“Change is inevitable; progress is optional” is a favorite quip of Andy Stern, who was president of the Service Employees International Union (SEIU) from 1996 to 2010. He took that adage to heart. Indeed part of his appeal to my generation of union organizers—who witnessed firsthand the weakening of labor over the last three decades—was his call for new models of unionism that would restore it as essential to a good society.

Stern’s new book, Raising the Floor, thus reflects a startling reversal. After leaving the SEIU, Stern spent half a decade talking to venture capitalists, CEOs, and worker advocates who were trying to envision the future of work. He came to believe that “unions, as important as they had been, were no longer the only or even the best way to achieve what would matter most to American workers in twenty-five years.” What happened?

The putative robot revolution. Stern, along with many business leaders and economists, believes that as robots and learning machines infiltrate more and more fields, there will be substantially less work to go around. This has led a growing chorus in elite foundations and Silicon Valley to embrace a policy traditionally associated with the left: a universal basic income (UBI). Stern joins the crowd, proposing a UBI of $12,000 a year, a level that will eliminate grinding poverty without discouraging work and liberate citizens to care for children and elders, start companies, get more education, or volunteer in their communities. Drawing on his experience pushing through Obamacare and working on the Simpson-Bowles commission, Stern also suggests that conservatives and progressives in D.C. could unite behind that proposal, and he maps out funding options.

The fight for basic income has found surprising support from a growing chorus of elite foundations and Silicon Valley.

I greatly admire Stern’s political creativity and his willingness to challenge labor movement orthodoxy. Several years ago union staffers contacted me looking for “big-think” ideas around low-wage worker organizing and economic inequality. I too suggested a UBI, among other things, but was told that idea was a nonstarter. Union members who work for a living would not want their dues spent on policy innovations for non-workers, they seemed to feel, and in any event the union had far greater and more urgent challenges to address. I found that response frustrating, and still do. If designed appropriately, a UBI would be an excellent antipoverty program and would reduce economic insecurity for those above the poverty line.

I am less sold on Stern’s broader political vision. Like many others in the future-of-work debates, he presents UBI as an urgent policy solution to a problem—massive technological unemployment—that has not yet materialized. Also like many others, he signals a willingness to cut welfare benefits in order to get conservatives on board. These tropes of mainline debate are unfortunate. They alienate progressives who might otherwise favor a UBI and cast the policy as a one-stop solution for economic inequality, which it is not. Mainline debate may also be politically naïve. Both to pass a UBI and to ensure its long-term viability, we need to limit rather than enhance market ordering.

Roughly speaking, there are two schools of thought about the risk technology poses to employment. The first says that we are at an inflection point in our economy or even in the history of capitalism, one driven by the use of advanced information technologies to remake ever more swaths of economic activity. Stern stands firmly in this category. So does journalist Martin Ford, whose Rise of the Robots (2015) is an excellent primer both on the underlying technologies and on their impact on blue-collar and white-collar jobs. The basic transformation, according to this camp, is driven by the automation of routine tasks. This process is, of course, as old as the steam engine; what’s new are the steadily lower cost of processing power, the gathering and analysis of enormous data sets generated by global internet use and mobile computing, and breakthroughs in computer science that enable new sorts of machine learning. The results include robots far more supple than those that populate assembly lines and AI processes that can mimic a wide variety of human judgments.

An oft-cited study by two Oxford professors estimates that such developments could displace 47 percent of all U.S. employment over the next twenty years. As Ford makes clear, the threat extends to many jobs that have in the past seemed immune to automation, including those in warehouse work, food services, customer support, translation, and of course taxi, bus, and truck drivers—who, importantly, are the largest source of male employment in many states. In fact, Ford argues, even doctors, lawyers, journalists, consultants, and other white-collar workers are vulnerable. Automated document review has changed litigation practice significantly in recent decades, and IBM’s Watson, the self-learning machine that won Jeopardy in 2011, has now spawned a product that claims to do much of the basic legal research through which early-career associates at large firms cut their teeth. If it works well, we may see less hiring by such firms.

The other school of thought, perhaps best exemplified by the work of MIT labor economist David Autor, points out that we’ve had this debate many times before—from the Luddites in the 1810s, to John Henry in the 1870s, through 1960s’ fears of computerization that even Martin Luther King, Jr., voiced in one of his final speeches. Each time, however, technological progress has led to creative destruction rather than ending work as we knew it. At the end of the Civil War, over half the country’s population labored on farms; by 1950 only around 12 percent did. The transition to an industrial society was very painful, but displaced agricultural workers typically could—and did—find jobs. The transition to a service economy over the last forty years has also been wrenching, but again opportunities have shifted, not evaporated.

The threat of automation may focus the mind, but technology is already harming workers in a thousand smaller ways.

Autor also asks that we acknowledge the limits of automation. Tasks that require tacit knowledge or emotional intelligence cannot be regularized as easily as assembly or data analysis can. Jobs requiring such skills may change thanks to technology, but they are very unlikely to vanish—including jobs in childcare, elder care, education, social work, personal services, and customer-oriented positions in hotels and restaurants. Lawyers and law schools will adapt to any artificially intelligent legal research tools by emphasizing other core lawyering skills such as writing, advanced legal analysis, and client counseling. Some of the more breathless writers on automation—not Stern or Ford—even imagine a future in which robots perform installation and maintainence tasks in homes and businesses. That prospect strikes me as extremely unlikely, since by definition houses and offices have different physical spaces, fixtures, wiring systems, and the like, making task regularization extremely hard, at least in the foreseeable future.

As is probably clear, I am more sympathetic to the second school of thinking about technology and jobs. While the threat of automation may focus the mind, it seems to me far more likely that advanced technology is harming workers through a thousand smaller, different cuts. Lower communications and monitoring costs make outsourcing, subcontracting, and offshoring easier, severing the legal and social links between firms and their workers. Big data analytics help retailers predict the bare minimum their suppliers will accept for goods, driving down profits and workers’ wages. The network effects in two-sided market platforms give firms such as Uber enormous power to drive down compensation for drivers or to terminate them without due process. Background checks can uncover minor criminal histories or financial difficulties that eliminate applicants from consideration. Algorithmic scheduling leaves baristas and fast food workers without predictable hours. And all this plays out against the backdrop of globalized trade and investment, financialization, and the erosion of the welfare state, all of which shift market risks to workers.

The result is that those with jobs are increasingly divided into haves and have-nots. The haves include those in the professions along with a small number of well-compensated managers at leading firms who develop production strategies. The have-nots comprise a growing number of freelance knowledge workers hired on a project-by-project basis and an enormous number of less skilled service workers. The have-nots are losing out along multiple axes: they lack decent compensation and reasonable hours; their work is precarious; and they have no formal representation in the workplace or the economy. Needless to say, women, people of color, non-citizens, and the poor suffer the most. Panic over automation just makes matters worse, allowing brick-and-mortar firms to resist worker demands for higher wages or paid sick leave: if workers start to ask for too much, the market reasoning goes, robots will look more and more attractive. The future of work is already here, in other words, and it looks grim.

If this is right, a UBI would be a godsend even if unemployment falls. By placing a floor below earnings, it would alleviate dire poverty and protect workers against the existential risks that accompany job losses today. It would also make it easier to move between jobs, relocate for work, or take time out of the workforce for training, education, caregiving, or just a salubrious break. Perhaps most important, it would sever the link between work and welfare, enabling all to obtain basic resources without the shame of an adversarial application process. And as Stern argues, if workers have a guaranteed minimal income, they’ll be better able to stand up for themselves at work—and in voting booths and civil society.

But a standalone UBI can only do so much for workers, as Robert W. McChesney and John Nichols elucidate in their contribution to this debate, People Get Ready: the Fight Against a Jobless Economy and a Citizenless Economy (2016). McChesney and Nichols survey the same landscape that Stern and Ford do, but they argue that the challenge we face runs much deeper than automation.

Their target is three decades of deregulated markets, which have created an oligopolistic political economy. Thanks to the weakening of antitrust enforcement and the explosive growth of tech companies with characteristics of natural monopolies, a handful of firms now dominate many industrial sectors: Walmart, Target, and major grocers in retail; Apple, Google, and Facebook in consumer tech; Amazon, Uber, FedEx, DHL, and UPS in logistics; fast food brands and restaurant chains in food service. That the finance sector sits atop this pyramid should by now be a matter of judicial notice. Without powerful unions or the support of the state, workers can’t hope to counteract such firms in either the economic or the political sphere.

The technological determinism of many UBI proponents only reinforces the neoliberal state.

In my view, our current path of technological development will only make matters worse. For one thing, robots themselves will become an important capital investment. As Ford notes, Uber’s move toward autonomous cars may heavily centralize car ownership, since many consumers will no longer feel a need to have an automobile. But Ford doesn’t follow that argument to its conclusion: that if one company controls urban transportation, that company may become politically untouchable. What’s more, Uber’s power pales in comparison to that of some other Silicon Valley firms. At a recent conference I argued that Google could soon become a state-within-the-state. Several technology policy experts laughed—not because they thought I was being dramatic, but because they felt we crossed that Rubicon some time ago. The massive investments in data-gathering and processing necessary for advanced machine learning make it the provenance of a handful of firms, who will then control who can use their AIs, and how.

McChesney and Nichols therefore propose broad structural reforms to create a more egalitarian political economy: rejuvenating antitrust enforcement to break up large firms, treating certain essential companies as public utilities, limiting the activities of financial firms, reforming campaign finance laws, and creating multiple paths through which citizens can build and exert countervailing power, including rejuvenated unions. And they call for nothing less than a basic restructuring of our economy to ensure that everyone enjoys basic goods—food, water, education, health care, child care, transportation, and wifi—by making them the state’s responsibility rather than leaving them to market forces.

We need more of this kind of thinking in the future-of work-debates. A stand-alone UBI looks remarkably tame in comparison—a variation on the traditional tax-and-transfer welfare state rather than a fundamental redistribution of property rights. In fact, McChesney and Nichols dismiss a UBI as little more than a means to bolster consumerism. If we’re going to have a major fight over work and welfare, they say, we should “fight for a world where [our] concerns are central, and not struggle to be extras in a world of, by, and for the rich.” That’s absolutely correct, but it shouldn’t rule out a UBI, which could be layered atop a liberal economy or a social democratic economy. Nothing in the policy precludes a more fundamental rethinking of social and political relations. Still, McChesney and Nichols are right to highlight some very problematic aspects of mainstream UBI support.

Many UBI proponents embrace market ordering and consumer welfare to the detriment of democracy and social equality. Historical support for the policy among neoliberal giants Friedrich Hayek and Milton Friedman is a frequent reference point, including by Stern, and leading advocates have often suggested that a UBI would enable labor market deregulation. While Stern endorses minimum wage hikes and enhanced collective bargaining rights in the short term, he echoes deregulatory logic in places, at one point stating that a UBI would make “a minimum wage less necessary.” The underlying argument is now canonical within welfare economics and legal scholarship: taxation is a more effective means of redistribution than regulation is, and it doesn’t contract markets the way regulations can. Policymakers can therefore ensure both efficiency and equity through market ordering and transfers.

We must fight for a UBI as one component of broad structural reform to create a more egalitarian society.

But that argument has glaring flaws. Applied to minimum wage laws, it disregards the emerging consensus among leading economists that raising the wage within historical limits will not reduce work opportunities. It also disregards the close relationship between economic power, political power, and social power that McChesney and Nichols highlight. And in some cases it just plain misses the point. As I’ve argued elsewhere, minimum wages do far more than just alleviate poverty. They help undermine status hierarchies and relationships of domination in the workplace, bolster the state’s legitimacy with workers, encourage norms of shared sacrifice, and deter certain very punishing forms of work. Such laws will be even more important if a UBI is limited to citizens; otherwise non-citizens will perform all menial work for meager wages, as citizens ascend to a sort of rentier class.

Support for a UBI coupled with labor market deregulation makes sense, however, if one thinks that markets are the optimal form of social cooperation—or at least the best we can realistically hope for—and if one distrusts the regulatory state. The technological determinism of many UBI proponents only reinforces this worldview by suggesting that employment regulations encourage further automation.

A similar faith in market ordering and skepticism toward democracy may explain Stern’s rejection of a “participation income”—a basic income conditional on performing some volunteer work or civic service. Such a program, he writes, would spark “endless debates on what we, as Americans, should value and do, instead of leaving those issues up to the more efficient, free-choice mechanism of people simply spending their money.” But isn’t that precisely the debate we should be having? Depending on its institutional design, a participation income could emulate workfare and poor laws, or it could be radically democratic. Moreover, even if we were to opt for a UBI rather than a participation income, there is no reason simply to shunt all social choice processes thorough the market. We could also encourage deliberation around social goals, including priorities for state funding and what sorts of goods should be provided by the state, the market, and the civic sector.

To be clear, I don’t think Stern has completely given up on democracy or egalitarian politics. But he is operating within mainstream UBI debate, which tends to endorse or at least presuppose a neoliberal state—one limited to creating and policing competitive markets—and is therefore prone to overemphasize entrepreneurship. Norms of shared responsibility, in this worldview, should not be enforced by the state, nor should the state concern itself too much, if at all, with “private” relationships like employment. This vision is both morally unappealing and contrary to lay conceptions of economic fairness, which affirm social equality and reciprocity among citizens. A distributive scheme that taxes income while leaving underlying property rights intact will be unstable, moreover, since those bearing the tax burden will surely resist it with all their substantial power.

These are not minor concerns. They get at deeper problems with our faith in markets: social organization always involves politics, and what looks like spontaneous ordering is always enabled and constrained by institutions. As Karl Polanyi showed, the most acute moments of conflict over technological progress occur when it enables economic practices to outrun social practices. But the solution is not to buy off a hoard of displaced workers armed with pitchforks. Rather, it is to fight for a UBI while rebuilding a robustly democratic state that can enact broader progressive reforms to tax and welfare policies, subject large firms to far greater oversight, and protect workers and the unemployed in the here and now. Such reforms aren’t just good policy but also good politics, since they can help rebuild faith in the state. A standalone UBI simply cannot.