Trends in unemployment insurance benefits
Daniel P. McCmurrer
The share of the unemployed receiving unemployment insurance declined slowly, but consistently, starting in the 1940’s, dropped dramatically during the 1980-84 period, and remains low
The Federal-State unemployment insurance (UI) system, created in 1935, was designed to provide temporary wage replacement for unemployed workers who have demonstrated a strong attachment to the labor force and to assist in stabilizing the national economy during cyclical economic downturns.
The nature of the system assigns different responsibilities to the Federal and State governments. Although broad Federal laws ensure consistency in areas where uniformity is considered essential, States determine most of the details of program operations and administration. As a result, many features of the system vary greatly among States.
Two separate, but interrelated, programs currently provide income support to qualified unemployed workers: the permanent, regular, State UI programs and the Federal-State Extended Benefits program. In addition, during every recession since 1958, emergency supplemental UI benefit programs have been enacted by Congress on an ad hoc basis. The characteristics of the three components of the UI system are discussed in more detail below.
Regular State unemployment insurance. Regular State UI programs generally provide up to 26 weeks of benefits to qualified unemployed
workers. The eligibility of an unemployed worker is determined by State laws regarding monetary factors (such as recent earnings history) and nonmonetary factors (such as the reason for separation from employment and current availability for work). The duration and amount of benefits for eligible individuals are based primarily on an individual’s recent earnings history.
State taxes on employers(1) finance most benefits paid by the program.(2) Tax rates vary among employers in the same State and are based partially upon the level of past UI claims that were made by an employer’s former employees. Federal taxes imposed by the Federal Unemployment Tax Act pay for the administration of State UI programs and the Federal share of the Extended Benefits program. The total amount paid by the regular program is cyclical with the level increasing as the number of unemployed increase during periods of economic downturn. In 1993, more than $22 billion was paid in regular benefits.
Federal-State Extended Benefits. The Federal-State Extended Benefits program provides up to 13 additional weeks of benefits to individuals who have exhausted their regular UI benefits. Half of the cost of extended benefits is financed by the Federal government and half is paid by the State distributing the benefits. Extended Benefit amounts are the same level as the State’s regular benefits.
Extended Benefits are available only when a measure–usually, the Insured Unemployment Rate–of State unemployment rises above a particular level. Most States currently use the insured unemployment rate as the only “trigger” for the program. Because this rate is determined by the number of regular UI claimants in a State, eligibility for extended benefits in most States is affected directly by States’ UI eligibility laws. As a result, a decline in the percentage of the unemployed who receive regular UI benefits has contributed directly to a drop in the number of States in which Extended Benefits are available.
Emergency benefit programs. The Emergency Unemployment Compensation program is a temporary benefits program that the Congress enacted in November 1991 and extended on several occasions. The Congress allowed the program to expire in February 1994. This emergency compensation program was similar in many ways to several previous emergency programs, which the Congress enacted during recessions. For example, the Federal Supplemental Benefits program paid benefits between 1975 and 1978, and the Federal Supplemental Compensation program paid benefits between 1982 and 1985.
The number of additional weeks of benefits that were available in the Emergency Unemployment Compensation program depended on three factors: when the claimant first applied for these benefits, a State’s unemployment rate, and the national unemployment rate. Claimants for Emergency Unemployment Compensation were required to meet their State’s eligibility criteria, in addition to Federal requirements recompensation while it operated.
Because the Federal Government finances all the costs of emergency unemployment benefits, but only 50 percent of Extended Benefits costs, States took advantage of the option to provide Emergency Unemployment Compensation. Following the most recent recession, the Emergency Unemployment Compensation program nearly replaced the Extended Benefits program entirely (payments of extended benefits since 1991 have been less than $400 million). In total, the Emergency Unemployment Compensation program cost more than $26 billion; a significant proportion was financed out of general government revenues.
Characteristics of the unemployed differ slightly in comparison with the civilian labor force. (See table 1.) In particular, younger individuals, men, and blacks are disproportionately represented among the unemployed. Individuals who seek UI benefits tend to be older than unemployed workers in general; men also are disproportionately represented.
Table 1. Characteristics of the labor force and recipients of
unemployment insurance, 1993
Characteristics Civilian Total Unemployment
labor force unemployed claimants
16 to 34 43 58 42
35 to 54 45 34 46
55 and over 12 8 12
Men 54 56 60
Women 46 44 40
White 85 75 –
Black 11 21 –
Other 4 4 –
Duration of benefits. In most States, the potential duration of UI benefits also is based on an individual’s recent earnings.(5) Maximum duration is uniform among States; all but two provided a maximum of 26 weeks of benefits in 1993.(6) In general, the average potential duration of benefits has increased gradually, as has the average duration of unemployment spells. (See chart 2.)
Trends in receipt of UI benefits
Two trends have become apparent in the UI benefits program. The percentage of the unemployed who receive UI benefits (referred to as “recipiency”) has declined slowly, but consistently, since the 1940’s; and the percentage of recipients has dropped dramatically between 1980 and 1984 and has remained at a low rate throughout the 1980’s and early 1990’s.
These declines are of considerable concern. They threaten to undermine the two primary functions of the UI system: partial replacement of wages for unemployed workers, and countering economic downturns by automatically injecting more money into the economy during periods of high unemployment. The insured unemployment rate is the primary method that activates the Extended Benefits program during recessions. Because the decline in the percentage of recipients is reflected in the insured unemployment rate, the decline also has had the effect of weakening the countercyclical effectiveness of the Extended Benefits component of the UI system.
The two declines have likely been caused by a combination of factors that tend to have similar effects on the UI system. The long-term decline is probably a partial result of broad shifts in labor market demographics, with industrial shifts such as the decline in manufacturing, and increases in UI coverage. To the extent that the percentage of the unemployed who receive UI benefits has decreased over the long-term, the UI program no longer responds to the needs of a growing portion of the unemployed population.
Several researchers have identified the causes of the recent, more short-term decline in recipiency nationwide. Four factors have been identified as the primary causes, although the results have not been wholly consistent and researchers have had substantial difficulty in separating the effects fully.
First, policy changes were made on the Federal and State levels that appear to have reduced the percent of the unemployed who receive benefits. Second, an increasing percentage of the unemployed live in States in which the percent of the unemployed who receive benefits is consistently below the national average. Third, the unionized percentage of the work force, in which rates of UI claims have historically been high, has declined. Fourth, the percentage of the work force employed in the manufacturing sector, in which rates of UI claims also have been high, has declined.
Who receives benefits?
Two primary statistics that generally measure recipiency are the ratio of the insured unemployment rate to the total unemployment rate,(7) and the ratio of UI claimants to the total number of unemployed.(8) The two ratios are highly correlated. (See chart 3.) The ratio of the insured unemployment rate to the total unemployment rate is more difficult to interpret than the ratio of UI claimants to the total number of unemployed because of various mathematical complications related to the definitions of the populations that are counted. This can result in a measure that is above 100 percent. Still, the ratio of insured unemployment rate to the total unemployment rate ratio is widely reported, and the insured unemployment rate is particularly important because it represents the primary trigger for the Federal-State Extended Benefits program.
Both ratios are based on a measure of the number of UI claimants, which is collected weekly by States. The total number of claimants, however, includes some individuals who do not receive UI benefits but are counted among the insured unemployed for a particular week. Three primary groups of individuals fall into this category: individuals who are on a 1-week waiting period before they begin to receive benefits; claimants who ultimately are denied benefits for nonmonetary reasons; and claimants who are disqualified from collecting benefits in a particular week for reasons that include the requirements that recipients be able and available for work and that claimants who are working do not exceed a particular level of income in a week. Including these groups has tended to inflate the measure of UI recipiency by 10 to 15 percent per year.
Trends and implications
Both recipiency measures have shown a long-term decline and a more short-term decline. (See chart 3.) The measures also vary considerably across States: in 1993, the ratio of claimants to total unemployed ranged from 15 percent in South Dakota to 64 percent in Alaska. (See table 2.) An additional measure, the ratio of UI claimants to total job losers, also has demonstrated long-term and short-term declines. See chart 4.
Table 2. Percent of total unemployed who are unemployment
insurance claimants, by State 1993
Furthemore, the decline of the insured unemployment rate relative to total unemployment has weakened the countercyclical effectiveness of the UI system: the insured unemployment rate is the primary mechanism to activate the Extended Benefits program during recessions. Thus, the decline in the insured unemployment rate has resulted in a significant reduction in the number of States in which extended benefits are available.
The long-term decline
Research suggests that the long-term decline is primarily a result of changes in the demographic composition of the labor force and that the decline in one measure (the ratio of the insured unemployment rate to the total unemployment rate) is partially the result of increases in UI coverage.
Broad demographic changes. A primary cause of the decline in the ratio of UI claimants to the total number of unemployed before 1980 was the changing demographic composition of the jobless, according to Burtless and Saks.(12) Throughout the 1960’s and 1970’s, as many women and young workers from the baby-boom generation entered the labor force, they also made up a higher percentage of the unemployed. As a result, men of prime working age, who are the most likely to receive UI benefits, declined considerably as a percentage of the unemployed. Burtless and Saks found that such demographic changes explain a large percentage of the decline in the ratio of UI claimants to the total number of unemployed before 1980.
While the impact of demographic changes described by Burtless and Saks declined after 1980, other demographic changes have continued or even accelerated in the 1980’s and 1990’s. Perhaps the most significant change is the continuing increase in the number of two-earner families. Although empirical research has not addressed this issue, the increase in two-earner households has most likely reduced the need among some workers to apply for UI benefits when they become unemployed. Thus, it is possible that various broad demographic changes continue to have a negative impact on the rate of UI recipiency.
Increases in UI coverage. Newly covered employees in the 1970’s were probably less likely to apply for compensation than previously covered groups.(13) As a result, the insured unemployment rate (the number of UI claimants as a percentage of jobs covered by UI), declined because of the increased coverage of the system. Burtless and Saks suggest that the insured unemployment rate may have declined by between 0.5 and 0.8 percentage points because coverage was extended twice in the 1970’s(14). Such a decline would account for a large percentage of the decline in the ratio of the insured unemployment rate to the total unemployment rate in this period, although it would not be expected to have the same effect on the ratio of UI claimants to the total number of unemployed.
Decline in manufacturing. Burtless and Saks also identified the shift of workers from manufacturing and other industries with high recipiency rates as a primary cause of the long-term decline in the number of recipients. They report that estimating with precision the magnitude of this effect is difficult. The decline in manufacturing also has been identified as a significant cause of the decline during the 1980’s.(15)
The short-term decline
Research examining the decline in UI recipiency that occurred in the early 1980’s continues to be inconsistent. The variability of the results is an indication of the difficulty researchers have had quantifying the impact of the four factors identified earlier: changes in Federal and State policy, population shifts, declining unionization rates and the decline in manufacturing. A combination of some or all of these factors probably contributed significantly to the short-term decline.
Policy changes. During the 1980’s, several changes in Federal and State law appear to have contributed to the drop in the percentage of the unemployed who received unemployment benefits. Overall, the Federal General Accounting Office found that policies designed to improve the solvency of State trust funds reduced the recipiency among unemployed individuals.(16) Most significantly, numerous State laws were changed to restrict eligibility and reduce benefit levels, partly in response to Federal policies that encouraged States to adopt more restrictive legislation for regular State unemployment programs. Several Federal laws, most notably the decision to tax UI benefits, also directly reduced the value of unemployment benefit levels.
Federal policies. During the 1980’s, Federal regulations governing State UI Trust funds were changed significantly. Beginning in 1982, States were required to repay with interest Federal loans to their trust funds. Previously, the loans were interest-free and repayment requirements were unclear. States with loans also were required to adopt other specific measures to ensure solvency.
Overall, these changes provided incentives to States to avoid the need for future loans by reducing the scope of State programs. In addition, States were given other direct incentives, linked to Federal Extended Benefits funds, to tighten UI eligibility requirements and to reduce UI benefits. Taken as a whole, State policy reflected these changes in Federal policy. Federal laws also were changed in ways that directly affected the recipiency rate. In 1979, UI benefits for the first time were partially taxed, and in 1986, all unemployment benefits became subject to taxation. States also were required to reduce or eliminate UI payments to unemployed workers who received pensions or Social Security payments. Walter Corson and Walter Nicholson found that, overall, between 11 percent and 23 percent of the total decline can be attributed directly to various Federal policy changes. Specifically, between 11 percent and 16 percent of the decline is due to partial taxation of benefits and up to 7 percent is the result of less generous Extended Benefits programs.(17)
State policies. The GAO reported that, between 1981 and 1987, 44 States adopted tighter monetary eligibility standards or stricter disqualification provisions for their regular UI programs. Many of these State changes probably were the result of Federal incentives to tighten eligibility, although determining the precise impact that changes in Federal legislation alone had on the policy decisions of States is impossible. Some research has found that these and other changes in State policy account for a significant percentage of the decline in recipiency.
Corson and Nicholson found that between 21 percent and 55 percent of the decline in the number of recipients is attributable to State policy changes. Specifically, the decline is due to:
* to 11 percent to increases in denial rates for disqualifying
* to 11 percent to increases in the minimum earnings
required to qualify for UI;
* to 11 percent to increases in the denial rate for misconduct;
* up to 13 percent to changes in voluntary separation standards;
* percent to reductions in maximum duration of benefits;
* to 4 percent to changes in wage replacement rates.(18)
Corson and Nicholson also found that the ratio of UI claimants to the total number of unemployed would have increased between 1 percent and 13 percent as the result of reductions in work test denials, partially canceling the effects of the other factors.(19)
Burtless and Saks also concluded that State legislative and administrative changes are the primary cause of the decline in rates of change in the number of recipients, but they did not present estimates of the magnitude of the effects of these changes.(20)
Marc Baldwin and Richard McHugh suggested that State policy changes account for 54 percent of the 1979-90 decline in recipiency.(21) An updated work by Baldwin, however, found sharp reductions in the apparent effects of State policy changes.(22) Baldwin and McHugh attributed the decline to:
* 21 percent to increases in the minimum earnings required
to qualify for UI;
* 16 percent to increases in the earnings required to qualify
for the maximum benefit;
* 8 percent to increases in the number of States with disqualification
periods for job quitters;
* 7 percent to increases in the number of States with disqualification
periods for refusal of suitable work;
* 1 percent to increases in the number of States withright-to-work
But Rebecca Blank and David Card found little evidence that State policy changes had any impact on recipiency. They found that individual eligibility for UI benefits appeared to decline slightly as the result of tighter State eligibility standards, although these effects were offset by increasing wage levels.(24)
Population shifts. An increasing share of U.S. unemployment is in Southern and Mountain states, where the ratio of UI claimants to the total number of unemployed has consistently been lower than the national average. As the percentage of national unemployment in these States increases, the national ratio of UI claimants to the total number of unemployed would be expected to fall accordingly. This is a long-term demographic trend, occurring throughout the last three decades and continuing into the present. Blank and Card found that these regional shifts in population accounted for approximately 50 percent of the decline in the national ratio of UI claimants to the total number of unemployed between 1977 and 1987.(25) Wayne Vroman asserted that 25 percent is a more appropriate figure,(26) and Corson and Nicholson attributed 16 percent of the variation to geographic population shifts.(27)
However, these analyses do not explain the underlying variations in ratio of UI claimants to the total number of unemployed across States that have caused the national rate to be affected by interstate migrations. Much of this variation can likely be attributed to differences in State policy, although the exact extent to which this is the case has not yet been determined.
Decline in unionization. Between 1979 and 1988, the percentage of unionized employees decreased 25 percent.(28) Because unions have traditionally represented a powerful source of information regarding available benefits for unemployed workers, the decline in union membership could have exacerbated problems related to distributing information among the unemployed. In addition, unions have often helped members file UI claims by guiding them through the UI system. Finally, many union members are eligible only for supplemental unemployment benefits paid by their union if they apply for regular UI.
Blank and Card attributed 25 percent of the decline in recipiency to the decline in unionization.(29) Baldwin and McHugh assigned 29 percent of the drop in recipiency to the decline in unionization.(30) Vroman also points to the potential importance of the unions’ information role by noting that the most important reason for nonapplication for UI benefits by unemployed individuals is their belief that they are ineligible for UI.(31) Inability to understand eligibility conditions may cause eligible workers to fail to apply.
Decline in the manufacturing sector. As noted above, Burtless and Saks suggested that industrial shifts contributed to the long-term decline in recipiency. This trend continued in the 1980’s as manufacturing as a percentage of total employment fell by 22 percent between 1979 and 1990. This factor also has been identified as a significant cause of the short-term decline. Corson and Nicholson found that between 4 percent and 18 percent of the decrease in the UI claims ratio can be attributed to the decline in the manufacturing sector.(32) Baldwin and McHugh attributed 16 percent of the total decline in the ratio of UI claimants to the total number of unemployed to this factor.(33)
In addition, Corson and Nicholson noted that an unemployed worker who had been employed in manufacturing is 25 percent more likely to collect UI than a similar worker from another industry. These findings are partially called into question, however, in analyses by Corson and Anu Rangarajan,(34) and Baldwin.(35) They found that a decline in manufacturing employment leads to an increase in the insured unemployment rate. Overall, it should be noted that because unions traditionally have been composed disproportionately of workers in the manufacturing sector, the decline in manufacturing is closely linked to the decline in unionization. As a result, the effects of the factors may be difficult to separate.
In sum, the percentage of the unemployed who receive Unemployment Insurance benefits has declined steadily, with a particularly sharp decline in the early 1980’s. This suggests that the relevance of the system to the needs of today’s work force has been eroded. A number of factors have contributed to this erosion, including Federal and State policy changes, broad demographic changes, and the decline in the manufacturing sector and in unionization. The resulting decline in recipiency has jeopardized the program’s capacity to carry out its two primary functions: wage replacement for involuntarily unemployed individuals and the countercyclical stabilization of the economy.
[ILLUSTRATIONS 1-5 OMITTED]
(1) Employees also pay UI taxes in Alaska, New Jersey, Pennsylvania, and West Virginia. In some of the four States, payment by employees depends on the status of the UI trust fund. (2) State and local governments and many nonprofit organizations do not pay UI taxes. They reimburse the UI system directly for benefits paid to their former employees. (3) Rebecca M. Blank and David E. Card, “Recent Trends in Insured and Uninsured Unemployment: Is There an Explanation?” Quarterly, Journal of Economics, NNovember, 1991.
Marc Baldwin and Richard McHugh, “Unprepared for Recession: the Erosion of State Unemployment Insurance Coverage Fostered by Public Policy in the 1980s,” Economic Policy Institute Briefing Paper, February 1992, also find results that are consistent with this conclusion. (4) Data produced by U.S. Department of Labor, Unemployment Insurance Service, Division of Actuarial Services. (5) In nine States, all eligible claimants have uniform potential durations. (6) Massachusetts and Washington allow benefits for up to 30 weeks. (7) The insured unemployment rate is defined as the number of regular UI benefit claimants divided by the average number of employees covered by UI over 4 of the last 6 completed calendar quarters. The total unemployment rate is defined as the number of all active unemployed job seekers divided by the total civilian labor force. (8) The specific measure of recipiency used by researchers in examining this question has varied. Walter Corson and Walter Nicholson, An Examination of Declining UI Claims During the 1980s, Unemployment Insurance Occasional Paper 88-3 (U.S. Department of Labor, 1988) examined both ratios, but focused upon the ratio of UI claimants to the total number of unemployed, which they call the UI claims ratio.
Blank and Card, in “Recent Trends,” also examined this measure, which they call the fraction of insured unemployment.
Wayne Vroman, The Decline in Unemployment Insurance Claims Activity in the 1980s, Unemployment Insurance Occasional Paper 91-2, (U.S. Department of labor, 1991) also focused on the ratio of UI claimants to the total number of unemployed.
Baldwin and McHugh, “Unprepared for Recession,” also examine the ratio of UI claimants to the total number of unemployed, but include Extended Benefits recipients in addition to regular State UI recipients. (9) “The Uncompensated Unemployment: An Analysis of Unemployed Workers Who Do Not Receive Unemployment Compensation,” Congressional Research Service, 1990. (10) The ratio of the insured unemployment rate to the total unemployment rate and the ratio of UI claimants to the total number of unemployed can be statistically predicted quite accurately for the years up to 1980 by knowing two variables: the year, which reflects the long-term decline of the system, and the unemployment rate, because the ratio tends to increase significantly during periods of high unemployment. Since 1980, however, the recipiency ratios no longer have the same statistical relationship to these two variables. (11) Gary Burtless and Daniel Saks. “The Decline in Insured Unemployment During the 1980s,” Unpublished Brookings Institution Report to the Department of Labor, March 1984, p. 42, (12) Burtless and Saks, “The Decline in Insured Unemployment During the 1980s,” p. 20. (13) This particularly was likely to be true for State and local government employees because they experienced low levels of unemployment in the early 1980’s. (14) Burtless and Saks, “The Decline in Insured Unemployment,” p. 17. (15) Ibid., p. 19-20. (16) Unemployment Insurance: Program’s Ability to Meet Objectives Jeopardized (Washington, DC, U.S. General Accounting Office, 1993), pp. 30-37. (17) Corson and Nicholson, An Examination of Declining UI Claims, pp. 119-20. (18) Any apparent discrepancy in totals is due to rounding. (19) Corson and Nicholson, An Examination of Declining UI Claims, pp. 119-20. (20) Burtless and Saks “The Decline in Insured Unemployment,” 1984, pp. 54-80. (21) To facilitate greater comparability between the findings of Baldwin and McHugh, “Unprepared for Recession,” and those of other studies, Baldwin and McHugh’s findings have been re formulated in the text. In particular, they report that State policy changes account for 97.4 percent of the total net change in ratio of UI claimants to the total number of unemployed, rather than 54 percent reported in the text. Overall, they find three primary factors that contributed to the decline in the ratio of UI claimants to the total number of unemployed and other factors that partially offset the decrease. As a result, when only the three factors that decrease the ratio are combined, they are larger than the net decline. Each of the factors independently appears to be a large percentage of the net decrease. To determine the relative impact of each factor, the percentage of the overall negative impact upon the ratio of UI claimants to the total number of unemployed that is attuributable to each of those factors that decreases in the ratio of UI claimants to the total number of unemployed must be calculated. These calculations indicate that State policy changes account for 54 percent of the decrease in the ratio of UI claimants to the total number of unemployed, declining unionization for 29 percent, and decreases in the manufacturing sector for 16 percent. The remaining 1 percent is attributable to the lagged unemployment level. (22) The research literature has not yet reconciled the variations in the results found by Marc Baldwin, “Benefit Recipiency Rates Under the Federal/State Unemployment Insurance Program: Explaining and Reversing Decline,” Unpublished Ph.D. diss., Massachusetts Institute of Technology, 1993), and by Baldwin and McHugh, “Unprepared for Recession,” 1992, p. 18. (23) Any apparent discrepancy in totals is due to rounding. (24) Blank and Card, “Recent Trends in Insured and Uninsured Unemployment,” p. 1166. (25) Ibid., p. 1177. (26) Vroman, The Decline in Unemployment Insurance, p. 13. (27) Burtless dismissed regional shifts as a possible explanation. However, later studies have appeared to confirm the merit of this factor. “Why is Insured Unemployment So Low?” Brookings Papers on Economic Activity (Washington, DC, Brookings Institution, 1983), pp. 225-49. (28) Michael A. Curme, et al. “Union Membership and Contract Coverage in the United States, 1983-1988,” Industrial and Labor Relations Review, October 1990, pp. 5-34, and Edward C. Kokkelenberg and Donna R. Sockell, “Union Membership in the United States, 1973-1981,” Industrial and Labor Relations Review, July 1985, pp. 497-542. (29) Blank and Card, “Recent Trends in insured and Uninsured Unemployment,” p. 1179. (30) Baldwin and McHugh, “Unprepared for Recession,” p. 18. (31) Vroman, The Decline in Unemployment Insurance, p. 25. (32) Corson and Nicholson, An Examination of Declining UI Claims, pp. 119-20. (33) Baldwin and McHugh, “Unprepared for Recession,” p. 18. (34) Walter Corson and Anu Rangarajan, “Extended UI Benefit Triggers,” (Princeton, Mathematica Policy Research, 1993, emphasize that this result is unexpected, and suggest that it should be viewed with caution. (35) Baldwin, “Benefit Recipiency Rates,” p. 201.
Daniel P. McMurrer and Amy B. Chasanov are policy analysts at the Advisory Council on Unemployment Compensation.
COPYRIGHT 1995 U.S. Bureau of Labor Statistics
COPYRIGHT 2004 Gale Group