The show must go on . . .

Despite the events of the past few days, the never ending stream of data, useless and informative alike, continues unabated.

My longstanding view is that most of the data in these series is not statistically meaningful. The collection process is imperfect; the models used to analyze this data are rather flawed. And the very human tendency to over-emphasize the most recent piece of data tends to increase the total amount of volatility, be it in equity prices or belief systems.

What does matter is the overall vector of a given economic sector. Vectors include the rate of acceleration or deceleration, persistency, direction etc.  Think overall “trend” and changes thereto. For employment, this means: Are we seeing an increase in the factors that lead to hiring? What is the ratio between hires at big firms vs small firms? Are Wages increasing, staying flat, or decreasing; Temp workers getting hired, total hours worked etc. What are the likely data and modeling errors? Collectively, those factors all add up to an issue of the employment situation roughly improving, maintaining a stability, or getting worse.

Hence, each data point should be looked at in terms of whether it is continuing the overall trend, or suggesting a reversal in trend. Everything else is noise.

Consider the transition over the past five years — 2006-2011. In 2007, the entire new job creation pool was a result of Birth/Death adjustments. That was a warning sign about a major change in trend. In 2008, we saw monthly job losses of half to three quarters of a million per month. That reversed after the market collapse and massive liquidity/bailouts of 2009. In 2010, job creation barely kept up with population growth. By the time we hit 2011, even that modest job creation began faltering.

For this month, the projected gain in U.S. payrolls in September is small:  The median forecast of 91 economists surveyed by Bloomberg News was that NFP Employment likely climbed by 55,000 workers after no change in August. Jobless rate is expected to remain unchanged at 9.1%.

Of course, economists have been proven themselves inadequate at making forecasts, which is yet another reason to look at the data, and remain focused on the overall trend.

Category: Data Analysis, Employment, Really, really bad calls

Please use the comments to demonstrate your own ignorance, unfamiliarity with empirical data and lack of respect for scientific knowledge. Be sure to create straw men and argue against things I have neither said nor implied. If you could repeat previously discredited memes or steer the conversation into irrelevant, off topic discussions, it would be appreciated. Lastly, kindly forgo all civility in your discourse . . . you are, after all, anonymous.

14 Responses to “NFP Report: Trend vs Single Data Point”

  1. [...] Barry on today's jobs report: Pay attention to trend, not one single monthly data point.  (TBP) [...]

  2. Orange14 says:

    We may have reached some type of stasis with 9% unemployment which is clearly not good in the general sense of well being. However, if you look at certain subgroups within the economy things are not so gloomy which is pretty weird. Some specialty retailers have seen same store sales pretty constantly up 10% over the previous year and I’ve observed this trend all year. There are apparel manufacturers likewise that have done well this year. What this means to me is that those who do have jobs are still out there spending money on the necessities and even some “semi-luxuries” despite the overall economy. I don’t pretend to know how much longer this can continue but it certainly could go on for several more quarters barring a huge spike in unemployment.

  3. VennData says:

    Some Unemployed Find Fault in Extension of Jobless Benefits

    ‘… In a recent survey of the unemployed by Rutgers University, more than one in four respondents was opposed to renewing the current extended unemployment benefits…’

    http://www.nytimes.com/2011/10/07/business/some-unemployed-find-fault-in-extension-of-jobless-benefits.html

    Where are the calls – ala Warren Buffett to “pay more in taxes if he wants to” – to these one-in-four to stop taking these $298 checks if they don’t approve?

    You’d think the GOP media machine would be consistent and make this a loud, ubiquitous talking point. Everyone hammering it on every radio show and TV gab chance… you know how they do it… I wonder if I just missed it

  4. murrayv says:

    I suspect we are seeing a “perfect storm” of factors impacting demand and employment. The most obvious effects are offshoring of jobs, household debt and the end of the boomer peak spending years. Boomers were 78 million births in 18 years, Gen-x was only 46 million in 14 years. That’s nearly a 25% drop. Also Gen-x seem to be less “exuberant” than Boomers. That represents a big drop in demand. Job offshoring increases unemployment and drops demand. Then there have been 1 to 2 million illegal immigrants who have returned home, and their demand is gone. The debt overhang also restrains new spending and demand.
    Finally there is a real saturation effect in consumer durables. In essence every one that wants one has one (or two, or three). That applies to housing, cars, appliances, etc. There was a recent Republican complaint that 98% of the poor people in the country have a refrigerator.
    The only hi volume items sustaining demand by rapid new model introduction are cell phones and tablet computers, and they don’t represent jobs in the USA.
    We are probably in a new structural economic phase where full employment means >6% unemployment.

  5. Moe says:

    September sales show luxury good sales still chugging along. Yet elsewhere I see a record 45 million Americans are on food stamps.

    You can torture data enough to make any point you want – sometimes it’s good to take things at face value. We live in a country of the haves – and have nots. If you think of those on food stamps as all a bunch of lazy, uneducated hippies living off the government – it makes it easier to dismiss.

    We live in a country where one Leo Apotheker at HP can exist. In a little over a year the company saw a huge 47% drop in share price under his leadership. So, he gets fired – as a parting gift, for his stellar service (I guess) he gets nearly $14 million in severance. Merrill Lynch fired CEO Stanley O’Neal in 2007 after the company reported over $2 billion in losses for that fiscal year’s third-quarter. Bloomberg reported that O’Neal received a $161.5 million parachute in stock awards and retirement benefits. Home Depot CEO Robert Nardelli was given about $210 million in severance pay after he was let go in 2006. During his five-year role as head of the company, the stock price declined 8 percent.

    The end-game of Capitalism (if enough of the “democracy” is peeled away and discarded) is not unlike poker – one or two will end up with all these chips.

    In 100 years people will look back on this as we now look back on the Tulip Mania in Holland in the 1600s – delusional and pure madness.

  6. RW says:

    1. Treating economic trends as a problem in vector analysis is extremely useful.

    2. One sector booming & at full employment with another sector depressed w/ high unemployment is what we would expect if unemployment was structural (ht B dL) …but we really aren’t seeing that, yet …so the problem is still primarily cyclical and amenable to improvement in aggregate demand …so far.

    3. What Moe said.

  7. [...] Reality check for today’s jobs number. (The Big Picture) [...]

  8. dad29 says:

    ….each data point should be looked at in terms of whether it is continuing the overall trend, or suggesting a reversal in trend. Everything else is noise.

    Precisely the recommended approach in sighting in a new rifle or rifle-scope. Five rounds downrange before you TOUCH the sight adjustments!!

  9. Greg0658 says:

    ok .. boys & girls wannabe moms & dads dont .. make capital crave your creation again
    but that really won’t be the fix .. economies will tank .. hoards will overtake our borders – take our stuff and kill us
    so shape pc thoughts and be 1st to others
    accept that is the way it is

    align yourselves with the earlier arrivers & get up to speed .. power to the t-party
    kill off the streetlevel moochers
    you know you have to be strong

    a rising tide floats all boats – not when its your streetlevel neighbor with the bigger boat :-|
    (or survives doing nadda)

  10. theexpertisin says:

    Good thing the 55,000 Verizon workers settled their contract and came off the dole.

    Regarding the “trend”…well, it sucks and anything less than 160,000 private sector jobs created per month means we are going towards 9.5% unemployed by March 2012. And climbing.

  11. Futuredome says:

    Yes, Murry is beginning to get it. The time of economic boom has ended. Maybe just not for a economic generation, but for several generations. Getting people off of FS adn JC should be something that the government looks at rationally rather than intellectually. Who needs the jobs(older middle ages and youth). Just kicking them off sets the stage for declining living standards and “soft-bolshevism” that the paleo’s want further creating more slums.

  12. willid3 says:

    thinking looking wages is a better sign of trend. wages go up if there is real growth in the number of jobs. and we can see a lot of how well sales are really going by looking at sales taxes

  13. [...] through knee-deep mud, but not in economic collapse. That said, prior to the report, Barry Ritholtz offered some wisdom regarding individual data points versus trends: What does matter is the overall vector of a given [...]

  14. [...] knee-deep mud, but not in economic collapse. That stated, prior to the report, Barry Ritholtz offered some wisdom regarding individual data points versus [...]