Thoughts on the Cordon Sanitaire
How could excluding the Radical Right from politics Actually Work?
A key question in all democracies is: how should mainstream parties react to the emergence of the radical right? A few years ago it would have been about populist parties, but only the radical-right ones survive to this day, thus I will keep the discussion about that.
The firewall/cordon sanitaire
The argument for the firewall (cordon sanitaire) rests on the idea that if other parties exclude radical right parties from the political game, it will prevent their emergence. Some add that it is intrinsically wrong to reach deals with parties that hold morally unacceptable positions. But let’s rest for now on the more pragmatic dimension: what should other parties do to confront the radical right?
Center-right parties face a dilemma. They share common interests with the radical right—such as avoiding left-wing governments and aligning on certain issues—and voters will reward them for fulfilling those interests. It would be natural for the center-right to accommodate these new parties on the basis of those shared goals.
However, the argument goes, doing so “normalizes” the radical right, improving its electoral outcomes and ultimately working against the center-right’s long-term interests. Why would this happen?
(Note: There is a well-established body of work in political science covering this, notably the excellent book by Vicente Valentim. I have not followed this literature closely for a few years, so I will keep the discussion mostly citation-free—apologies in advance).
One channel for normalization is strategic voting. Many center-right voters choose these parties as a second best; their preferred option is the radical right, but the majoritarian premium of electoral systems gives them an incentive to coordinate with moderate parties. This incentive holds if center-right parties credibly commit not to form coalitions with the radical right. If, however, center-radical coalitions are shown to be possible, these voters will simply move to the radical right, their first choice anyway.
A second channel is the change in voter perceptions. If the radical right enters the political game, the stigma they suffer breaks down. Voters begin to perceive them as just another party rather than outsiders. Governing experience makes them appear more competent, reducing the “risk premium” voters associate with them, and opening the door to protest votes over valence issues like corruption or poor economic management.
Another version of the normalization argument involves issue manipulation and agenda setting. The firewall supposedly prevents issues “owned” by the radical right—immigration, nativism, law and order—from dominating the agenda. When radical right parties enter the political game, they shift the agenda to make these issues salient. The assumption here is that moderate parties cannot compete on these issues because the radical right better represents voters’ preferences on them.
This logic extends to whether mainstream parties should adopt radical right issues to meet popular demands. The argument warns that when center-right parties incorporate nativist or economic nationalist policies, voters prefer the “original over the copy” and vote for the radical right anyway.
Closely related to all this is the reality that allowing radical right parties into institutions helps them accumulate power and resources—access to media, ministries, and funding—which fundamentally changes the balance of the political game.
Critique 1: Are voters that manipulable?
These arguments share a few core assumptions. Mainly, they rest on the plausible idea that voters’ choices are not exogenous, but manipulable, and thus inherently unstable in a collective choice sense (a dynamic formally captured by McKelvey’s chaos theorem).
Consider strategic voting. It assumes the center-right sits on an unstable electoral base, with many voters not supporting them out of genuine preference. It is hardly surprising that when a radical right party becomes viable, these voters defect to their first-best option.
Similarly, the argument about norms suggests a precarious coordination equilibrium where voters’ latent preferences for radical right issues are excluded from the system. The root cause is a deep discontent with center-right parties failing to serve large chunks of the electorate. As Alexander Kustov recently noted, these movements often succeed not by changing minds, but by activating these latent preferences. The root cause is a deep discontent with center-right parties failing to serve large chunks of the electorate. At best, the cordon sanitaire can tip the equilibrium temporarily, but in a democracy, large-scale voter preferences cannot be permanently walled off from representation.
Critique 2: How the cordon sanitaire backfires
Beyond these foundational problems, there are many practical reasons why the cordon sanitaire can backfire.
Firstly, enforcing the firewall often forces mainstream left and right parties into grand coalitions, blurring the ideological lines between them. This allows the radical right to completely monopolize the anti-establishment vote, as the mainstream options become indistinguishable from one another. This dynamic is especially problematic during difficult times—such as economic crises or periods of high inflation—where being in power generates inevitable wear and tear (desgaste), leaving the radical right as the only untarnished opposition.
Furthermore, while the conventional wisdom suggests that adopting radical right issues only fuels contagion (voters choosing the original over the copy), there are also notable cases of successful co-optation where mainstream parties have effectively drained radical right support by tightening their own policies.
Critique 3: What is the empirical evidence?
Finally, the evidence in favor of these firewall mechanisms is remarkably difficult to evaluate empirically. What we want to measure is how things play out, at the margin, when a moderate right party decides exogenously to maintain the cordon sanitaire versus dropping it. There is no clean experiment for this.
One major limitation is that mechanisms like agenda setting and normalization evolve at the national—or even supranational—level. This generates spillovers that make any policy evaluation extremely hard.
Perhaps the best evidence comes from close-election regression discontinuity designs, comparing what happens when the radical right becomes decisive by a tiny margin versus when it falls just short. Or we could look at cases where, for exogenous reasons, the radical right does not run in a particular district. In my view, the massive, society-wide impact on norms, expectations, and issue-setting is unlikely to be captured by these localized variations. Even if you look at the paper above, the causal effect is quantitatively rather small.
Similarly, survey experiments showing what happens when specific issues are made salient are suggestive, but they make it hard to deduce what actually happens when a radical right party effectively enters a national cabinet and normalizes its status.
Ultimately, even the observational record across countries is not very optimistic for the firewall. In places where the moderate right has strictly stuck to the cordon sanitaire—France and Germany immediately come to mind—the results have been far from a resounding success.
My dissatisfaction on this debate is that, while everyone has strong opinions about this particular topic, the empirical evidence is highly inconclusive, and, I would argue, will always be.
One final thing I should say. Many of these arguments are about the demand for issues. That is, they consider that voters have pre-established preferences over issues (immigration, etc) and vote for parties that represent these. Eventually, this will move the vote towards new parties. I will leave this to a different post, but I do not think the model that poses that issues are a driver of vote works very well, in general.