For liberals and progressives, the real shocker in the November election is less Trump’s victory than the support he received from the working class in the industrial Midwest, once the heart and soul of the Democratic Party. It is a clear signal that the party has failed to understand and address the concerns of a core constituency. The election thus obliges us—liberals and progressives who vote for Democrats—to critically evaluate our thinking about the problems of the working class. Dean Baker does that, but his contribution speaks only to the beginning of what will be a lengthy process.

The principal problems we need to grapple with are the pressures for structural adjustment and the burden they place on particular groups and communities. In the last several decades, these pressures have converged on the industrial heartland of the country. Trade is one of those pressures. Equally important have been technological change, high-skilled immigration, and “financialization,” the increasing role of financial intermediaries in corporate management. Of course, every society has to absorb pressures for change. Arguably the chief function of an economic system is to absorb pressures and facilitate adjustment to them. The problem isn’t so much change as the distribution of the burden of adjustment.

That burden has unrelentingly fallen on industrial workers. Finance has undermined institutions that traditionally slowed the pace of change and moderated its impact. Investors press firms to renegotiate collective-bargaining agreements, aggressively resist union organization, and relocate to jurisdictions at home and abroad with fewer work regulations. At the same time, while the country has been preoccupied with undocumented, unskilled migration, high-skilled immigrants have forestalled counter-pressures to redesign work and on-the-job training in ways that would limit displacement or allow displaced workers to move into new jobs in expanding sectors of the economy. The result is a large number of uncompensated industrial job losses.

The chief function of an economic system is to absorb pressures and adjust to them.

The textbook approach to structural changes of this kind is to weigh the social benefits against the social costs. On this view, changes are justified if the benefits are sufficient to compensate for the costs. In reality, however, compensation is almost never paid. Workers displaced by trade are eligible to receive help under the Trade Adjustment Assistance (TAA) program, but funding is limited and only a small proportion of displaced workers actually participate. (During the campaign, Hillary Clinton’s major proposal for ameliorating the bad effects of trade was an expansion of this program.) Data suggest that TAA participants do little better than similar workers trying to cope with displacement on their own. In other words, it is not clear that we know how to facilitate the adjustment process. The program does not even try to help workers find jobs comparable to those they lost.

In any case, we cannot address the problems created by structural change after the fact, through adjustment assistance. Recent change has been so rapid and thorough that it shatters the structures of work and local economies in which peoples’ identities are lodged. No amount of retraining or monetary compensation can replace the sense of self-respect that has been lost.

If anyone can recognize this, it should be the Democrats, who emphasize recognition of and respect for disadvantaged groups. We do not propose compensation or adjustment assistance when we resist the deportation of so-called dreamers, the undocumented children of immigrants whom Trump proposes “returning” to places they have never known. Displaced workers in the Rust Belt and Appalachia are rightly indignant at our willingness to protect the identities of the undocumented while we tolerate the destruction of the identities of the working class.

What would it mean to focus policies directly on the pressures for change, to moderate their impact and control their distribution across geographic and social space? It need not imply freezing the structure of the economy. But it probably does imply distributing more evenly the kinds of effects that have converged on the industrial Midwest.

In the case of trade, that means a moratorium on new deals until the impact of existing agreements has been absorbed. In retrospect, it seems sheer madness that the Obama administration was pushing major new treaties among Pacific and European countries while the plight of workers in the Midwest remained unresolved. Whether we should renegotiate current treaties, as Trump has proposed, is an open question. Any retreat from those agreements will have to alleviate pressures, not create new ones aggravating social conflicts, particularly in the impoverished regions along the Mexican border.

The problem isn’t so much change as the distribution of the burden of adjustment.

Responding to technological change is more complicated than handling the effects of trade. There is a broad consensus that innovation and autonomous trends in technology are “upskilling” work. Less-skilled labor is falling by the wayside, and the current generation of manufacturing workers is ill-equipped to fill new jobs demanding more education. This belief tends to become a self-fulfilling prophecy. Technological changes are innovations, their development an excursion into the unknown and surrounded by uncertainty. To resolve that uncertainty, managers and policymakers fall back on conventional wisdom when deciding where to place their bets. This is true not only in the United States but also in developing countries, where there has been overinvestment in higher education, producing a surplus of highly trained engineers and scientists. These workers in turn migrate in large numbers to North America and Western Europe. They staff research and development efforts, which hasten the overall pace of technological change, and mitigate labor shortages that might otherwise force employers to invest in more effective training technologies for reskilling workers and in less demanding technologies of production.

The conventional wisdom is mistaken in consequential ways. Not all recent technology developments require upskilling. From digital calculators to graphical user interfaces, many advancements have reduced skill requirements for workers. And the development of technology in recent decades has in no sense been autonomous. The major shifts in computing, telecommunications, navigation and logistics (e.g., GPS), and biotechnology have been driven by federally funded and directed research. For example, the U.S. Department of Defense has been investing heavily in robotics, which is sure to aggravate worker displacement in the industrial Midwest. All of this creates a major gap between returns to the private sector and the social cost of technological development. A tax on manufacturing robots might bridge that gap, yet this suggestion is jarring in the context of the current policy debate. That it seems beyond contemplation speaks to the distance between political and policy thought and the problems the country actually faces.

In retrospect, it seems sheer madness that the Obama administration was pushing major new treaties among Pacific and European countries while the plight of workers in the Midwest remained unresolved.

The point about the relationship between private and social costs is a general one, applying not just to trade and technology policy but perhaps especially to the institutional structures in which labor decisions are made. Institutions that previously worked to moderate the pace of structural change, to strike a better balance between private and social costs, have collapsed. The most effective of these institutions were trade unions, and not simply the unions themselves but also the parallel benefits to nonunion workers who could threaten to organize.

Those benefits included more than pay hikes. Collective bargaining generated rules limiting layoff, controlling the distribution of work, obliging companies to provide on-the-job training, and imposing high severance costs.

It is unlikely that labor unions will reemerge, but similar restraints could also be achieved by government regulations. Of course it is unlikely such policies will be pursued within Trump’s administration, but we should still think about better alternatives. Developing a program of economic adjustment responsive to worker welfare will enable liberals to recapture their allegiance.