Or, I know columnists have deadlines, but really...

In Tuesday’s NYT column, entitled “The Structural Revolution”, David Brooks writes:

[I] ... believe the core problems are structural, not cyclical. The recession grew out of and exposed long-term flaws in the economy. Fixing these structural problems should be the order of the day, not papering over them with more debt.

...

... we structuralists do not believe that the level of government spending is the main factor in determining how fast an economy grows. If that were true, then Greece, Britain and France would have the best economies on earth. (The so-called European austerity is partly mythical.) We believe that the creativity, skill and productivity of the work force matter most, and the openness of the system they inhabit.

I wonder if Mr. Brooks has ever taken a course in macroeconomics. Almost every intermediate macroeconomics textbook I know of uses a Solow growth model to determine long run growth, and a Keynesian type model for the short run, so I don’t understand this simple dichotomization.

I think there should be a rule that anybody who writes about macroeconomics should be forced to read a macroeconomics textbook; even if Mr. Brooks had read Barro’s textbook, at least he’d understand under what he was criticizing was a straw man.

Here is the CBO’s assessment of the structural and cyclical components of GDP and of unemployment.

Figure 1: Log GDP (blue) and log potential GDP (red), bn. Ch.2005$. NBER defined recession dates shaded gray. Source: BEA, 2012Q1 advance release, CBO, Budget and Economic Outlook (January 2012), NBER and author's calculations.

Figure 2: Unemployment rate (blue), and natural rate of unemployment - long term (green) and natural rate of unemployment - short term (red). Source: BLS vis St. Louis Fed FRED, CBO, Budget and Economic Outlook (January 2012), and NBER.

By these measures, most of the deviations are “cyclical”. There is a structural aspect, of course. The CBO’s measures take into account trends in capital stock, labor force, and multifactor productivity. If that ain’t structural, I don’t know what is.

The CBO’s assessment of the natural rate of unemployment, including the recent measures of short term natural rate, highlights the fact that structural does not necessarily mean long-lived, nor does it necessarily mean immune to measures implemented to address cyclical weakness. These measures include extended unemployment insurance, etc.

In fact, maybe it would be useful to quote from CBO’s Budget and Economic Outlook, released at the end of January 2012 (p.35-36):

Cyclical and Structural Unemployment. In CBO’s view, most of the 3.5 percentage-point rise in the unemployment rate since the onset of the recession can be directly attributed to a cyclical decline in the demand for goods and services, and hence for workers. However, CBO estimates that part of that rise—roughly 1 percentage point—reflects structural factors associated with the recession but not directly linked to the current level of aggregate demand. Those structural factors include a mismatch between the requirements of existing job openings and the characteristics of job seekers, including their skills and locations; the lasting effect of long-term unemployment on individual workers’ ability to find and hold a job; and the effect of extended unemployment insurance benefits on incentives to continue searching for work (as opposed to either accepting a job offer or dropping out of the labor force). Although quantifying the relative importance of these factors is quite difficult, CBO estimates that in late 2011 the rate of unemployment attributable to sources other than the current level of demand for goods and services—the so-called natural rate of unemployment—was about 6 percent, up from about 5 percent before the recession. In CBO’s projections, most of the effect of those structural factors on the unemployment rate fades by 2022.

Roughly half of the 1 percentage-point rise in unemployment that CBO attributes to structural factors reflects mismatches between the skills and locations of available unemployed workers and the needs of employers, CBO estimates. One important source of such mismatches is the decline in demand for construction workers that followed the collapse of the housing market. The effect of mismatches on the unemployment rate is projected to diminish gradually over the next five years—as people acquire new skills and, in some cases, relocate to faster growing regions and as some older workers who lost their jobs during the recession leave the labor force.

About a quarter of the 1 percentage-point increase due to structural factors can be attributed to the effects that extended unemployment insurance benefits have had on the supply of labor. Such benefits induced some unemployed people to search for work less intensively or to reject unsatisfactory job offers. The benefits also encouraged some unemployed people who would otherwise have stopped looking for a job and dropped out of the labor force to stay in it to remain eligible for benefits. If extended unemployment insurance benefits expire on February 29, as scheduled under current law, those effects will dissipate by the summer of 2012.

The remaining roughly one-quarter of a percentage point reflects the difficulties that the long-term unemployed (people who have gone without a job for at least six months) face in finding work. Such workers may encounter difficulties resulting from the stigma attached to long-term unemployment—that is, employers’ perception that the long-term unemployed would be low quality workers—and from the erosion of their skills while they are unemployed. As a result, some workers who have been unemployed for a long time, especially those displaced from a long-tenured job, are likely to have trouble landing another stable job. Consequently, they could remain unemployed for an extended period; moreover, even after they are reemployed, many will remain more vulnerable than before to additional future spells of unemployment. As a factor boosting unemployment, such difficulties for the long-term unemployed will, in
CBO’s view, increase in importance over the next two years (as some people who are currently out of work stay out of work longer) and then persist for several more years, before gradually diminishing but not completely disappearing by 2022.

Emphasis added -- MDC. This means a majority of the increase in unemployment is not "structural" in the sense that Mr. Brooks uses the term.

(By the way, I know there will be a bunch of commenters who claim the CBO’s approach is flawed – usually without understanding how those measures are constructed – but I find it telling that alternative approaches say using DSGEs and assuming monopolistic competition – obtain similar measures of the natural rate [1].)

Experts concur that most of the unemployment we see is cyclical in nature, and even a large component of the increase in “structural” unemployment is policy-driven. From the summary of the SF Fed’s recent conference on unemployment, by Neumark and Valletta:

The recent San Francisco Federal Reserve Bank conference on workforce skills examined labor market changes that may have accelerated during the Great Recession. These changes may have increased mismatches between employer needs and worker skills. In general, we find that this doesn’t appear to be the case. Estimates of the extent of skill mismatches in recent years indicate that it has been limited and is likely to dissipate. Moreover, the conference’s research presentations and a panel of workforce development specialists did not identify a noticeable increase in mismatches in recent years. Thus, concerns about growing skill mismatches may be overblown. On the other hand, successful integration of low-skilled workers into the workforce represents a continuing problem. Conference participants offered useful ideas on how to meet this challenge, stressing the roles of community colleges and well-designed training programs.

The conference website is here. But I think it might be too much to ask Mr. Brooks to read more broadly, given his predisposition to spout data-free assertions. See for instance this post on Mr. Brooks’ earlier column on long term unemployment. From Mr. Brooks' column:

There are probably more idle men now than at any time since the Great Depression, and this time the problem is mostly structural, not cyclical. These men will find it hard to attract spouses. Many will pick up habits that have a corrosive cultural influence on those around them. The country will not benefit from their potential abilities.

This is a big problem. It can’t be addressed through the sort of short-term Keynesian stimulus some on the left are still fantasizing about. It can’t be solved by simply reducing the size of government, as some on the right imagine.

And as I wrote before:

While surely there is a structural component, Brooks is making the common mistake of equating long term unemployment with structural unemployment. As discussed in this post, the two are related, but the bulk of the current unemployment is cyclical in nature.

My final observation: in the neoclassical synthesis, potential output is a function of the capital stock, the labor stock, and multifactor productivity, i.e., Y^Pot = AF(K,L), where A is MFP, K is capital, L is labor. K is the sum of private and public capital. Augmentation of public capital is primarily due to investment in infrastructure. I believe many critics of the ARRA pointed to investment in infrastructure as a problem, while many (including in the Administration) have asked for more investment in infrastructure. This makes me think that a lot of those of us who Mr. Brooks calls "cyclicalists" have indeed been thinking long and hard about structural factors.

As I've asked before, would it kill columnists who write about macro to read a textbook (even a neoclassically oriented one like Barro, although I use Blanchard in my intermediate macro courses).