Is Structural Unemployment 'Humbug' or Are Krugman and Baker Biased?

Were David Brooks and Ruth Marcus so wrong to talk about some form of "structural unemployment" when July's jobs numbers came out? Paul Solman disagrees with Paul Krugman and Dean Baker that there's an economic consensus behind a cyclical theory of unemployment.

By Paul Solman


Is there really an economic consensus that unemployment is cyclical, as Paul Krugman suggests? Photo courtesy of Justin Sullivan/Getty Images.

Liberal economist and much-respected friend Dean Baker, co-director of the Center for Economic and Policy Research, where he keeps the Beat the Press blog, has appeared on PBS NewsHour often over the years, and recently on these pages in "Don't Blame the Robots." He appeared here again Wednesday, decrying what he called the media's "mindless" budget reporting.

But when he wrote on his blog on Aug. 3 that "[t]he PBS Newshour won the gold medal for journalistic malpractice on Friday (Aug. 2) by having David Brooks and Ruth Marcus tell the country what the Friday jobs report means," he seemed curiously harsh and patently partisan.

"Brooks and Marcus got just about everything they said completely wrong," Baker continued. "Starting at the beginning, Brooks noted the slower than projected job growth and told listeners: 'Yes, I think there's a consensus growing both on left and right that we -- the structural problems are becoming super obvious...'"

BROOKS AND MARCUS On Unfinished Business as Congress Leaves for August Break

But, Baker insisted, "It's hard to know what on earth Brooks thinks he is talking about. There is nothing close to a consensus on either the left or right that the economy's problems are structural, as opposed to a simple lack of demand (i.e. people spending money). This is shown clearly by the overwhelming support on the Federal Reserve Board for its policy of quantitative easing."

In a post on his New York Times blog, "Conscience of a Liberal," another long-featured NewsHour commentator and key protagonist on the Making Sen$e Business Desk, Paul Krugman, enthusiastically seconded Baker's condemnation and called the previous day's Brooks/Marcus discussion on unemployment "structural humbug."

He began: "OK, this is really depressing. The PBS Newshour isn't always a good place to get the best analysis, but it's a terrific place to take the pulse of Washington conventional wisdom -- and as Baker notes, that conventional wisdom has clearly swung to the view that our high unemployment is 'structural', not something that could be solved simply by boosting demand."

"In short, the data strongly point toward a cyclical, not a structural story," Krugman concluded, "and there is broad agreement, for once, among economists on this point. Yet somehow, it's clear, Beltway groupthink has arrived at the opposite conclusion -- so much so that the actual economic consensus on this issue wasn't even represented on the Newshour."

"As I said, this is really, really depressing," he finished.

Now, as his readers know, Krugman is easily depressed. Indeed, in 2009 singer/songwriter Loudon Wainwright III serenaded an audience with "The Paul Krugman Blues," which features this verse:

"When Paul goes on the NewsHour, To talk to old Jim Lehrer, He looks so sad and crestfallen It's more than I can bear."

Baker is rather more chipper, even taking me on an economic silver lining tour during the depths of the Crash of '08. And there is no "Dean Baker Blues" of which I am aware.

Look, folks, there may indeed be no "consensus growing on left and right" about the predominance of structural unemployment, as David Brooks alleged. Just look at how vigorously Krugman and Baker took the other side. But I rather doubt Krugman's assertion that there is an "actual economic consensus" on the unemployment debate that favors his cyclical explanation to the exclusion of the structural. Unless, of course, Krugman means a consensus among economists he agrees with.

A confession: Brooks is a friend for whom I have great respect, as I do for Ruth Marcus. And, of course, I work for the NewsHour covering economics and have done so for 28 years now. So readers would sympathize, I trust, were I to reflexively defend all parties under attack, even though the attackers are also people I have long known and respected.

But quite apart from sentiments and loyalties, I think Baker and Krugman were not simply being ungenerous, but misleading. Is it really "humbug" to suggest that there's a mismatch between employers and job seekers in today's U.S. economy?

Krugman followed up the day after his critique with his definition of "structural unemployment": it is "the assertion that the 'full-employment' rate of unemployment, the level of unemployment at which prices and wages start to rise and you risk a wage-price spiral, has increased. When that happens, you can't solve the unemployment problem just by getting someone to spend more and thereby increasing demand; when it hasn't happened, you can."

Well, that is one definition, which Krugman elaborated on in 2010. But usually, when people speak of "structural unemployment," they mean something more casual: a mismatch between jobs and workers that's built into the economy, at least for some extended period of time.

Here's what Brooks went on to say on the NewsHour on August 3:

When "this recession started a number of years ago, you had 63, something like that, out of 100 Americans in the labor force. Now we're down, fewer than when the recession started (below 59 percent employment/population ratio). And so that suggests we have got some deep structural problems. It probably has a lot to do with technological change. People are not hiring -- companies are not hiring human beings. They're hiring machines.

"It probably has to do with a skills shortage, that as technology increases, skills have got to keep up and skills are just not keeping up. It has to do with some sociological changes, men dropping out of the labor force, women, and especially young women, never entering the labor force.

"And so these are deep structural changes. And I think there's a consensus growing that something really fundamental has shifted in the economy."

Now Brooks could be wrong about his points of emphasis, of course. His "probably's" acknowledge as much. He didn't mention that the baby boom has begun to retire, for example, and that surely lowers the employment-to-population ratio. It has also been pointed out that workers whose homes are "underwater," or worth less than their mortgages, may not be able to move and take jobs elsewhere in the country. Heck, employers may simply be unwilling to hire people at wages that Americans will accept because of what was once called "the reserve army of the unemployed," as I wrote here last year and chronicled on the NewsHour at a Mott's Apple Sauce factory strike soon after. But Brooks wasn't pretending to come up with a comprehensive list. And the employment-to-population ratio from before the Great Recession until today is just as he remembered it:


The above chart shows the Bureau of Labor Statistics' seasonally-adjusted employment-to-population ratio for individuals 16 and over from the Current Population Survey.

This isn't a debate that's likely to be resolved by whether or not inflation has reared its head. And of course Krugman and Baker are right in suggesting that government activism can solve the unemployment problem. All it has to do is create enough jobs that people are willing to take another batch of alphabet soup agencies like President Franklin Roosevelt's Civilian Conservation Corps (CCC), chronicled by PBS's "American Experience," or my own thought experiment, the Mass Massage Mobilization (MMM), which I suggested to Krugman during an interview this spring.

But CCC jobs paid a dollar a day plus minimal room and board, the equivalent of between $17 and $50 for an eight-hour day in today's dollars, according to the Measuring Worth calculator, which comes to less than the minimum wage by even the most liberal conversion formula. Would enough people take the jobs? At some wage, they would. But is the answer to "cyclical unemployment" really for government to create jobs and pay wages high enough to get us to "full employment"?

Unlike Krugman and Baker, my main job for 36 years now has been to interview not only economists like them, but hirers and hirees, firers and firees. I've done so through both recessions and recoveries alike. I wheedled soundbites out of the drearily downhearted high tech-workers of the late 1970s and spoke to the happily hopeful hires of the late 1990s.

THE SOLMAN SCALE Is Baby Boomer Retirement Behind the Drop in July's Unemployment Rate?

So yes, it's true: the economy has cycled up and down. Yes, there was "full employment" as recently as the late 1990s under President Bill Clinton -- down near 4 percent just before the dot.com crash, though you could hardly call Clinton's spending policies Keynesian at the time, since he was building budget surpluses (net of Social Security and Medicare), taking money out of the economy instead of adding to the annual deficit and thus raising our cumulative national debt.

And yes, it's true: the same thing happened at the end of President George W. Bush's second term -- we had near-full employment -- just before the Crash of '08.

But one counter argument is that the higher you fly, the harder you fall -- that the "full employment" of 2000 and 2008 were bubble-fueled mirages: the Internet bubble of '90s; the housing bubble of "the aughts."

Another counter is that for decades now, official unemployment has been severely undercounted, an argument that Baker has convincingly made himself. We first reported this on the NewsHour in a story called "Non-working Numbers" back in 2003 when the term "jobless recovery" first came into vogue. We followed up in 2009 and in January of 2011, inaugurated our own more inclusive measure of under- and unemployment, "U-7," which we update the first Friday of every month in lockstep with the release of the government's official unemployment number, aka "U-3."

None of the above is meant to argue that cyclicality plays no role in the unemployment rate. The chart below doesn't adjust for the official undercounting of unemployment, which began in the mid-'90s. If it did, the recent peaks would be much higher. But who would wish to contest the roller-coaster nature of the U.S. job market?


Bureau of Labor Statistics' historical unemployment rate reflects the roller-coaster nature of the U.S. job market.

Cyclical and structural unemployment are not mutually exclusive. A 2011 paper from the San Francisco Fed attributed 60 percent of long-term unemployment to cyclicality and 40 percent to structural factors.

Conservatives like Brooks may think that corporate tax cuts will address the structural problem. Liberals like Baker and Krugman may think that's nonsense. Fair enough. But to deride Brooks and Marcus for suggesting that structural unemployment plays an important role these days seems, if not "humbug," to be at the very least intellectually stinting. Scrooge-like, even. For more on the structural/cyclical debate, here's one of our Making Sen$e reports from 2011.

This entry is cross-posted on the Rundown -- NewsHour's blog of news and insight. Follow @paulsolman