Payroll Employment Report and the Birth/Death Model

How should we interpret today’s employment report, showing a growth in non-farm payrolls of 88K, a bit less than expected?  There is also an increase in the unemployment rate and a different picture of job growth from the survey of individuals.

This is the real story, with as little technical stuff as we can manage.  If traders or investors want to understand the data — viewed by many as most important — this is as good as we can make it for those without technical background.

Background:  The BLS in Action

The most important thing for investors and traders to understand is that the reports are prepared by civil service employees using professional statistical methods.  They do not change staff or opinions with shifts in partisan control.  They are not biased in favor of a particular outcome.

Nearly all of those commenting on today’s report, and particularly the Birth/Death Model, have never worked in a government agency and never done any forecasting using statistical methods.  Do not listen to them!  Since this is a subject where we can add value for the investment community, we have written extensively on this topic.

[Digression for the record — I was "loaned" from the University of Wisconsin to the Wisconsin Department of Revenue for a couple of years a long time ago.  The DOR did these forecasts for the state.  I also taught graduate classes to mid-career government  professionals.  The courses covered public finance and research methods, among other things.  So I know how these public officials think and what they try to do.]

Imagine that you are attempting to count all of the jobs in the U.S. economy, and you must do it EVERY MONTH.  There is no way to count them.  Compare this to voting, where in the 2000 election we had trouble actually counting the votes of people who showed up at the polls.  (Remember the "hanging chad"?)  There are 130 million jobs.  The result you get is an estimate based upon surveys and statistical techniques.

The monthly job count is the result of a survey, using as your sampling frame the businesses that existed at the start of the year.  This is subject to various sources of error, as follows:

  • Sampling error.  Even when you have a large sample size, there is always sampling error.  It is small in percentage terms, but large in the aggregate.  The sampling error for the jobs report has a standard deviation of about 60k jobs.  (Check here for the best standard English description of this you will see online via my colleague Allen Russell).  That means that 2/3 of the time the reported result will be +/- 60K.  If you want a 95% confidence interval, you need to double that.  This is the result after ALL businesses in the sample send in their forms.
  • Revisions.  As we discussed in our Employment Report Preview, the source of the revision is not any tinkering by the government.  It is the result of businesses reporting late.  Did you ever take an extension in filing your tax return?  It has no meaning, unless you think there is a systemic bias to businesses that are late in filing.
  • Non-sampling error.  The original sample does not include new businesses, and it includes other businesses that have failed.  Briefly put, this is a major problem unless there is an exact offset between the two.

So what does our earnest and professional government statistician do?  The only major place to improve is the non-sampling error.  The BLS discovered that they had serious errors without making some adjustment to the original survey.  This was the reason for creating the Birth/Death model.

They state in the description of the methodology that there is a fairly constant relationship between births and deaths of businesses.

Earlier research indicated that while both the business birth and
        death portions of total employment are generally significant, the net
        contribution is relatively small and stable. To account for this net
        birth/death portion of total employment, BLS is implementing an
        estimation procedure with two components: the first component uses
        business deaths to impute employment for business births. This is
        incorporated into the sample-based link relative estimate procedure by
        simply not reflecting sample units going out of business, but imputing
        to them the same trend as the other firms in the sample. 

It seems counter-intuitive to use business deaths to predict births, but they are looking at the data and we are not.  None of the critics are either.  Perhaps the BLS needs to share more about the underlying data.

The major point is that the BLS economists are making an honest and earnest effort to capture job changes missed by the survey.

They check their results with a benchmark revision.  Eventually, they have an actual count of jobs from state data, and they update their model regularly to reflect this.  They do an annual revision for the benchmark and a quarterly revision of their ARIMA time series modeling.  (This is a standard method.  Like other forecasters, we use it frequently.  Readers should look with suspicion on the comments of anyone who does not seem to understand this method).

In short, the BLS staff is using standard professional methods to make the best possible estimate of job growth.  It could be off by 60K on the sampling error, or more if the Birth/Death model is wrong.  None of the critics offer a better method.

What the Birth/Death Model Contributes

The Birth/Death Model has improved the estimates.  In fact, despite the constant criticism from those who claim that it adds "fictional jobs" the model actually underestimated job growth in recent years.  That is why the benchmark revisions added jobs.  It was an attempt to correct the non-sampling error, and it improved the estimate.  The critics who said that the model over-estimated job growth were proven wrong by actual state data, used in the benchmark revision.  The reaction of the critics was to ignore the actual count and criticize the BLS for "finding jobs."  These critics should walk a mile in the moccasins of someone trying to provide useful data.

Please note that the Birth/Death model does not make a single jarring adjustment to the number of jobs in a particular month.  The BLS goes back and spreads the jobs growth over the year, using the ARIMA model, in proportion to the year’s data.  Critics are challenged to offer a better approach….

The BLS warns that the methods are based upon the continuation of past trends.  They do not look at any other sources of data about economic growth.


  • Those drawing inferences from revisions do not understand the process, as we explained in the  preview for this report.  Doug Kass (we like him, we respect him, we read him, and we often profit from his short-term advice) made the point both on his daily blog and on Kudlow (video here) that there was the first downward revision in many months.  This is meaningless and Doug should know better.
  • There were various mis-statements about the Birth/Death Model.  Some analysts tried to add the adjustment to the final NFP report.  The BLS explicitly states that one should not add Birth/Death adjustments to seasonally adjusted data.  Anyone doing this does not understand the method.

Q: Can I subtract the birth/death adjustment from the seasonally adjusted
over-the-month change to determine what it is adding to employment?

A: No.  Birth/death factors are a component of the not seasonally
adjusted estimate and therefore are not directly comparable to the seasonally
adjusted monthly changes.  Instead, the birth/death factor should be
assessed in the context of its effect on the not seasonally adjusted estimate.

  • Some observers tried to focus on the construction job component, especially suggesting that the Birth/Death Model added an unrealistic number of jobs in this sector.  The model does not specify the sector for job growth, so this statement is simply wrong.  Many of those looking at the "internals" of the report lose track of a simple fact:  This is a survey.  There is an error band around any reported category.  They tried to count all of 9 million jobs in construction and mining and such in one month and compare it to the count in the next month.  There is a wide error band.

Our Conclusions

While we do think that the BLS is using a professional and technically sound method, it was obviously slow in catching growth during the expansion.  We would not be surprised to see some modest downward benchmark revision when we have an actual job count for this year.  This point was made by Tony Crescenzi today in his excellent blog on’s RealMoney (a paid site, but worth it for any serious investor).  Tony uses the household employment survey to assist his analysis, since he thinks it might catch factors missed in the payroll report.  We agree.

Econobrowser, a site that we read daily on RSS, takes a different approach.  James Hamilton blends data from several report estimates to get a different projection.  While this differs dramatically from our own forecast using Michigan Consumer Confidence, initial jobless claims, and the ISM manufacturing index, it is interesting and professional. 

Meanwhile, the Fed is looking at these data in a long-term trend, in conjunction with other indicators.  Chairman Bernanke, the Open Market Committee, and the 350 Fed economists all know how to interpret employment data.  They do it much better than those on TV.

We would all like to have a monthly read on employment.  I would also like to know how the Sox will do tomorrow or the Bulls will do versus the Pistons.  It is important to realize the limitations of the data.  There is a difference between the data we want and the data we need.  It requires interpretation.


Many of the interpretive mistakes cited here will appear in Alan Abelson’s column tomorrow…..

You may also like


  • theroxylandr May 4, 2007  

    Do I understand correctly that birth/death model makes serious mistakes when underlying trend makes sharp changes?
    For example, if actual job count contracted in April by 500k jobs after March +180k jobs. Will B/D model actually smooth the picture and show much less of the contraction (which will take months to count properly)?

  • Jeff Miller May 4, 2007  

    Good question!
    Our general point is that too much emphasis is place upon payroll employment as opposed to the collection of economic indicators.
    The other indicators do not suggest anything like a contraction of 500K jobs. Our research, and the work of the highly-respected ECRI (google it in our search box for specifics) suggests quite the opposite.
    Having said this, the BLS method does not catch dramatic turning points. Last year the benchmark added close to a million jobs over the course of a year. The methodology could also miss a negative turning point, although probably not as dramatic as you posit in your question.
    Thanks for your excellent question.

  • Barry Ritholtz May 5, 2007  

    As noted in yesterday’s post, its hard to ignore the ANNUAL contribution of the B/D model to overall job creation measures of the BLS.
    The B/D adjustment has become very significant. It has had a major impact on BLS Non Farm Payroll total numbers. In 2006, for example, of the 2.26 million new jobs that BLS reports as being created, 964,000 — nearly half (42.6%) — were due Birth/Death adjustments.
    That moves from the realm of counting and estimating to a different place, where we hypothesize and guesstimate — a process that is far less objective, reliable and accurate.
    As to the seasonal adjustments on the monthly data, consider this: Of those 317k new jobs hypothesized by BLS for April, 49,000 of those supposed jobs were in construction. Now what are the odds of that corresponding directly to reality?

  • RK May 5, 2007  

    Love the dig at Abelson’s column in Barron’s. I studiously avoid reading it. It is the single most useless column in Barron’s. You should write for Barron’s instead of Mr. Abelson. Barron’s readers would benefit.

  • Jeff Miller May 5, 2007  

    Barry —
    Thanks for your comment. The birth/death estimates have gotten large because of a fact: much of the economy was missed in the regular payroll survey. The increased size does not make the method more fictional; it makes it more necessary.
    On the 49K, I am most curious about where you got that number.

  • sk May 7, 2007  

    What I find difficult to accept is that, as you point out, the BLS on their webpage say that the “net contribution (of B/D portion )is relatively small and stable”. I wish they’d quantified that – by my reckoning, it certainly was NOT stable in 2006, and 2007 and it isn’t small either, if compared to the change in employment. It is perhaps small compared to total employment but people look at the MoM change !
    Talking of change, if they are using time series forecasting, then I’m really surprised they aren’t stationarizing the time series and using differences; perhaps their B/D model does use differences but we’ll never know since they don’t publish it !
    I hope they appreciate why this sort of ambiguity in wording and their refusal to publish the B/D model leads to considerable scepticism and conspiracy theories about “spin”, happy talk and so on. But given this government’s record on “truth , justice and the American Way” – perhaps its not just bureaucratic tone-deafness – but that’s for a different blog…

  • Mike May 8, 2007  

    Thanks for the post. On the BLS website, they do post the adjustment by sector. Also, on the website, BLS states, “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.” DO you have any thoughts or historical perspective on how birth/death adjustments may have skewed data in the past? Thanks.

  • Jeff Miller May 8, 2007  

    Thanks for taking the time to comment, K and Mike. The BLS provides a lot of information about methods and also raw data on the site. Those of us who like to dig in and do our own analysis always want more. I do not think they are intentionally hiding anything about the birth/death method, but it would be nice to have more description of the methodology.
    I cannot find a seasonally adjusted, birth/death contribution on the BLS site. Goldman Sachs is out with a report estimating the construction sector contribution at 19,000, and warning (as I did) about the error in these point estimates. The error bands overwhelm the estimates. Goldman also points to the Household Survey and quarterly unemployment records.
    There is no real history to analyze, since they have only been doing this adjustment for a few years.

  • Jene November 4, 2007  

    The bls jobs report for oct 2007 added 103,000 jobs as a result of the birth/death alchemy. That does not look ‘small and stable’ and neither do its other contributions.

  • Jeff Miller November 4, 2007  

    Jene –
    Our mission is to help investors make accurate interpretations of government data. Those of us engaged in that effort avoid terms like “alchemy.” It suggests that one’s viewpoint drives the interpretation of data rather than allowing the data to lead.
    The first thing that you should understand is that the job creation in the US economy exceeds 2.5 million jobs each month. The adjustment, is a rather minor factor that has improved prediction in every year since its implementation.
    I guess I need a follow-up post on this.
    Thanks very much for your comment. I am quite sure that your viewpoint — largely due to massive dis-information — reflects the feelings of many.