What Lies Beneath: The Jobs Report

Oh, that jobs report. It certainly did seem to refill a lot of helium-type objects. For example, some reports that said the payroll increase was more genuine because of the shutdown – but that the increase in the unemployment rate was less genuine because of the shutdown. Very artistic, that.

Then there was CNBC’s Joe Kernen, who has been blaming Obamacare for slow hiring. He concluded that hiring has now picked up because of the Obamacare deferral. The latter was announced on July 2nd, so I guess he thinks CEOs are slow on the uptake. Not to be outdone, the President then proudly tweeted the numbers on some company that just went public (TWTR). Really, there was something for everyone.

Don’t be misled by all the hype that’s going to fly around to justify buying stocks. The biggest noise is from the reports beating estimates, not that the economy underwent some sort of fundamental change. It didn’t. GDP did beat estimates for 2.5%, but the 2.8% initial estimate for the third quarter is the same as the third quarter print in 2012. The jobs report beat estimates, but according to the unadjusted data from the Labor Department’s Bureau of Labor Statistics (BLS), employment through the first ten months of 2011, 2012, and 2013 has grown at a remarkably similar rate: 1.47%, 1.46%, 1.47%. I’ve written before about the tighter data-fitting that’s been going on this year at BLS, ever since February.

You’ll also see plenty of stories over the weekend telling you that some (insert floral, over-excited adjective here) report led stocks to another weekly gain, but the truth is that at 3PM Friday, stocks were unchanged on the week. It took another one of those well-known late Friday tape-painting pushes to get stocks into the green.

Look at the data, not the stories. Since the beginning of 2012, or in other words over the last fifteen quarters, the Bureau of Economic Analysis (BEA) says that the economy has produced eight quarters of GDP growth of 2.8% or higher. Every one of these quarters has been followed by a rate that was the same or lower the following quarter, and a sharp slowdown either the next quarter or the one after. I’ve written about this before – all that really happens is that we get periodic episodes of inventory restocking and destocking. They give us quarterly excitement in both directions without the economy’s underlying annual growth rate ever changing. In fact, real GDP is up only 1.6% since the third quarter of 2012 – and we are due for another inventory rundown in the fourth quarter.

The jobs report was one of the oddest examples of the species I have ever seen, as you will see below in the Economic Beat, filled with contradictions. So was the analysis of the same, with some in the media excitedly reporting that investors are betting on an acceleration in corporate profits (with margins and percentage of GDP already at record highs), and that investors didn’t care about tapering.

They were probably the same ones who were saying before the report’s release that we shouldn’t count on this report – the pre-release hours were filled with columns explaining why we couldn’t. Others said instead that it was the increase in the unemployment rate and the underlying weakness in the jobs report that was causing stocks to rally, because that would mean deferred tapering. And good old Art Cashin thought it was just good old dip-buying.

My own reaction to the jobs report may be mirrored in ways large and small in the market over the coming weeks. My initial thought – after rubbing my eyes a couple of times – was to think gee, maybe the economy is really picking up here. This could be an inflection point. But digging through the numbers was such an unexpected, sobering experience that I felt obliged to conclude that nothing had changed at all. In fact, a thorough review of the week’s data has me more worried than I was before.

The biggest change will be in the headlines. The Wall Street Journal bubbled about “accelerating job growth” and a “solid corporate-profit season” (sic). For those of you keeping score, the third-quarter earnings growth rate is 3.4% – any more solid, we can’t afford. The number of companies cutting guidance for the fourth quarter is 73. The number that raised was 12 – any more solid, we might go broke.

The point is that the truth of the actual data may not matter in this environment, only the stock-market truth of estimates being beaten and the hurry-up-you’re-missing-out shrieking of the media. It was interesting that upticks in the German manufacturing PMI and the country’s new manufacturing orders crossed my news services many times, but nary a one mentioned the big September drops in EU retail sales and German industrial production. A day or two after the EU cut its outlook for 2013, the story had been largely forgotten in the fresh clamor for a reaccelerating Europe – one that is moving up so impressively that the central bank felt compelled to cut interest rates again (perhaps more importantly, instigating a quick devaluation of the euro).

I haven’t given up on my pre-Thanksgiving stock market fade scenario, but it’s admittedly more difficult to handicap now. The calendar is fairly empty next week, with the two noisiest events likely to come from the China plenary session (I predict that it will predict success in China) and Ben Bernanke’s speech on the 13th of November. My fade scenario was also balanced by my belief that the odds favor the usual Thanksgiving-to-New Year’s rally, which could help complete a blow-off top move we may be in the middle of right now. It’s possible we skip the fade and move right into a parabolic rally (silly) phase – and that will end in very large tears indeed. Here’s a chart from Factset that suggests why:

S&P 500 vs forward EPS projections. Source: FactSet

The last time price got ahead of forward projections like that was late 2007. The aftermath wasn’t very pleasant, as I recall. Have a good long weekend, those of you not toiling in retail trade and/or the stock market.

The Economic Beat
The jobs report caught the market by surprise, of course, and left many observers stunned, including this one. My surprise only increased as I went through the data – this is one of the strangest jobs reports I have ever seen. I kept looking back at the tables and staring, thinking I must have misread something.

There were contradictions at every turn, so many that I’m not sure where to begin. I’ll start with the fact that the household survey recorded a loss of 735,000 jobs. That stands in stark opposition to the establishment payroll figure of 204,000 jobs (seasonally adjusted). The household survey is more unpredictable and has bigger revisions, no doubt about it. I’ve seen big differences before. But a difference of nearly 1 million jobs – .94 million, to be precise – in one month, I cannot recall seeing before. It’s huge.

There’s a reasonable possible explanation – furloughed government workers may have been counted as unemployed in the household survey, but employed in the establishment survey, creating the disparity. Still, I don’t know how that can explain the civilian labor force shrinking by 720,000, or the huge decline in the participation rate, or the sharp increase in adults excluded entirely from the labor force. That’s an awful lot of work for a two-week furlough.

There’s more to say from the household survey, but let me turn first to the establishment figure of 204,000. I believe I’m on safe grounds by firmly saying that the total was largely driven by new seasonal adjustment factors, which added 15,000 jobs to the September count all on their own: The originally reported unadjusted number remained unchanged from last month. The additional 40,000 jobs the Labor Department now attributes to August was based on a unadjusted addition of only 14,000. In sum, the BLS counted 14,000 extra jobs over the previous two months and called it 55,000 more. The department did a preliminary benchmarking revision for the months since February, and seems to have emerged with a renewed emphasis on its mission to make 2013 look as much like 2012 as possible.

Coming back to the household survey, the civilian labor force shrank, the participation rate declined sharply to its lowest level since 1978, the employment-to-population ratio hit its lowest level since 1983, and the Not-In-Labor-Force (NILF) category soared to a new record (unadjusted or not) of over 91 million. All of this is completely at odds with the relatively strong picture given out by the establishment figure. Had the participation rate stayed the same, the unemployment rate would have skyrocketed to 8.1%! One of these surveys has a big problem.

Before you’re certain it was the admittedly more volatile household survey, the establishment survey was also full of weakness (apart from the seasonally adjusted job figures) that showed zero evidence of labor force pressure. It was rather the opposite. The average workweek was unchanged, and the non-supervisory workweek declined a tenth of an hour. Average hourly earnings rose by two cents. Aggregate weekly hours declined, and the weekly payroll index was unchanged. A quarter of the new jobs were leisure and hospitality and half were in entry-level jobs paying close to the minimum.

One pundit I saw on television was braying (there simply is no better term) about the large sustained increases in temporary help (which some had instead attributed to Obamacare) being the surest sign of employment strength. Well, the survey reported an estimate of 3,000 temp workers in October, the smallest increase in over a year. Transportation and warehousing employment, a category which should be rising with the economy, was unchanged.

All of these numbers can rightly be excused as representing only one month, but the real problem is that the establishment payroll estimates are on one side and nearly every other datum in the report is on the other, not to mention the recent pronounced move upwards in weekly claims. Construction job growth has been weak also. In sum, I’m inclined to take the side of those whose immediate reaction was that the number will be revised downward, but the evidence must be set against the BLS predilection this year for matching up the 2013 dynamics with 2012.

Indeed the best guide to November and December jobs announcements may be the data from November and December 2012. I wrote in Seeking Alpha two days earlier that BLS history suggested that the figure might come out on the high side at 150,000 (consensus was for 120,000), and that was because my own interpolation of BLS reporting and year-on-year pattern-matching had produced a number of 166,000. The latter seemed high to me in light of the previous week’s ADP tally of only 130,000; turns out it wasn’t high enough.

The extra jobs didn’t show up in September’s personal income and spending report, either. Although income rose 0.5% (seasonally adjusted), it was skewed upwards by farm income and one-time factors; private wages and salaries grew by 0.3%, a slowdown from August. Spending increased only 0.2%, 0.1% in real terms. If the jobs data is real, we should see significant pickups in both categories in October – or will it be November? October is likely to be given a free pass from the shutdown (unless it’s good, in which case we can trust it).

Personal spending slowed in the third quarter, according to the Bureau of Economic Analysis (BEA), which released the first estimate for third-quarter GDP on Thursday, an annualized rate of 2.8%. Most economic observers noted the underlying weakness in the report, as the rates of growth slowed in final sales (from 2.1% to 2.0%), spending (1.8% to 1.5%), exports (from 8% to 4.5%), imports (from 6.9% to 1.9%), and investment, with nonresidential fixed spending slowing from 4.7% to 1.6%, and business capital goods spending down 1.1% in September, according to the latest factory orders report (+1.7% total, minus 0.2% excluding transportation). The big difference in GDP was another large increase in inventories, which looks set to start declining after three quarters of build. As noted above, it’s been a familiar pattern and in this case is masking a slowdown nearly everywhere else.

Remember that pattern. 2.8% was the same figure as the third quarter of 2012, when nominal GDP rose at a 4.9% clip; that compares to the estimate of 4.8% for the quarter just ended. Because of the variability in inventory builds and other data that don’t follow seasonal patterns, I prefer to track year-on-year changes in GDP. Lately I’ve been dialed in on the year-year data for nominal GDP, because they’re a much better trend signal than the volatile quarterly headline figures.

I’ve warned about the danger lights that flash from annual growth in nominal GDP dropping below 3%; the first and second quarter year-year rates were 3.08% and 3.1% respectively. The third quarter rate was 3.07%. That is not at all the picture you’re slated to be bombarded with over the next few days.

Next week is a light one, as noted, with banks and bond markets closed on Monday November 11th, along with many European markets. Tuesday will bring the long-delayed Chicago Fed national activity index and the small business confidence index, both rarely good for more than a five-second mention in the press. International trade data is due on Thursday, and Friday livens things up with October industrial production, the New York Fed manufacturing survey, import-export prices and wholesale trade data (beloved by geeks like me). The highlight of the week domestically should be Brother Bernanke’s speech on Wednesday. May all enjoy Armistice Day/Veteran’s Day/Remembrance Day.