Non-Farm Payroll Prediction: Huge Beat Of Consensus Estimates, SPX Melt-Up
by Dave Kranzler, Investment Research Dynamics
As John Williams of Shadowstats.com reiterates every month when he presents his analysis of the monthly jobs report, the non-farm payroll report is the most useless, unreliable and politicized data report produced by the Government.
One my biggest pet peeves with this report is the so-called “Birth Death model,” which is nothing than a statistical plug used by the Government in order to pad the number of jobs created during the quarter. Of course, the headline report only shows the number of jobs created. You have to dig into the bowels of the report itself in order find the number of people who left the labor force. Typically the number of people who leave the labor force every month greatly exceeds the number of “new” jobs “created.”
I use the word “created” quite loosely there, because a preponderance of the jobs are created by a few computer keystrokes at the BLS.
My friend and colleague, Mark Kellstrom of Strategic Energy Research, has written an informative essay which describes why he (and I) think the payroll report a week from tomorrow will exceed the consensus estimates and trigger a melt-up in the stock market:
April Non-Farm Payrolls Preview: Look for a beat next Friday – Mark Kellstrom, Strategic Energy Research
In this Orwellian universe we presently inhabit, little to nothing currently carries less statistical legitimacy than the monthly U.S. Non-Farm Payrolls report. We know that the Bureau of Labor statistics has a host of measures used to manipulate the data including labor participation, removing headcount from the job seeker category, not separating part time versus full time jobs, counting temp workers, counting census and other transitory government workers, making seasonal adjustments, diverging from the household survey, fudging survey data, revising prior numbers and lagging the impact of items like the roll over effect of oil prices on the energy related jobs market, etc.
However, my favorite tool used by the BLS to massage the monthly non-farm payrolls report and the unemployment rate is the “Birth Death” plug model. The Birth Death model is used by the BLS to guesstimate how many jobs were added in a given month due to business start ups, offset by business failures. The Birth Death plug numbers, fluffing up the monthly report, rarely have basis in economic reality and are typically removed from the labor statistics with a large one time downward revision in January, before the imagined Birth Death job additions grow anew over the next 11 months of NFP reports.
In watching the reported “Birth Death” model over the last few years, I have noticed a very regular pattern in the quantity of jobs “added” each respective month using this dubious model. A look at the historical data since 2012 for January through June is included below. One can see a very distinct pattern in how the model “creates” jobs during that time frame. The reported monthly numbers can be revised to a small degree over the course of a year but not by a material amount until the January washout.
So, I am testing a theory this next week: If history holds, when the April Non-Farm payrolls are reported on next Friday (May 8, 2015), then the NFP number should be robust and beat expectations. The reason is simple. Current consensus forecasts (I believe) expect the April NFP report to show that approximately 250,000 jobs were added to the labor pool. However, if historical patterns repeat, then the BD model will add between 225,000 and 250,000 jobs in April alone, meaning that few or no real jobs are required to match consensus—the BD model sets up the beat by itself……unless of course, the BLS finally subtracts the many thousands of newly unemployed in the oil & gas space from the Birth Death gains.