From James Monroe and the Quing Dynasty to George W

How democracy and dictatorship affect economic growth: Evidence from James Monroe and the Quing Dynasty to George W. Bush and the Communist Party.

Abstract

This article provides solid evidence for the hypothesis that democracy is generally more conducive to economic growth than dictatorship. At present, there is no consensus among academics and policy makers on whether and how democracy affects economic growth. The article reviews and evaluates some of the most important theoretical arguments on why democracy and dictatorship might affect economic growth. Thereafter, the article presents themost extensive statistical study conducted on the effect from democracy on economic growth, with data from 1820 to 2003 on a global sample. Different panel data methods, 2SLS andmatching techniques are used to study the relationship. The estimated effect from democracy on growth is positive, significant and relatively robust. Barro’s (1991, 1997) claim that semi-democracies perform better than democracies is also tested, and rejected.

The question and the answers

Is democracyin general better at generating economic prosperity thandictatorship?[1] Can people have political freedom and a growing economy at the same time? Some argue that this best of all possible worlds-scenario is unrealistic. The “Lee-thesis” (Sen, 1999:15), credited to former Singaporean PM Lee Kuan Yew, postulates that particularly in developing countries, a strong authoritarian regime is necessary to push through economic development. The East Asian Tiger-states, Pinochet’s Chile and present-day Chinaare viewed as decisive empirical evidence for this assertion. Democracies, with their perceived political conflicts,gridlocks and electorally induced conspicuous consumption, are claimed unable to generate economic development. However, several analysts have to the contrary argued that democracy is beneficial for economic growth. There is a “democracy advantage” (Halperin et al., 2005). Democracies for example have institutions that provide vertical and horizontal accountability so that predatory rulers cannot extract resources in their own self-interest, and the open flow of information in democraciesimproves abilities for correcting policy mistakes and for diffusion of technology. However, rigorous statistical evidence for a “democracy advantage” in economic growth is still needed.

Although there are adherents to both the Lee-thesis and the democracy advantage view, the “skeptics” are the most prominent among academics studying the economic consequences of democracy. Two seminal contributions are Przeworski and Limongi (1993) and Przeworski et al. (2000). The first surveys earlier statistical studies, and finds that studies claiming positive, negative or no effect from democracy on economic growth are about equal in numbers. The second conducts a thorough statistical study of its own on the topic, and the general bottom-line is that there is no significant effect from democracy on economic growth.These results are accepted by many political scientists. Larry Diamond (2008:96), for example,asserts that the “evidence is murky”when it comes to the claim that democracies produce more rapid economic development, while George Tsebelis reports it as a surprising fact that there is no evidence of democracies producing superior economic outcomes (2002:70). However, this article does provide empirical evidence for the hypothesis that democracies on average produce higher economic growth rates than dictatorships. When analyzing the most comprehensive data-set available, incorporating data from 1820 - when James Monroe was the President of the US and the Quing-dynasty ruled China -to 2003, a clear and significant effect emerges. Not only are the adherents of the Lee-thesis wrong, the “skeptics” should also reevaluate their beliefs. According to the analysis below, democracy is good for economic growth.

Theoretical arguments

In their seminal study, Przeworski and Limongi (1993) evaluated four theoretical arguments on the relationship between political regime types and economic growth. These four arguments are only a subset of the existing arguments in the literature on democracy, dictatorship and effects on growth, but they are among the most important. The arguments highlight how regime type might matter for I) property rights, II) capital investment, III) autonomy of the state and IV) checks on predatory rulers. I will take a fresh look at these arguments, and include some new theoretical insight as well as relevant empirical findings. I also add a fifth argument on democracy and technological change. As Przeworski and Limongi (1993), I score the arguments after whether they indicate that democracy increases (for democracy) or decreases (against democracy) economic growth relative to dictatorship. However, my scorecard differs from Przeworski and Limongi’s.

I) Democracy and the protection of property rights – For democracy

Consider a hypothetical country where the median citizen’s property entitlement, pm, is below the average property entitlement, pa. Property can only be redistributed progressively. There are aggregate economic costs related to redistributing property, r, given by the function c(r), and these costs can be interpreted for example as economic loss due to distortion or transaction costs. The costs are equal for all citizens, and c(r) is expressed in per capita terms, with c’(r) and c’’(r)>0. Under democracy and the assumption of one-dimensional politics, the median voter’s preferred outcome would be to redistribute property until the marginal personal gain of redistribution is equal to c’(r). If we now compare the democracy to a right-wing authoritarian regime, where the median member in the regime’s group of backers, s, has a property entitlement ps≥paverage, there will be no property redistribution since c’(r) is strictly increasing. If income y is given by y = y0 -c(r), then dy/dr<0. Income will be lower under democracy in this stylistic economy, and the argument could easily be remodeled slightly to show that economic growth would also suffer from redistribution within democracy. The argument that democracy will lead to extensive redistribution to the poor, with subsequent negative effects for aggregate production is old, and was shared by for example John Stuart Mill, David Ricardo and Karl Marx. Przeworski and Limongi (1993) asses the debate on the economic consequences of democracy from the nineteenth century, and claim that the right to vote and freedom of organization were widely perceived to have adverse effects on private property rights protection, and further economic growth.

If we relax the assumption that property can only be redistributed progressively, there are strong counterarguments to the claim that democracy weakens property rights protection.Democracy is associated with power distribution, both horizontal and vertical, and these features provide checks against violations of property rights. In principle, any form of government implies concentration of coercive power and therefore the possibility of state-led confiscation of property. Apolitical position within a state apparatus “provides the opportunity for individuals with superior coercive power to enforce the rules to their advantage, regardless of their effects on efficiency. That is, rules will be devised and enforced on behalf of the interests of the politically advantaged” (North, 2000:50). However, democracies will have certain advantages over dictatorships. First, in democracies the politically advantaged will constitute a larger segment of the population. In Olson’s model (2003), a larger group will internalize more of the indirect negative incentiveeffects of property rights violation on the overall economy, even if they gain directly from redistributive activity: “[T]hough both the majority and the autocrat have an encompassing interest in the society because they control tax collections, the majority in addition earns a significant share of the market income of the society, and this gives it a more encompassing interest in the productivity of the society. The majority’s interest in its market earnings induces it to redistribute less to itself than an autocrat redistributes to himself” (Olson, 2003:122).Second, there is a larger degree of power dispersion in democracies, also between different state institutions, which makes the room for single actors enforcing their will at the cost of others more unlikely.Because of lack of protection of individual rights, poor political accountability and concentration of power, property rights will be less protected in dictatorships than in democracies, since these features will allow dictatorial elites to confiscate property.

Przeworksi and Limongi recognize the multiplicity of arguments on the matter and the different empirical implications. Their overall assessment is therefore that “[W]hile everyone seems to agree that secure property rights foster growth, it is controversial whether democracies or dictatorships better secure these rights” (Przeworski and Limongi, 1993:51), and they further conclude that “[T]he idea that democracy protects property rights is a recent invention, and we think a far-fetched one” (Przeworski and Limongi, 1993:52).

I disagree with Przeworski and Limongi in their conclusion. The fact that different arguments have different implications does not mean that the arguments are equally valid. I will not go into a theoretical evaluation here, but I have done so elsewhere (Knutsen, 2007). The most compelling argument for refuting Przeworski and Limongi’s conclusion is in any case empirical. Chile, Singapore and South Korea are the only recent empirical experiences of authoritarian countries that were not hostile to property rights (Przeworski et al., 2000:211). Boix finds that democracies empiricallyhave a lower degree of expropriation risk than dictatorships (Boix, 2003). In Knutsen (2007), I used different proxies for property rights protection, and generally found that democracy significantly increased the protection of property rights, even when controlling for an extensive set of variables. The effect was robust over different specifications when using OLS with panel corrected standard errors, Random Effects and Fixed Effects. The effect was not equally robust when using 2SLS, trying to incorporate the plausible possibility of endogeneity of regime type, but several specifications yielded a significant and positive effect. The conclusion is that democracies most likely protect property rights better “on average” than dictatorships.

II) Dictatorships and investments – Either way, but for democracy when human capital included

The highest average investment/GDP ratio in the period 1970-2000 belonged to Singapore. An estimated 45% of GDP went to capital investment over this period (Knutsen, 2006:411). The investment ratio for the Soviet Union in the 1930’s was equally high, and today China is often racking up annual ratios over 40%. Could these high investment rates have been viable under democratic regimes? Probably not! Dictatorships have several policy means that allow them to drive up investment rates, thereby increasing medium to long run growth rates (Solow, 1956, Romer, 1990). First, dictatorships often suppress freedom of association, thus crippling independent organization of unions. In the absence of unions, it is plausible to assume that wages are lower, and that capital ownerstake a larger share of the pie. When combined with the assumption that savings rates increase with income, this yields the prediction that aggregate savings and thereby probably investment rates will be higher in dictatorships. Political accountability is also lower under dictatorship, among others because of the lack of free and fair elections. This reduces the pressure on political rulers to channel resources to immediate public consumption over the national budget. Instead, dictatorscan channel resources to investments, independent of the desires of “short-sighted electorates”. A similar political logic underlies the argument that dictatorial governments need not provide as much social security to the population in general. The response of rational citizens living under dictatorial rule is to save privately in order to self-insure for the future.

The argumentthat dictatorships are better able to generate higher savings and investment rates is therefore founded on solid theoretical reasoning.However, political mechanisms that allow dictatorships to have a higher investment rate, should the regime want so, does not imply that most dictatorial governments have incentives to generate high investment rates.The empirical evidence often cited in favor of the argument above seems prone to selection bias. The relatively few historical dictatorships with extremely high savings rates are used as examples, and general inferences on high savings rates under dictatorship are thereafter drawn, giving rise to a systematic bias stemming from selecting cases on a specific value (high savings and investment rates) on their dependent variable (Geddes, 2005 and King et al., 1994). A quick growth accounting exercise on a global sample in the 1990’s showed that the estimated average effect from dictatorship on capital stock-growth was negative, albeit insignificant(Knutsen, 2006:348-350). Most dictatorships do not generate very high savings and investment rates, and there are several reasons for this. First of all, self-interested dictators might not see it in their interest to follow investment-induced growth, as will become clear from the argument below on predatory dictators. A second related point is thatinvestment, and particularly foreign direct investment, is sensitive to the protection of property rights, where democracies were argued to have an advantage. Third, a high degree of corruption deters investment, and dictatorship probably increases corruption (Hegre and Fjelde, 2008, Knutsen, 2008a). Therefore, even if it is theoretically plausible that certain dictatorshipsmight generate extremely high investment rates due to the large scope of possible policies under limited political accountability, most empirical dictatorships do not produce high investment rates.

Mankiw et al (1991) modified the traditional Solow-model of economic growth. They substituted the traditional Solow-model, which claims that income Y is a function of technology, A, labor, L, and physical capital, K, with a model where human capital, H, also enters. In more mathematical terms they go from Y = F(AL,K) to Y=F(AL,K,H). If we stretch the capital concept to include human capital, democracies have an extra advantage over dictatorships, since it is hypothesized that democracies invest more in schooling andhealth.Mankiw et al. (1991:417-8) estimated that human capital is at least equally important as an input to the economy as traditional physical capital. Even if Mankiw et al.’s estimation procedures have been criticized (Klenow and Rodriguez, 1997), human capital is widely agreed among economists to be an important type of investment.

Education and health are highly valued by most people. This means that one would expect more and better education and health care for most citizens in democracies, which are assumed more responsive to citizens’ preferences than dictatorships.According to Acemoglu and Robinson (2006:64), Lindert (2000) finds a strong and positive effect from democratization on educational expansion in Western Europe. Engerman et al.(1998) find the same effect in Latin America. Stasavage (2005) finds that democracy has a positive effect on primary education spending in Africa.More generally, Baum and Lake (2003) find a positive indirect effect from democracy on growth via human capital. Universal schooling and health care provision are likely to be redistributive. Some economists have appreciated this point and produced theoretical models where egalitarian pressures in democracies affect economic growth through increasing public education, which is a kind of redistribution (Saint-Paul and Verdier, 1993). This argument is generalized by Acemoglu and Robinson (2006:219-220), who claim that democracy might force the provision of public goods towards a more socially efficient level than in dictatorships.

In conclusion, it is not obvious that dictatorship on average increases investment in physical capital. However, it seems likely that democracy increases the accumulation of human capital. If we apply a broad definition of capital, one might therefore argue that democracy on average increases investment, thus contributing to increased growth rates.

III) Dictatorships and autonomy of the state – Against democracy

Scholars studying East Asia have, as Przeworski and Limongi (1993) noted, often linked the fantastic economic performances of some Asian dictatorships to the autonomy of the dictatorial state. “In this view, the key to the superior economic performance of the Asian "tigers" is "state autonomy," defined as a combination of the "capacity"of the state to pursue developmentalist policies with its "insulation" from particularistic pressures, particularly those originating from large firms or unions. This argument takes two steps: "state autonomy" favors growth, and "state autonomy" is possible only under authoritarianism” (Przeworski and Limongi, 1993:56). Olson (1982) argues that democracies are prone to capture from special interest groups. This might lead to policies that areincoherent with the interests of the broader populace. Economic growth might be sacrificed for the protection of specific business sectors or pivotal voting blocs whose interest is not aligned with economic growth. In any case, the lobbying process will be associated with wasteful rent-seeking, which will detract financial resources and focus from more productive ventures. The argument is that politicians and bureaucrats are insulated from such pressures under authoritarianism and are therefore better able to conduct “proper” policies. One important special case is that of economic reforms. Certain microeconomic reforms for example, which allow prices to adjust to a level where they better reflect marginal costs and utilities, improve the efficiency of resource allocation. However, an adjustment process towards an efficient equilibrium might be painful, andcertain earlier “privileged” groups might lose out. Under democracy, the potential losers might be important political groups, so-called “veto players” (Tsebelis, 2002), and might act to block reform. Trade liberalization might be a particularly fitting example, where protected industries might block liberalization, even if the expected result is an increase in national GDP. Even if a reform might carry a positive expected value with it for a group, group members might be risk averse and prefer the status quo even if their best guess is that they will gain from reform. Under dictatorship, the dictator is assumed to have the means, and in some instances also the incentives, to carry out “painful” reform. Reform is also assumed to be conducted more speedily under a dictatorial regime, since many of the procedural steps needed in democracy as well as complex and time-consuming negotiation can be skipped.