The truth about the BLS jobs report

There are a few things to keep in mind as you read the Bureau of Labor Statistics (BLS) job numbers each month and listen to the news reports about what these numbers mean:

1. You will hear a lot about the month-to-month net job creation which BLS says was 69,000 from April to May.  Don’t give too much emphasis on any single month.  The important thing is the TREND over time.  The following charts from the BLS report provide some perspective.  See the charts here or below: http://www.bls.gov/news.release/pdf/empsit.pdf

In Chart 1 you can see that the unemployment rate has been trending down since about October in 2010.  That’s the good news.  However, in Chart 2, you can see that even though we have been adding jobs since then during last year and this year the early months of the year have been better than the summer months.  We may be seeing the same pattern. 

2.  Recognize that the BLS numbers are ESTIMATES based upon a sample of establishments and a household survey.  Consequently, we have to consider the sampling error which BLS says is +/- 105,000 jobs.  That means that the actual jobs created in May could have been as low as a NET LOSS of 36,000 to a NET GAIN of 138,000.  This sizable sampling error is another reason NOT to pay too much attention to month-to-month changes particularly if they are LESS than 100,000 jobs.

3.  The sampling error for categories of employment (heath care, manufacturing, etc.) or worker (adult men, adult women, Blacks, etc.) are EVEN GREATER than the sampling error for the entire report.  Here again, trends matter more than individual month-to-month changes particularly if these changes are not substantial.

4.  Recognize that BLS makes adjustments to the numbers to correct for inaccuracies in the sampling and for other reasons.  The intention is to make the estimates more accurate but there is a lot of debate about whether that is the case.  For example, BLS uses a birth/death adjustment to account for business start-ups and closures.  See here: http://www.bls.gov/web/empsit/cesbd.htm  That model was changed in 2008, after the recession hit and some people claim that the 2008 changes cause the BLS estimates to be UNDER estimating new job creation. 

5.  Pay attention to other estimates of unemployment out there.  For example ADP provides an estimate for the non-farm private business sector each month which usually comes out just before the BLS report.  ADP said 133,000 were created in May, considerably more than BLS estimated.  Steven Hansen at Global Economic Intersection endorses the ADP method over the BLS result. He says ADP is a better estimate.  See: http://econintersect.com/wordpress/?p=22511.  Over time the trends shown by the  ADP and BLS estimates tend to similar but there can be a lot of month-to-month variation.  Also, Trim Tabs provides an estimate of job creation based upon an analysis of daily income tax deposits to the U.S. Treasury from all salaried U.S. employees which it says is historically more accurate than initial estimates from the BLS.  Trim Tabs estimated that 124,000 jobs were created in May, considerably higher than the BLS estimate.

BOTTOM LINE: When it comes to understanding what is happening in the jobs market focus on trends and take a look at multiple estimates. And, learn to live with uncertainty and imperfection in the jobs numbers.