Considering the Precedents on Payrolls

Explaining a freakishly low jobs number.
Publish date:

Pop Song 99

JACKSON HOLE, Wyo. -- Should we talk about the weather?

Your narrator fully expected all kinds of idiots to draw all kinds of idiotic slowdown conclusions from the March

employment report


And still they managed to surprise on the upside.

Know what it means when payrolls print waaay below trend?

It means they print huge a month later.

Monthly payroll gains were trending at 233K (on a 12-month moving average basis) as of February. Then March

delivered a paltry 46K increase. So payrolls printed 187K below trend last month.

Economists are exactly like lawyers in that most of them are delusional enough to think that it takes a particularly big or clever brain to do what they do. But it doesn't. Any simpleton can grow up to be either (witness the overpaid state-school pea-brain writing this column). Why? Because precedent does much of the work for both groups. And precedent usually ends up proving pretty persuasive.

Consider that a 150K below-trend payroll print in July 1998 led to a 322K-job increase in August. Consider also that a 199K below-trend payroll print in March 1998 led to a 320K-job increase in April. Consider also that a 173K below-trend payroll print in August 1997 led to a 386K-job increase in September. Consider finally that a 233K below-trend payroll print in January 1996 led to a 479K-job increase in February.

Is it possible that Friday's payroll print will go down as the first in a series of relatively weak job increases? Sure thing. Absolutely anything is possible in a world where sissy

American League

pitchers can get away with never having to pick up a bat.

But it's much more likely that the swing between the warmest winter since 1895 (seriously) and brutal weather during the March 12 week (the week during which the


collected employment-report data) ended up producing a freakishly low jobs number. Period.

Or look at it this way. Any slowdown type who really thinks the March payroll number is meaningful -- that it had very little to do with weather -- ought to be very happy to bet big that the April number is likelier to print south of 200K than north of 300K.

Put up or shut up. That's the Law around these parts.

Why is it that nearly everyone who writes up the employment numbers obviously has no clue as to how the unemployment rate is actually calculated?

Partly because labor markets are so tight that they need not fear being included in it (sorry job performance be damned). But mostly because idiotic economics writers quote the idiotic pronouncements of idiotic economists without ever verifying them.

The jobless rate is calculated by dividing the number of unemployed persons by the sum of the unemployed and the employed. In other words, it is calculated by dividing the number of unemployed persons by the labor force.

People tend to pooh-pooh decreases in the unemployment rate when they come against a backdrop of a decrease in the labor force; exactly such a reaction greeted the March numbers. The unemployment rate

fell to 4.2% last month from 4.4% in February while the labor force decreased by 455K (mostly as a result of a huge decrease in the number of unemployed), so most folks labeled the two-tick decrease meaningless. "The March employment report wasn't as rosy as the low unemployment rate suggests" is how one jackass put it.



, think again. Decreases in the unemployment rate are at their pinkest when the household

measure of the employed is accelerating, and that's exactly what's been happening since last autumn. It rose 0.4% during the third quarter; another 2.3% during the fourth; and another 3.1% during the first. This brought year-on-year employment growth to 1.7% in March from 1.0% in August.

Anyone pooh-poohing the drop in the jobless rate in March ought to be willing to bet that it's now likely to see 4.4% again before it sees 4.0%.

Take that action if you can get it.

Did the March employment report please Greenspan?

No. But it did make the hawks very happy because it upped the chances (however marginally) that the next move in the

funds rate

will be up.

Waste as much time as you like marveling at the fact that the employment-report measure of nominal

wages is rising at "only" a 3.4% year-on-year rate. But do keep the following in mind.

The number of people willing to work can be usefully defined as the unemployed component of the labor force plus those not actively seeking work, and thus not counted in the labor force, but who nonetheless say they would like a job if they could get one. This pool of potential workers aged 16 to 64 currently numbers about 10 million, or just 5 3/4% of that group's population -- the lowest such percentage on record, which begins in 1970, and 2 1/2 percentage points below its average over that period. The rapid increase in aggregate demand has generated growth of employment in excess of growth in population, causing the number of potential workers to fall since the mid-1990s at a rate of a bit under 1 million annually. We cannot judge with precision how much further this level can decline without sparking ever greater upward pressures on wages and prices. But, should labor market conditions continue to tighten, there has to be some point at which the rise in nominal wages will start increasingly outpacing the gains in labor productivity, and prices inevitably will begin to accelerate.

That's from Greenspan's February



Note that alternate measures of nominal wages are rising faster than the employment-report measure. The wages and salaries portion of the

employment cost index

is set to show a 3.7% year-on-year increase for the first quarter; wages in the nonfarm business

sector are rising at a 4% year-on-year rate; and wages in the nonfinancial corporations

sector are rising at a 4.3% year-on-year rate. Also note that every measure of nominal wages under the sun is now accelerating faster than productivity, which is rising at a 2.8% rate (and that's generous).

Also note that the key unemployment rate G mentions continues to fall. It has dropped to 5.1% from 5.6% over the past year and now sits at its lowest level ever. (This is the U-5

series in Table A-8. The BLS provides this rate only on a not-seasonally-adjusted basis, but applying the seasonal factor used for the regular unemployment rate seems reasonable enough.) Also note that job leavers as a portion of the unemployed hit 13.5% last month; that marks the highest such percentage of the cycle. Employees are now bolder than they have been at any time since December 1990.

Bring on the rainmakers.

Side Dish

Kevin Brown vs. Randy Johnson? Oh. My. God.

Most likely to win Series?