“Falling leaves hide the path so quietly.” – John Bailey
The good news about last week was that it was up (barely) for the sixth week in a row with an impressive close Friday, even allowing for the usual end-of-week gaming and jobs report rally. The bad news is that six weeks is about the limit, the rally isn’t earnings-based, and stocks are well overbought technically, so expect a bit of heavy weather ahead.
When I look at third quarter earnings, I see a current blended (actual results plus estimated ones) estimate from FactSet of a 2.2% decline for the S&P 500, with only 56 companies left to report. Allowing for the usual charade of earnings beats, that strongly suggests a loss of at least 1%, more likely north of 1.5%, leading to the first stretch of two consecutive quarterly decline since 2009. That’s a little misleading, as the declines this time around are quite mild (they were double-digit back then), but nevertheless it looks like 2015 overall will be in the red. The fourth-quarter estimate is already down to (-3.7%), suggesting that the present quarter will be much like the third. Declining earnings are no tailwind for stocks that are already at high multiples, and not a green light for a cycle that last peaked eight years ago.
Much of the decline is blamed on the dollar and energy, but the two are intertwined. The dollar isn’t going to drop near-term unless the outlook for the U.S. economy gets significantly (and obviously) worse, and without a falling dollar it will be tough for oil prices to move higher. Expect me to keep hammering away at this theme in the near term, along with the truth that the global macroeconomic environment is weakening.
You may wonder at this – how can the global economy be weakening in light of the recent strength in the stock market? Look at that October jobs report! Strength in the former has been buoyed primarily by a chart rebound that could not have taken place without a lot of dovish remarks from central banks around the world, in particular (in this instance) from the European Central Bank (ECB) – and a lot of short covering. There’s a reason that banks are dovish. Now that everyone seems to be convinced that stocks are fine again, there is fresh room for them to fall and I expect that prices will remain under some pressure until the end-of-the-year rally run that begins sometime around Thanksgiving.
As for the jobs report, I cover it at some length below. In twenty-five words or less, don’t jump to any new conclusions. Employment is a lagging indicator of economic activity, and the October report is 1) one month only; 2) possibly a bit skewed from attempts to eliminate underestimates from recent years; and 3) the longer trend still suggests that employment growth is waning and has been doing so since the first quarter. An eccentric-looking pop would look a lot better in the early stages of a recovery – say, in 2010 – than it does six years into a recovery.
The Economic Beat
The jobs report was the event of the week, along with being the surprise of the week as well. The headline total of 271,000 was far above the consensus of about 185,000 and exceeded even the most optimistic estimates.
Jobs reports are often contentious affairs, with many pundits wanting the report to read one way or another to validate certain points of views. The latest one was no exception. That said, I feel obligated to present some doubts about whether the month was really as good as advertised, and would add the admonition that the monthly data can be quite lumpy, with the better picture usually a three-month average that only obtains a couple of revisions later. Not much for real time decision-making, perhaps, but more realistic.
Two data points caught my eye right away. The first was that although the estimate (and by the way, all the numbers are estimates, present, past and future – there is never an actual count) of the unadjusted number of jobs went down in August and September – by 7K in August and 40K in September, for a net loss of 47K – the number of seasonally adjusted (SA) jobs (the number that gets reported) went up. That net loss of 7K in August was translated to a net gain of 17K when seasonally adjusted, and the September decline of 40K was reported out as a loss of only 5K when seasonally adjusted.
Estimating the labor force is never a fixed recipe – the dynamics of labor evolve over time. Data samples from 1985 and data samples from 2015 will lead to different places because the economies are different. The Bureau of Labor Statistics (BLS) is constantly striving to improve its current adjustment formula, as they should, and I have no axe to grind in that respect. The BLS would be the first to admit the limitations of what they do, but the numbers tend to get bandied about as gospel. Here, the upward adjustments to downwardly-revised estimates invite further scrutiny.
The second item that caught my eye were the percentage increases in the payrolls estimate (the one everyone talks about). Year-on-year, October stands at a provisional +1.94%, about the same as September’s +1.91% – which had also been 1.94% before revision. Year-on-year growth was well over 2% in the first half of the year, but has been steadily falling in recent months. In addition, the rate of growth in the year-to-date payroll estimate has fallen to +1.59%, about where it was two years ago and down quite a lot from last October (+1.93%). That suggests numbers that are not quite as strong as the 271K would imply, and so I went back to look at this time of year in 2014 and 2013.
October 2013 was originally reported as showing a gain of 204,000. Two months later the total stood at 200K, but November had been revised sharply upward, from 203K to 241K, and then again from 241K to 274K in January – a net change of 71,000! Do you think that the BLS wanted to avoid the same mistake in 2014? That year, the bureau reported an original estimate of 214K in October. That month was revised all the way up to 261K by December. November was originally reported as having a gain of 321K, but by January had leapt to 423K! (As an aside, do you think that maybe the Fed should have acted then?)
Clearly the Bureau would like to avoid having to issue such jumbo revisions. My suspicion is that the raw data for October looks no better than it did a year ago, but the seasonal adjustment factors are upgrading the totals to something much larger for the initial estimate. After all, the rate of jobless claims is certainly better than it was a year ago. That’s not to say that October wasn’t a good month for hiring, but that the change from September might be misleading when it comes to any real sea change in the labor markets. The year-to-date trend points to slowing rates of growth, and there is a strong possibility that October 2015 was no different in its original sample than 2013. If so, there will be no big later revisions upward, as the BLS may have headed them off.
The composition of hiring was also clearly seasonally biased. Most of the hiring came from entry-level jobs – 44K in retail trade, 41K in leisure-hospitality, 57K in health care and social assistance. At least manufacturing arrested its run of losses with an estimate of no change for October.
There were some other good bits. For once, the not-in-labor-force category did not show another outsized gain, in fact reporting shrinkage of nearly 100K. The household survey (very volatile data) also reported a good gain of 320K, the unemployment rate fell to 5% and the alternative U-6 rate dropped to 9.8% (SA) – it hasn’t been below 10% since 2008. The change in the rate was all in the teenage and less-than-high-school-diploma categories, suggesting tightness in the labor force (but also that other categories may be full up). Hourly earnings rose by 0.4% and 2.5% year-on-year, though weekly earnings remain at 2.2%. Business services also had a healthy reported gain.
In sum, employment looks fine, but I would again remind that it is a lagging indicator. I don’t expect any dramatic changes in hiring this quarter, despite my prognosis that the cycle has nearly ended, but there are a couple bits of evidence that the quarter is near the end of the full-employment road. It’s still too early to tell. The first quarter may look quite different, but I have my doubts about whether the BLS can catch it right away.
Elsewhere, the two ISM purchasing manager surveys reported widely different results. The manufacturing survey showed no change with a reading of 50.1, but also produced a danger reading of seven sectors growing versus nine contracting (two unchanged). The better non-manufacturing survey reported a strong reading of 59.1, up from 56.9 in September with a growth-contraction score of 14-1. However, its prices component remained slightly below neutral at 49.1, and prices are the best leading indicator for non-manufacturing.
Neither of the two surveys are highly reliable leading indicators, in particular the non-manufacturing or services survey, which can put out good numbers right up to a recession (i.e., 59.7 in November 2000). The manufacturing sector is definitely more sensitive, but historically not a reliable predictor of either recession or the stock market. It’s suggestive more than conclusive, though given the weakness in many other categories, it can be viewed as supportive evidence that the cycle is winding down. Factory orders were reported to be down 1% in October and are down 7.2% year to date. International trade showed continued weakness with September imports, and indeed the OECD further downgraded its outlook for this year and next.
Next week’s highlight is the retail sales report for October, released on Friday. As we near Christmas, I expect retail sales growth to pick up somewhat in the short-term – after all, we are still at full employment and holiday shopping hasn’t vanished. Thus, even though the intermediate-to-long term sales trend may not see any change, we are likely to see a couple of better months this quarter. It’s a quiet week otherwise in terms of headline data, one that shows import-export price indices on Tuesday, producer price changes on Friday, wholesale (Tuesday) and business (Friday) trade sales and inventories, and the JOLTS report (labor turnover) on Thursday. In light of the jobs report, one would expect the last item to be another decent number, which the stock market may not like, seeing it as further enhancing chances for a December rate increase.