An official website of the United States government

The .gov means it's official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you're on a federal government site.

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Bureau of Labor Statistics > Employment Projections > Data

Labor Force Projections Evaluation: 2012–2022

Labor force projections include population, labor force participation rates, and labor force level components. Population projections use Census Bureau resident population projections as a starting point and are benchmarked to civilian noninstitutional population (CNIP) estimates from the Current Population Survey (CPS). Participation rates are modeled using historic trends in the data. Participation rates and population projections are then multiplied together to estimate labor force levels. Errors in BLS labor force projections can therefore come from either source.

Limitations

In 2003, the CPS added a new race category, “Two or more races,” and adjusted the way data on ethnicity was collected. These changes had previously limited the ability for BLS to evaluate the accuracy of our labor force projections by race and ethnicity because of inconsistencies with the historical data. BLS now has enough consistent data by race and ethnicity to estimate the naïve model using nine years of historical data. The naïve models, by sex, use 10 years of historical data, but the race and ethnicity comparisons in this round of evaluations only use nine years of historical data to estimate the naïve model.

Measuring accuracy

How often did BLS correctly project growth and decline for labor force segments?

BLS correctly projected which labor force segments would grow and which would decline 88 percent of the time.¹ BLS incorrectly projected declines in the labor force for the youngest age group of ages 16-19 for both men and women. All other BLS projections matched the actual direction.

How much did BLS project the labor force to grow between 2012 and 2022?

BLS projected the labor force to grow 5.5 percent between 2012 and 2022.

How much did the labor force actually grow?

The labor force actually grew 6.0 percent between 2012 and 2022.

Accuracy of the CNIP projections

The 2022 actual CNIP was similar to the projected CNIP, with an absolute difference of about 0.5%, and dissimilarity indexes for all races of about 1.0%. (See Table 1.) The “all other groups” category for race had the highest dissimilarity indexes at 1.7% for men and 1.5% for women.² (See Table 1a.) For ethnicity, Hispanic men had the highest dissimilarity index at 2.5%.³(See Table 1b.)

Table 1. CNIP dissimilarity indexes
All	Male	Female
1.0%	1.1%	0.9%

Table 1a. CNIP dissimilarity indexes by race
Race	Men	Women
White	1.2%	1.0%
Black	1.5%	1.1%
All other groups	1.7%	1.5%

Table 1b. CNIP dissimilarity indexes by ethnicity
Ethnicity	Men	Women
Hispanic origin	2.5%	0.9%

Accuracy of the labor force projections

The BLS and naïve models performed similarly for the overall labor force. BLS slightly underestimated the total labor force, and the naïve model slightly overestimated the total labor force. The naïve model performed better for men, and the BLS projections performed better for women. (See Table 2.)

Table 2. Absolute percent error by sex, 2012–22 (Numbers in thousands)
Sex	Actual 2022 labor force	Labor force projections		Absolute percent error		Best performer
Sex	Actual 2022 labor force	BLS	Naïve	BLS	Naïve	Best performer
All	164,288	163,450	165,134	0.5%	0.5%	BLS
Men	87,421	86,913	87,280	0.6%	0.2%	Naïve
Women	76,867	76,537	77,882	0.4%	1.3%	BLS
Notes: For the naïve model, details do not sum to totals because separate models were run for Men, Women, and All. Actual 2022 data may not match CPS data due to rounding at different aggregate levels. The absolute percent error for the BLS and naïve models are 0.510% and 0.515%, respectively, for all sexes.

BLS outperformed the naïve model in every race by sex category. Both the BLS projections and the naïve model performed worse for Black workers and workers of all other races compared to White workers. BLS consistently overpredicted the number of White workers and consistently underpredicted the numbers of Black workers and workers of all other races. (See Table 3.)

Table 3. Absolute percent error by race and sex, 2012–22 (Numbers in thousands)
Race	Sex	Actual 2022 labor force	Labor force projection		Absolute percent error		Best performer
Race	Sex	Actual 2022 labor force	BLS	Naïve	BLS	Naïve	Best performer
White	Men	68,163	68,989	69,408	1.2%	1.8%	BLS
White	Women	57,795	57,934	59,979	0.2%	3.8%	BLS
Black	Men	10,259	9,547	9,360	6.9%	8.8%	BLS
Black	Women	10,977	10,700	10,513	2.5%	4.2%	BLS
All other groups	Men	8,998	8,377	7,713	6.9%	14.3%	BLS
All other groups	Women	8,094	7,903	6,909	2.4%	14.6%	BLS
Note: Actual 2022 data may not match CPS data due to rounding at different aggregate levels.

BLS performed better than the naïve model at estimating the Hispanic labor force. (See Table 4.)

Table 4. Absolute percent error by ethnicity and sex, 2012–22 (Numbers in thousands)
Ethnicity	Sex	Actual 2022 labor force	Labor force projection		Absolute percent error		Best performer
Ethnicity	Sex	Actual 2022 labor force	BLS	Naïve	BLS	Naïve	Best performer
Hispanic	Men	17,370	17,925	16,672	3.2%	4.0%	BLS
Hispanic	Women	13,232	13,254	12,847	0.2%	2.9%	BLS
Non-Hispanic	Men	70,051	68,988	69,674	1.5%	0.5%	Naïve
Non-Hispanic	Women	63,635	63,283	64,857	0.6%	1.9%	BLS
Note: Actual 2022 data may not match CPS data due to rounding at different aggregate levels.

Note

¹All detailed age groups for men and women.

²The "all other groups" category includes (1) those classified as being of multiple racial origin and (2) the racial categories of (2a) Asian (2b) American Indian and Alaska Native and (2c) Native Hawaiian and Other Pacific Islanders.

³There is no comparable non-Hispanic group for calculating the projected and actual CNIP dissimilarity indexes. The 2012–22 BLS Projections had data by sex and age group for White non-Hispanic origin. The 2022 actual data by sex and age group is only available for all non-Hispanic origin.

Return to Projections Evaluation Homepage

Last Modified Date: January 19, 2024