In new research, Marc Jacob, Barton E. Lee and Gabriele Gratton argue that legislative gridlock is not only a consequence of Congress’ polarization but also a cause of it. In sum, both polarization and gridlock fuel one another generating a vicious spiral toward political mire.

The United States Congress is breaking records for all the wrong reasons. By many accounts, the current Congress, which began in 2023, is the most polarized in recent decades: its members on each side express more extreme policy views and vote with members of the other party on legislation less often. Bipartisan legislation seems—on almost all issues—a thing of the past. In fact, this Congress seems particularly unable to pass any legislation at all: in its first year, it passed a grand total of 34 laws, putting it on track to be one of the most gridlocked Congresses in U.S. history.  

This pattern of gridlock and polarization is not new. Congressional polarization has been on the rise since the 1970s. At the same time, and likely as a consequence of polarization, legislative gridlock has been increasing. As a result, Congress has failed to legislate on an increasing share of salient policy issues, such as climate change mitigation and funding for Ukraine, raising concerns about the ability of American democracy to respond to some of the most urgent challenges of our time.

Together with concerns for democracy, the rise of Congress’ polarization also raises puzzling questions for political scientists. While members of Congress have continually polarized since the 1970s, a large body of scholarship argues that voters have not polarized to the same extent. Instead, many scholars agree that voters have maintained relatively moderate policy preferences. Yet, voters have evidently contributed to Congress’ polarization by repeatedly voting in more extreme legislators to Congress.

But if most voters are moderate, then why do they vote for politicians who express extreme policy views? 

In a recent paper, we develop a theoretical model and an experiment to provide new insights into this question. Our theory offers a new way of thinking about the relationship between polarization and gridlock. We argue (and provide evidence) that legislative gridlock itself can cause polarization by making moderate but partisan voters more willing to vote for extreme co-partisans (partisans from the same party). Our argument reverses—but does not substitute—the common view that polarization causes gridlock. Ultimately, our theory suggests that both polarization and gridlock are likely to fuel one another in a spiral that may be hampering American democracy.

Our theory may be best described with an example. Imagine a voter who is generally aligned with the goals of the Democratic Party. Suppose that, over a policy issue such as the federal minimum wage, this voter has liberal but moderate preferences. For example, this voter may prefer a small increase to the minimum wage, say from the current status-quo of $7.25 per hour to $10 per hour. However, this Democratic moderate voter may dislike very large increases and rather prefer to maintain the status-quo than to increase the federal minimum wage to $35 per hour (close to what some Democratic candidates have recently advocated for). 

Now suppose it is election time. The voter faces a choice between a Democratic candidate advocating for extreme increases in the minimum wage to $35 per hour and a moderate Republican candidate who intends to maintain the status-quo minimum wage.

Our moderate voter faces a dilemma. Although the voter is generally aligned with the Democratic Party, the Democratic candidate’s position on the minimum wage is too extreme for the voter—on that issue, the voter prefers the Republican candidate’s position. The fear of a $35 minimum wage being enacted may lead this voter to swing away from the Democratic candidate. If so, then moderate voters like this one work to limit polarization in Congress, by voting out candidates who propose too extreme policies.

Our theory and evidence suggest that this moderating effect may be somewhat muted in the modern Congress where gridlock is prevalent.

High levels of gridlock on the minimum wage means that, even if elected, the Democratic candidate’s $35-per-hour wage proposal is unlikely to be enacted. Indeed, since 2007, every proposal to increase the federal minimum wage has failed. Knowing this, our moderate voter’s tradeoff becomes less of a dilemma. Gridlock reduces the risk of the candidate’s extreme policy being enacted and makes the voter more willing to vote for them. In this way, gridlock causes moderate voters to discount extremism and, in turn, contribute to polarization in Congress, by voting in co-partisan candidates who propose relatively extreme policies.

Although the above example is highly stylized and overly simple, it illustrates the basic point that underscores our theory:  when voters perceive higher levels of gridlock, they may discount extremism and be more willing to vote for candidates whom they would otherwise view as too extreme. Importantly, this argument applies to Democratic (as in our example) and Republican voters alike.

Anecdotal evidence suggests that these considerations do, in fact, play an important role in reality. Consider, for example, the issue of abortion. Since the Supreme Court ruling of Roe v. Wade in 1973, and until 2022 when the Supreme Court overturned its precedent, this policy issue had largely been gridlocked: Congress was severely limited in its ability to legislate on the issue. But then, in 2022, legislating on abortion became suddenly more plausible—even if still unlikely. In this case, gridlock was induced not by Congress itself (or Congress’ polarization), but by an outside factor: the Supreme Court and its interpretation of the Constitution. Our theory suggests the overturning of Roe v. Wade will lead Republican candidates to moderate their stances on abortion because moderate Republican voters will now be more concerned about the possibility of these extreme positions actually being enacted. Consistent with our argument, Republican strategist Barrett Marson described exactly this consequence of the Court’s ruling in a recent Guardian article:

Over the years, it’s been OK to advocate for the strictest abortion regulations […] because abortion generally was protected by Roe v. Wade. Now it’s no longer theoretical. So now the most restrictive policies have real-life consequences. And suburban women are giving a candidate’s position on abortion greater weight as they consider who to vote for.

In our paper, we move beyond theory and anecdotes. We run a large-scale online experiment with almost 9,000 voting-age Americans to test our theory. We begin by eliciting respondents’ partisanship and preferences over a range of policy issues. We then randomly assign half of our respondents to an experimental treatment that reveals information to them about the high levels of legislative gridlock in the U.S.; the remaining half do not receive this information and, hence, act as the “control” group. We then ask respondents to what extent they believe policy change is possible for a range of policy issues. Finally, using a sequence of hypothetical elections that mimic the voter’s dilemma described earlier, we analyze respondents’ stated willingness to support an extreme co-partisan.

Our results provide compelling evidence for our main prediction—that gridlock can cause polarization—and for our theory more broadly. We show that “treated’’ respondents (those who were randomly assigned information about the high levels of gridlock) became systematically more skeptical about the possibility of policy change. Furthermore, among respondents with moderate policy preferences, those who were treated became more willing to support extreme co-partisan candidates—even when these candidates advocate for extreme policies that these respondents do not like. In short: as our theory predicts, higher perceptions of gridlock cause moderate voters to support extreme candidates more.

We also provide evidence that our results are not driven by other psychological or behavioral effects. For example, our experimental results are not explained by gridlock simply increasing respondents’ partisanship or causing respondents’ policy preferences to become more extreme. We also show that gridlock does not make moderate respondents prefer extreme co-partisans over moderate co-partisans. Overwhelmingly, our data suggests that moderate respondents want to elect moderates: extreme candidates always face an electoral penalty from moderate votes. However, gridlock reduces this electoral penalty, making it more likely that extreme candidates are elected.

Our paper raises several important questions about the relationship between polarization, gridlock, and American democracy. On the one hand, gridlock is an intentional feature of American democracy and its institutions, stemming from an intricate system of checks and balances (separation of powers, bicameralism, anti-majoritarian rules). From this view, our theory and experimental evidence suggest that a polarized Congress may paradoxically be a sign of voters trusting that these institutions are working well and as intended. Furthermore, if voters’ perceptions of gridlock are correct, then Congress’ polarization should not be expected to lead to severe policy consequences.

Yet, this view is likely too optimistic. Voters may systematically overestimate gridlock or fail to internalize (or predict) the broader societal and democratic costs of Congress’ polarization. For example, polarization within Congress may erode intangible assets of democracy or social capital, induce affective polarization among the public, or increase the risk of democratic backsliding. From this view, Congress’ polarization is a cause for concern and our results highlight a tradeoff for the design of democratic constitutions. Institutions that ensure more policy stability (and, hence, induce more gridlock) are likely to give rise to a more polarized Congress; weakening these institutions may reduce polarization but risks exposing policymaking to larger swings. With recent events documenting the growing democratic costs of polarization, this tradeoff is ever more relevant to the future of American democracy. 

Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.