The Labor and Economic Analysis Division (LEAD) prepares projections of employment growth by industry and occupation for the state and sub-state areas. Employment projections are widely used by North Carolina’s workforce, educational, and economic development partners for their planning in workforce development, programs and budgets, public policy, and career exploration activities.
The 2017-2026 Employment Projections at the state and sub-state areas are based on a 10-year projection (2016-2026) with an updated data point estimate for 2017. The long-term projections are revised every two years to maintain currency and incorporate economic changes that occur in the state and local areas. Statewide short-term projections are for a 2-year period, and are updated annually.
Projections are prepared using the methodology, software tools and guidance developed by the Projections Managing Partnership (PMP) in conjunction with the U.S. Department of Labor. The long-term industry and occupation projections are produced every two years while the short-term projection is prepared every year. Sub-state areas are prepared biannually with a focus on the Prosperity Zone sub-regions.
LEAD utilizes industry employment data derived from the Enhanced Quarterly Unemployment Insurance (EQUI) dataset. It is the most complete and timely source of monthly employment and quarterly wages information by detailed industry and county. The data contains a quarterly count of employment and wages report that is sent from employers based on the North American Industry Classification System (NAICS) code. Employment data on uncovered industries within the Unemployment Insurance (UI) program is collected from other sources such as Current Employment Statistics (CES), Census Bureau, and Railroad Retirement Board. The EQUI dataset also forms the base for federal data programs through the BLS.
The employment data passes through multiple phases of data processing and analysis. Historical data is first cleaned to ensure consistent formatting and validity then aggregated by NAICS for all detail levels. The data is also aggregated from the county level to sub state areas and statewide.
The second phase involves importing the historical data into the industry projections system. The industry projections system has multiple estimation models. The analyst chooses the model that best fits the historical data among the included shift-share, time series and regression models. Outside sources of information are also valuable in the projections process. Industry expert opinion, current events, and objective national and regional input all play a role in producing a reasonable estimate. Economic indicator variables, such as population or retail sales, are used in the projection process after analyzing the historical data series to determine which variables could be used to explain the particular industry historical data series. These variables become an integral portion of the projection models.
Lastly, the collection and analysis of the industry-staffing pattern is examined. An industry-staffing pattern is the ratio of the employment in each occupation to the total employment in the industry. Data used in the creating of the staffing pattern is collected from the Occupational Employment Statistics (OES) program. Micro data from OES is imported into the Projections Suite Software to produce an industry-occupation matrix that transforms the industry employment projections into occupational employment projections.
In 2018, we implemented two major changes in our Projections Methodology that impacted the way that openings were measured, and our base or first year of the projections. Now, total openings in national, statewide, and sub-state levels more accurately account for the way people work and change jobs, and will impact North Carolina’s projections going forward. Our projections time period is nine years instead of ten- starting with the 2017-2026 long term projections.
Beginning with the 2016-2026 U.S. projections the Bureau of Labor Statistics released in 2017, the national projections now use a different methodology called the ‘Separations methodology’. Our state projections not only employ the new Separations methodology, but also update the numbers to a more current time period.
The previous Replacement methodology relied on assumptions where people did not change occupations over the course of their career, and workers who need to be replaced due to retirement were replaced by workers from often younger cohorts.
The new Separations methodology captures a more accurate picture of workforce separations and distinguishes between workers who leave the labor force entirely (labor force exits) and those who change jobs and leave an occupation (occupational transfers). Let’s use an example below:
Sam grew up in Warren County. After graduating high school, she began working as a bank teller and stayed in that occupation for five years before leaving her job to further her education and enroll as a full-time college student at East Carolina University. She spent the next four years as a college student.
After graduating, she reentered the labor force and found a job as an accountant. Eight years later, she landed a job in a new occupation as a human resources specialist. She spent 15 years in HR before being promoted to financial manager, where she worked 12 more years before retiring.
The new methodology will account for the openings generated by these transitions more accurately. Explanatory resources and a fun video with Sam’s experience can be found here.
The new methodology has three key advantages: more clarity about what is being measured, more robust methods, and more reliable results.
While the numbers generated by the separations methodology strictly should not be compared with the earlier replacements methodology, we will see the average numbers of openings increase due to the separations figures often being significantly larger than the earlier replacement figures. While the new measures will give more information, please note that these openings should not be used as specific targets, but treated as trends meant to inform general growth or decline.
To help you get a visual sense of how the separations methodology operates, we have created a fictional example of an occupation:
Under the old replacement method on the left side, we have 100 growth openings based on new jobs created and 200 replacement openings to replace leavers. Together, these would add up to 300 total openings.
Under the new separations method, we still would have 100 growth openings based on new jobs created, but what had been considered ‘replacement’ openings would be measured as occupational transfers (400) and occupational exits (500). Together, these would add up to 1,000 total openings.
To use another example with real data from North Carolina, we can look at Human Resource Assistants (SOC 43-4161). Below is a chart comparing their Annual Projections Statistics using Replacement and Separations for 2014-2024:
As we can see, the new methodology yields a significantly greater amount of Annual Total Openings than the old methodology, but it more likely captures the potential openings based on those switching occupations and those that are leaving the labor force after working in this occupation.
As for the 9-year projections period- we decided to update the starting (or ‘base’) year of our long-term projections to a more recent year (2017 rather than 2016), representing a 9-year projection. Over the summer of 2018, we evaluated competing models and determined that the best way to produce updated projections was to use 2017 Annual QCEW data to inform our base year estimates. These updated 2017-2026 projections rely on the solid 2016-2026 projections, but provide greater value by being more current.