Discussion around the idea of a job guarantee has picked up of late (CBPP, The Nation, Sanders Institute I, II). As with any idea that attracts a sizable number of adherents, the term “job guarantee” means lots of different things to a lot of different people, which makes it a difficult thing to talk about in a comprehensive way. But with that limitation recognized, I thought it might be worthwhile to share some notes about the recent burst.
As part of trying to sell the idea of a jobs guarantee, advocates make claims about poverty that are untrue and contradicted by their other statements.
 What we believe is that anyone whose working full time in America should not be living in poverty.  And that means choosing a wage that will allow a family of three or four to have one person working and to pull that family out of poverty. So we start at $15 an hour.
 The federal job guarantee would provide a job, at non-poverty wages, for all citizens above the age of 18 that sought one. …  The federal job guarantee would provide a job at a minimum annual wage of $24,600 for full-time workers (poverty line for a family of four) and a minimum hourly wage of $11.83.
In both cases, statement (1) claims the job guarantee is going to ensure people who work full time are not in poverty, but then statement (2) clarifies that this is not true for anyone who lives in a family with more than four members.
To provide a wage that really ensured no person who works full time is in poverty, the amount would need to be $26 per hour, as that corresponds to the official poverty threshold for families with nine or more members. Even that is too low though. The Census poverty thresholds arbitrarily cut off at nine members, but families can definitely be larger than that. In the 2016 CPS ASEC file, the largest family unit is 16 members. If you extrapolate out the Census poverty thresholds to a family that size, a true non-poverty wage would need to be around $40 per hour.
Presumably the reason the JG advocates do not propose a $40 wage is because that would be impractically high for many reasons. But what this insight should also tell you is that this is not a particularly good way to eliminate in-work poverty. It is already the case that people who work full time and live alone are not in poverty. What makes workers fall into poverty is the existence of nonworker family members (children, disabled, students, unemployed, elderly) who live with them. The way super-low-poverty countries get that way is by ensuring each of these nonworker populations receives an adequate transfer income: pensions, child benefits, unemployment benefits, etc. Nothing else can get the job done and I’d say it’s broadly unhelpful to the project of poverty eradication to keep pretending otherwise.
Arguments for job guarantees often contain contradictory claims about what effect it would have on inflation, the labor market, and labor demand generally. In most JG advocacy, it is simultaneously claimed that the job guarantee would employ large numbers of people without creating inflation problems and that it will create more competition for workers and thereby bid up private sector wages. Needless to say, these two things are at odds with one another.
 The job guarantee would function as a de facto floor in the labor market, greatly increasing the bargaining position of workers throughout the economy. For private employers to attract employees, they would have to offer a job that is at least as good as the one offered by the government.
 With a job guarantee policy in place, the Federal Reserve can conduct monetary policy without promoting rising levels of unemployment. In this scenario, the job guarantee program can maintain employment and consumption expenditures while the Federal Reserve employs monetary policy to reduce private investment in order to cool the economy.
The job guarantee would function as a robust automatic stabilizer in the economy, maintaining levels of employment during economic downturns through direct hiring, and freely allowing workers to flow from the jobs program to the private sector during economic boom times.
The first statement suggests that the JG will actively compete for workers in the private sector. That is, the JG will lure workers out of their private jobs through superior compensation, thereby creating wage inflationary pressures. In his Nation piece, Sean McElwee makes this claim very explicitly, saying that nurses could quit their jobs at hospitals to work a JG job and that this would put upward wage pressure on the nursing profession.
The second statement suggests that the JG will act as a buffer that absorbs workers in economic downturns and releases them in economic upturns. In this construction, the plan is not inflationary as it does not compete with private sector employment, but only scoops up those who would otherwise not be employed.
This is a tension that runs throughout the JG world.
The last problem I will discuss here is the problem of coming up with suitable jobs. This to me remains the biggest challenge that is really not satisfactorily answered in JG programs. Hugh Sturgess has provided the most eloquent articulation of this problem I’ve seen so far. Sturgess argues that JG job ideas, justifications, and criteria combine to form an “impossible quadrilateral:”
[I]t is not possible for JG jobs to display all of the following qualities: requiring only basic skills (needed for universal eligibility); socially beneficial; indifferent to time (as JG work is countercyclical); and distinct from the rest of the public sector. Any attempt to achieve one criterion requires abandoning another.
These can be divided into three overarching tensions: (1) many suggested JG jobs require skills that are not commensurate with the minimum wage; (2) most social needs are ongoing, so should not be addressed through short-term jobs provided in a countercyclical fashion; and (3) it is difficult to keep the JG and the mainline public sector distinct.
To get more concrete about this point, it’s best to just look at the jobs being proposed. Here’s Paul:
 the repair, maintenance, and expansion of the nation’s infrastructure, housing stock, and public buildings;  energy efficiency upgrades to public and private buildings;  assistance with ecological restoration and services to reduce the country’s carbon footprint;  engagement in community development projects;  provision of high-quality preschool and afterschool services; provision of teachers’ aids;  provision of high-quality elder care and companionship;  rejuvenation of the nation’s defunded postal service;  support for the arts; and  other activities that shall support the public good
Ideas 3, 4, 7, and 9 are too vague to assess. Ideas 1 and 2 require a lot of skill and training, which is not suitable for a JG program. It is also hard to imagine how a local job guarantee administration would manage to pull off building a bridge or a house by itself. Ideas 5 and 6 require considerable amounts of skill and training and are also permanently needed, which make them unsuitable for a cyclical jobs program. Additionally, we generally expect people who work with children, elderly people, and disabled people to be vetted as most are not comfortable allowing just anyone to care for vulnerable groups. Idea 8 is the only one of the bunch that looks promising, though it’s debatable how much social value it carries and it is also hard to imagine that many of the kinds of people who would go through the JG program would want to do “the arts.”
Of course, the “impossible quadrilateral” only presents itself because of the claims made by JG advocates. The challenge posed by it is easily solved once you drop the “socially beneficial” claims, which frankly seem implausible. It seems possible to have a program that absorbs workers in downturns and assigns them to low-value, low-capital, low-skilled (“make-work”) jobs in order to keep them activated. But since this is not a very sexy thing to talk about, proponents inevitably exaggerate in ways that create incoherence.
The left has long said that full employment must be an economic priority. But a program where you create offices around the country that promise to find tasks for anyone who shows up is obviously not the only possible approach here.
For instance, consider a package where the Federal Reserve stops prematurely jacking up rates, the federal government funds ordinary active labor market policies (ALMP) that help match people to job openings, and provides free child care. This is how a more typical country seems to pursue high employment rates, with considerable success
These other approaches to full employment will presumably be deemed insufficient because they don’t truly eliminate all involuntary unemployment in the same way that a take-all-comers approach that directly employs any and all random labor flows does. But this is more of a semantical distinction than anything else. If you take anyone receiving an unemployment check and have them do some sort of task, you can declare that they are employed. Workfare does not traditionally get counted as employment but if you wanted to do that, you obviously could. Whether such people are usefully employed in any meaningful respect is a different thing altogether and the idea that thousands of administrations across the country will be able to usefully employ random flows of labor with random sets of skills in random durations is fairly implausible. This is not a knock on the bureaucracy. The top business geniuses in the country would struggle to run an enterprise like that.