar to those used in estimating monetary interest of financial corporations for years when tax return tabulations are not available.
Net interest paid by the rest of the world is estimated as part of the U.S. international transactions accounts. The basic approach for estimating both receipts and payments is to multiply the amount outstanding in various asset categories by appropriate interest rates.
In 1981, net interest was revised down $8.9 billion. The revision mainly reflected the incorporation of data from corporate tax return tabulations. In 1982, the revision in level was negligible and the revision in change was up $8.7 billion. Tabulations of tax returns for sole proprietorships and partnerships, revised data on assets and debts of nonfinancial corporations from the Federal Reserve Board's flow-of-funds accounts, and revised interest rates were introduced. For 1983, the upward revision in level was $9.1 billion and the revision in charge was $9.3 billion. A full breakdown of these revisions by component cannot be shown, because, as noted earlier, not all components are estimated separately for current quarters. Net monetary interest more than accounted for the revision in level ($11.8 billion) and revision in change ($12.5 billion). The revision reflected the substitution of the methodology used for annual estimates, which was just described, for the methodology used for estimates in current quarters, which is explained immediately below in connection with personal interest income.
The other major interest aggregate in the NIPA's is personal interest income. Because personal interest income cannot be estimated reliably on the basis of information either about persons' interest receipts or about the portions of interest payments by business, government, and the rest of the world that go to persons, an indirect estimating procedure must be used. The calculation used for annual estimates of personal interest income is shown in table 10: net interest (the component of charges against GNP) plus interest paid by government to persons and business less interest received by government plus interest paid by consumers to business.
For current quarters, this formula cannot be implemented because there are no data with which to estimate the portion of net interest that consists of the net monetary and imputed interest paid by business. The procedure that is used is, in effect, to apply in reverse the formula for the annual estimates: to estimate directly personal interest income, interest paid and received by government, and interest paid by consumers to business in order to derive net interest. Personal interest income is estimated in two parts: Monetary interest is extrapolated using flow-of-funds information on assets held by persons and appropriate interest rates; imputed interest received is extrapolated by past trends. Federal Government interest is estimated largely on the basis of Treasury Department reports; State and local government interest is estimated by extrapolation of past trends. Interest paid by consumers to business is extrapolated on the basis of Federal Reserve Board information on consumer debt outstanding and on appropriate interest rates.
Most of the revisions in personal interest income in 1981 and 1982--down $9.5 billion and negligible--reflect those in net interest. The revision in personal interest income in 1983--up $10.0 billion--cannot be attributed to the revision in net interest. The revision in that year reflects the replacement of the direct estimate by an estimate prepared by applying the formula for annual estimates. In addition, there were revisions in government interest, largely in interest paid and received by State and local governments. These revisions largely reflect the incorporation, for 1981 and 1982, of data from the Census Bureau's Governmental Finances.
5. Deflation of space rent.--Space rent is the rent of a housing unit exclusive of utilities furnished and of rent of equipment, such as kitchen appliances. It is a component of both owner- and tenant-occupied housing services in personal consumption expenditures and makes up the bulk of housing output. Beginning with the estimates for 1981, a new price index prepared by BEA is used in its deflation. The index is based on information from the American Housing Survey (and thus is consistent with the sample from which the mean contract rent that is used in estimating space rent is taken). The major methodological improvement is that the new index takes into account that landlords often raise (or lower) the rent they charge when they rent to a new tenant. Such changes are registered in survey data when a housing unit moves from a vacant to a rented status. Previously, space rent had been deflated using the rent component of the Consumer Price Index. The effect of the change in procedure was to lower constant-dollar space rent by $1.9 billion, $3.7 billion, and $5.6 billion in 1981, 1982, and 1983, respectively.
6. Quarterly farm estimates.--Two improvements were introduced in the quarterly farm estimates prepared by the Department of Agriculture. For farm inventories, the quarterly pattern of change in current dollars is derived by interpolating between constant-dollar annual estimated of inventory change and reflating the quarterly estimates with an aggregate index of market prices. The resulting series, which shows less volatility than the previous series, appears more reasonable. Previously, the quarterly estimates were derived by interpolating between annual estimates of current-dollar inventory change.
For cash receipts for crops, the new procedure relates to the level of detail at which seasonal adjustment is carried out. Cash receipts are the sum of open-market sales and net Commodity Credit Corporation loans. Under the new procedure, the two components are seasonally adjusted separately and then summed. Previously, the sum of the two seasonally unadjusted components was seasonally adjusted.
The revisions caused by these changes in procedure were small.
7. Quarterly pattern for rental income.--The quarterly estimates of rental income of persons from owner-occupied housing were introduced as the interpolator and extrapolator of total rental income of persons. Rental income of persons from owner-occupied housing, which accounts for about one-half of the total, reflects quarterly movements in housing completions; the Consumer Price Indexes for rent, maintenance and repairs, property insurance; mortgage interest; local property taxes; and capital consumption allowances. Previously, the quarterly estimates of total rental income of persons were derived by fitting a smooth curve to the annual estimates and, for current quarters, by extrapolating past trends. The revisions caused by the change in procedure were small.
|Printer friendly Cite/link Email Feedback|
|Publication:||Survey of Current Business|
|Date:||Jul 1, 1984|
|Previous Article:||The business situation.|
|Next Article:||The economy and the federal budget: guides to the automatic effects.|