What are seasonal adjustments to jobs numbers and why are they necessary?
There are times of the year when employment always goes down and unemployment always goes up. Vice verse, there are times of the year when employment always goes up and unemployment always goes down.
Christmas is the easiest of those times to understand. Employment is always higher in December due to Holiday retail hiring, in good economic times and in bad. Unemployment is always down in December. In January, when all of those retail establishments cut back, employment is always down and unemployment is always up.
There are also ups and downs due to seasonal variations such as outdoor construction work and landscaping work, staffing for summer and vacation venues, and layoffs associated with the end of the school year.
Seasonal adjustments (made by statisticians using these usual variations) are attempts to even out these normal ups and downs...
so that we can all have a better understanding of what is happening with the employment situation aside from these usual seasonal ups and downs. Without seasonal adjustments, we'd have to wait a whole year to determine if the job market is improving or deteriorating. However, when seasonal adjustments are made, we can get a pretty good idea as to what is happening to the job market from one month to the next.
What good are "raw" or "not seasonally adjusted (unadjusted)" numbers? Why are they reported?
Unadjusted numbers created confusion when McDonald's had its Hiring Day.
The lack of understanding of seasonal variations and adjustments created some confusion in May. Job growth was announced to be 54,000 at the same time that McDonald's was in the process of hiring 62,000. The rumors were that there was no job creation in May outside of the McDonald's hiring. I discussed this here: Did McDonald's Create all of the New Jobs in May 2011?
Why is there sometimes such a big difference between a seasonally adjusted job number and a non-seasonally adjusted jobs number?
In some months, so many jobs are created and so many people hired that the seasonal adjustment must also be very, very large. May is one of those months. As mentioned above, 680,000 new jobs resulted in only 54,000 jobs in seasonally adjusted numbers in May 2011 because May is traditionally a big hiring month.
Likewise, December to January is a time period in which one would assume that a large number of people typically hired in December but laid off in January would result in a much smaller job loss, or even a job gain, in January jobs report. In December 2009 to January 2010, 2,800,000 "real" jobs were lost, but in seasonal adjustments, only 20,000 jobs were lost. In December 2010 to January 2011, 2,900,000 "real" jobs were lost, but in seasonal adjustments, a gain of 36,000 was reported. In December 2008 to January 2009, 3,600,000 "real" jobs were lost, but in seasonal adjustments, the loss was "only" 600,000 jobs. It appears that, in January, a loss of about 3,000,000 jobs is expected.
The Bureau of Labor Statistics collects employment, unemployment, and jobs numbers in "raw" "not seasonally adjusted" numbers based on what people (in the Current Population Survey) and employers (in the Establishment Survey) report to the them. The BLS then converts these numbers, using algorithms developed from statistical historical models, to "seasonally adjusted" numbers.
Seasonally adjusted numbers in employment and jobs are necessary to give the people and the government a better picture of how the employment/job situation looks from month to month, without the usual variances caused by seasonal changes in employment. As mentioned above, seasonal changes in employment result from holiday shopping and shipping, seasonal construction and landscaping jobs, amusement and vacation hiring, and the like.
If you still don't understand, please leave a comment or send me an email, and I will try to explain things a bit more clearly.