Employment Report Disinformation

Each month the market reacts to the employment situation report, consisting of a survey of establishments and a survey of individuals.  There is a lot of information and the methodology is complex.  This creates a great opportunity for "spinning" the data.

Why  Spin?

Frankly, we do not know the motives of the various sources that consistently seek the worst possible interpretation of each new data point.  The difficulty for most traders and investors is that they will not study the data enough themselves.  Even if they did, almost no one has the methodological skill to evaluate what they see.  This leaves nearly everyone at the mercy of those purporting to be experts.  Some of these sources are well-intentioned, but mistaken.  Others may simply have another agenda.

At "A Dash" we are quite willing to be led by data.  Regular readers (and our investors) know that we will sell and even sell short stocks as the circumstances dictate.  We are bullish, but not perma-bulls.  We are consistent and objective observers of the  market, as our investors from the 2000 era know well.

Analytical Errors

The problems relate to three different issues:

  • The Birth/Death adjustment.  Monthly job creation is about 2.5 million jobs.  For many years the BLS assumed that establishments not responding to the survey were replaced by new establishments with the same number of jobs.  A few years ago they realized that this method under-estimated actual job creation, so they implemented an adjustment.  This was based upon many years of analysis and is reviewed twice a year, comparing it to actual data.

This method has worked very well.  It has improved the estimate of employment, by actual count, in every year since it has been implemented.

Ignoring these data, leading critics (including Ritholtz, Kasriel, Abelson, Maudlin, Norris, and others) focus on the adjustment and not the overall problem of estimating changes in establishments.  Some of the critics confuse total job creation with the increment from the adjustment.

We have a simple challenge for these critics:  In the years since the birth/death adjustment has been adopted, has it improved the estimate of overall payroll employment?  Has there been any single year in which the adjustment has made the overall estimate worse?

Check out this comment from former Dallas Fed Pres. Bob McTeer to confirm our analysis.

  • The Bar. Those wishing to minimize economic progress claim that the needed monthly job growth is a gain of 150K jobs.  Mauldin actually cites 166K jobs.  Who knows where they get these numbers?  Economists observing the change in demographics (including those at the Fed)  cite the needed monthly gain at 125K jobs or so.  Some estimates are as low as 110,000.  If the economy is experiencing a planned slowing, a monthly growth somewhat lower is to be expected.
  • Household Survey.  The household survey looks at people rather than businesses.  It captures those who are working as consultants and new small businesses.   Those bearish on the economy (wrongly) dismissed this information during the multi-year expansion of employment.  The survey has a 90% confidence interval of +/- 400K jobs.  John Mauldin embraced this reading in his regular commentary.  This month it showed a big decline.  Last month it showed a big increase.  Why comment on this only in the month when it supports one’s viewpoint?


Our conclusion, looking objectively at all of the data, is that employment is growing, but the pace is lower than past years.  We do not have a political agenda.  The results are consistent with the "soft landing" scenario that bearish pundits (now 20% losers on the S&P 500) told you was impossible.

Investors must either develop the skill to interpret data for themselves, or they must find the right sources.

You may also like


  • Bill aka NO DooDahs! November 4, 2007  

    Or, traders could ignore that data. Buffett is a trader who successfully ignores it.
    Quite frankly, to use economic data in trading implies not only that one has an accurate interpretation (either one’s own, or one from someone other than Ritholtz), but that one also has a model which tells them what sectors or stocks to be long or short in, BASED on the interpretation.
    Isn’t it simpler to just use a system that works over time, i.e., trend-following, value investing, CANSLIM, or whatever, and ignore the economic data?
    IMO the only way to play econ data is from the contrarian POV. My systems tell me to be long; it is confirmed by the plethora of obviously flawed analysis screaming to be “cautious” or predicting a recession. To the point that others actually place money on these flawed analyses, it increases my confidence in what my systems were already telling me.

  • Jeff November 4, 2007  

    As you know, I have read the description of your methods. I also follow your systems, and your approach to testing. It is a good method with strong potential for future profitability.
    There are many different ways to profit in the market. One issue is that many people like to focus on “macro-economic commentary.” Many of these potential investors have been sidelined for years, missing out on investment gains.
    Explaining this is part of my mission, but it is difficult. The criticisms are so easy and so believable for the average person — government lying or screwing up, etc.
    Thanks for an accurate comment.

  • Barry Ritholtz November 5, 2007  

    At best, its arguable to say that this B/D has improved the data — at worst, it is significantly distorting it.
    1) The new methodology has been in place for less than a full cycle (recovery, recession, recovery).
    I suspect it increases the accuracy of the number in the beginning of the cycle, but then is likely to get it wrong at the end of the cycle. We see evidence of this in the B/D creation of construction and financial jobs as of late.
    2) The percentage of B/D, at 80%, has dramatically changed the value of this data point. NFP has ceased to be measured; its now “modeled.”
    3) The recent BLs benchmark revisions have become increasingly large 800k+ in 2006, 300k in 2007 — these are much higher than they have been historically.
    4) Replacing a measured data point (payrolls) with a hypothesized one (estimated new companies hiring) dramtically allters what BLS is doing — from measuring to estimating.
    There is no conspiracy. To BLS’ credit, all of their data — CES/CPS — is right on their site or anyone who wants to do the math themselves . . .

  • Bill aka NO DooDahs! November 5, 2007  

    Hi Jeff, a clarification and a distillation:
    I loved the post, don’t get me wrong, and I understand your goal is to re-educate the mis-educated where macroecon-based speculation is concerned; I think it’s a laudable goal and one that I share, to the extent that I comment on econ.
    Macroecon-based speculation is a two-step process, yes, it can be successfully done, but it has additional complexities over other methods. Unless of course, one is just taking the other side of the bets made by the drastically misinformed who have flawed methodologies …
    I felt compelled to respond to the sentence “Investors must either develop the skill to interpret data for themselves, or they must find the right sources.” because there is a third choice, e.g. ignore the data. Perhaps inserting “who trade based on macroecon” right after “Traders” oh, whoops, you typed “Investors” would be good?

  • Jeff Miller November 5, 2007  

    Barry –
    It is always helpful to have your comments!
    I feel that you are not responding to the major point: Births and deaths of businesses have ALWAYS been modeled. Before this adjustment, the assumption under the old model was 1 to 1. You and I discussed this many months ago. You seem to think that worked just fine. Why?
    Your argument implies that it is OK to assume 2.5 million jobs each month from the “old” method, but you are going to highlight the 100K or so from the additional adjustment. Why?
    As to the method breaking down during the cycle, I welcome any critic showing a single year where the adjustment has worked worse than the “old” method. The BLS also points out that they monitor this twice a year.
    So much of the world follows your views on The Big Picture, that it is almost impossible to correct this erroneous conclusion. It is repeated by everyone.
    Thanks for stopping by to engage and explain.

  • Jeff November 5, 2007  

    Bill –
    While I do not mention it in every post, I have often mentioned the Warren Buffett attitude of ignoring the economy.
    There are many ways to trade and invest successfully. One of the factors leading certain stocks to be undervalued is an attitude of gloom about the economy. In my own investment management, I look for themes like this and select some stocks to fit. Others do the same, so they are the main audience for this type of story.
    Thanks for emphasizing this point.

  • Tim November 5, 2007  

    I like reading the comments here as much as the blog sometimes. Insightful from all involved. My belief is the macro information gets so skewed by the agendas/mind sets of those doing the reporting it provides no useful information in regards to my investing. Longer term trends will show themselves long before the talking heads give up their current beliefs. Jeff, your info on this site is very good and I hope some others get value from it. I surely do.

  • VennData November 5, 2007  

    I think another doozy is “the Fed added 41B in liquidity [the other day.]” so they must be nervous. That’s just repo rollovers.
    But wait, are you saying that the employment numbers are marked to model?

  • Jeff Miller November 5, 2007  

    You are so right! The Fed rollovers attracted a lot of attention. I guess there is a need for a blog that tracks all of this stuff. I certainly cannot keep up with it.
    Those who begin with an opinion about money supply, inflation, and the Fed, will find evidence everywhere, no matter what the reality.
    Eventually, the truth will out. Meanwhile, who knows?
    Thanks for a perceptive comment.

  • Aaron November 6, 2007  

    Very interesting post. I found this to be one of the best posts yet on A Dash. I think the point about all of these numbers being spun every which way is very important. It is amazing how some commentators or perma bulls or perma bears will use the household survey whenever they find it beneficial then the very next month say that the survey is useless. I also think that the bar is spun often too. The bar is set to get their point across to their audience, not to inform or truly guess what the bar is.

  • Blackvegetable November 11, 2007  

    Wow….did my dissent on Malpass, complete with incriminating citations, get scrubbed?
    So much for weathering challege, eh?

  • Blackvegetable November 11, 2007  

    My apologies……my comment remains on the board….

  • Jeff Miller November 11, 2007  

    Blackvegetable –
    I am sorry that I did not acknowledge your earlier comment. I have been fighting the flu bug, but trying to do my best to keep up:)
    I very much appreciate any comments, especially those that argue a viewpoint with evidence, as you did.
    I carefully separate my political views from my investment perspective. I am well aware of David Malpass’s background. I think that he has had a good read on the economy for some time, and have written about that. More recently he has become more cautious, although still not seeing a recession as likely.
    He is one of many good sources.
    Meanwhile, my policy on deleting comments is simple. I delete spam — that’s all.
    I hope that you will continue to join in with comments. We all benefit from a diversity of viewpoints.
    Thanks again.