Dr. Jan Eberly
Although the unemployment rate has moved down over the past year, it remains more than 3.5 percentage points above its average in the years before the crisis.
There is an ongoing debate among economists and policy-makers as to whether the high level of unemployment is due to weak demand or structural factors, such as unemployment benefits themselves or a mismatch between worker skills and the needs of employers. As we demonstrate below, the evidence points to cyclical demand-driven factors as the main impediment to reaching full employment. Structural factors explain at most a small portion of the unemployment challenge we face today.
That is why extending the payroll tax cut and unemployment insurance benefits makes sense both for families and for the economy. Not only do these policies put money in the hands of families struggling to cope in difficult times, but independent analysts have estimated that these programs will provide a substantial boost to demand – on the order of ¾ percentage point – at a time when our recovery is far from complete. While policies that combat structural unemployment also deserve consideration, the immediate focus should remain on strengthening demand and economic growth to help prevent what is predominantly a cyclical challenge from becoming structural.
The Current Unemployment Situation
As can be seen in Chart 1, the unemployment rate has been moving down over the past year but remains at a very high level. 1 Although the unemployment rate was higher in the 1980s, reaching a peak of 10.8 percent, the unemployment rate outside of recessions was also higher at that time, at least in part because there were so many young people in the labor market. 2 At present, the unemployment rate remains about 3.5 percentage points above the average rate from the mid-1990s until the start of the recession.
Meanwhile, long-term unemployment – measured as the number of individuals unemployed for more than 26 weeks – is still at 3.7 percent as a share of the labor force and near the record high it reached soon after the crisis.
Source: Current Population Survey (CPS) (Bureau of Labor Statistics via Haver Analytics).
As we have shown in previous posts, GDP remains well below potential
, capacity utilization is low
, and surveys of businesses and economists cite a lack of demand
as a key impediment to growth. Data from the labor market are also consistent with a demand-side explanation for the high unemployment rate. Starting with survey evidence, the Conference Board consumer confidence survey shows that the share of individuals responding that jobs are hard to get has not improved much after reaching its peak two years ago. And although the share of small firms reporting they have a hard-to-fill position has moved up since its low during the recession, it remains lower than at any other time since the early 1990s. This does not suggest a labor market plagued by a mismatch between worker skills and the skill requirements of available jobs, but rather one in which jobs are still scarce – an inference we explore in more detail below.
Source: The Conference Board Consumer Confidence Index and National Federation of Independent Business Small Business Optimism Index (via Haver Analytics).
Those who advance the case that structural unemployment has increased note that this recession has hit certain industries, such as construction and manufacturing, particularly hard. They argue that many workers in those industries must now move into others, for which they may not have appropriate skills. However, this line of reasoning ignores the fact that in every recession some sectors of the economy are hit harder than others. This phenomenon can be seen in Chart 3 below, which measures sectoral reallocation as the difference in employment growth across industries.
3 The chart clearly shows that sectoral reallocation is a common feature of recessions. The fact that this recession was accompanied by a particularly large amount of reallocation is not surprising given the severity of the recession. Moreover, as of earlier this fall, the level of sectoral reallocation had fallen back to about pre-recession levels, again similar to previous episodes, suggesting that the expansion in employment has not been particularly uneven. Since the sectoral reallocation in this recession and recovery has been in line with historical experience, and since recessions are not always associated with an automatic increase in structural unemployment, this evidence does not suggest that this recession has produced an unusual divergence between job requirements and worker skills.
The Relationship between Vacancies and Unemployment
Source: Kuang & Valletta (2010).
Data on job vacancies can also be useful for examining the potential for structural unemployment. According to the Job Openings and Labor Turnover Survey (JOLTS) data as of October, there were more than four unemployed people for each job vacancy, well above the average ratio of unemployed people to job vacancies of less than 2-to-1 in the years before the crisis.
The Beveridge curve provides a formal mapping of the relationship between unemployment and vacancies. In chart 4, the blue dots map points from 2000 to 2008. In comparison to the relationship suggested by these points, more recent readings suggest a higher unemployment rate for a given level of vacancies. The relationship between vacancies and the unemployment rate can change for a variety of reasons, including cyclical factors and demographics, as well as a change in the ease with which workers find jobs, known as match efficiency in economic parlance. Among the factors that might cause deteriorating match efficiency are a misalignment between the skills of workers and the needs of employers – in other words, a decrease in match efficiency could be a signal of an increase in structural unemployment.
To determine how much of the shift is due to a decrease in match efficiency as opposed to changes in labor supply or demand requires careful statistical analysis. Estimates vary somewhat, but the most careful research suggests that a deterioration in match efficiency (and hence a rise in structural unemployment) explains no more than perhaps a percentage point of the increase in the unemployment rate, including any effect of long-term unemployment benefits (discussed further below) on match rates. 4
Source: Job vacancies from the Job Openings and Labor Turnover Survey; Unemployment Rate from the Current Population Survey (Bureau of Labor Statistics via Haver Analytics).
The Relationship between GDP and Unemployment
The fact that the unemployment rate has moved about as much as would be expected by the movements in GDP, a relationship known as Okun's law, is another piece of evidence supporting a demand-driven explanation for the high unemployment rate. Chart 5 plots changes in the unemployment rate against the growth in GDP. The line maps the average relationship between the two variables over time. As can be seen, the cumulative error for this recession and recovery (that is, adding the errors for 2007-2011) is small and well within the pattern of historical errors.
Source: Unemployment rate data from the Current Population Survey; GDP data from the National Income and Product Accounts (Bureau of Labor Statistics and Bureau of Economic Analysis via Haver Analytics).
Has Emergency Unemployment Compensation Caused Higher Unemployment?
One criticism leveled against the extension of emergency unemployment compensation (EUC) is that the program itself might have pushed up the unemployment rate. A number of studies that have examined the experience of the past few years – consistent with many previous studies – have found relatively small effects. For instance, Farber and Valletta find that eliminating extended benefits would reduce the expected duration of unemployment by less than a day. Even for those unemployed longer than six months, they estimate the expected duration of unemployment would be reduced by only about a week. Similarly, Rothstein finds that the emergency unemployment compensation program has pushed up the unemployment rate by only a small amount. 5 Moreover, these estimates do not take into account the positive effect of unemployment insurance, including keeping people in the labor force, job creation, and growth.
Most importantly, both these studies also find that to the extent emergency unemployment compensation increases unemployment, it does so largely by reducing exit from the labor force rather than exit to employment. In other words, even these modest program-induced increases in unemployment result primarily from the fact that the benefits encourage individuals to continue searching for work, rather than turning down available jobs, so that they can continue to receive benefits. These findings are not particularly surprising given the weak labor market and the relatively low benefit level. The average weekly emergency unemployment compensation benefit is $293, equal to about 46.2 percent of the previous weekly wage earned by UI beneficiaries. To put this in context, if a person were to receive this benefit for an entire year, it would be worth $15,236, well under the poverty rate for a family of four. Given the opportunity to earn more in the workforce and to gain both work-related skills and other nonpecuniary benefits, people strongly prefer employment to unemployment benefits.
Some studies that have used data from previous episodes to calculate the impact of emergency unemployment compensation on the unemployment rate have found somewhat larger effects. However, these computations, which use parameters estimated in very different economic conditions, are likely to overstate the impact. 6 Mazumder (2011) uses an estimate of the impact of UI extensions on the duration of unemployment estimated from earlier episodes, but attempts to account for current policy and labor market conditions. His analysis suggests that the emergency unemployment compensation program can account for 0.8 percentage point of the increase in the unemployment rate, which, given the limitations of the approach, would appear to be an upper bound.
Overall, these data suggest that structural factors such as the mismatch between worker skills and job requirements and the unemployment compensation program, have contributed only modestly to the unemployment crisis, which is primarily due to a lack of demand. Even if structural factors were to explain 1.0 percentage point of the excess of unemployment, GDP would still have to rise in the neighborhood of 8 or 9 percent in order to eliminate the unemployment rate gap.
Additional policies to prevent structural unemployment also deserve consideration, because the longer the unemployment crisis persists, the larger an issue structural unemployment is likely to become as unemployed workers' skills become outdated and their attachment to the labor force weakens. For that reason, in addition to policies that strengthen near-term demand, the President has proposed a number of initiatives aimed at easing the transition back to work. These include a plan to reform the UI system that encourages training and employment, as well as support for sector-based training and other programs to improve employee qualifications and make hiring easier. We hope that these plans will receive serious consideration given the potential consequences of long-term unemployment. However, the more immediate focus should remain on boosting demand and therefore growth through policies like extending the payroll tax cut and emergency unemployment compensation. Stronger growth will help prevent a predominantly cyclical challenge from becoming structural.
The Effect of the Payroll Tax Cut and Emergency Unemployment Compensation on Growth
The nonpartisan, independent Congressional Budget Office (CBO) recently reported that increasing aid to the unemployed is the most effective program out of the 13 policies it analyzed for boosting economic and job growth, because families who receive the money spend almost all of it. The CBO estimates that every dollar spent on increasing aid to the unemployed generates up to $1.90 of additional GDP. Private forecasters estimate that extending emergency unemployment benefits alone will boost GDP growth by ¼ percentage point in 2012. The CBO also finds reducing employees' payroll taxes to be an effective stimulus, and private forecasters project that extending the 2 percentage point payroll tax cut will add approximately ½ percentage point to GDP growth in the coming year.
Taken together, these data paint a picture of a challenging labor market, suffering from a lack of demand in the economy. In this environment, extending the payroll tax cut and emergency unemployment insurance not only provides much-needed assistance to workers and their families in a difficult economy but also provides a boost to the broader economy at a time when the recovery is far from complete.
Dr. Jan Eberly is the Assistant Secretary for Economic Policy.
1. In 1994 the Bureau of Labor Statistics (BLS) implemented a redesign of the Current Population Survey (CPS). Evidence from the BLS suggests that the pre-redesign survey understated the share of the unemployed experiencing a spell of 15 weeks or longer by 17 percent. They do not provide a comparable statistic for the proportion of unemployed workers in a spell lasting over 26 weeks. However, it seems likely that the difference between the current episode and previous episodes is somewhat overstated. For more information see Polivka, Anne E. & Stephen M. Miller (1998) "The CPS After the Redesign: Refocusing the Economic Lens" in Haltiwanger, John, et al. eds. "Labor Statistics Measurement Issues,
NBER: Cambridge, Massachusetts, pages 249-289.
2. See for instance the discussion in Stiglitz, Joseph (1997) "Reflections on the Natural Rate Hypothesis," Journal of Economic Perspectives Vol. 11, No. 1 (Winter).
3. The measure is a simple "Lilien" dispersion index: , where sit is industry i's employment share at time t; git is employment growth at time t; and gt is total employment growth at time t. The chart is from Valletta, Rob and Katherine Kuang "Is Structural Unemployment on the Rise?" FRBSF Economic Letter, November 8, 2010, updated by the authors.
4. Barnichon, Regis and Andrew Figura. 2010. "What Drives Movements in the Unemployment Rate? A Decomposition of the Beveridge Curve" Finance and Economics Discussion Series No. 2010-48, updated in October 2011 by the authors.
5. Farber, Henry and Rob Valletta. "Extended Unemployment Insurance and Unemployment Duration in the Great Recession: The U.S. Experience, June 24, 2011. Specifically, Farber and Valletta find that the expected duration of unemployment would be reduced from 2.16 to 2.15 months for all the unemployed and from 12.61 to 12.38 months for the long-term unemployed. Rothstein, Jesse, "Unemployment Insurance and Job Search in the Great Recession." October, 16, 2011. An earlier version of this paper was presented at the Brookings Papers on Economic Activity, Fall 2011 conference.
6. Mazumder, Bhashkar, "How did unemployment insurance extensions affect the unemployment rate in 2008-10?" Chicago Fed Letter, April 2011, 285.