NOTICE The EP program is implementing a new method for measuring occupational separations that would replace the current method of measuring replacements needs. Information about the change is available at

Estimating Occupational Replacement Needs

Projections of job growth provide valuable insight into future employment opportunities because each new job created is an opening for a worker entering an occupation. However, opportunities also arise when workers leave their occupations and need to be replaced. In most occupations, replacement needs provide many more job openings than employment growth does.

To project the magnitude of replacement needs, the Bureau of Labor Statistics (BLS) calculates an estimate of openings resulting from workers retiring or otherwise permanently leaving an occupation.(1) Because workers entering an occupation often need training, these replacement needs, added to job openings due to growth, may be used to assess the minimum number of workers who will need to be trained for the occupation. This estimate of replacement needs does not count workers who change jobs but remain in the same occupation.

Developing historical replacement rates

To develop estimates of replacements, BLS uses occupational employment data from the Current Population Survey (CPS), a household survey that collects demographic and employment information about individuals. BLS measures the net change in occupational employment for 13 different age cohorts over a 5-year period.(2) For each occupation, employment data for each age cohort in 2009 were compared with corresponding data in 2014. For example, the number of medical and health services managers aged 30 to 34 years in 2009 was compared with the number of medical and health services managers aged 35 to 39 years in 2014. (See table 1.) The larger employment in 2014 indicated that there were more entrants than separations among individuals aged 30 to 34 years in 2009. The net entrants were recorded as a positive net change for this cohort. Employment of medical and health services managers in the cohort aged 50 to 54 years in 2009, however, showed a decline. Declines in employment for an age cohort measure net separations; increases measure net entrants..

After calculating net change by age cohort, BLS estimates historical replacement needs. Replacement needs are equal to net separations unless total employment for the occupation declined between 2009 and 2014. In these cases, declines in employment are subtracted from the net separations because not all workers who left the occupation were replaced. For each age group, replacement needs are divided by 2009 employment to calculate a historical 5-year replacement rate.

In most occupations, net separations occur mainly in the older age groups, usually above age 40. However, large numbers of net separations of young workers may occur in occupations that have relatively low entrance requirements and pay relatively low wages. Young workers may take jobs in such occupations while obtaining additional education or training and then leave when they qualify for higher paying occupations. Customer services representatives are an example of this type of occupation; net separations occurred for all except the youngest age cohort in this occupation. (See table 1.)

Developing projected replacement rates

Historical replacement rates are used to estimate replacement needs during the 2014–24 decade. First, replacement needs are calculated for the 2014–19 period. Then employment is estimated for 2019 by applying the cohort's historical rate of change to its 2014 employment. Next, the 2019 employment figure is multiplied by the historical replacement rate to calculate replacement needs for the 2019–24 time span. Finally, projected replacement needs for each 5-year period are summed to compute 10-year replacement needs for each occupation. Following are some examples of the process used to project replacement needs by occupation.

Replacement needs were calculated for 2014–19 by multiplying 2014 employment for each age group by its historical replacement rate. For example, in table 2 (numbers in thousands),

2014 employment
(age 50-54)
2009–14 replacement rate
(age 50-54)
2014–19 replacement needs
(age 50-54)
90 × .092 = 8

An estimate of 2019 employment for each age cohort is calculated by applying the cohort's historical rate of change(3) to its 2014 employment.(4) Because the 2014 workforce will age by 5 years by 2019, it is necessary to age the cohort into the next older age group. Continuing with the same example from table 2 yields the following calculation:

2014 employment
(age 50-54)
2009–14 rate of change
(age 50-54)
2019 employment estimate
(age 55-59)
90 × [1 - .092] = 82

The following example from table 2 illustrates a cohort with net entrants—that of workers aged 40–44 years in 2014:

2014 employment
(age 40-44)
2009–14 rate of change
(age 40-44)
2019 employment estimate
(age 45-49)
77 × [1 + .216] = 94

After estimating employment for 2019, BLS calculates replacement needs for 2019–24 by multiplying the historical replacement rate for each age group by the 2019 employment estimate for that age group. Continuing the earlier example of workers aged 50-54 years in 2014, who will have entered the 55- to 59-year-old age group in 2019, results in the following calculation:

2018 employment estimate
(age 55-59)
2009–14 replacement rate
(age 55-59)
2019–24 replacement needs
(age 55-59)
82 × .255 = 21

Summing the number of replacement needs for each of the 5-year periods provides an estimate of replacement needs over the 10-year projection period. The 2014–24 replacement rates are calculated by dividing replacement needs for 2014–24 by 2014 employment.

For some small occupations, the CPS sample size is too small to give reliable results using the cohort method. In these cases, the occupation was assigned the replacement rate of a summary occupational grouping (either the broad occupation or the minor group) which included the detailed occupation and had enough observations to provide more reliable data.

The CPS data used to calculate replacement rates are coded in accordance with U.S. Census Bureau occupation classification systems. However, the occupation classification system changed during the historical period for which data were used. To account for this, BLS matches the old coding system to the new coding system, sometimes combining occupations from one or both classification schemes to make groups which had comparable occupational coverage. Calculations were then performed using those groupings.

The occupational employment projections use occupational codes from the Standard Occupational Classification (SOC) system, which are often more detailed than the U.S. Census Bureau's occupational classifications. To apply CPS data to the occupations in the projections matrix, BLS identifies the CPS occupations that are equivalent to, or that included, a given detailed SOC occupation. The calculated replacement rate is then multiplied by the 2014 employment figure from the National Employment Matrix to determine projected replacement needs for SOC occupations.

Table 1.10 presents projected 2014–24 replacement rates, replacement needs, and job openings for the SOC-based matrix occupations.

Frequently asked questions about replacement needs

Q. How should replacement needs be used?

A. Replacement needs provide an estimate of opportunities that will be available to enter an occupation. They can be used to provide information on job opportunities for future workers and to help with career guidance.

Q. Why are the estimates of growth and replacement needs described as providing a minimum measure of training needs?

A. The number of new entrants needed in an occupation is underestimated because replacement needs are based on net changes and do not capture the gross flows into and out of an occupation. Even if growth and replacement needs perfectly measured the need for new entrants, training needs would still be underestimated because some people who complete training do not enter the occupation for which they qualify.

Q. Do the 2014–24 projected replacement needs assume that future labor market behavior will differ from past behavior?

A. No. The projected replacement needs assume that workers will continue to retire and otherwise exit an occupation at ages similar to those which have been observed in the recent past. Occupation- and age-specific rates for 2009–14 are used in calculating the projected rates. The 2009–14 rates are applied to current occupational age-distribution data to estimate replacement needs for the future. The result is an occupation-specific replacement rate that captures the impact of demographic, but not behavioral, changes.

Q. Are net separations the same as replacement needs?

A. In most occupations, yes. If employment declined during the historical period, however, net separations exceeded replacement needs by the decline in employment. When employment is declining, not all people who separate from an occupation are replaced.

Q. Should a projected decline in employment be subtracted from replacement needs to estimate job opportunities?

A. No. If employment is projected to decline, the number of opportunities resulting from growth is zero and replacement needs constitute the only source of opportunities. When employment declines, net separations increase both because it is more likely that individuals lose their jobs and because fewer are entering the occupation. Replacement needs already capture these effects by reducing net separations by declines in employment. They should not be further reduced by projected employment declines. In the 2014–24 projections, 216 occupations are projected to decline in employment; all are projected to have replacement openings.

Q. If employment is declining rapidly, is it possible for replacement needs to be zero?

A. In the extreme case, yes. For example, assume that, in a limited geographic area, a single firm is the sole employer of tool and die makers. If the firm ceases operation, all tool and die makers in the area will leave the occupation; net separations will equal the decline in employment, and there will be no replacement needs. On a national scale, however, a situation like this is highly unlikely because not all areas of the country share the same market conditions.

Q. Are there any data on replacement needs by industry?

A. No. Estimates of replacement needs are created only for occupations. The BLS does have a survey that collects data on current job openings and labor turnover by industry. (For more information, see the Job Openings and Labor Turnover Survey.) The BLS also has a program that estimates gross job gains and losses by industry. (For more information, see the Business Employment Dynamics program.) Note that the concepts and methods used by these programs differ from those used by the Employment Projections program.


  1 The method described here is used for the 2014–24 projections. BLS will be implementing a new method that measures occupational separations instead of replacements for the 2016–26 projections. Details on the new separations method is available from
  2 To increase the sample size and reduce cyclical fluctuations, employment data from 5 years (2005, 2006, 2007, 2008, and 2009) were averaged and the resulting average was used as the employment figure in the first reference year. Similarly, employment data from 2010, 2011, 2012, 2013, and 2014 were averaged to represent employment 5 years later. To simplify the discussion that follows, the 2005–09 averages are referred to as 2009 data and the 2010–14 averages are referred to as 2014 data.
  3 To apply the historical rate of change, multiply 2014 employment by the sum of 1 and the rate of change found in table 2.
  4 Note that, for this occupation, the replacement rates equal the absolute value of the negative rates of change. This is true for any occupation with increasing employment over the historical period. However, as discussed in the previous section, if employment in the occupation was declining, then the replacement rates were adjusted downward to account for the fact that not every worker who separated was replaced.


Last Modified Date: December 8, 2015