Summary: Today we examine the very beating heart of the US economy, where all the arteries converge — the jobs report. The picture it shows is clear, although most commentary obscures this beneath analysts’ hopes and fears. Also, the data warns that the US economy has changed in an important but ominous way.
“Unless you expect the unexpected you will never find truth, for it is hard to discover and hard to attain.”
— Heraclitus, the pre-Socratic “Weeping Philosopher” of Ionia
The bottom line
Here’s one version of the bottom line (at the end is a more accurate version) of the numbers in the January employment report: job growth continues at the same rate as during the past ten months. Not slowing, as some feared. No breakout, as some hoped.
The following graphs show the numbers on on a Year-over-Year basis, smoothing the lines but losing resolution of recent change. The numbers are too noisy and too revised to see the pattern otherwise.
Is this rate fast or slow? The monthly numbers don’t say, especially since population growth means that the same number of new jobs represents slower growth over time. Let’s look at job growth as a percent change. We see a slow — very slow — acceleration, perhaps even the start of a break-out: 2.3% YoY (2.1% YoY using the Household survey data, a second survey confirming the results). That’s fast, the fastest since May 2000 — at the peak of the tech bubble.
Here’s good news: the number of jobs is growing faster than the working age population!
Some not so good news
Wage growth is slowing. Profits high and growing plus a tightening labor market should produce rising wages for the average worker. It’s not happening, yet at least — although for several years we’ve been told that it’s going to happen real soon.
This is an important change, perhaps one of the most important structural changes of the past few decades. A strong recovery should bring strong wage growth. Corporations have learned how to break this relationship, what economists thought was a “law”. They’ve broken the unions, developed tacit (or even explicit) agreements to suppress wages, and many other means to shift money from wages to profits. See this example of these methods: firing your American workers, replacing them with cheap young foreign workers imported under H1-B visas. They’ve grown so powerful they no longer need to disguise these illegal actions.
Update: for a brief but powerful analysis showing why wage growth should be higher at this point in the cycle see Benn Zipperer at Equitable Growth: “U.S. wage growth remains sluggish despite employment gains.” Especially note the graph.
Where are the jobs?
The BLS Highlights gives the answer. Perhaps in the information sector (often described as our savior from automation)? Nope, the new jobs have all gone to machines.
What about mining? We hear doomsters talking about unemployment from the oil bust. Not a big deal; it’s a dot on the overall economy, job-wise.
Where is the job growth? Every month it’s the same: business services, retail, education & health, leisure & hospitality. As many have said, most of these are not the sectors we want the fastest growth in.
I’m surprised by the continued growth in retail. The shift to online plus slow growth in wages should (guessing) give this sector at best slow growth. We love to shop, even if we can’t afford to.
How do we make sense of this numbers blizzard?
To help us the Fed produces a Labor Market Conditions Index, using a model fed by 19 indicators. It paints a picture of job conditions consistent with the slow growth of most economic indicators since Summer 2012. Like most indicators it shows no acceleration. No signs of the coming boom predicted by so many during the past 3 years.
It’s a “bottom line” consistent with the simpler version given at the start of this post, so we can have confidence in its accuracy. Data as of December 2014.
For More Information
(a) See all posts about economics. Especially see this: Dreams of a boom fade & attention turns to secular stagnation.
(b) Recent posts about this economic cycle:
- Are we following Japan into an era of slow growth, even stagnation?, 18 November 2013
- Has the Fed blown another housing bubble?, 30 January 2014
- Has America’s economy entered the “coffin corner”?, 10 July 2014
- Economists forecast a boom soon. The numbers show slowing. Who is right?, 21 July 2014
- See the true trend of the US economy, hidden in the daily news, 1 August 2014
- It’s not too soon to worry about the US economy. There are things worse than slow growth., 18 September 2014
- Listen to the slowing US economy, hear echoes of Japan, 24 September 2014
- 3 graphs tell the story about the US economy, hidden amidst the noise of the jobs report, 6 October 2014
- Look at the economy. Fight the illusion of normality. Feel the weirdness., 8 October 2014
28 thoughts on “The jobs report shows the beating heart of the economy. See the good news, & the bad.”
Bloomberg calls it a Sea Change in today’s headline.
I think there really are a lot of people who should start a Blog and feature a big Tip Jar.
Best hope for income growth.
The so frequent “sea change” headlines result from people looking at the noisy monthly data. Except for specialists that is imo a waste of time. Endless sequence of We’re doomed1 and Pop the champaign! Clickbait to fill the space between advertisements.
As I have said for 3 years — in hindsight, correctly — the US is growing at a slow steady rate. Few signs of acceleration. The risk at these levels is that a slowdown will quickly tip us into a recession. Most commentators, even economists, do not understand this. Recessions do not quickly happen (except following large shocks), but that’s because “normal” growth is half-again faster than we’ve had since the crash. We’re flying low, so any dip risks hitting a tree.
Good point(s)….low flying risk. Oops….. Really wouldn’t take much to erase the smiles and then what would we read and see. More cheerleading? Probably, as it’s called “click bait”. And more lethargy from most. Yup
Remember that coins have two sides. We could slow, which probably would mean a recession, we might equally likely accelerate from here — as the top-line job numbers hint might be happening.
For lethargy see the US public’s waning interest in our wars, in our police, in the govt’s surveillance of us. The Deep State has exhausted us, and now begins the pursuit phase.
I thought it was understood that unemployment and job figures were extremely skewed due to all the people that have stopped looking for a job, people who’s unemployment benefits have run out and people who have been forced to work part time jobs because of increasing health insurance costs for employers. I’v read that the true unemployment figure is as high as 24%.
As I understand it, the rate is determined based on the results of something like 15,000 surveys completed by households and companies. That is an extremely small sample for saying with any confidence that you can determine what unemployment is in a country of…well…I’m not even sure what the US’s population is right now–but I have seen figures from 320 million to 340 million.
At a local level you can see huge amounts of lying over ‘unemployment’ (which is probably why the feds use a survey of households and companies anyway). Just take for example the county of PA I used to live in. They would shoot out huge headlines about how freaking low the unemployment was in the county, and how nothing needed to be done but more of the same. But the reality was that it was a bedroom community of 140,000 people, most of whom commuted OUTSIDE OF THE COUNTY to work. I could see this at the age of 16 and knew what crap government stats are, as I worked my fastfood $2.20/hour job (after Reagan had cut minimum wage to help all those untrained youths).
The liars are the people you rely on for information.
Most economic data comes in two or more releases of increasing accuracy. The final on the establishment survey is extremely accurate, based on tax records — not small surveys.
The other metric, the household survey, is relatively small — and so has larger error bands, and varies quite a bit from the establishment survey on a monthly basis. However over several quarters the two track well, and even more so over a 12 month period.
Saying that U.S. Unemployment is at levels similar to the Great Depression is nuts.
Thanks for mentioning this! I strongly recommend that you upgrade your information sources. Don’t pay attention to anyone who says such a thing. 24% unemployment is like that during the Great Depression. We’d know if that was true.
In fact there is no one true definition of “unemployment” (as with so many economic concepts, like “inflation”). The BLS provides accurate data allowing us to track six definitions of unemployment from narrow (U-1 at 2.7%) to broad (U-6 at 11.3%). The standard or headline definition is U-3. None is more correct than the others. No rational definition gives anything close to 24%.
A side note, perhaps not relevant. I’ve found that people who believe unemployment is sky-high also believe inflation is far higher than reported. It’s not.
On some politically oriented discussion lists, we used to argue about every such report and its significance. Here is what I remember from such discussions. First, the concept of ‘unemployment’ isn’t really what most people think it is. Two, the figure is determined based on statistical treatment of the results of two surveys, the two surveys being a very very very small sample of the population (a survey of households and a survey of companies). Third, labor force participation rates are very low and appear to be staying low. Fourth, the US hides a lot of unemployment of young people by counting them as students and military personnel. It even does this with older people by counting them as in ‘re-training’. Fifth, someone working a low-paying, part-time job who would take a full-time job if they could get it counts as employed. Sixth, some skeptics say if you want to know what the real unemployment rate is, always take an OECD country’s rate and multiply by 2–at the very least.
One additional thing I have noticed is that, for a country that prides itself on free markets, entrepreneurship, and capitalism, a large core of the full-time employed have GOVERNMENT JOBS. When I set out to get a firm figure on that though, it proved exceedingly difficult. I estimate that over 30 million Americans work in government full-time. That is not counting all those who rely on sales to/contracts with the government for their revenues.
Almost none of that is correct. Yes, there are two surveys — different methodologies which provide confirmation of trends. One surveys households; one surveys employers. Only the household survey is used for calculation of unemployment.
The establishment survey data is initially reported from a small sample, but refined as more data comes in. The final is done using tax data, and is as accurate as possible.
The rest is all quite wrong as well. I suggest you seriously upgrade your information sources, since the ones your trust are unreliable.
You haven’t really shown where unemployment ‘skeptics’ are relying on inaccurate information. Your thinking relies mostly on the US government’s concepts of unemployment, which many argue are more than simply flawed. Also, saying something is wrong and showing it is wrong are two different things. While I agree, perhaps some of us haven’t done a very good job showing where the government stats are bad or wrong, you haven’t either–you haven’t done a very good job showing where they are accurate. Saying one thing tracks another might simply mean they both track the same garbage. And saying an argument is wrong simply because you know it to be wrong is not really a convincing argument. Some of the unemployment skeptics said unemployment was 24% at the peak of the last recession (or the end of the free world as Hank Paulsen called it, until he got his bailout). I believe now they are reporting figures more like 16%-19% (except shadowstats still says 23%).
One thing is clear, whether you buy government stats on unemployment or not–labor force participation rates are well below the level they were their peak in 2000 or the levels of pre-recession, say around 2006. Also, the lack of growth in wages shows that labor can not make demands of capital on this front. Of course part of it is that no one is organized to do so. But the other side of it is that there is so many unemployed and under-employed around.
Finally, interesting that you should bring up inflation because that too is an area where the government needs to be challenged. You can see where during the past two decades Americans have been hurt by inflation in health care costs and in higher education. Meanwhile, a decade of artificially high oil prices combined with a weaker dollar for much of the period did impact their monthly budgets on even the most basic things, like filling up the car to go to the supermarket.
“Your thinking relies mostly on the US government’s concepts of unemployment, which many argue are more than simply flawed. ”
I lost interest here. Basic economic concepts are what they are. You can invent new ones, but using the standard names for your new concept is silly. Saying economists are deceitful for not using your definition is nuts.
All economic concepts are “flawed”. They are abstractions — simplifications of reality used to aid understanding and analysis. This is basic semantics. “The map is not the territory. The name is not the actual object.”
Is a nice review of issues surrounding secular stagnation by Steve Keen.
Excellent work collating a lot of detailed data. Since as FM points out this data is extremely noisy an the U.S. economic is enormously complex, the raw data often gets drastically revised up or down, so the important information involves smoothed long-term trends.
The long-term trends which hypesters like Bloomberg an the Wall Street Journal and Zero Hedge ignore are: nearly flat wage growth, very slow job growth, ever-increasing percentage of the overall population unemployed.
Some of this appears to be due to the boomer generation retiring. But a significant part of the slow job growth and the dropping labor participation rate seems due to automation.
For a good discussion including the federal reserve research report on reasons for the declining labor participation rate, see:
Intriguing discussion of automation and its effects, referencing a recent Scientific American article and an article from Foreign Affairs, here:
Summary: some of these long-term trends seem due to long-standing demographic developments in the U.S. population related to the baby boom retirement & women’s entry into & recent pullback from the workforce, while other of the economic trends FM discusses seem to be caused by more recent upheavals like the development of advanced robotics, AI, and so on.
“some of these long-term trends seem due to long-standing demographic developments in the U.S. population related to the baby boom retirement & women’s entry into & recent pullback from the workforce, while other of the economic trends FM discusses seem to be caused by more recent upheavals like the development of advanced robotics, AI, and so on.”
Yes, that’s the standard narrative. But economists have shown that the effect of demographics has been grossly exaggerated, and there is almost nil effect so far from “advanced robotics and AI” (it’s a big concern for the future, but not yet). The reason is obvious: American will go far to believe explanations that ignore political changes — and so excuse our apathy and neglect of our responsibility as citizens.
Also, when I said just take the unemployment stat that the media are reporting and double or triple it to get a more accurate figure…well, that is the fast way to get to U6. Really, does anyone believe that the unemployment rate of the USA is 5.6%? That is about as ridiculous as saying Great Depression levels. Sure, people shouldn’t just rely on their own experience of the economy, but when stats and what the media report are so far from actual experiences, skepticism prevails. I also refer back to my earlier example. Living in a county that bragged of its ultra-low unemployment rate, when the reality was retired federal workers and exurbanite commuters who traveled hours each way to get to their government jobs (in Harrisburg, in DC, and in the military bases of the region) OUTSIDE of the county.
I”m sorry that you cannot be helped. But as everybody over 18 knows, people who believe conspiracy theories are unreachable. Sad. You’re banned.
Well I’m fast losing interest in FabMax–you are being lazy and arrogant.
But for a different look at unemployment try Gallup. It’s something more like everyone can understand.
I don’t understand why you believe the Gallup poll supports your claims. It doesn’t.
(1) The Gallup survey is only half the size of the BLS Household Survey — 30,000 vs 60,000 (Gallup has nothing like the Establishment Survey, which provides the headline jobs number, and eventually the definitive employment numbers based on tax records). Odd that your so strongly criticized the BLS HH survey as too small, but cite one half the size.
Being so small, the number is far more noisy.
(2) You appear to be comparing the seasonally-adjusted U-3 number to the Gallup numbers. That’s wrong. I mentioned this in an earlier comment. Use the BLS numbers here:
(a) The Gallup definition of unemployment corresponds roughly to the BLS U-5 NSA number: “Total unemployed, plus discouraged workers, plus all other persons marginally attached to the labor force, as a percent of the civilian labor force plus all persons marginally attached to the labor force.”
Gallup: 7.1%; BLS 7.4%.
(b) The Gallup definition of underemployed corresponds roughly to the BLS U-6 NSA number: “U-6 Total unemployed, plus all persons marginally attached to the labor force, plus total employed part time for economic reasons, as a percent of the civilian labor force plus all persons marginally attached to the labor force.”
Gallup: 15.8%; BLS: 12.0%.
Differences result from the fact that they question different populations and ask different questions. Note that the government survey is not only 2x the size but has a MUCH higher response rate than Gallup’s. Lots of people — like me — refused to answer telephone polls — which might be the largest factor distorting Gallup’s results. People with jobs might not answer as often as people without jobs. Even small behavioral differences produce large swings in the results.
I changed your status to “moderated”. The comment about Gallup presented evidence, not a rant about government conspiracies. Questions and evidence, even wrongly interpreted, are always welcomed here.
“I’m surprised by the continued growth in retail.”
Anecdotally, I’ve heard of many people who are qualified to work in an advanced or specialized field but can’t get that job, so they settle for work in retail stores at minimum wage and with minimal benefits.
I think that might be the case for many Americans whose skills are undervalued or otherwise mismatched to the current job market. This could be the reserve army of labor, for whom retail is the employer of last resort, the employer by default, paying default wages and offering default benefits.
According to the BLS, “These four occupations [retail salespersons; cashiers; first-line supervisors of retail sales workers; sales representatives, wholesale and manufacturing, except technical and scientific products] are expected to account for two-thirds of jobs added by sales and related occupations over the 2012–2022 period. Retail salespersons is projected to add 434,700 jobs, the third most new jobs of any occupation.”
Maybe it’s not so much that retail is hiring more, as that other sectors are hiring comparatively less.
Retail jobs have expanded at roughly the same rate as jobs overall (click on the graph to enlarge). My surprise is that the shift to online and the slow growth of jobs & income has not slashed retail jobs. Big boxes crush smaller retailers. Automation reduces headcount. Lots of bankruptcies. Given these things I find this graph quite amazing, from the January 2015 BLS Highlights:
I’ve seen the same graph. I have no idea why this is, and I can only speculate.
My best guess is that the steady increase in Retail employment might be because of the steady shift from full-time to part-time workers.
Many retailers now prefer to hire two part-time workers instead of one full-time. There are a variety of reasons for this, including labor law and benefits cutoffs, and general scheduling flexibility. It’s a trend that’s been very apparent to anyone working in that sector, and it makes things much more difficult for those workers trying to earn a living.
If we saw the number of jobs in Retail has gone up, but the total hours worked in Retail has remained flat, then part-time workers would clearly be the culprit. I can’t find enough information about it to draw that conclusion, however.
That’s a logical explanation! Note that the full-time workers have not recovered to levels before the crash — in terms of % of the civilian non-farm population.
FM claims: “Yes, that’s the standard narrative. But economists have shown that the effect of demographics has been grossly exaggerated, and there is almost nil effect so far from “advanced robotics and AI” (it’s a big concern for the future, but not yet).”
That is provably false.
Robots have destroyed high-paying employment for high school graduates by eliminating the $20-an-hour entry-level jobs in manufacturing since the 1970s, particularly in the automotive industry.
Once upon a time, your typical high school graduate could get a very highly paid job at a Ford or GM or Chrysler assembly line or at a steel factory and raise a family and buy a house. Today, your typical high school graduate can’t afford to start a family because the available low-age service jobs don’t pay enough.
Since FM will of course deny these well documented facts, time to bludgeon him with the statistics:
Source: “A Guide to Statistics on Historical Trends in Income Inequality,” 10 December 2014.
Source: “Entry-level workers’ wages fell in lost decade,” 7 March 2012, http://www.epi.org.
Robots were introduced into automotive assembly lines between 1973 an 1995.
As always, FM will deny, deny, deny, and regardless, the facts remain clear:
Source: “Industrial robots moving to sectors beyond the automotive industry,” 4 December 2014.
These facts will simply result in more denial by FM, so let’s be pellucidly clear about what’s happened to auto manufacturing employment since the 1970s:
Notice in particular the chart labeled “Michigan’s share of U.S. auto engineers and production workers.”
Source: Federal Reserve Bank of Chicago, Detroit Branch
FM will predictably retort that despite these catastrophic job losses in auto and steel production, many more jobs were created since 1970 in other industries. That is true. But the jobs created since 1970 were in much lower-wage service industries, not highly-paid manufacturing jobs. Those highly paid jobs have been destroyed in the auto industry by robotics, and FM’s efforts to deny this undeniable fact merely deep-six his waning credibility.
“Robots have destroyed high-paying employment for high school graduates by eliminating the $20-an-hour entry-level”
Yes, I think that’s quite obvious. It depends on the definition of “robot”. In the context of the statement “advanced robotics and AI” I interpreted this to mean devices capable of sophisticated sensing and autonomous decision-making. As I and so many others have written, these are coming — and will be job-killers on a scale almost unimaginable.
You are referring to automation, aka mechanization. Yes, those have been called robots — but they’re not in any meaningful sense IMO.
That’s quite a daft interpretation of what I wrote, like interpreting someone’s statement about it being cloudy and dark as implying that the sun didn’t rise today. It’s quite troll-like, in the literal sense of idiotic belligerence. When you believe someone has said something obviously false, you should check your interpretation before launching into a proof that the sun in fact rose today.
“As always, FM will deny, deny, deny, and regardless, the facts remain clear:”
The Force is weak in you, padawan. Stick to facts and leave the precognition to others.
Pingback: Data mining job growth - Yerasoft