The Federal Reserve Bank of San Francisco has published an Economic Letter titled “Searching for Maximum Employment” by Sarah Albert and Robert G. Valletta.
Please, find below our review, (critical) comments, and possible implications for the Fed Board staff.
To keep in mind
The authors claim that the labor force participation gap -that is, the distance of the labor force participation rate from its trend, is still largely negative (same for employment-to-population ratio). Instead, we show that under more reasonable assumptions, the gap is closer to zero, suggesting that participation (and employment-to-population ratio) is sending a similar signal compared to the unemployment gap.
What the paper does
The authors assess the moving target of maximum employment by examining potential paths for two labor market indicators: the labor force participation rate (LFPR) and the employment to population ratio (EPOP) relative to their underlying trends, which are largely determined by population demographics. The reason is that the traditional indicator of labor market conditions -the unemployment gap- provides a partial measure of the labor supply, as reflected in jobless individuals actively searching for work. LFPR and EPOP provide additional information on labor supply and labor market conditions by including individuals who are employed, thus providing useful insights on where the labor market is relative to maximum employment. The Federal Reserve’s commitment to “maximum employment” is not captured by any single indicator but instead reflects assessments of a wide range of indicators; therefore, at any point in time, it is important to monitor not only the unemployment gap but all labor market indicators.
The authors measure the demographic drag on LFPR (and EPOP) using a straightforward “shift-share” analysis. Specifically, they assume that LFPR for seven age groups by gender (14 groups total) remain fixed at their early 2020 pre-pandemic peak values. Then, the authors calculate the demographic trend in the overall LFPR by weighting each group’s LFP by its projected population share, using annual Census Bureau population projections. For the trend EPOP rate, the authors combine the LFP trend projection with an unemployment rate of 3.5%, which is the low reached during the pre-pandemic expansion.
Results
Figure 3 of the paper shows the authors’ main results for LFPR (similar conclusions apply to EPOP). For convenience, we report Figure 3 below.
The authors estimate that LFPR trend has a downward slope of 0.25 percentage point per year. As such, according to the authors’ estimates, the level of the trend is expected to decrease by 1 percent from 2020 to 2023 (from 63.4 to 62.4 at the end of 2023). The authors’ estimates of the slope are similar to other studies (see Hornstein, Kudlyak, and Schweinert 2018) which find that overall LFPR has been pulled down roughly 0.25 percentage point annually due to the aging of the population in recent years.
The authors conclude that because FLPR (and EPOP) gap is still negative and projected to remain so until 2023, the labor market is short of the Fed’s maximum employment goal, despite very strong labor market conditions that partly reflect pandemic-related employment constraints.
Comment
In our view, the authors’ results are optimistic. In our estimates, the LFPR gap is currently closer to zero than suggested by the authors. Below we list the reasons why we disagree:
First. The authors assume that the level of LFPR trend was equal to FLRP itself before the pandemic (as shown in Figure 3 above). In other words, the authors assume that LFPR gap was zero in February 2020. However, this is a generous assumption given that several other indicators (unemployment gap, output gap, etc..) suggested a positive gap before Covid. On this point, we report below the estimates of LFPR gap by Hornstein, Kudlyak, and Schweinert (2018). A more realistic assumption about the level of LFPR trend in 2020 is lower than the authors’ estimates (at 63.4 percent) – please note that according to Hornstein, Kudlyak, and Schweinert (2018), the level of LFPR trend in 2020 was projected around 62, and below 62 in 2022.
Second. The authors’ “shift-share” analysis assumes that LFPRs for seven age groups remain fixed at their early 2020 pre-pandemic peak values. However (as the authors’ themselves acknowledge) this represents an optimistic assumption for maximum employment for two reasons. First, LFPRs by age group may not reach their pre-pandemic peaks (see Bauer and Edelberg (2021)). And second, most estimates of the natural rate of unemployment are closer to 4%.
Third. The authors’ sample does not seem to include the most recent data. The analysis does not mention explicitly the last month in sample but the published LFPR reading for January (at 62.2) is already very close to the authors’ projections of the level of LFPR trend in 2022 (at 62.65) and above the estimates of Hornstein, Kudlyak, and Schweinert (2018). Therefore, even assuming that LFPR gap was at zero in 2020 (as the authors do), the LFPR gap in January 2022 would be about 0.5 percent (half of the authors’ estimates), representing less than 1 million potential workers (and not 1.5 million as claimed).
We conclude by saying that our comments do not imply that LFPR (and EPOP) cannot increase going forward. They can and, hopefully, they will. The point of this note is to remark that under slightly different (and more realistic) assumptions, LFPR gap is not so negative as the authors claim but it is already possibly close to zero. As mentioned, going forward LFPR gap will hopefully turn positive soon, sending similar signals compared to other labor market indicators.
Implications for the Fed staff
In our view, the Fed Board staff relies on internal estimates of LFPR gap (and EPOP gap). In our view, the Fed Board staff estimates are closer to our own than the San Francisco Fed analysis. For this reason, in our view the San Francisco Fed analysis will not have any impact on monetary policy in upcoming FOMC rounds.