Input-output accounts of the U.S. economy, 1981.

Planting, Mark A.

Input-Output Accounts of the U.S. Economy, 1981

THIS article presents the U.S. input-output (I-O) accounts for 1981. These annual accounts are consistent definitionally with the 1977 benchmark I-O accounts published in May 1984 as modified by the subsequent comprehensive revision of the national income and product accounts (NIPA's) released in December 1985.(1) The 1981 I-O estimates of final demand--that is, estimates of GNP components-- differ from the NIPA estimates for that year because the former incorporates additional information. For the major components, these differences are shown in table A and mainly reflect an improved estimate of business purchases of computers.2 The previously published 1977 and 1980 I-O accounts have been revised to provide a consistent series of accounts.

1. For the 1977 I-O accounts, see "The Input-Output Structure of the U.S. Economy, 1977,' SURVEY OF CURRENT BUSINESS 64 (May 1984): 42-84. For a description of the NIPA revision, see "Revised Estimates of the National Income and Product Accounts of the United States, 1929-85: An Introduction,' SURVEY 65 (December 1985): 1-19.

2. See "Corrections to the Estimates of Purchases of Computers,' SURVEY 66 (March 1986): 10.

The full set of 1981 I-O accounts, at the two-digit industry/commodity level, are presented in five tables: (1) Use table, (2) make table, (3) commodity-by-industry direct requirements table, (4) commodity-by-commodity total requirements table, and (5) industry-by-commodity total requirements table. The structure of these tables is identical to those published and described in the May 1984 article except that, in tables 1 and 3, the components of value added are not shown and, in tables 4 and 5, "output multipliers' are added. Output multipliers express the cumulative effect on industries or commodities of a one dollar change in the final demand for a commodity. This article presents only tables 1 and 2. See the box for information about the availability of the other tables.

Annual I-O accounts are prepared using basically the same procedures as used in the most recent benchmark table, but with less comprehensive and less reliable source data. There are four major steps in the preparation of the annual accounts: (1) Determine industry and commodity output totals, (2) estimate the commodity composition of intermediate consumption for each industry, (3) derive each GNP component and its commodity composition, and (4) balance the table. Each of these steps is described below, with a focus on the differences between the procedures and source data used in the 1977 benchmark and the 1981 annual I-O accounts.

1. Industry and commodity output totals

Source data are available to estimate annual industry output at the same level of detail as used in the 1977 benchmark. These sources include the Annual Survey of Manufactures, the Service Annual Survey, the Annual Retail Trade Survey, the Annual Trade Survey, the American Housing Survey, the Current Construction Reports, and the Current Industrial Reports--all from the Census Bureau--and tabulations from the U.S. Department of Agriculture, the Internal Revenue Service, and Federal regulatory agencies. For most industries, the data used for the annual accounts are based on sample surveys; more complete and detailed data for industries covered by the quinquennial economic census are used for benchmark accounts. When necessary, the 1981 survey data are adjusted by BEA to approximate more closely census results based on differences between the 1982 census and 1982 survey data.

For each commodity, the output total is derived as the sum of the outputs of that commodity in each industry in which it is produced. For the 1977 benchmark, the commodity composition of each industry's output is available from the quinquennial economic censuses. For the annual accounts, estimates of the commodity composition of each industry's output are prepared using commodity-industry output proportions from the 1977 benchmark. This procedure assumes that the proportions were constant from 1977 to 1981. For industries in manufacturing, the commodity output estimated in this way is adjusted to be consistent with commodity output totals available from the Annual Survey of Manufactures.

2. Commodity composition of intermediate consumption

Source data to estimate the commodity composition of intermediate consumption for each industry are available only from the quinquennial economic censuses. For the annual accounts, initial estimates of intermediate consumption are prepared using industry consumption patterns from the 1977 benchmark. This procedure assumes that input requirements per unit of constant-dollar industry output remained constant from 1977 to 1981. Industry output for 1981 expressed in 1977 dollars (see next paragraph) is multiplied by the 1977 direct requirements per dollar of industry output to obtain 1981 constant-dollar intermediate consumption, and the results reflated by commodity to current dollars.

To derive annual constant-dollar industry output, current-dollar industry output is deflated with industry price indexes that are calculated by weighting commodity price indexes using proportions from the make table. For manufacturing commodities, the indexes are to same as those prepared by BEA to estimate constant-dollar GNP by industry. For most services, the indexes are the same as those prepared by BEA to estimate constant-dollar service components of personal consumption expenditures (PCE). For other goods and services, the price indexes are derived by BEA from various sources. These commodity price indexes also are used to reflate intermediate consumption.

3. GNP components and their commodity composition

PCE.--The estimate of PCE goods and services is derived using the commodity-flow procedure--that is, PCE is derived as domestic shipments or receipts, plus imports, less exports, intermediate consumption, inventory change, and government purchases all valued at producers' prices, to which transportation and trade margins are added to put the commodity total at purchasers' prices. The source data for, and thus the implementation of, the commodity-flow procedure for the annual accounts differs from that of the benchmark accounts in three ways. First, for domestic goods and services, and annual source data, which are the same as described in step 1, provide less detail than the economic census data used in the benchmark. Second, because the commodity composition of intermediate consumption is not available, constant-dollar consumption is estimated using the 1977 consumption proportions and reflated to current dollars. Third, information on transportation and trade margins is not available annually. For 1981, total trade margins are estimated using sales data from the Annual Trade Survey and the Annual Retail Trade Survey multiplied by 1977 margin rates. Trade margins on each commodity are estimated using 1977 commodity margin rates and adjusted to agree with total trade margins.

Gross private domestic fixed investment. --The estimates of producers' durable equipment and the mobile homes part of structures are derived using the commodity-flow procedure described above for PCE. The same differences in source data that exist for PCE apply to producers' durable equipment. The estimates of other residential structures and nonresidential structures for both the annual and the benchmark accounts are based primarily on data from the Current Construction Reports. The only major difference between the annual and benchmark source data is in the mining exploration, shafts, and wells component of structures, where the benchmark estimates were based on data from the census of mineral industries and the annual estimates on survey data.

Change in business inventories.-- Annual source data for change in business inventories by industry are from the Census Bureau annual surveys and Internal Revenue Service Statistics of Income. The same sources are used for the benchmark except for mineral industries, manufactures, and manufacturers' sales branches, for which economic census data are used.

Net exports of goods and services.-- Exports and imports in both the annual and benchmark are from the Census Bureau foreign trade statistics and BEA balance of payments accounts.

Government purchases of goods and services.--Federal Government purchases in both the annual and benchmark accounts are based primarily on various annual budget documents, the Current Construction Reports, and the Current Industrial Report "Shipments to Federal Government Agencies.' State and local government purchases in both the annual and benchmark accounts are based on the Census Bureau Governmental Finances series, the Current Construction Reports, and Bureau of Labor Statistics Employment and Wages.

4. Balancing the table

Total commodity consumption (that is, intermediate and final, in steps 2 and 3) is adjusted to equal total commodity output (step 1) by allocating the difference proportionally to all consuming industries. Total commodity consumption by industry is then subtracted from total industry output (step 1) to yield value added. For some industries (for example, farms), independent estimates of value added are used, and the commodity consumption is adjusted.

Table: A.--Comparison of GNP in the NIPA's and the I-O Accounts, 1981

Table: 1.--The Use of Commodities by Industries, 1981

Table: 2.--The Make of Commodities by Industries, 1981