Skip to main content
The FRED® Blog

Are jobs in education still recession-proof?

Studying employment data in the education sector

The U.S. economy is in recession, and the unemployment rate is above 10%. But the start of the school year is around the corner, and teachers are going back to work.

Historically, student enrollment in colleges and universities increases during recessions,* but what do the data on educational employment show us? Here, we look at two graphs—one for New York City (the most populous U.S. city) and one for California (the most populous U.S. state)—to see if employment in the education sector really is recession-proof.

Both FRED graphs above show there was no noticeable change in college, university, and professional school employment (the blue area) during the 1990-1991, 2001, and 2007-2009 recessions. The data, by the way, are seasonally adjusted to account for regular summer layoffs when schools are out. For reference, we also plot the employment in elementary and secondary schools (the green area). Student enrollment in those schools is mandatory, so one would expect a constant-size labor force, likely increasing due to population growth.

But before we call employment in the education sector “recession-proof,” we have to examine the latest data. In March and April, the mandated social distancing to combat the COVID-19 pandemic in the U.S. resulted in a decrease in employment in higher education both in New York City and California. By June, employment figures bounced back; but as the recession and pandemic continue, employment in the education sector may decrease again.

*For some summer reading, look at work from Harris Dellas and Plutarchos Sakellaris that shows college enrollment increases during recessions: When people are out of work, they choose to increase their education and accumulate “human capital.” (As economists would say, the opportunity cost of schooling decreases during recessions because there are fewer good alternatives.)

How these graphs were created: For both graphs: Search for a series in FRED, then click on “Edit Graph”; then open the “Add Line” tab and search for another series. Once done, open the “Format” tab and choose the graph type “Area” with stacking.

Suggested by Diego Mendez-Carbajo.

View on FRED, series used in this post: SMU06000006561110001SA, SMU06000006561120001SA, SMU06000006561130001SA, SMU06000006561160001SA, SMU36935616561110001SA, SMU36935616561130001SA

Location, location, location in house price data

Manufactured home prices help separate the house from the land

The FRED graph above shows the average price of single-family homes in the four Census regions. Homes in the Northeast are about twice as expensive as in the Midwest or the South, with the West in between. Why so? It could be that the houses have different characteristics (e.g., size and amenities), but it more likely has to do with the location.

The second graph shows prices for manufactured homes in the same four Census regions. These homes come in a fairly standard size and layout. But more importantly for our purposes here, they’re priced at the seller location: in the Northeast, Midwest, South, or West. And they’re priced without the land they’ll be on.

The graph shows there’s no systematic or notable difference in the level of prices in the different regions. Which leads us to conclude that the main suspect for the price differences across regions (in the first graph) is the price of land. Which, obviously, differs by location.

How these graphs were created: Use the release table for home sales, check the average price for each region, click on “add to graph” and start the sample period on 2014-01-01. For the second graph, go to the release table for average manufactured home prices, check the four regions, and click on “add to graph.”

Suggested by Christian Zimmermann.

View on FRED, series used in this post: ASPMW, ASPNE, ASPS, ASPW, SPSNSAMW, SPSNSANE, SPSNSASO, SPSNSAWE

Mapping U.S. unemployment claims

Did you know FRED has launched a new map feature?* It’s true. And today we’ll reveal some of the benefits of using data maps by looking at unemployment claims. 

Since the onset of the COVID-19 pandemic, unemployment has risen to extraordinary levels, which we can see in the continued claims (insured unemployment) data series. The map above shows these claims, state by state, for the second quarter of 2020.

At a first glance, it may seem there’s a lot of heterogeneity across states. The largest increases in claims occurred in California, Texas, Illinois, Michigan, Georgia, Florida, Pennsylvania, and New York—each with more than 2.7 million persons continuing to file for unemployment benefits.

The second map shows population by state. The largest states are California, Texas, Illinois, Ohio, Georgia, Florida, Pennsylvania, New York, and North Carolina, each with more than 10 million persons.

So let’s compare the claims and population maps:

  • Michigan is in the group of high unemployment claims but not one of the largest states. So, we can conclude that Michigan had a larger increase in unemployment claims relative to its population as compared with national averages.
  • In a similar way, Ohio and North Carolina are on the list of largest states but not on the list of most claims. So, we can conclude that the pandemic didn’t hit employment in Ohio and North Carolina as hard as it hit the overall U.S. economy.

This exercise shows that data on maps can be very useful for visual inspection. Of course, one has to take care when interpreting the units and control for other factors such as population by state. GeoFRED does have the ability to show some maps in per capita terms when the frequency of both series coincides, which isn’t the case here.

*All FRED data series with geographic characteristics now have a “View Map” button on the northeast side of the graph. Use it to create a new dimension of data visualization, from a time-series perspective to a cross-sectional perspective, with interactive functionality such as mouse-over and zooming.

How these maps were created: From GeoFRED, click on “Build new map” (green button, northeast corner of the screen) and then click on “Tools” (orange cogwheel button, northwest corner). Under “Region” select “State,” under “Data” select “Continued Claims,” under “Frequency” select “Quarterly, end of period,” under “Units” select “Number,” and under “Date” select 2020:Q2. Note that you can also edit the colors, legend, and labels. Repeat the process with population data for the second map, with the exception that the frequency remains “Annual.”

Suggested by Julian Kozlowski.



Subscribe to the FRED newsletter


Follow us

Back to Top