PARTY SYSTEMS IN THE TRANSITION FROM STATE-LED DEVELOPMENT TO MARKET LIBERALISM

In his seminal study of Latin American political development, Guillermo O’Donnell emphasized the importance of understanding period-specific political challenges that arise at particular stages in the process of social and economic development. According to O’Donnell (Reference O’Donnell1973: 65–6), ‘different levels of modernization . . . generate different constellations of issues that define each country’s problematic space. In turn, the set of political actors and their political responses (actors’ goals and coalitions, public policies, and political system types) are molded by these different constellations and by the different structures in which these constellations have emerged.’ In Latin America, the demise of state-led development in the debt crisis of the 1980s marked an epochal shift in political eras that thoroughly reconfigured the constellation of issues and the positioning of actors within the ‘problem space’. When initial efforts to respond to the crisis with heterodox adjustment measures collapsed under the pressures of hyperinflationary spirals, widespread capital flight and acute foreign exchange bottlenecks, the deepening crisis placed market-based austerity and adjustment measures at the forefront of the political agenda. This orthodox combination of austerity and structural adjustment posed serious challenges to party systems whose competitive alignments had been shaped by the rise of mass-based, labour-mobilizing parties during the era of state-led development.

Not every party system in the region, however, had been reshaped by historical patterns of labour mobilization and incorporation. In my book Changing Course in Latin America (Roberts Reference Roberts2014), I argued that the transition from state-led development to neoliberalism – a transition which I characterized as a new ‘critical juncture’ in the region’s political development – had different effects in countries that retained ‘elitist’ party systems from those that developed ‘labour-mobilizing’ party systems prior to the critical juncture of market reform. Countries were assigned to the labour-mobilizing category on the basis of two criteria: a relatively high level of trade union density, and the emergence of a major new populist or leftist party in the twentieth century that played a leading role in the labour movement and rose to become one of the two largest competitors in national elections. Countries retained elitist party systems where trade union density was low and ‘the largest union confederation was affiliated primarily with traditional oligarchic parties, or with leftist parties that remained electorally marginal’ during the era of state-led development (Roberts Reference Roberts2014: 71). In my account, labour-mobilizing cases – which typically attained higher levels of programmatic structuration – were plagued by more severe economic crises in the 1980s and early 1990s and, in the short-term, higher levels of electoral volatility.

In his aforementioned review article, Mainwaring (Reference Mainwaring2016: 707–12) challenged my assignment of cases to the elitist and labour-mobilizing categories and, hence, the theoretical inferences that I drew for explaining party system volatility and the severity of economic crises during the period of transition to market liberalism. A proper coding of cases, he claimed, would show that labour-mobilizing party systems were more, not less, stable during the period in question, and no more prone to severe economic crises as a result of acute distributive conflicts. While conceding that ‘the combination of profound economic crisis and deep economic restructuring had a huge impact on Latin American party systems’, Mainwaring (Reference Mainwaring2016: 714) concluded that ‘change in economic models in a critical junctures analysis does not offer a convincing parsimonious guide to understanding the dynamics of these systems’.

Clearly, an accurate coding of cases is critical for effective comparative historical analysis, particularly that which relies heavily on qualitative methods of casual inference. Mainwaring’s critique, however, is predicated on a questionable reinterpretation of the historical evolution and sociological foundations of party systems prior to neoliberal reform. Although these party system attributes and the varied state-centric development matrices in which they were embedded were ‘antecedent conditions’ in my critical juncture analysis, and thus not decisive for determining longer-term institutional effects, they nonetheless predisposed countries to experience the critical juncture in different ways. As such, Mainwaring’s critique warrants closer examination for those seeking to understand a critical period in the historical development of Latin American party systems.

The centrepiece of Mainwaring’s critique is a claim that my elitist and labour-mobilizing categories were static, and thus failed to account for shifts from one category to another in seven of the 16 countries included in the study (mostly from the labour-mobilizing to the elitist category). Mainwaring’s own critique, however, clearly demonstrates that these categories were not static in my analysis, as he proceeds to reject my updating and reassignment of Nicaragua from the elitist to the labour-mobilizing category following the Sandinista Revolution in 1979. As explained in my book, Nicaragua had a long-standing elitist (patrimonial) party system under the Somoza dynasty prior to its revolutionary transformation, but the country did not go through the economic crisis of the late 1980s and its subsequent process of market reform under Somoza’s patrimonial system, as Mainwaring’s recoding would have it. A revolutionary labour-mobilizing system was assuredly in place during the critical juncture of market-based structure adjustment.

Mainwaring’s other recodings reflect similar problematic interpretations of the historical record, in part attributable to his decision to make 1979 a uniform coding benchmark for party systems prior to the onset of the debt crisis and the critical juncture of market reform that followed in its wake.Footnote ¹ A standard challenge faced by critical juncture approaches is the identification of the time period when a critical juncture begins and ends, given that different countries may encounter similar strains, shocks or policy choices at distinct historical moments (see Collier and Collier Reference Collier and Collier1991). In my account (Roberts Reference Roberts2014: 10), I argued that national critical junctures ‘began to unfold’ with the exogenous shock of the 1982 debt crisis but reached their ‘truly decisive stage’ with the ‘adoption of orthodox market reforms’. In many cases this decisive stage did not occur until the late 1980s or early to mid-1990s, following failed attempts to achieve stabilization through heterodox or statist adjustment measures (in particular, wage and price controls). In restrospect it might have made sense for me to analytically separate the onset of the crisis from the occurrence of the critical juncture, when partisan alignments around the process of orthodox reform proved decisive for the structuring of post-adjustment competitive dynamics. The key point, however, is that Mainwaring’s 1979 benchmark fails to take into account ongoing evolutionary patterns in party systems during the 1980s that contributed, in many countries, to a deepening of the crisis at the end of the decade and the configuration of partisan actors during the critical juncture itself.

Mainwaring argues that five of my labour-mobilizing cases reverted to the elitist category prior to 1979, but these recodings misrepresent the competitive dynamics and sociological bases of national party systems as the crisis unfolded in the 1980s. In Peru and Bolivia, for example, Mainwaring is right – as my analysis recognized – that the traditional populist labour-mobilizing parties (Alianza Popular Revolucionaria Americana–Partido Aprista Peruano (APRA – American Popular Revolutionary Alliance–Peruvian Aprista Party) in Peru and the Movimiento Nacionalista Revolucionario (MNR – Revolutionary Nationalist Movement) in Bolivia) lost their leadership of the trade union movement over the course of the 1960s and 1970s. This hardly meant, however, that these countries rolled the clock back to the times of elitist party systems with weak labour movements. Indeed, the exact opposite occurred: traditional populist parties were progressively displaced in labour movements by more radical leftist rivals, labour and popular mobilization intensified, and leftist coalitions emerged as major electoral contenders heading into the crises of the 1980s.

In Peru, a Communist Party-led labour confederation overtook APRA’s union movement as the country’s largest confederation in the 1970s, while the Communist Party and five other radical left parties formed a new electoral coalition, the Izquierda Unida (IU – United Left), as Peru returned to democratic rule in 1980. The IU subsequently developed into the country’s second leading electoral force until the party system imploded in 1989–90 in a context of acute economic crisis and revolutionary violence (see Roberts Reference Roberts1998). In Bolivia, a leftist breakaway faction of the MNR forged an alliance in the late 1970s with other pro-labour leftist parties, including the Bolivian Communist Party and the Movement of the Revolutionary Left. This alliance won national elections in 1980 and eventually took office during a turbulent democratic transition in 1982.

As Bolivia descended into a hyperinflationary spiral in the middle of the decade, even this leftist government was unable to control the country’s powerful mining-based labour confederation, which had a tradition of militant Trotskyism and diffuse partisan sympathies. The tenuous nature of the organizational linkages between ruling leftist parties and militant labour unions, however, hardly warrants a reassignment of Bolivia to an elitist category comprising countries that possessed weak labour movements and electorally marginal left or populist parties. In comparative perspective Bolivia was an exceptionally strong case of left-labour sociopolitical mobilization and class conflict during the crisis of the 1980s. To conceptualize the country in elitist terms, alongside the likes of Honduras or Paraguay, is to miss the central purpose of the coding exercise: to differentiate countries according to the strength of working- and lower-class sociopolitical mobilization behind statist and redistributive policy platforms. This differentiation was critical for explaining cross-national variation in both electoral volatility and the depth of economic crises during the turbulent transition to market liberalism.

Similarly, Mainwaring’s recoding of Brazil as an elitist case is only applicable to the post-1964 military regime, when both electoral competition and labour mobilization were severely and artificially restricted by authoritarian repression. During periods of open democratic contestation, however – when social and political actors are free to mobilize core constituencies and their political weight can be more accurately assessed – Brazil’s party system had strong labour-mobilizing tendencies both before 1964 and after the regime transition in 1985. A massive strike wave against the military regime in the late 1970s led to the founding of the leftist Partido dos Trabalhadores (PT – Workers’ Party) in 1980, setting the stage for a rapid reconstruction of a labour-mobilizing party system as the country began its democratic transition. Indeed, the PT candidate finished a close second in the first national presidential election at the end of the decade, in the midst of a deepening hyperinflationary crisis. Brazil thus had a labour-mobilizing party system in place prior to the process of structural adjustment in the early to mid-1990s, when the critical juncture occurred.

Mainwaring also recodes Mexico and Venezuela as elitist cases, perhaps reflecting a semantic disagreement over my use of the term ‘labour-mobilizing’ rather than ‘labour-based’ party systems. As my analysis recognized (Roberts Reference Roberts2014: 70n.), the long-standing electoral dominance and governing responsibilities of a historical populist party in these countries led to a gradual shift towards corporatist labour control rather than militant forms of labour mobilization (thus helping Mexico and Venezuela avoid the severe hyperinflationary pressures that other labour-mobilizing cases encountered in the 1980s). This shift, however, did not give these countries the sociological foundations and competitive dynamics of elitist party systems. Semantics aside, their labour party attributes continued to meet the criteria I established for assignment to the labour-mobilizing category. Indeed, an illustrious body of research demonstrates that these parties’ strong labour bases generated distinct political challenges – as well as opportunities for control – during the period of neoliberal reform (Burgess Reference Burgess1999, Reference Burgess2004; Burgess and Levitsky Reference Burgess and Levitsky2003; Murillo Reference Murillo2000, Reference Murillo2001). An understanding of these challenges would hardly be advanced by reassigning these countries to the elitist category.

Finally, Mainwaring argues that Uruguay transitioned in the other direction – from elitist to labour-mobilizing – and thus received an improper coding in my analysis as well. The Uruguayan case is discussed at some length in my volume, and I emphasized that the country exhibited hybrid characteristics and may well have been in transition to a labour-mobilizing system until the 1973 military coup nipped the process in the bud. Under my coding criteria, however, Uruguay clearly belonged to the elitist category, as the country’s two traditional nineteenth-century oligarchic parties remained electorally dominant through the crisis of the 1980s and the critical juncture of market reform in the early 1990s. The leftist Frente Amplio (FA – Broad Front) did not become one of the two leading electoral contenders until the post-adjustment era, when it became progressively stronger.

In short, my categorization of cases was not static, and it was sensitive to the impact of major political upheavals like the 1952 Bolivian Revolution (which relocated Bolivia to the labour-mobilizing category), the 1979 Nicaraguan Revolution, and the authoritarian interludes that disrupted and distorted party system development in much of the region. Shifts between the elitist and labour-mobilizing categories, however, were not common beyond the initial process of mass political incorporation, and none of the countries that developed a labour-mobilizing party system can properly be said to have reverted to the elitist category for the critical juncture of market reform. The partisan leadership of and linkages to labour unions clearly evolved over time, but once strong labour movements had emerged, they heavily conditioned the political dynamics of market liberalization under democratic regimes. My book did not argue, however, that labour-mobilizing party systems per se caused severe economic crises or greater electoral volatility. Instead, I argued that these party systems belonged to a deeper state-centred matrix of development that collapsed in the 1980s, leading to severe economic crises and political upheaval. The party system component of this matrix was not present at all times in each country, for the reasons Mainwaring suggests, but the sociopolitical pressures that spawned these parties and complicated the process of economic adjustment were surely operative by the beginning of the 1980s in each case.

THEORIZING PERIOD EFFECTS: ECONOMIC CRISES AND PARTY SYSTEM STABILITY

The proper coding of cases is critical for drawing accurate theoretical inferences about the determinants of severe economic crises and electoral upheaval in the 1980s and 1990s. The demise of state-led development and the transition to market liberalism played out very differently in countries where partisan and electoral competition were heavily conditioned by formidable patterns of populist or left-labour sociopolitical mobilization. These sociological foundations of party systems sharply differentiated labour-mobilizing cases like Bolivia, Brazil, Peru and Nicaragua from elitist cases with very weak labour movements, such as Colombia, Honduras and Paraguay. Conflating them, as Mainwaring’s coding scheme would have it, does not advance our understanding of Latin America’s travails in the waning decades of the twentieth century.

In Mainwaring’s account (2016: 710–11), for example, severe economic crises in the 1980s appear to be attributable largely to political voluntarism – i.e. the adoption of ‘ill-advised’ or ‘misguided’ heterodox and nationalistic economic policies. That misguided policies abounded in the region is indisputable, but it is important to recognize how policy choices were shaped and constrained by underlying sociopolitical pressures and the conflicts they generated. Where labour was well-organized and pro-labour populist or leftist parties were major electoral contenders, even conservative rulers were reluctant to impose orthodox austerity and adjustment measures that posed serious risks of strike waves, mass social protest and electoral backlashes. Argentina’s Peronist labour movement, for example, declared 13 general strikes against the tepid stabilization measures introduced by the non-Peronist government of Raul Alfonsín over the latter half of the 1980s – surely a major constraining factor in his government’s ‘ill-advised’ policy choices. Similarly, as Bolivia’s ill-fated leftist government struggled in vain to contain rampant inflationary pressures, the national labour confederation ‘staged nine general strikes and a staggering 1,799 protest events that included strikes, marches, roadblocks, demonstrations, and hunger strikes’ (Silva Reference Silva2009: 106n.).

It is not surprising, then, that many countries – especially those in the labour-mobilizing category, including Argentina and Bolivia – initially responded to the debt crisis with heterodox adjustment measures that only postponed the day of reckoning. Such measures ultimately led to acute hyperinflationary spirals and more wrenching, comprehensive, neoliberal ‘shock treatment’ once labour had been weakened by the crisis (as in Peru) or pro-labour parties were in office to contain union resistance (as in Argentina under Carlos Menem). Among the labour-mobilizing cases, the scourge of hyperinflation was only avoided in the two countries, Mexico and Venezuela, where populist parties had long-standing governing responsibilities that allowed them to develop state corporatist control over labour mobilization. In contrast, none of the elitist cases with weak labour movements experienced hyperinflation, and most moved gradually towards market orthodoxy (starting from less statist baselines) over the course of the 1980s and 1990s.

More ad hoc explanatory factors introduced by Mainwaring to account for different levels of crisis add little to our theoretical understanding, and they work against the aspiration for theoretical parsimony. The US-funded war against the Sandinistas’ revolutionary government in Nicaragua undoubtedly contributed to that country’s economic crisis, as Mainwaring suggests (2016: 710), but that war was endogenous to the revolutionary transformation that relocated Nicaragua to the labour-mobilizing category. The US surely would not have waged a proxy war against the friendly patrimonial regime of the Somoza dynasty, had it still been in place during the critical juncture. Major civil wars were fought in five Latin American countries during the 1980s – Colombia, El Salvador, Guatemala, Peru and Nicaragua – but only the two labour-mobilizing cases, Peru and Nicaragua, experienced hyperinflation and severe economic crises, by regional standards. Likewise, it was only in these two countries where domestic political alignments and economic policies clashed with US prescriptions for managing the transition to market liberalism, leading to international financial pressures that exacerbated national crises.

Similarly, a proper coding of the cases is essential for our theoretical understanding of the period-specific factors that influenced party system stability and volatility during Latin America’s transition to market liberalism. My account seeks to explain why countries with nineteenth-century oligarchic party systems (Colombia, Uruguay, Paraguay and Honduras), as well as several other elitist party systems of more recent formation (Costa Rica and the Dominican Republic), weathered the storm in the 1980s and 1990s with relatively moderate economic crises and low-to-moderate levels of electoral volatility. Other than in Uruguay, where a leftist opposition party progressively strengthened over time, partisan competition in these countries was not well-structured by ideological or programmatic differences. This was largely attributable to the historic weakness of labour and popular mobilization, and the resulting political dominance of elitist or multi-class parties with shallow trade union constituencies and business-friendly policy orientations.

Political upheaval was more common in labour-mobilizing cases, however, where competitive alignments were more sociologically grounded in class distinctions, corporatist forms of interest representation, and state-centred development policies. In the latter cases party systems typically entered the critical juncture with greater programmatic structuration and more densely organized popular constituencies, which are conventionally understood to be correlates of electoral stability (Bartolini and Mair Reference Bartolini and Mair1990; Mainwaring and Zoco Reference Mainwaring and Zoco2007). These were, however, the systemic attributes most directly threatened by the collapse of state-led development, the informalization and de-unionization of the workforce, the shift from corporatist to pluralist forms of interest representation, and the technocratic policymaking convergence around variants of market liberalism. As such, the destabilizing effects of the transition to market liberalism were heavily concentrated in countries with labour-mobilizing party systems.

In O’Donnell’s (Reference O’Donnell1973) terms, however, the ‘constellation of issues’ created a very different ‘problematic space’ for party systems once Latin America entered the post-adjustment political era by the second half of the 1990s, following the adoption of neoliberal reforms and the defeat of hyperinflation across the region. As Polanyi (Reference Polanyi1944) would have it, societal resistance to market orthodoxy strengthened in the post-adjustment era, producing a leftward shift in voting behaviour and a revival of popular (often non-labour based) movements that repoliticized social and economic inequalities. Party systems varied dramatically, however, in their capacity to institutionally channel this societal resistance, depending on the configuration of partisan actors around the process of market reform during the critical juncture. Indeed, the programmatic structuring of partisan competition emphasized by Mainwaring did become paramount in this post-adjustment era, creating a strikingly different set of period effects. Where conservative actors had imposed neoliberal reforms over the principled opposition of a major leftist party, the latter could channel societal resistance in ways that aligned and stabilized partisan competition. In countries like Brazil, Chile and Uruguay, therefore, relatively moderate and institutionalized leftist parties – the PT in Brazil, the Socialists in Chile and the FA in Uruguay – gradually strengthened and won national elections in the post-adjustment era, but centrist and conservative rivals remained major contenders, and party systems did not break down.

By contrast, ‘bait-and-switch’ market reforms imposed by traditional labour-based populist or leftist parties de-aligned party systems programmatically, leaving them vulnerable to a wide range of destabilizing ‘reactive sequences’ (Mahoney Reference Mahoney2001) in the aftermath to the critical juncture. Such patterns of reform caused mainstream parties to converge on neoliberal orthodoxy – an archetypal example of Lupu’s (Reference Lupu2016) ‘brand dilution’ – and channel societal resistance into extra-systemic or anti-systemic forms of social and electoral protest, as seen in countries like Venezuela, Bolivia, Ecuador and Argentina. Reactive sequences in these countries included the toppling of presidents by mass protest movements, the partial or complete breakdown of traditional party systems and the election of new populist or ‘movement’ contenders on the left flank of the old order.

Theorizing such period effects – in short, recognizing causal heterogeneity – is admittedly not parsimonious. It cuts across the grain of efforts to identify general causal propositions that are universally applicable across time and space. Such universal propositions, however, are necessarily thin, inaccurate or incomplete in a real world marked by historical complexity and spacio-temporal variation. To say, for example, that the prospects for party system institutionalization are weakened by severe economic crises is undoubtedly true, but a relatively superficial first-cut at theoretical explanation. We need to know whether some types of party systems are embedded in political economies that are more susceptible to crises, at particular junctures or stages of capitalist development, than others. Similarly, the argument that strong programmatic brands are conducive to party system institutionalization is intuitively appealing, but questionable as a universal causal proposition. The Latin American experience suggests that the causal nexus may be time-bound and period-specific, as historical junctures may exist where certain types of party brands or programmatic and sociological alignments – such as those found in many labour-mobilizing party systems in the 1980s – are incompatible with evolving market constraints.

That said, the central theoretical arguments of Changing Course in Latin America (Roberts Reference Roberts2014) can be succinctly stated: party system instability in the region was more likely to be found in labour-mobilizing party systems during the transition from state-led development to neoliberalism, and in programmatically de-aligned party systems in the post-adjustment era. That is a reasonably parsimonious starting point for explaining party system variation across 16 countries and four decades of political development that entailed multiple shifts in the political winds – or ‘constellations of issues’, as O’Donnell might have put it. The theoretical reasoning that undergirds such sweeping arguments, however, necessarily requires that the study of political institutions be coupled with an analysis of their shifting economic moorings, an approach that O’Donnell pioneered and Luna et al. (Reference Luna, Murillo and Shrank2014) have more recently urged the field to revive. It also requires far more attention to the sociological underpinnings of partisan alignments than most contemporary scholarship affords.

In short, explaining period effects on the stability of Latin American party systems requires the theoretical integration of multiple social fields, something that a narrow focus on parties as political institutions rarely affords. Such integration makes it possible to explore how party systems are embedded in larger political economies, how and why these institutional settings change over time, and how economic crises or reforms alter the issue agenda, transform the social landscape, and reconfigure the actors and interests that compete for partisan representation. An integrative approach may also generate novel theoretical insights in other domains of inquiry; although Changing Course was primarily a study of party system change and continuity, it spawned derivative propositions regarding patterns of crisis and reform during the transition to neoliberalism, the political and economic structuring of social protest, variation in Latin America’s post-adjustment ‘left turn’, and patterns of oligarchic and populist ‘careening’ under democratic regimes in the neoliberal era (see Roberts Reference Roberts2016).