Initial unemployment claims appear stable over past several months
June 28, 2022
The number of initial claims for unemployment insurance is a timely and closely followed indicator of layoffs, labor market strength and overall economic activity. Movement in this number has been of particular interest over the course of the COVID-19 pandemic, as claims spiked to historically high levels in spring 2020 and then declined steadily through March 2022.
In recent weeks, initial claims have reversed course and trended up at a meaningful pace (Chart 1). The four-week moving average of initial claims increased by 53,000 from April 2 to June 18. Such a rise over an 11-week period is relatively rare outside of recessions and has raised concerns about a cooling labor market and an increased rate of worker layoffs.
Seasonal adjustment complications in claims data
It is likely that the latest rise in initial claims reflects difficulty adjusting the data for seasonal patterns in the wake of the COVID-19 pandemic, rather than a deterioration in underlying economic conditions. This reading of the data would also be consistent with data on job openings, quits and layoffs from other sources, which suggest a stable and tight labor market.
The Department of Labor releases both a seasonally adjusted and a raw (not seasonally unadjusted) measure of initial claims. The adjustment removes predictable seasonal fluctuations in the data to provide a clearer reading on underlying economic conditions.
Like any statistical technique, seasonal adjustment is not exact, and complications can arise. Large fluctuations in the data like the COVID-19 layoff spike in spring 2020 can lead to a change in the estimated seasonal patterns, even if it was a one-off event—distorting the adjusted data in several years surrounding the atypical event.
Chart 2 plots the raw data (without seasonal adjustment) over the past several years, where the horizontal axis is the specific week within a year (first week of January is 1, last week of December is 52).
Initial claims in 2022 (red line), covering the first 24 weeks of the year, have been roughly flat, similar to the pattern of the several years before the pandemic. In fact, the most recent week of data shown is lower than the same week over 2017–19. This suggests initial claims are not rising more than usual in a normal April–June period, counter to the seasonally adjusted data.
Chart 3 plots this same data in a different way, with the red line representing the three-year change in the nonseasonally adjusted data. The 2022 values are then relative to the same week in 2019. Other than some volatility due to holiday effects, this line has been roughly flat in 2022, not displaying the upward trend in the seasonally adjusted data (beige line).
Assessing COVID’s impact on estimates of seasonality
The seasonal factors have undergone several rounds of revision in recent years to deal with the unprecedented impact of COVID-19 on unemployment insurance data. These seasonal factors—equal to the ratio of the unadjusted data to the adjusted data—are plotted in Chart 4.
The red line shows the presence of the factors as they appeared in February 2020, before the pandemic, and the beige line is how they appear today. Through 2014, they are identical but move further apart as they approach 2020.
Seasonal adjustment estimation allows seasonal factors to change over time, so years close to 2020 could be distorted by COVID-19, while those further away would not be. The gap between the current factors and those in February 2020 is largest in late March 2019 and then narrows through June.
This revision to the seasonal factors reflects the impact of the spike of initial claims in March and April of 2020. The seasonal adjustment algorithm likely interpreted it as partially reflecting a change in the seasonal pattern—relatively high layoffs in a typical March and April—expected to be recurring in the years before and after the pandemic.
Of course, COVID-19 was a unique event specific to 2020 and should not influence the seasonal factors. As a result, the seasonally adjusted data for the years 2018, 2019, 2021 and 2022 show a meaningful decline in March and April and then a rebound in May and June. This, however, reflects COVID-19’s distortion of the seasonal factors rather than changes in the underlying state of the labor market.
As a result, the takeaway is that through mid-June 2022, initial claims for unemployment insurance were not trending upward but rather remained stable at low levels. This is consistent with other indicators of the labor market that point to very tight conditions, such as firms having difficulty hiring and being very reluctant to lay off workers.
About the Authors
The views expressed are those of the authors and should not be attributed to the Federal Reserve Bank of Dallas or the Federal Reserve System.