The FRED Blog has discussed how the COVID-19 recession reduced the demand for services and boosted the demand for goods, whereas in previous recessions it was the inverse. Today we examine the same dynamic from a different angle: how these changes in consumption patterns have affected industry-specific employment.

The FRED graph above shows the large initial declines in employment for goods-producing industries (e.g., construction and manufacturing) and service-providing industries (e.g., leisure and hospitality). We changed the units of the data into an index, with a base period set at the start of the latest recession, to make it easier to measure and compare changes over time. (This post from October 2020 also uses an index to track unemployment by age during recessions.)

After the initial double-digit declines, employment in goods-producing and service-providing industries gradually bounced back. At the time of this writing they were 5.6% and 7% below pre-recession levels, respectively. For reference, we also plot employment data in the service-providing government sector. This includes federal, state, and local governments, which had similar but smaller declines.

For comparison, the second FRED graph shows changes in employment during the Great Recession, from December 2007 to June 2009. At that time, the largest losses in employment were registered in the goods-producing industry—specifically, in construction. Employment in service-providing industries declined more gradually and by a smaller amount. And employment in the government sector remained effectively unchanged.

In that recession, of course, there was no pandemic and no widespread use of masks and social-distancing measures. To learn more about how recent changes in work and consumption habits might impact future economic activity, read the Economic Synopses essay by Julian Kozlowski.

How these graphs were created: From FRED’s main page, browse data by “Release.” Search for ”Employment Situation” and select “Current Employment Statistics (Establishment Data) > Table B-1. Employees on nonfarm payrolls by industry sector and selected industry detail, Seasonally adjusted.” Select the series “Government,” “Goods-Producing,” and “Private Service-Providing.” From the “Edit Graph” panel, select the “Edit Lines” tab. In the “Units” drop-down menu, select “Index (Scale value to 100 for chosen date)” and choose “2020-02-01” for the first graph and “2007-12-01” for the second graph. Adjust the date range to mirror the dates shown in the blog post.

Suggested by Diego Mendez-Carbajo.

One constant throughout economic history is that, as an economy develops, its service sector keeps growing. The graph shows that this is certainly true for the United States. It divides nonfarm payrolls into three categories: government (at all levels), goods-producing industries (mining, manufacturing, construction…), and service-providing industries. Although government is roughly constant, services have far surpassed goods.

Is this bad? Of course not. The standard of living has clearly improved since 1939, when the graph starts. Indeed, goods can now be produced with fewer people—thanks to technological progress and automation…and perhaps also automatization. This transformation allows the economy to direct more of the labor force to enhancing our lives in other ways, such as tourism and entertainment, advanced health care, and anything related to the Internet, all of which are services that were either nonexistent or luxuries in 1939.

How this graph was created: Using the nonfarm payrolls by industry sector release table from the establishment survey, check the series and click “Add to Graph.” From the “Edit Graph” panel, open the “Format” tab and select graph type “Area” and “Stacked.”

Suggested by Christian Zimmermann.