Well, at least market performance was relatively equal across sectors this week, with no need for anybody to feel left out. Most equity markets dropped by 2% to 3% for the week, commodities were down about 3%, and bonds gained as the yield fell on the U.S. 10-year bond from 1.93% to 1.84%.
Whatever the headlines may say, we believe this type of performance suggests fears of slowing economic growth around the world. Fixation on daily oil price movement continues, too. At least when markets were focused on the Fed, the volatility was limited to a few days or a few months of the year. Almost daily announcements of one oil metric or another, as well as world markets for energy have kept markets gyrating from day to day. And some of this movement set off other program trades (rules such as sell stock when oil falls 5%) in other markets, and then momentum kicks in. It's a self-perpetuating process that has been painful to watch. First, predicting exactly how and when computer programs kick in is way beyond our capabilities. Second, for now, the market has lost its grounding in fundamentals and seems to care only about daily oil prices. Some posit that oil demand is sending implicit signals about economic growth. Oil traders may be good at many things, but forecasting economic growth is not one of them.
Economic data was a mixed bag, as usual. Some China purchasing manager data was down while other reports were up; however, none of them was ebullient. U.S. auto sales were greatly improved in January after December data was skewered by the statisticians, who have apparently misplaced their seasonal adjustment calculators. Reported sales have now moved from 18.1 million units in November to 17.2 million units in December, before rebounding to 17.54 million units in January. The bounce came despite a major snowstorm in the Northeast that closed some dealerships for days during January. Plus, consumers didn't seem so afraid of the January market sell-off, based on relatively strong auto sales, a key concern on investors' minds.
Headline job growth on Friday also came in below expectations at 151,000. A falling unemployment rate, faster hourly wage growth, and longer hours offset some of the sting, though. Year-over-year employment remains very stable at around 2%. The manufacturing sector even stabilized some, with a higher U.S. PMI reading for January from the ISM (and manufacturing industrial production has been moving up since July). With wages and employment doing so well and manufacturing stabilizing, it's very hard to envision a U.S. recession.
We think the underlying U.S. and maybe European economies are stronger than many believe. It doesn't seem that investors are taking a nuanced look at recent economic data, which has become increasingly difficult to analyze. Headlines haven't necessarily looked great, but stripping out the seasonal factors and weather, the economy seems to be acting quite well. We continue to believe a lot of the poor December headline data will be reversed, perhaps as early as the releases of January data (auto data has already reversed some of the problems). However, higher wages and currency issues might continue to weigh on corporate earnings and stock market performance.
Employment report contains unmistakable signs of strength despite weak headline jobs number
On the surface, headline job growth of 151,000 was a disappointment relative to our estimate of 200,000 jobs and the consensus estimate of 185,000. The number was also lower than the 12-month average job growth rate of 222,000. We warned in our column last week that this report will contain a bunch of oddball factors, and we weren't a bit disappointed on that front. Less-favourable weather conditions, annual restatements, and known problematic categories, including volatile courier numbers, were all issues. We just didn't know the magnitude. We will walk through some of the intricacies of the report later.
More important than the headline number were the falling unemployment rate, much higher-than-expected wage growth, and even upward movement in the number of hours worked per employee. Hardly signs of a softening economy.
Year-over-year employment growth holds steady at just under 2%
Our more reliable, year-over-year data also more than held its own, despite some unusually strong growth in employment in the waning months of 2014 and early 2015 (though some of the extreme months were reduced in the revised data set released today). So despite all the to-and-fro data in the monthly reports, the year-over-year employment growth rate of just about 2% hasn't changed to any degree of statistical significance for the past four years.
Employment growth unusually stable, though slow
The same pattern holds for the faster-growing private sector data. This data is not held back by slow growth in government employment. We also stepped back to show that such long periods of stability in job growth are unusual. Though the growth rates are lower than in the past, we note that slower population growth and greater baby boomer retirements are holding back further expansion. These demographic factors will continue to limit employment growth, in our opinion. We might add that we would gladly trade a few growth points for greater stability.
Perhaps one of the more exciting portions of the employment report was a nice jump in the hourly wage rate from month to month. However, we caution this rate is subject to a lot of fits and starts, with good months often followed by months of no increases. As usual, the year-over-year data provides a clearer picture and strips out some of the noise. The data shows a clear and unmistakable upward trend in hourly wages. The second data column below shows that year-over-year averaged hourly wage growth moved from 2.0% to 2.5% over the course of the year. That is particularly impressive when there was hardly any inflation. For a change, workers pocketed most of the growing wage increases. Given ongoing labour scarcity issues, we suspect that wage growth will remain elevated. Great news for workers--not so much for employers, at least the labour-intensive ones.
The wage growth wasn't quite equally distributed, with some industries doing better than others, as shown in the chart below.
The top few categories are typically well-paying jobs already. However, the leisure and hospitality and retail sectors showed above-average growth (average is 2.5%), too.
Monthly quirks make it tough to make conclusion on January
We expected this to be an odd month, but we never guessed so many categories would be messed up at the same time. Our overall view is that the positive oddities largely offset the negative. However, the strange monthly data points to the importance of using the year-over-year data we discussed earlier. That data remains unequivocally positive. Below is a laundry list of some monthly issues in January.
- Construction jobs swung from plus-48,000 new jobs in weather-aided December to just 18,000 in January.
- Manufacturing, in a difficult environment, added 29,000 jobs. This may prove difficult to sustain.
- Holiday couriers swung from a plus-10,000 to a minus 15,000
- Retail "added" an amazing 59,000 jobs, as fewer jobs added in late fall for the holidays meant fewer layoffs in January
- Temp jobs remained on a roller coaster, adding 25,000 jobs in December and losing 25,000 jobs in January. The category had a bad year as more employers are using more full-time employees instead.
- Government managed to lose jobs again, this time about 7,000. We thought the worst was over in this category.
- Layoffs at a number of large, for-profit schools hit at the same time, costing the economy 39,000 jobs, basically the full amount of the employment downside surprise. This industry has been under a lot of government pressure, and a stronger economy hurts this sector.
There weren't too many categories without some type of anomaly. Nevertheless, growth in healthcare, finance, information services, professional and business services (excluding temp workers), and leisure and hospitality were all consistent, if not improved from recent trends.
Utility, auto and gas data weigh on consumption report
Consumption data represents almost 70% of U.S. GDP, so it is one of the most important economic reports we receive every month. It also is one of the few monthly reports that adjusts for the variable effect of deflation and inflation on different goods categories.
Unfortunately, it is also one of the last economic reports we receive, so the data tends to be a bit stale. Still, more timely reports, such as retail sales, aren't adjusted for inflation, which has become truly problematic. Seasonal factors to adjust both retail sales and consumption are being confounded by changing consumer purchasing patterns and weather. All that said, we will try to explain why we believe the consumer is alive and well, seeming almost oblivious to headline news concerning China.
First, though we really hate to analyze month-to-month data because of normal volatility, we will start there to avoid the appearance of avoiding a piece of data because it doesn't fit our thesis. The headline monthly consumption growth rate was 0.1% for the month of December versus November. To set some context, the average monthly growth rate for the past 12 months was 0.2% within a range of 0.0%-0.6%, as shown below:
In that context, the 0.1% December growth rate looks a little soft, but certainly no disaster. Most consensus economic forecasts for 2016 have a monthly average consumption growth of 0.2% (or 2.4% or so annualized).
However, the December consumption data doesn't really tell the whole story without some adjustments. First, our old nemesis motor vehicles took 0.12% off the growth rate despite one of the very best, if not the best, December in history, before day count and seasonal adjustments. In fact, more automobiles were sold in December 2015 than any month of 2015. Yes, there were technically more selling days, but we are not so sure how many
people were racing to buy a car the Monday after the New Year, which was one of the extra days. As discussed later, the adjusted sales came bouncing back in January.
Utilities managed to take 0.07% off of the December report as unusually warm weather limited natural gas and utility usage. Weather also affected the sale of cold-weather gear. That caused the apparel category to take 0.02% off of the December consumption report. With colder weather in January, these two categories should be additive in the next report.
Grocery sales also took an unusual fall in December, taking another 0.03% off of the calculation. Most likely faulty seasonal factors and perhaps more meals out weighed on the data. However, some of the downside goofiness was offset by a huge help from the tiny luggage category and a boost from the "Star Wars" movie, which was big enough to be clearly seen in the data. When you roll it altogether, it looks like consumption growth would be in the 0.1%-0.2% range, pushing growth closer to normal. As we indicated, many of those same factors will turn additive in January.
Year-over-year consumption data now back at trend
Our normally reliable year-over-year data seems to have dropped off a bit, too. However, there are starting and ending point issues that have managed to confuse even that data. Last fall was unusually strong, as we have said many times before (at least in terms of employment, incomes, and consumption). Consumption growth got over 3% for a few months. Now, the December anomalies were enough to pull the full-year growth rates down some. However, the current 2.6% year-over-year consumption growth rate is still above the post-recession average of 2.2%. It seems likely to us that while year-over-year consumption might dip for one more month, consumption will be moving back up from its current 2.6% growth rate by the time of the February data release in late March or early April, when some of the more extraordinary 2014 data drops out of the analysis.
Income growth should support continued consumption growth
The data continues to show consumer spending undershooting total wage growth and the more comprehensive real disposable growth metric as shown below:
The gap between consumption growth and income growth has widened considerably since August. After being nearly identical, income growth of 3.4% exceeds consumption growth of 2.6% by almost 0.8%, suggesting fuel for future growth. That said, consumption has not been a disaster and is being hit hard now by the rough December consumption report. Again, we expect some of that damage to be reversed as early as January.
The myth of the battened-down consumer is just that, a myth
The media loves to talk about consumers not spending their gasoline savings on much of anything, or that they have even cut back. The data shows otherwise. Admittedly, we are cherry-picking a bit, but here are some of the fastest growth categories, fourth-quarter 2014 to fourth quarter 2015--hardly items for the conservative consumer.
So what is holding back consumption? A lot of the large consumer staples categories, again cherry-picked, are shown below.
Many of these are very large categories that suffer from the law of large numbers; it’s just hard to grow things this big. Housing, besides short-term weather-related issues, which are real and significant, is being held back by the slow expansion of the housing market. Each year about 1.1 million new homes are started on a base of 110 million homes, or just 1%. Since many of those replace existing homes, the natural growth rate of this category is even lower. Throw in some warm weather and energy conservation, and you have a serious threat to the consumption calculation.
Food varies a bit, but you can grow it only so much, and more of those meals are eaten out, again limiting growth in a key category. And unless you really want to spend your life in a car, miles driven--the driver of gasoline usage--can't really go up that much. There is at least one large consumption category that is doing considerably better: restaurant sales. Growth here, even adjusted for inflation, is up 4.0%. Even more telling, the alcohol portion is growing at 5.4%.
Auto sales data confirms December and November data were opposing, self-cancelling blips
Recently, auto data, both on a sales and production basis, have been a train wreck for statisticians. This fall the timing of summer shutdowns ruined the production data base (more limited shutdowns in the summer made that data look strong while the late fall data made it look like manufacturers were on strike). The sales data is being messed up by antiquated seasonal factors from an era when light truck sales were primarily to businesses and by selling-days calculations. A low sales day count in November plus seasonal adjustments made November look like a barn burner. December looked like it fell off a cliff because there were so many sales days in the month. Therefore, an outstanding December looked awful when adjusted for a really large number of alleged selling days, several of those just after New Year’s.
Fortunately sales came bouncing back in January as shown below and were well above year-ago levels. These improved sales came despite some snowy weather along the East Coast, which forced many dealerships to close for several days.
It was another great month for high-margin light-truck sales but not so good for conventional autos. Most of those vehicles are typically made in the U.S. because of tariff rules. Prices were generally higher, too, more because of mix than price hikes. Incentives still looked disciplined, too. All in, it certainly doesn't look like consumers are panicking.
Next week, just retail sales
While there are a number of smaller reports next week, the only major fireworks should be the retail sales report for January. As usual, the report may not mean much until we can see the inflation data to adjust those sales figures. Because of sharply falling gasoline prices, expectations are that growth may come in at 0.1%, following the not-so-great negative 0.1% in December. Depressed prices are affecting both sets of data. Some of the weekly figures did look a little better in December and a return to more normal winter weather in addition to some old-fashioned hoarding in the Mid-Atlantic states (in front of the snowstorm) make us believe that retail sales could do slightly better than expectations.
We will take a quick look at the Job Openings and Labor Turnover report next week to see if it supports our bullish assessment of the employment report. Openings have been over 5 million for some time, indicating that poor availability of qualified candidates is holding back employment growth.
