The publication of Scott Mainwaring’s new edited volume, Party Systems in Latin America: Institutionalization, Decay and Collapse (henceforth PSLA), is a major event in the fields of Latin American politics and comparative politics. In conjunction with its predecessor, Mainwaring and Timothy R. Scully’s Building Democratic Institutions: Party Systems in Latin America (1995, henceforth BDI)—PSLA is likely to become the leading reference on Latin American party systems for the next few decades.

When BDI came out, most theoretical literature on parties was based on studies of advanced Western countries, whose party systems importantly differed from those in Latin America. In much of Latin America, parties were not permanent fixtures in politics and did not structure elite and voter behavior. Voters did not identify with parties, and politicians did not need them to win office. Most parties were disposable, short-lived, personalistic vehicles. No concept in wide use captured these phenomena.

BDI changed that. Mainwaring and Scully argued that some Latin American party systems were not institutionalized, and that low party system institutionalization (or low PSI) undermined democratic quality and stability. The concept of PSI caught on quickly; today, virtually all party scholars, Latin Americanist and otherwise, understand and employ it. BDI also made a valuable empirical contribution, placing Latin American party systems on a spectrum from low to high PSI and paving the way for future scholars to examine the causes and effects of high and low PSI in their countries of interest.

After BDI’s publication, the term PSI took on a life of its own. In BDI, Mainwaring and Scully define PSI as a cluster of variables (i.e., electoral stability, organizational strength, societal rootedness, legitimacy), but scholars who later used and operationalized the term tended to equate PSI with the first of these. PSI and electoral stability became almost interchangeable. In chapter 1 of the new volume, Mainwaring, Fernando Bizzarro, and Ana Petrova pare down the definition of PSI, making electoral stability central and removing the other three variables as criteria.

This is the right move, for three reasons. First, the new definition, as suggested, more closely approximates scholarly usage. Second, most negative effects of low PSI are, in fact, negative effects of party system instability. When parties are electorally unstable, outsiders and novices win office more easily; voters find it harder to make electoral choices and determine who is accountable; politicians have shorter time horizons; policy outcomes are more volatile; institutions weaken; government performance worsens; and executive-legislative conflict and democratic breakdown are more likely. (In chapter 3, Mainwaring helpfully enumerates and elaborates on six effects of PSI. In chapter 14, Gustavo Flores-Macías elaborates on one of these effects, policy stability. Using data from all Latin American countries since the onset of the Third Wave, Flores-Macías argues that PSI leads to greater economic policy stability, with positive macroeconomic consequences.)

Third, as Mainwaring points out, variables such as organizational strength, societal rootedness, and legitimacy are often determinants of party stability. The original definition, then, may have obscured not only the relative importance of, but also the causal relationships between, its components.

BDI did not include programmatic or ideological stability as a defining feature of PSI. PSLA does. Samuel Huntington might have preferred the old definition; he argued that a party’s ability to change its program and remain electorally stable—something many clientelist and machine parties in Latin America have achieved (e.g., Peronism in Argentina, Mexico’s PRI, Peru’s APRA)—was a mark of institutionalization. On the other hand, as Mainwaring, Bizzarro, and Petrova note, stable programs and linkages “are integral to political actors’ ability to understand and form stable expectations about the party system” (24). Which definition is preferable? This question I leave open.

Nearly a quarter-century has passed since BDI’s publication. During this time, Latin American party systems have undergone considerable change. The main contribution of PSLA is to generate new empirical knowledge about this change. The volume reoperationalizes PSI, provides updated PSI scores for 18 Latin American countries (with the United States as a benchmark), compares these countries’ 2018 and 1995 scores, and categorizes the various pathways that Latin American party systems have taken in the last few decades. What do we learn?

Mainwaring finds that overall, PSI in Latin America has not declined since the mid-1990s. Party system trajectories have varied widely, with some party systems becoming significantly more institutionalized (Brazil, El Salvador, Panama), others significantly less (Venezuela, Argentina, Colombia), and still others experiencing relative continuity in their high (Chile, Uruguay) or low (Peru, Guatemala) PSI levels.

Mainwaring offers important qualifications, however. Although Chile’s party system remains institutionalized, for example, the main parties have suffered a recent erosion of partisanship. He also notes that several cases do not fit neatly into a single pathway. In Venezuela, Bolivia, and Ecuador, party systems collapsed in the 2000s but later underwent processes of asymmetric institutionalization as powerful governing parties emerged. Mexico, interestingly, has maintained an institutionalized party system since the early 1990s, but the system has shifted from hegemonic and authoritarian to multiparty and democratic.

What accounts for the varied trajectories of Latin American party systems? More broadly, what are the determinants of PSI? Instead of offering a single big answer, the volume offers multiple small ones. Mainwaring and Fernando Bizzarro, in their admirably cautious empirical analysis, mostly present null findings that run counter to established literature. Using data on the same 18 Latin American countries since the Third Wave, they find little to no relationship between PSI and hypothetical explanatory variables, such as institutional design, government performance, a country’s previous history of democracy, the longevity of a country’s current democratic or semidemocratic regime, party system polarization, and programmatic (as opposed to clientelistic or personalistic) party-voter linkages.

Interestingly, though, the other chapters in the volume—most of them single-case studies—often tell a different story. Take, first, the variable of institutional design. Mainwaring, Timothy J. Power, and Bizzarro attribute the increasing institutionalization of Brazil’s once inchoate party system partly to the implementation of new electoral laws (e.g., concurrent elections, greater public financing) in the mid-1990s. Conversely, Carlos Gervasoni ascribes the deep erosion of Argentina’s party system partly to institutional features, such as nonconcurrent elections and federalism.

Second, take government performance. Mainwaring, Power, and Bizzarro partially credit decades of low inflation and economic growth for Brazil’s increasing institutionalization. Conversely, case studies on Argentina (Gervasoni) and Colombia (Juan Albarracín, Laura Gamboa, Mainwaring) impute party system erosion partly to poor government performance: in Argentina to the 2001–2 economic crisis and in Colombia to the traditional parties’ “mismanagement of the security situation … and a severe security threat in the early 2000s” (228).

Third, take party system polarization. Mainwaring, Power, and Bizzarro argue that programmatic differentiation between Brazil’s major parties contributed to increasing institutionalization. Conversely, in his small-n comparative study, Noam Lupu posits that strange bedfellow alliances between once-differentiated parties contributed to the collapse or eroding partisanship of Argentina’s UCR, Bolivia’s MNR, and (less dramatically) Chile’s Concertación coalition.

Fourth and finally, take programmatic linkage strategies. Samuel Valenzuela, Nicolás Somma, and Scully argue that Chile has sustained high PSI for decades because Chilean voters can gauge the main parties’ distance from the rich and the Catholic Church and vote largely on this basis. Conversely, Gervasoni argues that in Argentina, programmatic inconsistency, brand dilution, and the personalization of politics under the Kirchners have contributed to party system erosion.

The volume offers additional insights. Kenneth Greene and Mariano Sánchez-Talanquer suggest that the nature of Mexico’s authoritarian regime favored high posttransition PSI. Unlike repressive authoritarian regimes, which “stunted genuine and independent opposition parties,” Mexico’s more permissive, competitive authoritarian PRI regime created space and generated incentives for opponents to invest in party building, craft distinct platforms, and develop constituency links (203). Jana Morgan finds that in Venezuela, the shift from ISI to neoliberalism weakened the programmatic, sectoral, and clientelistic linkages of Venezuela’s traditional parties, AD and COPEI, and that organizational and resource constraints prevented AD and COPEI from adapting. Jason Seawright finds that in Latin America as a whole, high PSI levels are associated with ethnic homogeneity, the availability of desirable careers in the public bureaucracy, and, most prominently, low socioeconomic development.

In Latin America today, many party systems are inchoate or semi-institutionalized (e.g., Argentina, Bolivia, Colombia, Ecuador, Guatemala, Peru, Venezuela), and relatively few have high PSI scores that rival those of the United States (e.g., Uruguay, Mexico, Chile). Noting similar phenomena in East and Southeast Asia and Sub-Saharan Africa, Allen Hicken and Rachel Beatty Riedl argue that inchoate and semi-institutionalized party systems are normal, not exceptional. Furthermore, they suggest that many such party systems are in, not out of equilibrium. Why might this be? In his chapter on Peru, Steven Levitsky argues that once parties collapse, extraordinary events (e.g., civil war, revolution, authoritarian repression, intense populist conflict) are usually necessary to rebuild party systems. In the meantime, politicians adapt to democracy without parties, and eventually, their alternative practices (e.g., partisan free agency, party switching, use of party substitutes) become informally institutionalized.

So we may have to get used to a world without parties. Can democracy function in such a world? Unfortunately, this is a question future scholars must tackle.