Recent work in political science, drawing on the comparative politics literature on features held in common by countries that have exhibited democratic breakdowns, and using tools such as qualitative classification analysis (QCA), have offered opposing points of view on whether the United States exhibits the features historically found to be associated with democratic breakdowns. I have great sympathy for the notion of seeing the United States as but one case among many (see, especially, Taylor et al. Reference Taylor, Shugart, Lijphart and Grofman2014, or Grofman Reference Grofman, Grofman, Lee, Winckler and Woodall1999). Nor do I dispute the relevance of the two potentially causal factors highlighted by Weyland (Reference Weyland2020)—an economic or other crisis and weakness in institutional supports for democratic decision making—although I also agree with a point made by Lopez and Luna (Reference López and Luna2021) that increased inequality and increased polarization should be taken into account.Footnote ¹ I do not attempt here to resolve the differences between these two essays about either the general relevance of QCA analyses for the study of causality or the degree to which QCA findings support the claim that the United States displays many of the indicators whose combination could predict a high likelihood of democratic breakdown. Instead, I focus on institutional differences between the United States and other major democracies that suggest care in seeking to generalize from historical cross-national comparisons.

The United States differs from most other democracies in that it is a federal polity, and this is a factor neither Weyland (Reference Weyland2020) nor Lopez and Luna (Reference López and Luna2021) incorporate into their models. If we think about the time path of US democracy and take as our slogan “Remember the Civil War” and recall, too, that the southern name for that war is the “War between the States,” it is easy to see why the federal nature of the United States should be taken into account in any attempt to prognosticate democratic breakdown. Although I focus in this article on the implications of federalism—including factual aspects of party control at the legislative level and legal aspects of deference by federal courts to state-level authorities—I also discuss important changes in the structure of electoral competition in the United States, especially in terms of the increasingly minoritarian nature of major US political institutions, on the one hand, and the nature of mass beliefs about the integrity of the electoral process, on the other, that can substantially affect the prospects for democratic breakdown.

Before the January 6, 2021, certification vote in Congress of the 2020 presidential election and the continuing challenges to the legitimacy of reported presidential vote tallies in Arizona, Pennsylvania, and elsewhere, the possibility of democratic breakdown seemed remote. Yet the combination of the following factors makes nondemocratic outcomes easier to implement, more likely to happen, or both: (1) incredibly high levels of partisan polarization; (2) recent legislation passed by states under Republican trifecta control that makes partisan misreporting of ballot outcomes easier; (3) the high level of expected pro-Republican gerrymandering of state legislative and of congressional districts in the 2020 redistricting round that will make it much harder for Democrats to win majority control of many state legislatures or of Congress and that will insulate legislators from reprisals at the ballot box; (4) the high aggregate level of closeness in control of the House, the Senate, and the presidency, which makes each election result more consequential; and, perhaps most importantly, (5) the belief in the big lie that Biden is not the legitimate president, which is held not just by tens of millions of ordinary voters but also by a high proportion of Republican elected officials.

Although national-level partisan polarization and polarization along racial lines are obviously important,Footnote ² also of relevance is the degree to which polarization creates geographic divides, especially in terms of partisan control of the states, which has consequences for fair and free elections. Many election law specialists (see especially Hasen Reference Hasen2020) have warned about the potential for breakdown of the electoral process at the state level, not just at the federal level, and about state-level attempts to minimize the vote from disfavored groups.

The two most likely scenarios of democratic breakdown in the United States would be (1) when a Republican-controlled legislature in a state with a Republican governor decides that a close congressional or presidential election in the state was “illegitimate” because of allegations of vote fraud committed by Democrats (almost certainly false but also almost impossible to disprove)Footnote ³ or (2) when a Republican-controlled US House or US Senate likewise challenges the certifications delivered by a state, as was unsuccessfully attempted on January 6, 2021. Such a legislative postelection reversal of the popular will, especially if it affected either the presidential outcome or control of Congress, could trigger a constitutional crisis of a form last seen in 1876. The resolution of the 1876 crisis was a compromise that ushered in nearly 100 years of Jim Crow suppression of minority rights. Footnote ⁴ A similar resolution of a modern equivalent would not be an event to be cheered.

But we may ask, Aren’t such scenarios highly implausible? I think not. First let us consider whether they could happen and then turn to how likely they are to happen.

Weyland (Reference Weyland2020, 389) argues that our still intact system of checks and balances insulates the United States from breakdown because it prevents antidemocratic forces from capturing all the keys to the kingdom. In this context, it is useful, however, to be reminded just how successful President Trump was in eroding checks and balances.Footnote ⁵ And let me emphasize two points: because of the federal nature of the US political system, antidemocratic forces can easily capture many states—and states have real legal power. Thus, even without a national victory for an autocrat, democracy could still be severely eroded. Reading the Federalist Papers, it is natural to think of the existence of local centers of power as a check on tyranny from the center, and the assertion that the states are “laboratories for democracy” reinforces a rosy picture.Footnote ⁶ But even the most cursory study of the history of conflict over state’s rights will reveal that federalism is not always a friend of democracy. And there are features of US federalism that can generate particular problems today.

First, unlike most countries, election administration in the United States is decentralized to the state level and often below that, such that counties count and certify votes. Second, malapportionment in the Senate is, by almost all measures, high relative to its historical average (Cervas and Grofman Reference Cervas and Grofman2020); partisan gerrymandering in the House is at its highest level in the post–one person, one vote era (McGann et al. Reference McGann, Smith, Latner and Keena2016; Reference McGann, Smith, Latner and Keena2021; Wang Reference Wang2013); and the distribution of electoral strength has made partisan bias in the Electoral College (EC) in the elections of 2016 and 2020 among the highest in recent history (Cervas and Grofman forthcoming). Moreover, rather than canceling each other out, each distortion is in a pro-Republican direction. For example, only a roughly 52% or higher popular vote for the Democratic candidate creates near-certainty that the EC winner and the popular vote winner will be the sameFootnote ⁷; similarly, Democrats currently need more than 50% of the national two-party vote to capture either the House or the Senate. Such distortions can allow a minority of a state (or of the nation) to determine which party will be in power at the state or national level.

Third, and relatedly, when the Supreme Court in Rucho v. Common Cause 588 U.S.__ (2019) rejected the view that federal courts had any role to play in placing a constitutional check on partisan gerrymandering, it in effect wrote a blank check to state legislatures by allowing them to engage in even more egregious gerrymandering in the 2020 redistricting round than they had in the 2010 round—which was already, as noted earlier, the most egregious in the post–Reynolds v. Sims history. Unlike in all other major democracies (Handley and Grofman Reference Handley and Grofman2008), electoral boundary drawing in most states is a partisan process, with a near-certainty of partisan gerrymandering when there is trifecta control of state government.

Furthermore, the relative balance of Republican trifecta and Democratic trifecta states is unfavorable for Democrats going into the 2020 redistricting round. There are presently 15 Democratic and 23 Republican trifecta states, but that statistic is misleading because a number of these states are either single-seat states (where no redistricting is necessary) or they are states where a commission, rather than the legislature, is doing the redistricting. In terms of congressional redistricting, eliminating those states with commissions and excluding single-seat states, there are 7 states where Democrats are in full control and 19 where Republicans are in full control. But an even more relevant statistic for the potential for pro-Republican gerrymandering is that, excluding single-seat states, in 2020 Democrats fully controlled 72 seats in the House, whereas Republicans fully controlled 184.Footnote ⁸ Moreover, the congressional seats that are being shifted after the 2020 census-based apportionment are disproportionately in states whose legislatures are under Republican control.

Fourth, in some states with Republican trifecta control, such as Georgia, legislation has already been passed that would enable the state legislature to take over control of elections from local election boards and from the secretary of state; similar legislation is under consideration in other states. Such legislation is clearly intended by Trump supporters as a “muzzle the future Brad Raffenspergers” scheme. Although most of the public attention has been on the spate of legislation in many Republican-controlled legislatures that makes it harder to vote—at least as compared to very recent elections (see, e.g., Wines Reference Wines2021)—the changes in ultimate administrative reporting authority are potentially far more consequential.Footnote ⁹

Fifth, state legislatures already have the power to replace direct election by the voters of a slate of presidential electors committed to a given candidate with some other form of selection, although this change needs to be made sufficiently far in advance of the actual election (Anderson Reference Anderson2020a; Reference Anderson2020b). Under the US Constitution, how to select presidential electors is a matter for state discretion.Footnote ¹⁰ In particular, a state could decide to have the state legislature select presidential electors, as was the case in some states in the early period of US history.Footnote ¹¹ Gerrymandering of the state legislatures thus makes it quite possible that, in states that opted for this option, the result would be a partisan outcome opposite from that of the statewide popular vote.

Sixth, and more generally, Supreme Court precedents suggest strong deference to state election administration in the absence of clear constitutional violations.Footnote ¹² The strong pro-Republican bias at the state level, especially with respect to gerrymandered state legislatures, thus has important implications for outcomes in which the majority of voters are powerless to achieve their will. For example, if there is a challenge in federal courts to the decisions made by state legislatures (or governors) about which slate of electors to report in the case of dueling slates, or when legal challenges are unresolved in state courts or recounts are still underway after the “safe harbor” reporting deadline has passed, the legal issues can get murky and depend a lot on the exact details of state law, as well as on the stance of the legislature and the governor (Anderson, Reference Anderson2020a; Reference Anderson2020b).Footnote ¹³ Moreover, Congress has the power to decide its own membership and has chosen between competing tallies reported from the same state, and it is hard to see the Supreme Court reversing congressional choices in the case of a congressional election, no matter how ill-founded these choices might be (see the discussion in Congressional Research Service 2021).

Of course, the fact that a gerrymandered state legislature or Congress could use a blatantly false excuse of ballot fraud (in elections won by a narrow margin) to substitute its will for the preferences of the voters, or that a state legislature could even do so directly in advance of any presidential election, does not mean that it would. And if you had asked me about the likelihood of such behaviors even as recently as a year ago, I would have pooh-poohed the idea. But there are four reasons why such behavior is today far more plausible.

First, incredibly high levels of partisan polarization, especially in the form of affective polarization, have caused some recent authors to view contemporary partisan competition as a new kind of “holy war” (Wuffle Reference Wuffle2021). Arguably, in the aftermath of the 2020 election this polarization has only grown worse. To paraphrase Barry Goldwater, in a war against the forces of Satan, it is much easier to think that “extremism in the defense of one’s way of life is no vice.” In other words, the ends justify the means. Think of the 2021 decision by the Republican-controlled Texas legislature to foster private citizens acting as bounty hunters to be paid $10,000 to chase down anyone involved in aiding and abetting abortions, under the theory that private action is not state action and thus cannot easily be prevented by federal courts’ enforcement of Roe v. Wade and the constitutionally protected rights of women. This is a clear instance of the unbelievable lengths to which state legislators are prepared to go to avoid federal oversight of their actions.

In the election context, the view that the ends justify the means is enhanced by righteous anger. Whether in the Capitol or the state house, a high proportion of Republican elected officials endorse the big lie that Biden is not the legitimate president.Footnote ¹⁴ If you believe passionately that the 2020 election was stolen, then it may seem “only fair” to steal back elections from the other side in 2024 (or 2022).

Second, and most obviously, we have the events of January 6, 2021. Virtually all the news coverage of January 6, especially coverage after that day, has focused on the storming of the Capitol and its implications for democratic breakdown. But, like the misdirection of a stage magician, worrying about the crowd initially outside the Capitol distracts attention from what was really important about January 6: the behavior of the Republican House members and senators inside the chamber. Having a hard core of white supremacists prepared to engage in violence and antigovernment behavior should come as no great shock, especially because they were emboldened by the president’s speeches and tweets about a stolen election and the need for action; these pronouncements gave them a pretext, aided by the absence of sufficient police presence to be a real deterrent and being partially sheltered by having a crowd around them of people who were there because they truly thought that the election was stolen.Footnote ¹⁵

In my view, what was truly shocking about January 6 was not the behavior of the demonstratorsFootnote ¹⁶ or even that of President Trump, which was more or less Trump being Trump, but the votes on accepting the tallies reported from two states that Trump had won in 2016 but had gone for Biden in 2020. In the case of Arizona, 6 Republican Senators and 121 Republican House members voted against accepting the result; the corresponding numbers for Pennsylvania were 7 and 138. In other words, in the US House, as many as 69% of the Republican members accepted the big lie that the election was stolen from Biden and were prepared to vote to reverse the “unacceptable” outcome.Footnote ¹⁷ Votes were held on only two states because each had a challenge to results from both a senator and member of the House from that state, which is required to force a congressional vote. But the percentage of votes objecting to Arizona and Pennsylvania results would almost certainly have been similar had Georgia, Michigan, Nevada, and Wisconsin—the four other states to whose results at least one Republican member of Congress had lodged an objection—also come up for a vote before the Capitol was evacuated due to the insurrection. No matter how pernicious and even deadly was the storming of the Capitol building, its repercussions or those of similar acts can be dealt with through the channels of ordinary governance; that is, through arrests and convictions for unlawful actions. The same is not nearly as true for illegitimate actions by state officials when these actions can be given the color of law.

Third, pragmatic considerations of office-seeking/office retention are of limited relevance. Most Republican state legislators and most Republican members of Congress occupy districts that are highly insulated from electoral challenge by a Democrat.

Fourth, the high aggregate level of closeness in control of the House, Senate, and the presidency raises the stakes for each and every congressional election and outcomes at the state level for president and senators (Fiorina Reference Fiorina2017; Lee Reference Lee2016).Footnote ¹⁸ The higher the stakes, the higher the incentives to cheat, especially if you think that, because you have control of the key organs of government in your state or in Congress, you can get away with it.

In 2020, President Trump encouraged Republican governors and legislatures to override the popular vote outcome in close states that he lost. In 2020 they disregarded him. In 2020 President Trump urged (ordered?) the Republican vice president presiding over the Senate to disregard the results of reported outcomes in close states won by Joe Biden. He did not comply. What will happen in 2024 is far less certain, because moderates within the Republican Party are being purged in favor of Trump loyalists and a clear majority of Trump supporters still believe in the lie of a stolen election, despite all the evidence to the contrary.

Views about the resilience of US democracy need to be tempered by analysis of how what is happening at the state level can affect national-level politics. Yet the degree of belief in Trump’s big lie about his election loss in 2020 can only serve to exacerbate previous fault lines in American politics,Footnote ¹⁹ much the same way that economic shocks, such as extraordinary high levels of unemploymentFootnote ²⁰ or runaway inflation, or defeat in war in a way that staggers national confidence have been argued to help trigger democratic breakdowns in Europe or elsewhere.

The reader may object to a seeming lack of even-handedness in this focus on the likelihood of a democratic breakdown coming from the Republican side of the aisle, because certainly Democrats in the past have also been accused of stealing electionsFootnote ²¹ and there has been violence on the part of some Black Lives Matter supporters. But it is Republicans who have embraced the big lie of a stolen presidential election, which is one of the driving forces that could trigger a democratic breakdown today.

Yet, a democratic breakdown is not inevitable. Consider one possible scenario that leads to a democratic breakdown:

1. 1. If a state legislature takes over from a county election administration (as the legislature gave itself the direct power to do in Georgia this June), and

2. 2. if the legislature or governor brings in a totally unqualified and highly partisan firm to do a recount (as the Arizona legislature did), and

3. 3. if the recount reversed the original result in which a Democrat was elected, and

4. 4. if the legislature decided that this recount was to become the official result, and

5. 5. if the state’s highest court decided that this was a purely political question that needed to be left to the legislature and governor, and

6. 6. if the Supreme Court, on legal challenge, buried its head in the sand regarding political reality (as legal scholars such as Stephanopoulos [Reference Stephanopoulos2019] and Hasen [Reference Hasen2021] have argued it is now wont to do) and said that this was entirely a matter of state election administration and that the state decision did not raise federal constitutional questions that mandated a reversal of the state’s reported election outcome,

then, if such a sequence of events were to occur, it is hard to argue that this would not be a democratic breakdown. For such a scenario, this article has identified the pivot points that would allow this breakdown to happen.

In thinking about this scenario, however, it is important to remember the Anna Karenina principle. Footnote ²² That principle states that, in any endeavor that rests on a sequence of necessary conditions, the failure of any one of these conditions dooms it to failure. So, if any YES on this “if sequence ladder” is actually a NO, then somewhere down the ladder, a breakdown involving an illegitimate reversal of an election outcome will be averted.

On the other hand, almost any YES on this “if sequence ladder” is a point of democratic vulnerability, and some might even be regarded as democratic breakdowns in and of themselves, even if the winner of an election is eventually legally acknowledged as such. Moreover, cliffhangers are only fun if you are at the beach on a blanket in the sand reading a mystery thriller; they are not at all fun when the consequences of falling off the cliff are real, especially if you are worrying that you might be an ostrich burying your head in the sand about the true magnitude of the threat.