In new research, Giovanna Invernizzi and Federico Trombetta build a model that shows how factional disagreement from a national leader’s own party can reveal to voters the leader’s strength and deter them from enacting harmful reforms.
In January 2019, British Prime Minister Theresa May brought her Brexit deal to the House of Commons and suffered the largest parliamentary defeat of any British government in a century, inflicted in large part by members of her own Conservative Party. The rebellion did not just kill the deal. It broadcast to every voter in the country that May could not control her own party, and her approval ratings collapsed in the months that followed. Within a year she was gone.
Contrast this with Matteo Renzi, who, as Italy’s prime minister, launched a sweeping labor market reform over the open opposition of his Democratic Party’s left wing. The reform was approved and his party won nearly 41% of the vote at the subsequent European elections, its best result ever. Or consider United States President Donald Trump’s 2018 steel and aluminum tariffs, imposed despite the written protests from more than 100 Republican lawmakers. The tariffs delivered little measurable economic benefit to the regions they were meant to help, but they paid off politically.
These episodes share a consistent pattern. Voters reward leaders who get things done, and they punish leaders who appear too weak to govern through divisions in their own parties. This observation is usually taken as an argument for stronger leaders and more unified parties. A prevalent view among political scientists treats internal party divisions as a pathology: factions obstruct legislation, raise the costs of decision-making, and divert resources to internal power struggles. Voters seem to agree: parties perceived as divided are systematically penalized at the polls. This stigma goes back to U.S. Founding Father James Madison, who warned about the “mischiefs of faction.”
In a recent paper, we develop a model of policymaking inside a divided party that suggests reality is more complex, and empowering leaders may not always be the best solution. The very fact that voters reward leaders who get things done gives strong leaders a reason to do things that should not be done, while internal opposition, precisely because it is an obstacle, is what keeps that temptation in check.
When proving strength means over-reforming
Our model features a party leader who must decide whether to pursue a reform and an internal faction that can resist it. The leader knows if the reform is genuinely needed and wants the same policy outcomes voters want. The faction disagrees with the leader about the policy, and its dissent can obstruct the reform’s implementation. The effectiveness of the faction’s dissent depends on the leader’s strength: strong leaders push their agenda through internal resistance largely intact; weak leaders see theirs watered down.
Voters observe the policy outcome, but they do not know if the reform is actually needed or how strong the leader is. That second uncertainty is what generates the central force in the model. A reform that emerges (almost) fully implemented reveals a leader strong enough to govern through division, exactly the kind of leader voters want in charge. As a consequence, rational voters reward successful reformers at the ballot box.
This, however, has important consequences. Because pushing a reform through is what distinguishes strong leaders from weak ones, strong leaders may choose to implement reforms even when they know that no reform is needed. Maintaining the status quo would be the right policy, but it would make them look indistinguishable from a leader too weak to act and cost them re-election. Weak leaders, who cannot fake strength, as any reform they attempt comes out visibly diluted, simply do the right thing. The result is what we call over-reform: a world in which the most capable leaders are the ones behaving “badly,” and voters reward them for it because they value electing strong leaders. No voter irrationality is required; the distortion survives even when voters see the consequences of every policy clearly.
What this implies for voter welfare
What does this mean for voters’ welfare? A natural first guess is that stronger leadership is simply good: strong leaders implement needed reforms more effectively. Our analysis shows the answer instead runs through two channels that policy debates tend to blur together.
The first channel is the voter’s prior belief that the incumbent leader is strong. A leader more likely to be strong means needed reforms are implemented more effectively, and it is optimal to keep the leader in power in the future. However, it also means unnecessary reforms become more likely. Whether voters gain or lose from a stronger incumbent depends on which of these effects dominates.
The second channel is the strength of the challengers that may win the elections (and therefore choose policy) if the current incumbent is voted out. This second effect is unambiguous: a better pool of potential replacements always benefits voters.
The two channels are substitutes. The better the bench of replacements, the less voters gain from discovering their incumbent is strong, and so the less a strong incumbent is worth. This substitutability drives our second result. When both governing and opposition parties recruit from the same pool of political talent, and therefore anything improving the overall quality of the political class strengthens incumbents and challengers together, voter welfare is highest at an intermediate level of that quality. Up to a point, a stronger political class helps: reforms get implemented, and replacements are reliable. Beyond that point, the cost of over-reforms dominates the benefit from having a pool of stronger candidates.
Why infighting may help voters
Intraparty factions in our model are not passive players. They choose whether to fight the leader openly or keep quiet. And a faction’s dissent may be costly to the party: it exposes the leader’s weakness to voters. So why would a faction ever go public rather than preserve a facade of party unity?
In the model, a faction trades off advancing its policy preferences with the cost of retaliation from the leadership and the overall effect of intra-party argument on the electoral prospects of the party. After all, factions care about their party winning.
From the point of view of the voter, factional silence does not need to be good, even when there is policy alignment between voters and party leaders. In fact, while factional silence improves the likelihood and integrity of a necessary reform, it also reduces how much voters learn about the strength of their leader. Moreover, silence worsens the outcome of unnecessary reforms, as leaders can take their measures to the extreme.
When the incumbent party is trailing, voters could be strictly better off with a faction that fights than with one that keeps quiet, even though the faction’s policy goals conflict with voters’ own.
What this means for party reform
Our results on the role of factions has concrete implications for recurring debates about party organization. Reforms that concentrate power in leaders’ hands, such as centralized candidate selection, tighter discipline over dissenters, and party resources allocated at the leadership’s discretion, are routinely defended as ways to help parties govern. Our analysis says these reforms carry a hidden cost: they dismantle the internal resistance that keeps leaders from governing for show, and they silence the dissent that tells voters who their leaders really are—and if their policies are really for the best.
The broader lesson extends beyond parties. Voters’ admiration for leaders who get things done is precisely what tempts capable leaders into doing unnecessary things. Institutions that make governing harder can protect voters from that temptation. A political system should think twice before trying to maximize leadership strength.
Authors’ Disclosures: The authors report no conflicts of interest. You can read our disclosure policy here.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
Subscribe here for ProMarket’s weekly newsletter, Special Interest, to stay up to date on ProMarket’s coverage of the political economy and other content from the Stigler Center.
