The hallmarks of so-called “neoclassical” economics are: (1) a faith that, no matter what hits the economy takes, it will always grow itself back to full employment, at least eventually; (2) changes in the distribution of wealth and income (inequality) have no macroeconomic effects; (3) the amount of an economy’s aggregate income (GDP) and employment are determined by supply-side factors; and (4) aggregate demand, the cornerstone of Keynesian theory, is a tag-along, a mere reflection of the supply of goods and services produced.
This perspective has emerged since the advent of the industrial revolution and the dominance of corporations, beginning around 1850 in Europe, and around 1880 in the United States. Today, a fully developed neoclassical ideology dominates mainstream economics. The problem with this “don’t worry, be happy” ideology, from the standpoint of ordinary citizens with ordinary incomes (roughly the bottom 99%), is that it obscures what is really going on: A slowly worsening depression within the bottom 99% of the U.S. economy and the rapidly rising incomes of very wealthy people (roughly the top 1%) are masked by the image of slow, across-the-board recovery from a retreating recession.
It has been my expectation that, as the deterioration of the bottom 99% economy continues, the bankruptcy of neoclassical thinking would necessarily reveal itself, the only question being when and how. Earlier this year we witnessed the demise of the alleged academic underpinnings of the “austerity doctrine” when the Reinhart/Rogoff study “Growth in a Time of Debt” (GITD) was discredited, forcing the researchers to withdraw their thesis. See my detailed analysis in two earlier posts in the “Finding a New Macroeconomics” series: “Reinhart, Rogoff, and Reality” (here) and “Reinhart, Rogoff, and Ideology” (here).
Now another revealing and important development has been called to our attention. On November 8, 2013, in his New York Times Op-Ed “The Mutilated Economy” (here), Paul Krugman described his reaction to a study released at the recent IMF conference “Crises: Yesterday and Today” (here). The study, he reports, has persuaded him that the United States has been “sacrificing the future” with policies of “economic self-mutilation” that have created a “terrifying” amount of long-term damage.
The level of Krugman’s alarm, in my opinion, is more than justified.
The thing is, you wouldn’t know it from studying the report. The views of its authors are more equivocal and reserved, and for what are likely ideological and political reasons they do not convey the same overall sense of urgency. Still, their discussion unavoidably reveals weaknesses they themselves perceive, or suspect, in the “supply-side neoclassical” paradigm.
The study, “Aggregate Supply in the United States: Recent Developments and Implications for the Conduct of Monetary Policy” (“the FED study,” here), is the work of Federal Reserve Board econometricians Dave Reifschneider, William L. Wascher, and David Wilcox. These researchers are neoclassical economists in the tradition of Ben Bernanke, a well-known disciple of Milton Friedman, who recently argued that inequality in America is mostly a matter of people with college and advanced degrees earning higher incomes than high school graduates (New York Times, 12/5/2010, here). These researchers do not refer at all to income and wealth redistribution: Indeed, their paper exemplifies all of the hallmarks of neoclassical thinking listed at the top of this post.
Paul Krugman, on the other hand, is a hybrid. We might call his economic school “neo-Keynesian,” as opposed to the “supply-side neoclassical” paradigm exemplified by the Federal Reserve Board researchers. As a Keynesian, Krugman rejects hallmarks (3) and (4), but he has subscribed to (1) and (2), a point frequently emphasized in this blog. Krugman is America’s chief spokesman for what remains of Keynesian economics in today’s neoclassical quagmire. I am certain that his heart is in the right place, and that he genuinely wants economic recovery for the sake of the American middle and lower classes. But he limits himself by tending to regard U.S. economic history as little more than a series of “financial shocks” interrupted, every 80 years or so, by a deep depression, with unemployment and stagnation the consequences of these financial shocks.
I don’t know whether the Fed study’s authors, who appear to have a financial orientation and background similar to Krugman’s, share his deep concern for the middle and lower classes. What I do know is that the neoclassical belief system has been deeply ingrained in mainstream economics for decades, and that fallacies in that system of thought are typically overlooked, and challenges to it dismissed. But as Harvard economist Raj Chetty candidly conceded in a recent New York Times Op-Ed (10/20/13, here), for mainstream economists “the answers to many ‘big picture’ macroeconomic questions – like the causes of recessions or the determinants of growth – remain elusive.” They don’t yet have, as he put it, “a precise understanding of how the economy works.”
That’s why continuing to evaluate mainstream beliefs and conclusions is essential. This study and Krugman’s reaction to it offer a rare opportunity for such an evaluation. There are signs in this study that the U.S. economy is performing more poorly than mainstream economists had expected. There are also signs that the authors have not found in the “conventional” neoclassical belief system adequate explanations for the prolonged and continuing decline.
The Fed study focuses on two issues: (1) Potential GDP, with a finding of declining growth in the components of aggregate supply; and (2) labor market performance, with a finding of “some structural damage in the labor market,” and “a modest rise in the natural rate of unemployment” (pp. 2-3).
Let’s start with Krugman. His article begins:
Five years and eleven months have now passed since the U.S. economy entered recession. Officially, that recession ended in the middle of 2009, but nobody would argue that we’ve had anything like a full recovery. Official unemployment remains high, and it would be much higher if so many people hadn’t dropped out of the labor force. Long-term unemployment — the number of people who have been out of work for six months or more — is four time what it was before the recession (here).
These dry numbers translate into millions of human tragedies — homes lost, careers destroyed, young people who can’t get their lives started. And many people have pleaded all along for policies that put job creation front and center. Their pleas have, however, been drowned out by the voices of conventional prudence. We can’t spend more money on jobs, say these voices, because that would mean more debt. We can’t even hire unemployed workers and put idle savings to work building roads, tunnels, schools. Never mind the short run, we have to think about the future!
The bitter irony, then, is that it turns out that by failing to address unemployment, we have, in fact, been sacrificing the future, too. What passes these days for sound policy is in fact a form of economic self-mutilation, which will cripple America for many years to come. Or so say researchers from the Federal Reserve, and I’m sorry to say that I believe them. * * *
It’s pretty clear . . . that the blockbuster paper of the conference will be one that focuses on the truly ugly: the evidence that by tolerating high unemployment we have inflicted huge damage on our long-run prospects.
How so? According to the paper . . . , our seemingly endless slump has done long-term damage through multiple channels. The long-term unemployed eventually come to be seen as unemployable; business investment lags thanks to weak sales; new businesses don’t get started; and existing businesses skimp on research and development.
What’s more, the authors — one of whom is the Federal Reserve Board’s director of research and statistics, so we’re not talking about obscure academics — put a number to these effects, and it’s terrifying. They suggest that economic weakness has already reduced America’s economic potential by around 7 percent, which means that it makes us poorer to the tune of more than $1 trillion a year. And we’re not talking about just one year’s losses, we’re talking about long-term damage: $1 trillion a year for multiple years.
That estimate is the end product of some complex data-crunching, and you can quibble with the details. Hey, maybe we’re only losing $800 billion a year. But the evidence is overwhelming that by failing to respond effectively to mass unemployment — by not even making unemployment a major policy priority — we’ve done ourselves immense long-term damage.
This is the graph of long-term unemployment from FRED (St. Louis Fed. Reserve Bank) that Krugman linked to his article:
This shows that after the Crash of 2008, long-term unemployment, defined as people unemployed for 27 weeks or longer, sharply increased after the Crash and throughout the Great Recession. The question is what this abnormally high level of long-term unemployment portends.
Krugman implies that the problem is simply one of job creation. Keynes made this same mistake, expressly assuming that full employment would bring full prosperity because at full employment labor output would be optimal. But putting aside some messy qualifications to that assumption, the discussion of which has barely moved beyond the early thought-provoking arguments of the English economist William H. Hutt in The Theory of Idle Resources (1939), this assumption becomes impossible to make when we take a deeper problem — growing income and wealth redistribution — into account.
With growing income and wealth concentration, there occurs a reduced ability to achieve what had previously been regarded as “full employment” — a new “full employment” condition is established with higher unemployment than before, and wages and salaries across the bottom 99% are reduced, so that previous levels of comfort and prosperity become un-achievable.
From a demand-side perspective, lower demand and reduced employment go hand-in- hand. Growing unemployment is a symptom of declining bottom 99% wealth and income: Unemployment is high for the bottom 99% in deep recessions or a depression because less money is available then for sufficient consumer demand, and the jobs to meet such demand are simply no longer there. Keynes overlooked the redistribution problem, and now Krugman is making Keynes’s mistake all over again: The problem is not failing to address unemployment — the problem is failing to address inequality.
The Fed Study
The authors of the Fed study, I must note, do not, in so many words, attribute the decline they found in potential GDP solely to labor force problems. However, as their discussion unfolds, it becomes apparent that they envision economic decline generally to be a consequence of lower employment, and continuous high long-term unemployment generally to result from of “structural” labor market constraints (i.e., the difficulties associated with matching unemployed people with new jobs). Their opening summary begins:
The recent financial crisis and ensuing recession appear to have put the productive capacity of the economy on a lower and shallower trajectory than the one that seemed to be in place prior to 2007. Using a version of an unobserved components model introduced by Fleischman and Roberts (2011), we estimate that potential GDP is currently about 7 percent below the trajectory it appeared to be on prior to 2007.
The Fed study proceeds under the presumption that declining income and growing unemployment are typically short-term phenomena attributable to shocks experienced in financial crises. (This is the grist in the Reinhart/Rogoff econometric mill, and a common neoclassical theme.) The economy is normally expected to recover fairly quickly from a recession, as it generally has in the past. The problem is, however, the Great Recession and its aftermath have been much deeper and longer lasting than all others since WW II, and this fact has turned the economics world upside down looking for explanations.
The Fed study’s explanation for the prolonged and deep unemployment is revealing. Focusing on long-term unemployment, the authors hypothesize “labor market damage.” They include a version of the FRED graph, but only going back to 1980 (p. 78):
This shows a base level of long-term unemployment of about 0.5%-0.8%, occurring just before recessions, until 2008. Their discussion of this development is brief:
As shown in Figure 2.8, the number of individuals unemployed for more than 26 weeks as a share of the labor force rose to 4.3 percent in April 2010 and has since fallen only to 2.7 percent, as compared with ¾ percent in 2007; likewise the share of the unemployed who have been out of work more than 26 weeks peaked at about 45 percent in early 2011 and remains above one-third today, well above the levels experienced during any previous post-World War II recession. Long-term unemployment is of particular concern because individuals out of work for extended periods of time may find that their skills, reputations, and networks deteriorate, resulting in a persistently higher level of structural unemployment or a steeper downtrend in the labor force participation rate. Although such effects do not appear to have been important in the United States in the past, they have been evident in other advanced economies and the unprecedented durations of unemployment during the present episode in the United States may reduce the relevance of historical experience in this country. (Emphasis added, p. 22)
At this point, we need to pause, and ask ourselves why the authors of the Fed study chose to focus on long-term unemployment and discuss structural unemployment. Their supply-side ideology, they no doubt have presumed, provides an adequate explanation for the huge jump in total unemployment, which ballooned to 10%: This was, after all, the worst market crash in history and the worst financial crisis since WW II, and their models show a 7% decline in output over the last five years.
What they are looking for, however, is an explanation for the lack of a more rapid recovery. There really are only two possibilities: (1) Either we’re in a gradually developing depression — and, unlike Keynesian demand-side theory, supply-side theory provides no explanation for a depression, or; (2) there is a supply-side factor prolonging the eventual recovery to full employment and prosperity — and the only supply-side factor that could prevent the return to full employment, under the supply-side assumption that potential GDP has not been materially compromised and the needed jobs are for the most part out there waiting to be filled, would be structural unemployment.
Attempting to blame the continuing level of high long-term unemployment on structural unemployment, however, encounters an immediate snag: Since structural unemployment has not been an important factor following past recessions in the United States, why should it suddenly become a major problem after the Crash of 2008? The authors don’t know, and can only suggest that the historical experience may somehow have become less “relevant.”
Their discussion turns to the matter of “permanent job loss,” that is, the record of jobs that are permanently eliminated, as shown by the red line on the following graph (p. 74):
The discussion of permanent job loss, however, only further undermines their argument: The implicit reasoning here is that if the degree to which jobs are permanently eliminated hasn’t gone up much, the higher number of long-term unemployed must be due to a higher level of structural unemployment. Put another way, the economy hasn’t shrunk, it’s just being restrained by a more inefficient labor market. But the problem remains: Why would the labor market suddenly be so much more inefficient?
The Fed study builds its argument concerning “long-term damage to the labor market” (p. 16) around the concept of “hysteresis,” apparently defined as “a permanent increase in structural and frictional unemployment and a higher natural unemployment rate” following a significant increase in unemployment, when “unemployed resources are permanently changed, through loss of job skills or seniority, making them less employable” (here).
We can think of the “natural” rate of unemployment as the amount of unemployment associated with full employment and, frankly, I think the version of hysteresis that associates the natural unemployment rate with structural unemployment is something of a manufactured concept. Hysteresis used to mean the level of unemployment below which inflation increases due to increasing consumer demand (here). 
The Fed study argues that “permanent separations of any type” could involve substantial relocation costs and a slow pace of finding new jobs:
Figure 2.4 shows that the rate of permanent job loss—the red line—rose sharply during the recession, briefly reaching a level more than twice as high as it reached in the aftermath of the relatively mild recession during the early 2000s and as high as that during the 1982 recession. Although the rate of permanent job loss has trended downward during the past four years and is currently close to its pre-recession level, the stock of persons still unemployed following a permanent job loss (the black line) remains noticeably higher than prior to the recession. This suggests that many permanent job losers continue to experience difficulties in finding a new job, consistent with the hypothesis that structural unemployment may have increased. [fn. omitted]. (p. 18)
There are unanswered questions about what data, exactly, the red line tracks and how the permanency of job loss is determined. The Fed study also fails to explain why the “permanent separation” factor would affect the slow decline from this enormous increase in long-term unemployment when its own data show no material increase in permanent job losses.
Regardless, the continuing high level of long-term unemployment has a much better, and more obvious, explanation: The economy — to be more precise, the bottom 99% economy within which jobs are offered and filled — has shrunk. With total unemployment jumping to 10% after the Crash, and only gradually falling thereafter, long-term unemployment had to increase, and then fall gradually as well. There is an adequate demand-side explanation for the failure of both total and long-term unemployment to rapidly return to normal: With the collapse of the housing bubble, there was an estimated loss of $10.2 trillion of wealth during 2008 ($3.3 trillion in home equity value and $6.9 trillion in shareholder wealth, according to Business Insider, 2/3/09, here). Consumer spending necessarily declines with such a massive reduction in net worth, so the most likely explanation for the huge spike in unemployment after the crash is that the economy was much weaker and fewer jobs were available.
Supply-side economics has trouble recognizing such higher unemployment as simply a reflection of garden variety stagnation, because it does not comprehend the effects of reduced consumer demand resulting from a huge increase in unemployment. But neo-Keynesian modeling stumbles here as well, as will become clearer as we work through this in Part 2 of this post, because aggregate GDP data includes the huge growth of top 1% income, masking the decline in income (and jobs) in the bottom 99%.
In dubiously seeking to lay off the continuing high rate of long-term unemployment mainly on “structural unemployment,” the Fed study seems to deflect the possibility that the recession isn’t over, as it is officially claimed to be, and to avoid discussing the possibility of a still-deepening stagnation. It seems to me, however, that the report reflects genuine confusion, because this sort of long-term effect is not what supply-side theory contemplates to result from a “financial shock.” Regardless, the discussion of the hysteresis issue and labor market damage mysteriously peters out, ending ambiguously:
Our analysis on this point [labor market damage] suggests that there has been a modest rise in the natural rate of unemployment and a steepening of the downtrend in labor force participation in recent years, but the evidence on the likely persistence of labor market damage is less conclusive. (p. 3)
In these circumstances, however, the authors still felt compelled, if reluctantly, to visit “the somewhat blurred distinction between ‘supply’ and ‘demand’ shocks” (p. 3):
We go on to argue that a significant portion of the recent damage to the supply side of the economy plausibly was endogenous to the weakness in aggregate demand—contrary to the conventional view that policymakers must simply accommodate themselves to aggregate supply conditions. (p. 1)
This is a significant concession for conventional mainstream economists, but one they apparently found unavoidable:
The fallout from these events on credit availability, balance sheets, and confidence continues to weigh on aggregate demand, restraining the pace of recovery in the housing market, firms’ willingness to hire and invest, and spending by consumers and state and local governments. In addition, these demand effects have probably diminished the productive capacity of the economy. (p 2)
Potential GDP and Monetary Policy
The “conventional” neoclassical position, as noted at the top, denies that consumer demand has any material relevance to growth or decline, so it is notable that the Fed report challenges that belief somewhat:
In many macroeconomic models, aggregate supply shocks are viewed as exogenous—and specifically as outside the range of influence of monetary policy. However, if—as we suggested earlier—some elements of aggregate supply are significantly influenced by changes in aggregate demand they may also be susceptible to influence from monetary policy. . . [But] demand shocks can also have long-lasting effects on unemployment duration and labor force attachment that, in principle, activist monetary policy might be able to check. And finally, demand shocks and monetary policy may even be able to influence potential output over the medium term through their effects on new business formation and research and development. (p. 3)
They conceded up front that their argument “that a significant portion of the recent damage to the supply side of the economy plausibly was endogenous to the weakness in aggregate demand” is “contrary to the conventional view that policymakers must simply accommodate themselves to aggregate supply conditions.” (p. 1)
Significantly, conceding that demand “shocks” can affect employment and capital formation establishes a rationale for stimulating consumer demand, and brings Keynes back into the conversation. In Keynesian theory, expectations of future demand are a crucial component of the estimation of expected returns on capital investment, and therefore crucial to investment decisions, which is why Keynes suggested that expansionary monetary policy might help stimulate a sluggish economy.
Since the Crash of 2008, however, monetary policy hasn’t worked, as Paul Krugman has tirelessly pointed out in his column. There has been a continuing “liquidity trap” even with the prime interest rate approaching zero and, the Fed furiously pumping more money into the economy, a policy that is under continuous review. Not surprisingly, the Fed study makes no claim that monetary policy will actually work, and refuses even to make an educated guess:
Thus, in an uncertain world, a policymaker’s choice of policy will depend not only on the extent to which he or she believes a demand shock is likely to affect potential GDP and employment, but also on his or her view of the risks associated with actively trying to offset these adverse supply-side developments through accommodative monetary policy. (p. 4)
What more could these researchers say, considering the extreme uncertainty they identify in their results as well as the results of others? For example:
While other analysts also marked down their estimates of potential output growth [in 2010], the timing and the extent of these markdowns varied considerably (Table 1.2). * * * In contrast, U.S. officials were somewhat slower to recognize the decline in potential output growth. * * * Of course, considerable uncertainty attends all of these estimates of potential output growth and the natural rate of unemployment. (p. 12)
The comparable confidence band around the estimated natural rate of unemployment (middle-left panel) ranges from about 4½ percent to 7 percent. Moreover, these ranges undoubtedly understate the true uncertainty surrounding our model-based estimates as they do not account for uncertainty about data revisions, the specification of the state-space production-function model, or the possibility that other altogether-different frameworks might yield different estimates of supply-side damage. (p. 12)
Not surprisingly, frameworks that differ from ours in more fundamental ways can also yield different estimates of potential output and economic slack. (p. 14)
Looking forward, the trajectory of potential GDP is even more uncertain. (p. 14)
This your-guess-is-as-good-as-mine perspective is not the stuff of a blockbuster paper. Indeed, the Fed study’s conclusions could not be more equivocal:
In the labor market, matching efficiency seems to have been somewhat impaired, the natural rate of unemployment appears to have risen somewhat, and trend labor force participation appears to have moved noticeably lower relative to what would have been expected based on pre-crisis trends. In addition, the capital stock and trend multifactor productivity are appreciably lower than what would have been predicted in 2007. Our point estimates suggest that, in combination, these developments—whose eventual magnitude was arguably apparent only in hindsight—shaved almost 7 percent off the level of potential output relative to its pre-crisis trend. That said, the uncertainty about this estimate is extremely high and the implications for future growth are quite uncertain (pp. 51-52).
Translation: It looks like the economy is 7% smaller than before the crash, but we’re not really sure.
[P]olicymakers may appropriately be restrained from pursuing a highly aggressive response to a deep recession if they fear the attendant risks to financial stability, or are concerned that inflation expectations may become unanchored. More generally, the pervasive uncertainty in which policymakers operate may encourage them to proceed with caution. (p. 52)
Translation: Since we don’t really know what’s going on or where we’re headed, maybe it would be best not to rock the boat too much.
Today I Googled the Fed paper, and found no mention of it in the press. I’m not surprised. Anyone in the mainstream press taking a look, curious about what was scaring Krugman so much, would scarcely know what to make of this study, or even be able to figure out what it says. Credit the authors with an outstanding job of keeping their report under wraps. The Fed is trying to build confidence in our economy, so it would hardly do to provoke headlines like “Fed study finds the Great Recession isn’t over,” or “Fed study reports 7% decline since the crash.”
Regardless, the Fed study has demonstrated a lot about the inadequacy of supply-side neoclassicism and its modeling. If as Raj Chetty argued in his recent article (cited above) — “yes, economics is a science” — is it too cynical to suggest that a new paradigm capable of producing something better than mass uncertainty is long overdue? Beyond that, is it too cynical to conclude that institutional and psychological barriers to correcting our decline are so powerful that removing that uncertainty would accomplish nothing? I hope not.
In Part 2 of this post I’ll discuss what such a new distributional perspective might look like. It’s clear to me that we can’t lay these problems entirely off on the supply-side neoclassical paradigm. Within the narrow range of conditions in which Keynesian policy could work, his demand-side theory operates reasonably well; but demand-side theory, in the hands of the neoclassical hybrid “neo-Keynesianism,” likely fares little better than supply-side modeling itself. Part 2 will begin by highlighting evidence for a distributional paradigm embedded both in the Fed study and in Krugman’s response to it.
To Paul Krugman and others who find evidence of significant “permanent” economic decline terrifying, I say: “Welcome to the club!” I have felt this terror for over three years. Increasingly, I am not alone.
JMH — 11/15/13 (“The Fed Report” edited 11/18/13; ARC edits 12/11/13)
 Ironically, I discussed that concept in a paper on the incompatibility of full employment and price stability I wrote for a labor seminar in Ann Arbor in 1970. I don’t remember exactly what I wrote, but I do remember the thrill of meeting with guest reviewer Gardner Ackley, a former Chairman of Economics Department at the University of Michigan, a member of Kennedy’s Counsel of Economic Advisers, and Chairman of the CEA under Lyndon Johnson, shortly after his return from Italy.