I like maps.  Randy Olsen recently linked to an interesting map of European unemployment:

Before seeing this map I knew that Germany had higher lower unemployment than France.  And indeed you see a dramatic change in the unemployment rate right where the two countries meet—at the Rhine.

My assumption has been that the lower German unemployment reflects better labor market regulations. But this graph hints at another difference—language.  Let’s suppose that Germanic language areas have lower unemployment than Romance language areas.  How would we test whether language or government policies are more important?

One approach would be to note that some countries have multiple languages.  Thus southern Belgium and southwestern Switzerland speak Romance languages, and the rest of those countries mostly speak Germanic languages.

Now look at the map.  I see a sharp unemployment divide in Belgium, right on the line between French and Flemish.  And in Switzerland, I see much higher unemployment in French-speaking Geneva, Vaud, and Valais, as well as Italian-speaking Ticino.  Below is a map of Switzerland, so you can see the cantons more clearly.

I also notice an area of very high unemployment from Greece to southern Italy to Sardinia to southern Spain. So maybe it’s not language, maybe it’s latitude.  Or maybe both language and latitude are proxies for culture.

Before we fall in love with this theory, however, there is just one problem.  As recently as 2005, Germany had about 11% unemployment and was regarded as the “sick man of Europe”.  We should be suspicious of cultural explanations that apply in one decade, but not the next. Surely culture doesn’t change very much in 12 years!

In the end, I’m included to think, “it’s complicated”.  There are probably lots of factors that are playing a role here.  Southern Belgium was once richer than northern Belgium, but today it’s more of a rust belt, as its old heavy industry is less competitive than the newer firms in the Flemish areas to the north.  Perhaps there are similar explanations for Switzerland. Perhaps culture and government policy interact—thus people in the far south of Europe have a culture that leads to political outcomes associated with excessive labor market rigidity.

So no easy answers, but the map is certainly food for thought.

PS.  Unemployment is related to wealth, and indeed probably more closely related than in earlier decades.  But the correlation is still far from perfect.  Northern Italy does better on wealth than employment (check out my “rich heart of Europe” post.)  For instance, Lombardy has more unemployment than northern England, despite being richer.  Or compare (rich) Paris with the Czech Republic.

Some people argue that the 4.7% unemployment rate is misleading.  In fact, our PEOTUS makes that argument, citing figures as high as 30% or 40%. There are tens of millions of workers who are not even trying to find work (so we are told) because they are so “discouraged” by 4.7% unemployment and the Obama record of 200,000 new jobs a month for the past 7 years.  And they lack the education to do work like computer coding. But then I read articles like this:

The development boom that’s changing downtowns across the country—and adding new units in hyper-competitive markets—has also led to an acute shortage of qualified construction workers, which is starting to weigh heavily on future projects and planning. As of April 2016, there were over 200,000 unfilled job openings in building construction, according to the Bureau of Labor Statistics Job Openings and Labor Turnover Survey(JOLTS).

Wait, aren’t construction jobs supposed to be the sort of work that is appropriate for non-college educated workers?  What am I missing?  Yes, some construction jobs are highly skilled, but not all.  Heck, even I have done painting, drywall and roofing.

There simply is not much “slack” left.  Many of the long-term unemployed will never work again, even if we have a 1969 or 1999-style economic boom. 

Trump loves big projects, and wants to spend a lot on infrastructure:

“There isn’t much capacity left in the construction industry,” says Julian Anderson, President of Rider Levett Bucknall, North America, a property and construction firm. “There’s a big labor shortage, and construction unemployment is down to 4 percent. It’s so nuts in LA and San Francisco, it’s gotten to the point where it’s probably turning off development.” Anderson says that proposals to deport undocumented immigrants, who make up a sizable portion of the construction workforce in some markets, may severely exacerbate the issue.

Confused?  Join the club.

How about a career in law enforcement?

Economic and social changes have made it harder for police departments to keep their forces fully staffed, and lead to increasingly desperate recruitment.

The Los Angeles Police Department was short of nearly 100 officers as of mid-December—only 1% of its total workforce, but still enough to be felt on the ground, says Captain Alan Hamilton, who runs recruitment for the department. Philadelphia had 350 vacancies, largely due to a spate of retirements. Last spring, Dallas cancelled two academy classes for lack of applicants; its preliminary applications dropped by over 30% between 2010 and 2015. In 2012, the ratio of police officers to population hit its lowest level since 1997, according to Uniform Crime Reporting Programme data published by the FBI.

The dynamics underpinning the shortages vary by department, but there are national trends making it harder for police forces to attract applicants. The first is a strong economy. Nelson Lim, a researcher at the RAND Corporation, a think-tank, says this is nothing new. When plenty of jobs are available, people are usually less motivated to enter dangerous professions. Police forces as well as the armed forces tend to field less interest in boom times.

(The article mentions a slight uptick last year in police fatalities, but they are still down sharply from 10 years ago.)  You see similar stories in manufacturing. Yesterday I heard on CNBC that there is a severe shortage of welders in manufacturing, with 300,000 positions unfilled.  This contributes to manufacturing going to places like China.  Companies in all sorts of “working class” sectors are having trouble finding qualified employees at current wage rates.  Is the problem a tight labor market, or a lack of technical education?  I’m not sure, but I don’t see much evidence for the “robots stealing our jobs” claim.

The implications are clear, if America wants to create substantially more jobs, we don’t need monetary and fiscal stimulus—we need deregulation, supply-side tax reform, better technical education, and most importantly we need to boost the rate of immigration.

PS.  I recall that a substantial share of police officers are now college grads, so I don’t mean to suggest they are all low-skilled jobs.

PPS.  Off topic, but people wonder why I have such a low opinion of Trump on foreign policy.  Here’s his pick for Secretary of State:

At his confirmation hearing before the US Senate Foreign Relations Committee on Wednesday, Mr Tillerson said China’s construction of artificial islands in the contested sea was “akin to Russia’s taking Crimea” from Ukraine in 2014.

“We’re going to have to send China a clear signal that, first, the island-building stops and, second, your access to those islands also is not going to be allowed,” he said.

Yeah, dumping some gravel on an uninhabited reef is sort of like invading a European country of 40 million people and annexing a part of its territory.

And is Russia’s access to the Crimea also “not going to be allowed”?  These people are truly evil—their worldview boils down to “Russia good, China bad”.

I feel like I fell asleep for 20 years and woke up in a country I don’t even recognize. I gather Marco Rubio feels the same way, given his exasperated tone when questioning Tillerson.

PPPS.  My grandmother’s surname was “Van Winkle.”  Seriously.

In recent years, the unemployment rate has fallen to low levels (4.7%).  Some people continue to argue, however, that there is a lot of labor market slack.  They point to data such as the U-6 unemployment rate, as well as the labor force participation rate. But with today’s jobs report, it’s becoming almost impossible to continue arguing for labor market slack.  The labor market is clearly tightening:

There is no longer any respectable argument for monetary or fiscal stimulus.  There are arguments for stimulus that deny the natural rate hypothesis, but having lived through the 1970s once, I’d just as soon not experience it again.

Time for supply-side policies, beginning with a repeal of those idiotic overtime rules.  Then on to tax reform.

PS.  I am at the AEA meetings, so blogging will be sporadic for the next few days.

Over at Econlog I have a post that suggests the answer is no, CA deficits do not cost jobs.

But suppose I’m wrong, and suppose they do cost jobs.  In that case, trade has been a major net contributor to American jobs during the 21st century, as our deficit was about 4% of GDP during the 2000 tech boom, and as large as 6% of GDP during the 2006 housing boom.  Today it is only 2.6% of GDP.  So if you really believe that rising trade deficits cost jobs, you’d be forced to believe that the shrinking deficits since 2000 have created jobs.

So why have manufacturing jobs plummeted since 2000?  One answer is that the current account deficit is the wrong figure, since it also includes our surplus in trade in services.  If you just look at goods, the deficit is closer to 4.2% of GDP.

But even that doesn’t really explain very much, because it’s slightly lower than the 4.35% of GDP trade deficit in goods back in 2000.  So again, the big loss of manufacturing jobs is something of a mystery.  Yes, we import more goods than we used to, but exports of goods have risen at about the same rate since 2000.  So why does it seem like trade has devastated our manufacturing sector?

Perhaps because trade interacts with automation.  Not only do we lose jobs in manufacturing to automation, but trade leads us to re-orient our production toward goods that use relatively less labor (tech, aircraft, chemicals, farm produces, etc.), while we import goods like clothing, furniture and autos.

So trade and automation are both parts of a bigger trend, Schumpeterian creative destruction, which is transforming big areas of our economy.  It’s especially painful as during the earlier period of automation (say 1950-2000) the physical output of goods was still rising fast.  So the blow of automation was partly cushioned by a rise in output.  (Although not in the coal and steel industries!)  Since 2000, however, we’ve seen slower growth in physical output for a number of reasons, including slower workforce growth, a shift to a service economy, and a home building recession (which normally absorbs manufactured goods like home appliances, carpet, etc.)  We are producing more goods than ever, but with dramatically fewer workers.

Update:  Steve Cicala sent me a very interesting piece on coal that he had published in Forbes.  Ironically, environmental regulations actually helped West Virginia miners, by forcing utilities to install scrubbers that cleaned up emissions from the dirtier West Virginia coal.  (Wyoming coal has less sulfur.)  He also discusses the issue of competition from natural gas.

A commenter named dwb left this comment:

The “technological change” that killed coal jobs is the 1-2-3 punch of cheap natural gas, low electricity demand, and Obama’s war on fossil fuels.

At least he doesn’t blame trade.  Even so, this is basically false—except for very recently, coal jobs have been lost to automation. Here’s employment in the coal industry:

It’s even worse than it looks, as office jobs were added in 1973, creating an artificial surge in the data.  If you just count actual miners, the job losses have been far worse.  But even this graph shows a loss from 870,000 jobs to about 110,000, slightly worse than in steel.

So you might assume that our coal industry is being overwhelmed by imports, right?  No, over the past 5 years we’ve been net exporters of coal, in the range of 7% to 12% of total production.

If it’s not imports, then production must be being hammered by competition from oil and gas, right?  Not really, as the following graph shows, the coal industry has been increasing production in recent decades, until the past few years when competition from oil and gas really did eat into production:

So why have so many coal jobs disappeared?  The answer is simple, automation. We are producing nearly twice as much coal as when I was young, and we are doing so with far fewer workers.

Some commenters think that job loss due to automation is less painful than job loss due to trade.  In fact, they are equally painful.  Jobs lost to automation don’t occur gradually over time, through attrition, they occur in waves, often during recessions. Thus in steel, 1000s of jobs are lost when US Steel or Bethlehem shut down old mills, and Nucor and Chaparral open new more efficient mills in other parts of the country. Lots of steel jobs lost in Pittsburgh are replaced with a smaller number gained in Texas.

Something similar happens in coal.  Big new strip mines in Wyoming use huge shovels that replace 100 workers in a West Virginia mine that shuts down.  Here’s the production of coal by state:

If Wyoming were another country, the West Virginia miners would be screaming at their representatives that they need “protection” from cheap Wyoming imports. But because Wyoming is as American as apple pie, nobody advocates tariffs, even though the economic issues are exactly the same as when Ohio steel is impacted by Chinese imports.

I see an orgy of sanctimonious commentary in the media about how we have to pay more attention to the suffering of Ohio steel workers and West Virginia coal miners.  OK, but how many of those pundits realize that the interests of those two groups are diametrically opposed?  If Trump pursues a protectionist policy to help steel, it will hurt American coal exports.  TPP would be a boom to West Virginia, while threatening Ohio manufacturers.

But at a deeper level, the problems facing coal and steel are exactly the same.  In the US, and indeed almost everywhere in the world, automation is rapidly reducing employment in mining and manufacturing.  That problem is not going to go away, indeed with advances in robotics it will get even worse.  Trump can make a few symbolic moves (Carrier, weaker environmental laws, etc.) which will save a handful of jobs, and cost other jobs that are invisible to the public, but it won’t change anything fundamental. It will just give us a dirtier, hotter planet.  And rust belt workers will still be angry.

It’s always comforting to demagogue the issue by blaming foreigners for our woes; but they are doing the same—blaming other foreigners, including us.