Taking Back America!
Monday's MOC post on the March BLS employment statistics for Wisconsin caused a bit of a stir, and since the April data will be published in between the recall primaries and the recalls themselves, this may be our last chance to cool heads down so they can prevail.
So, one last time explaining BLS data and that’s it from me. Apparently, some people thought my link to a historical table in the other post was an attempt to trick the reader or invent my own data. Believe me - I have neither the time nor the inclination to trick you. So this time I will use only the data found on the very same BLS press release that the media quoted last week when they ran their stories that Wisconsin led the nation in job losses over the past year. Here’s the link, nothing up my sleeves.
You will see it starts with an executive summary that is very familiar, many of its phrases were lifted verbatim by the media who reported its findings. And if you scroll down, you will see a whole section on methodology and notes on what the data means and how it is gathered and corrected for sampling error and other statistical “noise”. Nothing sinister about them; routine statistical caveats.
One key definition is the term “labor force” which is defined as “the sum of employed and unemployed persons”. Because the table only displays two of those three data points, it was necessary for me to subtract in order to derive the other. Subtracting is not inventing numbers; please do not start.
And after methodology comes the data – ten tables and two maps that show both seasonally adjusted non-unadjusted data that the executive summary describes. Tables 2 and 3 are labor force tables, unadjusted and seasonally adjusted - they are the numbers of people in the labor force and people not working. Tables 4 and 5 are the establishment tables that estimate the number of jobs people work at in firms.
So, how many jobs have been gained or last in Wisconsin from March of 2011 to March of 2012? That is a very good question that has many correct answers.
If you ask people, the answer is a gain of 21,600.
If you ask establishments, the answer is a loss of 13,900.
If you apply “seasonal adjustments” the people answer changes to a gain of 7,200.
If you apply “seasonal adjustments” the establishment number becomes a loss of 23,900.
That last number is the only one you read about in all the headlines last week. Pick the one you like and knock yourself out; whatever you believe is true for you.
Nobody has said much about those seasonal adjustments and they should. It is only after the seasonal adjustments have been applied to the establishment data that Wisconsin’s losses exceed the range of statistical error. BLS fully discloses that it uses different adjustment factors for each state so any state-to-state comparisons are tricky at best. It’s not their fault people don’t read their warning labels – i.e. the methodology notes - and it doesn’t help that their own executive summaries disregard the cautions of the statisticians who develop the data sets.
Since economic data influences the stock market, the investment community has long been critical of BLS inaccuracy and contradictory data, and no one outside of BLS really understands how data is adjusted for seasonality.
When the economy is moving, little wobbles in seasonal adjustment method were insignificant, but now that we have been basically stagnant for three years, fluctuations in seasonal adjustments suddenly become very important to understanding the relevance of the summary data reported.
How relevant? In March of 2012, the results of labor force survey were “seasonally adjusted” to add 15,100 people to Wisconsin’s labor force, reduce the number of unemployed by 22,100 and increased the number employed by 37,200. So the adjusters are more powerful than Scott Walker.
“Hello, this BLS, are you employed?”
“No, I’m not.”
“I’m adjusting you to yes.”
“But I’m not employed!”
“You are now.”
“Because it’s March.”
This year’s adjustment is big, but last year’s was even bigger - BLS March 2011 seasonal adjustment increased the number of employed by 51,600. So Wisconsin “lost” 14,400 “jobs” year to year just in the difference of seasonal adjustments to the labor force survey. And if that is not confusing enough, different adjustment factors were used to adjust the CES totals than were used to adjust labor force data. Advil, please.
Remember, we are counting two different things – two way-different things. Before any seasonal adjustments, there were 144,500 more people who said they were working (labor force table) in March of 2012 than the number of jobs the establishments (CES table) reported. That in itself seems unlikely.
But after seasonal adjustments, that gap shrinks to 126,100; two different adjustment factors are being used for seasonality. How does that work? Does spring come earlier for people working than to the establishments they work at?
People matter; and the number of them that are working is the important metric – that and how much they make, which nobody is even talking about anymore. Wisconsin’s real job growth is bad; there is no need to pile on Governor Walker with adjustments and statistical deviations that turn slight gains to slight losses and confuse everyone.
It would be terrific if we had a diligent press that would at least inform us that the BLS provides contradictory data instead of only reporting one number out of four as if it were precise right down to the individual job. I don’t want to rely on campaign ads to get my economic news and you can bet the next few weeks the outside groups will be citing data all over the place about Walker, Barett, and Falk. There is a good reason that you have to swear to tell the truth AND the whole truth in court.
Ditto for President Obama at the national level, by the way – the April job numbers came out today, and he is getting pilloried for a paltry 115,000 net job growth, less than the rate of population increase. But that is the seasonally adjusted CES number. According to the non-adjusted population (CPS) survey, there are 169,000 more people who say they are working.
So all you guys who say that I am too hard on President Obama, call him and tell him I said to use the CPS data from now on – no hard feelings. It’s bad enough that three years into his recovery plan we are still treading water despite claims to the contrary; there is no reason to pile on him with adjusted numbers that make it even worse.
No statistical analysis - adjusted or non-adjusted - is going to change the obvious: Wisconsin is stuck, the nation is stuck, and the world is stuck. The debt bomb that went off in 2008 is still raining shrapnel on us and neither party has a plan that will make it stop.
It won’t stop until government spending is brought to heel and we quit running deficits that have choked the economy into a coma that seasonal adjustments don't cure.
“Moment Of Clarity” is a weekly commentary by Libertarian writer and speaker Tim Nerenz, Ph.D. Visit Tim’s website www.timnerenz.com to find your moment and order Tim’s new book, “BRING IT!