Testimony

Reforming Unemployment Insurance for the 21st Century Workforce

by Howard F. Rosen, Peterson Institute for International Economics

Testimony before the Income Security and Family Support Subcommittee, House Ways and Means Committee
March 15, 2007

 


Note: This statement is based on Lori Kletzer and Howard Rosen, “Reforming Unemployment Insurance for the Twenty-First Century Workforce,” The Hamilton Project ( Washington: Brookings Institution, 2006).


INTRODUCTION


CHANGES IN THE US LABOR MARKET

Over the last few years there have been changes in the nature of unemployment in the United States. After rising between the 1960s and the 1980s, the average unemployment rate began falling in the 1990s, reaching a low of 4 percent in 2000 and remaining moderate over the past six years (figure 1).

Despite overall declines in the unemployment rate, the average and median duration of unemployment has increased (figure 2). These two conflicting trends suggest a change in the source of joblessness—from temporary layoff to permanent displacement.


Figure 1 Unemployment rate

Figure 1 Unemployment rate
Source : Bureau of Labor Statistics.


For most of the past century, employment and unemployment were highly correlated with the business cycle. This relationship appears to have changed in recent years. First, with the exception of the early 1980s, there has been a decline in the official length of recessions. Second, there has also been a decline in the magnitude of job losses occurring during economic slowdowns. Third, employment declines have continued for at least one year after the end of the last two recessions and employment recovery has taken longer. Taken together, these three developments suggest that something has changed in the underlying structure of the US labor market in recent years.

Data presented in figure 3 suggest that there has been a significant decline in variation across state unemployment rates over the past 30 years. During the late 1970s, states in the Northeast and Midwest—regions with high concentrations of traditional industries such as automobile manufacturing, textiles and apparel, and steel—experienced significantly higher unemployment rates than states in other regions. Beginning in the 1980s, state unemployment rates began converging toward the national average, reflecting a slow decline in overall unemployment and more similarity in state unemployment rates. This convergence suggests that, during the past 20 years, unemployment has been explained more by national factors than by state or regional factors.


Figure 2 Duration of unemployment

Figure 2 Duration of unemployment
Source : Bureau of Labor Statistics.


Figure 3 Variation in state unemployment rates

Figure 3 Variation in state unemployment rates
Source : Author’s calculations from Bureau of Labor Statistics data.


To summarize, the US labor market has experienced three major developments in recent years:

The original UI program was designed to offset income losses during cyclical periods of temporary involuntary unemployment. By contrast, current workers face long-term structural unemployment. The existing UI system is inadequate in responding to these current labor market conditions.

The current UI system does not assist workers who seek part-time employment, workers who voluntarily leave one job in order to take another, or workers who experience long-term unemployment. New entrants and reentrants into the labor market are not currently eligible for UI, since these two groups of unemployed do not fit well with one of the program’s original objectives, i.e., insuring against the risk of involuntary job loss. Covering these workers would raise issues concerning the amount and duration of assistance, since they may not have relevant work experience.

Underlining these macroeconomic changes to the US labor market is a shift from traditional employer-based full-time employment to an increased reliance on contingent and part-time employment. The shift to these nontraditional forms of employment reflects additional shortfalls in the current UI program. A system designed to provide income support during temporary layoffs for workers who were permanently attached to a single employer is not well designed for a labor market with considerable self-employment and contingent, part-time, and low-wage employment.


THE CURRENT UI PROGRAM

Federal law established the UI program in 1935 in order to provide temporary and partial wage replacement to workers involuntarily separated from their jobs. It was believed that UI would serve as a countercyclical mechanism to help stabilize the economy during economic slowdowns. In the more contemporary language of the economic analysis of insurance, the primary goal (or benefit) of UI is the ability of the government to smooth income and consumption during unemployment spells.

The UI program was established as a federal-state system. The federal government sets rules and standards, primarily on minimum coverage and eligibility criteria, and imposes a minor tax to finance the overall administration of the program. Individual states set their own benefit amounts, duration of assistance, and means of financing that assistance.


Coverage and Eligibility

The existing eligibility criteria for receiving assistance, listed below, are based on monetary and nonmonetary determinations; the application of these criteria varies by state:

Monetary eligibility is essentially a sufficient work history prior to job loss. Each state determines its own sufficient work history, relying on earnings during a base period.1 Most state programs assist only those workers who lose their jobs through no fault of their own, as determined by state law. In more detail, reasons for ineligibility of UI include the following:

There is enormous variation across states in the definition of good cause for voluntary separation, i.e., leaving to accept other work, compulsory retirement, sexual or other harassment, domestic violence, and relocation to be with spouse. Program discretion in setting these standards results in numerous inconsistencies. For example, workers who quit to move with a spouse and meet the monetary eligibility criteria are eligible to receive UI benefits in some programs—including California, Kansas, and New York—but not in others—including Connecticut, Delaware, the District of Columbia, and Massachusetts.

Workers who quit because they have been victims of sexual or other harassment are potentially eligible for UI benefits in all programs except six: Alabama, Georgia, Hawaii, Missouri, New Hampshire, and Vermont. Workers who voluntarily leave their jobs in anticipation of a plant closing in order to accept another job are potentially eligible for UI in many states, including California, Minnesota, New York, and Pennsylvania, but are ineligible in North Carolina, South Carolina, Tennessee, and West Virginia. In a highly mobile society, with integrated labor markets, it is difficult to imagine a plausible argument in support of these differences in state programs.

The base period monetary criteria are used as an imperfect proxy for labor market attachment. One unfortunate consequence is that some workers have insufficient work experience to meet the base period requirement, i.e., reentrants into the labor market who are actively seeking employment are not eligible for UI. As a result, women who decide to postpone returning to work after childbirth and workers who return to school or who take up training following a job loss can be ruled ineligible for UI. This is true despite the fact that their current or former employers paid UI taxes, and despite the likely satisfaction of monetary eligibility requirements for the immediate base period prior to the job loss.

The percent of total unemployed workers receiving assistance, the recipiency rate, has declined over the past two decades. The recipiency rate peaked in 1980 when half of all unemployed workers received UI. The rate fell to as low as 30 percent in 1984, before rebounding to 39 percent in 1991. Receipt of benefits increased to above 40 percent in 2001, 2002, and 2003, before falling back in 2004 (figure 4). The average recipiency rate over the past 27 years is approximately 37 percent. In other words, in recent years only a little more than one-third of unemployed workers actually have received assistance under the UI program.


Benefit Levels

One of UI’s initial goals was to replace half of lost wages. Because of the federal-state nature of the program, each state sets its own minimum and maximum weekly benefit amounts. Although several states have set their maximum weekly benefit at approximately two-thirds the state weekly wage, currently only one state—Hawaii—has achieved the initial goal of actually replacing, on average, half of lost wages.

Almost all states set their maximum weekly benefits somewhere between $200 and $500, with the largest concentration of states between $300 and $400. Puerto Rico has the lowest maximum weekly benefit ($133). States with the highest maximum weekly benefits include Massachusetts ($551 to $826), Minnesota ($350 to $515), New Jersey ($521), and Rhode Island ($492 to $615). The average weekly benefit in 2004 ranged from $106.50 in Puerto Rico to $351.35 in Massachusetts. The average weekly benefit for the entire country was $262.24. This average is almost 10 percent less than the weekly equivalent of the poverty level for a family of three that was set by the US Census Bureau.2

The replacement rate, defined as average weekly benefits as a share of average weekly earnings, is a useful measure of benefit sufficiency.3 The District of Columbia has the lowest replacement rate, less than one-fourth of average earnings. As mentioned above, Hawaii’s UI program comes closest to replacing half of unemployed workers’ average weekly earnings. Thirty-eight states have an average replacement rate of more than one-third but less that one-half of their workers’ average weekly wages. The states with the lowest replacement rates include Alabama, Alaska, Arizona, California, Connecticut, Delaware, Louisiana, Maryland, Mississippi, Missouri, New York, Tennessee, and Virginia. The average replacement rate for the United States between 1975 and 2004 was 0.36, reaching as high as 0.38 in 1982 and as low as 0.33 in 1998 and 2000.


Duration of Benefits

In the early years of the program, the duration of UI benefits was 12 to 20 weeks. Starting in the 1950s, a period of relatively low unemployment, a sizable number of states increased their UI duration to 26 weeks. By 1980, 42 states had a maximum duration of 26 weeks, and the duration for the 11 remaining programs was between 27 and 39 weeks. By the 1990s, 50 states had a uniform maximum duration of 26 weeks, with two jurisdictions at 27 to 39 weeks. Currently, all jurisdictions except three have a maximum duration of 26 weeks.4

Over the past 30 years, the average duration for receiving UI has ranged from a low of 13 weeks in 1989 to a high of 17.5 weeks in 1983, hovering around 15 weeks for most of the period (figure 5). A sizeable fraction of UI beneficiaries exhaust their benefits, i.e., remain unemployed beyond the period for which they can receive UI, ranging from a low of 25.8 percent in 1979 to a high of 43.9 percent in 2003. On average, approximately one-third of UI recipients exhaust their benefits before finding new jobs.


Figure 4 Unemployed workers, job losers, and unemployment insurance recipients, 1972–2003

Figure 4 Unemployment workers, job losers, and unemployment insurance recipients, 1972-2003
Source : Congressional Budget Office 2004, figure 3.


Figure 5 Average duration of unemployment insurance receipt, with periods of recession highlighted, 1957–2005

Figure 5 Average duration of unemployment insurance receipt, with periods of recession highlighted, 1957-2005
Sources: Bureau of Labor Statistics, US Department of Labor, and National Bureau of Economic Research.


With the trend increase in the average duration of unemployment, the maximum period that workers can receive UI has fallen from two times to a little more than 1.5 times the average duration of unemployment. As with benefit levels, there does not appear to be any significant relationship between benefit duration and local labor market conditions.


Extended Benefit Program

The UI system proved unable to respond to surges in unemployment during most of the cyclical downturns over the past half century. Increases in the duration of unemployment during and immediately following those recessions were the primary impetus for extending statutory UI beyond its base period. Congress enacted the first temporary extension of UI during the 1958 recession. In 1970, Congress enacted the Extended Benefit (EB) program with automatic triggers to provide assistance in a more orderly fashion. High rates of regular UI exhaustion, problems with the automatic triggers, and political pressures resulted in the need for subsequent congressional action to deal with heightened levels and prolonged duration of unemployment during recessions.

Under the current program, UI benefits can be extended for an additional 13 weeks when the unemployment rate of those workers covered by the program, i.e., the Insured Unemployment Rate (IUR), for the previous 13 weeks is at least 5 and 20 percent higher than that rate for the same 13-week periods in the previous two years. Since states are required to finance half of the extended benefit programs, they are free to adjust this trigger.

Changes in the labor market combined with the static nature of the triggers, have produced an extended benefit system that is not automatic. As a result, Congress has occasionally found it necessary to extend UI through the Temporary Extended Unemployment Compensation program. Since the 1980s, the standard extended benefit program has provided a smaller share of assistance to unemployed workers than the emergency extensions of UI enacted by Congress.

Although helpful to millions of workers, these temporary stopgap measures have politicized unemployment, thereby undermining one of the initial goals of the UI program. These temporary programs have proven to be clumsy, typically being enacted after hundreds of thousands of workers have already exhausted their UI. In addition, the sunset provisions are arbitrarily set and usually fall before employment has recovered. Overall, the nation’s UI program has become less automatic and more dependent on congressional action in response to prolonged periods of economic slowdown.


Financing UI

UI is financed by a combination of federal and state payroll taxes. Revenue from the federal payroll tax is used to finance the costs incurred by federal and state governments in administering the UI program and to cover loans to states that exhaust their regular UI funds. States are required to raise the necessary revenue to finance regular UI benefits paid to their unemployed workers. Federal and state governments share the costs of financing benefits under the automatic extended benefit program. Currently, federal taxes finance 17 percent of the UI program. The remaining 83 percent is financed by state taxes. Temporary extended UI programs enacted by Congress have typically been financed by federal budgetary expenditures without any specific revenue offset.

The federal tax established by the Federal Unemployment Tax Act (FUTA) is currently 6.2 percent on the first $7,000 of annual salary by covered employers on behalf of covered employees.5 Employers must pay the tax on behalf of employees who earn at least $1,500 during a calendar quarter. Employers in states with federally approved UI programs receive a 5.4 percent credit against the tax, making the effective FUTA tax rate 0.8 percent. The bottom line is that the federal tax is trivial: A maximum of $56 is collected annually for each worker who is covered under the program.

There have been few adjustments in the FUTA taxable wage base since it was first established in 1939. The wage base, originally set at $3,000, remained fixed for 32 years, until 1972, when it was raised to $4,200. That increase kept the taxable wage base in line with its real value in 1960. Congress raised the federal taxable wage base to $6,000 in 1978 and to $7,000 in 1983, where it has remained for the past 22 years. Had the taxable wage base been adjusted for inflation over the past 65 years, it would currently be approximately $45,000 (figure 6).


Figure 6 Federal taxable wage base, 1940–2004

Figure 6 Federal taxable wage base, 1940-2004
Sources : US Department of Labor and author’s estimates.


If the taxable wage base were adjusted to $45,000, the net federal tax rate, i.e., the tax rate minus the credit, could be reduced by half, to 0.4 percent, and generate the same amount of revenue that is currently being collected. Although it is unrealistic to expect an adjustment of this magnitude anytime soon, any increase in the wage base to make up for the erosion in its real value over the past two decades could provide additional funding for providing assistance to workers in need, or could enable the federal government to reduce the FUTA tax rate, or both. Most importantly, adjusting the wage base upward would reduce the regressive nature of the tax. Under the current structure, the FUTA tax accounts for a larger share of lower income workers’ wages. Adjusting for inflation alone, as many states have been doing for their own UI taxes, would increase the federal taxable wage base fivefold, make the system more progressive, and provide additional revenues to the system.

Federal guidelines dictate that states have in place UI payroll tax systems that are experience rated. With experience rating, firms that lay off fewer workers face a lower tax rate on their payroll. States have the discretion to structure their own experience rating system, and those systems, as with the tax rates, vary considerably among the states.

Some aspects of the current UI system work well and deserve to be highlighted. UI constitutes an important source of income for unemployed workers and their families, particularly for the long-term unemployed. The Congressional Budget Office (2004) reports that UI benefits played a significant role in maintaining the family income of recipients who experienced long-term spells of unemployment in 2001 and early 2002, particularly for those families that had only one wage earner. Before becoming unemployed, recipients’ average family income was about $4,800 per month. When recipients lost their job, that income—excluding UI benefits—dropped by almost 60 percent. Including UI benefits reduced the income loss to about 40 percent.6


REFORMING UI

In recent years, the US labor market has come under increased pressures from intensified domestic and international competition. These pressures have changed the nature of job turnover in the United States. Unlike the cyclical job losses that characterized the labor market and economy from 1945 to the 1980s, job losses are now related more to structural factors, with workers simultaneously changing jobs, industries, and occupations. The existing UI program, though, is fighting the last battle, one of widespread temporary layoff, where workers were attached to a single employer.

As discussed above, current labor market conditions differ a great deal from those that existed in 1935, suggesting that it is time to revisit some of the fundamental elements of the original UI program. The reforms outlined below maintain the basic structure of UI, while enhancing its efficiency, reach, and impact to reflect the changes in the labor market since the program was designed. Although each proposal can be evaluated and implemented separately, it would be preferable to enact them all.


Strengthen the Federal Leadership Role in UI

As documented above, the nature of unemployment in the United States has shifted from cyclical to structural. Although there clearly remain some differences in local labor market conditions, the current pressures on the US labor market are becoming more national. State differences in the incidence and experience of unemployment have narrowed considerably. Local labor market conditions primarily affect the prospects for reemployment. Given the increasingly national nature of the labor market, UI would better meet its original objectives if the federal government played a more prominent role in this partnership.

In addition to inequities created by disparate rules across states, a significant downside of the current federal-state partnership is the states’ real or perceived fears that program generosity will result in adverse changes to their business environment. Increased federal leadership would avoid interstate competition and a “race to the bottom” in program benefits.

An increased leadership role for the federal government would be characterized by expanding standards for eligibility, duration, and level of benefits and for financing the program.


Eligibility

The share of unemployed workers who actually received assistance under the UI program averaged 37 percent between 1980 and 2005. The proposals outlined above are designed to increase the number and share of unemployed workers eligible to receive assistance. Given the difficulties associated with precise estimation of how much each of the individual proposals would contribute to increasing the number of potentially eligible workers, the costs associated with raising the recipiency rate in increments to 50 percent is estimated (table 1), which is a reasonable objective for the changes delineated above.


Table 1 Estimated costs associated with increasing the recipiency rate

Recipiency rate

Increase in number of workers eligible * (thousands)

Increase in total benefits paid * (billions of dollars)


40 percent

220

1.6

45 percent

620

4.5

50 percent

1,000

7.4


* = increase in workers and costs (benefits paid) relative to 25-year average.

Source : Kletzer and Rosen (2006).


Benefit Levels and Duration of Benefit Receipt


Table 2 Estimated costs associated with increasing the replacement rate

Replacement rate Average weekly benefit at new replacement rate (dollars) Increase in average weekly benefit (dollars) Increase in total benefits at new replacement rate (billions of dollars)

40 percent

295.67

34.00

0.3

45 percent

332.63

70.96

0.7

50 percent

369.59

107.92

1.1


Source: Kletzer and Rosen (2006).

Note: Estimates are based on the following assumptions: The average replacement rate between 1980 and 2003 was 35.4 percent; the average weekly benefit in 2003 was $261.67; the average weekly wage in 2003 was $739.18; the total number of weeks of compensation in December 2005 was slightly fewer than 10 million.


Financing

Local or regional wage differences, or both, would be respected under this plan, because the harmonization of benefits would be in percentages of earnings, not dollar levels. Treating workers more equally, in terms of program standards, would remove differences that have little or no justification, other than tradition. Given their long experience in providing these services, local and state providers would remain primarily responsible for reemployment assistance, job training, intake, and administration of benefits.


Augment UI with a Program of Wage-Loss Insurance

On average, dislocated workers pay a heavy price as a result of unemployment. According to the Displaced Worker Survey only two-thirds of unemployed workers find a new job within 1 to 3 years after layoff (figure 7). More than 40 percent of workers experience earnings losses and only approximately one-fourth of workers experience no earnings loss or an improvement in earnings after reemployment.


Figure 7 Reemployment and earnings experience of displaced workers

Figure 7 Reemployment and earnings experience of displaced workers 
Sources : Displaced Worker Survey, Bureau of Labor Statistics; author’s calculations.

Wage-loss insurance offers assistance that is tailored to actual earnings losses. In order to be effective, wage-loss insurance must be a complement to traditional UI, since it only assists those workers who find new jobs. Under the program, eligible workers would receive some fraction, perhaps half, of their weekly earnings loss over a specific period.

For example, the average weekly wage before layoff for workers displaced from manufacturing industries was $396.88 between 1979 and 2001, and the average weekly age for those laid off from nonmanufacturing jobs was $368.65. For those workers who found new jobs, the average percent loss in earnings was 29.2 percent for manufacturing workers and 18.6 percent for nonmanufacturing workers. Had a wage-loss insurance program been in place, manufacturing workers would have received approximately $6,000 over a two-year period, or 15 percent of their prelayoff wage. Nonmanufacturing workers would have received approximately $3,600 over a two-year period, or 9 percent of their prelayoff wage.

The Trade Act of 2002 expanded Trade Adjustment Assistance (TAA) to include a limited wage-loss insurance program. Under the TAA program, workers who are more than 50 years old and earning less than $50,000 a year may be eligible to receive half the difference between their previous and new earnings, subject to a cap of $10,000, for up to two years. Workers must find a new full-time job and enroll in the Alternative Trade Adjustment Assistance (ATAA) program within 26 weeks of job loss and cannot receive other income support or training under TAA.

Despite its benefits, wage-loss insurance is not a perfect solution to addressing the costs associated with unemployment. Structuring a program with a relatively short eligibility period, starting with the date of job loss, may create a reemployment incentive, addressing one of the most commonly expressed UI concerns, but it also limits the compensatory nature of the program. Displaced worker earnings losses are long term (i.e., earnings losses exist five to six years after job loss), well beyond the two years covered by ATAA.

In order to avoid any adverse effect on wages, wage-loss insurance must be provided to workers, not employers. In fact, there is no reason for employers to even know that workers are receiving assistance under this program.

The cost of a wage-loss insurance program depends on the number of eligible workers, the earnings losses of those reemployed at lower pay, and the duration of unemployment prior to reemployment. Other critical program characteristics include the duration of wage-loss insurance payments, the annual cap on program payments, and the replacement rate. It has been estimated that the cost for a program with a two-year duration, a 50 percent replacement rate and a $10,000 annual cap for all dislocated workers would be around $4 billion.

An expanded wage-loss insurance program could be financed through general government revenues or by raising the FUTA taxable wage base or tax rate. Augmenting UI, with assistance tailored to the size of reemployment earnings losses, is possible with relatively small changes in UI program parameters.

In general, the current UI system has a limited relationship with efforts to transition workers back to employment. The worker profiling system targets resources on workers at risk of exhausting benefits. Workers receiving UI are required to prove that they are actively seeking employment, primarily by documenting job inquiries and interviews. Most unemployment spells (and benefit receipt) are too short for serious training, but job search assistance can be short term with high return, given its relatively low cost. With the rise in structural unemployment, training needs are likely to expand. As a result of the bureaucratic wall of separation between UI and federally supported training programs in the United States, the amount of funds appropriated are inadequate to provide any kind of serious training to all long-term unemployed workers.


CONCLUSION

The current federal-state structure of UI is a relic of its 1935 establishment. The program has not undergone any major reforms, despite significant changes in the US labor market over the last few decades. The current UI program was created to assist workers experiencing transitional unemployment due to cyclical factors. Today’s workers are experiencing longer spells of unemployment and larger earnings losses due to structural factors like technological change and intensified competition resulting from globalization.

Changes necessary to move UI into the 21st century require strong federal leadership. The very basic structure of UI must be reformed, broadening from the single-employer, full-time worker, temporary layoff model to an approach that accommodates permanent job loss, part-time or contingent work, self-employment, and the incidence of job loss and national, rather than local or regional, unemployment. Reforming the nation’s UI program is necessary in order to make it relevant to the labor market of the 21st century.

Congressman Jim McDermott’s draft legislation being considered by the subcommittee incorporates most of the recommendations outlined above. The congressman’s proposals would go very far in improving the relevance of the UI program to current labor market conditions. I strongly encourage members of Congress to seriously consider these proposals and to enact them as soon as possible. Delaying their adoption will result in raising the costs that unemployed US workers already face.


Notes

1. See Kletzer and Rosen (2006) for a complete discussion of the base period used to determine UI eligibility.

2. Annual incomes at and below $14,974, for a family of three, with one child under the age of 18, were defined as poverty level for 2004 (US Census Bureau, 2005).

3. Only average weekly earnings for UI recipients are used in calculating the replacement rate.

4. Washington and Massachusetts have a maximum duration of 30 weeks.

5. The 6.2 percent includes a 0.2 percent surtax initially passed by Congress in 1976, designed to replenish the UI trust fund. The surtax is scheduled to expire on December 31, 2007.

6. Long-term recipients are defined in this report as unemployed workers who received UI benefits for a spell of at least four consecutive months, in 2001 or early 2002.



© 2014 Peter G. Peterson Institute for International Economics. 1750 Massachusetts Avenue, NW.
Washington, DC 20036. Tel: 202-328-9000 Fax: 202-659-3225 / 202-328-5432
Site development and hosting by Digital Division