Questioning Backsliding

Nancy Bermeo, Journal of Democracy, october 2022

Nancy Bermeo is senior research fellow in politics at Nuffield College, Oxford University. Her essay “On Democratic Backsliding” appeared in the January 2016 issue of the Journal of Democracy.

Jason Brownlee and Kenny Miao deftly question widespread views about the connections between democratic backsliding, democratic breakdown, and a global wave of autocratization. This brief response highlights the practical political questions that emerge from their findings and from the structural arguments they use to justify their relatively positive forecasts. The questions involve: backsliding’s breadth, location, and assessment; backsliding’s connections with the military; how recent changes in capitalism and party competition affect democratic resilience and, most important, why democracy’s defenders succeed or fail. Tracing and naming trends is useful, but the comparative study of how individual countries resist or reverse backsliding is essential.

The political theorist Jan-Werner Müller began a recent talk on contemporary Hungary by noting that he “never makes predictions, especially before the fact.” His clever remark came back to me as I read Jason Brownlee and Kenny Miao’s essay. Like many political scientists, they do hazard predictions “before the fact.” They assert that contemporary democracies are more “sturdy” than imagined, and that democratic backsliding is much less likely to lead to autocracy than is currently assumed. Without denying some sinister similarities between the twentieth century’s interwar years and the past decade, they argue, in essence, that the global “specter of democratic collapse” is just that—a ghost emanating from illusions about the past and a clouded view of the present. Tallying cases of democratic backsliding versus actual democratic breakdowns since 1920, they conclude that the new hazards facing democracy have so far had no net impact, and that “the tale of backsliding preceding breakdown has been much less common than the story of backsliding—and recovery—within electoral democracies.”

These conclusions counter scores of dire forecasts, and may console many who fear democracy’s collapse, but the case that Brownlee and Miao make will certainly roil those who see democracy’s present and future in grimmer terms. The debate about whether the world faces viral neofascism or simply a phantom will go on, not least because, as David Waldner and Ellen Lust have noted, backsliding “involves relatively fine-grained degrees of change.” They deemed the study of backsliding a “new research frontier” in 2018.¹ While research has proliferated since then, the term’s meaning and measurement still vary, as much or more than before.

We see this variation in meaning within the Brownlee and Miao essay. Drawing on my own early attempt to grapple with the concept, the authors define backsliding as “the state-led debilitation or elimination of any of the political institutions that sustain an existing democracy.” By “political institutions,” I meant only what James March and Johan Olsen called “concrete” institutions such as legislatures, electoral and legal systems, militaries, and political parties.² Brownlee and Miao say that backsliding can mean “norm erosion” as well. This is simply a difference rather than an error—many fine scholars view norms as “institutions”—but it highlights how elusive a common definition can be.

It also shows how backsliding’s different meanings can complicate causal stories. I currently view norm erosion less as a sign of backsliding than as a precondition. Formal democratic institutions become hollowed out or disassembled because the norms that sustain them weaken. In any case, if backsliding includes norm erosion, then whose norms erode and to what effect? Whose norm erosion matters most? My study of democratic breakdowns in interwar Europe led me to conclude that elites were the guilty parties. I suspect this is also the case with backsliding today. As Larry Bartels has shown, the democratic norms and values of ordinary Europeans have been surprisingly constant since 2004 and have “generally not been decisive in accounting for toxic politics … or democratic backsliding.”³ Of course, Brown-lee and Miao cannot be expected to deal with all this in a brief essay. My point is that the broader our conceptualization of backsliding, the greater the challenge of assessment.

The subject of assessment brings us to the difficulties of measuring backsliding and to the highly consequential differences among scholars when it comes to measurement tools and formulas. A useful study by Hans Lueders and Ellen Lust compared five indicators of backsliding and found their average pairwise correlation coefficient to be only 0.33, even lower than the 0.46 correlation for regime transition.⁴ Unless measures of both backsliding and democratic breakdown are identical, even careful scholars using similar datasets can draw radically different conclusions about how these two phenomena are associated.⁵

Given the difficulties that surround questions of definition and measurement, perhaps the work of naming trends that everyone agrees are troubling should take a backseat to learning how to check or even reverse them. We know too little about what works or does not work in various institutional landscapes, and how existing democratic guardrails can best be defended.

Happily, Brownlee and Miao give us ample material to expand our strategic inquiry. Although they insist that “practical means for preventing democratic breakdown” are beyond their essay’s scope, they do offer some leads.

One arises from their historical comparison of paths to breakdown. They find that while executive takeovers are up, military coups are down, leaving the overall number of electoral democracies stable. This should lead us to ask: When do armed forces favor and defend democratic institutions? Appeasing a military’s interests and concerns can work wonders, as Brownlee and Miao note, but as we learned in Venezuela and elsewhere, democracy’s disassemblers can bid for military support too. We have much to learn about when and why military elites may tolerate backsliding, or even support it.

Two even broader areas of inquiry come from Brownlee and Miao’s bold argument about the structural conditions that will hold backsliding in check. The first involves their data-driven conclusion about the protective power of national wealth (outside of rentier economies) above a certain economic-development threshold. This argument might comfort the anxious in rich democracies. It might also be used to bolster the case for development aid, which has begun to weaken as rising economic distress and feelings of national insularity have taken a toll on donor budgets in recent years. Yet, as the authors admit, the core of their argument has a “storied” history. Any brief article is simply a new chapter and I suspect the story is not over.

Weaving an economic-growth argument from a century’s worth of cases has undeniable merits and is far superior to the unsubstantiated claims that the authors seek to counter. But what if the nature and implications of economic growth are changing in relevant ways?

After all, GDP per capita is not an institution or an actor. It is simply a numerical proxy for the complex web of socioeconomic relations that Lipset and others saw as the actual mechanisms joining classes and social groups to a democratic form of rule. The correlations presented by Brownlee and Miao raise pressing questions about how these mechanisms are being shaped by the changed nature of economic development and wealth today. Our two authors point to the “socioeconomic conditions” that have “tamed” those who would be autocrats, but we must ask if those conditions can remain in place amid de-industrialization, globalization, increasing precarity, and levels of wealth concentration that were unanticipated when modernization theory was born.

The macro-level analysis presented here also invites more research about the timing and terrain of the taming process. Vanessa Boese and coauthors offer us a useful lens for examining the timing question by separating resilience to the onset of backsliding from resilience to breakdown after backsliding sets in.⁶ In line with Brownlee and Miao, these researchers find that economic development is associated with resistance to the onset of backsliding. Yet they also find, unlike Brownlee and Miao, that there is no positive correlation between a higher level of development and greater resilience to breakdown after backsliding has begun. Here, as in so much of our work on backsliding, different conceptual frameworks and statistical treatments yield different results.

If the question of when economic development constrains backsliding remains contested, the question of where development constrains erosion remains underexplored. Brownlee and Miao are right to assert that “a country’s macroeconomic profile communicates important information for interpreting its internal politics,” but applying their reasoning below the national level will also bring rewards. Regional economies within countries—even wealthy countries—vary enormously. Some have facilitated regional backsliding and even regional authoritarianism. Whether the phenomenon of regional backsliding within countries supports or undermines Brownlee and Miao’s argument demands further research.

So too does the fact that ordinary people are most affected by their local economies. If norm erosion is, indeed, a form of backsliding, we must analyze how local economies change or stabilize democracy-relevant norms including tolerance, solidarity, threat perception, participation, propensity for violence, and more. In sum, Brownlee and Miao present a strong correlational analysis that raises questions about how, when, and where economic development safeguards democracy—or fails to do so.

Brownlee and Miao perform the same agenda-setting role with their argument about the stabilizing “momentum of multipartism.” Here they align with a vast and less “storied” literature linking democratic stability to party competition. Most important, they point to a democratic guardrail that can be built much faster than a developed economy. If this argument holds, it is of broader applicability.

But where and when does the argument hold? A short section of a short essay does not allow an answer. There are certainly cases where backsliding executives lose elections but what distinguishes these from others where power-grabbers are reelected or followed by handpicked successors?

I suspect that part of the answer lies with the order in which democratic institutions get disassembled. If aspiring autocrats manage, before seeking reelection, to craft biased electoral laws or new constraints on associational life, then multipartism may not matter much. I also suspect that the positive role which “competition” plays in the authors’ argument is far from constant. Party competition can be a lethal liability for a troubled democracy if it splits an aspiring autocrat’s democratic opposition.

Finally, I suspect that multiparty competition (like economic development) is undergoing changes that may have permanently altered the dynamics of the past. With social media and relatively little time, would-be autocrats can now create new, personalist parties, win executive power, and then engage in the older practice of using state resources to expand their party further. Will the momentum of multipartism remain powerful with the rise of social media alongside media silos?

Each of these suspicions and the many other questions raised above invite more intensive, country-based research. Fortunately, the cases that Brownlee and Miao sort into survivors and fatalities provide a useful place to start. Once we update their tallies and distinguish between countries that reversed backsliding and countries that are simply holding it in check, the work can begin.

Moving from counting cases to systematically comparing individual countries will enable us to examine the authors’ arguments about the military, development, and multipartism more closely. It will also help us to answer the questions that I have posed above. Most important, perhaps, it will enable us to help craft the “Resistance Playbook” that Luca Tomini, Suzan Gibril, and Venelin Bochev have proposed.⁷ Despite its brevity, Brownlee and Miao’s argument contributes to a wide range of tactical as well as academic debates. In so doing it raises the probability that some of our most consequential future predictions may be correct.

Notes