Moreover, it is easy to see that ϕ** < ϕ* derived in Section 7.1. Taxes in democracy are higher when the median voter is a poor agent (i.e., δp ≥ ½ ) which corresponds to the case in which the middle class is small or when the median voter is a middle-class agent (i.e., δp < ½ ) but is relatively poor and likes higher taxes. Both of these cases make coups more attractive to the rich. Therefore, a relatively large and affluent middle class makes democracy less costly for the rich and acts as a buffer between the poor and the rich, making coups less likely and thus helping to consolidate democracy.
8. Conclusion
In this chapter, we introduced a third group into our analysis, the middle class. We focused the discussion on situations in which sociopolitical conflict was along the lines of socioeconomic class. We used this model to generate some interesting new insights consistent with some of the claims made in the political science and sociology literature on the importance of the middle class for democracy. Although we investigated various phenomena in this chapter, including how the introduction of the middle class allowed us to provide a simple model of how “splits in the elites” might work, there are two main results that we believe may be the most significant.
First, a strong and large middle class may aid democratization because it is less in favor of radical policies than the poor. Hence, if the rich are convinced that democracy is controlled by the interests of middle-class agents, they have less to fear from democracy and are less inclined to use repression to avoid it. This may occur either because the middle class grows numerically and, thus, the median voter becomes a middle-class agent or the median voter is already middle class and the middle class becomes richer (and, thus, prefers less redistribution). It is also interesting that other theories of the distribution of power in democracy, such as those discussed in the appendix to Chapter 4, emphasize that the political power of the middle class in democracy is often greater than its number would indicate. The most famous example of this is “Director’s law of income redistribution” (Stigler 1970), which claims that the preponderance of redistributive policies in democracy actually favor the middle class and not the poor. Although there are no definitive microfoundations for this claim and it is contentious empirically, it can be formalized in various ways. For instance, Persson and Tabellini (2000, pp. 57-8) show in a probabilistic voting model (see the appendix to Chapter 4, Section 2) how, if the middle class is less ideological than the poor and the rich, then its preferences are critical in determining the policies adopted in democracy. An alternative approach would be to assume in the context of a lobbying model (Grossman and Helpman 2001; appendix to Chapter 4, Section 3) that the middle class is better able to solve the collective-action problem than other groups.
Second and related a strong middle class may encourage democratic consolidation. The mechanism by which it does so is similar to that by which it helps to promote democratization. If the middle class is sufficiently influential in democratic politics, then democracy is not too costly for the rich; as a result, coups are less attractive. Hence, a strengthening of the middle class, in terms of an increase in either its size, political power, or relative income, may lead to the consolidation of a previously unconsolidated democracy. There is, of course, a natural caveat here: if the middle class becomes too rich, then it becomes indistinguishable from the rich and, therefore, will not be able to play the critical role of buffer between the rich and the poor.
Overall, the models in this chapter suggest that a third group, the middle class or the bourgeoisie, can play important roles in democratization. It can do so because it alters the nature of political conflict. Although scholars in the democratization literature have certainly suggested that the middle class can play an important role in democracy, they have not provided microfoundations for these claims. In this chapter, we showed that extending our framework to three groups provides natural microfoundations for the importance attributed to the middle class.
9
Economic Structure and Democracy
1. Introduction
So far, we have taken the determination of the level and distribution of income as exogenous. In this chapter, we endogenize the level and distribution of income. Instead of being directly endowed with income, people have various endowments of assets: land, labor, and physical and human capital. We introduce a technology, an aggregate production function, that determines how these factors of production can be combined to produce output. We also introduce key economic institutions, specifically property rights and competitive markets, that determine the rates of return on various assets.
Why would any of this matter? Intuitively, the structure of the economy or economic institutions could be important if they influence the trade-off between democracy and nondemocracy for the elites or the benefits of democracy as opposed to revolution for the citizens. There are many reasons why this might be so. First, the structure of the economy might influence the costs of revolution, repression, or coups. Second, the structure of the economy may also influence the nature of redistributive politics between different groups, something that our framework links to the creation and consolidation of democracy. We investigate both sets of ideas in this chapter. The models analyzed also allow us to consider some of the most salient claims in the political science and sociology literature about democracy. For instance, the claim that democracy can never be sustained in a primarily agrarian society, or at least one where the elites are large landowners, is common in the literature from Moore (1966), and Dahl (1971), to Rueschemeyer, Stephens, and Stephens (1992). Yet, the microfoundations of this claim are unclear. The models in this chapter help isolate mechanisms that may induce such a connection.
There seem to be a number of plausible reasons for why the costs of coups and repression might be related to the structure of the economy. Most important, repression and coups are costly because they disrupt economic life. Production in a modern capitalist economy requires input from many diverse firms, and much of this is not coordinated centrally or consciously but rather organized by the invisible hand of the market, as well as the visible hand of established firms. Moreover, most of these economic relationships are based on some type of implicit trust. At the simplest level, the employer knows that the workers will show up the following day and the workers know that when they show up, there will be work for them and they will be paid. More important, each firm trusts that its suppliers will provide it with the materials necessary for production and that customers and firms downstream are there to purchase those products. Even more important, there is an implicit trust in the quality of the goods and services provided. Employers believe that workers will not only show up but also exert appropriate amounts of effort, and suppliers will supply not just any odd materials but materials of sufficient quality to enable production. Finally, customers trust that they will be buying relatively high quality products, not things that would be unattractive for them to consume. Any sudden eruption of violence, any turmoil transforming the political system, any situation heightening the already existing conflicts in society also disrupts the economic structure, the relationships of trust, the cooperation that is the essence of capitalist production. In a related discussion, Dahl (1971) notes the
... enormous limitations, costs and inefficiencies of violence, coercion and compulsion in managing an advanced society where incentives and complex behavior are needed that cannot be manipulated by threats of violence. (p. 79)
There are also analogous ideas in the literature on military politics. For instance, Finer (1976) argues that military governments cannot run complex industrial societies because the costs would be too high. He notes
as an economy advances, as the division of labour becomes more and more extensive, as the secondary and then the tertiary services expand, and as the society requires the existence of a professional bureaucracy, of technicians ... so the army ceases to be able to rule by its own resources alone. (p. 17)
These ideas suggest why repression a
nd coups are costly. The reasoning emphasizes the breakdown of complex economic relations, which are important for capitalist production. Although the same relationships are present in agrarian production, they are clearly less important. Quality issues are less paramount when it comes to agricultural products than in manufacturing. In a less developed and less industrialized economy, there are fewer complex relationships of buyer and supplier networks and less reliance on investments in skills and relationship-specific capital. These considerations naturally suggest that repression and coups become more costly in economies where production techniques are capital-intensive, both physical and human, rather than land-intensive.
Of equal importance is that the structure of the economy may influence the form of political conflict and the redistributive implications of democracy for the elites. For instance, landowners may have more to lose in democracy than industrialists. Recall that our entire approach is based on the presumption that the citizens have the political power to set policy to favor themselves in democracy. One factor that may limit the ability of the citizens to get the policies they want in democracy is that the elites may have power out of proportion to their numbers (e.g., through lobbying or the control of parties, as analyzed in the appendix to Chapter 4). But, there is an equally important economic factor that limits what the citizens can do, which we referred to as the Laffer Curve in Chapter 4. It is easiest to discuss this in the context of income redistribution. If taxes are very high, this stifles economic activity and creates such deep economic distortions that there is not much output left. Therefore, democracy is naturally restrained in applying high taxes, trying to ensure that these taxes do not distort the allocation of resources too much and do not induce the elites to withdraw their assets from economic activity, thus reducing tax revenues. But, the extent to which these considerations apply to capital and to land differs significantly. A high tax rate on land at most encourages landowners to leave their lands empty, but there is not much more they can do. In contrast, physical capital is more elastically supplied or more mobile: tax capital at a higher rate and there will not be much accumulation ; capital holders will invest their money in nontaxable sectors or take it abroad where it will not be taxed (a possibility we discuss in detail in Chapter 10). Human capital is probably in the most elastic supply because it is useless unless people exert effort which they will not be prepared to do if tax rates becomes too high. This implies that democracy naturally applies higher taxes on landowners than physical or human capital owners in an effort to maximize redistribution without creating too many distortions. Similarly, democracy in many unequal societies at first turns to asset redistribution and, because land is much easier and probably less distortionary to redistribute than physical capital, land reform is a way of dealing with the most severe inequities. Human capital is, of course, impossible to redistribute. Again, landowners have more to lose from democracy than capitalists or industrialists.
These ideas imply that elites who are heavily invested in land are typically more willing to use force to preserve nondemocracy or ensure a transition back to nondemocracy than elites who are invested in physical or human capital. This may be because repression and coups are less costly in such a society and, therefore, the costs of opposing or undermining democracy are lower relative to the benefit from doing so (which is the avoidance of pro-citizen, anti-elite policies such as redistributive taxation). Or it may be that the benefits of avoiding democracy are greater for landowners because they expect their income to be taxed at a higher rate or even their assets to be redistributed in land reforms.
A final consideration may be that landowners are typically richer than industrialists or those with human capital, especially in relatively poor countries that are at the margin of becoming democratic or at the relevant threshold of economic and political development where they could be democratic but still have not consolidated their democracies. Therefore, in terms of our analysis in Chapter 8, landowners correspond to the rich and industrialists and people with high human capital correspond to the middle class. Landowners lose more from taxation because they are richer; hence, everything else being equal, they are more in favor of actions to prevent democracy. Nevertheless, in this chapter, we examine the implications of economic structure holding inequality constant so as to focus more clearly on the other mechanisms described previously.
We conclude this chapter by outlining how the ideas discussed herein might be useful in understanding the relationship between economic and political development.
2. Economic Structure and Income Distribution
We now introduce an explicit economic structure that enables us to endogenize income distribution and discuss the political implications of different factor endowments. We want this structure to include labor (as the source of income for the citizens), physical capital, and land. To simplify, we abstract from human capital in most of the analysis and return to its implications in Section 10. We consider a fully competitive economy with a unique final consumption good, produced via the aggregate production function:
where K is the capital stock, L is total amount of productive land, and N is the labor force. Y is aggregate output, that is the physical quantity that people have to consume. All of these factors are fully employed and we assume that the production function F exhibits constant returns to scale, so that when all three factors are doubled, total output is doubled. Constant returns to scale is important because it implies that all revenues from production are distributed as income to the factors of production, capital, land, and labor. Fully competitive markets imply that all factors of production will be paid their marginal products. Holding institutions constant, inequality results because these marginal products differ and there are different scarcities for different factors.
The simplest way to provide a microfoundation for the framework used so far is to assume that the aggregate production function takes the special Cobb — Douglas form:
(9.1)
where 0 < θ < 1 and σ > 0. As will become clear when we calculate the distribution of income in this model, the choice of 1 — θ as the power to which N is raised is chosen deliberately to relate this model to those used so far.
Two features are implicit in this function. First, there is a limited amount of substitution between labor and the other factors of production (more precisely, the elasticity of substitution between labor and the other factors is exactly equal to 1). Second, there is a much higher level of substitution between capital and land. Both of these assumptions are plausible. For instance, they imply that the share of labor in national income is constant when income grows as a result of capital accumulation, whereas the share of land falls and the share of capital rises. This is roughly consistent with empirical evidence.
Let us examine this in greater detail. First, we assume like before that there are 1 — δ citizens and now these agents correspond to wage earners. Hence, N = 1 — δ, and we can write the production function as:
Moreover, the remaining δ agents, who constitute the elites, do not own any labor, and each of them holds a fraction δ of total capital stock, K, and a fraction δ of total land stock, L.
We assume that the final good Y is the numeraire (i.e., its price is normalized to 1). Throughout, all other prices are therefore relative to the price of the final good. Exploiting the fact that in competitive markets all factors of production are paid their full marginal products, we have the following expressions for factor prices:
(9.2)
Here, w denotes the wage rate, r the return to capital, and ν the rental rate of land. These prices are all “real” or relative prices because they are measured in terms of the final good.
The shares of national income accruing to three factors are given as:
(9.3)
It is interesting that the share of national income accruing to labor is a constant equal to 1 — 0. This stems directly from the functional form of the Cobb-Douglas production function (9.1). For example, even if ca
pital accumulates and from (9.2) real wages increase, the share of labor in national income is nevertheless constant. At the same time, the share of capital in national income increases and that of land declines.
Now, total income is (K + σ L )θ (1 - δ)1 — θ and because total population is 1, this is also average income. Hence:
(9.4)
Exploiting the fact that citizens only have labor income, we can derive an expression for the income of a citizen, denoted yp:
(9.5)
which is the exact expression for yp used throughout the book.
Recall that, for now, we are assuming that all members of the elite are homogeneous and own both capital and land. Therefore, we have:
(9.6)
as the expression that gives the income of a member of the elite.
We assume that the parameters are such that average incomes are less than the incomes of the rich or, in other words, δ < θ, which is identical to the assumption made in the model in which incomes are exogenous.
3. Political Conflict
We now show how our previous analyses of policy determination in democracy can be adapted to this more complex economic model. As before, all individuals have utility functions that are linear in consumption and, because people consume all their income, they aim to maximize income. Again, we assume that there are two policy instruments: a tax rate proportional to income and a lump-sum transfer that all agents receive. As before, it is costly to redistribute income. Although we now have a model with a richer set of underlying institutions, we assume these to be exogenous in the analysis of this chapter, although in Section 9 we discuss how they could enter into the analysis.
Economic Origins of Dictatorship and Democracy Page 41