“All that glitters is not gold.” – William Shakespeare, The Merchant of Venice
So our main engine, the jobs report, misfires, and equities rally anyway. Shades of 2013? No, last week was more like 2007 or 2000. The headline jobs number was indeed a big miss, but the report itself was not as bad, and indeed not as bad as it could have been, a familiar Wall Street rallying cry. It also revived a familiar parlay – either the economy is really better than the headline, or if it isn’t, the Fed will be obliged to rescue us.
After the Dow’s worst day in years (Monday), the week’s big reversal rally began with Wednesday’s bounce from a Fibonacci line, beloved amongst technicians (the decline from the top was 50% of the move from the October low, and 50 is probably the Street’s favorite Fibonacci number), ending the week in rip-rebound mode, technical factors leading the way. I had indeed written that the market would try to rally last week, but it did so in reverse fashion: the ISM manufacturing survey that was supposed to be good wasn’t, leading to a brutal sell-off, while the jobs report that was supposed to be back to trend wasn’t either – and so of course the market rallied.
Much of Friday’s rally was based on a hopeful, if flawed calculus. On the one hand, private sector payrolls were up 142,000, compared to the headline number of 113,000 (the government sector was assigned a loss of 29,000). Traders had starting giving the number a weather-related haircut days in advance of the release, so add back another thirty or fifty thou from the weather and hey, you’ve got yourself a near-200K number, right?
Another fallback was that the increase in the household survey employment was 638,000, seasonally adjusted. So back to the parlay: Either the report is right in the right places, and the economy is fine and pay no mind to that weather, or the report is wrong in the wrong places, and the Fed must come to the rescue (with the unemployment rate now at 6.6%, no less). Don’t overlook the desire to ride a rebound from the bounce; that it was a rebound rally can be seen in the movement of the ten-year bond and the price of gold – if the employment report really was alright, you wouldn’t have seen those two moving up at the same time.
Economists were fairly unanimous in their immediate takedown of the weather excuse – not easy for traders who had bought into the opposite theory for days – but then, isn’t Janet Yellen going to testify next week? Maybe the Fed will “taper the taper,” good for equities. That notion didn’t get any support either with Fed officials and observers, but they were done speaking by mid-morning and who listens to those pointy-heads anyway. A steady parade of bulls proclaiming time to buy filled the airwaves the rest of the day.
The market never did get oversold on an intermediate or long-term basis – there is a long way to go for that, and a potential trip to the 1700 region, where both the 200-day and moving 12-month average lurk, is still waiting down the road. But first should come a further attempt to retake the highs, as I have been writing in recent weeks. Indeed, the violence of the rebound suggests that momentum trading will want to look for some triumphant technical plays – first the 50-day average on the S&P (just above Friday’s close), followed by the January high.
The calendar does favor the bulls for the next few days. Whenever it’s as light as it is in the first half of next week, that favors the prevailing trend, so prices could keep on rolling in the absence of some fresh fright from overseas. The next major report due up is the January retail sales report on Thursday – it isn’t going to be pretty, but the hidden beauty of it is that it comes with very low expectations (consensus is for a decline of 0.1%).
Despite all this, the early evidence indicates to me that without some major economic developments to the upside, a bear market may be imminent. But bear markets rarely begin with a bang – it’s usually a choppy affair in the early going, with lots of counter-rallies, often punctuated with defiance moves back to the previous high-water mark. The really serious damage usually doesn’t begin until much later – three months, six months, twelve months – when the capitulation trade finally begins in earnest.
Sentiment has clearly been damaged – indeed, one of the most frequent remarks heard amid Friday morning’s rally was, “who’s doing the buying?” The only way that the Fed is going to pull back from the taper any time soon is if markets quickly burn off twenty percent or more. There is no firm evidence either for the annual lure of the accelerating economy, reminiscent of Charlie Brown and Lucy with her football (“this time I really, really mean it”). Yet there will be standard-bearers ready to raise the flags and charge at every whiff of a weather rebound or a gentle remark from central banks, so don’t expect that narrative to die just yet. Even so, let Friday’s rally in gold and the ten-year bond be a warning sign.
The Economic Beat
The debate over the weather effect will go on for some time, but the report may have actually benefited from the weather, a possibility many suggested in the days leading up to the report. The measurement period that comprises the 12th day of the month came during the one mild week of January, and probably nowhere was that more apparent than in the estimate for construction jobs.
I heard people talk all day long about the addition of 48,000 jobs in construction, usually with a befuddled look that expressed less-than-total understanding about how construction hiring could pick up so much in a month not conducive to the making of anything not based on snow.
The answer is that we didn’t add 48,000 construction jobs, of course; in fact we didn’t add any jobs at all in January. We never do. The end of the year is a major separation moment in the employment world, such that the number of counted jobs in the establishment survey falls by about 2% every year from December to January. Whether people are missed in transition, retire, or have to find another job, it usually takes until April for the actual number of workers to recover to December levels. It all makes for a very good argument for seasonal adjustments.
So it would be more accurate to say that, “the initial estimate of the latest annualized run rate for construction employment count rose by 48,000 from December,” and the reason for that is that the decline from December to January was less than usual. Why? Because weather and other factors suppressed construction employment towards the end of the year, so there were naturally fewer people to lay off: The payroll decline from August to December in construction was 49,000 higher (unadjusted) than it was a year ago. Remarkably similar to that January increase, isn’t it? Fewer workers leads to fewer layoffs, presto-changeo, the headline says 48,000 new jobs. I don’t think so. A second factor is that the balmy weather for the measurement week would have meant some work being squeezed in between bouts of tough weather, also inflating the sample with people who may very well not have worked the rest of the month.
But construction is cyclical, don’t you know, so maybe things are really better than they seem. That’s what the stock market – which as a rule loves good news in construction – would say. This 48K is just statistical noise, though, and most unlikely to be borne out later on by construction spending data. That didn’t stop a certain amount of chest-beating over the increase in the goods production category – nearly two-thirds of which came in construction. Something similar happened a year ago with the retail sector – pre-Christmas hiring was below average, leading to a smaller number of post-holiday layoffs, a big increase in the seasonally adjusted January 2013 total for retail workers, and a month or so of buzz about strength in retail that never really happened.
I’m afraid that the big increase in household employment – 638K – wasn’t from any new trend either. The establishment and household surveys are supposed to converge over time, and they do. However, at the end of 2013 there was quite a disparity between the two: The establishment survey estimated an increase of 2.19 million jobs for the year (pre-revision), nearly 50% higher than the household increase of 1.46 million. The revised establishment number for 2013 is now at 2.26 million, so the household figure had some catching up to do.
A secret acceleration in hiring not evident in either the establishment survey or weekly claims data is much less likely than a simple catching up by the smaller, more volatile survey. The year-on-year January increase for the establishment survey rose to 2.32 million, while the household survey moved up to 1.91 million, narrowing the discrepancy considerably. Of course, there was no actual increase, just a smaller decline, probably because – similar to construction – the household count declined from August to the end of the year, instead of rising as it had the prior two years. Fewer workers mean fewer year-end losses. The U-6 underemployment rate fell to 12.7%, which cheered traders quite a bit, but that was a seasonal adjustment as well – the actual rate rose half a percent to 13.5%. That is less than usual for January, which I also attribute to the catch-up effect. By comparison, the January 2008 rate (unadjusted) was 9.9%.
The initial conclusion is that the labor market dynamic isn’t really changing at all. The pre-revision jobs increase for 2013 was 1.64%, and now stands at 1.69%. 2012 was 1.72%. The year-on-year December increase was at 1.63%, and now stands at 1.71% – you can forget about that 75K revised December number, it’s an artifact of warped seasonal adjustments (combine November and December and average them out for a better picture). The changes in some of the household numbers – unemployment rate, participation, working population – are mostly month-to-month noise. January establishment payrolls shrunk by 2.08% from December this year, compared to the 1981-2014 average of 2.06%, or the average of 2.1% for the last five years.
In fact, I would say that the headline numbers from the last two month – 75K, 113K – look set to create some real mischief. By making employment appear weaker than it really is, they’re setting the stage for a goofball rally when the headline numbers appear more normal again. It’s going to briefly revive the song of an accelerating economy in 2014, with what is likely the real message of the data getting lost – namely, that the underlying trend of the economy hasn’t shifted gears at all. The disappointment that ensues when the market realizes that we’re in for another year of 2% or less GDP, this time with a Fed halfway through its withdrawal – will come later. It rates to be keen.
The ISM manufacturing survey that precipitated Monday’s short-term selling climax came in at 51.3, close to three percentage points below consensus. The survey included the largest drop in new orders in four years. The results were at odds with the regional surveys that had led up to it, leading to speculation that the timing of the surveys – just before the bad weather, right after it – might have been responsible for the divergence. Certainly responder comments made it clear that weather had played a large role. The Markit Economics PMI number (53.7) didn’t show the same drop-off.
The report was accompanied by other bad news that day – weakness in auto sales, soft construction spending (and I would look for the latter to get a negative December revision too, much like its November sister).
Tuesday brought some relief with factory orders for December – they fell 1.5%, but that was not as much as expected (-1.8%). Even so, the underlying trend in business cap-ex spending continues to soften.
Wednesday helped the market start to rally again, first with an ADP employment number of 175,000 that was only slightly short of estimates, then with an ISM services PMI that at 54, slightly topped consensus for 53.9. The ADP number also revised December’s original estimate upward to 227,000, which may have encouraged dismissal of the BLS 113,000 number two days later. Thursday followed with same-store sales weakness, an increase in planned layoffs and a slightly wider trade deficit in December.
As noted earlier, next week starts off slowly, though I will pay serious attention to the wholesale sales and inventory data set for release on Tuesday. The report of the week is retail sales on Thursday, and Friday will feature December industrial production, along with the price indices for imports and exports. We’ll also get some production and inflation data during the week from the eurozone and China.