In Depth Analysis: CalculatedRisk Newsletter on Real Estate (Ad Free) Read it here.

Monday, September 07, 2009

Comparing BLS Job Losses and DOL Unemployment Claims

by Calculated Risk on 9/07/2009 01:05:00 PM

A frequent question is how do the 570,000 initial weekly unemployment claims, as reported by the Dept of Labor (DOL), correspond to the 216,000 in monthly job losses as reported by the Bureau of Labor Statistics (BLS).

If about 2.4 million people filed initial weekly claims in a month (570,000 X 4 weeks), how come the economy only lost 216 thousand net jobs in August?

First, I think it is helpful to look at total hires and separations each month. The BLS has a survey called "Job Openings and Labor Turnover Survey" (JOLTS) that provides this information. The difference between JOLTS hires and separations is similar to the CES (payroll survey) net jobs headline numbers.

Note: Remember the CES (Current Employment Statistics, payroll survey) is for positions, the CPS (Current Population Survey, commonly called the household survey) is for people. See Jobs and the Unemployment Rate for a comparison of the two surveys.

The following graph shows hires (Green Line), Quits (blue bars) and Layoff, Discharges and other (red bars) from the JOLTS. Red and blue added together equals total separations.

Unfortunately this is a new series and only started in December 2000.

Job Openings and Labor Turnover Survey Click on graph for larger image in new window.

Notice that hires (green line) and separations (red and blue together) are pretty close each month. When the green line is above total separations, the economy is adding net jobs, when the green line is below total separations, the economy is losing net jobs.

Although initial claims are for people and JOLTS is for positions, this does show why initial claims are so high. In the first six months of 2009, an average of about 2.8 million jobs were lost involuntarily each month. If all of these people applied for unemployment claims, the average initial weekly unemployment claims would have been about 650,000 per week (2.8 million divided by 4.3 weeks per month). In fact weekly claims averaged just over 600,000 per week for the first six months of 2009. Note: "quits" don't receive unemployment insurance.

So even though there were about 4.2 million new hires each month during the first six months of 2009, people who lose their jobs involuntarily during a recession have a difficult time finding a new job right away, and most apply for unemployment benefits.

In better times, like 2005, about 2.26 million jobs were lost involuntarily each month, but weekly claims only averaged 330,000 per week (2.26 million divided by 4.3 week is 525,000). This shows when the economy is adding net jobs, a larger percentage of people can find new jobs right away and don't apply for unemployment insurance. But many people still do file for benefits.

Although we don't have JOLTS data for the '90s, even in the best of times for employment (like 1997), the U.S. averaged about 230 thousand initial unemployment claims per week - even though the economy added almost 3.4 million net jobs for the year. This just points out there is significant employment turnover in the U.S. economy, and many people lose their jobs involuntarily even in good times.

Final Note: Since weekly initial unemployment claims are related to involuntary separations - and the overall strength of the job market (Can people find a job right away?), there is no magic formula between initial claims and net jobs. It does appear that initial weekly claims will have to fall to about 400,000 per week before the economy starts adding jobs, see from Brad DeLong: Payroll Employment Starts Growing When Seasonally-Adjusted Unemployment Claims Fall Below 400K per Week or so...

Brad DeLong This is the graph from DeLong's post (click on graph for larger image.)