PREVIOUS LITERATURE

In this section, we survey the literature on the effect of democracy on redistribution and inequality. Our emphasis will be on the empirical literature, though we also discuss some of the theoretical ideas that have played an important role in this literature (several the­oretical contributions have already been discussed in the previous section).

21.3.1 Democracy, Taxes, and Redistribution

In the basic model of the policy effects of democracy proposed by Meltzer and Richard (1981), an expansion of democracy should lead to greater tax revenues and redistribution. We first consider the tax and spending part of this. While Gil et al. (2004) found no cor­relation between tax revenues and different components of government spending and democracy in a cross-sectional specification, as we discuss below, there are many studies which do find such results.

This is certainly true of the more historical studies, for example, Lindert (2004), Gradstein and Justman (1999a), and Acemoglu and Robinson (2000). Aidt et al. (2006) and Aidt and Jensen (2009b) examine the impact of democratization measured by the proportion of adults who could vote in a cross-national panel consisting of 12 Western European countries over the period 1830-1938, and in a sample of 10 West­ern countries over the period 1860-1938, respectively. The latter paper, for example, finds robust positive effects of suffrage on government expenditure as a percentage of GDP and also tax revenues as a percentage of GDP.

One would expect that democracy not only changes the total amount of tax revenues, but also what taxes were used for. For instance, one might expect democracies to move towards more progressive taxation. Aidt and Jensen (2009b) investigated the impact of suffrage on tax incidence. They found, somewhat paradoxically, that suffrage expansion led to lower direct taxes and higher indirect taxes.

Scheve and Stasavage (2010, 2012) also adopt a long-run approach using data from OECD countries and find no correlation between democracy and either tax progressivity or the rate of capital taxation. Instead, consistent with Tilly (1985) and Besley and Pearson (2011), they emphasize the importance of warfare, a topic to which we return later.

An important study by Lindert (1994) found an impact of democracy on various types ofsocial spending in a panel data consisting ofEuropean and North American countries as well as Japan, Australasia, Argentina, Brazil, and Mexico and spanning the period from 1880 to 1930. In his 2004 book, Lindert summarizes his findings as: “Conclusion #1: There was so little social spending of any kind before the twentieth century mainly because political voice was so restricted” (Lindert, 2004, p. 22).

A lot of research is consistent with this. Huber and Stephens (2012) build a panel data­set for Latin America between 1970 and 2007 and measure democracy by the cumulative years a country has been democratic since 1945 and estimate pooled OLS models without fixed effects. They find the history of democracy is significantly positively correlated with education spending, health spending and Social Security, and welfare spending. In a panel data of 14 Latin American countries for 1973-1997, Kaufman and Segura-Ubiergo (2001) show that democracy, as measured by the dichotomous measure introduced by Przeworski et al. (2000), is positively correlated with government expenditure on health and education but not with other components of spending. Brown and Hunter (1999) also focus on Latin America using a panel between 1980 and 1992.

Using a broader set of countries and a panel between 1960 and 1998, Persson and Tabellini (2003) also find some evidence that democracy, as measured by the Gastil index and the Polity score, has positive effects on government expenditure and government revenues as well as welfare and Social Security spending as percentages of GDP.

Though most studies tend to focus on a broad measure of democracy, an interesting literature has examined female enfranchisement more specifically. The main focus of this research has been on whether enfranchising women has an additional or differential impact on government taxation or spending. Lindert (1994) showed that female enfran­chisement had an independent effect on social spending and this finding has held up well (see Aidt and Dallal, 2008, for similar results for a later period). Lott and Kenny (1999) studied the expansion of women’s voting rights in the United States between 1870 and 1940 and found that it coincided with increases in per capita state revenues and expen­ditures. Miller (2008) also examined this process showing that female suffrage increased health spending and led to significant falls in infant mortality.

Of all the research on this topic, only the paper by Aidt andJensen (2013) provides an identification strategy to tackle the fact that democracy is endogenous. Building on the theoretical ideas in Acemoglu and Robinson (2000, 2006) and their previous work (Aidt andJensen, 2011), they argue that “revolutionary threat,” measured by revolutionary events in other countries, is a viable instrument for democracy in a panel of Western European countries between 1820 and 1913.

In this light, the paper by Gil et al. (2004) appears an outlier in finding no effects of democracy on tax revenues as a percentage of GDP and spending. Nevertheless, there are econometric problems with all of these papers. Specifically, there is little attention to identification problems and most studies that use panel data do not include country fixed effects, thus confounding the effect of democracy with country-specific factors poten­tially correlated with democracy and redistribution. Though the important study of Aidt and Jensen (2013) moves the literature a long way forward, their empirical model controls for many endogenous variables on the right side and does not deal with the pos­sibility that revolutionary events in other countries might capture other correlated effects impacting the outcomes of interest (see the discussion of this possibility in Acemoglu et al., 2013a).

21.3.2 DemocracyandInequality

There is an even larger reduced-form empirical literature on the relationship between democracy and inequality, most of it by sociologists and political scientists rather than economists. This has typically delivered ambiguous results. Early work, which consisted mostly of simple cross-national regressions of measures of inequality (usually the income Gini coefficient) on various measures of democracy, was surveyed by Sirowy and Inkeles (1990). They concluded “the existing evidence suggests that the level of political democ­racy as measured at one point in time tends not to be widely associated with lower levels of income inequality” (p. 151).

Much of this literature, however, also suffers from the econometric problems of the type discussed in the last subsection. Most importantly, there is the possibility that omit­ted factors are affecting both inequality and democracy, and that reverse causation from inequality to democracy may be present (e.g., Muller, 1988).

Muller (1988), using a larger dataset than the previous literature, found that there was a negative correlation between the number of years a country had been democratic and inequality, which he interpreted as evidence that democracy had to be in place for long enough for inequality to fall. Yet the robustness of his results were challenged by Weede (1989) (see the response by Muller, 1989). Others, such as Simpson (1990), Burkhart (1997), and Gradstein and Justman (1999b) claimed that there was a nonlinear reduced-form relationship between democracy and inequality with inequality being low at both low and high levels of democracy and higher for intermediate levels. The plethora of results is what led Sirowy and Inkeles to be skeptical, though they do suggest that there may be some evidence in favor of the relevance of the history of democracy for inequality (Muller’s original finding has been replicated in many subsequent studies, e.g., by Huber et al., 2006; Huber and Stephens, 2012, table 5.10). Nevertheless, there are good reasons for being skeptical about these findings, since the impact of the history of democracy is identified in models that do not include fixed effects, and obviously, it will capture the impact of these omitted fixed effects. More generally, this is just a spe­cial case of the difficulty of identifying duration dependence and unobserved heterogeneity—a difficulty that this literature neither tackles nor recognizes.

Three more recent studies used better data and exploited the time as well as the cross­sectional dimensions to investigate the impact of democracy on inequality. Rodrik (1999) showed that either the Freedom House of Polity III measure of democracy was positively correlated with average real wages in manufacturing and the share of wages in national income (in specifications that also control for productivity, GDP per capita and a price index). He illustrated this both in a cross section and in a panel of countries using country fixed effects.

Though this research has been dominated by studies that examine the average effect of democracy, Lee (2005) uses a panel data random effects model to argue that there are heterogeneous effects of democracy on inequality. The panel is unbalanced and covers 64 countries between 1970 and 1994. In particular, he argues that there is a significant interaction between the size of government as measured by tax revenues as a percentage of GDP and democracy. The paper finds that, although there is a significant positive cor­relation between democracy and inequality, the interaction between democracy and the size of government is significant and negative, suggesting that for large enough levels of government, democracy reduces inequality. Lee interprets this as measuring state strength (similarly to Cheibub, 1998 and Soifer, 2013).

21.3.3 Education and Democracy

The impact of democracy on education has also been examined both historically and using contemporary cross-national data and some of the results were noted in the last section. The work of Lindert (2004, chapter 5) is again central and, as with his work on social spending, Lindert presents evidence that the historical emergence of democracy is connected with educational expansion. A complementary historical study by Engerman and Sokoloff (2005, 2011) points out that within the Americas there is a close connection between the extent of democracy, measured by voting rights, the proportion of adults that voted and an effective secret ballot, and measures of education such as lit­eracy rates.

A great deal of econometric work supports this research using various measures of education. Baum and Lake (2001), for example, found that secondary-school gross enrollment rates also increased with democracy across the developing world, “particularly among regimes that have experienced large changes in democracy” (p. 613) (see also Baum and Lake, 2003). Brown and Hunter (2004), focusing on 17 Latin American countries between 1980 and 1997, find that the Polity index is positively cor­related with total educational expenditures per capita and also with the share of expen­ditures going into primary education. This finding mirrors the earlier one of Brown (1999) who finds that various dichotomous measures of democracy created from the Pol­ity dataset and the measure of Przeworski et al. (2000) were positively correlated with primary school enrollment. Huber and Stephens (2012) also find robust evidence in Latin America for a positive correlation between the history of democracy and educational spending (see also Avelino et al., 2005).

These issues have also been intensively studied in sub-Saharan Africa. Stasavage (2005a) examined the impact of democratization in the 1990s in Africa on education, using a measure of democracy similar to Przeworski et al. (2000), and presented evidence that democracy increases total educational spending as a percentage of GDP. He also found evidence of increases in spending on primary education as a percentage of GDP, though this was not robust to the use of country fixed effects. Stasavage (2005b) provides a case study of democratization and educational expansion in Uganda. More recent research by Harding and Stasavage (2013) reconfirms the impact of democ­racy on primary education, this time looking at primary enrollment, and shows that the likely channel runs through a greater probability that democratic governments will abol­ish primary school fees.

Gallego (2010) presents one of the few attempts to develop an identification strategy to examine the impact of democracy on education. There are many reasons why this is important. Most obviously, there is the issue of whether or not there is reverse causation from education to democracy. Though the results of Acemoglu et al. (2005) reduce this concern, the above papers deal with this at best by using lagged democracy as an explan­atory variable. Gallego follows Acemoglu et al. (2001, 2002) and uses their data on the historical settler mortality of Europeans and indigenous population density in 1500 as instruments for democracy and finds that democracy in 1900, measured by the Polity score, has a significant causal effect on primary school enrollment in 1900. Gallego rec­ognizes that the exclusion restriction of his instrument may be violated but provides a very careful discussion of the potential biases that this involves and how this works against the findings he focuses on, arguing that he estimates a lower bound on the effect of democracy on education.

Using a broad sample of over 100 countries between 1960 and 2000, Ansell (2010) uses panel data regressions with and without country fixed effects to examine the impact of democracy, measured by the Polity score, on various components of educational spending. He also instruments for democracy using lagged democracy and the levels of democracy in neighboring countries. He finds that democracy has a positive and sig­nificant effect on total educational spending as a percentage of GDP, and on educational spending as a percentage of the government budget. Using cross-national regressions he also finds a negative correlation between democracy and private educational spending as a percentage of GDP and also between democracy and primary school expenditure per student by the government. He argues, contrary to Stasavage, that democracy tilts edu­cational spending away from primary and toward secondary and tertiary education.

The likely reconciliation of all these results is that the type of education democracy produces depends on what forces democracy unleashes and who wields power in democ­racy. In Uganda, when President Museveni allowed democratization, he did so in a soci­ety lacking a large middle class who could dominate educational spending decisions. Hence as Stasavage showed, primary school enrollment increased. But in a large cross-national sample, the relationship may be dominated by dictatorships that spend more on primary schooling and democracies that focus on secondary schooling (see also Gradstein et al., 2004; Ansell, 2010, for relevant models).

This may also account for the results in recent work by Aghion et al. (2012), which uses a long but unbalanced panel of 137 countries between 1830 and 2001 and reports a negative correlation between the Polity score and primary school enrollment.

21.3.4 Democracy and Health Outcomes

There is also some other work on the impact of democracy on health outcomes. These are potentially related to inequality, because rapid improvements in health outcomes tend to come at the bottom of the distribution. Many studies, for example, find that democ­racy is positively correlated with life expectancy (see McGuire, 2010, for an overview and case study and econometric evidence). Besley and Kudamatsu (2006) show this in a panel data model for the post-war period but without using country fixed effects. Wigley and Akkoyunlu-Wigley (2011) in a complementary study have shown that life expectancy is positively correlated with the history of democracy of a country. Kudamatsu (2012) showed in the context of democratic transitions in Africa that health outcomes improved in countries that democratized compared to those that did not.

Blaydes and Kayser (2011) looked at the relationship between democracy and average calories per capita interpreted as a proxy for inequality, because calories con­sumed decline very quickly with income. Using a trichotomous measure of democ­racy based on the Polity IV dataset (where greater than 7 is a democracy, less than —7 is an autocracy, and everything in between a “hybrid regime”), they show in a panel data model with country fixed effects that democracy is positively correlated with average calorie consumption.

Gerring et al. (2012) find using panel data from 1960 to 2000 that, although the cur­rent level of democracy, as measured by the Polity score, is not robustly correlated with infant mortality, there is a strong negative correlation between the history of democracy and infant mortality—the more a country has experienced democracy in the past, the lower is infant mortality currently. Contrary to these findings, Ross (2006), using panel data from 1970 to 2000, the Polity score, the Przeworski et al. (2000) dichotomous mea­sure of democracy, and the history of democracy as independent variables, finds no robust correlation between any of them and infant and child mortality. A possible reconciliation of these findings is that, as mentioned above, the history of democracy is nothing but a proxy for the omitted fixed effects, and Ross obtains different results from Gerring et al. because he controlled for fixed effects. Another confounding factor is that this literature in general does not control for the dynamics of democracy and GDP per capita and the endogeneity of democratization (see Acemoglu et al., 2013).

21.3.5 The Intensive Margin

All the papers discussed so far use various national-level measures of democracy, usually based on well-known databases created by political scientists. An important complemen­tary direction is to investigate within-country variation exploiting other measures of “effective” enfranchisement.

In this context, particularly interesting is Fujiwara’s (2011) study of changes in the voting technology in Brazil in the 1990s. These, by making it much simpler and easier for illiterate people to vote, massively enfranchised the poor. Fujiwara estimates the effect of this change by exploiting differences in the way the policy was rolled out. He shows that the consequence of the reform was a change in government spending in a pro-poor direction, particularly with respect to health expenditures, and that infant mortality fell as a result. Baland and Robinson (2008, 2012) examine another related reform, the intro­duction of an effective secret ballot in Chile in 1958. Though they do not directly study any policy outcomes, they do show that the reform led to large increases in the vote share of left-wing parties, which, they argue, is consistent with this democratizing reform mov­ing the political equilibrium towards more pro-poor policies. They also find that land prices fall, which illustrates that the price of land capitalized the value of controlling workers’ votes under the open ballot.

Martinez-Bravo et al. (2012) study the effects of elections in China on redistribution and public good provision. They use variation in the introduction of village elections in China, controlling for village and year fixed effects as well as province-level trends. They find that village chairmen experience higher turnover and become more educated and less likely to be Communist Party members following the introduction of elections. They also find that taxes and public goods increase as a result of the elections. In particular, irrigation increases more in villages with more farmland, and public education increases in villages with more children. They also find that income inequality is reduced, and less land is leased to elite-controlled enterprises.

Naidu (2011) examined the impact of the disenfranchisement of blacks in the US South via poll taxes and literacy tests in the period after the end of Reconstruction. He finds that this reversal of democracy reduced the teacher-student ratio in black schools by 10—23%, with no significant effects on white teacher-student ratios. Also, consistent with Baland and Robinson’s results, disenfranchisement increased farm values.

Relatedly, using state-level data Husted and Kenny (1997) examine the impact of the abolition ofliteracy tests and poll taxes in the United States over the period 1950-1988 and find that this was associated with a significant increase in welfare expenditures but not other types of government expenditures. Using county-level data, Cascio and Washington (2012) find that expansion of voting rights in the South resulted in increased state transfers to previously disenfranchised counties. Besley et al. (2010), on the other hand, show that the abolition ofliteracy tests and poll taxes was associated with increased political competition in US states. Increased political competition between the Repub­licans and Democrats reduced government tax revenues relative to state income and increased infrastructure expenditure relative to other components of government expenditure.

21.4.