Economic Origins of Dictatorship and Democracy

by Daron Acemoglu

  Figures 3.19 and 3.20 show the likelihood of a democracy-to-nondemocracy transition and the likelihood of a nondemocracy-to-democracy transition They are constructed in a manner similar to Figures 3.9 and 3.10. The first figure comprises countries that were not democratic in 1970, 1975, 1980, 1985, 1990, and 1995 and separates them according to whether they had an economic crises in the preceding five years. An economic crisis is defined as an annual growth rate of GDP per capita less than — 5 percent in any one of the preceding five years. The figure then shows the fraction of nondemocracies with and without economic crisis that have transitioned to democracy. Figure 3.20 performs the same exercise for transitions from democracy to nondemocracy. Both figures show that economic crises make transitions more likely. Overall, we interpret this pattern as supportive of the notion that regime transitions are more likely during times of crisis or turbulence.

  5. Social Unrest and Democratization

  Our approach to democratization, in fact, stresses not only the role of crises but also the importance of social unrest, the threat of revolution, and generally the de facto power of those without de jure political power in inducing a transition to democracy. In this section, we return to the historical discussion of the emergence of democracy in nineteenth-century Europe and twentieth-century America to discuss this issue (see Therborn 1977; Rueschemeyer, Stephens, and Stephens 1992; and Collier 1999 for overviews).

  Figure 3.19. Transitions to Democracy and Crises 1970-1995.

  Figure 3.20. Transitions to Nondemocracy and Crises 1970-1995.

  5.1 Democratizations in Nineteenth-Century Europe

  In France, although democracy had flourished briefly after the revolution, it was quickly ended by the rise of the Jacobins and then Napoleon. After the fall of Napoleon, the absolutist monarchy was restored. Absolutism began to weaken after the 1830 revolution, which led to a highly restricted democratic regime in which property restrictions limited the electorate to about 0.75 percent of the population (Cole and Campbell 1989). The collapse of the Orleanist monarchy in the 1848 revolution led to the Second Republic, with the introduction of universal male suffrage in 1849 (Collier 1999, pp. 41-2). The effect of this was cut short, however, first by restrictions on voting rights introduced in 1850, disenfranchising 2.8 million men, and then by the coup of Louis Napoleon in 1851. Historians split this subsequent period into two phases: the “authoritarian” phase from 1852 to 1860 and the “liberal” phase from 1860 until the defeat of the French armies in the Franco-Prussian War of 1870. The defeat in the war led to further unrest (in particular, the Paris Commune) and to the collapse of the regime, making way for the Third Republic (Zeldin 1958; Plessis 1985; Price 1997). Finally, 1877, democracy with complete male suffrage was established, although other reforms, such as the secret ballot, were only introduced later in 1912 (Kreuzer 1996).

  The history of modern democracy in Germany starts with the 1848 revolution, when nearly all German states significantly increased popular participation in government, again in the face of revolutionary pressures (Blackbourn 1998, Chapter 3). The effects of this democratization were strongly mitigated by institutional restrictions, however. This regime featured a three-class voting system and was controlled initially by Junker landlords, followed in the 1870s by the coalition of “iron and rye”; the Parliament could not appoint ministers or discuss foreign policy, and voting was oral. Although after 1870 all adult males over the age of twenty-five had the right to vote, voting was controlled in rural areas by the landlords (Gosnell 1930; Goldstein 1983). As Abrams (1995, p. 10) stated during this period “the German Empire was, in theory, a constitutional monarchy, yet in practice it was governed by a Prussian oligarchy.” The final emergence of German democracy, the Weimar Republic in 1919, was in response to the severe threat of social disorder and revolution triggered by the collapse of the German armies on the Western Front in August 1918 (e.g., see the classic accounts in Gerschenkron 1943 and Mommsen 1981).

  In Sweden, democracy arrived via a series of gradual franchise extensions, starting in 1866 with the creation of a bicameral parliament with First and Second Chambers. Universal male suffrage was introduced in 1909 in the First Chamber; however, true parliamentary government arrived only in 1918, when the political power of the Conservative Party and the monarchy were limited - once again, an outcome of unusual turbulence spurred by the end of the First World War and the severe economic crisis (Verney 1957). Tilton (1974) argues that

  ... neither [of the first two reform acts] passed without strong popular pressure; in 1866 crowds thronged around the chamber while the final vote was taken, and the 1909 reform was stimulated by a broad suffrage movement [and] a demonstration strike. (p. 567)

  The reform in 1909 had been preceded by strikes and demonstrations and, even though Sweden was not a participant in the First World War, the revolution in Russia and the situation in Germany forced the concession of democratic rights. In 1917, the Liberals and Social Democrats formed a coalition government and proposed full male suffrage, which was defeated by the Conservative-dominated Second Chamber. Collier (1999) explains that

  ... it was only after the economic crisis of 1918 and ensuing worker protests for democracy led by the Social Democrats that the Reform Act was passed. Indeed, in November 1918, labor protests reached such a point as to be perceived as a revolutionary threat by Sweden’s Conservative party and upper classes. (p. 83)

  In all of these cases, the driving force behind political liberalization and the introduction of democratic measures is the threat of social disorder and, ultimately, revolution. Disorder was heightened by wars and other shocks to the social order.

  5.2 Why in the Nineteenth Century?

  Our approach so far explains the emergence of democracy but, in the European context, it does not answer the question of why the wave of democratizations started in the nineteenth century. The notion that democracy was a feasible set of political institutions goes back to ancient Greece and Rome and, at least by the seventeenth century in England, particularly during the Civil War, there were consistent demands for universal suffrage. A possible explanation for this emerges toward the end of the book, but even at this stage, it is worthwhile to see if the available evidence is consistent with the comparative statics we have already derived.

  Before the nineteenth century, the disenfranchised segments of society were scattered in rural areas; therefore, we may think of the threat of revolution as less severe because it was very difficult for them to organize. Therefore, the combination of increased urbanization and factory employment may have been a key factor in initiating the wave of democratization in nineteenth-century Europe. Changes in the structure of society and the economy during the early nineteenth century altered the balance of political power - in particular, making the exercise of de facto power by the politically disenfranchised much easier (Thompson 1963; Tilly 1995; and Tarrow 1998).

  Figure 3.21. Gini Coefficient (United Kingdom).

  It is also undoubtedly true that the ideological changes that occurred during the Enlightenment, the French Revolution, and the U.S. War of Independence had the effect of changing people’s ideas about the proper nature of government and the legitimacy of the old political order.

  In the context of our approach so far, another potential answer is that inequality was more limited before the nineteenth century. Recall that when inequality is low, revolution is not a threat and even if it is, the elite can prevent revolution by promises of redistribution. Only with a sufficiently high level of inequality does democratization become a necessity. The limited data that exist on nineteenth-century inequality are consistent with the notion that inequality was rising until democratization (and then it started declining because of the redistribution following democratization). Much of this literature focuses to trying to discover whether there was a “Kuznets curve” historically, following Kuznets’ (1955) conjecture that inequality first rises and then falls with economic development.

  Data on
income inequality for the nineteenth century are not extremely reliable. Figure 3.21 plots three different estimates of the historical evolution of the Gini coefficient in Britain. There is consensus among economic historians that income inequality in Britain fell sharply after the 1870s. There is also consensus that inequality rose in the century before this, although different scholars with different datasets found different timing for this rise. Lindert and Williamson (1982, 1983) found that this increase occurred before 1800 (see also Lindert and Williamson 1985; Lindert 1986, 2000a); Williamson (1985) found that it happened between 1800 and 1870. Other evidence (e.g., O’Rourke and Williamson 2002) is consistent with Williamson (1985). Whatever the case, the data on inequality are clearly consistent with the idea that inequality had risen in the century before 1867 and the Second Reform Act, and it may well have risen even before the First Reform Act. The evidence also suggests that inequality fell substantially after political reform.

  Data for other countries are even more scarce. Morrisson (2000) surveyed the existing evidence and argued that Germany, France, and Sweden all went through a Kuznets curve. In Germany, inequality rose during the nineteenth century; most researchers place the peak around 1900. For example, Kuznets (1963) found that the income share of the top 5 percent went from 28 percent in 1873-80 to 32 percent in 1891-1900, stayed at 32 percent during 1901-10, and declined to 31 percent in 1911-13. Dumke (1991) found the same income share to be 28.4 percent in 1880, rising to 32.6 percent in 1900, and falling to 30.6 percent in 1913. During the Weimar Republic, inequality fell rapidly. Kraus (1981) records that by 1926, the income share of the top 5 percent had fallen by 6.2 percent. Overall, Morrisson (2000) argues that the Kuznets curve in Germany peaked in 1900, went flat, and started to fall in the 1920s. This date corresponds closely to the major democratization of 1918-19. Bourguignon and Morrisson’s (2002) data show exactly this pattern.

  For France, Morrisson (2000) and Morrisson and Snyder (2000) argued that inequality rose until 1870, with the income share of the top 10 percent peaking at around 50 percent. Inequality started to fall, however, in the 1870s; in 1890, the income share of the top 10 percent was down to 45 percent, falling further to 36 percent by 1929. The major political reforms of 1860-77 in France are, therefore, approximately around the peak of the Kuznets curve. The conventional wisdom about France has, to some extent, been challenged in recent research by Piketty (2003) on the twentieth century and by Piketty, Postal-Vinay, and Rosenthal (2003) on the nineteenth century. Using data on taxation returns, these authors found that inequality rose monotonically in the nineteenth century and only fell during the First and Second World Wars in the twentieth century.

  Finally, Söderberg (1987, 1991) recorded that income inequality grew in Sweden, peaking just before the First World War, leveling off or falling slightly during the 1920s, and then falling rapidly thereafter. Once again, there is close correspondence between the decline in inequality and the extension of the franchise.

  Overall, therefore, in Britain, France, Germany, and Sweden, the peak of the Kuznets curve appears to have followed democratization, which is in line with the mechanism proposed in this book.

  5.3 The Latin American Experience

  The evidence from the Latin American experience with democratization also supports the notion that the threat of revolution and social unrest was important and is broadly consistent with the comparative statics with respect to inequality. In Chapter 1, we discussed the Argentine case in which social conflict was important in the passing of the Sáenz Peña Law. In Chapter 2, we noted the views of Bushnell (1993) on the democratization in Colombia in the 1930s. We now look briefly at some other cases in more detail.

  Historical studies of the movement toward democracy suggest an important role for social conflict. In Venezuela, a long period of caudillismo and political instability was ended by the dictatorship of Juan Vicente Gómez between 1908 and 1935. His military successors ruled until the first modern democracy was created in 1945. Levine (1973) describes the events leading up to democratization as follows:

  ... after several days of fighting, a provisional revolutionary government was formed, with four members from Acción Democrática, two military officers, and one independent civilian. The three years that followed marked the introduction of a party system into Venezuela, abruptly ushering in an experiment with mass political democracy. (p. 89)

  Democracy fell to a coup in 1948 but was reinstated in 1958 when the regime of General Pérez Jiménez collapsed in the midst of a widespread uprising. Levine (1989, p. 256) argues that redemocratization was in response to the unrest following economic depression and writes, “underground political forces, now united in a Junta Patriótica, mounted a wave of demonstrations and street fighting.”

  In Central America, the threat of social conflict and outright revolution has been a significant factor in inducing political elites to accede to democracy. For example, in Guatemala, General Jorge Ubico’s thirteen-year dictatorship ended in 1944 when he was replaced by a junta led by General Federico Ponce. He was deposed the same year by an upsurge of pro-democratic sentiment and a student revolt, leading to the election of Juan Arévalo as president in 1945. He was followed in 1950 by Jacobo Arbenz, who was ousted by the coup of 1954. Redemocratization in Guatemala followed the same pattern and was a direct response to the eruption of conflict. Starting in 1982, the military acceded to a gradual redemocratization: Marco Cerezo was elected in 1985, followed by Jorge Serrano in 1990. This process continued after Serrano’s attempted coup was foiled in 1993. Although these regimes were closely constrained by the military, the political liberalization was due to massive social unrest (Trudeau 1993).

  In El Salvador, the picture is similar except without the brief early period of democracy (Baloyra 1982; Paige 1997). Rule was ceded by the coffee oligarchs to the military after the matanza insurrection of 1932. After 1962, democratic elections began but were closely controlled by the military and were subject to massive fraud. After a brief military interlude, redemocratization occurred in 1982 but in the midst of an extensive civil war, which only ended in 1992; 1994 was the first election in which the main left-wing group, the Frente Farabundo Marti para la Liberación Nacional (FMLN), contested power.

  It is perhaps clearer that social conflict, often class and distributive conflict, has been behind most of the democratic collapses and coups in Latin America. This was a central theme of O’Donnell’s (1973) seminal book, and Stepan’s (1985) analysis of military coups in Chile, Uruguay, Argentina, and Brazil also echoes the same conclusion. He writes:

  The new authoritarianism in all four countries ... was installed in an atmosphere of growing class conflict. In each country the bourgeoisie provided the social base for the new authoritarian regime, whose first political acts were the use of the coercive apparatus of the state to dismantle ... working class organizations. (p. 318)

  Drake (1996) similarly argues in his analysis of the role of labor in the dictatorships of Argentina, Brazil, Uruguay, Spain, and Portugal that

  most of the dictatorships arose out of the distributive struggle between capital and wages ... Before the dawn of the dictatorships, working-class militance had begun to frighten property owners, who therefore abandoned liberal democracy. Losing profits, power, and legitimacy, the economic elites were rescued by the military ... These right-wing, military based governments defended capitalism from populism, socialism, or communism by suppressing demands from the lower classes. They favored the private over the public, the wealthy over workers, capital accumulation over redistribution, hierarchy over equity. (pp. 3-4)

  Another way of getting at the same issue is to look at how coups influence the value of different assets. For example, to the extent that democracy leads to redistribution and taxation of the assets of the rich (land and capital), we would expect the prices of these assets to fall with democracy and rise after a coup. Figure 3.22, constructed from data in Couyoumdjian, Millar, and Tocornal (1992), shows the re
al value of the stock-market index in Chile from 1928 to 1978. The real value of stocks declined continuously from the 1930s through to the coup of 1973, reaching its nadir with the election of Salvador Allende in 1970. The authors relate this secular decline to the increased intervention of the government in the economy, commenting that

  ... the 1930-1960 period was scarcely auspicious for stock-market operations. It began with a deep depression, which finished in 1932.... From then on began an unequivocal process of deterioration, which had to do with the increasing state intervention in the economy, which, directly or indirectly, constrained free enterprise. It was limited in its development by price controls, tax increases, high inflation and other measures of distrust ... Stock-market activity was not more than a reflection of the decreased participation of the private sector. (p. 309)

  Figure 3.22. Chilean Real Stock Market Index, 1928-1978.