On the Theory, Influence and Evidence
At the heart of O’Donnell’s work is a puzzle: although at first glance Latin American countries seem to comport with the predictions of modernization theory, if one re-specifies the explanatory variable to capture industrial development, then modernization seems to have a dramatically different effect than the one predicted by Lipset. Indeed, O’Donnell writes, in South America “political authoritarianism – not political democracy – is the more likely concomitant of the highest levels of modernization.” [O'Donnell, 1973 #103: 8] Based on this observation O’Donnell ripostes what he calls the “optimistic equation” (in which development leads to democracy), and outlines a model in which authoritarianism is the byproduct of development.
O’Donnell begins his essay by reevaluating the methodological basis of standard modernization theory, arguing that mean GNP per capita is a highly flawed measure of modernization. Comparisons, he notes, are more appropriately drawn between the modern sectors of states – to this end he outlines a measure based on the degree of modernization (low, intermediate, and high) of these national ‘centers.’ Arranged according to this measure Latin American states present a striking pattern: democratic governments are only found in states with intermediate levels of modernization – countries with high and low degrees are uniformly non-democratic. According to this superior measure, he concludes, conventional modernization theory not only fails, but seems to be diametrically opposed to the truth.
Armed with this evidence O’Donnell presents a theoretical mechanism by which increased ‘modernization’ leads to authoritarian government. Within his model he categorizes political systems dichotomously: as “excluding” (in which the government attempts to deny influence to an “already activated” urban popular sector), and “incorporating” (in which the government seeks to activate and give voice to this group). Interestingly, there is a perfect overlap between excluding political systems and highly developed states in his sample – the only two countries in the former category (Brazil and Argentina) are also the only two in the latter as O’Donnell defines it.
Using a general case study of the two countries, he presents a model of shifting coalitions as the driving force behind the evolution from oligarchy to protectionist populist regime, and finally to bureaucratic-authoritarianism. Initially it is a multi-class coalition of urban and industrial interests which rises in opposition to the free trade policies of oligarchies, seeking to implement protectionist policies in order to decrease reliance on foreign firms while increasing industrialization and domestic production. The initial, “easy,” stage of industrialization (focused on replacing imports with consumer goods) serves a broad spectrum of interests among coalition members as long as this “horizontal” expansion continues to be possible. Modernization increases “social differentiation” and political pluralism, and high levels of public mobilization (a byproduct of modernization) lead to increased political activation among the popular sector.
Once demand for consumer products is exhausted – and the country is unable to “deepen” industrialization to reduce its dependence on imports for “intermediate” and “capital” goods - economic woes (a foreign exchange shortage, inflation, and reduced growth) and the possibility of reform highlight divergent interests within the populist coalition that cause it to splinter. Without the power of the coalition, the popular sector becomes increasingly politically isolated – forced to engage in ever-higher levels of mobilization for shrinking returns on this investment. The high levels of mobilization lead to a Huntingtonian “mass praetorian” stalemate in which naked power and political survival assume chief importance, and the desire to placate threatening actors trumps the ‘problem-solving’ role of the government.
Under this stalemate, new coalitions arise between actors seeking to further their interests by instituting a more efficient government. O’Donnell places particular emphasis on what he calls “technocratic roles,” filled by people trained in techniques of production, planning and control, and created by the increased importance of technology in a social structure made increasingly complex by modernization. Where high levels of modernization have resulted in the dense penetration of crucial social sectors with such technocrats, they are likely to be “highly confident” of their capabilities, and easily coordinated (by virtue of conditioning around a similar world-view). It is this group that is likely to band together to form a coup coalition in order to overcome the deadlock impeding effective political action.
In staging their coup, O’Donnell argues, these actors will attempt a “drastic re-shaping” of the social structure in an attempt to create more extensive opportunities for them to apply their expertise and expand the influence of those sectors which are heavily technocratic. They will be expected to employ substantially coercive tactics to overcome mass praetorianism by repressing the mobilized public sector. It is this particular form of exclusionary government by a group of efficiency-minded technocrats which O’Donnell terms bureaucratic-authoritarianism.
This model was quite influential in the contemporary literature on Latin America, in large part because it synthesized a rebuttal of conventional modernization theory with dependency theory. As O’Donnell was writing, twenty years of economic growth in Latin America had not produced the democratizing effects predicted by Lipset and his followers. The quest for economic development had been, writes Hirschman, “so dominant a theme” in Latin American politics that a systematic connection between it and major political trends had “an inherent intellectual appeal.” [Hirschman, 1979 #104: 71]
Military coups in two of the region’s more developmentally-promising countries (Brazil in 1964 and Argentina in 1966) posed a conundrum for modernization theory: not only had growth not led to stable democracy, but the authoritarian regimes were now taking aggressive steps to achieve growth by attempting to attract foreign investment. It looked, according to Hirschman, “increasingly as though the effort to achieve growth, whether or not successful, brings with it calamitous side effects in the political realm.” [Hirschman, 1979 #104: 62] Kaufman argues that it is these governments’ combination of exceptional repression and economically rational attempts to establish “pragmatic and predictable relations with the private sector” which makes the subject particularly compelling [Kaufman, 1979 #108: 166].
Indeed, this result took on particular salience in a state system divided between capitalist and socialist states. Cardoso argues that the coups were essentially a move by conservative, capitalist elites who saw themselves as “the defenders of Christian and occidental values against the world wide menace of communism” in the face of a “powerful threat” from the left [Cardoso, 1979 #106: 45]. In many ways it does not seem unreasonable to see the causal arrow that O’Donnell draws as support for the common pseudo-socialist arguments that domestic redistribution in Latin America ought to precede integration into the international capitalist system.
Dependency theory – and, particularly, the impact of foreign interests on domestic politics and economics in Latin America – is a particularly important aspect of this debate. In a critique Serra categorizes the literature on bureaucratic-authoritarianism as arguing that this regime-type is “the result of a dependency on international monopoly capital.” [Serra, 1979 #107: 99] Cotler examines the role of what he calls “enclave” societies, in which foreign capital firms penetrated and aligned themselves with the domestic bourgeoisie, creating a “total subordination” of local landholders in the name of the creation of opportunities for national development [Cotler, 1979 #109].
In spite of its tremendous influence, however, O’Donnell’s theory does not seem to hold up very well in light of contemporary statistical evidence. As an aside, before we examine the model’s predictive accuracy it is worth leveling a methodological criticism – especially as the first chapter of his book is largely spent, as Przeworski puts it, “drag[ing] Lipset over the coals for various methodological transgressions.” [al, 2000 #110:99] Rather than developing and testing his model on separate cases, O’Donnell limits his discussion to Latin America – and particularly the ‘southern cone’ countries. His initial intuition is based on the distribution of modernization and regime-type in a sample of ten Latin American countries including Brazil and Argentina. He goes on to use these two countries as case studies for the development of his theory – and then essentially ‘tests’ the theory with a detailed narrative about Argentina. This problem proves significant when we look at the theory’s ability to ‘travel’ – Przeworski et al. note that Argentina “turns out to be a distant outlier,” and take this result as evidence that Lipset’s exogenous modernization argument is correct [al, 2000 #110: 101].
Indeed, the most current large-n studies of regime transitions and coups provide substantial evidence to rebut O’Donnell’s theory. Przeworski et al. note a very strong relationship between democratic stability and per capita GDP – while there is a substantial chance that a democracy will die at a GDP/cap of less than $1000, no democracy has ever died in a country with a GDP/cap level above $6,055 (a maximum point, interestingly, set by O’Donnell’s main case). O’Donnell would likely parry this attack by reprising his criticism of the use of cross-sectional means, rather than comparing absolute values (in this case, the degree of modernization of the most modern area of the country) over time. Nonetheless, although this type of defense might offer O’Donnell’s theory some protection, it would seem to be missing the forest for the trees: in spite of whatever shortcomings GDP/cap may have, it is unlikely that the measure fails to capture the variable to such an extent that we can substantially discount the statistical evidence which Przeworski et al. provide.
Similarly, Londregan and Poole present evidence showing that coups themselves are a phenomenon essentially confined to poor countries – and, even more interestingly for O’Donnell’s theory - non-constitutional rule seems to have a substantial retarding effect on economic growth. This latter point is significant in so far as it seems to attack the very heart of O’Donnell’s assumption that bureaucratic-authoritarianism springs from attempts to encourage growth through foreign investment. Here O’Donnell could reasonably defend himself by noting that his theory was meant to explain only a particulartype of non-constitutional rule. Bureaucratic-authoritarianism, he writes, differs from other types of authoritarian rule (such as populist authoritarianism) in important ways – notably the policies of each system (in addition to the social problems to which it responds and the coalition on which it is based) [O'Donnell, 1973 #103: 108]. In some respects this criticism of Londregan and Poole’s study is valid - since they do not make any attempt to make this distinction – although in many ways the distinction is worrying in itself. If we are attempting to explain variation in political outcomes by using economic explanatory variables, then there would seem to be a substantial danger of endogeneity by using aspects such as economic policies to limit our sample.