How exceptional are current unemployment rates?
One of the key variables that signal the current strength of the economy is the very low unemployment rate in the United States, but also in the UK. At an official rate of 3.6%, the unemployment rate in the United States hovers around levels last seen in the late 1960s. In the UK, you have to go back to the mid-1970s to find unemployment rates as low as today.
But are these numbers really that exceptional? In the chart below, I have used an HP-filter to de-trend the US unemployment rate and isolate the cyclical component of unemployment. As you can see, the cyclical component is low at 1.1 percentage points below trend, but not exceptionally low. In fact, the cyclical deviation in unemployment is less than it was at the height of the economic expansion in 2000 and 2007 and about the same as in the year before the 1990/1991 recession.
US unemployment trend and cyclical component
Source: St. Louis Fed.
A similar picture emerges once I de-trend the UK unemployment rate. The cyclical deviation from the trend is about one percentage point, less than in 1989 and 2005/2006.
UK unemployment trend and cyclical component
Source: ONS.
Looking at it this way, the current cycle looks rather ordinary and similar to previous cycles, which, on the positive side, means that unemployment can decline even further from current levels. Of course, the cyclical deviation depends on the actual unemployment rate as well as the trend. And trend unemployment is declining in both countries for quite some time now.
But why should this declining trend be real? After all, just because an econometric filter shows a declining trend that doesn’t mean that trend unemployment is really declining. Andreas Hornstein and Marianna Kudlyak from the Federal Reserve Bank of San Francisco have shown that this declining trend might indeed be real. They used the monthly Current Population Survey of the Census Bureau to look at the movement of individual US citizens between being employed (E), unemployed (U) and out of the labour force (N). This detailed micro data enabled them to estimate the true flows between these three states and construct a real-time estimate of the unemployment rate.
The first chart below shows the movement of people from being unemployed to being out of the labour force or being employed. These flows remain relatively stable over time with no visible long-term trend.
Flows of unemployed people out of the labour force (UN) or into employment (UE)
Source: Hornstein and Kudlyak (2020). Note: NBER recessions are indicated in grey.
The second chart shows the flows of people between being employed and out of the labour force (EN) as well as from being out of the labour force to being employed (NE). Again, no long-term trend visible.
Flows of employed people out of the labour force (EN) or people newly gaining employment (NE)
Source: Hornstein and Kudlyak (2020). Note: NBER recessions are indicated in grey.
But now take a look at the flows of people who are out of the labour force and become unemployed (i.e. ready to get a job but can’t find one right away, NU) and people who are no longer employed and become officially unemployed (UE). Since 2008, there is a steady decline in these flow rates. This means that unlike in the past, people no longer claim unemployment benefits. They are far more likely than in the past to drop out of the labour force altogether after losing their job (or voluntarily leaving their job) and they are far more likely to move from being out of the labour force back into employment.
Flows of people out of the labour force back into unemployment (NU) and employed people into unemployment (EU)
Source: Hornstein and Kudlyak (2020). Note: NBER recessions are indicated in grey.
This is the impact of the gig economy on the labour market. If people no longer have standard work contracts, they are no longer eligible for unemployment benefits when they lose their jobs. Hence, they will not bother claiming unemployment benefits but instead drop out of the workforce and go back to school etc. Similarly, if they eventually do find a job, it is unlikely to be a job that they find through traditional job searches but instead, they might just become an Uber driver or get another gig somewhere without any decent work benefits. In this case, they go straight from out of the labour force to being employed. The net effect of these trends is a technical decline in the unemployment rate that has nothing to do with the strength of the economy. Given that we can expect the gig economy to grow in size, we should also expect that unemployment rates will continue to experience a long-term trend lower with cyclical variations around this declining trend.