Productivity is a measure of economic efficiency which shows how effectively economic inputs are converted into output.
Advances in productivity — the ability to produce more with the same or less input — are a significant source of increased potential national income. The U.S. economy has been able to produce more goods and services over time, not by requiring a proportional increase of labor time, but by making production more efficient.
Productivity is measured by comparing the amount of goods and services produced with the amount of inputs which were used in production. BLS produces measures for both labor productivity and total factor productivity. Labor productivity is the ratio of the output of goods and services to the labor hours devoted to the production of that output. A change in labor productivity reflects the change in output that cannot be accounted for by a change in labor hours. Total factor productivity relates output to the combination of inputs used in the production of that output, such as labor and capital or capital, labor, energy, materials, and purchased business services (KLEMS). Capital includes equipment, structures, inventories, and land. A change in TFP reflects the change in output that cannot be accounted for by the change in combined inputs. As a result, TFP measures reflect the joint effects of many factors including research and development (R&D), new technologies, economies of scale, managerial skill, and changes in the organization of production. Total factor and multifactor productivity are the same, see Division of Productivity Research FAQs.
Output per hour of all persons—labor productivity—is the most commonly used productivity measure. Labor is an easily identified input to virtually every production process. In the U.S. nonfarm business sector, labor costs are more than sixty percent of the value of output produced. Output per hour in the nonfarm business sector is the productivity statistic most often cited by the press.
The most commonly cited total factor productivity measure encompasses the private business sector of the economy. This sector measures the for-profit sector of the economy and is the broadest sector for which total factor productivity measures are available.
An index series is simply a way of expressing, in percentage terms, the change in some variable from a given point in time to another point in time. For example, let's say that output increased by 10 percent from an initial year (1987) to a subsequent year (1988). The index for our arbitrarily chosen base year of 1987 would be 100.0 while the index for 1988 would be 110.0. Conversely, if output had declined in 1988 by 10 percent, the 1988 index value would be 90.0.
The index numbers published in the Productivity and Costs news release by BLS are rounded to one decimal place. All percent changes in the press release and on the BLS website are calculated using index numbers to three decimal places. These index numbers are available at the BLS Data Overview website
The percent change for the year is NOT equal to the average of that year's four quarterly percent changes from the previous quarter of the same year. The quarterly percent changes calculated from the previous quarter of the same year do not capture information on the year-to-year movement of the index number.
The percent change for the year should closely approximate the average of that year's four quarterly percent changes from the corresponding quarter of the preceding year. Because the percent change values presented in the official data are rounded to one decimal point, and because of minor variations in the year-to-year values used for seasonal adjustment of the quarterly data, results of a calculation based on press release data will reflect a small difference between the percent change for the year and the average of the quarterly percent changes from the preceding quarter a year ago.
The Division of Productivity Research and Program Development (DPRPD) works on strengthening and improving Bureau productivity measures and on understanding the sources and effects of productivity and technical change. The Division’s economists work on clarifying input and output concepts for productivity measures, using methods from microeconomic and macroeconomic theory, labor economics, industrial organization, econometrics, and statistics. Staff time is devoted partly to individual, long-term research on theoretical and empirical topics and writing working papers and publications. For more on how to obtain research papers or opportunities to be a researcher at BLS, please see: the Division of Productivity Research FAQs.
The BLS’s in-house journal, the Monthly Labor Review (MLR), has a program for publication of original articles from outside authors. You can learn more about the MLR publication process and how to submit your work at this web page: Submitting papers to Monthly Labor Review: guidelines for non-BLS authors. The editors of the MLR can also provide you with additional guidance on this process. They can be reached at the email address firstname.lastname@example.org.
BLS measures of productivity and costs are based on underlying series from a variety of sources. These data sources are frequently revised as additional data become available. When any of these underlying series are updated or revised, the productivity and costs measures are revised to reflect the new information.
Generally, a recent quarter of data is revised two or three times after its initial release. Historical revisions occur when the source data used in their construction are revised. Because many of these sources are revised independently, the productivity measures undergo frequent revision. Revisions to source data include the following:
Output and compensation data: A multi-year revision is generally incorporated into the August press release. This revision occurs when the Department of Commerce revises the National Income and Product Accounts.
Employment and hours data: A five-year revision is generally incorporated into the March press release. This revision occurs when the Bureau of Labor Statistics benchmarks the national establishment employment data.
Manufacturing output data: Revisions to manufacturing output affect a variable number of years and occur on a somewhat variable schedule. Revisions occur when the Board of Governors of the Federal Reserve Board revise their data and are usually incorporated into the December press release. Revisions also occur due to changes in data from the Economic Census and Annual Surveys of Manufactures and to the input-output tables prepared by the Bureau of Economic Analysis.
Some source data needed to construct such measures are not available quarterly.
The total economy includes several sectors in which it is difficult to measure productivity. These include general government, nonprofit institutions, paid employees of private households, and rental of owner-occupied houses. The broadest measure of productivity published by the BLS therefore excludes these sectors and is known as the business sector. Business sector output accounted for about 75 percent of the value of gross domestic product (GDP) in 2010.
The measures of both output and labor input differ at the private business/private nonfarm business level versus the manufacturing sector. As a result, the measures of total factor productivity are conceptually somewhat different for these sectors. To summarize: Total factor productivity for the private business/private nonfarm business sectors relates value-added output to the combined inputs of capital and labor hours. Total factor productivity for the manufacturing sector and manufacturing industries relates sectoral output to the combined inputs of capital, labor hours, and intermediate inputs.
Output for the private business/private nonfarm business sectors is measured as “value-added” output. Value-added output is defined as the total value of goods and services sold in final-demand markets. In other words, the value of goods and services that are used in the production process of other goods and services (known as “intermediate inputs”) are subtracted from the total amount of goods and services produced to determine value-added output.
For the manufacturing sector and manufacturing industries, output is defined as “sectoral output.” Sectoral output is defined as the total value of goods and services sold outside the industry/sector. Sectoral output includes goods and services sold to both final consumers as well as other businesses outside the industry/sector.
The measurement of labor input for the private business/private nonfarm business sectors includes an adjustment for the effect of changing labor composition (which is a measure of the level of skill of the workforce). For the manufacturing sector and manufacturing industries, labor input is a direct aggregate of hours, and changes in labor composition are not measured.
While not available at the industry level, state and regional labor productivity and costs measures are available for the private nonfarm sector starting in 2007. Our sector or industry total factor productivity measures are only available at the national level.
State and regional labor productivity and costs measures are available for the private nonfarm sector starting in 2007. Total factor productivity measures for detailed sectors and industries are only available at the national level. See differences between state and national productivity measures subsection of BLS publishes experimental state level labor productivity measures.
No. BLS productivity measures are based on aggregate national measures of outputs and inputs. These data sources do not provide the information BLS would need to construct occupational measures. There are also conceptual obstacles to disaggregating these national measures. For example, the output of a factory may require both white-collar and blue-collar labor inputs, and it is therefore unclear how to allocate the output to the two groups separately.
While our productivity data does capture the labor composition of the workforce and many inputs affecting productivity (see General Questions and Answer #3), it does not capture where the worker is doing the work.
Our research office has published some older articles you may find interesting related to productivity and working from home.
“Bringing Work Home: Implications for BLS Productivity Measures”, (with Lucy Eldridge). Monthly Labor Review , Vo1. 133, No. 12, December 2010, pp. 18-35.
“Are Those Who Bring Work Home Really Working Longer Hours? Implications for BLS Productivity Measures”, (with Lucy Eldridge). In Productivity Measurement and Analysis: Proceedings from OECD workshops, Julien Dupont and Pierre Sollberger (Eds.), Swiss Federal Statistical Office, Neuchâtel, Switzerland, 2008, pp 179-209.
The Bureau of Labor Statistics (BLS) also has many Monthly Labor Review (MLR) articles related to telework/remote work you may find interesting.
The primary source of hours and employment data is the BLS Current Employment Statistics (CES) program, which provides data on total employment and average weekly hours of production and nonsupervisory workers in nonagricultural establishments.
For the quarterly productivity measures, information from the National Compensation Survey (NCS) is used to convert the CES hours to hours at work by excluding all forms of paid leave. Average weekly hours for nonproduction and supervisory workers are estimated by using data from the Current Population Survey (CPS), the NCS and the CES. Because CES data include only nonagricultural wage and salary workers, data from the CPS are used for farm employment as well as for nonfarm proprietors and unpaid family workers. Government enterprise hours are developed from the National Income and Product Account estimates of employment combined with CPS data on average weekly hours.
For the industry productivity measures, average weekly hours for nonproduction and supervisory workers are estimated using data from the CPS and the CES. The industry labor input measures also include estimates of the hours of proprietors and unpaid family workers from the CPS.
Hours are the number of hours worked by all employed persons, including wage and salary workers, self-employed persons, and unpaid family workers. Hours for wage and salary workers are primarily from BLS Current Employment Statistics (CES) and hours for self-employed and unpaid family workers are from the BLS Current Population Survey (CPS). The hours are adjusted from an hours paid basis to an hours worked basis using data from the BLS National Compensation Survey (NCS). See hours subsection of BLS publishes experimental state level labor productivity measures.
Unit labor costs are calculated by dividing total nominal labor compensation by real output or—equivalently—by dividing hourly compensation by productivity.
That is, unit labor costs = nominal labor compensation \ real output; or equivalently,
unit labor cost = hourly compensation / productivity = [nominal labor compensation / hours] / [output / hours]
Thus, increases in productivity lower unit labor costs while increases in hourly compensation raise them. If both series move equally, unit labor costs will be unchanged.
Capital input is measured as “capital services” — the flow of services from the physical stock of capital assets. The stock of capital is measured using a “perpetual inventory method” as the sum of past investments adjusted for depreciation and retirements. The stocks of different types of capital assets are combined into a single measure by cost share weighting. Capital stocks are multiplied by implicit rental prices to yield cost share weights. For more on intangible capital see Division of Productivity Research FAQs.
The rental price is the amount of rent per year a dollar's worth of capital stock earns. Since the owner and user of capital goods are often the same, the rental price of capital services must be implicitly estimated.
Compensation is a measure of the cost to the employer of securing the services of labor. It includes wages and salaries, supplements (like shift differentials, all kinds of paid leave, bonus and incentive payments, and employee discounts), and employer contributions to employee-benefit plans (like medical and life insurance, workmen's compensation, and unemployment insurance).
Compensation also includes an imputation of the earnings of self-employed workers.
The ECI and the hourly compensation measures may differ because they are designed to measure two different things. There is no formal analysis of the differences between the ECI and those in the hourly compensation series, largely because most of the differences are not quantifiable.
One quantifiable difference, however, is the presentation of percent changes between adjacent periods. Both series present their data as index numbers, one-quarter or 3-month percent change, and four-quarter or 12-month percent change. The index numbers and year-over-year comparisons are consistent; however the one-quarter or 3-month percent changes are presented differently. ECI presents these data as the actual change over the 3-month period while the hourly compensation measures are presented at a compound annual rate of change (i.e., as if the same percent change were to continue for four quarters).
Some of the other differences between the ECI and hourly compensation for all persons in the nonfarm business economy include:
Ownership—The ECI covers private industry and state and local government. Data are available separately or combined. Hourly compensation measures cover private industry plus federal, state and local government enterprises.
Employment coverage—The ECI covers most private employees but excludes persons working for token wages as well as business owners and others who set their own wage (for example, corporate CEO's). The ECI also excludes family workers who do not earn a market wage. The hourly compensation measure excludes employees of nonprofit institutions serving individuals (about ten percent of private workers, most in education and medical care). However, the hourly compensation measures do include an estimate for the unincorporated self-employed because they are assumed to earn the same hourly compensation as other employees in the sector. Implicitly, unpaid family workers also are included in the hourly compensation measures with the assumption that their hourly compensation is zero.
Compensation coverage—The ECI includes wages and benefits that employers provide. The hourly compensation measures include wages and benefits that employees receive. Therefore, some types of compensation (such as tips received by employees) are included in the hourly compensation measures but not in the ECI. Also, the ECI does not include stock options as an employer-provided benefit. For the hourly compensation measures, the accrued employee income from stock options is incorporated into wages and salaries when the option is exercised.
Weights—The ECI uses a fixed set of industry and occupational weights to obtain a measure of the change in labor cost that is not influenced by changes in the industrial and occupational structure of the economy. These weights are updated periodically. The hourly compensation measures are influenced by changes in employment distribution. That is, even if the wages for all jobs in the economy were unchanged, hourly compensation could change if the distribution of jobs in the economy changed. In addition, the ECI holds hours worked constant, unless there is a plan change. Benefit usage is also usually held constant in ECI. Hourly compensation measures will show changes due to changes in hours at work and overtime.
Time period covered—The ECI collects data for specific months: March, June, September, and December. The hourly compensation measures are average compensation for the three months in the quarter.
Framework—The ECI is a sample survey specifically designed for the measurement of changes in labor cost and is subject to the usual sampling and nonsampling errors. The hourly compensation measures are constructed using information from the national income and product accounts as well as the employment and hours of persons working in the business economy. Errors in these data may also come from sampling and nonsampling errors.
Data from government enterprises is included in the calculation of labor productivity measures. Government enterprises are defined as government agencies that cover a substantial proportion of their operating costs by selling goods and services to the public and that maintain their own separate accounts.
Capital measures are generally unavailable for government enterprises. Therefore government enterprises are excluded when total factor productivity is calculated.
Different products are aggregated into one output measure by weighting (multiplying) the relative change in the output of each product by its share in the total value of output. Thus, the products that require more resources to produce are given higher weight.
Business sector output includes a subset of the goods and services included in GDP. The business sector excludes those portions of the economy for which productivity measures are difficult to calculate. General government, the nonprofit institutions, the employment of private households, and the rental value of owner-occupied real estate are excluded.
Within the U.S. business sector, outsourcing of production or services from manufacturing industries to other domestic industries alters the distribution of production among firms. Since firms can differ in their productivity, domestic outsourcing can affect business sector productivity if the contracting firm differs in its productivity from the original firm. Similarly, outsourcing from U.S. manufacturers to businesses located abroad (or offshoring) can affect business sector productivity if the productivity of the production lost to offshoring differs from the productivity of remaining and any new U.S. business sector production. Any effect of outsourcing or offshoring on business sector productivity change is expected to be modest.
Outsourcing and offshoring have the potential for greater effect on labor productivity measures for the manufacturing sector. BLS measures output for the manufacturing sector using the sectoral output concept: gross output less sales between establishments within the manufacturing sector. Unlike the value-added measure for the business sector, this output measure includes the value of intermediate inputs purchased from outside the manufacturing sector, whether purchased from domestic or foreign suppliers. As manufacturing firms outsource or offshore production of intermediate inputs the value of manufacturing output is unchanged, but the shift to U.S. nonmanufacturing or to imported intermediate inputs is accompanied by a reduction in labor hours and therefore an increase in the measure of labor productivity. It is estimated that the growth in imported intermediate inputs contributed 23 percent (0.92 percentage points) of the 3.96 percent average annual growth in labor productivity in the manufacturing sector from 1997-2006. Further discussion can be found in the June 2010 Monthly Labor Review article "Effects of Imported Intermediate Inputs on Productivity."
If data were available on the number of haircuts, shaves, etc. performed, these data could be used just as we would have used data for tons of steel. We might weight haircuts and shaves differentially, but the concept is the same.
Generally, data for the quantities of output produced or the number of times a service has been performed are not available. However, an alternate methodology is available. Barbershops may not know how many haircuts, etc. have been performed, however, they will know how much revenue they have received from these services. Changes in revenues reflect changes in both quantity of output and its price. Price changes are removed by dividing an index of revenue by a price index, the remainder being an index of quantity.
BLS state-level measures of output for the private nonfarm sector are created using GDP by state and industry data published by the Bureau of Economic Analysis (BEA). BEA does not produce a private nonfarm sector measure of real output by state. To create the necessary output series, BLS subtracts several industry components — the farm sector, private households, and owner-occupied housing — from GDP by state using a Fisher ideal index formula. See output subsection of BLS publishes experimental state level labor productivity measures for more information on BEA's methods for nominal and real measures of GDP by state and industry.