An official website of the United States government
The Department of Labor's Employment and Training Administration (ETA) contracts with the Bureau of Labor Statistics (BLS) to run the annual programs for weekly unemployment insurance (UI) claims seasonal adjustment.
ETA collects the weekly UI claims reported by each state's unemployment insurance program offices and publishes a weekly news release. ETA uses the set of seasonal factors BLS provides annually and applies them to the unadjusted data from the regular UI program during that year. Concurrent with the implementation and release of the new seasonal factors, ETA incorporates revisions to the UI claims historical series caused by updates to the unadjusted data. ETA is the technical expert regarding UI claims data, and maintains these data at https://oui.doleta.gov/unemploy/claims.asp.
Once a year, BLS updates the models used to calculate seasonal adjustment factors for weekly UI claims. This process includes using all of the regular UI claims data available at that time, updating the parameter files, and identifying potential outliers. BLS is the technical expert regarding the seasonal adjustment process.
Seasonal adjustment is a statistical technique that attempts to measure and remove the influences of predictable seasonal patterns to reveal how weekly UI initial and continuing claims change from week to week.
Over the course of a year, the amount of UI claims undergoes fluctuations due to seasonal events including changes in weather, holidays, and school schedules. Because these seasonal events follow a more or less regular pattern each year, their influence on statistical trends can be eliminated by seasonally adjusting the data from week to week. These seasonal adjustments make it easier to observe the cyclical, underlying trend, and other nonseasonal movements in the series.
The current factors most recently were revised from 2018 forward in conjunction with the news release on April 6, 2023. See below for additional information.
Prior to 2002, a method of seasonal adjustment that was developed by staff of the Federal Reserve Bank (FRB) and BLS had been used for at least two decades. Data users voiced concerns about the volatility in the weekly seasonally adjusted initial claims estimates during highly seasonal periods. BLS evaluated an alternate method developed by the FRB and found it to improve the weekly seasonal adjustment over such periods. BLS and ETA introduced this alternate method to develop seasonal factors on April 11, 2002, effective with the release of claims data for the week ending April 6, 2002.
The pre-2002 method of seasonal adjustment assumed that the claims series had a fixed seasonality. That is, the claims data reflect a holiday or regular seasonal event the same way each year and the seasonal factors change only from the effects of the calendar. The alternate method assumes that the claims series exhibit variation in response to a seasonal event (moving seasonality). The alternate method allows the coefficients that determine the factors to change over time, in addition to reflecting the change based on calendar effects. (As part of testing the alternate method, it was confirmed that the claims series does in fact exhibit moving seasonality.)
In the development of seasonal factors, the alternate method uses claims data from the first few weeks of January. Extending the period into the current year more fully accounts for claimant activity during the holiday period and better captures seasonal movement. The pre-2002 method incorporated data only through December of the prior year in the development of new seasonal factors.
Beginning with the Unemployment Insurance Weekly Claims News Release issued on April 6, 2023, the methodology used to seasonally adjust the national initial claims and continued claims reflects a change in the estimation of the models.
Seasonal adjustment factors can be either multiplicative or additive. A multiplicative seasonal effect is assumed to be proportional to the level of the series. A sudden large increase in the level of the series will be accompanied by a proportionally large seasonal effect. In contrast, an additive seasonal effect is assumed to be unaffected by the level of the series. In times of relative economic stability, the multiplicative option is generally preferred over the additive option. However, in the presence of a large level shift in a time series, multiplicative seasonal adjustment factors can result in systematic over- or under-adjustment of the series; in such cases, additive seasonal adjustment factors are preferred since they tend to track seasonal fluctuations more accurately in the series and lead to smaller revisions.
Before the pandemic, the seasonally adjusted UI claims series used multiplicative seasonal adjustment models. Starting with March 2020, BLS specified these series as additive. After the large effects of the pandemic on the UI claims series lessened, the seasonal adjustment models once again were specified as multiplicative. This is important since statistical tests show that the UI claims series should, in normal times, be estimated multiplicatively.
As long as the pandemic period is within our 5-year scope for revision, the UI claims series will use a hybrid adjustment. For the most volatile portion of the pandemic for these series, covering the week ending March 21, 2020, through the week ending June 19, 2021, the series are additively adjusted. These additive adjustments were not revised on April 6, 2023. Nor will they be subject to further revision. Before and after the highly volatile portion of the pandemic, the series are adjusted multiplicatively and in scope for regular revision. For consistency, all seasonal factors displayed in the factor files are shown as multiplicative, as the additive factors were converted to implicit multiplicative factors effective with the revisions on April 7, 2022.
Now that the pandemic impacts on the UI claims series are more clear with the benefit of hindsight, modifications have been made to the outlier sets in the seasonal adjustment models for both of the claims series, generally replacing level shifts with temporary changes. This led to larger than usual revisions for many weeks in scope over the five years. However, these changes should more accurately indicate claims levels and patterns for both the initial and continued claims series.
Eurostat Handbook on Seasonal Adjustment, Chapter 28 Weekly Seasonal Adjustment: A Locally-weighted Regression Approach
Please submit inquiries to the Local Area Unemployment Statistics program.
Last Modified Date: April 6, 2023