Está en la página 1de 17

Corruption, Democracy, and Economic Growth

Author(s): A. Cooper Drury, Jonathan Krieckhaus, Michael Lusztig
Source: International Political Science Review / Revue internationale de science politique, Vol.
27, No. 2 (Apr., 2006), pp. 121-136
Published by: Sage Publications, Ltd.
Stable URL:
Accessed: 02/03/2010 05:40
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact

Sage Publications, Ltd. is collaborating with JSTOR to digitize, preserve and extend access to International
Political Science Review / Revue internationale de science politique.

Is P IPSRI Corruption. 2002).JONATHANKRIECKHAUS. and Economic Growth ANDMICHAELLuSZTIG A. mitigate Although corruption certainly occurs in democracies. while non-democracies suffer significant economic harm from corruption.Scholars have long suspected that political processes such as are important factors in determining and corruption democracy economic growth. and political educational expenditure. although democracy indirect influences on growth. a number relationship and economic between democracy appears to have growth. CA and New Delhi) . stability (Baum and Lake. Studies show.' While is more intuitively appealing.1177/0192512106061423 SAGE Publications (London. we show that corruption has no significant effect on economic growth in democracies. 121-136 . explanation facilitates wealth by stimulating namely that democracy economic growth. COOPERDRURY. of studies find no direct. Kurzman of the matter. Vol 27. statistically significant Indeed. ABSTRACT. Since the end of World War II. This does not put an end to the 2003. while corruption is generally accepted by scholars as having a direct and negative impact on economic perfor mance. however. Democracy. of course. No.. due to its positive effect on such things as important life expectancy. Keywords: * Corruption * Economic * Democracy economy * States growth * Political It is no great insight to proclaim that liberal democracies tend to be wealthier than non-democracies. reality suggests the relationship complicated. democracy. 2. ( 2006 International Political Science Association Thousand Oaks. a great deal of scholarly effort has gone into exploring the relationship between economic growth and liberal with many pursuing an obvious for their association. the electoral in corrupt acts that mechanism inhibits politicians from engaging and thereby jeopardize their damage overall economic performance political survival. 1994. et al. Helliwell.International Political Science Review (2006). We argue that one of democracy's indirect benefits is its ability to the detrimental effect of corruption on economic growth. that democracy has only indirect effects on growth. It simply suggests that greater understanding is needed the most robust system of government apparently symbiotic role played between DOI: 10. Using time-series cross-section data for more than 100 countries from 1982-97.

primarily because consumers of graft may be innumerable. the understanding We attempt to enhance of the indirect effects that democ racy has on economic our focus is on just one of these indirect growth. allows citizens to evict politicians mechanism that engage in particularly damaging in other words. 1992) and the economic growth and efficiency that appears to sustain it. and not exclusive.122 International Political Science Review 27(2) ever developed (Fukuyama. In this article. Nye. 1999: 3-4. activity. which levels. rather intuitively. the legislative branch." whether We define corruption or in terms of status. Our expectation (discussed below) will mitigate since the electoral democracy the negative effects of corruption. Lambsdorff. by examining democracy's indirect effects on economic is that growth. nepotism. 1993: 599). a foreigner must bribe every agency involved in foreign investment. where to spur investment. and while this disagreement is somewhat intuitive. Corrupt activity includes bribery. including the foreign investment office. albeit at differing is present corruption. some argue that corruption has beneficial effects for an economy. Democracy. although to economic as both a tax on productivity is that it is damaging and a performance market distortion. the relationship with economic growth. 1967: 419. . Post-communist Instead. pecuniary or to those closely associated with such an individual or group. it is one is substantively that. rapacious Russia illustrates the point nicely: To invest in a Russian company. and other misappropriation (see Bardhan. explain how democracy on Economic Growth on economic growth and then The Ill Effects of Corruption "as the abuse of public office for private gain. As the literature review below suggests. impact on economic performance Our unique contribution. as is clear from the discussion below. is to explore further these relationships however. negative in non-democracies. may exhibit no direct statistical forms of corruption. but it clearly serves to militate against negative economic effects of corruption. We are hardly the first to delve into the role that corruption plays with respect to economic growth. Mauro that corruption (1995) finds empirically reduces private sector invest ment even in countries cumbersome economic featuring regulations. the finance ministry. The predominant. Shleifer 1997: 1321. Although effects. the central bank. be expected Shleifer and Vishny corruption might (1993) is more suggest that one reason for this is that corruption than simply a tax on economic there is no central mechanism for collection. The Effects of Corruption and Democracy of corruption's We now turn to a discussion effect ameliorates this effect. We concentrate on political in all regimes. we use time-series cross-section data from 100 countries over a 16 year period and find. of public resources theft. some of our findings are unexpected and shed new light on and economic the connection between democracy performance. view of corruption Vishny. that corruption has a significant. the executive branch of the local government. The gain may accrue to an individual or a group. We disagree. the relevant industrial ministry. important and exists worldwide to varying degrees. the state property bureau.

Democracy. the result of which incentive for all officials taxation on productivity. 1993: is that foreigners result 615. may promote initiatives not to satisfy social need. and unethical its effects need not be economically deplorable Leff (1968) argues that where government sets for itself the task of detrimental. Thus. Krueger. or firm-specific favors. This second because the position is to get into the corruption game. a crude market for supply may be assumed firms most able to favors emerges. the literature effects of corruption yields two traditional and accepted position is that corruption has positions. but perhaps better than a . That is.Corruption. honest bureaucracy. 1968). The second view is ciently high transaction costs to limit significantly in states that are serves to create an economic that corruption equilibrium the weakest firms from the marketplace and excessively bureaucratic. corruption efficient than the alternative. to be low and demand high). one to receive bribes also does mere to the market for exists. overcentralized. there is an underappreciated supply-side One manifestation is that policymakers rent-seeking. provides and public spending. dishonest one with a rigid. few see taxes as spurs to economic ismore growth. 1982. the construction of a market costs in that it dissipates resources high opportunity that could otherwise be used on productive activity (Bhagwati. with the richest (and perhaps most efficient) in this firms must become more efficient to compete outbid their rivals. The first. it dissipates investment. The (Shleifer obvious and Vishny. As such. their point is that. Rather. The success of these base of taxation a broader firms. excessive Further. as the literature on rent-seeking and directly unproductive activity for political influence and favors generates suggests. Weaker sector or exit the productive black market. that while The other view of corruption suggests corruption itself may be to moralists. moreover. and generates suffi worse. but (public works projects are an excellent example) for bribes. graft provides an alternative for private sector interests otherwise to influence not well represented channel (Nye. assuming at least some of the monies Even those that do argue that corruption has economic benefits do not suggest that corruption is efficient per se. 1988. Among others. see also Bardhan. Brooks and Heijdra. commodities (whether import licenses. 1967: 420). corruption draws off funds that would otherwise be available for economic growth. 1324-6) Rose-Ackerman (1996) also argues that corruption generates more distortion than taxation. Tullock. more it renders otherwise bad and bad government few virtues: good government resources that could be used productively. on the economic In sum. because such projects increase opportunities Moreover. Many forms of corruption take the form of the sale of limited these are policies. (Leff. Huntington (1968: 69) states it even more boldly: "the only bureaucracy is thing worse than a society with a rigid. are reinvested by the state (Nye. Just as an incentive to bribe exists. and Economic Growth DRURY/KRIECKHAUS/LUSZTIG: and so on. Put differently. Leff (1968) characterizes as a tax on economic corruption activity. overcentralized. 1967). of the right or left are excellent economic modernization exam (dictatorships economic ples) graft may promote growth. 1967: 420). 1974. most of those arguing the benefits of corruption regularly point out that it is not the ideal." beneficial also can be economically because it tends to favor the Corruption most efficient firms. Lien. do not 1997: 123 invest in Russia. under some circumstances. rationalizing for that provided economic substituting private-sector decision-making by the is problematic it does not consider state. 1990.

ate high levels of government and Nelson spending. political pluralism acts to release energies economic and foster conditions conducive to change.. and invest . a second and complementary evict politicians set of focus on the microeconomic arguments effects of a democratic climate. of democracy We now turn to the discussion and economic performance. The IndirectBenefit ofDemocracy As a type of government. democracies for investment. Bueno et al. Olson. argues that authoritarian elites will prey upon societies unless constrained de Mesquita by democratic institutions. democracy both are not entirely clear. This who argued that in newly view was popularized pessimistic by Samuel Huntington. Therefore. A second critique of democracy stems from the neoclassical political-economy for instance. along with Przeworski and Limongi (1993). for instance. which generally suggests that authoritarian avoid and politically better regimes rent-seeking motivated policy mistakes are argued to reduce the surplus available (Haggard. popular political participation not only has the consequence of breaking down the privilege and vested interests of a few but also feeds a participative mentality that carries over into the economic arena and greatly increases the flow of information so essential to effective and efficient governments. the extension and protection of civil liberties and basic freedoms are thought to generate the security of expectation necessary to motivate citizens to work. where we argue that democracy has the indirect benefit of mitigating corruption's harmful impact on an economy. 1960) argues that a symbiotic relationship between wealth and democracy he suggests exists. is touted as having many benefits. nomic) The more pessimistic view of democracy is rooted in an older literature. political these effects: Sirowy and Inkeles (1990: 133-4) nicely summarize Overall. however. arguments. The economic political benefits Several writers have argued that democracy has positive effects on economic growth for a variety of reasons. we hold that corruption will have a negative impact on GDP growth. Third. In addition to this general idea that democracy allows citizens periodically to who hurt the economy. democracy allows for the eviction of bad leaders. holding other factors constant. and development. and economic. to . class retains a strong stake in a system that provides freedom of choice sufficient (political and eco to permit the creation of more wealth. First. at least that one response in order to promote economic Similar arguments can temporarily. argues that special interest groups tend unduly literature. provide analogous. In sum. save.International Political Science Review 27(2) 124 rigid." be found in the literature on East Asia. development. Huntington (1976: 23) argue is that "political participation must be held down. the middle industrial producers. 1990). Specifically. citizen demands will rapidly escalate and gener democratic developing countries.. inefficient bureaucracy. with a consequent negative effect on economic growth. albeit more complicated. In sum. Olson (1993). North (1990). In addition. (2001) similarly argue that authoritarian leaders have few checks on their power and thus engage in cronyism and corruption. that democracy is most likely to is generated occur in an industrialized society in which wealth by a large number of (middle-class) In turn. Lipset (1959. entrepreneurial risk.

it becomes more plural istic and consequently less efficient. however. The ability of the society to react is largely . 2001. in older democracies there is more time for to overcome interest groups the difficulties associated with collective action (Olson. politicians (Bueno de Mesquita 1990). Moreover. reaping Olson (1982) argues 125 particularistic privileges that damage the overall that as a democracy ages. by extension. Our argument is grounded effects. 1990). Corrupt behavior yields obvious benefits. The cost to politicians is primarily determined hurts by how a given act of corruption to this societal actors. Our second assumption is that these Corruption costs vary substantially across types of corruption and types of political system. North. both of which facilitate economic growth (Baum and Lake. and Karl. state reflects. the composition Our argument rests upon two plausible assumptions. that democracy may not merely reduce the level of corruption. state policy. Democracy. (2000) conclude (1994) and Przeworski that there is simply no statistically significant relationship between democracy and growth. Using different methodologies and a fairly sophisticated set of techniques for dealing with endogeneity. that democracy has no significant impact on growth at all. Evans. 1991). 2003. Scholars 1975. 1982). This does not mean. First. tradition develop this argument arguing that authoritarian regimes. democracy has no direct effect on growth. ismore Democracy. political in make-up there are more voices represented government. both Helliwell et al. in East Asia. These potential exist for most benefits in most political systems. 1989. literature suggests that neither perspective empirical is accurate. but scholars are increasingly realizing that democracy does important have indirect effects. Finally. are better able to resist special interest groups' distributive especially demands and rent-seeking pressures (Amsden. We further pursue this insight that democracy might have important indirect to corruption. because the democratic the of its constituents. however. We argue. decreased economic performance (see also Bell. politicians also entails costs. Przeworski and Limongi (1993) resurrect the 19th-century argument that democracy undermines the security of property rights by providing the dispo ssessed with a powerful tool for expropriating political the wealth of property holders. therefore. but also change of corruption. namely the claim that since citizens are better able to remove corrupt facilitates democracy growth et al. or perhaps both are accurate and they balance each other out. politicians weigh the costs and benefits of specific acts of corruption when they are faced with the in an illicit act. Moreover. While both positive and negative the more recent findings have been argued. 1994). in the developmental Brittan. Instead. there are ever-more demands on the resources of the state. and how capable those actors are of responding particular damage through the political system. for instance. especially in one pertaining variant of the compatibility interesting perspective. As a result. which known to be good for economic growth. leading to political sclerosis.DRURY/KRIECKHAUS/LUSZTIG: Corruption.. 1976. facilitates political Helliwell. Simply put. The result is decreased governmental efficiency and. choice of engaging inclu and the ability to gain political support from those ding both personal enrichment groups benefiting from corruption. a result that could lead to considerable economic uncertainty and thus lower economic growth. at least to some degree. likely to lead to greater spending on education and health. and Economic Growth influence economy. This "political" inefficiency leads to decreased economic performance. Wade. democracy is also stability. Schmitter state in depth. 1995.

Our dependent (1982-97). because citizens will demand at least keep that politicians their corrupt behavior from influencing what is probably the most important means of legitimacy in modern nations . Data We now turn to a discussion of the data we use to test our argument. Summary variable. The data are cross-section of more for 16 years arrayed as a time-series than 100 countries statistics for the data appear in Table 1. This will reduce of corrupt in a the total number activities is that this reduction in corruption democracy. When these costs outweigh the benefits of any given corrupt act. Given that authoritarian leaders will not suffer retribution from society. Our expectation corruption. with lower scores indicating that: . For our first independent we rely on the International variable.126 determined International Political Science Review 27(2) by regime type. which is costly for politicians. ability to punish elected officials provides a powerful incentive for politicians to confine their corrupt activities to economically irrelevant activities. but we believe that in a democracy effect will be much weaker. and growth. Instead. In authoritarian systems. This is a rather common-sense for intuition.economic growth. will not be even across all forms of corruption. therefore. However. corruption. deciding quite rationally that this severe misallocation of resources from infrastructure to corruption not would threaten his ability to maintain power (Evans. In democratic systems. is that corruption will have a negative effect on growth in authoritarian regimes. (2001) note. Consequently. they can engage in extremely costly forms of corruption. politicians will be deterred from corruption. growth of GDP. politicians will avoid those that cost society dearly. may continue unabated because the benefits continue to outweigh the minor political costs. for our purposes. less costly forms of corruption. Risk Guide's of corruption in a wide Country (ICRG) assessment range of countries between 1982 and 1997. Corrupt activities that impose a level and composition large cost on society will annoy voters. both the of corruption will be lower. given that such acts are most likely to types of corruption . In sum. The index ranges from six to zero. where Mobuto allowed 90 percent of the road network to erode away. but it has interesting implications the relationships between democracy. A democracy might but this experience high levels of corruption. More interesting.namely. such as nepotism or bribes for expedited access to govern ment officials. removal from office. corruption will be restricted to those activities and sectors that have relatively little economic impact on national performance because voters will definitely act to in significant remove politicians that engage growth-impairing This corruption. the costs of corrupt behavior imposed upon the majority of the population can be safely ignored. our expectation is that at any given level of corruption. A good example of such systematic corruption is Zaire from 1962 to 1994. citizens can remove politicians and. Corruption have severe political consequences that in physical impedes important investments infrastructure and education will not the political costs outweigh be pursued because the benefits. is measured by the World Bank's World Development Indicators (2003). 1995). as per conventional this negative theory. the effect of that on economic corruption than under author growth will be lower in a democracy itarian rule. as Bueno de Mesquita et al. the supporting or ruling coalition is relatively small.

from 1982 until 1997.629 4.441 0 -0. They rate countries is elected. growth and mitigating of democracy.606 6. (1996) argue that democracy as a dichotomous scale. levels of corruption. democracy.019 8. to examine up Given data limitations for other variables. Bangladesh. tax assessment. We dichotomize ment that countries with a score of less than 5. Summary Statistics Variable Mean Standard deviation Growth Level of corruption Life expectancy Trade openness Population growth Logged GDP per capita Tropical climate Government spending 1. We dichotomize and to measure the effect a democratic regime has on economic performance that we expect to see beneficial It is in democracies effects on corruption.346 "high government officials are likely to demand special payments" and "illegal payments are generally expected throughout lower levels of government" in the form of "bribes connected with import and export licenses. indices used in the literature." than 5. policy protection or loans. First. exchange controls.5 are "not free.500 20.5. itwould only be possible six years of data. 1995: 225) We recode the original data so that the least corrupt countries (for example. which would leave merely the ICRG data exists for a much longer period. Haiti.813 1. as Polity and Freedom House do. which measures the latter from the former. by Alvarez. (2) the legislature .013 1.639 0 1. in an index that runs from 2 to 14. 2000). but it is only available for one year.030 0. and Economic Growth DRURY/KRIECKHAus/LuSZT1G: 127 TABLE1.578 138.211 51. while the most corrupt example. Sweden.692 1 180. A somewhat provides a measure better measure is Transparency which provides data for 1996-2003. we use the Polity IV data (Marshall and a country's level of democracy and autocracy and Jaggers. By comparison.Corruption. but rather be measured variable in which countries either are or are not democratic.897 6 4. but have severe limitations (1995) of corruption.464 0. Our second independent is captured by the most common variable. we utilize an check on the robustness Third. Mauro as compared to the ICRG measure. and so on) score a six.734 2.500 0. The result is creates an overall measure by subtracting this variable because we want a score that ranges from -10 to 10.430 0. Limongi.090 67. with lower scores indicating more resulting this index at 5. Finland.174 5. and so on) score a zero. as democratic if: (1) the chief executive is elected. until 2001. (for Australia. Thus.391 440. Freedom House measure Second. Niger.342 Minimum Maximum -52. as an additional index of democracy created and Przeworski Cheibub. should not be rated along a (ACLP).162 0." whereas countries with a score of more of the results. based on Freedom House's judg democracy.922 0. Alvarez et al. economic effects on corruption. higher values mean higher measures of corruption Alternative also exist.174 10.478 13. Democracy.096 0 3. International.5 are either "free" or "partially free." (Knack and Keefer. we use the equally prominent which consists of a combined score of a country's political rights and civil liberties.

is logged life expectancy. According state-induced from free Ricardo's theory of comparative advantage. . recognized since rapid rates of (Bleaney and Nishiyama. We did. The cross is much sectional time-series literature on growth et al. the capital stock per worker will fall. We also include six control variables there are Theoretically. of initial GDP is suggested First. we include a dummy variable identifying is tropical. When the rate of of new workers entering the work population growth is high. rather than a true independent as such.. see Munck and Verkuilen (2002). however. sparser. population growth may inhibit economic growth. Fourth. 1997). First. 1996. run all of the analyses with education in the analysis. of a country that the proportion Sixth. 1997. and found little change in the results. and without more able to work diligently. strong reasons to believe that each influences economic growth. but Kurzman (2002) find that these variables perform reasonably well in pooled samples. in the model. 2002).. 2000). trade openness growth positively. Empirically. as a control Finally. It is likely that these factors are particularly important since much labor is physically strenuous and citizens' overall developing countries. More generally. We did not include education as a control variable because a number of African countries to retain as fell out of the analysis due to missing data. the inclusion by basic neoclassical theory. by Sachs and Warner produc (1997). is expected to influence to Fifth. Second. and this variable is accordingly control variable. Economists Our second control variable argue that the overall health of workers since workers are allows for greater productivity. as defined by the proportion of the country that lies between the tropic of Cancer This variable has been popularized and the tropic of Capricorn. the large number force serves to dilute total capital per worker. For any given level of investment.128 International Political Science Review 27(2) and (3) there ismore than one party (Alvarez et al. rich countries Given diminishing should grow less rapidly than poor countries. resulting in lower levels of economic productivity. government may retard growth since government consumption expen ditures entail higher levels of taxation and thereby reduce private sector actors' to work or produce. succumbing in or debilitation. causality is ambiguous. these in most previous variables have generally been found to correlate with growth cross-sectional analyses (Barro. deviation trade will merely employ the world's resources and reduce world out inefficiently studies find that greater trade openness put. to disease for longer hours. returns to capital. For a critical review of these three different measures of democracy. government shifts willingness consumption resources to the public from the private sector sector. Przeworski et al. Barro (1991) used the log of initial GDP as a proxy for the capital stock. of health is the log of average life expectancy quantitative measure (Barro. and most economists than the public believe that the private sector more efficiently allocates resources sector. this proxy has become a staple of statistical analyses of growth. while earlier growth studies frequently included investment it is increasingly that this represents a suboptimal control variable. The typical health ismore likely to be salient than with respect to white-collarjobs. and we wished large a sample as possible. who note that in a variety of ways agricultural tivity and health is lower in tropical climates. it is not intervening variable variable. Bleaney and Nishiyama. Third. lead to higher investment constitutes an growth levels of investment. Most empirical does in fact facilitate a common growth. 2002).

The insignificance of the corruption. the model corruption growth.1 percent increase in the growth rate. We use panel-corrected standard errors to correct for this bias that might otherwise inflate our significance measures (Beck and Katz.DRURY/KRIECKHAUS/LUSZTIG: Corruption. respectively. democracy the results support our argument that democracy mitigates the negative impact of on economic at Table 2. We correct with a panel-specific dependence AR(1) model (Achen. the 0. 638-640). Unlike the Polity and Freedom House that includes only We then separate the data into two models democracies. in a democracy the same increase leads to a marginal 0. 2000) 2 To test for the differences between non-democracies and democracies. (2002) report an even lower R2 when noting that Kurzman examining are inherently annual data. Levin and Lin. would essentially control for the very effect we are trying to uncover. Hadri. . Because we report the analyses for each of these measures in Tables 2-4. although on the strong. The in all three results tables provide for our almost uniform support in Tables 2-4 report our results for the interaction of argument. There are many more cases than are very likely to have nonspherical time periods. private or public investment. The first columns and corruption. 1994: 78). to control for investment Therefore. One means by which appropriate corruption influence for instance. Looking that for each standard deviation in the level of corruption. The of responses from more comprised data are arrayed as a time-series cross-section. This strategy permits us to the impact of corruption on growth in democracies compare versus non . non-democracies. As they note. 2000. and democracies) (interaction. the ACLP data end in 1990. For both the Polity and Freedom House measures. non-democracies and the other that contains democracies. (Banerjee..0001 their performance is not overly significant beyond level. 1993. However." given business cycles and other short-run factors are accounting for much of the annual variation in growth. 1997. economic all other variables growth decreases by nearly holding constant.4 A nearly identical effect is found in Table 3. Democracy. for this temporal 2002). depending a higher R2 would be preferable. and Economic Growth 129 to control for its effects (King et al.17. Analysis for which data were Selecting all of the countries the data set is available. for example. 1995: 636. that corruption These results clearly support the argument is a drag on economic performance only in non-democratic regimes.. increase predicts 1 percent. with the R2 statistics ranging between 0. This approach to view the differential effects corruption provides a more intuitive means has on and non-democratic democratic we have three measures of regimes. democracy. than 100 countries over a 16-year period. the analysis by seven years and cutting out almost half of the effectively truncating data. is to reduce the quantity of might growth. and data with this characteristic errors. and interaction in democracy. annual models that "noisy. Im et al. Drukker.3 the models are all Overall. Woodridge. variables Table 4 is most likely the result of the limited scope of the ACLP democracy measure.07 and 0. we first interact the democracy and corruption variables. measure of democracy used. 2003. Diagnostic tests revealed that GDP growth is autoregressive 1999. While it is worth et al.

252) 4.088** (0.174) 0.493) -0.017** (0. significant -0.232) -0.International Political Science Review 27 (2) 130 TABLE2.752** (0.419) -1.0188* (0.132** (0.004) -3.956* (0.430** (0.550** (0.022) -3.207 (3.422) -0.006) -72.447) -0.193 (12.400) 0.998 (13.016) -0.064) 761 0.84 (7.625** (0.543) 1435 0.615) -2.341) 1.131) 0.358) -1.111) -0.688** (0.61 (16.363) -1.583** (0.017** (0.345 (0.05 TABLE3.424 (2.900* (0.061 (19.224) 2.623** (0.148) 0.814) -1.37 (0.021) -5.162 (19.727) 0.053 (0.232** (0.126) 602 0.508 (0.134 (0.031 (0. The Effects of Corruption on Economic Growth inNon-Democracies and Democracies.922) -0.507) 0.593) -0.067* (0.081** (0.07 833 0.024) 7.090** (0.243) 0.743) 0.399) -0.005 (0.962 (2.09 at 1 percent Non democracy Democracy -0.5 (7.555* (0.018** (0.07 Observations R2 Non democracy Democracy -0. 1982-97 (Polity fVDemocracyData) Democracy/corruption interaction Level of corruption Democracy Corruption / democracy interaction Life expectancy Trade openness Population growth Logged GDP per capita Tropical state Government spending Constant -0.262) 3.097 (3.006) -11.44 (0.462** (0.419** (22.151) -1.099 (2. 1982-97 (FreedomHouse DemocracyData) Democracy/corruption interaction Level of corruption Democracy Corruption / democracy interaction Life expectancy Trade openness Population growth Logged GDP per capita Tropical state Government spending Constant Observations R2 errors in parentheses Notes: Standard * ** significant at 5 percent.284) 0.181) -1.018** (0.027) -2.863 (33.595) -0.627) -0.900** (16.08 745 0.090** (0.003) -43.072 (0.007) -35.018** (0.902* (0.242) 4.371 (9.813) 1506 0.173) 0.630) -1.513 (11.004) -23.09 .946) 0.271 (2.016) -0.908) -0. The Effects of Corruption on Economic Growth inNon-Democracies and Democracies.383) -1.

242) -0. corruption for democratic never attains or even approaches statistical significance countries.476) -1.110** (0.508** (0.831) -0.078 (25. Unlike their non-democratic brethren. If there was little to no in corruption for democracies.093** (0. the economically itself seems to mitigate of Democracy damaging effects Unlike whose economic corruption. mitigates for the lack of a correlation and between alternative explanations corruption is that the effects are simply however.147 3. As each of the tables report.232) 0.163 (0. The first concern growth in democracies.977 (0. The tables results for non-democracies all show that corruption has a deleterious in effect on economic performance As the interaction model a one standard deviation non-democracies.332) 3. clearly could not have an .191) 0.017** (0.514) -2.17 389 0. and Economic Growth DRURY/KRLECKHAUS/LUSZTIG: 131 TABLE4.845 (2.015) 3.124** (0.488 (2.366 -0.029) -5. performance significantly democracies suffers from corruption. growth in demo however.522) -2. our contention that democracy but not democratic supports regimes. the second and third columns and democracies as two separate samples. Democracy.10 Note: Standard errors in parentheses * significant at 5 percent.247** (0. has a negative effect in authoritarian Our results show that corruption regimes.129) (38.069* (0. predicts.164) 0.966 (5.025) 0.281 (0. non-democracies.337 (7.005) -21. due to a lack of variation in the independent then corruption variation variable.046) 0.979) -3. which are three There the negative effects of any given level of corruption.023** (0.286 (16.094 (0.14 399 0.240) 2.Corruption. ** significant at 1percent in Tables 2-4 report our As an alternative test.640) -0. The Effects of Corruption on Economic Growth inNon-Democracies and Democracies.538) (41.006) 3.494) -1.551 * (0.357 (0.363 (9. increase in corruption leads to nearly a full point decrease in the annual growth rate.020** (0.171 * (0.067** (0.005) -17.599) -0.501) -0.676) Democracy/corruption interaction Level of corruption Democracy Corruption / democracy Life expectancy Trade openness Population growth Logged GDP per capita Tropical state Government Constant Observations R2 spending interaction 788 0. 1982-90 (ACLPDemocracyData) Non democracy Democracy -0.042) 0. corrupt grow just as fast as apparently democracies with little to no corruption. economic cratic regimes is unaffected by corruption.

but while the signs are as expected. and our findings confirm the typical cross-sectional finding. Our main goal in the above analyses was to note that democracy mitigates the negative effects of a given level of corruption. however. reflecting diminishing returns to capital in richer countries.0. the mean level of corruption be lower in democracies. while higher facilitates generally it. Therefore. there are plenty of data to falsify our hypothesis. and our results confirm this standard prediction for the sample with non-democracies. many democracies exhibit very high levels of corruption. expectations The effect of a tropical climate has not been previously tested in annual time series cross-sectional analysis. Moreover.3. Economists analogously argue that trade openness should enhance growth. might causing to have no impact on the economy. the variables generally do not reach conventional levels of significance. and that there is therefore no way to falsify our hypothesis. These free-market are confirmed in most of the analyses. however. fully 29 of the 52 democratic in our sample did at one point or another countries expe rience a level of corruption greater than or equal to the mean level of corruption in authoritarian regimes. corruption it is true that the average Indeed. the maximum including score of six. there is an additional nonetheless positive effect by reducing the impact of corruption at any level. In fact. In fact. . Thus.6 For precisely this reason. the fact that the term is statistically interaction and positive confirms significant that even after for democracy's controlling possible effects through lowering corruption. Although democracies we certainly would not deny that democracy might improve economic performance by reducing the level of corrup tion within countries. A second potential issue is that while corruption might vary in both types of regime. Population growth impedes economic life expectancy growth.5 These models control for the level of corruption in each country. corruption does vary substantially in democracies. namely that the countries in a more tropical climate do in fact suffer less economic significantly growth. Neoclassical growth that initial GDP has a negative theory predicts effect on growth. A third and more is that despite subtle possibility of the high variance in democracies.132 International Political Science Review 27(2) effect on growth even if democracy has no ameliorative effect whatsoever. however.4 in democratic regimes and only 1. It is important to note that previous extensions economic of cross-sectional analysis to pooled data have generally some rendered so we do not necessarily control variables insignificant. Economists also argue that government hurts growth by taking consumption resources away from the (efficient) them within the private sector and placing (less efficient) public sector (Barro. Finally. and even varies more than in authoritarian regimes: the standard deviation of corruption is 1. expect all control variables to be significant. but their effects do not generally reach statistical significance. 1997). while the average corruption corruption score in non is somewhat higher at 3. column one of both Tables 2 and 3 nonetheless confirms that democracy mitigates the effect of any given level of corruption. there simply might not be enough corruption in any corruption to reduce substantially democracy economic growth. the last two control variables also have the anticipated effect. given efficiency gains from comparative and greater opportunities advantage for technology transfer. score in democracies is 2.1 in authoritarian regimes. it is interesting to examine which control variables are robust to cross-sectional time-series analysis. but it is also worth discussing the influence of the control variables.

Given that some nations are rife with corruption. punished results suggest the need for further research on the interaction Our empirical that a democracy's and growth. Another causal story may be that institutions an authoritarian from that constrain individual politicians political parties put in place mechanisms in growth-impairing regimes are likely to Finally. One research should explore that process would the impact of a free press. on economic growth. Our theoretical corruption of the reconceptualization entails a significant evidence. below. by merit. a define structure that liberal substantive output.mitigating corruption's has yet another benefit democracy promoting effects. Most studies of democracy relationships between complex studies of test its direct impact on economic growth and find no result. We cant. We dependent place lagged presented no were to those results The correction. and Economic Growth DRURY/KRIECKHAUS/LUSZTIG: 133 Conclusion and that political processes such as democracy Scholars have long suspected argument. mechanism. democratic corruption. popular in variables of the below with also ran the models 2. 3. another check on the quantity and composition between politics Finally. authority equal elites of the of the individual. in the and where government protection rights grounded ensure must are selected the respon The selection moreover. it is necessary for the electorate to the ability in a democracy occurs. ported by empirical these three variables. and that democracy will mitigate political process in which corruption in the political effects of corruption. page. In addition economic growth. engaging yet for the judiciary. Specifically. effect. we argue that the negative or occurs. that the causal relationship argue effect of corruption is mediated by the Specifically. not only their general human them may enhance rights. changes command files (Stata 8) are available from the first author's web . We argue corruption. between democracy. opportunity Notes a as one to the that conforms government following: where all and the autonomy of all individuals. Future causes corruption electoral mechanism specification of the nature of this causal process. To understand reduce that negative the political to take into account context within which the short. to have no impact on its economy. Many democracies refrain from growth-impairing corruption lest they be but their leaders must at the next election. sup are important for economic growth. which provides greater political independence provide of corruption.Corruption. the law. democracy the preserves and appeared. corruption has the stunting effect corruption remove its leaders from office seems to mitigate exhibit significant levels of corruption. We impact on growth and find a negative test its aggregate corruption between these variables is more complex. Most effect. with universal elections (near) voting rights. AR(1) signifi presented comparable 1. to civil of representatives selection and must entail siveness of elites through society. given that in a demo emphasize activities than in cratic context journalists are freer to publicize growth-impairing such as context. Democracy. our findings suggest that democracy large role in determining ill economic to recommend it . Data. our results provide new insight into the relationship to showing generally that political factors play a and economic growth. where individuals before transparent. rights state are and is limited. but also their within for prosperity.

NJ: ." American of Economy 47: 333-47. Political Journal of Political to Do "What (1995). have undergone more the by the of which statistically is lower societies hence indicated this 2-4. Contradictions of Capitalism. Kurzman 6. Explanatory of Regimes. of (Friedrich. Katz. B. The not Additional significant. The Political and Gordon Tullock (eds)." Journal Countries. but for that richer are lack insignificant. Correlation Panel-Data Models. Princeton (1995)." "Panel Data Unit Roots (1999). small increase coefficients is in impact corruption the appears variable is in non-democracies than the also be relationship might substantial and social political democratic simultaneously less and one for instance. Jonathan Data in Comparative and Series-Cross-Section Growth: Empirical Study.." (1988). corruption beneficial effects and we add democracy. BenJ. Science American N. Special and Development: "Corruption 1320-46. CA. J. Smith. (DUP) Activities. cross-sections with pooled data using (2002). a similar Barro become (1997) reports phenomenon. Brooks. Studies M. "The (2003). "Political Bueno Siverson." Banerjee. Daniel The Cultural (1976). British Journal in Charles of K. (2001). "Directly Unproductive Journal ofPolitical Economy 90: 988-1002. Bulletin Anindya of Economics Robert (1991). Michael Rowley. Morrow. Pranab Bardhan. R. "Explaining Growth: A Contest between Models. Contradictions of Democracy. (1975). University Embedded Press. and A. compare our in find and that R2 falls while control (as many analysis). in growth of Tables states. H. New F. Bleaney. Asia's Next (1989). they sharply.." Journal of Democracy Competition for Serial The in Linear D. in democracies but to be likely effect and corruption of corruption summed significant that show the of democracy. "Why Lagged Los Angeles. J. and A. Developmental and Cointegration: An Overview. Growth." the Political Methodology University Alvarez. Autonomy: States and Industrial Transformation. Cheibub. third second and be might to the due given spurious modernization.A. (3)2: StataJournal Peter Evans.134 International Political Science Review 2 7 (2) To 4. de Mesquita. David Capital. New York: Basic Books. Heijdra... Robert D. Press. (1997). Jagdish N. "Economic A in a Cross Growth Review Section of of Issues. corrupt. (and Not Political Growth: to Do) with Time 89: Science Review 634-47. Christopher Political Oxford Statistics. Michael and Nishiyama. Oxford (2000). MA: MIT and A." of Quarterly Journal ofEconomics 106: 407-43. Boston. and Economic 12: 58-72. and Human Democracy Nathaniel Beck. Political Science 5: 129-60. Alice Amsden. Robert Barro." "In Search of Rent-Seeking. Literature35: Economic Paper of at Industrialization. (2003). Independent Variables Dependent Variables." Profit-Seeking Bhagwati. Section Can Suppress York: Press. J. intervals year variables References Achen. and A. Determinants of Economic Lake.D. J. "Testing 1-10. July H.M. Matthew Baum. of Other Power Meeting Barro. is a near-zero result The 1982). "Classifying in Comparative International (2): 3-36. Tollison Growth Economic 7: 43-56. Limongi. Cambridge. Politics. Akira (2002). et al. Association. a in corruption evidence columns in democratic positive The that average fact 5." A Cross-Country A. (1996). insignificant. and (1997). Giant: and Late South Korea Issue: 607-29. Academic." Drukker. Przeworski. Bell. calculate the interaction effect coefficients in a democracy." of Economic Journal "The Samuel Brittan. presented the American the the Annual Science Political 20-22. Rent Economy of MA: Kluwer Seeking. Princeton. (1982).

San and Politics of in "Unit Tests on Economic International Corruption Perceptions in Brazil. (2000). (1968). http://www. Nye. Democracy. (1995). Paolo and Growth. Press. Donald Da-Hsiang Economics Development (1993)." and Stagflation.A. Data: at San of California American Comparative and Allocation "Corruption Society. NJ: Princeton Qualitative Yale S. in Roots http://www. Time-Series Political Designing P.econ." Analysis. Economic and "Conceptualizing Haven. Economy Krueger.S. Limongi. New (1992).E. Paper Economics 1951-1980. Cambridge. Burkhart. John for 135 (1994). "Political A Cost-Benefit Regimes J." (1990). R. Panels. Andrew Levin. in Participation Social Inquiry: Scientific Inference Press. S. (2000). 64: 291-303." Fukuyama. Tests Societies. Equations. Industrializing Helliwell. CT: Haven. and A Pooled (1997). Y and unpublished Cambridge.E. University Between Linkages and Democracy Economic Growth.. Research. and Institutions Press. "Dictatorship.S. Joan M. andjaggers. J. Characteristics Marshall. in Stationarity Regression Multiple York: Free Press. Document. Samuel Huntington. Przeworski. A. Journal Stephan D. University New Economics. Order in Changing P. American Journal of Francis The End ofHistory and the Last Man. Kaddour Hadri. in the World. of Nations: Press." Economic and Development Policy-Making Chien-Fu Lin. Comparative Political Studies 35 ( 1 ) :5-34. MA: Harvard Press. Developing Im. Measuring and Democracy. Economic Review Lien. University Diego: (1990). Political (1968). Werum. Cambridge New Rigidities. Press. Robert Friedrich. Discussion and 93-56. 1950-1990. Monty Polity TV Project: Political Regime and Transitions." Political 53: 69-105. (2000). "Institutions Political University (1993). J. (1982). Social Political Journal Democracy Cambridge: of and .DRURY/KRIECKHAUS/LUSZTIG: Corruption. "The Political (1974). University and Economic Performance: Cross-Country R. Mancur Olson. Efficiency. Science Review Przeworski. Perspectives A. Princeton. Hashem Pesaran. Political Science 26: 797-833. 87: 567-76. Cheibub. North. Haggard. working Cambridge: faculty/pesaran/lm. Social of Requisites American Democracy. Mancur (1982). York: Cornell "Empirical Press. (1995). Political Man: The Social Bases of Politics.umd. Institutional Institutions." Journal of 153-64. University (1999).. Countries. Development. Development: Cambridge Change. No. and Economic Growth in "In Defense of Multiplicative Terms J." Diego. Institutional Measures. K. Johann 1999: Framework Index Leff." British Journal ofPolitical Science 24: 225-48. Development Graf Lambsdorff." F. Well-Being and Growth. G. in Panel New 1947-1964. Panel Heterogeneous The Data." Economic Performance. M.cidcm. 7(3): the Rent-Seeking (2002). "Some (1959).pdf. F. New York: Martin (1960).L. "Testing Econometrics 3: 148-61. H. Shin. URL: 1800-1999. and 37: 3-33. "Corruption Development: Political Science Review 61: 417-27. Kurzman. Nathaniel York: Wiley. University and Political (1967). (2002). Quarterly Journal of Economics "Corruption G. M. Douglas Cambridge: Olson. Growth: New No Easy Choice: (1976). "Testing for Unit URL: of Applied Department King. F. (1990) Limongi. Lipset. Seymour Science Review of "The Transparency International URL: www. University Nelson. 7: 51-69. and Samuel Huntington." 110: 681-712. and Keohane. and Verkuilen. Mauro. P. Economic and Democracy. (1994). Seymour Doubleday. Munck. C." Heterogeneous Knack. and Decline CT: Yale and Alvarez." and Economic American Analysis. paper. American Growth." using Alternative Anne O. 33: Martin Lipset." Effect "Democracy's in Studies Root 207-28. The Rise University (1993). Keith G. Pathways from thePeriphery: The Politics of Growth in theNewly New Countries.transparency.

6: 335-76. USA [email: krieckhausj@missouri. Dallas. Andrei and Vishny." Paper. USA [email: drury@missouri. Journal Robert Wade. Philippe summer Journal of Democracy.136 International Susan Rose-Ackerman. Washington. Comparative Politics. Washington. and specifically sanctions. ADDRESS:Department of Missouri. Sirowy. ADDRESS: University Columbia. Economies Terry 75-88. as well as numerous articles such as the Review of International Political Economy.smu. A Review. His most in the Journal of Politics. "Sources (1991). Columbia. COOPER DRURY is Assistant at the University of Professor of Political Science on foreign Missouri. of African and Karl. DC: World Bank. Economies. Wooldridge. 27(2) Harmful?" of Slow Is Quarterly in African Growth Democracy "Corruption. World Politics. MO 65211-6030. M. He is the author of Dictating Development: How Europe Shaped the Global Periphery in the British Journal of Political His work has also recently appeared (forthcoming). Cambridge. University NJ: Princeton and World Bank (2003). and Publius. We thank Jay Dow. USA [email:mlusztig@mail. in journals Protectionism. Doh Shin. Risking Free Trade and The Limits of He University. and Panel Data. Taylor Drury for their helpful . (2002).]. Acknowledgments. J. MO Science and World Development. Econometric Analysis of Cross Section the Role of Government in East Press. and Melanie comments on previous drafts of this]. 65211-6030. Journal of Peace].And Journal Is Not. Shleifer. University International Studies Perspectives. TX 75275. ." Studies in Comparative International 25: 126-57. He is the author of Economic Sanctions and Presidential Decisions: Models of Political Rationality (2005). Columbia and conducts research policy and political on economic economy. Bank. Larry and Inequality: Gordon on Economic Alex "The Effects of Democracy Growth Inkeles. Columbia growth. Andrew. DC: World Political Review Corruption (1997). "What Lynn (1993). Working . of Political Science. Asian Industrialization. (1990). . MICHAEL LuSZTIC isAssociate Professor of Political Science at Southern Methodist is author of two books. "Welfare Costs and Theft." of Economics 108: 599-617. World Development Indicators 2003. MIT Press." Western Economic (1967). Science is "When (1996). Robert W. ADDRESS:Department Southern Methodist University." Schmitter. Monopolies 5: 224-32.M. theMarket: Economic Governing Theory Princeton. isAssistant Professor of Political Science at theUniversity of KRIECKHAUS JONATHAN and conducts research on the politics of economic Missouri. Development of Tariffs. (1990). of Missouri. and recent articles appear of Political Science. Journal C. MA: Biographical Notes A. Jeffrey and Warner. and Tullock.