Thursday, June 11, 2020

The (Semi-Official) End of the Longest US Economic Expansion

The longest economic expansion in US history (or at least back to 1854, before which time the data gets not just shaky but exceedingly shaky) ended in February at 128 months. It beat the record established during the 120-month expansion of the 1990s. Third place is the 106 month expansion of the 1960s, with the 92-month expansion of the 1980s in fourth place. These figures are from the Business Cycle Dating Committee of the National Bureau of Economic Research, which is to say it's from a group of academic economists who have an affiliation with a certain prestigious research institute. 

(Specifically, the economists involved in this decision were Robert Hall, Stanford University (chair); Robert Gordon, Northwestern University; James Poterba, MIT and NBER President; Valerie Ramey, University of California, San Diego; Christina Romer, University of California, Berkeley; David Romer, University of California, Berkeley; James Stock, Harvard University; and Mark Watson, Princeton University).

It sometimes comes as a surprise to non-economists, but although the US government publishes a wide array of economic data, it does not attempt to pronounce on when a recession has started or ended. Given the political implications of setting such dates, the choice of leaving this task to an outside group is probably a wise one.  In a press release earlier this week, the NBER committee of economists announced that they had chosen February 2020 as the month for the peak of the previous business cycle, and explained why. The committee focuses on monthly data about domestic production and employment.

For domestic employment, there is a Current Employment Statistics (CES) survey of employers. Each month, "CES surveys approximately 145,000 businesses and government agencies, representing approximately 697,000 individual worksites." It does not include, for example proprietors running their own firm, the self-employed, farm-workers, or those employed by households. But the Total Nonfarm Payroll that it does cover includes about 80% of all workers. As you can see, it peaks in February.

However, one problem with the CES data is that because it's collected from employers, it includes workers who are being paid--but are on furlough and thus might be thought of as "unemployed." So for confirmation, the NBER committee also looked at the the Current Population Survey, which surveys about 60,000 households each month. The advantage here is that the survey can ask whether you are on furlough. The downside is that data from payrolls is a pretty solid measure of how many people are getting paid, while surveys that ask people may be subject to more measurement error. But that said, the CPS data peaks in February, too. 
On the production side, estimates of total US economic output like gross domestic product are not available on a monthly basis, but only quarterly: moreover, GDP numbers are first released as an "advance" estimate with preliminary data at the end of a given quarter, updated several months later with second and third estimate as more data become available. For making a decision now about the month the recession started, reliable if less complete monthly data is needed. Thus, the NBER turns to measures of personal consumption expenditures and personal income. Personal consumption expenditures is part (about 70%) of total GDP; personal income is part of the gross national income calculation. 

The data on real personal consumption expenditures is compiled by the Bureau of Economic Analysis from a variety of sources,  including the Monthly Retail Trade Survey, but also other government agencies (like the Departments of Energy, Transportation, Health and Human Services) as well as private associations and trade groups).  
Instead of measuring production by expenditures, an alternative is to measure the income that people received for producing. However, the issue here is that one needs to look at transfers in a way that does not include transfer payments, using sources like the ongoing data collected for the Quarterly Census of Employment and Wages for income, and government data on payments for estimating transfers. The monthly data on income is thought of as being a little more subject to later revision than the data on expenditures, but it also shows a peak in February 2020. 
These four measures are not always so synchronized in their timing. But in this case, it's of course not a surprise that the various measures are neatly aligned, because the pandemic that brought on the recession hit the economy with full force in March.

All definitions of "recession" are unofficial, but a definition one sometimes hears is that a recession is two quarters of negative economic activity. Thus, it may seem odd to have the NBER declaring a recession in June, with only three months of data available since February 2020. In response to these issues, the NBER committee writes: 
[I]n deciding whether to identify a recession, the committee weighs the depth of the contraction, its duration, and whether economic activity declined broadly across the economy (the diffusion of the downturn). The committee recognizes that the pandemic and the public health response have resulted in a downturn with different characteristics and dynamics than prior recessions. Nonetheless, it concluded that the unprecedented magnitude of the decline in employment and production, and its broad reach across the entire economy, warrants the designation of this episode as a recession, even if it turns out to be briefer than earlier contractions.
Finally, I'll add as a mignardise that the four longest economic expansions have all happened since the 1960s and three of the four longest happened since the 1980s. It's not much comfort for the current economic problems, but one of my best friends, Casual Empiricism, suggests that the US economy has become less recession-prone over time.