STANFORD – US President Donald Trump’s tyrannical and lawless behavior must not obscure the key factor underlying the decline of contemporary American democracy.
To provide an antidote to Trump’s brand of politics, we must understand where his MAGA (Make America Great Again) movement came from. Not only has a combination of new technology and government policy allowed for the widespread destruction of workers’ livelihoods over the past few generations, but this desolation has often been accepted as a normal, even necessary, economic outcome.
The problem is that the elimination of millions of jobs entails a profound loss of human and social capital. Families were torn apart as incomes collapsed, and as drug addiction and suicide spread.
So savage have been the results that workers without college degrees now have shorter lifespans than in the recent past, a finding that the economists Anne Case and Angus Deaton have examined rigorously and persuasively.
In a follow-up study to their 2020 book, Deaths of Despair and the Future of Capitalism, Case and Deaton show that life expectancy for college-educated Americans in 2021 was 8.5 years longer than for the two-thirds of American adults who lack a bachelor’s degree.
Even more to the point, while the first group’s life expectancy has risen since 1992, despite a fall during the COVID-19 pandemic, the life expectancy of less-educated Americans has fallen.
These pathologies are the result of job destruction over the past half-century. In my forthcoming book, Private Power and Democracy’s Decline, I show that since the 1970s, job destruction has caused growing economic and political inequality, as well as increased social polarization, with workers who lack a college degree pitted against more educated Americans. It is these trends that are driving the precipitous breakdown of democratic institutions that we are now witnessing during the second Trump presidency. The rise of MAGA populism is a direct outcome of these destructive effects on blue-collar labor.
Making matters worse, we now face even larger-scale, faster-paced changes to our economies with the widespread use of AI. The technology is raising both hope and fear: hope that it will boost economic productivity and improve human well-being; and fear that it will be designed to replace human labor, diminish human creativity, and cause extensive job losses.
How should society respond to these negative effects of job destruction? The compensation principle proposes that the winners from any public policy should compensate the losers, so that no one will be harmed by it. The problem is that economic reasoning lacks the tools to make comparisons between people’s gains and losses and thus cannot justify such a policy. Indeed, it cannot justify any redistribution. Because there is no compelling economic reason to justify this cause-no-harm principle, it is usually ignored (or, at best, paid lip service).
From a worker’s perspective, the loss of livelihood is a grossly unjust outcome, attributable to choices made by indifferent and misguided elites. The answer to this dilemma is to accept that the malign impact of massive job destruction is a political problem that can be addressed with economic tools. An economic-policy intervention is then justified by the political aim of preserving the stability and legitimacy of the democracy that produced the malign outcome.
Thus, I would propose an economic-policy intervention designed to achieve a clear political objective: to preserve the legitimacy of democracy. While democratic societies should not resist technological and economic changes that might accompany AI, they must neutralize the political impact of such changes with a livelihood-preservation policy.
Such a policy would have two parts: the first would create an incentive to spur innovations that support labor; the second would ensure a restoration of livelihoods for workers whose skills are eliminated because of technological and economic changes that public policy supported.
Labor-Enhancing Innovations
The first part encourages AI developers to pursue innovations that support workers instead of replacing them. Since all AI algorithms involve information creation and processing, these two are crucially different. For example, the prevailing approach is to let an AI algorithm run independently, even though we do not know precisely what it will do and how it will perform. But one can also imagine an algorithm that is designed to provide a human operator with the information he or she needs to make decisions and perform their tasks more efficiently.
When it comes to protecting workers and, thus, democracy, this difference is crucial. The first approach displaces human operators and carries the inherent risks common to all AI algorithms, which still have many unresolved defects that make their output and performance unreliable, and potentially downright dangerous. The second approach, by contrast, involves a machine cooperating with humans, and AI not only increasing labor productivity but also reducing many of the risks associated with machine autonomy.
To change the incentives, I propose that all AI-based products and services sold on the market be taxed at 10% of their market value.
The producer can apply to have the product or service classified as “labor-supporting,” but must provide proof for such a claim to a public board made up, in equal parts, of technologists, government officials, and academic engineers. Any product or service approved as “labor-supporting” would not only be exempt from the tax but would also be awarded a 5% subsidy.
The justification for the tax lies in the fact that a private firm’s decision-making or strategy does not consider the potential adverse effects on society. This is in contrast with Silicon Valley’s belief that firms should be free to innovate anything they want, and that society should simply deal with all negative externalities as they come.
The above point of view is bolstered by the current tax code, which taxes labor income heavily, creating an extra incentive in support of labor-replacing automation. It is obvious that, in addition to introducing new taxes on labor-saving innovations and subsidies for labor-supporting ones, the present negative approach to labor taxation should be reconsidered.
Livelihood Restoration
The second part of the policy is a livelihood-restoration agenda, which would address the loss of human capital and degraded livelihoods caused either by public policy or private acts supported by public policy. Restoring livelihoods entails far more than simply compensating a worker for lost wages. In fact, the evidence shows that workers who are eligible for such programs do not want to become welfare burdens. Rather, they want recognition that they are innocent victims of a historical process for which they are not responsible, and they seek a way to restore their dignity, income, and quality of life.
Consequently, livelihood restoration requires several parallel actions. The first is professional counseling and a selection process to determine one’s suitability for acquiring a new skill. For workers who are too old to be retrained, dignity in retirement is the only remedy.
The policy’s central goal is retraining. This requires a program that develops community colleges, technical schools, local vocational institutions, or apprenticeships at the former employer or with any other firm where the worker can be trained. Such workers will need not only support for living expenses while training, which may take as long as a year, but also family counseling, medical insurance (either by the health plan of the prior employer or by securing a new one), and coverage for daycare (if there are children) and moving expenses if the family decides to relocate elsewhere to take a better job. Finally, I also propose an employment subsidy of 20% for one year to encourage firms to hire retrained workers.
Knowhow Workers
The rationale for pursuing a restoration policy, rather than allowing market forces to solve the problem of displaced and unemployed/unemployable workers, is straightforward: Work experience is a significant source of income for workers without a college degree, who account for the majority of all workers (and voters). When these workers lose their jobs, they lose most of their human capital as well, because it is tied to their acquired skills.
Such specialized skills cannot simply be transferred from a declining industry to a new career in an expanding industry.
The evidence shows that if the problem of job destruction is left to be solved by the market, two-thirds of displaced workers end up taking low-skill and low-wage jobs with lower income, while around one-third leave the labor force altogether.
Moreover, the permanent loss of a job and career is often a cause of emotional trauma that destabilizes family life and, in a significant number of cases, results in poor health and shorter life expectancy, as noted earlier. In a democratic society, those who suffer such losses from public policy will naturally consider the policy unjust, and when enough people fall into this cohort, they will form a critical mass with political consequences.
As a powerful electoral group that is hostile to the system itself, they tend to exacerbate the social and political polarization that ultimately weakens democratic institutions. When their ranks grow as large as the MAGA constituency that has twice propelled Trump to the presidency, their desire for retribution can produce truly destructive results.
The evidence presented in my forthcoming book shows that democratic countries that made greater efforts to preserve the livelihoods of their workers have enjoyed greater social and political stability. For example, Japan, Germany, and the Scandinavian countries all approached the problem in slightly different ways, but achieved similar overall results in terms of safeguarding their institutions.
Brace for Impact
AI will have a very different impact on employment than have previous waves of technology. It will be concentrated in the service sector, rather than in manufacturing (which has already been heavily automated), and it is likely to affect technical occupations that comprise activities based on extensive empirical records: medical, legal, scientific, or simply bureaucratic.
The reason is that AI excels at the rapid synthesis of recorded data. Since these skills require higher education and significant investment, the economic impact of future job elimination by AI may be more significant than the cost of lost manufacturing jobs due to automation. That makes a policy to restore livelihoods even more urgent than it already was.
After all, those who become economic casualties of the AI revolution will have every reason to blame public policy, which has placed no safeguards or hurdles before this rapidly evolving technology.
On the contrary, the Trump administration has gone out of its way to untie Silicon Valley’s hands, including by refraining from antitrust enforcement, facilitating deal-making to build data centers and other infrastructure, and issuing executive orders and public statements rejecting the very idea of regulation or AI governance.
That means workers of all stripes will be on their own. While that is not a new experience for US workers, many more could soon be displaced. Unless robust livelihood protections are introduced, the prospects for preserving or reviving democratic institutions are grim.
Mordecai Kurz, Emeritus Professor of Economics at Stanford University, is the author of The Market Power of Technology: Understanding the Second Gilded Age (Columbia University Press, 2023), and the forthcoming Private Power and Democracy’s Decline: How to Make Capitalism Support Democracy (MIT Press, 2026).
Copyright: Project Syndicate, 2025.
www.project-syndicate.org
