Birth/Death Model - Reader Responses

I received a question about the birth/death model from Eric on Minyanville and a second similar reply on my blog Construction Jobs Puzzle from an anonymous responder. Here is the email from Eric and my response (to both inquiries):

Prof. Shedlock,

Last week in Five Things Kevin Depew wrote: 'Meanwhile, the Labor Department reported the economy added 157,000 payroll jobs in May, up from 80,000 in April. Miraculously, the Birth/Death Model was responsible for the addition of 40,000 construction jobs in May, which comes to an annualized rate of 480,000 construction jobs... even as housing starts are declining at a 15.7% annualized pace!'

First, he is comparing SA-numbers with NSA numbers. NSA, the economy added 880k new jobs. Did he not know that difference or didn't he care to explain he was comparing apples and oranges? Seems odd, to compare NSA and SA-numbers, no? Prof. Depew has harped on this stat before without pointing out the difference. So have you. I wonder why no one has figured this obvious stat out yet?

Secondly, there are plenty of construction jobs not in housing - hotels, retail, roads, oil/gas platforms, bridges, tunnels, other government work, ship building [huge boom there also], heavy marine, etc. Nominal spending on nonresidential construction was up 15% in 2007:1Q
relative to 2006:1Q. So total construction running at 5% of new jobs. Seems reasonable.

Sincerely,
Minyan Eric

Minyan Eric,

We appreciate the opportunity to clarify the Birth/Death model for inquiring Minyans. For starters, let me say that we are very much aware of the discrepancy between seasonally adjusted figures and non seasonally adjusted figures.

Proof of this is easy enough to find. On Friday, March 09, 2007 in my blog entitled February Jobs / Employment Data I wrote:

With subprime lenders blowing up everywhere (going out of business) the BLS is assuming 11,000 new jobs were added in financial activities. The total number of jobs added by such assumptions for February was 118,000 jobs.

No doubt you will see some who will subtract 118,000 jobs from 97,000 jobs and conclude that 21,000 jobs were lost as opposed to gaining 97,000 jobs. Unfortunately such math is inaccurate because the establishment numbers are seasonally adjusted and the birth/death assumptions are not.

Here is the pertinent snip from the BLS on Birth/Death Methodology.

• The net birth/death model component figures are unique to each month and exhibit a seasonal pattern that can result in negative adjustments in some months. These models do not attempt to correct for any other potential error sources in the CES estimates such as sampling error or design limitations.
• Note that the net birth/death figures are not seasonally adjusted, and are applied to not seasonally adjusted monthly employment links to determine the final estimate.
• The most significant potential drawback to this or any model-based approach is that time series modeling assumes a predictable continuation of historical patterns and relationships and therefore is likely to have some difficulty producing reliable estimates at economic turning points or during periods when there are sudden changes in trend.

It just so happens that I was discussing the birth/death model with economist Paul Kasriel at the Northern Trust last Friday before your email was passed on to me. None of us find those revisions reasonable. The reason is simple. The idea behind the birth/death model is the birth and death of businesses (supposedly in relation to the business cycle).

Look at all the subprime lenders that blew up. Look at the problems in housing itself. The BLS is assuming not only that jobs were added, but that new unaccounted construction businesses were created in this environment (where business capex spending has been weak, housing has been horrid, over 50 lenders have gone out of business or stopped writing loans, and the GDP has collapsed from 3.5% to .6%).

Are new construction companies really being created in this environment or are they going out of business? If the latter, perhaps 40,000 jobs should have been subtracted (not added) to construction businesses. That is precisely what Kevin and I (and now Paul Kasriel) have been taking a critical look at.

Please consider Paul Kasriel's latest article Will the Real Private Nonfarm Payrolls Please Stand Up?

Each month the BLS heroically makes an adjustment to private nonfarm payrolls for the estimated hiring by new businesses not yet included in the BLS survey and the firing by closed-down businesses not captured in the BLS survey. This is called the “birth/death” adjustment. The birth/death adjustment does not take into account cyclical influences on business start-ups and failures. So, at a time when economic growth is slowing, there conceivably could be a slowdown in start-ups and an increase in failures that would not be incorporated in the birth/death adjustment. That is, late in a business cycle, the birth/death adjustment could bias upward the BLS estimate of private nonfarm payrolls.

Chart 1 shows that small businesses currently do not think that now is a good time to be expanding their operations. If now is not a good time to expand the operations of an existing business, it would be reasonable to assume that now would not be a particularly good time to start a business either.

Chart 1



The BLS reported that seasonally adjusted, residential construction employment was down only 1,300 in May. Given that home builders have cut back their production and that many are laying off staff, it is curious that such a small number of residential construction jobs were lost in May. The birth/death adjustment might have played a role in limiting this decrease inasmuch as the adjustment added 40,000 construction jobs (both residential and nonresidential) to the total of not-seasonally-adjusted private nonfarm payrolls in May.

I am a big believer in the notion that the private sector can do most things more efficiently than can the government sector. And this extends to measuring economic activity. Perhaps the ADP estimate of payrolls, which is based on actual company payroll data processed by ADP and which is not adjusted for the birth and death of small businesses, presents a more accurate picture of employment growth in America than does the BLS Establishment and Household Surveys.

There you have it: a timely third opinion. Thanks Paul. And thanks to you Eric as well for giving us the chance to clear up some of the confusion that may have been lingering about the birth/death model.

This post originally appeared on Minyanville.

Mike Shedlock / Mish
http://globaleconomicanalysis.blogspot.com/