A federal job guarantee (JG) is one of the hardy perennials of American politics. Such a guarantee would offer public-service employment (PSE) to anyone who wanted it, with government at some level or an approved nonprofit organization as the employer.
The idea of a job guarantee traces its roots to the Great Depression, when the federal government created thousands of jobs through programs like the Works Progress Administration and Civilian Conservation Corps. Later, in his 1944 State of the Union address, President Franklin Roosevelt put the right to a job at a living wage at the very top of his “Second Bill of Rights.” In the 1970s, a job guarantee was again proposed as part of the Humphrey-Hawkins Full Employment Act, although it was dropped from the draft before that bill was passed.
Today, the idea is undergoing a revival as part of the Green New Deal, introduced in Congress in 2019 with more than 100 co-sponsors. Some might find the timing odd, with the unemployment at a 50-year low. In reply, however, JG proponents point to three gaps that persist even when official data point to a tight labor market:
- A hidden
unemployment gap: A gap between the number of people counted in
the labor force and the number who would seek work if jobs were available
at a higher wage.
- A pay
gap: A difference between what people are now paid and the
maximum that their employers would be willing to pay if necessary.
- A public
service gap: A large, unmet need for labor-intensive public
services that would generate benefits equal to or greater than the cost of
This three-gap model of the labor market will serve as a convenient device for organizing this commentary on guaranteed jobs.
Our discussion will focus on two of the most fully developed JG proposals currently in circulation. One is outlined by Mark Paul, William Darity Jr., and Darrick Hamilton (PDH) in a report commissioned by the Center on Budget and Policy Priorities. The other comes from L. Randall Wray, Flavia Dantas, Scott Fullwiler, Pavlina R. Tcherneva, and Stephanie A. Kelton of the Levy Economics Institute.
The two proposals have much in common. Both would offer a full- or part-time job to anyone over the age of 16 (Levy Institute) or 18 (PDH) who wants one. The federal government would cover the full cost, but in most cases, it would not be the actual employer.
In the Levy Institute plan, the federal government would finance and supervise the program, with state and local governments and nonprofit organizations as the employers of record. The PDH plan calls for the formation of a permanent National Investment Employment Corps, which in turn would administer employment grants to eligible entities, including state, county, and local governments, as well as Indian Nations, to engage in direct employment projects.
Both proposals specify a generous package of benefits, including health insurance, child care, retirement, and parental leave, among others. Both call for a living wage — that is, a wage high enough to ensure that one earner can support a family at an income above the poverty level. Under the PDH plan, the wage would vary with experience and qualifications. In the Levy Institute version, all participants would receive the same wage.
The following table gives estimated costs per job for the two plans:
In many respects, both proposals reflect the economic conditions that prevailed at the time they were published – 2018 for the PDH version and 2019 the Levy Institute. The following are among the assumptions that both proposals make, sometimes implicitly and sometimes explicitly:
· A JG would be useful as an automatic stabilizer in times of economic downturn, but it would also serve important structural purposes even if introduced at a time when the labor market, by conventional measures, were close to full employment.
· The JG wage at the time the program goes into effect is assumed to be significantly higher than the prevailing national minimum wage.
· Both before and after implementation of the JG, employer sponsored insurance will be the predominant source of health care coverage, both for PSE workers and those who continue to work in the private sector.
From time to time in the following discussion, I will point to ways in which these assumptions are important for assessing the merits or problems of a JG program.
JG and the hidden employment gap
Estimating uptake for a JG program. The number of applicants for guaranteed jobs would depend on the size of the hidden unemployment gap — the difference between the number of people now working and the number who would take a guaranteed job if it were offered. Here are the Levy Institute’s estimates, based on data for the third quarter of 2017, at which time the official unemployment rate was 4.1 percent:
million to 6.1 million workers from the officially unemployed — that
is, those who are not working but actively looking for work.
million to 5.7 million people from those currently out of the labor force
but who say they would like a job.
million to 6 million people who would like to work full-time but are
currently working part-time either because they cannot find full-time work
or because of child care responsibilities.
The total comes to around 12.6 to 17.8 million people, or about 8 to 11 percent of a labor force that currently numbers about 160 million.
Using 2018 data, the PDH provides a somewhat lower estimated uptake of 10.7 million people. Most of the difference is attributable to some 3.5 million people who say they want a job but have not looked for work in the previous 12 months, and are thus not classified by the Bureau of Labor Statistics (BLS) as “marginally attached to the labor force.” The Levy Institute counts them as likely candidates for PSE, but PDH does not.
In practice, there seems to be little difference between the willingness to work of this group and those who have looked for work in recent months. One analysis of labor market transition data, for instance, notes that people who want a job but have not looked within the past year are almost as likely to find employment in any given month as are those who are officially “marginally attached.” If both groups are included, there is little difference between the uptake rates expected by PDH and the Levy Institute.
For the sake of discussion, let’s accept some number in the range of 12 million to 18 million as the number of people likely to apply if guaranteed public-service jobs were offered to all who wanted them. That is not the whole story, however. If we want to evaluate JG as anti-poverty policy, we need to look more carefully at the characteristics of likely applicants, not just at their numbers.
Nonemployment and financial stress. There is no question that many people who want to work but can’t find a job suffer serious financial stress, but that is not true for all of them. Consider, for example, the results of a survey of the nonemployed conducted in 2014 by the Kaiser Family Foundation, the New York Times, and CBS News. Here are some key findings of the survey:
all nonemployed prime-age individuals, 36 percent said their employment
situation was not a source of stress and 20 percent said it was only a
minor source of stress.
those who said they were nonemployed but able to work, 26 percent
self-identified as homemakers. Among self-identified homemakers, 77
percent said their nonemployment was not a source of stress and another 15
percent said it was only a minor source of stress.
all prime-age nonemployed adults, 51 percent reported they were very or
somewhat financially secure. Among self-identified homemakers in this
group, 76 percent reported they were very or somewhat financially secure.
The nonemployed, nonstressed people in this survey were prime-age adults whose financial needs, in most cases, were presumably being met by the earnings of other members of their households. The survey covered only people of prime working age. It did not include teenage or college-aged people living at home but looking for part-time work while studying. Nor did it include older people who wanted part- or full-time work to supplement an already minimally adequate retirement income, or simply to get out of the house and do something interesting.
It is likely, then, that at least some guaranteed jobs would be taken up by people who were motivated by a desire for self-fulfillment or extra pocket money rather than a need to escape poverty. Providing those jobs would increase the cost of a JG program and, at the same time, would dilute its rationale as an antipoverty measure. Additional surveys or results of JG pilot programs might shed light on the number of people in each category. Meanwhile, the reported survey suggests that the issue is far from trivial.
The hard-to-employ. The long-term unemployed and those who want to work but have given up looking are key target categories for JG policy. By their nature, they include many people who are, for one reason or another, hard to employ.
JG advocates recognize that the workers they seek to attract are more likely to have less education and lower skills than the average employed worker. The proposals we are looking at do include at least a modest training component. However, there is more to employability than education and training. The hard-to-employ also include people with criminal records, unstable housing, substance abuse issues, family situations that interfere with regular work schedules, borderline mental and physical conditions that fall short of actual disability, and other problems that make it hard to hold a job.
Studies of past and present employment programs show that such people can benefit from work opportunities outside the normal labor market, but they also suggest reasons for caution. Consider, for example, the results of a set of social-enterprise experiments in California, as reported by Mathematica Policy Research. Social enterprises are mission-driven, nonprofit businesses that are specifically established to hire the hard-to-employ. The enterprises that Mathematica studied offer work in service positions similar to those for JG beneficiaries.
An important finding of the Mathematica report was that success required extensive support services. Those services included not only job-related training (as recognized by JG advocates), but also general training in job-readiness and personal finance skills; assistance with clothing, transportation, and housing; food pantries; nutritional education; counseling to avoid relapse into drug dependency or criminal behaviors; education regarding public benefits; and tax preparation.
Another takeaway is that even with extensive support services, the employment gains of the hard-to-employ, although statistically significant, were quantitatively modest. A year after entering the program, the rate of employment among participants rose from 18 percent to 51 percent; the number in stable housing rose from 15 percent to 53 percent; and the share of income from government fell from 71 percent to 24 percent. To put it concisely, at best half of the participants actually achieved the goal of self-sufficiency.
Similar findings emerge from studies of work requirements, which use a stick rather than a carrot to move people from welfare to work. The gold standard for evaluating work requirements is a set of 11 controlled experiments known as the National Evaluation of Welfare-to-Work Strategies (NEWWS), which were conducted as part of the welfare reforms of the 1990s. The experiments each lasted five years and were conducted in various cities around the country. Each compared a group of people whose benefits were conditioned on work requirements with a control group who continued to receive welfare as usual.
Some of the NEWWS experiments did produce modest gains in employment. Importantly, though, they did so only where they were backed by intensive administrative support, adequate funding, and well-trained staff. Case workers and other administrators had to do more than simply monitor eligibility and compliance. In the most successful experiment (Portland), case workers interacted one-on-one with participants to cajole them into jobs or training programs, or to coerce them to make greater efforts by threatening withdrawal of benefits. Experiments in Oklahoma City and Detroit, where staffing and administrative funding were lower, reported no statistically significant increase in employment.
The authors of the PDH and Levy Institute plans recognize that some applicants may have difficulty adjusting to the JG program. However, rather than looking to supportive case work for such employees, they emphasize discharge or discipline. For example, in a reply to critics, L. Randall Wray, a key member of the Levy Institute team, writes:
The JG should not devolve to either workfare or welfare … Workers can be fired for cause — with grievance procedures established to protect their rights, and with conditions on rehiring into the program.
Similarly, the PDH report says:
For employees to receive their compensation, they must show up to their job and perform the tasks assigned to them. As was the case with the WPA, a Division of Progress Investigation (DPI) should be established to monitor shirking or corruption. If workers are found to be negligent, or generally disruptive to the workplace, disciplinary action can be taken by the DPI.
In my view, neither of the JG proposals considered here pay enough attention to the potential challenges posed by the hard-to-employ. Either these plans should budget substantial additional amounts for administrative support and one-on-one case work, or they should cut back their expectations for the number of people that their programs will succeed in moving from poverty to self-sufficiency.
JG and nonmarket production. Finally, JG advocates need to give more attention to the fact that many potential applicants are already working, but not in the market economy. Both the PDH and Levy Institute teams point to an increase in GDP as one of the benefits of a JG program. However, to the extent such a program simply replaced nonmarket production with market-based production of similar goods and services, changes in measured GDP would overstate those benefits.
Caregiving is one example of a service that is already being produced, but not for pay, by JG candidates. BLS data tell us that many of the people who are out of the labor force but want a job are unavailable for paid employment because they are caregivers for children or other family members. The same is true of many who are working part-time but would prefer full-time jobs.
Consider, for example, two parents, in two separate households, who have each been staying home to care for two young children. If a JG program spent $56,000 to hire one of them to work in a community arts program and hired the other as a day care worker to look after all four children, measured GDP would rise by $112,000. However, the only new services produced are those of the arts program. The day care services were already being produced by the parents themselves, but were not counted in GDP, since they were outside the market economy.
Additional examples can be found in the vast volunteer sector of the U.S. economy. A BLS report on volunteering found that as of 2015, more than 62 million people aged 16 and older participated in volunteer work. Of those, 5.9 percent devoted more than 500 hours a year to volunteering — a huge amount of work.
Much of that was spent on exactly the kind of jobs that would be offered by JG programs. In some cases, people who are now out of the labor force and who volunteer as wildlife monitors would happily take paid jobs as JG wildlife monitors. In other cases, people who are already employed but volunteer to pick up trash in the local park on weekends would find they are no longer needed, since a JG worker is already doing the job. Either way, at least part of the work done by JG participants would displace something already being done by volunteers.
Conclusions regarding the hidden unemployment gap. JG advocates are right when they say that the official unemployment rate, recently at a 50-year low, understates the degree of slack in the labor market. However, an abundance of potential applicants is not enough to conclude that guaranteed public-service employment would be a cost-effective way to move people from poverty to self-sufficiency. To make a convincing case, JG advocates need to pay more attention to three characteristics of the nonemployed and underemployed population they hope to attract:
all of the nonemployed and underemployed suffer financial stress. Many
live comfortably on the earnings of other family members, on savings, or
on other resources.
practice, many people who are not working but say they want a job have
problems that make them hard to employ, such as unstable housing,
substance abuse, criminal records, and mental health conditions that fall
short of full disability.
potential participants in a JG program are already working, but outside
the market economy.
These issues alone are enough to require a rethink of the costs and benefits of introducing a full-scale federal job guarantee at a time when the economy is close to full employment by conventional measures. However, it is only fair to note that problems discussed here would be of less consequence in times of economic downturn. The workers released from private employment during a recession would, by assumption, have demonstrated employability and greater motivation to work than the “hidden unemployed” who remain on the sidelines even as the economy approaches a cyclical peak. On balance, then, the considerations discussed in this section bear more strongly on the merits of JG as an antipoverty program than on its merits as an automatic stabilizer.
JG and the pay gap
Labor migration. An essential part of the case for a job guarantee is the assumption that there is a large gap, at the margin, between what workers are now being paid and the maximum that private sector employers would be willing to pay to keep them on the job. Without such a gap, private employers, if forced by external circumstances to increase wages, would reduce the number of workers they hire or the hours of work they offer. The result could be large-scale migration to guaranteed public-service employment (PSE).
JG advocates do not dispute that a few low-wage workers might make the switch to public-service employment. The Levy Institute proposal estimates that 80,000 to 160,000 minimum-wage workers would make the switch, but that is less than 2.5 percent of the total estimated JG uptake, and less than 0.2 percent of private sector workers who now earn less than the proposed JG compensation. That kind of small-scale migration to public-service jobs is not alarming. As L. Randall Wray, one of the Levy Institute team, writes in a response to critics, “Firms with business models that require that their workers live in abject poverty should find a new business model — or be driven out of business.”
In defending the notion that most employers of low-wage workers would respond to JG mainly by raising wages rather than cutting payrolls, both the PDH and Levy Institute teams rely on studies of minimum-wage increases. Those studies have used data from changes in the federal minimum wage and also cross-border data from similar states and cities where the minimum wage differs. Although there are exceptions, the consensus of the literature seems to be that recent changes in minimum wages have caused little if any change in employment.
However, there are caveats in applying those studies to a broad federal job guarantee. One is that most recent changes in minimum wages have been relatively small. In a comprehensive review of the literature for the Center for Economic and Policy Research, John Schmitt notes that although “the minimum wage has little or no discernible effect on the employment prospects of low-wage workers,” the most likely reason is that “the cost shock of the minimum wage is small relative to most firms’ overall costs and only modest relative to the wages paid to low-wage workers.”
By comparison, the cost shock of a broad JG policy would be massive. The federal minimum wage currently stands at $7.25 per hour, with no requirement that benefits be paid. The version of JG outlined by the Levy Institute envisions a wage of $15 per hour and benefits equal to 20 percent of the wage. (Even 20 percent is probably an underestimate if JG is assumed to offer full health benefits.) The version presented by PDH calls for a starting wage of $11.83 per hour and an average wage of $16.25, with benefits equal to 30 percent of the wage. With benefits included, both would represent more than a doubling of the current federal minimum — a far greater increase than those considered in the empirical literature.
It is worth noting that the importance of these problems depend on the context in which JG is assumed to be introduced. For example, the national minimum wage might first to be raised to $12 or $15 per hour, as some already propose. If time were allowed for the labor market to adjust to the higher minimum, then later introduction of a JG at $12 or $15 would be less of a shock. Similarly, there are a number of proposals on the table for introducing universal health insurance in a form that would no longer be linked to employment. If such a system were in place, then migration from private to public employment would bring no gain in health benefits. If so, introducing JG would, again, be less of a shock to the labor market than we have assumed.
Effects of JG on state, local, and nonprofit employment. A further issue in assessing the scale of potential labor migration to JG from existing jobs is that the minimum wage literature is mainly concerned with effects on private-sector employment. However, a JG wage of $15 an hour plus benefits would also have a big impact on workers in state and local government and in the nonprofit sector. Together, these sectors employ almost 35 million workers, many of them in the same kinds of public-service jobs that a JG program would offer. The migration of even a small percentage of these workers to newly created guaranteed jobs could be quite disruptive.
As of 2019, state governments employed some 5.3 million people and local governments 14.7 million. According to another study from the CEPR by John Schmitt, when differences in age and education are taken into account, those jobs pay about 4 percent less than jobs in the private sector.
JG advocates recognize the potential conflict between existing government jobs and newly created, guaranteed public-service positions. PDH, for example, caution that “it is vital that the program is designed to avoid state and local governments utilizing the program to pay for existing state and local government jobs which would normally be financed through local tax revenues.”
In practice, though, it might not be easy to draw a clear line between existing state and local workers and newly created guaranteed jobs. For example, it is hard to imagine that currently employed teachers’ aides or custodians would continue to work contentedly at $11 per hour, elbow-to-elbow with newly hired JG workers doing similar work for $15 an hour and more generous benefits.
What is more, unlike a fast-food restaurant or retail stores, state and local governments could not simply raise the wages of existing employees and then pass the increase along to customers. Instead of customers, the sources of government revenues are voters. Voters might not agree to having their taxes raised in order to allow states and localities to raise wages to match those of federally-funded guaranteed jobs.
A similar situation exists in the nonprofit sector. Nonprofits currently employ some 14 million workers, nearly as many as local government. Although some workers in this sector are well paid — for example, health care professionals at nonprofit hospitals and professors at private colleges — others are not. According to Nonprofit Quarterly, pay is especially low in the large social-assistance sector, which includes community, food, child and youth, elderly, and disabled services, and in arts and entertainment organizations. The reported average pay for such workers falls well short of that envisioned by the PDH and Levy Institute JG proposals. Yet JG advocates specifically single out this sector as a source of employment opportunities.
Like state and local governments, nonprofits would be unable simply to pass the cost of higher wages and benefits through to their customers. Instead, they would have to solicit additional donations, a task that might be no easier than for state and local governments to raise taxes.
JG and the public service gap
We turn now to the third feature of the labor market model used by JG advocates — the public service gap. This gap posits a large, unfilled need for public services that could be filled at a cost per job at least equal the benefits produced.
To be sure, it is easy to believe that some potentially worthwhile public-service jobs now go unfilled. In the government sector, that might occur because median voters in state and local elections are insufficiently sensitive to the needs of their communities. For example, voters without young children, a majority in many communities, might be reluctant to approve spending on schools. In the nonprofit sector, it might be that wealthy donors like to contribute to prestigious causes like art museums and elite college endowments rather than to less glamorous but possibly more urgent needs like soup kitchens and hospice care.
The real question, though, is not whether some worthy public-service jobs go unfilled, but how many there are. Both the PDH and Levy Institute proposals envision that the new JG positions would be created by local governments and nonprofits. Those two sectors together currently employ in the neighborhood of 30 million workers. Could they really create half-again as many new positions with value-added equal or greater to their cost?
Remember, those jobs do not come cheap. The PDH proposal estimates the cost of each new PSE position to be about $56,000 per year, including wages, benefits, payroll taxes, capital, and supplies. The Levy Institute, which uses lower estimates for benefits and overhead, estimates the cost per job at $45,800. As noted before, these estimates do not include costs of administrative support or case work with hard-to-employ job candidates.
Neither proposal offers an explicit estimate of the value added by new PSE jobs. However, we can get an idea of what to expect if we look at some of the projects that the Levy Institute gives as illustrations:
The city mobilizes men and women with varied skill levels for a cleanup of vacant lots and abandoned public spaces, rehabilitation of infrastructure, and reclamation of materials. People with disabilities who may have difficulty with physical work but have basic computer skills create a database, documenting the cleanup efforts, cataloguing the reclaimed materials, and offering office-based logistical support. At-risk youth help with park cleanup and apprentice with skilled workers in building, painting, and landscaping skate parks and basketball courts.
A former coal-mining community experiences city blight, mass unemployment, and a high incidence of health problems. The PSE program organizes a comprehensive project for restoring the natural habitat based on existing best practices. Workers are employed to plant appropriate tree species that restore the ecosystem, stem soil erosion, and reintroduce important lost wildlife to the region. The municipalities organize food insecurity, water quality, and malnutrition surveys. They launch a comprehensive community garden program.
It is easy to imagine that some of the jobs described in these examples would produce benefits in excess of their $40,000 to $50,000 annual costs, while others would not. Until JG advocates provide more convincing support for the idea that the backlog of cost-effective but unfilled public jobs is commensurate with the number of jobs their programs aim to create, the verdict regarding the public service gap must be, “not proved.”
Our three-gap model of the labor market shows how a job guarantee would work if everything went as hoped: The guarantee would attract millions of long-term unemployed, together with others who have dropped out of the labor market but still want to work. The wage for guaranteed public service employment would be high enough to lift participants out of poverty. Private sector employers would raise the pay of their own employees to match. Services rendered by JG workers would improve the lives of the vulnerable, strengthen communities, and safeguard the environment. The program would pay for itself through an expanding economy and decreased welfare spending.
However, if things were to go wrong — if the gaps were not as large as advocates think — the results might fall well short of expectations.
First, although no one disputes that many people remain on the fringes of the labor force even when official unemployment is low, not all of them fit the target profile for guaranteed jobs. While a job guarantee would raise some people from poverty to self-sufficiency, it would also attract other applicants who are not financially stressed and still others with characteristics that make them hard to employ. Too many of the former would reduce the benefits of the program as an antipoverty measure; too many of the latter would lower retention rates and raise administrative costs.
Second, the success of a JG program would depend heavily on the assumption that employers would increase the pay of their current low-wage workers without greatly reducing their payrolls. JG advocates justify that assumption by reference to the literature on the effects of past minimum-wage increases. However, a full-scale JG program would represent a greater raise for low-wage workers than most recent changes in minimum wages. That could lead not only to more layoffs by private employers, but also to adjustment problems for local governments and nonprofits. Larger-than-expected migration from existing low-wage jobs to guaranteed employment would increase the costs and reduce the benefits of a JG program.
Third, the entire rationale of guaranteed jobs rests on the assumption that there is a large, unfilled need for public services that could be produced by low-skilled workers at a cost per job that would at least equal the benefits produced. JG advocates have not adequately demonstrated that such a backlog exists on the required scale. Large numbers of hard-to-employ applicants; inadequate budgets for administrative support, training, and case work; and restricted allowances for capital and supplies would make it all the more challenging to create 10 to 15 million new, cost-effective, public-service jobs. If pressures to meet employment targets were to lead to the creation of unproductive, make-work positions, the entire rationale for a job guarantee would collapse.
In short, although a strong case can be made for improvements in America’s social safety net, there are many reasons to question how large a role guaranteed jobs should play. Even the most detailed plans, like those from PDH and the Levy Institute, raise more questions than they answer. Perhaps some of those questions might be addressed through small-scale demonstration projects. Meanwhile, it would seem wiser to focus limited resources on more conventional policies like tax credits, training, employment subsidies, and reductions in the work disincentives inherent in traditional in-kind welfare.
For a graphical rendition of the three-gap model of the labor market, see this slideshow.
This commentary is based on a series previously posted at NiskanenCenter.org [Part 1 Part 2]. This version includes certain additions and revisions based on comments by members of the PDH and Levy Institute teams, whom I thank for their input.