CES is a statistical program of the U.S. Bureau of Labor Statistics. BLS collects this information through a monthly survey of employers; they provide funding to the Center for Workforce Research and Information, an agency of the Maine Department of Labor, to provide local knowledge, feedback, assistance with annual benchmark revisions, and to concurrently publish estimates.
- Each month CES surveys a sample of about 2,700 nonfarm private employers (plus federal, state, and local government employers) in Maine asking them to report the number of jobs, hours, and earnings of workers on their payroll during the week including the 12th day of each month. The sample accounts for just 7% of employers, but 34% of nonfarm jobs in the state because a high share of large employers are surveyed.
- This data provides the basis for estimates for the state and three metropolitan areas: Bangor, Lewiston-Auburn, and Portland-South Portland-Biddeford.
- Preliminary estimates for a month are published in the latter part of the following month (the release schedule is here). Estimates for that month are revised the following month and those revisions are issued along with the release of new preliminary estimates. For example, June preliminary estimates are published in July. June revised estimates are published in August concurrent with the release of July preliminary estimates.
- Monthly estimates are further revised each spring to the complete job count of payroll jobs reported by employers through their quarterly unemployment insurance tax. These “benchmark revisions,” published each year in March, use the payroll data (which lags by six months), to replace monthly estimates for the 12-month period from October through the most recent September. Once those figures are established, estimates for the three month period from the most recent October through December are re-estimated (actual payroll data for that period is not yet available in March).
- Though the sample from which these estimates are derived is statistically valid, job trends of reporters will not perfectly match the trend of all employers. CES estimates for some industries have a large margin of error.
- Seasonal adjustment is used to eliminate normal seasonal changes to examine the underlying trend in jobs across seasons. Because the timing of holidays, poor weather, the start and end of school semesters, and other events are sometimes different than in the past, and because the week including the 12th of the month may be in the second or third work week, seasonal adjustment sometimes causes the change in jobs to be over or understated for a given month.
- Not all employers in the sample report every month. Non reporters can cause too much weighting to those that did report in a given month, skewing the estimates.
- Weighting factors are applied to many small employers in the sample to make changes in their employment representative of other small employers in the same industry. A small employer that added 4 jobs with a weighting factor of 80 pushes the estimate for that industry up 320. This is not always indicative of other employers, of course. A few small employers moving in the same direction (adding or subtracting jobs) with high weighting factors can have an outsize influence on the monthly change in the estimate of jobs in an industry.