Bureau of Labor Statistics > Productivity > Methods > Technical Notes

New Method for Estimating Labor Composition for Total Factor Productivity Measurement

Starting with data released on November 18, 2022, the Total Factor Productivity for Major Industries release will incorporate new methodology that uses a combination of the Census American Community Survey (ACS) and BLS Current Population Survey (CPS) to estimate labor composition. This improvement is the result of a multiyear partnership with the Bureau of Economic Analysis to investigate more robust and innovative ways to measure demographics that involve more than one dataset.

The New Method

The following approach is used to construct employment, hours, and compensation estimates cross-classified by the demographic groups used to estimate labor composition. First, initial estimates for each year, industry, and demographic group using the 1-year American Community Survey are constructed. Second, small area estimation (SAE) is used to refine the direct estimates by also using the 5-year American Community Survey. Small area estimation constructs new estimates of the complete cross-classification of workers as a weighted average of the direct data-based estimate and a model-based estimate, where the weights depend on the uncertainty in direct data-based estimate relative to the model-based estimate. Third, second job information from the Current Population Survey (CPS) is used to adjust workers across industries to account for multiple job holders. Fourth, iterative proportional fitting is applied to balance the industry-level hours and wages to aggregate-level hours and wages based on the CPS. Lastly, industry-level hours are scaled to OPT published hours.

ACS data is available starting in 2005 thus the new methodology is only applied to 2005-forward labor composition indexes. For 1987-2004, indexes based on the new methodology are linked to indexes estimated that solely use the CPS for demographic information. The following table details the estimation steps for both time periods.

Table 1. 1987-2004 and 2005-forward Method Processes
Step	1987-2004 Method	2005-forward Method
1	Monthly CPS data is used to construct annual industry hours and wages for detailed age, education, and sex groups.	The American Community Survey (ACS) 1-year survey is used to construct annual industry hours and wages for detailed age, education, and sex groups.
2	Industry hours are then scaled to BLS productivity published hours.	Small area estimation is applied to the ACS 1-year estimates to address high sampling variance estimates.
3	Multiple imputes regression methodology is applied to fill in missing industry demographic data.	Industry hours and wages are controlled to aggregate hours and wages from the CPS.
4	2-year industry estimates are calculated using only the cohorts that exist in both years.	Industry hours are then scaled to BLS productivity published hours.
5	A 3-year moving average is applied to the series to reduce volatility.	A full time series of industry cohorts is used to calculate labor composition.

An additional change with the new method is a change in the demographic groups for each industry. The following groups are used in both the 1987-2004 and 2005-forward method and reflect a merging of the methods between the BEA and the BLS.

Table 2. Demographic Groupings
Demographic Group	Previous Method	Current Method
Age Group	NA	<16
Age Group	16-18	16-17
Age Group	19-24	18-24
Age Group	25-34	25-34
Age Group	35-44	35-44
Age Group	45-55	45-54
Age Group	55-64	55-64
Age Group	65+	65+
Education group	NA	Grades 0-8
Education group	High school but no diploma	Grades 9-12
Education group	High school graduate	High school diploma
Education group	Some college -Associate	Some college
Education group	Bachelors	Bachelors
Education group	Advanced degree	More than a bachelors
Class of Worker	Not broken out	Employee
Class of Worker	Not broken out	Self-employed
Note: NA indicates a breakout equivalent does not exist

The Impact of using the New Labor Composition Methodology

The new methodology is a notable improvement over the previous methodology. The large sample size of the ACS and the use of small area estimation to address high sampling variance estimates are combined with CPS data that more closely correspond to the nationally produced estimates of employment that serve as the basis of our labor input measure. The new methodology discontinues the use of three-year moving averages and reflects more quickly changes in industry labor composition. This is particularly evident with the onset of the COVID-19 pandemic in 2020, where the improved method more accurately captures the speed of change in many industries. The new estimates also align more closely with the national accounts at BEA and provide a more consistent measure that spans productivity and national accounts estimates, while resulting in only minimal revisions over time.

See Table 3 for comparison of growth rates between the original and current methodology for the time periods. 1987-2020, 1987-2005, and 2005-2020.

Table 3. Revisions of labor composition growth rates
Industry	1987-2020 Original	1987-2020 Current	1987-2005 Original	1987-2005 Current	2005-2020 Original	2005-2020 Current
Private business sector	0.5	0.5	0.5	0.5	0.5	0.6
Private nonfarm business sector	0.5	0.5	0.5	0.5	0.5	0.6
Crop & animal production (Farms)	0.2	0.4	0.4	0.3	0.0	0.4
Forestry, fishing, and related activities	0.2	0.3	0.4	0.3	0.0	0.3
Oil and gas extraction	0.5	0.3	0.3	0.3	0.7	0.2
Mining, except oil and gas	0.5	0.3	0.4	0.3	0.7	0.3
Support activities for mining	0.2	-0.1	-0.6	-0.6	1.1	0.5
Utilities	0.4	0.4	0.5	0.5	0.3	0.3
Construction	0.3	0.4	0.3	0.3	0.4	0.5
Manufacturing sector	0.6	0.6	0.6	0.7	0.5	0.6
Durable manufacturing sector	0.6	0.6	0.6	0.6	0.5	0.6
Nondurable manufacturing sector	0.7	0.7	0.7	0.7	0.7	0.8
Wood products	0.4	0.5	0.5	0.5	0.3	0.4
Nonmetallic mineral products	0.4	0.5	0.4	0.4	0.2	0.5
Primary metal products	0.4	0.5	0.4	0.4	0.4	0.6
Fabricated metal products	0.4	0.4	0.4	0.5	0.4	0.4
Machinery	0.5	0.6	0.6	0.6	0.3	0.6
Computer and electronic products	0.9	0.8	1.1	1.1	0.6	0.4
Electrical equipment, appliances, and components	0.6	0.6	0.5	0.5	0.7	0.7
Motor vehicles, bodies and trailers, and parts	0.5	0.5	0.5	0.5	0.4	0.5
Other transportation equipment	0.6	0.5	0.7	0.7	0.4	0.3
Furniture and related products	0.6	0.5	0.6	0.6	0.5	0.4
Miscellaneous manufacturing	0.7	0.8	0.8	0.8	0.6	0.8
Food and beverage and tobacco products	0.5	0.5	0.4	0.4	0.5	0.6
Textile mills and textile product mills	0.7	0.8	0.8	0.8	0.6	0.9
Apparel and leather and applied products	1.2	1.2	1.2	1.2	1.3	1.3
Paper products	0.5	0.5	0.5	0.5	0.5	0.5
Printing and related support activities	0.4	0.3	0.3	0.2	0.5	0.4
Petroleum and coal products	0.4	0.4	0.4	0.4	0.3	0.4
Chemical products	0.6	0.6	0.6	0.6	0.6	0.6
Plastics and rubber products	0.7	0.6	0.5	0.6	0.8	0.6
Wholesale trade	0.4	0.5	0.6	0.5	0.3	0.5
Retail trade	0.2	0.3	0.3	0.3	0.2	0.4
Air transportation	0.5	0.5	0.6	0.6	0.3	0.4
Rail transportation	0.2	0.2	0.1	0.1	0.3	0.3
Water transportation	0.5	0.3	0.3	0.3	0.7	0.3
Truck transportation	0.3	0.3	0.3	0.4	0.2	0.2
Transit and ground passenger transportation	0.3	0.2	0.3	0.3	0.3	0.2
Pipeline transportation	0.2	0.1	0.2	0.3	0.1	-0.1
Other transportation and support activities	0.3	0.3	0.5	0.5	0.1	0.1
Warehousing and storage	0.1	0.0	0.1	0.0	0.1	-0.1
Publishing industries, except internet (includes software)	0.9	0.9	0.9	0.9	0.8	1.0
Motion picture and sound recording industries	0.6	0.6	0.6	0.5	0.5	0.7
Broadcasting and telecommunications	0.4	0.4	0.3	0.3	0.5	0.6
Data processing, internet publishing, and other information services	0.5	0.8	0.4	0.3	0.6	1.3
Federal reserve banks, credit intermediation, and related activities	0.8	0.8	0.7	0.6	1.0	1.0
Securities, commodity contracts, and other financial investments and related activities	0.6	0.6	0.7	0.8	0.4	0.4
Insurance carriers and related activities	0.5	0.5	0.5	0.5	0.5	0.6
Funds, trusts, and other financial vehicles	0.6	0.6	0.8	0.8	0.4	0.4
Real estate	0.3	0.3	0.3	0.3	0.3	0.4
Rental and leasing services and lessors of nonfinancial and intangible assets	0.3	0.3	0.1	0.1	0.5	0.5
Legal services	0.4	0.3	0.4	0.4	0.4	0.3
Computer systems design and related services	0.7	0.6	0.9	0.9	0.3	0.2
Miscellaneous professional, scientific, and technical services	0.4	0.4	0.6	0.6	0.2	0.2
Management of companies and enterprises	0.3	0.5	0.6	0.5	0.1	0.6
Administrative and support services	0.3	0.4	0.2	0.2	0.3	0.5
Waste management and remediation services	0.2	0.4	0.2	0.2	0.3	0.6
Educational services	0.3	0.4	0.4	0.3	0.2	0.4
Ambulatory health care services	0.0	0.0	-0.1	-0.1	0.1	0.1
Hospitals and nursing and residential care facilities	0.7	0.7	0.7	0.7	0.8	0.7
Social assistance	0.3	0.4	0.0	0.2	0.6	0.7
Performing arts, spectator sports, museums, and related activities	0.1	0.3	0.3	0.4	-0.1	0.1
Amusements, gambling, and recreation industries	0.4	0.4	0.4	0.3	0.5	0.6
Accommodation	0.4	0.5	0.5	0.5	0.3	0.6
Food services and drinking places	0.2	0.3	0.3	0.2	0.2	0.4
Other services, except government	0.2	0.3	0.3	0.3	0.1	0.4

Date Last Modified: November 18, 2022