How Low Can Employment Growth Go without Boosting the Unemployment Rate in Fourth District States?
The nation continues to add jobs as the economic recovery continues, but employment growth is slowing, and even reversing, in some states, including those in the Fourth Federal Reserve District. How will this impact the District’s unemployment rate?
Current employment growth in the Fourth District appears to be above the level required to hold steady the District’s unemployment rate at today’s historically low levels. But the unemployment rate, or the fraction of people in the labor force who do not have a job, will hold steady only if employment grows at the same rate as the labor force, or the number of people who either have a job or want one.
How low can employment growth go before we would expect unemployment rates to rise, and, alternatively, what rate of employment growth would it take to hold the District’s unemployment rate steady?
The answers to these questions depend on state-level measures of growth in the working-age population, net in-migration to a state, and labor force participation. This article examines these factors for states in the Fourth Federal Reserve District, which comprises Ohio, western Pennsylvania, the northern panhandle of West Virginia, and eastern Kentucky.
The first variable is a matter of demographics: The faster the working-age population, or the population aged 16 and older, grows, the higher the number of jobs required in order for the unemployment rate to remain steady.
Population dynamics vary across the country.
Data from the US Census Bureau show that the working-age population for the country as a whole increased by 0.98 percent per year from 2011 through 2015. Compare that to Ohio’s rate of working-age population change: an increase of just 0.33 percent per year during the same period. Of Fourth District states, West Virginia had the most negative working-age population change during this period: −0.06 percent per year.
The Census estimates incorporate realized migration among states, but it’s useful to think about this factor separately in order to look at a variety of migration rates in the estimates below. District states differ widely in how much net in-migration contributes to changes in each state’s labor force.
Ohio, for example, experienced a steady outflow—more workers moving out than moving in—from 2000 through 2015, while Kentucky and West Virginia have had intervals of both positive and negative net in-migration. When more people migrate into a state, that state needs to add jobs in order to hold the unemployment rate steady.
Changes to the labor force participation rate (LFPR), or the fraction of the working-age population in the labor force, also factor into the overall equation. In the simplest terms, when more working-age people enter the labor force, the LFPR rises. If there are not enough new jobs to absorb the new labor force entrants, then the unemployment rate will rise as a result.
There are predictable and unpredictable factors that affect the LFPR. Demographic factors such as an aging workforce and baby boomer retirement are relatively well anticipated. Business cycles, however, are not.
The Cleveland Fed’s Bruce Fallick, vice president of research, in a joint paper with other Federal Reserve economists, modeled the demographic and cyclical aspects of labor force participation at the national level. Their estimates suggest that between 2011 and 2015, change in the LFPR as a result of these factors was −0.54 percent on average per year, with projected change in the near future at similar levels. For a variety of reasons, working-age people are participating in the labor force at a lower rate as time goes on, a situation which decreases the number of jobs needed to hold the unemployment rate steady.
Using historical trends as a guide, then, allows one to examine how different assumptions about future trends in each of these factors affect the employment growth needed to hold the unemployment rate steady.
Five employment growth scenarios
While the calculations are relatively straightforward, several assumptions are required. Varying the assumptions about LFPR and migration trends offers different levels of employment growth that correspond to a steady unemployment rate, but there’s no single “right” scenario. Let’s look at 5 possibilities.
- Constant LFPR, average net in-migration. Keeping the LFPR constant and net in-migration at the average level seen from 2011 through 2015, this relatively optimistic scenario implies that employment growth needs to be 0.98 percent per year nationally to hold the unemployment rate steady throughout the nation. But because the working-age population is growing more slowly in Fourth District states than it is in the nation, and more working-age residents have been leaving than entering, the employment change needed per year in the District is lower, ranging from 0.50 percent in Kentucky to −0.06 percent in West Virginia.
- National LFPR trend, average net in-migration. If we instead presume that the recent national LFPR trend growth rate of −0.54 percent applies evenly to all states, including those in the Fourth District, while keeping the same assumption about average net in-migration, the employment growth needed to hold the unemployment rate steady in the nation falls to just 0.45 percent per year, acknowledging the reality that fewer working-age people are entering the labor force because of changes in age composition and other factors. District states can tolerate negative employment growth under this assumption and still maintain steady or falling unemployment rates.
- State-adjusted LFPR trend, average net in-migration. But let’s adjust the national LFPR trend growth rate to account for different LFPR trends at the state level, using the Census’s American Community Survey (ACS) to obtain state-level LFPR trends for 2011 through 2015 to construct our data. Ohio’s LFPR growth rate is less negative than the national rate in the ACS by 0.08 percentage points. Adding these 0.08 percentage points to the national LFPR trend growth rate of −0.54 percent per year brings Ohio’s state-adjusted LFPR trend growth rate estimate up to −0.46 percent. Similar adjustments are applied to other District states. Doing so raises the level of employment growth needed to hold the unemployment rate steady in Ohio and Pennsylvania and lowers it in Kentucky and in West Virginia.
- State-adjusted LFPR trend, low net in-migration. Let’s keep the state-adjusted LFPR trend and examine what happens if states experience low net in-migration, applying the lowest level of net in-migration seen in each state between 2000 and 2015. If Fourth District states entered a period of low net in-migration, they could tolerate even further declines in employment growth without facing increases to the unemployment rate. West Virginia in particular could tolerate negative employment growth of nearly 1 percent per year.
- Constant LFPR, high net in-migration. The final scenario considers the most optimistic of assumptions: returning to a constant LFPR and taking the highest net in-migration for each state in the 2000 through 2015 period. If this were to occur, states in the District would need positive employment growth ranging from 0.36 percent per year in West Virginia to 0.85 percent per year in Kentucky to hold the unemployment rate steady.
Comparing these estimates to the District’s November employment growth rate of 0.8 percent per year, we see that employment growth has been notably above the level required to hold the unemployment rate steady under reasonable assumptions. And under reasonable assumptions, District employment growth could slow further without boosting the unemployment rate. District states can accommodate lower employment growth than the rest of the country, primarily because of their lower population growth.
One might ask, then, why did Ohio’s unemployment rate just rise from 4.8 percent in August to 4.9 percent in October if employment growth is above the hold-steady rate?
There could be several reasons. First, the unemployment rate is measured using a survey of people, but the level of employment is measured using administrative data on jobs. Relating the two data sets is thus an imprecise exercise.
For one, survey samples don’t always match the characteristics of the underlying population, so the unemployment rate in a survey sample may be different from the rate one might calculate by surveying the entire population. The statistical uncertainty means that a 0.1 percentage point change in the unemployment rate at the state level is likely not significant.
Also, when a person holds more than one job, each job counts in the employment measure. For example, a person working one part-time job might accept a second part-time job from an employer who decides to add part-time workers. The employer’s expansion increases employment. Taking the second job does not decrease unemployment, however, because the person was already employed.
In other words, if 10,000 jobs were added to the economy in a month, and all those jobs were second jobs for existing workers, employment growth would be positive, but the unemployment rate wouldn’t change at all.
Finally, these calculations are based on recent trends in population growth, LFPR, and net in-migration. They aren’t projections of future trends. If any of the trends change, such change moves the level of employment growth needed to hold steady the unemployment rate. A sudden rise in working-age population, say, or a change to the LFPR away from the trend used here can have significant impacts.
Sum and substance: Employment growth appears to be above levels required to hold the unemployment rate steady in the District the Cleveland Fed serves. But trends can change, and change affects the level of employment growth needed to hold unemployment steady.