A stumble even in the Establishment Survey is not at all unexpected even in the best of times (in reality as opposed to purely statistical assumptions), however the furor unleashed over the March “weakness” is everything you need to know about how much surety rests within the estimate to begin with. Like human psychology, anyone who possesses a high degree of self-confidence is not so easily bothered by even the most efficient criticism. Likewise, confidence in the unemployment rate and the Establishment Survey seems inversely proportional only to the degree to which its adherents feel forced to defend it.

If the Establishment Survey’s trend-cycle benchmark has been established mostly by jobless claims rather than a broader set of economic indications, then March’s data may actually be worse than it appears. Historically, low levels of initial jobless claims correlate with above-average gains in the Establishment Survey which can reasonably assumed solely an artifact of cyclical assumptions that haven’t held this turn. Thus, if the BLS is taking jobless claims as a signal to amplify upward variation in payrolls, as I believe the statisticians are doing, then March suggests that economic conditions so far in 2015 may be much worse than feared. If whatever is actually taking place in the real economy is enough to knock even the Establishment Survey off its straight line and causing vociferous and heightened densification of those rushing to deny any chance of a downward drift, then it may truly be nothing short of dislocation in progress.

At this point, that is just an amplified possibility as one month proves absolutely nothing. It would be, however, consistent among nearly all other data points – which is precisely the point.

In just those cyclical terms, the labor force has shrunk since January’s massive “discontinuity” (the BLS hasn’t confirmed as much as a population adjustment, but the only other months with +1 million changes in the labor force were January 2000 and January 1990). In February and March, the counted labor force has declined by a rather sharp 274k combined. Ignoring January, the labor force has shrunk by 325k going back to March 2014.

ABOOK April 2015 Payrolls LF

That disparity alone gives us a sense of both cyclical delineation as well as lingering structural deficiency. Going back to October 2012 when I maintain the elongated cycle peak occurred, the total labor force apart from January 2015’s anomaly has only gained 364k out of an increase in potential labor of 6.5 million! Even including January 2015, only 1 in 5 potential workers has bothered to look for work? It just doesn’t add up to what the unemployment rate and Establishment Survey are projecting.

The problem of any cyclical inflection is obvious, but it is compounded exponentially by this lack of recovery. We cannot but speculate as to how that may turn out as this would be an entirely new occurrence in economic history: large recession, no recovery and then another recession. In one sense, that might offer a bit of hope for reprieve since labor utilization has remained at such a low level throughout, and thus there is so very little “fat” to further trim in a cyclical downturn.

On the other hand, with an already weakened structure never reinforced by expected wage and (actual) job gains, spending levels may not just be affected and further weakened but do so in a non-linear fashion.

ABOOK April 2015 Payrolls FTABOOK April 2015 Payrolls FT Recovery

The actual accounting of peak-to-peak “gains” makes this a frightening possibility, starting with, of course, full-time employment. The current estimates (which will undoubtedly be revised and likely lower if the turn is at work) show that there are still 851k fewer full-time jobs in March 2015 than November 2007. But even that understates the extent and proportion of the economic difficulty that remains.

ABOOK April 2015 Payrolls Accounts

The problem with this recovery is assigned not so much by what has been evident and chronicled but more of by what is missing. The Yellen version of the economy believes only in the most recent “gains” without addressing the trend-cycle assumptions which should have been wholly invalidated by 14 million missing souls.

ABOOK April 2015 Payrolls Structure

Since November 2007, the Establishment Survey shows a gain of 2.93 million payrolls – which seemingly “satisfies” the increase in the labor force of about 3.07 million (including January 2015). Already there is a weakness inherent in that cyclical re-linking, as, again, the count of full-time jobs remains decreased by the Great Recession. So of the total marginal additions in labor, all of it (more than all) went into part-time positions.

Setting aside the disparity already with the Household Survey which shows 1.2 million fewer cyclical gains, that total recovery only accounts for a small fraction of the then 17 million increase in labor potential. This is the crux of the “participation problem” which has very little to do with an aging population as is so often asserted (older workers are clearly remaining workers in historical proportions rather than retiring).

ABOOK April 2015 Payrolls Missing

What that shows, to me, is an economy that has shrunk considerably since 2007 in overall systemic productive capacity – which is exactly the worry in terms of the next recession whenever it may strike. Economic behavior was terribly altered by the Great Recession itself, which is a fact incorporating much of these “missing” to begin with, so it is impossible to suggest and predict how that might turn again if another recession was to “launch” from such a shriveled base. If a recession causes the “bunker mentality”, what happens if a second recession appears with consumers still inside the bunker?

Again, this would be entirely unique, as even double-dip recessions in the past offer little as far as direct comparisons. In 1981, the economy was far, far worse the second time than the prior 1980 recession – the unemployment rate peaked at 7.8% in July 1980, rebounded only to 7.2% and bounced around that level until July 1981 before again surging all the way to 10.8% by the end of 1982. In 1959, the unemployment rate peaked that November at 5.8% before dropping to 5.1% in May 1960, and then surging to 7.1% by the trough of the 1960-61 double dip. Both of those occurrences are suggestive of a worse run the second time, made all the more worrisome as the double-dip recession process was essentially a single “cycle” carried out over two “events.”

I think that is what the 14 million “missing” might suggest about the current “cycle”, as it would mean that the Great Recession never really ended and we are still within the throes of a singular process suggested by/linked to labor woes.

Fortunately, that is a consideration about future conditions which may never come to pass, as there is enormous uncertainty how even current circumstances might evolve let alone if they do at all. So far in 2015, economic indications are well beyond the “normal” and expected weakness inherent in a highly deficient cycle and the March Establishment Survey is relevant insofar as it may confirm that interpretation. If April comes out as weak, then, given the trend-cycle amplification through the jobless claims comparison, the lack of buoyancy, economically-speaking, will be shown a much closer possibility.