One of the indicators that economists monitor is consumer confidence, the degree of optimism consumers feel about the health of the economy. To get a pulse on worker confidence, meanwhile, it’s worth keeping an eye on labor market churn—the job-to-job movement that typically happens at a faster rate when jobs are plentiful and workers are optimistic they can find new employment that might pay more or better fit their skill sets.
In a new report, we examined the rate of labor market churn for every non-farm occupation, both nationally and for the most populous 75 metropolitan statistical areas. To perform the analysis, we used a new addition to EMSI’s data offerings: year-by-year numbers on hires and separations that come via the U.S. Census Bureau’s Quarterly Workforce Indicators (QWI). The following are some of the major takeaways of the study:
While national employment has mostly returned to its pre-recession peak, the churn rate has scarcely budged from the depths of the downturn. In 2013, churn stood at 68.1%, a modest improvement from the low of 64.8% in 2009.
Job churn has slowed across every major occupation group … and almost every large metropolitan area. Only Boston, among the 75 largest metros, had a faster churn rate in 2013 (88.3%) than in 2003 (87%). From 2010 to 2013, the churn rate dropped in 15 of the same 75 metros; New Orleans and Denver led the way in post-recession declines.
Low-wage occupations tend to have the highest churn rates. Nuclear power plant operators had the lowest average annual churn rate from 2010 to 2013 (22.5%), and they make $37.67 per hour in median earnings. Fast food cooks had an average yearly churn rate of 113.3%, and they make $8.88 per hour.
IT occupations, on average, didn’t see as dramatic of drops in churn as other occupations, and they recovered significantly faster. One example of this: The churn rate of web developers in the San Jose metro mushroomed from 47% in 2003 to 93% in 2013.
Definitions and Data Sources
Churn is not a measure of employment growth or decline, but rather a measure of job-to-job movement among workers within a labor market. For example, when a worker leaves an engineering job for a higher-paying role across town, the move counts as one separation and one hire. Unless her old firm decides not to replace her, someone new will be hired to take her spot.
EMSI calculates the annual churn rate by finding the average of hires and separations in an occupation, then dividing that number by that year’s average employment figure. Churn can exceed 100 percent in certain industries, occupations, and regions when average hires and separations throughout the year are greater than the number of people employed.
Data on hires and separations comes from EMSI’s labor market database, a compilation of more than 90 federal and state employment sources, and is based primarily on Quarterly Workforce Indicators. QWI connects unemployment insurance forms from businesses with the unique Social Security numbers of each employee. Any time a worker shows up on a company’s payroll in one quarter when she didn’t show up in the previous quarter, she is counted as a hire. Likewise, if for any reason a worker stops working for an employer, she gets counted as a separation.
Employment data used in this report comes from EMSI’s 2014.3 Class of Worker dataset and includes only wage-and-salary employees.
Two notes of importance:
QWI releases hires and separations, among many other types of data, by industry at the county level, making it the most comprehensive source for industry churn. EMSI combines QWI with other datasets to produce up-to-date hires and separations data for all occupations using staffing patterns, which show the percentage occupational makeup of jobs within each industry. QWI offers more industry and geographic detail—and more accuracy because it’s based on administrative records—than the Job Openings and Labor Turnover Survey (JOLTS) from the BLS, which sends a voluntary survey to 16,000 businesses every month. QWI, however, does not break out voluntary separations, also known as quits, like JOLTS does.
Farming, fishing, and forestry occupations (those in the SOC 45-0000 occupation group) were excluded because of the large concentration of seasonal workers in those fields.
A note to subscribers of Analyst, EMSI’s labor market research tool: Hires data by occupation is part of the job posting analytics report. If you’re interested in hires and separations by occupation (or industry) for your region, please reach out to your client services representative or email Josh Wright.
Employment Rebounding; Churn Not So Much
Total non-farm jobs in the U.S. stood just shy of 133 million at the end of 2013, about 1.4 million short of the high in 2007 but 6.1 million more than in 2010. Solid progress. But the churn rate through 2013 only eked up from a low of 64.8% to 68.1%—far short of the mid-2000s, when the churn rate was at 85% and 86%.
Why the stubbornly slow rebound? It could be that workers aren’t as confident in the labor market as they were before the recession. There’s clearly a psychological element in play when workers weigh whether to leave a company or stay with the job (or jobs) they have. In September, the number of voluntary separations that is measured by JOLTS was at the highest level since April 2008. Encouraging, yes, but we’re now only back to level seen during the heart of the recession.
Slow wage growth could also be a factor for the low churn rate. Inflation-adjusted earnings, as we analyzed in March, fell 1.3% nationally from 2010 to 2013. Wages have grown in some regions and some industries, but without the promise of a bump in salary, workers could be less hesitant to bounce from one job to another. It might not be worth their time to switch jobs without a financial incentive.
Visualizing Churn and Wages for Every Detailed Occupation
The highest average annual churn rate from 2010 to 2013 for a non-farm occupation belonged to actors, at 171.4%. The lowest average annual churn rate, as we mentioned earlier, belonged to nuclear power plant operators, at 22.5%.
There are striking differences in churn from occupation to occupation. But it’s not advisable to compare the churn rates of different occupations (or occupation groups) to determine that one is healthier than another. The type of work performed in an occupation and the nature of the industries that employ the occupation tend to dictate overall turnover.
This is illustrated by the groups of related occupations that begin to emerge when looking at the annual average churn rate for every occupation from 2010 to 2013. Entertainment, construction, and restaurant jobs each form their own groups at the high end; power plant, specialty health care, and airline/aerospace do the same at the low end.
Why is there such a big discrepancy in churn from plant operators to entertainment workers? Again, it goes back to the type of work and the industries that staff these workers. As the report explains:
In the case of power plant jobs, the small number of employers who need workers with these niche skill sets is likely the primary explanation for low churn. There are only 100 nuclear power reactors spread across 31 states in the U.S., and they employed less than 8,000 reactor operators in 2013, according to EMSI data. A lack of job demand, an inherent incentive for plants to retain experienced employees, and distance between prospective employers are all likely explanations for why nuclear power reactor operators have the lowest churn rate of all occupations.
The industries representing the high end of occupation churn rates couldn’t be any more different. Consider the restaurant industry, wherein hundreds of thousands of establishments employ nearly 10 million workers across the U.S. in jobs that are often part-time, low-pay, or seasonal. The transient nature of this type of work is why it’s possible to have churn rates above 100 percent. This is completely normal in some occupations within the entertainment, restaurant, hospitality, and construction industries. In fact, 79 of the 768 detailed occupations studied for this report had churn rates above 100 percent as of 2013. For comparison, there were 147 in 2003.
A general rule of thumb is that the pay in an occupation is a solid way to predict whether it has high or low churn. This is clearly displayed in the following interactive scatterplot—the bulk of high-paying occupations (to the right along the x-axis) are low-churn. From 2010 to 2013, only three occupations that pay above $30 an hour had average annual churn rates of greater than 100%; all were entertainment-related: multimedia artists and animators, producers and directors, and agents and business managers of artists, performers, and athletes.
Drops in Churn Across All Occupation Groups
From 2003 to 2013, no major occupation group escaped the slowdown in churn. Food preparation and serving occupations experienced the largest decline in churn during the recession (a 39-point drop) and from 2003 to 2013 (a 31-point drop). Personal care and service occupations recovered at the slowest rate after the recession. Construction and extraction occupations had the second-biggest drop from 2003 to 2013 (from 122% to 98%) and recovered only marginally.
On the positive side, production occupations—a key part of manufacturing and other skilled trades-reliant industries—tied for the largest rebound in churn rate from 2009 to 2013. Production was joined by arts, design, entertainment, sports, and media occupations with a 35-percent bounce-back.
Churn by Metro: New Orleans, Denver, 13 Others See Post-Recession Drop
Before the Great Recession, it was common for some of the largest metro areas in the U.S. to have churn rates approaching (or exceeding) 100%. But like the slowdown of churn nationally from 2007 to 2009, most big metros saw a sharp drop in churn when labor market conditions worsened—and a handful have continued to see declines post recession.
The churn rate for non-farm occupations in 15 of the 75 largest metros dipped from 2010 to 2013. The most dramatic drop-off came in New Orleans, which went from a churn rate of 89.7% in 2010 to 78.8% in 2013. New Orleans’ decline of 11 percentage points was followed by Denver (-7.3), Greenville-AndersonMauldin, South Carolina (-4.9), and Albuquerque, New Mexico (-4.6).
Meanwhile, from 2003 to 2013, 10 large metros saw churn slowdowns of at least 30 percentage points. Seven of the top 10 were in the Southeast, including the top three: North Port-Sarasota-Bradenton, Florida (-54.1); Virginia BeachNorfolk-Newport News, Virginia (-46.9); and Tampa-St. Petersburg-Clearwater (-39.8). These coastal metros also experienced labor market slowdowns or declines over the last decade. Wage-and-salary jobs in Northport-Sarasota-Bradenton, for example, sank 4% from 2003 to 2013.
Boston, on the other hand, is the only metro among the 75 largest that had a larger churn rate in 2013 (88.3%) than in 2003 (87%). The increase in Boston is partly a result of faster churn in core information technology jobs, which we explored in the report. Another Massachusetts metro, Worcester, and Raleigh, North Carolina, had the smallest drops among the rest of the 75 metros (-1.6 points and -3.6 points, respectively). New Haven-Milford, Connecticut, and Grand Rapids, Michigan, had the next-smallest declines from 2003 to 2013.
Tech-focused Raleigh had the highest churn rate in 2013 (90.7%) and joined Bakersfield, California, and Indianapolis as the metros with the biggest increases in churn from 2010 to 2013. Raleigh came in just behind Boston in average annual churn rate from 2010 to 2013. Both were over 87% for the four-year period, followed by Baton Rouge, Louisiana.
The five cities with the best average annual churn since the recession grew their employment base by at least 4% from 2010 to 2013. However, in each the average annual churn was smaller the last four years than it was before the recession (2003-2006). The shrinking average annual churn is especially noticeable in New Orleans, Baton Rouge, and Raleigh.
For more findings and churn analysis, see our full report.
For more on EMSI’s employment data—available at the county, MSA, and ZIP code level—or to see data for your region, email Josh Wright. Follow EMSI on Twitter (@DesktopEcon) or check us out on LinkedIn and Facebook.
