The Wrong Time to Extend Unemployment Insurance

Report Jobs and Labor

The Wrong Time to Extend Unemployment Insurance

May 4, 2004 19 min read Download Report

Authors: Paul Kersey and Tim Kane

Although the perceived weakness in the labor market has led to calls for another extension of temporary unemployment insurance (UI) benefits, this is a political cure in search of a problem. Resuming Temporary Extended Unemployment Compensation (TEUC) is ill-advised.

An extension not only would increase spending by roughly $1 billion per month, but also is unnecessary given the demonstrated strengthening of the labor market. Most damning of all is that well-established economic research shows that any extension of jobless claims generates exactly the wrong incentives and increases the average duration of unemployment for those who qualify.

The economic context surrounding this debate is much more encouraging than the overheated political rhetoric suggests, especially when one considers these facts:

  • Initial jobless claims are historically low. The magic number on Wall Street is the initial claims number of 400,000 per week: Anything higher means a contracting labor market. Initial claims are declining and have been below 400,000 since October 2003. The number of claims relative to the total U.S. population is even more striking (See Chart 1.) The current level of population-weighted claims is at the same level as 1997-1999, when the economy was "irrationally exuberant."
  • The rate of unemployment is now 5.7 percent--far below the rate when the original TEUC was cancelled under President Bill Clinton.
  • The labor force is growing, not shrinking. The idea of a shrinking or discouraged workforce is a myth: 1 million people joined the labor force in 2003, and total employment increased by a net 1.34 million workers during that same period, which means that the ranks of unemployed declined by 340,000.
  • The Bureau of Labor Statistics household survey shows that over the past two years (March 2002 through March 2004), the number of employed Americans increased by nearly 2.2 million--robust job growth that renders comparisons between the current economy and the Great Depression absurd.
  • Although the BLS payroll survey may be grossly underestimating job creation, it shows seven consecutive months of job growth, with 308,000 in March alone. These figures are consistent with a solid economic recovery that is creating jobs and putting Americans back to work.

The facts of a strong economy cannot obscure the fact that many workers are suffering, but policymakers should keep in mind that free markets constantly experience business failures and worker dislocations, even in the best of times. The current labor economy is not in a recovery so much as in a restructuring boom. Most workers gain, but a few suffer acutely from permanent job loss as productivity converts factories full of workers into new factories full of automation.

As a consequence, the median duration of unemployment has been unusually high since April 2002. This is arguably the single valid reason to extend UI, but even it fails to convince. The statistics on unemployment duration are exaggerated because new jobless claims are extremely low. With relatively few newly unemployed persons, the weight of long-term unemployed is proportionally higher. Nonetheless, the solution to longer periods of unemployment is not to extend benefits that have been shown to lengthen spells of unemployment.

Emergency Benefits

In spite of strong growth in payroll jobs during March 2004 and the positive adjustment for the months of January and February, not to mention the long-running jobs growth indicated by the household survey, it is likely that lawmakers will face continued pressure to extend unemployment insurance benefits. In the midst of a perceived (by some) weak labor market, some political and labor leaders have called for a further extension of TEUC. Members of Congress should resist this pressure. Resuming TEUC, or establishing any other emergency unemployment insurance benefits program, is ill-advised.

Unemployment insurance is provided jointly by the state and federal governments. States provide basic unemployment benefits, lasting up to 26 weeks, funded by taxes levied on employers. Extended benefits, lasting another 13 weeks, are funded jointly by state and federal governments and may be offered on a state-by-state basis when certain unemployment rate triggers are reached. Emergency benefits, funded entirely by the federal government, have been offered periodically to lessen the impact of joblessness during economic recessions.

The last such emergency benefits program was the TEUC program, which took effect in April 2003 and was allowed to expire in December. Workers who qualified for TEUC before that date may still be receiving TEUC payments, but these are scheduled to expire in April.

The current unemployment insurance system, while certainly well-intentioned and valuable, negatively affects both employers and the unemployed. Providing further emergency benefits can aggravate unemployment and slow down an economic recovery because those benefits weaken the incentives to search for and accept employment. These incentives are less important during a recession, when jobs are scarce, but become more important as the economy and employment market recover and jobs become available.

As both the household and payroll jobs figures indicate, the labor market has been recovering since the recession ended. Under these conditions, emergency benefits are unnecessary and probably counterproductive.

Basic Purposes Behind the UI System

The basic impulse behind unemployment insurance is sound: Maintain some stability for workers, and their families, who find themselves without a job through no fault of their own. Extended and emergency benefits are advocated on similar grounds: When economic conditions are such that workers cannot find appropriate jobs despite their best efforts, some provision should be made for their continuing economic needs.

However, unemployment insurance was never intended to serve as a permanent entitlement program. The unemployment insurance system is set up so that benefits represent a form of compensation for past employment and are limited to those who intend to remain in the workforce. For this reason, benefits are denied to those who leave work voluntarily or are fired for cause, and continued eligibility is contingent on recipients' continuing to search for work.

Can Unemployment Insurance Stimulate the Economy?

Unemployment insurance benefits are often justified on the grounds of economic stabilization and stimulus. The Economic Policy Institute continues to advocate for extended unemployment benefits, claiming that:

Unemployment insurance not only helps workers and their families during a time of need, but it also helps the economy. When workers become unemployed during an economic downturn, people have less money to spend on consumer items, which further hurts the economy. Unemployment insurance puts money in families' pockets, and when they spend it, the economy gets a boost.1

This special-interest advocacy fails Economics 101 on three counts. First, an individual's consumption behavior is widely known to follow his or her lifetime income path, not temporary income swings. Second, this simplified Keynesian "demand stimulus" view of unemployment insurance does not fully account for its effects on government finances. Third, the argument assumes the economy needs stimulating, which defies credibility with the growth in gross domestic product (GDP) averaging 6.1 percent annually over the past two quarters.

On the broadest issue of overspending, one must question the effectiveness of any program based on costs and benefits. Policymakers know that extending UI is expensive--up to $1 billion per month. Almost certainly it does not provide an equal amount of stimulus: Directly, it generates incentives not to work. Indirectly, the payments must be paid for by higher deficits.

During periods of low unemployment, revenue from unemployment taxes exceeds benefit payments, allowing state and federal governments to build a reserve that can be used to fund other government programs. In periods of high unemployment, governments restore funds to the UI program in order to pay UI benefits. Thus, paying unemployment insurance has a self-canceling effect because the additional money received by the unemployed is matched by increased pressure on government to borrow or tax to cover the additional cost.

Admittedly, to say that UI benefits have no effect would be a mistake. They do provide economic stability to families of unemployed workers on an individual level and, in the process, provide some stability to businesses that those formerly employed workers had patronized. However, on a macro scale, the impact is too small to make an impact.

The key date to watch was April 29, 2004, when the GDP growth figures of 4.2 percent were released for the first quarter. The expansionary surge was confirmed, neutralizing the case for UI.

Making Unemployment Last Longer

Unemployment insurance also makes unemployment spells last longer. By making unemployment more attractive--or at least less unattractive--than it would otherwise be, UI benefits tend to increase the "reserve wage" of unemployed individuals who are considering new job offers--a possible employer will have to offer a higher wage or some other inducements before a new job is accepted.

In theory, this is not a negative. If the additional time is spent in an active job search, and if that additional time yields additional job offers, the benefits recipient will likely find a job that better matches his or her skills and material needs--boosting wages, productivity, and job stability. However, recent research indicates that the additional time for job search does not necessarily result in higher wages upon re-employment.2 In other cases, unemployment insurance benefits may entice recipients into becoming complacent in their job search, relying on benefits to "get by" until they are recalled by their prior employer, delaying a serious job search until their benefits are nearly exhausted.

For good or ill, there is little doubt that UI benefits tend to lengthen unemployment. Working with Department of Labor data covering the employment histories of young men in the late 1970s and 1980s, Professors Mark Gritz and Thomas MaCurdy found that the mean duration of unemployment was 14.6 weeks for those who received unemployment insurance benefits versus 7.6 weeks for workers who were eligible but did not file a claim and 6.9 weeks for those who were ineligible for unemployment benefits.3

Research has also shown that a UI recipient's likelihood of finding a job rises dramatically as exhaustion of benefits nears. Economists Lawrence Katz of Harvard University and Bruce Meyer of Northwestern University reviewed income and employment data from the Panel Study of Income Dynamics, collected by the University of Michigan. They found that an unemployed person's likelihood of finding work tapered off in the early weeks but climbed from the 20th to the 26th week as exhaustion drew near.4

Dr. Meyer later refined this method, working with the Continuous Wage and Benefit History, a compilation of federal and state unemployment insurance records, and counting weeks back from the expected date of exhaustion of benefits. Meyer found a more pronounced spike in likelihood of finding work in the weeks leading up to exhaustion. He also found that the mere expectation that benefits would expire increased the likelihood that an individual would find work in a given week, even if further benefits were in fact available.5

Unimployment Insurance Affects Employers As Well

Recipients are not the only parties affected by unemployment insurance benefits. Employers appear to consider the duration of unemployment benefits when deciding to lay off and recall workers.

Along with a UI recipient's likelihood of finding a new job, Katz and Meyer also found that a worker's likelihood of being recalled by his or her former employer also increased in the last few weeks before benefits were exhausted. This suggests that employers may rely on UI to keep their workforces from seeking other jobs during layoffs, making layoffs more attractive to firms that are experiencing a drop-off in business.6


Delaying Necessary Changes

Finally, lawmakers should consider the effect of further unemployment benefit extensions on restructuring within the national economy. America is rightly acknowledged as the world's wealthiest economy, an achievement due in large part to its robust economic freedom, especially flexibility in labor markets. During the typical quarter, the American economy eliminates 7.71 million jobs while at the same time creating another 8.11 million.7

Even the healthiest of times will see some workers losing jobs as new industries and methods make old facilities obsolete. Ideally, new opportunities and employment openings will come along quickly, keeping unemployment spells short. However, to take advantage of the new job openings, workers may need to learn new skills, look into new companies and industries, and even move to another area. The alternative is an economy frozen in time, losing no jobs but completely stagnant in incomes and declining in competitiveness.

The employment dynamics of a strong economy can be an unsettling process, and workers understandably resist disrupting changes. However, when restructuring is permanent, the sooner the new job quest begins in earnest, the sooner it will end with a new work opportunity. Extended unemployment benefits allow many workers to delay accepting new jobs, or even looking for work. In other words, a temporary change in benefits creates perverse incentives that are potentially destructive to the long-term incomes of the poorest Americans.

In the midst of a high-unemployment economy in which jobs are simply unavailable, extended unemployment benefits may ease the difficulties faced by those who cannot find work. However, Members of Congress should remember that extended benefits also aggravate all of the negative consequences listed above: Recipients can put off a serious job search and hold out longer for higher-paying jobs even when suitable employment is available. At the same time, firms may keep workers laid off for longer periods without worrying about losing personnel to other companies.

Lawmakers should balance their legitimate concerns for the well-being of families with a healthy respect for the dangers inherent in paying workers while they are not working. In balancing these factors, lawmakers should start with an accurate and complete picture of the state of the economy, particularly the job market.

The Context for New Benefits

The case for extending benefits rests on the premise of a weak labor market, but the numbers speak for themselves. Jobless claims have fallen for the past six months to a level 10 percent below the historical average. Once claims are considered as a percentage of the population, the current level of new job losses is clearly back at the level of the strong economy of the late 1990s.8

The larger picture of labor markets shows real strength. Unemployment is low at 5.7 percent and has declined for six of the past nine months. The notion that Americans are discouraged and leaving the labor force in unusually large numbers has little basis in fact.9 The dynamics of a declining unemployment rate are driven by growth in the labor force of 2.4 million people over two years and growth in the number of employed workers by nearly the same amount. (See Chart 2.) Since June 2003, the number of unemployed Americans has declined by nearly a million.


Output per worker is higher in America than in any other advanced economy, exceeding most European countries by more than 10 percent. Over the past three years, this measure of labor productivity has accelerated even further in what the Congressional Budget Office calls "the most striking economic development of the last three years."10 From the third quarter of 2001 to the third quarter of 2003, productivity grew at an average annual rate of 5.6 percent, the highest rate over any eight quarters since 1950. Productivity continued to surge by an annualized 2.6 percent in the fourth quarter. The higher value of each labor hour has a positive effect on real wages, which have also grown in real terms by 3 percent over the past three years.

Only two indicators show weakness in the labor market. The first sign of weakness is the slow payroll jobs growth that prevailed through most of last year. During 2003, the payroll survey indicated that the American economy lost 61,000 jobs, but current research from The Heritage Foundation reveals that lost payroll jobs may have been a statistical illusion caused by miscounts and earlier overcounts in the payroll survey.11

Because the payroll survey focuses on firms' employment records, it does not count contract workers, such as consultants, even if they work primarily for one company and have a working relationship with that company similar to that of a regular employee. Changes in the employment turnover rate can also affect the payroll survey. When turnover--the rate at which workers change jobs--slows down, as happened after the September 11 attacks, payroll jobs will appear to shrink without any real change in the number of jobs available.

The higher-quality household survey showed a gain of 1.34 million jobs during the same year. This discrepancy between the two main measurements of employment casts doubt on the whole notion of a weak job market and a "jobless recovery."

The payroll survey has itself begun to indicate improving conditions in the job market, showing a gain of 308,000 jobs during March 2004. Payroll job figures for January and February have also been revised upward. If the current positive trend continues, there will be little reason to doubt that economic recovery has reached employment markets.

The only valid indicator of a weak labor market is the rising duration of unemployment spells. Many unemployed workers are suffering from permanent job loss. The median duration of unemployment has been nine weeks or longer--significantly higher than average--since April 2002. Manufacturing employment has declined by 3 million in five years, and even though some of these jobs have simply been recategorized as business services, many are real layoffs--painful and likely permanent.


The 2004 Economic Report of the President explains how this mirrors the experience of U.S. agriculture in the past century:

The evolution of U.S. manufacturing from 1970 to 2000 mirrors, in important respects, that of U.S. agriculture from 1940 to 1970. Total real farm output increased more than 60 percent from 1940 to 1970. Over the same period, employment in farming declined nearly 6 million...a decline in agriculture's share of total employment of 15 percentage points, from 19.4 percent in 1940 to 4.4 percent in 1970.12

Yet extending unemployment insurance benefits will not solve the problem of longer-duration unemployment. UI benefits have been shown to increase the length of unemployment. Workers laid off from old factories cannot afford to wait for their old jobs to return, because the new factories are using more automated capital equipment and fewer workers. As a result, the long spells of unemployment of many permanently displaced workers are driving up average unemployment durations.

This is not a sign that more temporary benefits are needed. It is quite the opposite: Extending benefits beyond 26 weeks would lead too many workers to wait for re-employment in their old lines of work.

The cure for that dislocation is faster growth in new sectors--something only private markets can meaningfully provide. Already, the demand for labor in the health and education sectors is far outpacing supply. There are jobs in the economy, just not in the old places.

Alternatives to TEUC

Decades ago, Congress established the jobless benefits programs, under which any worker displaced from a job can receive benefits for six months. Given the strong recovery in total employment, productivity, overall growth, and even real after-tax incomes, a temporary extension of benefits is the worst among many alternative options that Congress could consider.

The best option is to do nothing. A Congress that imagines every bill to be a jobs bill will eventually mimic the make-work charades of socialist government. It is a simple fact of history that governments are unable to create real jobs. Only a free-market economy that encourages entrepreneurs is capable of genuine employment growth.

The second best option is to experiment with alternative programs. States and communities should be free to try a variety of programmatic unemployment insurance reforms. One reform suggested by Nobel prize-winning economist Gary Becker would permanently extend jobless claims for a full year but not allow them to begin until two months into a jobless spell.

The White House has advocated personal re-employment accounts (PRAs), which essentially pay a bonus to unemployed workers who take new jobs quickly. Early results indicate that workers find equally good jobs much faster when PRAs are in place. However, caution is in order: PRAs should be designed carefully, particularly in terms of eligibility criteria.

Another idea worth more exploration is to privatize UI, or more precisely, to personalize it. A personalized UI system would give individuals some control over their unemployment insurance and allow them access to funds that they do not use, changing the incentives that surround UI and encouraging a speedier return to work.


Neither extending unemployment benefits nor any of these other options will create jobs, which all sides in this debate agree should be the goal. Fortunately, recent evidence, in the form of both payroll and household jobs statistics, indicates that the job market is improving.

In general, policies that create perverse incentives against hiring and seeking work--both of which are demonstrated marginal effects of unemployment insurance--are not helpful. Such policies are especially risky in the early stage of economic recovery. The bottom line, therefore, is that Congress should not extend unemployment benefits, but instead should maintain the regular benefit programs.

Paul Kersey is Bradley Visiting Fellow in Labor Policy in the Domestic Policy Studies Department, and Tim Kane, Ph.D., is Research Fellow in Macroeconomics in the Center for Data Analysis, at The Heritage Foundation.



1. Economic Policy Institute, "Unemployment Insurance: Frequently Asked Questions," at

2. Congressional Budget Office, Family Income of Unemployment Insurance Recipients, March 2004, p. 17.

3. Mark Gritz and Thomas MaCurdy, "Measuring the Influence of Unemployment Insurance on Unemployment Experiences," Journal of Business and Economic Statistics, Vol. 15, No. 2 (April 1997), pp 130-152, esp. p. 136.

4. Lawrence F. Katz and Bruce D. Meyer, "The Impact of the Potential Duration of Unemployment Benefits on the Duration of Unemployment," Working Paper No. 241, Industrial Relations Section, Princeton University, revised November 1988.

5. Bruce D. Meyer, "Unemployment Insurance and Unemployment Spells," Econometrica, Vol. 58, No. 4 (July 1990), pp. 757-782.

6. Ibid., Appendix, Table A1.

7. U.S. Department of Labor, Bureau of Labor Statistics, Business Employment Dynamics Data Series, 1992 to 2003.

8. For current unemployment claims data, see U.S. Department of Labor, Employment and Training Administration, Office of Workforce Security, "UI Weekly Claims," at (April 26, 2004).

9. Tim Kane, "Will the Real Unemployment Rate Please Stand Up?" Heritage Foundation WebMemo No. 456, March 24, 2004, at

10. Congressional Budget Office, "The Economic Outlook," Chapter 2, p. 5, in The Budget and Economic Outlook: Fiscal Years 2005 to 2014, January 26, 2004.

11. Tim Kane, "Diverging Employment Data: A Critical View of the Payroll Survey," Heritage Foundation Center for Data Analysis Report No. 04-03, March 4, 2004, at

12. Economic Report of the President, 2004, p. 77.


Paul Kersey

Former Visiting Fellow

Tim Kane