Democrats have lost ground with the working class for a simple reason: They became “globalist shills.” Under Bill Clinton’s leadership, the party enacted the North American Free Trade Agreement (NAFTA) and normalized trade relations with China, policies that had devastating consequences for the American worker. Barack Obama had a chance to chart a different course, but sought the Trans-Pacific Partnership and failed to renegotiate NAFTA. Such betrayals — and a wrenching experience of economic decline — pushed many working-class voters toward a Republican Party that gave voice to their contempt for liberal elites.

Or so goes one prominent explanation of the Democrats’ difficulties with working-class voters. This narrative has been in wide circulation for years, if not decades, but it attracted renewed attention after Donald Trump won a second term on the back of a multiracial, working-class coalition. (There’s no universal definition of who counts as a “working-class voter,” but for this piece, voters without college degrees serve as a reasonable proxy, even as the demographic includes a tiny minority of wealthy individuals.)

In a recent post on X, the labor economist Arin Dube argued that “30 years of evidence” showed that “free trade policy” had led “many working class voters to abandon Democratic Party.” The political writer John Ganz endorsed a similar perspective, suggesting that Clinton’s free trade policies had caused “the industrial base of the United States” to disappear and that this inevitably cost the party in the Midwest.

Matt Yglesias takes issue with this narrative, arguing that progressives overstate the salience of trade policy to voters’ shifting allegiances as a means of evading inconvenient truths — namely, that working-class voters largely disagree with the left on immigration, crime, climate policy, and various other social issues. In this account, there are inescapable trade-offs between adhering to progressive orthodoxy and building a Democratic majority. The anti-globalization narrative invites Democrats to ignore these trade-offs by telling them, in so many words: You don’t need to be less deferential to progressives on social policy so long as you become more deferential to their views on trade and economics.

I think there is some merit to both these perspectives. But I also think that the discourse around this subject elides some important facts and draws faulty conclusions from the genuine harms and benefits of globalization.

On the one hand, the negative economic and political consequences of free trade are often overstated. The median US worker has higher real earnings today than when NAFTA was signed, which is partly a result of trade liberalization. For this reason, among others, it’s unlikely that Bill Clinton’s trade policies are the primary driver of working-class defections from the Democratic Party.

On the other hand, deindustrialization has devastated many working-class communities in the US, while undermining organized labor. And there’s reason to think that Democrats have paid a political price for their role in accelerating the decline of American manufacturing employment. Specifically, as the party’s claim to represent labor’s interests lost credibility in trade-vulnerable communities, working-class social conservatives became more inclined to vote Republican.

In light of deindustrialization’s most pernicious effects, Democrats should make American workers’ access to remunerative employment and collective bargaining rights less contingent on the market’s whims, while rebuilding the party’s reputation for sound economic management.

My thinking on all this can be broken down into four points.

1) Globalization did not immiserate America’s working class

There’s a version of the “free trade cost Democrats the working class” thesis that paints contemporary economic life as a dystopian nightmare. In this account, globalization and neoliberal domestic policies have immiserated American workers. Faced with economic decline, poverty wages, and acute material insecurity, many working-class voters drifted toward right-wing populism out of desperation.

This story has kernels of truth. Opening up the US economy to exports for lower-wage nations like China and Mexico did reduce manufacturing employment in the United States, and this process of deindustrialization was further exacerbated by automation. Some working-class communities suffered gravely as a result. And for significant stretches of the past 40 years, wages really did stagnate or fall.

Nevertheless, as an account of how American workers are doing today, the dystopian narrative is badly wrong. In reality, by a wide variety of metrics, typical US workers are now much better off materially than they were when NAFTA was signed or Ronald Reagan was first elected (or when most any other milestone on the road to neoliberal globalization was passed).

As the economic analysts Adam Ozimek, Benjamin Glasner, and John Lettieri noted in a report earlier this year, the median US worker earns higher real wages and more benefits than ever before. These higher earnings have allowed Americans’ workweeks to shrink, even as their incomes have risen. A higher percentage of workers enjoy paid sick and family leave, and hold a retirement account. And fewer than ever are getting injured on the job.

Ozimek and co-authors back up their thesis with a wide array of Labor Department statistics, and their basic story is affirmed by additional data sources. For example, the median worker’s real personal income — in other words, annual income adjusted for inflation — was 38 percent higher in 2023 than it had been when NAFTA took effect in 1994 and 60 percent higher than it was in 1980.

That Americans have grown more prosperous over the past three decades shouldn’t be entirely surprising, even to those immersed in studies of free trade’s negative consequences. Even the most critical scholarly assessments of NAFTA and normalized trade relations with China have found that those trade policies increased real wages for a large majority of Americans. Which makes sense: In the early 1990s, only about 15 percent of US workers labored in manufacturing. The vast majority of Americans were consumers of manufactured goods, not producers of them, and therefore stood to benefit from greater access to cheap foreign wares.

It’s important to be clear about what all this does and does not mean. On the first count, it does not follow from any of this data that Americans should be content with the prevailing economic order. US workers have grown more prosperous in recent decades. But their wages have risen at a much slower pace than they did in the 1950s and 1960s. Meanwhile, income and wealth inequality have swelled over the past three decades, and the cost of many vital goods and services — including housing and health care — has soared in recent years.

In any case, precisely because America has never been wealthier, working people should expect more from their economy. Wages have risen but they should be higher. Poverty has fallen but it should be lower.

It also remains true that some working-class communities have suffered wrenching economic decline over the past 30 years. In many areas, globalization and automation eliminated well-paid, unionized manufacturing jobs and replaced them with far less remunerative employment opportunities. This devastated not only manufacturing workers but myriad businesses that depended on the patronage of such workers. Those on the losing end of globalization are overrepresented in Rust Belt swing states. Their justified economic grievances are therefore politically relevant.

But Democrats have lost ground with working-class voters in a wide array of areas, including places where scarcely anyone works in manufacturing. In other words, they have lost the support of many working-class “winners” of free trade. Thus, while the material hardships imposed by deindustrialization — and voters’ reaction to such hardships — may explain part of the Democrats’ problem with the working class, it cannot explain all of it.

2) Democrats’ embrace of free trade undermined them with some working-class voters, especially in the industrial Midwest

As just noted, a politically powerful minority of American workers has suffered as a result of deindustrialization. This reality has plausibly cost the Democratic Party some working-class support.

Before Bill Clinton’s presidency, the Democrats were unambiguously the more protectionist party in American politics. After Clinton signed NAFTA and normalized trade relations with China, the Democrats’ position on trade policy became less distinct from the (pre-Trump) GOP’s.

Theoretically, this could have hurt the party’s standing with working-class voters who favored protectionism but agreed with the Republican Party on some social issues. Granted, the GOP in the 1990s and 2000s was also enthusiastic about free trade. But if the difference between the two parties on protectionism seems small, then it might make sense to start voting on the basis of other topics — such as abortion or gun control — where culturally conservative workers and Republicans see eye-to-eye.

A 2024 paper in the American Economic Review suggests that, following Clinton’s signing of NAFTA, this dynamic played out in practice. Examining survey data from the American National Election Studies, the researchers found that voters who favored putting “limits on foreign imports to protect American jobs” were more likely than other voters to shift toward the GOP between 1992 and 1996, even after controlling for other demographic variables.

The study also shows, however, that this shift among protectionist voters was concentrated overwhelmingly among respondents who opposed abortion and/or attended church weekly. In other words: The paper suggests that NAFTA specifically caused some socially conservative, protectionist Democrats to become Republicans, seemingly because the differences between the parties on trade policy narrowed.

More broadly, globalization undermined organized labor in the United States, as the manufacturing sector had a higher unionization rate than the economy. And the decline of trade unions has likely hurt the Democratic Party with working-class voters in multiple ways.

For one thing, it shifted power in the Democratic coalition away from labor organizations — which is to say, groups accountable to a mass, working-class constituency — and toward business and advocacy groups accountable only to their highly educated staffers and wealthy donors. This plausibly led the Democrats to embrace priorities and issue positions that were less congenial to working-class voters than to affluent, college-educated people.

For another, unions have historically shown some capacity to politically influence their members to the Democrats’ benefit (this capacity is sometimes exaggerated but surely exists in some contexts), as well as to help mobilize working-class voters who lean Democratic. Advocacy groups lack a similar capacity.

3) It seems unlikely that trade policy was the primary cause of working-class defections from the Democratic Party

This said, there are reasons to doubt the centrality of Clintonian trade policies to the Democrats’ difficulties with working-class voters.

First, there’s an issue of timing. In 1996 — three years after signing NAFTA — Bill Clinton won a landslide reelection with the most working-class coalition Democrats had assembled since 1968. By contrast, Joe Biden was arguably the most protectionist Democratic president in modern history, and he nevertheless presided over a decline in the party’s working-class support.

Granted, the full effects of NAFTA and normalized trade relations with China were not felt until the early 2000s. And yet, between 2000 and 2008, Democrats gained ground with working-class voters, even as America hemorrhaged manufacturing jobs.

Second, as already mentioned, Democrats lost support in recent decades from many working-class voters who were never employed in manufacturing and likely benefited from globalization.

Most progressives attribute Democrats’ declining working-class support to the party’s broad failures of economic governance, including its culpability in the deregulation of finance (and, thus, the 2008 crisis). And I think it’s quite likely that the Democratic Party would enjoy more working-class support today if it had managed to deliver better economic outcomes during the Clinton, Obama, and Biden presidencies.

Still, I think Yglesias’s theory of the Democrats’ declining fortunes with working-class voters is more credible than many on the left wish to admit. The basic gist of that story is that working-class voters have always been, in aggregate, more socially conservative than educated professionals. The Democratic Party long managed to win significant support from working-class social conservatives anyway, but over the past three decades, Democrats moved left on virtually every social issue, from LGBTQ rights to immigration to crime.

Over this same period, voters — including working-class ones — became more conscious of the ideological divisions between the parties, as they gained access to Fox News and the internet. Studies have found that the expansion of broadband internet into rural areas leads to lower levels of ticket-splitting (i.e., more ideologically consistent voting) and higher levels of partisan hostility, seemingly because voters start consuming more partisan political outlets.

Thus, as Democrats have grown more socially liberal and culture war divides have grown more salient, the party has had a harder time winning working-class social conservatives.

This theory makes sense of Clinton’s success with working-class voters in 1996: Although he was a free trader, Clinton ran an exceptionally conservative campaign that year, touting his support for giving the death penalty to “drug kingpins,” cracking down on illegal immigration, and imposing work requirements on welfare programs.

Yglesias’s account is also consistent with Democrats losing ground with working-class voters who do not work in manufacturing.

To be sure, the culture war does not exist independently from the US economy. There is a complex interplay between economic conditions, social attitudes, and voters’ priorities. Nevertheless, I suspect the fundamental problem for Democrats is that working-class voters are not that progressive, particularly on social issues. The question is whether the party could have mitigated this issue by embracing a different set of economic policies in recent decades.

4) Democrats should take the negative economic and political consequences of deindustrialization seriously

Democrats should draw both economic and political lessons from the worst consequences of free trade.

On the economic front, the party should lament the economic decline of many manufacturing towns. The fact that most Americans benefited materially from globalization does not make the economic devastation of Rust Belt communities acceptable.

But that doesn’t mean we should try to rebuild the manufacturing sector of yesteryear by placing high tariffs on foreign exports. Today, more than 90 percent of US workers are not employed in manufacturing. Imposing high taxes on foreign-made goods would therefore make the vast majority of Americans worse off, even if such a policy succeeded in raising wages and employment in our nation’s manufacturing sector (which is far from a given).

Seeking to increase working-class living standards through higher manufacturing employment is a fool’s errand in the long run anyway. Manufacturing as a share of overall employment has been falling for decades in every rich country, and has even ticked down in China, despite its myriad industrial policies. Automation, combined with growing consumer demand for services, tightly limits how much manufacturing labor America requires.

The lesson of deindustrialization is not that we should seek to maximize the number of Americans who spend their days on an assembly line. Rather, it’s that creative destruction in a capitalist economy can impose terrible hardships on unfortunate workers. This doesn’t mean we should try to lock today’s pattern of production in amber. But it does mean that we should make economic displacement less painful for workers by making the welfare state more generous, job retraining more accessible, and access to collective bargaining universal.

In an America where all workers enjoyed the benefits of union — irrespective of their industry — deindustrialization would have had less grievous consequences for laid-off manufacturing workers. The same can be said of an America with more generous unemployment benefits and universal health care.

Of course, making the United States into a social democracy is easier said than done. At present, Democrats will need to win multiple elections by a landslide just to eke out a bare Senate majority by decade’s end. In Gallup’s polling, only a small minority of Americans say they would be willing to support higher taxes in exchange for more government services. On the merits, though, we should respond to deindustrialization by insulating workers from the market’s arbitrary whims.

On the political front, meanwhile, Democrats must recognize that they are reliant on the support of working-class voters who reject many of their cultural values. The party’s capacity to win majorities is contingent on its reputation for representing ordinary Americans’ economic interests. This is partly why Biden-era inflation was so politically devastating.

It is likely that Democrats could gain working-class support by moderating on various social issues. But in many cases, such moderation entails sacrificing the interests of vulnerable social groups, and such substantive costs must be weighed carefully against the hypothetical political benefits. Regardless, it’s imperative for Democrats to develop an economic agenda that resonates with popular concerns, and then deliver strong real-wage growth once in power (something Biden failed to do).

The forces that eroded US manufacturing employment and increased the political salience of America’s cultural divides are not going away. Democrats must figure out how to navigate them better. That will be easier if all concerned can agree on the basic contours of the political and economic landscapes.