¹ , Przeworski and , Limongi, “Modernization: Theories and Facts,” World Politics 49 (January 1997)Google Scholar.

² , Brown, “Review of Przeworski et al.'s Democracy and Development,” Comparative Political Studies 34 (June 2001), 576 Google Scholar.

³ , López, “Sanctions on Cuba Are Good, but Not Enough,” Orbis 44 (Summer 2000), 349 Google Scholar.

⁴ Przeworski, Adam, Alvarez, Michael E., Cheibub, José Antonio, and Limongi, Fernando, Democracy and Development: Political Institutions and Well-Being in the World, 1950–1990 (New York: Cambridge University Press, 2000)CrossRefGoogle Scholar.

⁵ Przeworski and Limongi (fn. 1); the version appearing in Przeworski et al. (fn. 4) drops the theoretical discussion altogether (chap. 2).

⁶ Przeworski and Limongi (fn. 1), 166. See also Adam Przeworski, “Why Democracies Survive in Affluent Countries” (Manuscript, Department of Politics, New York University, New York, 1996).

⁷ According to Przeworski and Limongi, the second mechanism that reduces the likelihood of a democratic breakdown in rich countries concerns the way in which capitai stock recovers from a war, since capital stock goes back to its steady state at a faster rate in a poor country than in a rich country. Although this reasoning fits squarely with the predictions of the classical Solowian growth model, which is based on a standard production function with diminishing returns, their conclusion that a faster catch-up rate should increase the value of being a dictator in a poor country is unwarranted. What should matter is not the rate at which the economy returns to its steady state but rather total national output, which determines the return (income flow) that the expropriator will obtain. Accordingly, the value of being a dictator is much higher in a wealthy country (ignoring any possible effect of a declining marginal utility of income).

⁸ Przeworski and Limongi (fn. 1), 167.

⁹ Przeworski et al. (fn. 4), 106. Summarizing the results of a multivariate dynamic probit analysis of transitions, Przeworski and his associates acknowledge “the impact of per capita income … is apparent for both regimes, but it is orders of magnitude larger for democracies” (p. 123). But, as the quote cited above suggests, they seem reluctant to embrace the finding that income growth causes democratization.

¹⁰ In this article we follow Przeworski and his coauthors' terminology in equating “development” with growth of per capita income. Yet our discussion below implies that other aspects of development, especially growing income equality, are probably more relevant dimensions of development for political change than are growing incomes. We also follow these authors in using the terms “dictatorship” and “authoritarianism” as synonyms and in treating both as equivalent to nondemocracy. Nondemocracy becomes the more accurate term as we shift our analysis back in time, when regimes at risk of democratization included monarchies and parliamentary regimes with limited franchises, neither of which would fit today's concept of dictatorship.

¹¹ Przeworski et al. (fn. 4), 93, table 2.3.

¹² All figures are international prices of 1985.

¹³ Przeworski et al. (fn. 4), 160.

¹⁴ In Przeworski et al. (fn. 4) the number of dictatorships rises to twenty.

¹⁵ Data are from Przeworski et al. (fn. 4), 93, table 2.3.

¹⁶ Their estimation appears in Przeworski et al. (fn. 4), 124, table 2.17.

¹⁷ In table 2.17 in Przeworski et al. (fn. 4), the authors report the coefficients in two columns: beta (the coefficient of transition to dictatorship) in the first one; and beta plus alpha (the latter being the coefficient of remaining authoritarian conditional on being authoritarian in the previous period). We have opted, instead, to report beta and alpha in separate columns.

¹⁸ Bairoch, Paul, Economics and World History: Myths and Paradoxes (Chicago: University of Chicago Press, 1993)Google Scholar, chap. 9; Maddison, Angus, Monitoring the World Economy, 1820–1992 (Paris: Organization for Economic Co-operation and Development, 1995)Google Scholar.

²² Running the same regression with different periods does not alter the main results of Table 4, model 3. The dummy for the interwar period has been dropped from model 3 to avoid collinearity.

²³ The income variables are defined as the corresponding per capita income above a given threshold and zero below. To choose the thresholds for this estimation, we have first examined the variation in our coefficient for different per capita segments through separate functions.

²⁴ Przeworski and Limongi (fn. 1), 163.

²⁵ Matthew Cleary, “Testing Endogenous and Exogenous Modernization Theory” (Paper presented at the annual meeting of the American Political Science Association, Atlanta, September 2–5,1999).

²⁶ As Przeworski and Limongi note, Chile broke through the $4,115 threshold twice but underwent a transition only on the second breakthrough.

²⁷ The list, with the year they achieved $4,115 first and the year of the transition second, is Brazil (1980, 1978), South Korea (1985, 1988), Greece (1970, 1974), Poland (1985, 1989), and Portugal (1973,1975). Poland also achieved the threshold income in 1974, without democratizing.

²⁸ Boix, Carles, Democracy and Redistribution (New York: Cambridge University Press, 2003)CrossRefGoogle Scholar.

²⁹ When we extend our analysis back to the mid-nineteenth century, other fixed assets suggest themselves. Yet what is critical is not just that an asset is fixed and that it plays a large role in a country's exports, but also that it accounts for a large portion of the country's GDP. Britain's coal industry in the nineteenth century, for example, would not fit these criteria.

³⁰ For evidence on the negative impact of oil on democratic transitions, see Boix (fn. 28), chap. 3.

³¹ We exclude all the Soviet-dominated countries (rather than employ a dummy variable for these cases) because the variable gauging its conditional effect on regime transition is perfectly collinear with the variable democracy and drops out of the estimation.

³² We do not pretend here to exhaust the list of omitted variables that might reduce the number of democratic transitions among a subset of wealthy nations. A third possibility would be that there are different types of dictatorships with differential probability rates of breakdown. For example, military governments may be less resilient than civilian juntas to development effects. We thank Adam Prze-worski for suggesting this possibility to us.

³³ For the literature on the effect on savings and investment, see Galenson, Walter, “Introduction,” in , Galenson, ed., Labor and Economic Development (New York: Wiley, 1959)Google Scholar; Schweinitz, Karl de Jr., “Industrialization, Labor Controls and Democracy,” Economic Development and Cultural Change 7 (July 1959)CrossRefGoogle Scholar; Huntington, Samuel P., Political Order in Changing Societies (New Haven: Yale University Press, 1968)Google Scholar. For the literature on insulated elites, see Haggard, Stephan, Pathways from the Periphery: The Politics of Growth in the Newly Industrializing Countries (Ithaca, N.Y.: Cornell University Press, 1990)Google Scholar.

³⁴ Przeworski, Adam and Limongi, Fernando, “Political Regimes and Economic Growth,” Journal of Economic Perspectives 7 (Summer 1993)CrossRefGoogle Scholar.

³⁵ Barro, Robert, Determinants ofEconomic Growth (Cambridge: MIT Press, 1997)Google Scholar.

³⁶ Przeworski et al. (fn. 4), 178.

³⁷ Although democracies may not affect growth too much, we do not deny that particular constitutional structures matter for growth, for example, having some form of liberal structure with an independent legislature. Indeed, there seems to be growing evidence that a constrained executive leads to higher levels of development. North, Douglass, Institutions, Institutional Change and Economic Performance (Cambridge: Cambridge University Press, 1990)CrossRefGoogle Scholar; DeLong, Bradford J. and Shleifer, Andrei, “Princes and Merchants: European City Growth before the Industrial Revolution,” Journal of Law and Economics 36 (October 1993)Google Scholar; Acemoglu, Daron, Johnson, Simon, and Robinson, James A., “The Colonial Origins of Comparative Development: An Empirical Investigation,” American Economic Review 91 (December 2001)CrossRefGoogle Scholar. It is difficult to claim, however, that the introduction of democracy, understood as competitive elections and universal suffrage, was at the root of the industrial takeoff of the West. A more plausible argument, and one that we partly pursue in the following section, is that once certain liberal institutions, sustained by a social and economic balance of power, led to growth, this generated particular conditions, such as growing income equality, which in turn opened the door to democratic constitutions.

³⁸ Boix (fn. 28). Recent data collected by Deininger and Squire on income inequality, consisting of 692 comparable observations (587 of them with Gini coefficients), show that, at low levels of economic development, the degree of inequality is highly variable across countries. For economies under a per capita income of $5,000, the mean Gini index is 42.5 with the values ranging from 20.9 to 66.9 and a standard deviation of 10.4. At higher levels of economic development, the occurrence of inequality diminishes. In economies with a per capita income of more than $10,000 (constant prices of 1985), the average Gini index is 34.2 with a standard deviation of 3.6. Deininger, Klaus and Squire, Lyn, “A New Data Set Measuring Income Inequality,” World Bank Economic Review 19 (September 1996)Google Scholar. Boix's fuller discussion also examines the effect of economic development on capital mobility and of capital mobility on democratization.

³⁹ The long-term trend toward income equality as economies develop suggests other plausible mechanisms linking development with democratization. For example, assume as a starting point a dictatorship in which poor people are excluded from participation. As the country develops, incomes become more equal. If the desire to participate grows among the poor and middle class as their incomes begin to catch up to the those of the wealthy (as Lipset long ago claimed it would) and if their organizational capacity also grows, then the costs of repression will rise as a function of economic development. In this case even if—contra Przeworski and Limongi—the marginal returns of capital remain stable, development would cause transitions to democracy.

⁴⁰ Boix (fn. 28).

⁴¹ The distribution of agricultural property is measured by the area of family farms as a percentage of the total area of holdings. This measure, gathered and reported by Vanhanen, defines family farms as those “that provide employment for not more than four people, including family members .. . that are cultivated by the holder family itself and ... that are owned by the cultivator family or held in ownerlike possession”; Vanhanen, Tatu, Prospects of Democracy: A Study of 172 Countries (London: Routledge, 1997), 48 Google Scholar. This definition, which aims at distinguishing family farms from large farms cultivated mainly by hired workers, is not dependent on the actual size of the farm—the size of the farm varies with the type of product and the agricultural technology used. The percentage of family farms captures the degree of concentration and therefore inequality in the ownership of land. The data set, reported in averages for each decade, covers the period from 1850 to 1979. It varies from countries with 0 percent of family farms to nations where 94 percent of the agricultural land is owned in family farms; the mean of the sample is 30 percent with a standard deviation of 23 percent. An extensive literature has related the unequal distribution of land to an unbalanced distribution of income.

⁴² To measure the level of human capital, Boix (fn. 28) relies on Vanhanen's “index of knowledge distribution,” which consists in the arithmetic mean of the percentage of literates in the adult population and the “level of students.”The level of students is the number of students per 100,000 inhabitants, normalized so that 1,000 students per 100,000 inhabitants corresponds to a level of 100 percent. The Vanhanen index of education, which also covers the period 1850–1979, varies from 0.5 to 99 percent with a mean of 29.2 and a standard deviation of 22.7.

⁴³ The Gini index is taken from Deininger and Squire (fn. 38).

⁴⁴ The introduction of the index of occupational diversification (without the variables of family farms and education) only reduces the statistical significance of the beta coefficient of per capita income. This estimation is not displayed in Table 7.

⁴⁵ Data on income inequality are too scarce before 1950 to test this hypothesis. However, if we look at its proxies, such as education and the distribution of agrarian property, inequality seems to have been less acute in advanced countries than in developing countries for similar levels of per capita income. For example, taking our country-year observations with per capita income lower than $2,000, the mean literacy index is around 40 percent of the population in developed countries and 20 percent in developing nations. For observations with a per capita income between $2,000 and $4,000, the average literacy index is 50 percent in advanced countries and 41 percent in developing countries. The average percentage of family farms is 39 percent and 28 percent in developed and developing countries for per capita incomes lower than $2,000. The difference gets larger for per capita incomes between $2,000 and $4,000: the average percentage of family farms is 42 percent in advanced countries and 25 percent in the rest. The universe of advanced countries includes North America, Western Europe, Japan, Australia, and New Zealand.