|
Article Excerpt
Abstract We ask what redistributions of income and assets are feasible in a democracy, given the initial assets and their distribution. The question is motivated by the possibility that if redistribution is insufficient for the poor or excessive for the rich, they may turn against democracy. In turn, if no redistribution simultaneously satisfies the poor and the wealthy, democracy cannot be sustained. Hence, the corollary question concerns the conditions under which democracy is sustainable. We find that democracies survive in wealthy societies. Conditional on the initial income distribution and the capacity of the poor and the wealthy to overthrow democracy, each country has a threshold of capital stock above which democracy survives. This threshold is lower when the distribution of initial endowments is more equal and when the revolutionary prowess of these groups is lower. Yet in poor unequal countries there exist no redistribution scheme which would be accepted both by the poor and the wealthy. Hence, democracy cannot survive. As endowments increase, redistribution schemes that satisfy both the poor and the wealthy emerge. Moreover, as capital stock grows the wealthy tolerate more and the poor less redistribution, so that the set of feasible redistributions becomes larger. Since the median voter prefers one such scheme to the dictatorship of either group, democracy survives.
Keywords Redistribution * Democracy * Dictatorship
JEL Classification Numbers P16 * P48
1 Introduction
We ask what redistributions of income and assets are feasible in a democracy, given the initial assets, their distribution, and some features of the political environment. The question is motivated by the possibility that if the redistribution is insufficient for the poor or excessive for the wealthy, they may turn against democracy. Moreover, if no redistribution simultaneously satisfies the poor and the wealthy and if either group has any chance to establish its dictatorship, democracy cannot be sustained. Hence, the corollary question concerns the conditions under which democracy is sustainable.
In a simple model of production and accumulation, where agents are heterogeneous in their initial wealth, the median voter chooses a sequence of redistributive tax rates. Decisions to save are endogenous, which means that they depend on future tax rates and thus future growth rates. We assume that decisions about redistribution are made in elections and show that no majority coalition of poor and wealthy leaves both better off than the decision of the median voter (see Theorem 2). Moreover, given a linear redistribution scheme, the identity of the median voter does not change over time. Hence, the same median voter is decisive at each time with regard to the entire path of future redistribution. To be accepted, however, the decision of the median voter must leave the poor and the wealthy at least as well off as they expect to be if they sought to establish their respective dictatorships, where they would choose their best redistribution scheme unilaterally. Hence, if democracy is to survive, any redistribution must satisfy the constraints originating from the possibility of rebellion. While we do not solve for the optimal redistribution scheme of the median voter, we investigate these constraints.
We find that democracies survive in wealthy countries. Conditional on the initial income distribution and the capacity of the poor and the wealthy to overthrow democracy, each country has a threshold of capital stock above which democracy survives. This threshold is lower when the distribution of initial endowments is more equal and when the revolutionary prowess of these groups is lower. In the extreme, democracy survives at any income if its distribution is sufficiently egalitarian or if neither group can establish dictatorship. Yet in poor unequal countries there exists no redistribution scheme that would be accepted both by the poor and the wealthy. Hence, democracy cannot survive. As endowments increase, redistribution schemes that satisfy both the poor and the wealthy emerge. Moreover, as capital stock grows the wealthy tolerate more and the poor less redistribution, so that the set of feasible redistributions becomes larger. Since the median voter prefers one such scheme to the dictatorship of either group, the outcome of electoral competition is obeyed by everyone and democracy survives.
These results are driven by an assumption about preferences. The cost of dictatorship is the loss of freedom. We follow the argument of Sen (1991) that people suffer disutility when they are not free to live the lives of their choosing. Specifically, even if we allow that the losers in the conflict over dictatorship suffer more, we also allow that everyone may dislike dictatorship to some extent. This preference against dictatorship (or for democracy) is independent of income: as Dasgupta (1993, p.47) put it, the view that the poor do not care about freedoms associated with democracy "is a piece of insolence that only those who don't suffer from their lack seem to entertain" (see also Sen 1994). Yet since the marginal utility of income declines in income, while the dislike of dictatorship is independent of income, at a sufficiently high income the additional gain that would accrue from being able to dictate tax rates becomes too small to overcome the loss of freedom. (1) However, this straightforward intuition valid for a static model is a significant oversimplification. First, to show that when income is low, there exists no redistribution profile over time that can sustain democracy, we must rule out the possibility that some redistribution sequence may generate future growth and well-being sufficient to forestall revolt today. Second, to show that when incomes are high there always exists redistribution profiles that sustain democracy, we must demonstrate that democracy is sustainable not only today but at every moment along the growth path, which of course depends on the redistribution scheme that is implemented. The prospects of growth that depends on current as well as future redistributions complicate the analysis, and therefore the proofs are relegated to the appendix.
Explanations in terms of preferences are justifiably suspect. We are driven to it because of two facts discussed in the next section: democracy is more likely, indeed certain, to survive in wealthy countries, and no plausible rival hypothesis eliminates the role of income in sustaining democracy. Hence, income matters and income is not a proxy for something else. (2)
2 Per capita income and the survival of democracy
The probability that a democracy (3) would survive rises steeply in per capita income. Between 1950 and 1999, the probability that a democracy would die during any year in countries with per capita income under $ 1,000 (1985 PPP dollars) was 0.0845, so that one in twelve died. In countries with incomes between $ 1,001 and $ 3,000, this probability was 0.0362, for one in twenty-eight. Between $ 3,001 and $ 6,055, this probability was 0.0163, one in sixty-one. And no democracy ever fell in a country with per capita income higher than that of Argentina in 1975, $ 6,055. This is a startling fact, given that throughout history about seventy democracies collapsed in poorer countries, while thirty-seven democracies spent over 1,000 years in more developed countries and not one died.
Is income a proxy for something else? Following the classical book of Lipset (1960), an enormous literature sought to explain the observed prevalence of democracies in the developed countries and their paucity in the less developed ones. This literature did not distinguish the factors that lead to the emergence of democracy from those that cause it to survive once established, and these factors are different (Przeworski and Limongi 1997; Przeworski, Alvarez, Cheibub and Limongi 2000; Przeworski 2004). Yet several arguments offered in this literature apply to the role of factors other than per capita income in sustaining democracy.
In Table 1 we show probit regressions in which the dependent variable are deaths of democracies and the column headings specify the rival hypotheses. The conclusion is clear: while some of the rival factors do matter in the presence of income (4), none of them eliminates the crucial role of income in sustaining democracy.
The most obvious candidate for a rival explanation is education. We take the years of education of an average member of the labor force (from Bhalla...
|
