THIS article presents revised estimates of new plant and equipment(P&E) expenditures by business in the Unites States for 1947-83.Revised estimates for the first three quarters of 1984 will bepublished, together with estimates for the fourth quarter of 1984, inthe April 1985 SURVEY OF CURRENT BUSINESS. The estimates are of bothactual and planned expenditures compiled from the quarterly P&Eexpenditures survey conducted by BEA.
This comprehensive revision is the fourth in the history of theP&E survey. Comprehensive revisions have two primary purposes: Theintroduction of “benchmarks,” to which the P&Esurvey-based estimates are tied, and the retabulation of the P&Esurvey reports. Benchmarks are universe estimates for each industrybased on data from the quinquennial economic censuses and, especiallyfor industries not covered by these censuses, from a variety of othersources that provide comprehensive, detailed information. Retabulationsincorporate information not available when the previously publishedestimates were prepared; industry and size classifications are updated,and company reports received too late to be included in the previoustabulations are included.
In addition, the quarterly survey samples arereedited. In this revision, benchmarks for 1977 are introduced, and estimatesback to 1973 and forward to 1983 are revised to reflect them. Theretabulation is carried out for 1977 and later years.
The revised seriesalso reflect definitional revisions, which affect the estimates for1947-83, and recalculation of seasonal factors, bias-adjustment factors,and price deflators. This revision was done in the context of a number of changesresulting from an in-depth review of the P&E survey over the lastseveral years. These changes were made to reduce respondent burden andto improve the quality of the estimates. Quarterly estimates no longerinclude selected nonmanufacturing industries: Real estate; professionalservices; membership organizations and social services; and forestry,fisheries, and agricultural services. Data are now collected onlyannually for these industries, which account for about 11 percent ofP&E expenditures. Also, separate quarterly estimates of plant andof equipment expenditures are no longer published. Instead, annualestimates now appear once each year. The revision raised the expenditures estimates for all years for”all industries,” which is the total for the industriessurvey.
Although the revision extends back to 1947, the largestrevisions are for the benchmark year 1977 and for later years. For1977, the estimate was raised 5.8 percent, to $184.8 billion, and for1983, 13.2 percent, to $304.8 billion (table 7 and chart 5). Inmanufacturing, the 1977 estimate was lowered 2.
5 percent, but the 1983estimate was raised 4.2 percent. In nonmanufacturing, the 1977 estimatewas raised 11.3 percent and the 1983 estimate was raised 19.
6 percent. In constant (1972) dollars, the revision raised the expendituresestimate for “all industries” 6.6 percent, to $126.
0 billion,for 1977, and 14.6 percent, to $146.4 billion, for 1983 (table 7 andchart 5).
In manufacturing, the estiamte was lowered 2.6 percent for1977, but raised 2.5 percent for 1983. In nonmanufacturing, theestimate was raised 12.
5 percent for 1977 and 23.0 percent for 1983. The first section of this article discusses the sources of therevisions. The second compares the previously published and revisedestimates and describe the revised estimates for total P&Eexpenditures, for plant and for equipment expenditures separately, andfor planned expenditures. Primary emphasis is on the period 1977-83.Three technical notes follow. The first describes the P&E surveyand the procedures used to prepare the P&E expenditures estimates.The second summarizes the methodology used to prepare the revisedP&E series.
The third compares the P&E series with thenonresidential fixed investment component of GNP. The revised seriesfollow the text: Annual P&E expenditures, in current and constantdollars, in table 7; quarterly P&E expenditures, in current andconstant dollars, in table 8; annual plant expenditures and equipmentexpenditures, separately, in current and constant dollars, in tables 9and 10; and planned P&E expenditures as a percentage of actualexpenditures, quarterly and annually, in tables 11 and 12. Sources of the Revisions The revisions in the current-dollar annual estimates had twosources: Statistical revisions and definitional revisions. These areshown for selected years, including 1977-83, in table 1. Therecalculation of seasonal factors affected the quarterly, but not theannual, estimates; the recalculation of bias-adjustment factors affectedthe estimates of planned, but not actual, expenditures; and therecalculation of price deflators affected the constant-dollar, but notthe current-dollar, estimates.
Statistical revisions The statistical revisions stem from the two primary purposes of therevision: The introduction of the 1977 benchmarks and the retabulationof the sample reports for 1977-83. The statistical revision for 1977due to the 1977 benchmarks was $8.7 billion; manufacturing was lowered$1.
7 billion and nonmanufacturing was raised $10.4 billion. Thestatistical revisions for 1973-76 and for 1978-83 also reflect the 1977benchmarks. For 1973-76, the estimates were revised using thepreviously published series and the 1977 benchmarks. For 1978-83, theestimates were prepared by extrapolating the 1977 benchmarks; therefore,the size of the statistical revisions for these years varies inproportion to P&E spending.
For 1983, the statistical revision dueto the 1977 benchmarks was $12.0 billion, 34 percent of the totalrevision. the retabulation of the sample reports affected the estimates for1977 and later years, although for 1977 it affected only the pattern ofthe quarterly estimates. For 1983, the statistical revision due toretabulation was $21.
9 billion, 62 percent of the total revision. Definitional revisions Three nonmanufacturing industries were affected by two definitionalrevisions. Unlike the statistical revisions, which affected only theperiod 1973-83, the definitional revisions affected all years back to1947. First, certain railroad track maintenance expenditures that werepreviously defined as current expenses are now defined as capitalexpenditures. Railroads were required by the Interstate CommerceCommission to make this change beginning in 1983; the P&Eexpenditure series for 1947-83 was revised to obtain a consistent timeseries. Second, expenditures for leased railroad equipment andaircraft, previously assigned to the railroad and air transportationindustries, are now assigned to the finance industry (in most cases, theindustry of the original purchaser). This revision, which affected theestimates for 1959 and later years, makes the treatment of leased assetsconsistent across all industries–that is, the P&E expenditureseries is now entirely on an owner basis. For 1947, the revision due tothe definitional revisions was $0.
5 billion. For 1973, it was $0.7billion, 49 percent of the total revision; for 1977, $1.4 billion, 14percent of the total revision; and for 1983, $1.7 billion, 5 percent ofthe total revision. The Revised P&E Expenditures Series On the revised basis, P&E expenditures for “allindustries” increased at an average annual rate of 7.
8 percent from1947 to 1983 (table 2). The previously published series increased at arate of 7.5 percent. During this period, manufacturing expendituresincreased at a rate of 7.5 percent, and nonmanufacturing increased at arate of 8.1 percent. In the previously published series, the comparablerates were 7.
3 percent and 7.7 percent, respectively. During 1977-83, P&E expenditures for “all industries”increased at an average annual rate of 8.7 percent, compared with 7.
5percent for the previously published series. Manufacturing expendituresincreased at a faster rate, 9.5 percent, than nonmanufacturing, 8.
2percent. In the previously published series, the comparable rates were8.3 percent and 6.9 percent, respectively. In manufacturing, nondurable goods industries increased at a faster rate, 10.
5 percent, than durablegoods, 8.4 percent. The largest growth rates in manufacturing were in”other nondurables” (18.
3 percent), aircraft, and electricalmachinery. The smallest growth rates were in blast furnaces-steel works(2.1 percent) and fabricated metals. In nonmanufacturing, the largestgrowth rates were in finance and insurance (18.7 percent), airtransportation, and wholesale and retail trade. Railroads and personaland business services had smaller growth rates, and “othertransportation” declined. On the revised basis, real (constant-dollar) P&E expendituresfor “all industries” increased at an average annual rate of3.
5 percent from 1947 to 1983, compared with 3.2 percent for thepreviously published series (table 2). For 1977-83, the revised seriesin creased at an annual rate of 2.5 percent, compared with 1.
3 percentfor the previously published series. During this period, realexpenditures in manufacturing increased at a faster rate, 3.0 percent,than in nonmanufacturing, 2.3 percent. Durable and nondurable goodsindustries increased at about the same rate. In nonmanufacturing,”commercial and other” and public utilities increased, whiletransportation and mining declined. The revised series of real P&E expenditures follows a cyclical pattern similar to that of the previously published series; both showtwo contractions during 1977-83 corresponding to the two business-cyclecontractions in the U.S.
economy (chart 6). Real P&E expendituresfor “all industries” peaked in the first quarter of 1980 at$154.1 billion; they declined 3.0 percent over the next three quarters.Real expenditures then rose from the first through the third quarters of1981, when they reached $157.7 billion. They then declined 11.9 percentover the next six quarters, to $138.
9 billion. The 1980 contraction inreal P&E expenditures was mild compared to the average for the eightpostwar business-cycle contractions (the postwar average was afive-quarters decline of 11.7 percent); the 1981-83 contractioncorresponded more closely to the average. The peaks and troughs for various industries differed in timing andseverity from those for “all industries”. During the mostrecent contraction in real P&E spending, manufacturing declined 15.6percent, and nonmanufacturing, 10.2 percent, from their peaks in thethird quarter of 1981 to their respective troughs in the fourth quarterof 1982 and the second quarter of 1983 (table 3). In manufacturing, thecontraction was more severe in durables, which declined 17.
9 percentfrom peak to trough, than in nondurables, which declined 14.u percent.In durables, the largest declines were in nonferrous metals (45.1percent), motor vehicles, blast furnaces-steel works, and fabricatedmetals. Stone-clay-glass, which had never recovered from the previouscontraction, declined an additional 27.2 percent.
In nondurables, thelargest declines were in paper (26.9 percent), food-beverage, andtextiles. The smallest declines in manufacturing were in rubber (10.
7percent), “other nondurables,” and machinery (exceptelectrical). During the most recent contraction, the largest declines innonmanufacturing were in mining (37.5 percent) and communication.Spending in transportation, which had never recovered from the previouscontraction, declined an additional 24.4 percent. The smallestdeclines–5.1 percent to 6.1 percent–were in public utilities,wholesale and retail trade, and finance and insurance.
Expenditures for plant and for equipment Once each year, companies are asked to separate their annualP&E expenditures into plant and equipment. However, not allcompanies that report total P&E expenditures provide such abreakdown. As a result, the two components are less reliable than thetotal, and separate estimates for plant and for equipment are presentedonly for the major industry groups shown in table 4.
On the revised basis, for the bench-mark year 1977, plantexpenditures were $62.5 billion, 0.3 percent lower than previouslypublished; equipment expenditures were $122.3 billion, 9.2 percenthigher than previously published. For 1983, plant expenditures were$107.4 billion, 0.4 percent lower than previously published; equipmentexpenditures were $197.
4 billion, 22.3 percent higher than previouslypublished. From 1947 to 1983, plant expenditures and equipment expendituresfor “all inudstries” increased at about the same averageannual rate–7.8 percent and 7.9 percent, respectively. For 1977-83,plant expenditures increased faster, at an average annual rate of 9.
4percent, compared with 8.3 percent for equipment expenditures. From 1977 to 1983, manufacturing expenditures for equipmentincreased at a slightly faster average annual rate, 9.6 percent, thanthose for plant, 9.
1 percent. Equipment expenditures increased fasterfor both durable and nondurable goods industries. In nonmanufacturing,in contrast, in contrast, expenditures for plant grew at a faster rate,9.6 percent, than those for equipment, 7.4 percent. Plant expendituresincreased faster for each major industry group in nonmanufacturing. Plant expenditures as a proportion of total P&E expendituresfor “all industries” varied little over time; they were 36.
3percent of the total in 1947, 33.8 percent in 1977, and 35.2 percent in1983. In manufacturing, the proportion declined from 35.6 percent in1947, to 28.4 percent in 1977, to 27.9 percent in 1983.
The largerdecline was in durable goods; nondurable goods showed relatively littlechange. In nonmanufacturing, the proportion changed little from 1947 to1977, then increased slightly. The largest increase was in mining, from26.1 percent in 1947, to 45.8 percent in 1977, to 55.8 percent in 1983.
Planned expenditures The estimates for planned P&E expenditures were revised to beconsistent with the revised estimates for actual P&E expenditures.The planned expenditures estimates incorporate adjustments forsystematic reporting biases due to factors other than cyclical changesin economic and operating conditions. The mean absolute percentage deviation between planned and actualspending levels for 1955-83 was 1.
8 percent for one-quarter-ahead plansand 2.6 percent for two-quarters-ahead plans. For manufacturing, thedeviations were 2.8 percent and 3.
9 percent, respectively; fornonmanufacturing, they were 1.7 percent and 2.7 percent, respectively. The mean absolute percentage deviations for major industry groupswere larger than the deviations for “all industries” and varyfrom industry to industry. The deviations were smallest for”commercial and other” (2.5 percent for one-quarter-ahead and3.5 percent for two-quarters-ahead plans), nondurable goodsmanufacturing, public utilities, and durable goods manufacturing.
Theywere largest for mining (5.6 percent and 7.3 percent) and transportation(3.9 percent and 7.9 percent).
For plans reported 1 year ahead in the fourth-quarter survey, themean absolute percentage deviation between planned and actual spendingfor 1955-83 was 3.0 percent for “all industries.” Formanufacturing, the deviation was 4.
5 percent, and for nonmanfacturing,2.6 percent. Among major industry groups, the deviations were smallestfor public utilities (3.4 percent), nondurable goods manufacturing, and”commercial and other.” They were largest for transportation(7.
7 percent), mining, and durable goods manufacturing. TECHNICAL NOTES 1. The P&E Survey This note describes the P&E survey–the data reported, thecollection schedule, and the sample–and the procedures used to preparethe P&E expenditures estimates.
Description of the P&E survey BEA’s quarterly P;E survey collects data on expendituresby business for new plant and equipment for installation or use in theUnited States. The survey covers expenditures both for new facilitiesand for expansion or replacement of existing facilities that arechargeable to fixed asset accounts and for which depreciation oramortization accountsa re ordinarily maintained. The distinction betweenstructures–that is, plant–and equipment is not always clear-cut.However, a useful guideline is that the former are not movable, but thelatter are.
Detailed definitions of plant and equipment expenditures,showing specifically what is included and what is excluded, as well asother instructions to respondents are given on the buck of the reportforms. A sample form is provided at the end of this article. Companies generally report expenditures for the quarter in whichpayment is made to the supplier; construction work performed by acompany’s own employees–force-account construction work–isgenerally reported for the quarter in which costs are incurred. Companies are instructed to report expenditures for structures andequipment acquired for lease to others. Thus, the reporting is on anowner, rather than on a user, basis. Expenditures are included in theindustry of the company retaining title, even if the capital good is foruse by, or even capitalized by, a company in another industry. Companies are instructed to report, whenever possible, on a fullyconsolidated basis–that is, for all their domestic operations,including those of their majority-owned subsidiaries. BEA classifieseach company in a two-digit Standard Industrial Classification industryon the basis of its primary activity, which is the activity with thelargest volume of sales or business receipts.
All of its capitalexpenditures–for its primary activity as well as for its otheractivities–are assigned to that industry. Company classifications arereviewed during comprehensive revisions using responses to surveyquestions and outside information; changes are reflected in theretabulated samples. In addition, classifications of companies involvedin major mergers are reexamined between revisions. P&E expenditures–both actual and planned–have been collectedquarterly since the survey began in 1947.
Each quarterly surveycollects the following date: Actual expenditures for the previousquarter and planned expenditures one quarter ahead, two quarters ahead,and three quarters ahead. The third- and fourth-quarter surveys collectplanned expenditures for the coming calendar year (year-ahead plans).(Plans for the second half of the year are derived in the fourth-quartersurvey by subtracting the sum of the first-quarter and second-quarterplans from the year-ahead plans.) At various times during each year,annual data are collected on the following: P&E expenditures,separate plant and equipment expenditures, sales, assets, price changes,type of business activity, and P&E expenditures for assets for leaseto others. BEA currently requests quarterly data from a nonprobability sampleof about 12,000 companies; about 9,000 additional companies are sampledin the industries survey only annually. In 1977, for “allindustries,” the proportion of P&E expenditures represented bycompanies responding to the survey–that is, the sample”coverage”–was 54 percent; the corresponding proportion formanufacturing was 68 percent and for nonmanufacturing, 45 percent.
Survey coverage is highest in the most concentrated industries,characterized by a relatively small number of large firms making a largeshare of the industry’s capital expenditures. Table 5 shows thatsample coverage in 1977 was above 80 percent in blast furnaces-steelworks, nonferrous metals, motor vehicles, aircraft, petroleum, electricutilities, and communication, but below 25 percent in mining, wholesaleand retail trade, finance and insurance, personal and business services(including construction), and in the nonmanufacturing industriessurveyed only annually. Estimating procedures For each quarter, universe estimates for actual and planned P;Eexpenditures are made for each industry by extrapolating the P;Euniverse estimates forward from the previous quarter on the basis ofmovements in the quarterly sample. (The starting points for the seriesof quarterly universe estimates are the benchmarks. The methodology fordetermining the benchmarks is described in technical note 2.) The foursteps in the procedure for making the quarterly estimates are asfollows. First, after each, quarter’s report forms have been reviewedfor accuracy and consistency, the sample data are summarized by”tab group.
” For most industries, the tab groups areasset-size classes; for the other industries, the tab group is theentire industry. For each tab group, expenditures for companiesreporting in both the current and the preceding quarter are totaled andused to calculate a ratio indicating the relative change from thepreceding quarter. Second, each tab group’s sample is edited. Companiesreporting relative changes in investment spending that are noticeably different from other companies in that group are identified, and someare classified as “outliers.” Third, a tab-group universe estimate is derived in which theoutliers are treated separately using their reported values; theremainder of the universe estimate is based on the relative change ininvestment spending by the other companies. (The universe estimates forplanned expenditures are made as just described except that thetab-group summaries are based on sample data for the current quarter andthe three successive quarters.) Fourth, several adjustments are made to the universe estimates.
Theplanned expenditures estimates are adjusted for systematic reportingbiases, and the actual and bias-adjusted planned expenditures estimatesare adjusted for seasonal variation. Then, the resulting set of actualand planned current-dollar estimates is adjusted to remove the effectsof inflation; the result is a set of actual and planned constant-dollarestimates. Descriptions of these adjustments follow. Bias adjustments.–Comparison of planned expenditures with actualexpenditures for the same quarter reveals systematic reporting biases inthe planned expenditures. Because there are well-established patternsin these biases, for most purposes it is desirable to adjust the plansdata for them.
Bias-adjusted plans estimates are prepared by dividinguniverse estimates for planned expenditures by bias-adjustment factors.These factors are calculated by industry for each planning horizon. Fora given quarter, the bias-adjustment factor is the median of the ratiosof planned to actual expenditures for that quarter in the preceding 8years. Seasonal factors.–Seasonal factors for adjusting the P&Eexpenditures by industry are computed using the Census Bureau’sX-11 program.
The seasonal factors for actual P&E expenditures arealso used to seasonally adjust the bias-adjusted planned expenditures. Price deflators.–The actual and planned quarterly expendituresestimates are adjusted by BEA to remove the effects of inflation usingimplicit price deflators derived from unpublished detailed estimates inthe national income and product accounts (NIPA’s) of current- andconstant-dollar nonresidential fixed investment (NRFI). Because NRFIdiffers from the P;E series (see technical note 3), the NIPAestimates must be adjusted before price deflators for P;Eexpenditures can be calculated. First, capital-flow matrixes are usedto transform the NIPA current- and constant-dollar estimates ofstructures and producers’ durable equipment by type into current-and constant-dollar purchases by establishment-based industry;adjustments are made to conform the NIPA estimates to P&E surveyindustry definitions.
Second, the adjusted establishment-basedestimates are transformed to a company basis. Implicit price deflatorsare then derived by dividing the current-dollar estimates by theconstant-dollar estimates. Constant-dollar P&E estimates are calculated by applying theimplicit price deflators to the quarterly current-dollar estimates.Final constant-dollar estimates are derived by constraining the industrydeflators so that the growth rate for the total P&E deflator for”all industries” equals that of the nonresidential fixedinvestment deflator (on a P&E survey basis). To adjust planned P&E expenditures, the price deflator for eachindustry is projected using its growth rate over the latest fourquarters. Once the deflators are projected, they are applied to eachindustry’s seasonally adjusted, planned current-dollarexpenditures.
2. Revision Methodology The methodology used to prepare the revised P;E series issummarized in five sections: 1977 benchmarks, definitional revisions,quarterly expenditures series, planned expenditures, and expendituresfor plant and for equipment. 1977 benchmarks Previous benchmarks were compiled for 1948, 1958, 1963, 1967, and1972; the current revision introduces benchmarks for 1977. The currentbenchmark year was selected primarily because of the availability of1977 Enterprise Statistics data prepared by the Bureau of the Census.(Benchmarks for 1982 will be prepared after 1982 Enterprise Statisticsdata become available.) The sources and methods used to prepare the 1977 benchmarks variedamong industries. The 1977 edition of Enterprise Statistics was theprincipal source of the benchmarks for the manufacturing industries, andfor the mining, trade, personal and business services, and constructionindustries. Specifically, the benchmarks for the manufacturingindustries, for mining, and for construction were based on published andunpublished tabulations of data from Enterprise Statistics, the Censusof Manufactures, the Census of Mineral Industries, and the Census ofConstruction Industries.
For wholesale trade, retail trade, andpersonal and business services, benchmarks were constructed from acombination of enterprise and establishment statistics prepared by theBureau of the Census. Specifically, the benchmarks were derived fromcapital expenditures and employment data for establishments from theCensus of Wholesale Trade, the Census of Retail Trade, and the Census ofservice Industries, using employment matrixes from Etnerprise Statisticsto transform the establishment data to a company basis. For transportation, a combination of data from the InterstateCommerce Commission (ICC), the Securities and Exchange Commission (SEC),other regulatory agencies, and the P;E sample was used inconjunction with data from the Statistics of Income compiled by theInternal Revenue Service (IRS). (The ICC data were supplemented withdata from other agencies to include companies engaged in intrastate transportation, which are not regulated by the ICC.) The ratio ofuniverse gross depreciable assets reported by the IRS to the grossdepreciable assets reported in the combined data was multiplied by theexpenditures reported in these data to obtain benchmarks. For public utilities, communication, real estate, and thefor-profit portions of professional services, benchmarks were based on acombination of data from the P;E sample and data from the IRSStatistics of Income. The ratio of universe gross depreciable assetsreported by the IRS to the gross depreciable assets reported bycompanies in the P;E sample was multiplied by the expendituresreported in the P;E sample to obtain benchmarks.
For finance, the basic procedure just described was followed,supplemented with balance sheet and income data compiled by the Board ofGovernors of the Federal Deposit Insurance Corporation, and the FederalHome Loan Bank Board. This basic procedure was also used for insurance,except that total assets were used instead of gross depreciable assets. For membership organizations and social services and for thenot-for-profit portions of professional services, IRS data were notavailable. Benchmarks for those industries were based on data from theCensus Bureau’s County Business Patterns on employment and numberof establishments (instead of IRS gross depreciable assets) inconjunction with P&E sample data. The benchmarks for most of theseindustries also incorporated data from other government agencies andvarious professional associations. For those industries stratified by tab group, benchmarks wereallocated by tab group based on the size distribution of total assets.
Asset data were obtained from the IRS Statistics of Income, fromtabulations published and unpublished Enterprise Statistics data, andfrom responses by the P&E sample. Definitional revisions In addition to the benchmarks, the revised estimates incorporatethe two definitional revisions. For railroads, the estimates of newlycapitalized track maintenance for 1979-82–years for which railroadswere required to restate their financial reports on the new basis–wereconstructed by multiplying previously expensed track maintenance by theproportion of that expense that was capitalized in the restated reportsto the ICC.
For earlier periods, the 1979 proportion was applied totrack maintenance expenditures. For the definitional revision involvingaircraft and railroad equipment for lease, the expenditures werereassigned, for 1959 and later years, from air and railroadtransportation to the finance industry. Quarterly expenditures series The preparation of the revised quarterly expenditures series wascarried out by tab group in three steps: estimating 1977 quarterlyexpenditures, interpolating the quarterly series between the 1972 and1977 benchmarks, and extrapolating the quarterly series forward from the1977 benchmarks. Estimates of expenditures for each quarter of 1977 were based onthe 1977 benchmarks and the quarterly pattern of the retabulated P&Esample. The quarterly samples of company reponses were retabulated toincorporate the following improvements: (1) Reports received too late tobe included in the previously published estimates were included; (2)each company in the sample was reclassified by industry and size on thebasis of data reported in 1980; and (3) the quarterly samples werereedited. A preliminary revised estimate for the fourth quarter of 1976was constructed by multiplying the previously tabulated estimate forthat quarter by the ratio of the 1977 benchmark to the previouslytabulated 1977 estimate. Preliminary revised estimates for the quartersof 1977 were obtained by linking them to the preliminary estimate forthe fourth quarter of 1976 using a chain of link relatives derived fromthe retabulated samples.
These revised quarterly estimates for 1977were then forced to the 1977 benchmark using the ratio of the 1977benchmark to the sum of the preliminary revised estimates for the fourquarters. The expenditures estimates for the quarters between the previousbenchmark year (1972) and the current benchmark year (1977) were revisedusing an interpolation procedure. The percentage difference between thepreviously published estimate for the first quarter of 1977 and thefinal revised estimate for that quarter was distributed geometrically tothe quarters from the first quarter of 1973 through the fourth quarterof 1976. The expenditures estimates for most industries were not revisedprior to 1973. However, as noted earlier, the estimates for airtransportation and finance were revised back to 1959, and the estimatesfor railroad transportation were revised back to 1947. In these cases,the revised quarterly estimates for each tab group for years prior to1973 were obtained by distributing revised annual estimates using thepreviously tabulated quarterly series. Expenditures estimates for successive quarters starting in thefirst quarter of 1978 were revised by extrapolating forward from thefinal revised estimates for the fourth quarter of 1977 using a chain oflink relatives derived from the retabulated samples. Expenditures datafrom sources other than the P&E survey were used as checks,especially when the sample was weak.
Planned expenditures The revised estimates for planned expenditures were prepared in twosteps. For plans reported in 1972-76, ratios of previously reportedplanned to actual expenditures for each tab group for each planninghorizon were multiplied by the revised actual expenditures to obtainrevised planned expenditures. For plans reported in 1977 and later years, the universe estimatesfor plans by tab group were revised in the same way as those for actualexpenditures for that period. The revised planned expenditures for eachperiod were then adjusted for systematic reporting biases using theprocedure described in technical note 1. Expenditures for plant and for equipment The methodology for preparing separate benchmarks for plant and forequipment expenditures was similar to that used in preparing benchmarksfor total P&E expenditures. In most cases, the same sources thatprovided benchmark information on total P&E expenditures providedinformation on the breakdown between plant and equipment. There were afew exceptions–for example, for public utilities, the breakdowns werebased primarily on reports from the Federal Power Commission and theAmerican Gas Association, and for petroleum pipelines, they were basedon the Census Bureau’s Value of New Construction Put in Place.
The interpolation back to 1972 and extrapolation forward from 1977were prepared using the separate data for plant and for equipment fromthe P;E sample in the same way as described earlier for totalP;E expenditures. Because the response rate for reporting plant andequipment expenditures separately is lower than that for reporting totalP;E expenditures, the estimates for plant and for equipment wereforced to equal the total P;E estimates for each tab group. 3. Comparison of the P;E Series with the NonresidentialFixed Investment Component of GNP The nonresidential fixed investment (NRFI) series, which is acomponent of GNP, differs from the P;E expenditures series in typeof detail, data sources, coverage, and timing.
First, the NRFI seriesprovides estimates of investment by type of structure and by type ofproducers’ durable equipment. The P&E series providesestimates of investment by industry. Second, the NRFI series is estimated primarily from Census Bureau surveys of the value of new construction put in place andmanufacturers’ shipments of equipment, with adjustments formerchandise exports and imports and for government purchases. TheP;E series is estimated primarily from BEA’s surveys of plantand equipment expenditures.
Third, the coverage of expenditures in the two series differs.Investment in the farm sector is included in NRFI, but not in theP&E series. Certain industries–real estate; professional services;membership organizations and social services; and forestry, fisheries,and agricultural services–are surveyed only annually in the P&Eseries, and, therefore, are not included in the quarterly “allindustries” total; they are included in NRFI each quarter. Inaddition, certain outlays not capitalized by business (and thus notincluded in the P&E series) are included in the NRFI estimates–forexample, outlays for new motor vehicles held for less than 1 year, forsome oil and gas well drilling costs, and for some other outlaysassociated with mining.
NRFI also includes, but the P&E seriesexcludes, reimbursable expenditures for new motor vehicles by employeesfor business use, and net purchases of, and broker/dealer margins on,used plant and equipment. The P&E expenditures series includes, butNRFI excludes, expenditures for certain types of residential structures,such as dormitories (which are included as residential investment in theNIPA’s). Fourth, the timing of the two series differs. The NRFI seriesreflects the value of new construction put in place and shipments ofequipment; the P;E series reflects expenditures. On balance,expenditures tend to lag.
Table 6 shows NRFI and P;E expenditures (adjusted to thedefinitions and coverage of NRFI) for P;E benchmark years and foryears since the current benchmark year. (Chart 7 shows the two seriesfor 1960-83.) For 1977-83, adjusted P;E expenditures increased 60.4percent, compared with 72.0 percent for NRIF.
As a percentage ofadjusted P;E expenditures, the absolute difference averaged 2.1percent for 1977-83, down from 6.6 percent for the interbenchmark period1972-77.