New Jobless Claims: 2013 Has Been a Good Year for Seasonal Adjustment

Headlines have highlighted the fact that  new jobless claims have fallen to 335,000 and 330,000, respectively, in the past two weeks.  News reports state that new unemployment insurance claims have dropped to their lowest levels in five years.  These statements are based on seasonally adjusted data.  The following table shows seasonally adjusted and unadjusted new jobless claims data in the first three weeks of 2012 and 2013.

	New Unemployment Insurance Claims
Date	Seasonally Unadjusted	Seasonally Adjusted
January 7, 2012	646,219	390,000
January 14, 2012	525,422	364,000
January 21, 2012	416,880	372,000
Avg First 3 weeks 2012	529,507	375,333
January 5, 2013	557,798	375,000
January 12, 2013	556,710	335,000
January 19, 2013	436,766	330,000
Avg First 3 weeks 2013	517,091	346,667

 The table indicates that while seasonally unadjusted new UI claims have been 2.3% lower in early 2013 than early 2012, seasonally adjusted new UI claims  dropped by 7.6% from one year ago.  The average seasonal adjustment factor in the first three weeks of 2013 reduced claims by 33% while the adjustment factor was 29% over the same three weeks in 2012.  Had the same adjustment factors been used in both 2012 and 2013, new UI claims would have averaged 366,533 per week in January 2013.

An average of 366,533 new UI claims per week is better than we have seen for much of the recovery, but it is certainly not the lowest level in the past five years.  As recently as late September and early October of 2012 the average number of seasonally adjusted new UI claims fell below 360,000 per week.  The last few weeks have seen unusually large seasonal adjustment factors making the difference between seasonally adjusted and unadjusted new UI claims larger than we have seen in recent memory.  The news on jobless claims is good, but the large seasonal adjustment may be overstating the improvement over the past few weeks.

Hurricane Sandy Likely to Increase Unemployment Rate to 8.0% or 8.1%

The November jobs report (released on December 7^th) will be the first one to include household and payroll survey data gathered after Hurricane Sandy.  It is likely that November’s unemployment rate will jump from its current level of 7.9% to 8.0% or 8.1% due to Hurricane Sandy.  Sandy had a devastating impact on the tri-state area of New York, New Jersey and Connecticut where about one eighth of U.S. output is produced.  A leading indicator of the unemployment rate is the weekly report of new unemployment insurance claims.  A spike in new jobless claims means that a large number of workers were displaced from their jobs.  As I explain below, Hurricane Sandy displaced 145,000 workers as measured by new jobless claims in the first full week after the storm hit.

During the week of November 10^th (the most recent week for which detailed state data are available) over 63,000 jobless workers in New York filed for first-time unemployment insurance benefits, compared to about 21,400 new claims one year ago.  Similarly, over 46,000 jobless workers in New Jersey filed for first-time unemployment insurance benefits during the week of November 10^th, compared to just over 12,000 new claims one year ago.  The rate at which workers lost their jobs nearly quadrupled in New Jersey and nearly tripled in New York compared to November 2011.

The following charts compare the year-to-year change in new unemployment insurance claims the week of November 10^th, the first report to reflect Hurricane Sandy effects, and four-week moving averages of year-to-year changes in new claims over the previous 20 weeks.  For example, the annual percentage change in new claims for November 3^rd is based on a comparison of data for the week of November 3^rd and the three previous weeks to the corresponding weeks in 2011.  The charts indicate that, for New York and New Jersey, new jobless claims were consistently below 2011 levels until Hurricane Sandy hit.

Hurricane Sandy caused about 80,000 people to lose their jobs and file for first-time unemployment insurance benefits in one week in New York and New Jersey alone.  Although the effect of Hurricane Sandy on the rest of the country is smaller, it isn’t negligible.  The following chart compares the year-to-year change in new jobless claims the week of November 10^th to four-week moving averages of year-to-year changes in new claims for the rest of the United States (excluding New York and New Jersey).  The chart indicates that new jobless claims were up about 12% in the first full week after Hurricane Sandy, or an increase of 65,000 claims.

Hurricane Sandy’s displacement of 145,000 workers in one week is enough to increase the U.S. unemployment rate by 0.1 percentage point, from 7.9% to 8.0%.  The November unemployment rate is based on worker’s labor force status for the week ending November 17^th.  That means that one more week of new jobless claims data will factor into November’s unemployment rate.  The preliminary new claims data for the week of November 17^th shows a smaller increase in displaced workers, probably half as many as the 145,000 displaced in the prior week.  We will know more on November 29^th when more detailed and complete data for the week of November 17^th are released.  At this point it is most likely that the November unemployment rate will jump to 8.0% or 8.1%.

New Jobless Claims Continue to Signal Labor Market Problems

There have been 15 weeks of new jobless claims reports released in 2012.  The average number of new unemployment insurance claims has been 392,200 per week (not seasonally adjusted).  Between 2004 and 2006, a period of modest annual employment growth of 1.7% or 2.24 million per year, the average number of new jobless claims was 328,800 per week.  Weekly jobless claims are substantially (19%) higher than they were in 2004-2006, when the economy was creating about 187,000 net new jobs per month.

New jobless claims are filed by people who have lost their jobs, so it is difficult to imagine a scenario in which new jobless claims could be much higher than they are now, given the current state of the labor market.  Data from the BLS JOLTS survey indicate that in the six months between September 2011 and February 2012, about 420,600 workers were laid off (or discharged) from their job each week, on average.  This means that in 2012 new jobless claims have been about 93.2% of layoffs per week, on average.  Between 2004 and 2006 new jobless claims were about 78% of layoffs, on average, per week.  We know that the labor market is struggling when almost everyone who loses their job files for unemployment insurance.