Reporting/Market Focus from the February 2005 Edition of the SGS Newsletter
Even though the regular Thursday morning reports on initial claims for unemployment insurance are often market movers, the series is heavily misunderstood by the public and frequently abused by those hoping to hype a particular tale on the economy. Nonetheless, when the data are smoothed over three months (17 weeks), the numbers can be used as a meaningful coincident-to-leading economic indicator.
The series is reasonably straightforward. Individuals who have lost their jobs and are eligible for unemployment insurance file a claim with their local unemployment office. Each state aggregates those claims and reports them to the U.S. Department of Labor (DOL). The DOL publishes a weekly total each Thursday for claims in the previous week ended Saturday. The DOL also publishes a total of ongoing claims, the number of individuals receiving unemployment benefits.
As quick clarification, these numbers have no direct relationship to the unemployment rate. The unemployment rate covers all people who are unemployed (in theory), regardless of claims status and is determined by a monthly survey sampling of households (see background paper on employment unemployment).
Weekly variation in the numbers often is touted, falsely, as evidence of a weakening or a strengthening economy, where a lower claims number means a stronger economy and vice versa. Weekly variability usually has nothing to do with economic activity. Claims-level volatility is affected by weather, holidays and other seasonal factors.
In an effort to improve the series, the DOL seasonally adjusts it for holidays, etc. Unfortunately, though, the DOL never has demonstrated an ability to seasonally adjust weekly data. As a result, weekly reports that encompass a legal holiday, such as July 4th, Labor Day, Christmas, New Year’s Day, etc., often show unexpected and market-moving swings that have nothing to do with underlying economic activity.
The new claims series, however, can be used as a solid coincident-to-leading indicator of economic activity, when the seasonally-adjusted, weekly numbers are smoothed over 17 weeks (three months) and compared year-to-year.
As discussed earlier in the Initial Claims section, the current environment is one where the annual decline in new claims is steadily narrowing, headed for an upturn by mid-year.
A final note of caution as to data distortions: Occasionally, the government and certain states will change unemployment claims rules, often extending claims in times of recession. Those filing for extended claims tend to get counted as new claims, but the weekly report usually is pretty good in highlighting any such factors.