“Turn off your mind, relax and float down stream.” – Lennon-McCartney
Over the last weekend, the battle lines became clear – go with the somewhat overstated July jobs report (a gain of 255,000, seasonally adjusted, or SA) as the better measure of the economy, or the anemic GDP number (+1.2% SA)?
Much of Wall Street, you will not be surprised to hear, prefers the employment data. Indeed, voices were heard proclaiming it the best period of the cycle (though July 2016 was below July 2015) and many others confidently predicted that with growing employment, spending would improve and the cycle would be good for another two years at least. Of course if that were true, there would never be a business cycle – contractions don’t start from bottoms, they start from tops, but try telling that to the Street.
So the stock market put on its obligatory rally, fueled by the double-barreled surprise, beginning with the ADP report two days earlier that estimated a smaller gain of 179K (it doesn’t count government jobs), compared with the private-sector component of the Bureau of Labor Statistics (BLS) report (217K). The other part of the surprise is that for all the bravado that the bull market will never end, the stock market is holding its breath every month for a fatal signal from the jobs report. Better-than-expected is the mantra of the market.
The BLS number may in time approach the ADP number. There was some misguided talk in the wake of the report about the seasonal adjustment factors that allowed the issue to be dismissed with a smirk when it shouldn’t have been. Critics chose to focus on the fact that the actual number of counted jobs declines in July, compared to the announced gain, and were quickly dismissed with the retort that it’s an annual process. Indeed it is, just as it is every January (though the latter is on a much larger scale). That was the wrong battle to pick.
Where the arguments should have focused is on whether the adjustments will hold up over time, and why May and June were revised upward on a seasonally adjusted basis even as the unadjusted numbers were revised downward. See the Economic Beat below for a fuller discussion, but it looks to me more like attempts to fit the data to the regression line than anything else. Big conclusions from any one month of payroll data are almost always a bad idea with payroll data, but you need to be especially careful with January and July. It really takes three to six months after the turn of each half-year for the BLS to catch up to a good estimate of what the current job totals might be.
At any rate, the GDP data is more consistent with all of the other data on spending and investment. For example, four-quarter nominal GDP (that is, unadjusted for inflation) stood at 2.4% through the end of the second quarter, roughly in line with the 2.67% growth rate in retail sales (also not adjusted for inflation) over the same period. Business investment was negative over the same period. On Wall Street, such nuances are ignored – what most talk about was that the estimated June sales bump of 0.6% (SA) beat expectations! Hallelujah, and never mind that most of the beat came from a downward revision to May.
Heading into the rest of the month, the stock market is overbought with one real report left on the calendar – July retail sales, due this Friday. After that we enter the financial markets’ vacation doldrums, followed by the silly season (the last week of August and the first week of September). I don’t know if prices can hang on through all of this. Complacency is extremely high, leaving us ripe for a big pullback, but whether a catalyst will appear (basically, any unexpected piece of bad news) is anyone’s guess. Earnings keep declining, the GDP run rate keeps declining, but so what? These prices are running on opiates. That doesn’t mean that they can’t grind still higher before the eventual hangover. Caveat emptor, because the risk-reward ratio is not on your side these days.
The Economic Beat
The jobs report dominated the week’s news as well as the weekend, but it wasn’t as special as many are making it out to be. The larger impact of the report was its surprise value and the somewhat dubious-looking upward revisions to the SA counts of the prior two months.
Let’s go to the unadjusted data. The preliminary July estimate is for a sound increase, that appears certain. Actual payroll counts don’t really increase in January and July, however, they decrease, and then seasonal adjustments follow. It’s the right thing to do, but that can translate into several revisions for the months in question, or more commonly (it says here) a delayed effect on latter months down the road as the BLS statisticians wrestle with their smoothed regression lines. So the July unadjusted decrease at (-0.71%) was smaller than the recent historical average (-0.91% over the last 35 years), but it was not as good as last year (-0.66%) and not far either from the year before (-0.77%) – in fact, the last three years look to be about the same trend, don’t they? Whether that really translates into actual hires isn’t known yet, but we can say that twelve-month payroll growth through the end of July is a provisional 1.7%, down sharply from the roughly 2% that prevailed over the last two years.
Something that makes me suspicious is why both May and June were revised downward in the unadjusted counts, but upward in the seasonally adjusted results. I’m not quite ready to buy into those numbers yet, as they may just be efforts to push everything back to the regression line. The weekly claims data is getting weaker too – pundits like to talk about the absolute level being so low, but the year-year comparisons are flattening out more and more often: the last three weeks of July were slightly higher than the year-ago period. So my conclusion is that one should not quite jump on the employment-is-better-than-GDP bandwagon yet, because it isn’t borne out in any of the rest of the major barometers – retail sales, wholesale sales, inventories, and business investment, to name a few.
Much was made out of an increase in hourly earnings, but that can also be a mix issue – I like to look at the year-year comparison in weekly earnings, and that remains at a subdued 2.3%, in line with the year-year increase in personal income (2.2%). Average weekly hours are down a tenth from last July, probably due to the auto business. The unemployment rate was steady month-to-month at 4.9%, though the alternative U-6 rate edged up to 9.7% from 9.6%. The participation rate edged up to 62.8%, but the number of discouraged and marginal workers also increased (seasonally adjusted). Finally, through the end of both May and June, the year-to-date data had looked distinctly weaker than recent years and was comparable to 2007. July’s initial estimate improves those totals, but I wonder it it’s sustainable. At any rate, I’m not going to hang a lot on one month quite yet.
The other major headline reports of the week were the national purchasing manager (“ISM”) surveys for the manufacturing and service sectors. Manufacturing eased from 53.2 to 52.6 (50 is no change), while services were similarly off a bit, from 56.0 to 55.5. Neither change has much significance, and neither report contained exceptional information. The main leading indicator for the non-manufacturing or service sector is prices, not new orders, and their growth rate did appear to slow significantly, from 55.5 in June to 51.9 in July. The new orders rate in manufacturing was unchanged at a very good 56.9, but the factory orders number was (-1.5%) for June and the inventory-to-sales ratio is still too high, at 1.35. Business cap-ex orders continued their string of year-on-year declines (-4%).
Construction spending appears to be slowing, though these numbers are subject to heavy revision. June currently stands at a maybe-decline of 0.6% (SA), as the year-on-year rate eased to only 0.3%. I would look for both of those numbers to come up (and if they don’t, we should be worried).
Personal income and spending for June was unsensational, with income a bit less than expected at 0.2% (SAAR) and spending a bit higher at 0.4%. The year-year spending rate move up smartly to 2.8%, which next to the 2.2% rate for income, suggests consumers are using more credit.
The main report of interest for next week is July retail sales, which are not strong going by the weekly data. Consensus is for a gain of (0.4%). Despite the apparent “surprise” in the June report, the year-on-year data from the Commerce Department showed feeble, below-average growth.
Wholesale sales and inventories are Tuesday, followed by business inventories on Friday. Wednesday has the latest labor turnover survey (JOLTS), while Thursday brings import-export price data and Friday producer prices.