How Good was the Payrolls Report Really?


A Lagging and Imprecise Indicator

As we have pointed out in our most recent update on manufacturing data, last Friday’s payrolls report will eventually be revised out of recognition (in previous months, a number of reports have at first been subject to an upward revision, only to be revised significantly downward again a month later. The final numbers will take still longer to arrive, up to a year if memory serves).


workersHammer guy is included the non-farm payrolls report, Ms. Sickle isn’t …

Photo via



This fact and the fact that employment is a lagging economic indicator makes it extremely odd that it is held to be the most important data point on the radar of the central planners at the Federal Reserve. However, the FOMC’s own statements indicate as much, and this policy focus is after all in line with the Fed’s bizarre mandate, which enjoins it to both keep price inflation low and employment high, by means of manipulating interest rates and the money supply.


1-PayrollsChange in non-farm payrolls, monthly – total (black bars) and private sector payrolls (red bars). Last month the government was quite busy hiring – private sector employment actually grew at a slower pace than in the two downwardly revised previous months. Government drones are not creating wealth – they are consuming it – click to enlarge.


This is of course in keeping with the Keynesian philosophy – Keynes’ system can be called a “labor-based macro-economics”, as Roger Garrison points out in Time and Money. Primarily Keynes was concerned with fighting unemployment by means of inflation – i.e., by lowering the real wages of workers by pulling the wool over their eyes.

It is noteworthy that Keynes admitted that lowering unemployment required an adjustment in wage rates – he just didn’t want wage rates to be lowered in nominal terms, as he believed workers would be more inclined to accept a decrease in their real rather than in their nominal wages. This belief was thoroughly shattered in the 1960s and 1970s, when unions began to watch price indexes like hawks.

Since then, the calculation of price indexes has been changed beyond recognition too, with the result that the same price increases nowadays result in much smaller increases in CPI than was previously the case. This is achieved by methods that are quite questionable in many respects (such as the use of “owner-equivalent rent” instead of actual rents charged in the market, various basket weightings based on the “substitution principle” and of course “hedonic indexing”).


Last Friday’s Payrolls Report

Anyway, even though it stands to be revised significantly, there are a few observations that can be made about Friday’s report. A hat tip to our friends Michael Pollaro and BC for providing some background information and charts (for a detailed presentation of the data as reported, the regular update by Mish is most useful). First of all, it is worth noting that the so-called “birth/death model” included the largest addition of jobs from thin air of the entire year (+165,000).

While these numbers cannot be simply added to or deducted from the total, the fact that they have reached an annual high still indicates a significant upward distortion in the reported data. Why is it likely to be a distortion? According to the Census Bureau and Gallup (you can read about the details here), more companies are dying than are being born in the US since 2008. One would therefore assume that the B/D model should actually deduct from reported payroll gains instead of adding to them.


2-Business birth and deathBusiness closing vs. start-ups, via Gallup.


It is also noteworthy that the rate of change of the growth in withholding taxes received by the treasury has clearly turned down (and sits currently at zero year-on-year). If jobs are growing, why isn’t the treasury getting more money than a year ago? There are two possibilities here, and both probably impinge on this datum. For one thing, actual jobs growth may well turn out to be slower than it currently appears once all the revisions have been concluded. Secondly, many jobs are in “low quality” categories with lower pay, including many part-time jobs designed to side-step the impositions of the Unaffordable Care Act. Below are three charts provided to us by BC that are illustrating the withholding receipts situation.


3-withholding-1Treasury withholding receipts y/y change rate – currently at zero – click to enlarge.

4-withholding-2A 12 month average of the 20-month annualized rate in withholding receipts – this is showing longer term trends more clearly – click to enlarge.


5-withholding-3Withholding receipts in millions, wage and salary disbursements in billions, and the 6-month average of withholding receipts, long term – click to enlarge.


As you can see, these data are quite volatile, but a spot of weakness has begun to creep in that indicates that all may not be well in the labor market. There is also the curious fact – once again in evidence in Friday’s report – that employment growth has shifted strongly to the older portions of the population, while young people have ever more problems finding a job.

As we have shown in Bill Bonner’s Diary update posted earlier today, the youngest and presumably most unskilled young workers are actually in the worst position by far. In an unhampered free market economy this would not be the case. Institutionalized unemployment of this sort is solely due to so-called “pro-labor” legislation, which keeps people from voluntarily entering into labor contracts in line with actual market conditions. The most unskilled and youngest members of the labor force are disproportionately suffering from minimum wage and similar regulations.

According to last month’s household survey, job gains for people 55 years or older amounted to a strong 378,000 – while those aged between 25 and 54 years actually lost 35,000 jobs. Below we show the employment levels of the two age cohorts in indexed form (01/01/2000 = 100), which provides us with a stark illustration of the diverging trends that have been in force for quite some time now (note that employment for the 25-54 age group is of course larger in absolute terms).


6-employment levels comparedEmployment levels of the 55 years+ and the the 25-54 age cohorts indexed – click to enlarge.


In short, those who want to find work apparently can’t find enough, and those who thought they could retire by now actually can’t afford to do so.


Forward-Looking Indicators by Comparison – Mind the Gap

The next chart compares total business sales to employment data. As you can see, an ever larger gap is opening up between the two – and not surprisingly, business sales tend to lead (there was an odd out-of-character spike in 2007/8 after they initially led to the downside, possibly due to some quirk in the data we are not aware of).


7-Business sales and employmentTotal business sales vs. total non-farm payrolls and total civilian employment – click to enlarge.


Along similar lines, here is a carefully annotated chart of a total output composite via Lance Roberts of This confirms our contention that there is already quite an effect from the fracking bust, in spite of the fact that it is only beginning to enter “phase 2” (the initial phase was primarily characterized by a collapse in rig counts; phase 2 is when the marginal players will run out of funds to pay their debt).


8-Output compositeThe STA output composite index with annotations – click to enlarge.


A renewed “restocking cycle” is not in sight yet, as the next chart by Michael Pollaro indicates. It shows inter alia the year-on-year change rate in manufacturing sales and the inventories to sales ratio, which remains near recent highs:


9-Manufacturer's salesAnnual change rates in new manufacturing orders, inventories, sales and the ratio of inventories to sales – click to enlarge.


Lastly, initial claims remain at an extremely low level historically, but as we have previously discussed, this is actually a contrary indicator. Recently claims have perked up a bit from their lows, but this may still be a seasonal quirk rather than the beginning of a new trend. In fact, these data are seasonally adjusted and therefore don’t show the raw numbers, which are often diverging significantly. However, once an upward trend does emerge, it will be a “red alert”. Initial claims correlate extremely well with the stock market’s trend (historically, they are essentially a coincidental indicator of the stock market).


10-Initial claimsInitial unemployment claims, weekly – perking up a bit from an extreme low – click to enlarge.



We haven’t discussed the poor labor force participation rate here, as that topic has been belabored (so to speak) many times already. However, one shouldn’t lose sight of it: it shows that unemployment is in reality much larger than the official unemployment rate suggests (once people have “given up” looking for a job and are therefore no longer part of the statistic, the fact remains that they are unemployed. If not, then what are they?).

The main point we wanted to make is simply that there are very good reasons to doubt the quality and durability of recent employment gains. Moreover, it seems likely that further downward revisions will eventually be made. Obviously, lower quality jobs are better than no jobs at all, but their relative growth is at least in part a reflection of the negative long term effects of the malinvestments undertaken in previous business cycles. As we often stress, it is not true, as Keynes averred, that “we are all dead in the long run” – unfortunately many of us tend to still be among the quick when the long delayed long term finally becomes the present.


heroicRelief in Slovakia showing heroic workers – a relic of the communist era

Photo credit:


Charts by: St. Louis Federal Reserve Research, Gallup, BC, Michael Pollaro, Lance Roberts/STA, StockCharts