Labor Market Conditions Index LMCI

The LMCI is a factor model. (A factor model is a statistical tool intended to extract a small number of unobserved factors that summarize the comovement among a larger set of correlated time series.)

The LMCI is derived from a dynamic factor model that extracts the primary common variation from 19 labor market indicators:

1. Unemployment rate

2. Labor force participation rate

3. Part time for economic reasons

4. Private payroll employment

5. Government payroll employment

6. Temporary help employment

7. Average weekly hours (production)

8. Average weekly hours of persons at work

9. Average hourly earnings (production)

10. Composite help-wanted index

11. Hiring rate

12. Transition rate from unemployment to employment

13. Insured unemployment rate

14. Job losers unemployed less than 5 weeks

15. Quit rate

16. Job leavers unemployed less than 5 weeks

17. Jobs plentiful v. hard to get

18. Hiring plans

19. Jobs hard to fill

It is possible that the history of the LMCI may revise each month. There are three sources that contribute to revisions:

1. New data that were not available at the time of the employment report. In particular, at the time of the Employment Situation report each month, the quit rate and hiring rate will be missing for the last two months of the sample because the Job Openings and Labor Turnover Survey is published with a longer lag than the model's other indicators. In subsequent months, as these data become available, the LMCI will revise.

2. Revisions to existing data. Many labor market indicators are subject to revision as additional source data become available or to incorporate annual benchmark revisions or updated seasonal adjustment factors. Prominent examples in the LMCI include the three payroll employment series from the Current Employment Statistics program.

3. Inherent to the model. The LMCI is derived from the Kalman smoother, meaning that the estimate of the index in any particular month is the model's best assessment given all past and future observations. Thus, when a new month of data is added to the sample, the model will revise its estimate of history in response to the new information. In practice, these revisions tend to be modest and concentrated in the most-recent six months of the sample.