And will we be so lucky next time?
I began writing this post on the morning of January 6, when I expected some drama in Congress around the certification of the Biden-Harris victory, along with some possible street violence by pro-Trump protesters. I was not expecting this.
The seizure of the US Capitol by extremists marks the climax (I hope!) of President Trump’s sustained efforts to overturn the 2020 presidential election. While these efforts do not meet the narrow definition of a coup in political science, there is no doubt that they represent a concerted, multi-pronged attempt to illegitimately seize power after January 20, 2021. In terms the general public uses, ‘coup’ is appropriate shorthand for such an attempt, so I’ll use that language here. And for social scientists, it behooves us to ask why the attempt failed, so that we can better understand the underlying danger to democracy in America. In other words, given that President Trump’s coup failed, how likely is the next attempt to succeed?
That there was indeed such an attempt to remain in power regardless of the election’s outcome will not surprise many readers who find their way here. But to understand how the attempt failed, and whether we are likely to be so lucky in the future, it is necessary to lay out the dimensions of the attack on our democracy. President Trump’s efforts to overturn the election took place more-or-less simultaneously on four tracks.
- Legal efforts in state and federal courts to throw out Biden-Harris ballots
- Private and public pressure on state election authorities, urging them to falsify results
- Pressuring Congress, and Vice President Pence, to choose the winner of the election by extra-constitutional means
- A disinformation campaign aimed at encouraging Americans to believe the election was rigged, and to protest as such
Success on any of these fronts could have granted Trump the presidency. There is next to no chance that President Trump remains president on January 20. Still, as the world saw yesterday, the last of these was by far the most successful. What accounts for these failures? And how likely are they to be repeated? For each, I’ll walk through some of the social science on the subject, and then give my own highly authoritative ranking of the risk of democratic breakdown on a scale of 1-5 Lincolns (5 Lincolns = the republic is safe, 1 = buckle up).
Attacking the election in court
The Trump campaign brought over 50 lawsuits before state and federal courts, and lost nearly all of them. The pages and pages of affidavits alleging fraud waved about by Trump surrogates on cable TV may have looked sinister, but they quickly fell apart in one of rare places where evidentiary standards still matter. In a Pennsylvania case in which the plaintiffs alleged they were not allowed to observe the county of ballots, for example, the federal judge reminded of the Trump campaign lawyer that lying to the court runs the risk of bar sanctions; the lawyer subsequently admitted that Republican observers had been permitted as the law allows. Many of these unfavorable decisions have come from Republican-appointed, and even Trump-appointed judges. While this is certainly disappointing for Trump, there are clear lessons from the comparative courts literature on why these challenges failed across the board.
US federal courts, at least, have among the highest levels of de jure independence in the world. They hold office for life, after all, and enjoy comfortable salaries. This insulates them, in theory, from outside pressures, allowing judges to rule according to their own view of the merits of the case More importantly, however, American courts also exhibit high levels of de facto independence–only two countries scored higher than the US on Linzer and Staton’s (2015) measure of de facto independence. This high degree of formal and behavioral independence makes it difficult for the President to pressure judges to rule against the evidence before them. There are two important sources for this high degree of independence.
The first of these is the underlying competitiveness of US politics at the national level. Political competition has been shown to increase courts’ behavioral independence in democracies (Epperly 2018, Aydin 2013, Randazzo et al 2016). The predominant explanation for this pattern is insurance theory, the idea that when governments cannot be sure whether the next election will put them in the minority, they insulate courts from political pressure. That independence, a sometime hindrance for governments, becomes a boon for same party when it enters the minority.
The second factor is the relative absence of patronal politics in the United States. The American president does not preside over a mass network of sub-patrons and clients with their own resources that can be mobilized in defense of the boss as needed. In non-democracies, political competition may worsen judicial independence (Popova 2012), as incumbents feeling electoral pressure lean on courts to rule unfavorably against their opponents. Even judges who enjoy formal independence can find their de facto independence undermined in a variety of unsavory ways: threats, blackmail, demotion within the court system, and more. But the US president cannot lean on senior judges to threaten to demote judges hearing his cases; he has no means to pressure a mayor to threaten a judge with criminal investigations or a surprise tax bill; he has no ability to use the security services to threaten a judge with the revelation of personal secrets. The American political system does not run on presidential patronage; as a result, the president’s ability to informally pressure courts is limited largely to the bully pulpit.
Summary: Courts did not back Trump’s claims because they were specious, because the courts are highly insulated from formal pressure, and because the presidency lacks the tools to successfully engage in informal pressure.
Risk rating: 4 out of 5 Lincolns. US courts are primarily protected from these kinds of attacks by the lack of patronal politics in the US; this is not likely to change. However, the US political system is not perfectly competitive. The Republican Party enjoys structural advantages in the Senate, and in the drawing of house districts after the 2010 (and likely 2020) Census. At least for now, it also enjoys an advantage in the Electoral College. All this means that the Democratic Party has to significantly outperform the Republicans nationwide in order to win power. If these trends persist, some strategic decision-making by judges in favor of the dominant party is possible.
Postscript: It is important to note here that we are discussing here the courts’ tossing of Trump’s post-election efforts to overturn the results. Courts were much more friendly to pre-election, legal attempts at biasing the outcome. For example, the US Supreme Court’s decided in 2019 that partisan gerrymandering is A-Okay, while the Texas Supreme Court decided this fall to uphold the governor’s order that each county have only one ballot drop-off box–that’s one box each for Harris County’s 4.7 million residents and Loving County’s 190 residents. Fair is fair! In general, this is the sort of election manipulation that parties and courts are much more comfortable with in the United States; no one undertakes any personal risk when carrying them out. Which leads us to…
Undermining state election procedures
At the same time as he was engaged in flurry of lawsuits, the president was also personally calling state and local election officials in battleground states, attempting to cajole them into changing the results. This was most dramatically documented in the recorded phone call with Georgia’s Secretary of State, in which the President urged that Republican elected official to “find” just enough votes for Trump to be declared the state’s winner. But the President had similar calls and meetings with at least 31 other officials as well as Congressional allies. To be very clear: the President was asking these individuals to engage in election fraud, either by falsifying the results or by discarding Democratic ballots. Thankfully, this effort was unsuccessful.
How do you get someone to subvert an election for you? In principle, it’s easy: offer them sufficient rewards to overcome the possible risk of exposure and punishment. However, a key wrinkle here is that the expected value of rewards and punishments depends in part on whether the attempted fraud is successful (Rundlett and Svolik 2014). To put it bluntly, if you forge the results to keep Trump in office, he can reward you with a post in the administration, advance your career in the Republican Party, or land you a high-profile media gig. He can also shield you from federal law enforcement, and lean on state officials to protect you there. But if the effort is unsuccessful, all those goodies vanish, and you may be facing years in prison for violating election law. So, your calculation about the costs and benefits of committing fraud depend in part on what you think other election officials will do.
In this case, Trump’s electoral defeat did not depend on one state, to say nothing of a handful of counties. This means that no single election official’s decision to commit fraud would be decisive; falsifying the results in Trump’s favor was clearly a collective action problem. Moreover, the US has a robust system of electoral observation and safeguards at all stages of the process, and the two major parties are armed to the teeth with lawyers for a general election. Illegal falsification would almost certainly be detected, and its perpetrators face both state and federal charges. This is a very high bar to clear for a president seeking to suborn election fraud from low-level officials.
Summary: Despite a concerted effort, no state or local officials falsified election results in Trump’s behalf. This is because of the collective action problem involved, and the associated inability of Trump to protect his supposed agents from punishment.
Risk rating: 3 out of 5 Lincolns. The collective action problem here would be much reduced in a scenario where only one state was needed to flip the result. In this case, it is conceivable that the president’s party could line up behind an effort to falsely change the results, and protect state officials from the consequences.
Attempts in Congress to disenfranchise voters from several states
I am very far from being a scholar of the US Congress, so this segment will be brief, and I look forward to reading how others interpret the events of the past few days in terms of executive-legislative relations. Cynically adopting the language of election integrity, Trump supporters in Congress objected to the Electoral College votes of some states won by Biden. Procedurally, this effort was doomed to fail, since objections require the assent of both houses of Congress.
Summary: This effort failed because it did not generate majority support in the Senate, and because the House was held by the opposition party.
Risk rating: 3 out of 5 Lincolns. We learned through this episode that there is a hearty appetite among the Republican Party for disregarding the will of Democratic voters–a majority of House Republicans voted in favor of the objections. Due to Republicans’ structural advantages in the Senate and the House going forward, it is at least conceivable that a challenge to a Democratic candidates electoral votes could survive, even if weakly grounded in facts, especially if it were done in coordination with the legislature from that state.
The battle for hearts and minds
The last prong of President Trump’s campaign to undermine the election has been far and away the most successful, though it will not save him from vacating the Oval Office in two weeks. Yet this the attack that is likely to be the most damaging and the most likely to accelerate democratic backsliding in the United States.
A rapid YouGov survey found that attitudes toward the storming of the Capitol break down clearly, and predictably, along partisan lines. Nearly 70 percent of Republicans do not see it as a threat to democracy–indeed, given the widespread belief in election rigging among that group, they likely see it as an effort to save democracy.
Indeed, a post-election survey by YouGov also found that fully 90 percent of Trump voters believed that mail-ballot fraud benefiting Joe Biden was definitely (66%) or probably (24%) occurring (survey data here). And that was before months of disinformation from the White House, Congressional allies, and the President himself. Here again we have lots of social science to help us understand what is happening, and what may happen in the future. Perhaps the most salient fact is that election fraud does not have to be real to result in potentially destabilizing protest. This is the central finding of a key paper by Daxecker, Salvatore, and Ruggeri (2019), who use survey data to show that subjective belief in fraud helps predict protest participation, while proximity to actual observed fraud does not. Indeed, the fear of protests founded on false allegations of fraud may be one of the reasons that non-democratic governments establish independent election commissions (Magaloni 2010).
Moreover, the subjective perception of fraud hinges strongly on partisanship and on the party’s electoral fortunes. Using a survey experiment, Beaulieu (2014) shows that both Republicans and Democrats were much more likely to perceive election manipulation in suspicious circumstances when an out-partisan was the winning candidate (Figure 2 from her paper reproduced below).
Similarly, in a panel survey before and after the 2012 Mexican presidential election, Cantu and Garcia-Ponce (2015) found that respondents across all major parties generally had the same distribution of expectations for electoral fairness prior to the election. After the election, however, supporters of losing candidates become much more likely to say the election was “very clean,” and much more likely to consider it “not clean.” This is mirrored in my own survey of Americans prior to the 2020 election, which found no appreciable difference in how Republicans and Democrats assessed election integrity in the US.
There is more to say on the literature here, but I’ll close with three final points. First, allegations of fraud make people angry. Results from the same 2020 survey experiment show that when fraud is posited as the cause of a candidate’s victory, people get mad. Republicans’ average anger nearly doubled from the condition where Biden wins cleanly, to one where he wins fraudulently (as is the case in the Trumpian epistemic bubble). Anger, in turn makes people more likely to participate in costly political behavior (Groenendyk 2011), such as dangerous protest or violent action.
Second, supporters of losing parties are considerably less satisfied with democracy than are supporters of winners (Howell and Justwan 2013). Third, Daxecker et al (2019) show that those who subjectively perceive fraud were more likely to protest if they were also closely embedded in communities–this may have analogous implications for the sense of community and networked connections found among many of the most die-hard Trump supporters.
Summary: This is the doozy. President Trump has been playing with fire, exploiting well-known (to scholars of election manipulation) tendencies for people to believe in and act upon unfounded allegations of fraud, if they fit with their pre-existing identities and beliefs. This likely represents an ongoing and durable threat to the survival of US democracy, as large numbers of Republican supporters believe the election was stolen; efforts to “restore” the rule of the democracy will likewise be met with fear and anger by supporters of Democrats. Already, 6 US senators and a majority of Republican representatives in the House have voted to overturn the will of the voters. I fear this will only get worse.
Risk rating: 1 out of 5 Lincolns. This is a poisoned well which I’m afraid we’ll be drinking from for years.
Coming (someday) soon
Part 2 of this series will consider how we as a society might be able to undo some of this damage.