Source: Board of Governors of the Federal Reserve System, Capacity Utilization In Manufacturing, Mining, Utilities, and IndustrialMaterials, G.3., monthly. (Based on data from Federal Reserve Board,Commerce, U.S. Bureau of Labor Statistics, and McGraw-Hill InformationSystems Company, New York, NY; and other sources.)

http://www.federalreserve.gov/releases/*

Industrial Production

Coverage. The industrial production (IP) index measures output in themanufacturing, mining, and electric and gas utilities industries; the reference periodfor the index is 1992. For the period since 1992, the total IP index has beenconstructed from 264 individual series based on the 1987 Standard IndustrialClassification (SIC). These individual series are classified in two ways: (1) marketgroups (shown in table 1), such as consumer goods, equipment, intermediateproducts, and materials; and (2) industry groups (shown in tables 2 and 6), such astwo-digit SIC industries and major aggregates of these industries-for example,durable and nondurable manufacturing, mining, and utilities.

Market groups. For purposes of analysis, the individual IP series are grouped intofinal products, intermediate products, and materials. Final products are assumed to bepurchased by consumers, businesses, or government for final use. Intermediateproducts are expected to become inputs in nonindustrial sectors, such asconstruction, agriculture, and services. Materials are industrial output requiringfurther processing within the industrial sector. Total products comprise final andintermediate products, and final products are divided into consumer goods andequipment.

Timing. The first estimate of output for a month is published around the 15th of thefollowing month. The estimate is preliminary (denoted by the superscript "p" intables) and subject to revision in each of the subsequent three months as new sourcedata become available. (Revised estimates are denoted by the superscript "r" intables.) After the fourth month, indexes are not revised further until the time of anannual revision or a benchmark revision. The last three benchmark revisions werepublished in 1990, 1985, and 1976.

Source data. In annual or benchmark revisions, the individual IP indexes areconstructed from a variety of source data, such as the quinquennial Censuses ofManufactures and Mineral Industries and the Annual Survey of Manufactures,prepared by the Bureau of the Census; the Minerals Yearbook, prepared by theDepartment of the Interior; and publications of the Department of Energy. On amonthly basis, the individual indexes of industrial production are constructed fromtwo main types of source data: (1) output measured in physical units and (2) data oninputs to the production process, from which output is inferred. Data on physicalproducts, such as tons of steel or barrels of oil, are obtained from private tradeassociations as well as from government agencies including those listed above; dataof this type are used to estimate monthly IP where possible and appropriate. Whensuitable data on physical product are unavailable, estimates of output are based oneither production-worker hours or electric power use by industry. Data on hoursworked by production workers are collected in the monthly establishment surveyconducted by the Bureau of Labor Statistics. The data on electric power use aredescribed below. The factors used to convert inputs into estimates of production arebased on historical relationships between the inputs and the comprehensive data usedto benchmark the IP indexes; these factors also may be influenced by technologicalor cyclical developments. Especially for the first and second estimates for a givenmonth, the available source data are limited and subject to revision.

Weights. In the index, series that measure the output of an individual industry areweighted according to their proportion in the total value-added output of allindustries. The industrial production index, which extends back to 1919, is built as anannually weighted chain-type index since 1977. The components of IP are combinedusing estimates of value added per unit of output. For months from January to June,the weights are drawn from the year containing the month being estimated and thepreceding year; for months from July to December, the weights are drawn from thecurrent and following year. The IP proportions shown in column 1 of tables 1A, 2A,and 6 are estimates of the industries' relative contributions to overall growth in thefollowing year. For example, a 1 percent increase in durable goods manufacturing in1997 would account for an increase in total IP of nearly 1/2 percent.

Seasonal adjustment. Individual series are seasonally adjusted by the X-11 ARIMAmethod, developed at Statistics Canada. For series based on production-workerhours, the current seasonal factors were estimated with data through October 1996;for other series, the factors were estimated with data through at least June 1996. Insome cases, series were preadjusted for the effects of holidays or the business cyclebefore using X-11 ARIMA. For the data since 1977, all seasonally adjustedaggregate indexes are calculated by aggregating the seasonally adjusted indexes ofthe individual series.

Reliability. The average revision to the level of the total IP index, without regard tosign, between the first and the fourth estimates was 0.28 percent during the 1987-96period. The average revision to the percent change in total IP, without regard to sign,from the first to the fourth estimates was 0.21 percentage point during the 1987-96period. In most cases (about 81 percent), the direction of change in output indicatedby the first estimate for a given month is the same as that shown by the fourth estimate.

Rounding. The published percent changes are calculated from unrounded indexes,and may not be the same as percent changes calculated from the rounded indexesshown in the release.

Capacity Utilization

Definition. Capacity utilization is calculated for the manufacturing, mining, andelectric and gas utilities industries. For a given industry, the utilization rate is equal toan output index divided by a capacity index. Output is measured by seasonallyadjusted indexes of industrial production. The capacity indexes attempt to capturethe concept of sustainable practical capacity, which is defined as the greatest level ofoutput that a plant can maintain within the framework of a realistic work schedule,taking account of normal downtime, and assuming sufficient availability of inputs tooperate the machinery and equipment in place. The 76 individual capacity indexesare based on a variety of data, including capacity data measured in physical unitscompiled by trade associations, surveys of utilization rates and investment, andestimates of growth of the capital input.

Groups. Estimates of capacity and utilization are available for a variety of groups,including primary and advanced processing industries within manufacturing,durable and nondurable manufacturing, total manufacturing, mining, utilities, andtotal industry. Component industries of the primary and advanced processing groupswithin manufacturing are listed in the note on tables 2 and 3 of the release.

Weights. Although each utilization rate is the result of dividing an IP series by acorresponding capacity index, aggregate utilization rates are equivalent tocombinations of individual utilization rates aggregated with proportions that reflectcurrent capacity levels of output valued in current-period value added per unit ofactual output. The implied proportions of individual industry operating rates in therate for total industry for the most recent year are shown in the first column of table 3.

Perspective. The historical highs and lows in capacity utilization shown in the tablesabove are specific to each series and did not all occur in the same month. Industrialplants usually operate at capacity utilization rates that are well below 100 percent:none of the broad aggregates has ever reached 100 percent. For total industry andtotal manufacturing, utilization rates have exceeded 90 percent only

*