Is Democracy a Normal Good? Evidence from Democratic Movements.

• Author: Minier, Jenny A.

Jenny A. Minier [*]

A frequent assertion of economists and political scientists is that democracy is a normal good, or that higher incomes lead citizens to "desire" more democracy. This assertion, however, has been difficult to test directly. I introduce a data set of democratic movements, and use it to address the relationship between income and the demand for democracy. Logit analysis of the estimated probability that a democratic movement occurs in an authoritarian county suggests that this probability is increasing in income per capita up to a level of approximately $5000. Unlike previous results, this does not suggest that all countries will become democratic once they pass some income threshold.

1. Introduction

   Is democracy a normal good? In other words, do people desire higher levels of democracy (more participation in government) as their incomes increase? This paper uses the existence of a democratic movement to solve a missing variable problem in the current literature. Works such as Barro (1996), Huntington (1991), and Lipset (1959) provide evidence of a strong, positive correlation between levels of democracy and income; this is frequently interpreted as evidence that higher incomes lead people to "demand" more democracy.

   I address this issue in a more literal sense. Specifically, if our goal is to infer what causes people to "demand" democracy, we cannot ignore the persuasive evidence that citizens of countries like South Africa under apartheid, Burma, or China--countries with low levels of democracy that have experienced widespread movements for democracy--demand democracy. Instead of democracy levels, I use the existence of a democratic movement as evidence of countries' "demand" for democracy.

   In this paper, I define democratic movements as widespread, public demonstrations that are explicitly pro-democracy (i.e., participants' demands must include either "democracy" or "direct elections"), and in which participants express these views physically. [1] Some of the well-known democratic movements that are included are: the Chinese student demonstrators in Tiananmen Square in 1989; the work of the African National Congress and others in South Africa to end the nondemocratic system of apartheid; and the 1986 "People Power" demonstrations in the Philippines that forced Ferdinand Marcos to concede victory to Corazon Aquino.

   The main goal of the paper is to use a new data set on democratic movements to estimate the relationship between income and democracy, testing more directly than has previously been done whether democracy is a normal good. A secondary goal is to investigate the conditions under which democratic movements are most likely to occur in an authoritarian country.

2. Background

   Two of the most frequently cited sources of the factors giving rise to widespread desire for democracy are Lipset (1959) and Huntington (1991). Both argue that economic development generates increased demand for democracy. Lipset (1959) argues specifically that as incomes increase, citizens become more receptive to "democratic political tolerance norms," desiring more democracy.

   Lipset (1959) and Huntington (1991) agree on the role of education, identifying it as a facet of economic development that facilitates the desire for democracy. Lipset writes:

   Education presumably broadens men's outlooks, enables them to understand the need for norms of tolerance ... and increases their capacity to make rational electoral choices.... The higher one's education, the more likely one is to believe in democratic values and support democratic practices. (1959, p. 79)

   Lipset's definition of economic development encompasses industrialization and urbanization, in addition to wealth and education. [2] Huntington's definition is also defined fairly broadly, incorporating the volume of international trade and the development of a sizable middle class. He argues that involvement in the world economy exposed societies to the democratic ideas practiced in other parts of the world.

   Huntington argues that economic growth not only increases the speed at which a country attains a level of income 'suitable' for democracy, but also may act to destabilize the authoritarian regime. [3] Growth, "raises expectations, exacerbates inequalities, and creates stresses and strains in the social fabric that stimulate political mobilization and demands for political participation" (p. 69).

   The population of a country may also be relevant: Olson (1965), among others, argues that larger groups are less capable of coordinating the provision of a collective good (such as democracy).

   The primary goal of this paper is to address whether or not democracy is a "normal good"; that is, whether an increase in incomes leads citizens of nondemocratic countries to demand democracy. The question is framed as a macroeconomic one: As per capita income rises, is an authoritarian country more likely to experience a democratic movement?

3. Data on Democratic Movements

   To construct the data set of democratic movements, I consulted a wide range of historical and political sources. To be included in this study, a democratic movement must satisfy three criteria. First, it must be fairly large in scale, relative to the population of the country. (This excludes coups d'etat, whether successful or not, which, almost by definition, involve only small numbers of insurgents; that is not important here, since coup leaders rarely have democratic objectives in mind.)

   Second, the people must express their desire for democracy physically; voting against a military regime's referendum or boycotting one-party regime elections are excluded, unless those campaigns are accompanied by widespread protests, demonstrations, or strikes. This restriction is arbitrary to some extent, but it clarifies the identification of such movements: The people's desire for democracy is strong enough that a number of them are willing to endanger their physical safety to demonstrate for democracy.

   Finally, the movement must be explicitly pro-democracy. Operationally, this means that the leaders or the participants of the movement must call for democratic reforms (specifically, democracy, direct elections, or political liberalization). General anti-regime movements, or protests against economic conditions, are not included unless the participants' demands include increased democracy. (In fact, many democratic movements began as protests against economic conditions or strikes in one sector of the economy; they are included in this study once their demands include democratic reforms.) Overall, 32 democratic movements in 22 countries between 1960 and 1990 are identified. Appendix A provides brief descriptions of these events. [4]

4. Empirics

   Previous empirical studies have generally focused on either the effects of democracy on growth, or on the determinants of democracy; this paper is concerned only with the latter. [5] Lipset (1959) identifies the more democratic countries within regions and compares them with the less democratic countries, concluding that, on average, democracies have higher levels of per capita income, industrialization, urbanization, and education. Barro (1996, 1999) uses the Freedom House index of political rights (see Gastil 1990), which consists of a ranking of democracy levels on a seven-point scale since 1972; Barro finds that the level of democracy is correlated with GDP, education, and other measures of standards of living. This paper differs from these previous studies in that the dependent variable is not the level of democracy, but the occurrence of a large-scale democratic movement within an authoritarian country.

   Formally, the demand for democracy [X.sup.D] in country i can be expressed as

   [[X.sup.D].sub.i] = [[X.sup.D].sub.i] ([[P.sup.D].sub.i], [Y.sub.i], [H.sub.i], [D.sub.i], [Z.sub.i]) (1)

   where [[P.sup.D].sub.i] is the "price" of democracy in country i (the cost to a country of becoming or remaining democratic in the next period), [Y.sub.i] is income in country i, [H.sub.i] is the education level in country i, [D.sub.i] is the existing democracy level in country i, and [Z.sub.i] is a vector of other possible determinants of democracy identified in section 2 (such as openness to trade, recent economic growth, or the size of the middle class). As discussed in section 2, it is frequently asserted that the demand for democracy is increasing in both income and education. It is further expected that the demand for democracy is decreasing in its price. The overall relationship between the current level of democracy and the demand for democracy is unclear: Citizens of more democratic regimes have less to fear should they demand more democracy, but they also have arguably less to gain from further democratization, and more freedoms to lose in the event of repression.

   Previous studies such as Barro (1996, 1999) have regressed levels of democracy on variables like those included in Equation 1. Barro (1996) controls for lagged democracy levels, male and female primary school attainment, and infant mortality, and presents a coefficient estimate of 0.051 with a standard error of 0.022 on the log of initial GDP. [6]

   This positive and statistically significant correlation between income and democracy levels has been interpreted as evidence that democracy is a "normal" or "luxury" good, or that as incomes increase, people demand democracy more strongly. [7] Following his analysis, Barro (1996) concludes that

   improvements in the standard of living ... substantially raise the probability that political institutions will become more democratic over time. Hence, political freedom emerges as a sort of luxury good. (pp. 23-24)

   However, this interpretation is incomplete for two reasons. First, citizens of authoritarian countries who participate in democratic movements that are repressed surely "demand" democracy. But the dependent variable in these regressions is the observed level of...