Editor’s Note: This article is adapted from Oren Cass’s new book, The Once and Future Worker: A Vision for the Renewal of Work in America.
The controversial subsidies that New York and Washington offered Amazon to attract its “HQ2” are not some novel approach to economic development. Last year, Wisconsin offered a larger package of incentives to entice electronics supplier Foxconn, assembler of iPhones, to build a $10 billion manufacturing facility in Wisconsin. Annual payroll for 13,000 workers would exceed $700 million, and Wisconsin expected the plant to generate annual state and local tax revenue of $181 million and lead to the creation of 20,000 additional jobs. Critics panned the deal as corporate welfare, to which Governor Scott Walker fired back, “That’s fine, but I think they can go suck lemons.”
Such subsidies defy conventional analysis. Are they liberal? Conservative? Free-market? Statist? Pro-business? Pro-worker? Under the strain of a global economy pulling from one side and a social safety net pulling from the other, the meanings of these terms are colliding. If other countries offer such packages to attract manufacturers, does the U.S. government promote a “free market” by refusing to go that route — or by doing the same? The free-market supporter might prefer an undistorted market, but by endorsing free trade with mercantilist economies, he has already allowed the distortions of foreign governments into the United States. For the unemployed Wisconsinite, “smaller government” is poor consolation when a firm chooses to hire workers in some other country because their government provides it a subsidy while his would not.
Wisconsin got things half right — paying for jobs makes sense. Work has enormous social value for the individuals who engage in it and for the formation and stability of their families, the opportunities of their children, and the vibrancy of their communities. Ideally, the labor market would settle in a place where productive, family-supporting work was available to all people in all places. But nothing in the theory of economics guarantees such an outcome, so, where the labor market delivers an inadequate result, policymakers should seek to alter the conditions in which that market operates.
But Wisconsin’s approach is the wrong one. The state’s offer amounts to a single-user tax cut, eliminating tax liability for a particular investment contingent on its creating the desired jobs. That may be defensible in isolation, but as public policy, it is unwise. Why should Foxconn pay lower taxes than Wisconsin’s many other manufacturers? Why should it pay lower taxes than the many other businesses that will be located in the same town? Perhaps unsurprisingly, costs to taxpayers appear to be rising far beyond estimates, while many of the promised jobs now appear in doubt.
If we really want to “pay for jobs” — and we should — then we should do it directly. As most workers recall with dismay from their first payday, a nasty little line item called “FICA” deducts payroll taxes from every check. What if another line, titled “Federal Work Bonus,” showed that the government had put an additional $3 into your check for every hour worked? That would be a wage subsidy.
* * *
The premise of any subsidy is to encourage economic transactions that would not otherwise occur; a wage subsidy aims to produce that effect in the labor market. Workers unwilling to sell their labor for less than $12 per hour may be worth only $9 per hour to an employer. No job will emerge in that scenario. With the insertion of a $3-per-hour wage subsidy, by contrast, the employer can pay the $9 per hour that the work is worth and the worker can receive the $12 per hour that he demanded. Thus will appear a job where none existed before. A wage subsidy is doubly beneficial because we care not only about how many transactions occur but also about how much workers receive. If a worker earning $9 per hour can take home $12 per hour instead, we may have lifted him out of poverty.
The federal earned income-tax credit (EITC) already operates something like a wage subsidy, offering low-income households large tax refunds that can exceed what they paid in taxes to begin with. But the EITC gets paid long after the income is earned — at tax time the following year — based on an opaque formula. It creates none of a wage subsidy’s immediate, transparent effect in the labor market. How many people, after all, realize that the federal minimum wage is not really $7.25 per hour because, counting the $5,600 credit for which he is eligible, the parent of two children working full-time at that wage eventually ends up with more than $10 per hour worked?
The EITC also skews its benefits heavily toward households with children. A single person working full-time at minimum wage would get a credit of $41, less than 1 percent of what his colleague with kids can expect. As an antipoverty initiative, such targeting may be defensible; as a labor-market intervention, it is not. In many cases, the single, childless worker is precisely the person whom policymakers should be most eager to connect with a job.
As in the $9-per-hour example, the ideal wage subsidy would instead operate as a per-hour payment to the worker, based on his market wage. The value of the subsidy would be set relative to a “target wage” of, say, $15 per hour and would close half the gap between the market wage and the target. A worker would initially receive a subsidy of $3 per hour in this case, equal to approximately $6,000 per year if he worked full-time.
This differs from most programs that transfer resources to lower-income households, including the EITC, which phase out as the household’s total income rises; for every additional dollar earned by the household, the worker loses some of the benefits he was receiving. With the direct wage subsidy, the worker receives the same subsidy for every hour worked at a given wage, no matter how much total income he earns. He can take a second job and earn the subsidy for each of those hours. His wife can take a job and earn her own subsidy, too. The value of the subsidy declines only as workers become more productive, earning promotions and raises.
Subsidizing wages is a particularly well-tailored response to the challenges that globalization presents for American workers. First, the wage subsidy is the appropriate mechanism for redistributing gains from the economy’s “winners” to its “losers.” It comes closest to doing this directly, by taking tax revenue drawn from higher earners and inserting it directly into the paychecks of lower earners. As a result, it demands the least of government and introduces the fewest opportunities for inefficiency and distortion. Perhaps most important, it ties the redistribution to productive employment rather than to its absence.
Second, the wage subsidy offsets subsidies given to foreign producers and moves the cost to employers for domestic workers closer to parity with what firms pay foreign workers living in sharply different social and economic contexts. The benefit is largest for industries where the work is most labor-intensive and relies on the lowest-cost labor — in other words, the industries under greatest pressure from globalization. But it does this through a neutral structure, not through politicians choosing when to intervene.
Third, the wage subsidy helps to sustain communities that lose their tradeable sector. A community lacking the ability to export (even to the rest of the nation) must rely on government transfer payments to fund the resources it requires from the outside world — the community is literally exporting need. The existing American safety net conditions those transfers on very low incomes — often, no work at all — and channels them primarily toward consumption of health-care services. With a wage subsidy, work, rather than unemployment, draws government support, and that support can flow to a fuller range of productive activities in the community. In this model, a services economy can still thrive disconnected from a tradeable sector — not an ideal arrangement but one far better than today’s.
This invites the question, Isn’t the wage subsidy just another form of redistribution, like all the safety-net programs we already have? Yes and no. Yes, it is redistribution. And yes, high-income taxpayers will finance it. But unlike with government assistance disconnected from work, the value of a productive job through which someone supports her family and contributes to her community is not diluted if it yields a paycheck into which the government has put in more than it takes out. Certainly a society of thriving and perfectly self-sufficient families would be preferable. But America is nowhere near such a reality today, and for some people, it may never happen. If we can at least make redistribution a tool for creating jobs and promoting work, we will be moving the labor market in the right direction and delivering better outcomes for those who need support.
* * *
Wage-subsidy proposals have existed for decades, emerging from disparate points on the political spectrum. In 1974, the Urban Institute’s Robert Lerman — then a congressional staffer — prepared a report for the Joint Economic Committee that proposed one. In 1997, Nobel Prize–winning economist Edmund Phelps published a book that proposed one. In 2017, Senator Marco Rubio proposed legislation to implement one in Puerto Rico. Yet the mainstream of both political parties has preferred to retreat behind firmly entrenched positions.
True, the EITC already exists. But it began as a temporary program in the 1970s, expanded in fits and starts, and, as described earlier, is poorly suited to the task of subsidizing wages. Both parties profess support, but neither will contemplate any sacrifices to bolster it. In 2014, President Obama and Representative Paul Ryan both proposed expanding the meager credit offered to childless workers, but Obama would not agree to shifting money from any other program to fund the expansion, and Ryan would not agree to raising any taxes. In 2017, Representative Ro Khanna, a Democrat representing Silicon Valley, proposed more than doubling the credit at a cost of $1.4 trillion over ten years. Lest anyone think him serious, he explained neither how to pay for this nor how it makes sense to deliver assistance to a family with $20,000 of annual income in the form of a $10,000 check the following year.
For many Republicans, any wage subsidy remains a form of redistribution and therefore inherently suspect. Furthermore, it intervenes in the market, encouraging employers to hire workers whom they otherwise wouldn’t under current market conditions. They might accept a subsidy as a replacement for existing safety-net programs, but if cutting safety-net spending is on the table, many would prefer to spend that savings on a growth-generating tax cut.
This became eminently clear in debates over the 2017 tax-reform package, which ultimately increased the ten-year federal deficit by $1.5 trillion for the sake of reducing the corporate tax rate, while failing to deliver even the small EITC increase for childless workers that Ryan had once championed. Indeed, while the Khanna proposal in its 2017 form is not a serious one, even it could have been implemented more cheaply than the tax reform that ultimately passed. The deficit spending would have been equally costly, but at least the labor market and its low-wage workers would have been the chief beneficiaries.
What really infuriates Democrats, meanwhile, is the possibility that employers might benefit. Factually speaking, they have a point. If the government offers a $3 subsidy atop a $9-per-hour job, the result will not necessarily be a $12-per-hour job. The employer might instead cut the market wage to $8, to which the government would add $3.50 — half the $7 gap to the target wage of $15 — leaving the worker with $11.50. Both worker and employer are better off than without the subsidy, but the entire benefit is not the worker’s. How workers and employers respond to the subsidy will vary based on labor-market conditions. What we do know from studies of the EITC and a similar program in the United Kingdom is that, in those instances, roughly 75 percent of the financial benefit accrued to workers. In general, employers have to benefit at least somewhat. A central premise of the wage subsidy is to pull more prospective workers into the labor force. Other things held equal, if the supply of workers increases, then employers will be able to offer lower wages — even as, thanks to the subsidy, workers take home more.
Democrats don’t like this because they view low-wage employers as culpable in creating the conditions that necessitate the support to begin with. Rounding out one of the less coherent agendas in recent memory, Representative Khanna has proposed — along with his massive EITC expansion — a tax on low-wage employers that would “levy a direct fee equal to the public assistance each corporation’s employees are eligible to receive.”
This badly misunderstands how public assistance affects the labor market. Welfare programs cannot subsidize low-wage employers because the programs do not increase in value when low-wage employment relationships are formed. To the contrary, all those benefit recipients would be receiving the same or higher benefits in the absence of their low-wage employment. If anything, public assistance operates as a tax on low-wage employers because workers are likely to lose a share of every dollar earned as their benefits phase out with higher income. The exception to this rule would be the EITC or, better, an actual wage subsidy, which subsidizes work by increasing in value as workers take jobs and earn more.
This leads to the more important point: To the extent that a government program does subsidize low-wage employment, and employers share in the benefit, that is a good thing. Remember, the wage subsidy’s goal is not only, or even primarily, to transfer resources into the pockets of low-income households. It is also to connect more workers with employers in permanent jobs. The task requires employers to do the hard work of hiring and training certain employees whom they otherwise would not, and this benefits society greatly. A central premise of the wage subsidy is to reward employers sufficiently so that they choose to do more. By contrast, just wishing that firms would create more and better jobs when they have no economic incentive to do so is futile; it has zero bearing on what will happen in the actual labor market.
Note that this need to create incentives for the employer is no different from what happens in any other effort at assisting low-income households in a market economy. When people use food stamps at the supermarket, the supermarket benefits. When they use housing vouchers to pay the rent, the landlord benefits. Unless the government wishes to produce everything itself, or order market participants to take actions against their own interests, efforts to deliver results that the market will not deliver for low-income households always benefit the businesses that choose to participate in the transactions. Otherwise, they wouldn’t participate! It is a strange consequence of our commitment to individuals as consumers that we unthinkingly pay hundreds of billions of dollars each year to hospitals and universities to provide treatment and education to customers whom they otherwise would turn away but that we shrink from the idea that society might pay anything to an employer to hire someone he otherwise would not.
* * *
Just as the Republican party’s relative disinterest in the labor market is made apparent by its preference for a tax cut over a wage subsidy, a good distillation of the Democrats’ core attitude toward the labor market emerges from comparing a wage subsidy to their preferred approach: the minimum wage. Raising the minimum wage is the quintessential left-of-center labor-market policy. Unsatisfied with the market outcome, Democrats suggest decreeing a different one. The outcome it professes to deliver is widely desired. It seems “free.” And then it damages, rather than strengthens, the labor market.
The minimum wage and the wage subsidy both aim to raise the earnings of low-wage workers, but whereas the wage subsidy asks taxpayers to make up the difference, the minimum wage asks employers to. “Who pays” is not only about fairness; it also dictates what happens in the labor market. The wage subsidy injects funds from outside the labor market to boost the formation of employment relationships and encourage greater investment in labor-intensive businesses. The minimum wage does the opposite, operating as a tax on low-wage employment that employers have to pay for every low-wage hour they use. We should expect them to respond by employing fewer low-wage workers. Maybe not right away — changing a business model takes time, and investments already made cannot be undone. Rather than immediately cutting back or closing down, it often makes sense to raise some prices, accept some lower profits, and try to recoup fixed costs. A small minimum-wage increase won’t trigger a sudden surge in unemployment.
But if the law’s effect is to tax the employment of low-productivity workers, it will influence which business models succeed, which investments get made, and therefore the trajectory on which the labor market evolves. If the minimum wage is $12 per hour instead of $8, an employer must decide whether to boost the productivity of his $8-per-hour workers by 50 percent, seek different workers who are already more productive, or find a way to make do without those positions. The higher the minimum wage is, the less likely it is that the low-wage workers will benefit. Capital will move away from businesses that employ them and toward ones that use different workers, implement different processes, or operate in different countries.
The roughly $200 billion price tag for a wage subsidy might require some new tax revenue, but its funding could come largely from the existing safety net, which already dedicates more than $1 trillion annually to low-income households — including many with workers. Those workers would benefit greatly from receiving that support in their paycheck instead of through a maze of government programs. Other benefit recipients might start to work if a wage subsidy were offered — for them too, shifting money in that direction is a clear improvement. And then some benefit recipients may see their support decline so that more can go toward supporting work. This is a trade-off, and a good one if the goal of fighting poverty is actually to reduce the number of people in poverty. The safety net’s overwhelming focus on poverty maintenance has always been a mistake; there is no reason to continue it indefinitely.
Any proposal to reduce spending on any anti-poverty program is invariably criticized as cruel to the poor, even when that money would be reallocated to other programs with the same goal. This default creates a bizarre political frame in which the more “generous,” liberal position is to spend new money on new programs, while the “miserly,” conservative position is to reallocate money toward more-effective programs. But allocating money effectively is not a proposal about spending at all, and proposals to reallocate are not stingy or anti-poor. Safety-net reforms could proceed much further if the spending debate between those wanting to increase and those wanting to decrease overall budget levels were put to one side and people across the political spectrum were to focus on the question of how best to spend the money already slated to go out the door. Anyone who believes that the random agglomeration of programs over 50 years of unfocused legislation has produced precisely the ideal is, of course, welcome to defend the status quo. Most people of good faith would, one hopes, agree that some recalibration might be in order.