Few questions in comparative politics feel as urgent as the one animating Susan Stokes’s (Reference Stokes2025) The Backsliders: why has democratic erosion become a global wave precisely in the first quarter of the twenty-first century? Where much recent scholarship explains backsliding through the acts and omissions of particular leaders and parties, Stokes poses a deliberately historical question (why now?) and answers it with an ambitious structural account. The book is a major achievement, synthesizing a formidable range of cases, methods, and evidence. Yet its central explanatory chain, which runs from rising inequality through polarizing and institution-denigrating rhetoric to the psychology of voters, leaves surprisingly little room for the organizational machinery through which incumbents actually build and sustain mass support. In Latin America, where the durability of would-be autocrats so often rests on dense distributive networks, that omission matters.
Stokes’s argument unfolds in three movements. The roots of erosion, she contends, lie in the income inequality bequeathed by late-twentieth-century deregulation and globalization. Borrowing Dani Rodrik’s contrast between an “identity cleavage” and an “income cleavage,” she traces two routes to a common destination: right-wing ethnonationalists in the wealthy democracies and left-wing populists in the Global South, with Latin America foremost among the latter. A train metaphor organizes the causal logic. Presidents and prime ministers may disembark before their systems erode, or remain aboard; those who stay tend to govern unequal, polarized, and institutionally distrustful societies, and either to lack democratic commitments or to fear prosecution once out of office.
Conceptually, Stokes follows Melis Laebens in defining erosion as the simultaneous decline of vertical accountability (voters, press, and civil society) and horizontal accountability (courts, legislatures, and electoral authorities), operationalized with Varieties of Democracy data that yield 23 episodes across 22 countries between 1998 and 2022. Chapter 1 distills the shared tactics into what scholars now call “the playbook.” The book’s most original move comes in its second part. Confronting publics that have not, on the whole, abandoned democracy as an ideal, backsliders deploy three strategies to retain support: weakening elections as instruments of accountability, exploiting partisan polarization, and, in Stokes’s signature concept, “trash-talking” democracy, the rhetorical denigration of institutions as corrupt and incompetent. A closing chapter argues that reducing inequality and defending democracy are a single task, and weighs the dilemmas of constitutional hardball.
Two features deserve particular praise. First, the distinction between polarization (attacks on partisan rivals) and trash talk (attacks on institutions) is analytically crisp and, more impressively, measurable. Stokes and her collaborators hand-code the speeches and tweets of López Obrador, Chávez, Trump, and Erdoğan, demonstrating that the categories are not merely conceptual but recur, in differing proportions, in actual political rhetoric; the finding that Chávez leaned on institutional trash talk whereas Trump favored polarization is genuinely illuminating. Second, the final chapter is unusually concrete for a book of this scope. It engages directly with Gamboa’s (Reference Gamboa2022) account of why institutional opposition strategies tend to outperform radical ones, and it refuses easy answers on whether pro-democratic forces should answer hardball with hardball.
The book’s structural and rhetorical emphasis, however, carries a cost that readers of Stokes’s own earlier work will feel keenly. The coauthor of Brokers, Voters, and Clientelism (Stokes et al. Reference Stokes, Dunning, Nazareno and Brusco2013) here constructs an account of mass support that runs almost entirely through inequality, rhetoric, and voter psychology, while the intermediary machinery she did so much to theorize, including brokers, party organizations, and distributive networks, nearly vanishes from the story of erosion. This is more than a matter of emphasis. As Roberts (Reference Roberts2006) argued, Latin American populism is not organizationally uniform: the intensity of conflict between populist leaders and economic, political, or military elites shapes whether a leader governs through dense popular organizations or through unmediated, plebiscitarian bonds. Chávez built organization; others rule by direct appeal. Trash talk and polarization operate at the level of persuasion, but they do not explain how incumbents convert distributive goods into durable electoral loyalty, the mechanism at the core of clientelism. For the Latin American cases in particular, an explanation of why voters remain “on the train” that brackets the distributive and organizational infrastructure of incumbency feels incomplete.
A related conceptual question is worth pressing. Roberts (Reference Roberts2022) distinguishes constitutive, spatial, and institutional dimensions of populist polarization, the last involving the contestation of democratic rules and institutions themselves. It is not obvious that trash talk is more than a sharpened, codable relabeling of this institutional dimension. If so, the contribution lies less in the concept than in its operationalization, a real advance but a narrower claim than novelty.
A second reservation concerns the singularity of the income-cleavage route. Riedl et al. (Reference Riedl, McCoy, Roberts and Somer2024) show that backsliding proceeds along distinct institutional pathways, among them legislative capture, plebiscitary refoundation, executive coups, and elite collusion, each opening and foreclosing different avenues of resistance. Chávez’s constituent assembly is not López Obrador’s legislative supermajority, which is not Bukele’s security-driven concentration of power; folding them into a single redistributive narrative risks flattening variation that is consequential for both explanation and resistance. The concern compounds a measurement issue raised by Knutsen et al. (Reference Knutsen, Marquardt, Seim, Coppedge, Edgell, Medzihorsky, Pemstein, Teorell, Gerring and Lindberg2024): whether a given case registers as erosion depends heavily on how democracy is operationalized. Under a thick V-Dem standard, Mexico under López Obrador counts; under a thinner standard centered on electoral competition, the verdict is less clear. Stokes is entitled to her choice, but the reader is owed more reflection on how much of the “wave” is an artifact of the measure. Weyland’s (Reference Weyland2024) caution against alarmism lingers as well: does the train metaphor, in which many leaders disembark before the terminus, already concede the institutional contingency he stresses, or does the book overpredict erosion? None of this diminishes the book’s accomplishment. The Backsliders is the most ambitious attempt yet to give the contemporary wave a single, historically grounded explanation, and its account of inequality and institutional trash talk will frame debates for years to come. But the pivot toward structure and rhetoric leaves the meso level of organizations, brokers, and distributive linkages undertheorized precisely where Latin American incumbents secure their support. That is best read not as a defect but as a generative limit: it marks an agenda for the comparative study of erosion, one that returns, fittingly, to the terrain Stokes herself helped to open.
