DOI: http://dx.doi.org/10.15665/rde.v13i2.558

Democracy and corruption¹

Democracia y corrupción

Democracia e corrupção

Frédéric Boehm²

¹ Artículo de investigación financiado por German International Cooperation (GIZ), Bonn y Eschborn, Alemania, http://www.giz.de. La investigación se adelantó desde enero 2013 hasta marzo de 2014.
² Doctor in economics from the TU, Berlin, and master in economics of regulation, Universidad de Barcelona, Barcelona, España. Economist and policy analyst OCDE, http://www.ocde.org Paris. Email: frederic.boehm@yahoo.fr

Fecha de recepción 10/02/2015, fecha de aceptación 5/05/2015

Citation
Boehm, F. (2015) Democracy and Corruption. Dimensión Empresarial, 13(2), p. 75-85, JEL: K42, P16, P17, DOI: http://dx.doi.org/10.15665/rde.v13i2.558

Abstract

It's been revised the arguments linking democracy to the control of corruption starting from the often heard call for a “strong man”, an autocrat that could give a country the needed stability and that could take the decisions required to get rid of corruption. Clearly, the call does not survive empirical scrutiny. Rather, the evidence shows a robust negative relationship between level of corruption and level of democracy. However, the evidence also shows that the potential of democracy in fighting corruption hinges upon a series of conditions, and that the mere existence of formal political competition through regular elections on its own is not enough.

Keyword: Democracy, corruption, elections, politicians.

Content: 1. Introduction, 2. linking between democracy and corruption, 3. Empirical evidence, 4. Conclusion.

Resumen

Se han revisado los argumentos que relacionan la democracia con el control de la corrupción, comenzando desde la llamada frecuente por un “hombre fuerte”, un autócrata que pudiera darle al país la estabilidad necesaria y que pudiera tomar las decisiones requeridas para acabar con la corrupción. Claramente, la llamada no sobrevive a un escrutinio empírico. Más bien, la evidencia muestra una relación negativa sólida entre el nivel de corrupción y el nivel de democracia. Sin embargo, la evidencia también muestra que el potencial de la democracia en la lucha contra la corrupción depende de una serie de condiciones y que la mera existencia de una competencia política formal a través de las elecciones regulares, no es suficiente en sí misma.

Palabras clave: Democracia, corrupción, elecciones, políticos.

Resumo

Tem sido revisto os argumentos que ligam a democracia ao controle da corrupção a partir da chamada muitas vezes ouvi de um “homem forte”, um autocrata que poderia dar a um país a estabilidade necessária e que poderia tomar as decisões necessárias para se livrar da corrupção. Claramente, a chamada não sobreviver a um escrutínio empírico. Em vez disso, a evidência mostra uma relação negativa entre o nível robusto de corrupção e nível de democracia. No entanto, a evidência mostra também que o potencial da democracia no combate à corrupção depende em cima de uma série de condições, e que a mera existência de uma concorrência política formal através de eleições regulares sobre a sua própria, não é suficiente.

Palavras chave: 1. Introdução, 2. Vinculação entre democracia e corrupção, 3. A evidência empírica, 4. Conclusão.

1. Introduction

Controlling corruption requires often not only profound changes in the system in a given society but usually also requires overcoming the more or less open resistance from individuals or organised interests that are gaining from the status quo. In this context, it is not uncommon to hear the call for a “strong man”, usually accompanied by a loose reference to the “success” of Singapore. Can we really trade-off political rights and liberties for lower levels of corruption? The arguments in favour of such a view are the same as sometimes put forward with respect to the promotion of economic growth. Allegedly, the “strong men” can give stability to the country enabling thereby coherent long term policies with “vision”. A further argument is that this good dictator is able take fast decisions without the need to go through the process of democratic dialogue and accountability that is often perceived as cumbersome and diluting.

But betting on authoritarian problem-solvers is risky for various reasons and cannot stand up to rigorous empirical tests (Easterly, 2014, Rodrik, 2007). Indeed, the argument of the “strong man” is flawed because of a fundamental problem: the autocrat cannot credibly commit himself, or herself, to a given set of decisions. An autocrat can take decisions without significant control. This may enable him to impose “good” laws and policies, but he can also change them again the next day for whatever reason. Stability in autocracies is thus more likely to be an exception than the rule. Moreover, an autocrat is highly vulnerable to become a kleptocrat; that is, a ruler that uses the state for its private enrichment. Even if there may be some saints out there - the probability to end up with one is quite low. And even this saint may end up succumbing to the temptations of its power after some time³.

Put differently, a look at the Control of Corruption Index from the Worldwide Governance Indicators compiled by the World Bank⁴ shows clearly that the twenty least corrupt countries are all democracies scoring a perfect 1 in the Political Rights indicator compiled by Freedomhouse, which includes criteria related to electoral process, political pluralism and participation, as well as on the functioning of government⁵. The exception to the rule in the top 20 is Singapore, scoring an intermediate 4. The correlation graph in figure 1 between the Control of Corruption Index (0 to 100) and the Political Rights indicator (1 to 7), including data from 193 countries, shows indeed a negative relationship: Fully functional democracies (score of 1) tend to be less corrupt, and the countries with the lowest control of corruption scores in this category are Nauru, Mongolia, and Tuvalu, with scores around 40. In turn, countries with bad scores in Political Rights (6 and 7) in general exhibit high corruption levels. The exceptions here are oil producing countries (United Arab Emirates, Qatar, Bahrein, Brunei), as well as Rwanda and Cuba. We will return to this correlation later when reviewing the empirical evidence.

Democracy therefore seems to come along with lower levels of corruption. But despite being suggestive, a simple correlation graph doesn't permit to draw conclusions with respect to causality. Does democracy cause less corruption? To answer this question the next section discusses the theory onto the mechanism through which democracy may influence the level of corruption. To do so, it is first of all necessary to discuss briefly the concepts of democracy and corruption. The third section then proceeds to review empirical evidence gathered by research in order to see whether the theoretical claims can be observed empirically. The article concludes with some final thoughts on the relationship between corruption and democracy.

2. Linking democracy and corruption

2.1 What exactly is corruption and what is democracy?

Corruption is a complex and multifaceted problem, and there have been many trials to find a definition (see for instance Gardiner (2005) for an overview). The most cited definition, used by Transparency International, the United Nations and in many research publications, defines corruption as the abuse of entrusted power for private benefits. The United Nations Convention against Corruption (UNCAC) does not define corruption, but defines in its second chapter practices that are considered as corrupt according to the global consensus and that have to be translated into the national laws of UNCAC member states and implemented by effective and coherent public policies (UNCAC, article 5).

The definition of corruption as an abuse of entrusted power for private benefits emphasises that corruption involves the delegation of power and a conflict between the private and the public interest. The concept of delegation, independently from the political regime, is quite clear within the public administration. Any regime needs an administration, and corrupt practices at these levels may occur both in democracies and autocracies. At first sight, an autocrat may even have more interest in controlling administrative corruption, as these are rents he cannot reap himself (see Lambsdorff, 2007: chapter 4, on this claim and why it doesn't hold). In turn, politicians elected for public office in representative democracies have been delegated power by the citizens for a limited period of time. They too can abuse of their delegated power. But the issue of such high level political corruption becomes less straightforward in countries without elections, as no formal rules of delegation of political powers can be violated there. In these cases, it is necessary to fall back to less precise notions such as public interest and legitimacy.

The concept of democracy is perhaps even more complicated and complex. An often used definition is the one proposed by Schumpeter (1942: 250): [Democracy is] the institutional arrangement for arriving at political decisions in which individuals acquire the power to decide by means of a competitive struggle for the people's vote. As highlighted by Acemoglu and Robinson (2005: 48), this means that in practice the concept of democracy is associated with a particular set of institutions, which are free and fair elections, the accountability of politicians to the electorate, and free entry into politics.

Sodaro (2004: 31) in its introduction to comparative politics defines democracy as follows:

The essential idea of democracy is that the people have the right to determine who governs them. In most cases they elect the principal governing officials and hold them accountable for their actions. Democracies also impose legal limits on the government's authority by guaranteeing certain rights and freedoms to their citizens.

Interestingly, the author also provides a list of “ten conditions for democracy”, where he enumerates the following points: state institutions, elites committed to democracy, a homogeneous society, national wealth, private enterprise, a middle class, support of the disadvantaged for democracy, citizen participation, civil society, and a democratic political culture, education and freedom of information, and finally a favourable international environment (Sodaro, 2004: 207-220).

Whether we agree in detail with this list or not, democracy is definitely not an “either or not” issue but a matter of degree. Countries can exhibit different levels of democracy. The view of democracy as a matter of degree is reflected by the existing indexes that try to measure democracy. With exception of the approach followed by Przeworski et al (2000), who construct an index classifying countries in either democracies or non-democracies⁶ all other indexes use a classification on a scale based on a range of criteria. The most widely used indicators of that type are the already mentioned Freedomhouse Indicator, the Polity IV index compiled by the Center for Systemic Peace, and the Democracy Index computed by the Economist Intelligence Unit.

In the end, the idea that politicians should be accountable to the citizens is definitely a key aspect, if not the most important ingredient of representative democracies. This accountability through votes is usually referred to as vertical accountability, in order to differentiate it from horizontal accountability, where some state entities, such as supreme audit institutions, regulators or the judiciary, control and sanction other government entities (checks and balances).

a. Why democracy may help in controlling corruption?

This idea of different “degrees of democracy” is of profound relevance for the link between democracy and corruption. It suggests that the alleged positive impact of democracy on corruption control is not so much due to the compliance with the existence, de jure, of some democratic institutions on paper, but hinges upon the effective implementation of these democratic institutions. Such an effective implementation may take time and may suffer from reversals as well.

But what exactly is the logic behind the idea that democracy helps in controlling corruption? The argument goes, basically, as follows. Even though corruption is sometimes referred to as a victimless crime, this view is wrong. Certain corrupt practices do involve a direct victim; this is the case whenever bribes or favours are extorted by public officials. In other words, citizens have to pay (either cash or through favours) for something they are entitled to get without the illicit payment, or to avoid an undue cost that can be imposed by the public official. The practices of embezzlement, bribery, favouritism or fraud, in turn, may not involve a direct victim, but these practices generate external effects imposing costs on third parties or the public interest in general. For instance, corruption in procurement of government construction contracts usually translate in lower quality and higher prices that have to be paid by general taxes, i.e. by everyone. Corruption in the health or education system is likely to translate in lower quality of these public services. Studies and examples on these indirect costs of corruption in sectors abound in the literature (see for instance, OECD, 2015).

If citizens are confronted with these indirect costs of corruption or are direct victims of extortion they could, through their votes, punish the government responsible for letting such practices happen, or reward an honest government with re-election. If a politician is directly involved in a corruption case, the reason for sanctioning the politician by not voting for him again is even more evident.

Under this logic, in theory, corrupt politicians or politicians turning a blind eye on corruption would not survive in a political contest for votes. In other words, it is the accountability of politicians to the citizens that control corruption. In authoritarian states this is not the case: the prediction therefore would be that corruption will be more ruthless in such countries (Drury et al, 2006). Recalling corrupt practices of dictators such as Mobuto, Suharto or Sani Abacha, and the evidence of the top 20 least corrupt countries being all democracies, seem to provide some anecdotal evidence for this.

However, at the same time recent evidence from countries that switched to or returned to democracy seems to show that this is not the whole story. As Sung (2004: 181) notes: (…) political liberalization has made matters worse in most of those countries that embarked on the democratic transition in the 1980s and 1990s. For instance, electoral competition in some countries seems rather to have incremented corrupt practices such as vote-buying in order to gain elections. The public office is then abused to extract as much public money, e.g. through corrupt practices in public procurement, as possible and perhaps to secure political support for a re-election, e.g. through the use of clientelistic practices.

Some simple considerations therefore suggest that this positive effect of democracy depend on certain conditions beyond only allowing for political competition and organising elections. Firstly, in order to punish misbehaviour, the misbehaviour obviously has to be observed by the citizens in the first place. But corruption, with the exception of low-level administrative extortion and bribery, characterises itself as a crime where the two (or more) involved parties gain directly from the deal. Corruption thus remains undiscovered unless internal conflicts arise that lead to denunciations, or the deal is discovered externally by investigators, auditors or journalists.

Horizontal accountability institutions that could uncover corrupt practices should, in principle, work better in a democracy than in autocracies. Note that investigated cases have to be publicised or at least accessible under freedom of information laws in order to enable citizens to take them into account when voting. Nevertheless, horizontal accountability seems to be just as unlikely to work in corrupted democracies as in autocracies. In turn, control institutions in autocracies are likely to overlook corruption where the autocrat or his clan is involved, and even may be abused by the autocrat to control political opponents or to get rid of them through false or true but selected corruption charges. Also, democracies are more likely to guarantee the right of freedom of speech, enabling the media to uncover, investigate and denounce cases of corruption, and thereby enable citizens to take electoral decisions based on these reported cases. We would therefore expect that freedom of the press potentiates the impact of democracy in controlling corruption.

Secondly, assuming that corruption is known, citizens must be able and willing to actually act based on this information and sanction corrupt government during elections. But, on the one hand, the decision to vote for one candidate or another depends on many factors such as personal or ideological preferences, tradition, or expected benefits from the policies promised by the candidates, which may also include working opportunities in the public administration or straightforward corrupt practices such as vote-buying. Past achievements or misbehaviour, for that matter, is only one part of information taken into account and may not be the determining factor. On the other hand, where corruption is endemic, voters may not be able to identify honest politicians anymore, perceiving all alternatives as being more or less equally corrupt, or the citizens may themselves become morally blunted, cynic and accept corrupt behaviour just as a normal aspect of political life. Additionally, the level of economic development and the general level of education of the electorate may be of relevance too.

3. Empirical evidence

Based on these or similar theoretical considerations, various studies have tried to establish empirically whether democracies do contribute to controlling corruption or not. In a first step we review studies using corruption perception data as well as some indicators of democracy or related indicators, as in the correlation graph above. After this review, in a second step, I review the evidence with respect to the central question of whether voters do actually punish corrupt politicians.

3.1 The relationship between democracy and corruption in the data

One of the first studies exploring empirically the causes of corruption across countries is Ades and DiTella (1999). They find only a weak negative relationship between political rights and corruption. Treismann (2000) in turn, also interested in the determinants of corruption, finds no significant evidence that the current degree of democracy predicts lower levels of corruption. However, the author finds that long exposure to democracy significantly predicted lower levels of corruption. This result is strongly confirmed by Serra's (2006) robustness analysis of 16 variables previously identified in the literature as determinants of corruption⁷. From these 16 variables, only 5 survive: economic development, protestant religion, colonial heritage, uninterrupted democracy, and political stability. This seems to confirm that the benefits of democratic institutions do not impact immediately, but requires time to unfold until these institutions stabilize.

Sung (2004) and Rock (2007) both are interested in the hypothesis that the negative relationship between corruption and democracy may not be linear. Such a non-linear relationship seems plausible taking into account Treisman`s (2000) result and the observation that many countries that recently started to introduce democratic reforms seems to be experiencing an increase in corruption. Nevertheless, overall democracies in the end seem to be less corrupt, suggesting that the relationship between both variables could look like an inverted U-shape. That is, at the beginning of democratization, countries may experience a rise of corruption levels, but when democratic institutions become stronger over time and a culture of democracy develops, democracy begins to control corruption and corruption levels start to fall.

In their regressions, both Sung (2004) and Rock (2007) find that such a quadratic inverted U-shaped function improves the predictive quality of the model. Rock (2009) finds that the …turning point in corruption occurs rather early in the life of new democracies - between 10-12 years. Sung (2004) went a bit further and showed that actually a cubic function does even a better job in predicting corruption levels than the quadratic function. The author finds that there may be …eruptions of corruption among intermediate democracies, but that …the consolidation of advanced democratic institutions eventually reduced corruption. Sung (2004) therefore provides evidence that both the initial political conditions and the final democratic achievements determine the level of corruption in a country.

Nur-tegin and Czap (2012) ask the interesting question whether it is worth, in terms of corruption, to trade a stable autocracy for democracy with political freedom but transitional instability. The authors thus ask a similar question to the one I raised at the beginning, whether it may be better to give up some political liberties in favour of less corruption due to the stability an autocracy can allegedly offer. However, Nur-tegin and Czap (2012) clearly find that this trade-off doesn't exist. Their results show that the level of corruption is lower in unstable democracies than in stable dictatorships. Interestingly, they only include stable autocracies and unstable democracies in their sample; as they argue, the other cases are quite unambiguous. This empirical evidence should put an end to the idea of having a “strong man” doing the job of reducing corruption - in general, he simply doesn't do it.

A recent study contributing to exploring this general relationship comes from Kolstad and Wiig (2012). The author observe that there are endogeneity problems of the proxies used to measure democracy in the previous contributions, as the level of corruption is taken itself into account when computing the level of the democracy indicator. The author therefore propose to use instead a dummy variable reflecting whether a country has been at war with a democracy in the period 1946-2009, while controlling for the extent to which countries have been at war in general. Kolstad and Wiig (2012) still find that democracy significantly reduces the level of corruption, and that the effect is even considerably larger than suggested by the previous studies. The authors thus conclude that …democracy may hence be more important in combating corruption than previous studies would suggest.

Concerning the hypothesis that there may not only be more corruption in autocracies, but that the impact of this corruption also may be stronger because corrupt practices will be more “ruthless”, Drury et al. (2006) find indeed that corruption has no significant effect on economic growth in democracies, but that corruption significantly reduces growth in non-democracies. Along the same line, Bhattacharyya and Hodler (2010) find that natural resources in a country can feed corruption when the quality of the democratic institutions is weak, while countries with strong democratic institutions do not suffer from this perverse link between natural resources and corruption. Both studies thus suggest that effective democratic institutions work as a safeguard impeding that corrupt practices have overly serious negative impacts for the country. More research is however needed in this area.

a. Do voters really punish corrupt politicians?

Let us now turn to empirical evidence concerning the conditions under which democracy is likely to work in controlling corruption. We saw that the logic hinges upon two aspects: Firstly that corrupt cases become known to the citizens, and secondly that the citizens are able and willing to use this information to sanction a government during elections. Only then, politicians will not only face incentives to be less corrupt, but may even compete in honesty introducing anti-corruption reforms.

Unfortunately, I am not aware of a study comparing differences in the effectivity of horizontal accountability institutions depending on the level of democracy. However, various studies have established empirically that press freedom and democracy together have a significant impact on controlling corruption (Chowdhury, 2004, Bhattacharyya and Hodler, 2012, as well as Kalenborn and Lessmann 2013). The emphasis here is on the finding that democracy and press freedom develop their positive impact jointly - they are complements in the fight against corruption, not substitutes. This confirms that citizens one the one hand need to know about corruption cases in the first place (through the media), and that on the other hand citizens need to have political rights (through democratic institutions) to be able to use this information obtained by the media and politically sanction the governments through their votes.

This brings us to the last but key question: do citizens actually punish corrupt politicians during elections? The data presented until now suggest that this indeed seems to be the case, but the crucial step in the theoretical causal chain leading to lower corruption is the sanction imposed on corrupt politicians: the sanction provides the incentives to change behaviour. Anecdotal evidence of corrupt politicians that have been re-elected into office at least cast some doubt on this logic; but what does the evidence so far tell us?

First of all, there are some country-specific studies. For Italy, Golden and Chang (2001) find that …electoral competition with other parties did not significantly affect the extent of charges of corruption in the country after the Second World War. In a follow-up study, Chang, Golden and Hill (2010) find for the same time period (1948-94) that judicial investigations do not discourage politicians from presenting themselves for re-election in Italy's large electoral districts. Interestingly, the authors also find that voters started to punish corrupt politicians beginning in the early 1990s. In order to understand this change in voter behaviour, the authors focus on the roles of the judiciary and the press and confirm their importance; their results are consistent with the theory that an independent judiciary alone is not effective in ensuring electoral accountability if the public is not informed.

For Spain, Costas-Pérez et al. (2012) study the effects of the availability of information in the press about corruption scandals on electoral outcomes. The authors use a database covering the period 1996 to 2009 with information on press coverage and judicial intervention in corruption scandals that affected Spanish local governments. Therefore, the authors could check how the voters reacted to the amount of information in the press (e.g., number of news) and to information regarding the seriousness of the case (e.g., judicial charges). The authors show that an incumbent involved in a corruption scandal can lose up to 14 % of votes in cases the politician has been charged with corruption and press coverage has been extensive. Cases that were reported in the press but did not lead to judicial charges, in turn, were not punished by the voters. It seems that Spanish voters distinguish between founded and unfounded corruption accusations.

In Brazil, Ferraz and Finan (2008) make use of random local government audits conducted before the 2004 election in order to study whether the information affected the vote for the incumbent. The results from these audits were publicly available. Their results show that corrupt mayors can lose from 10 to 30 % of their vote share. Their probability of re-election drops as much as 17 %. In turn, honest mayors saw an increase in votes and re-election probabilities. Just as mentioned in the studies on Italy after the 1980s and Spain, the authors find that the existence of a local radio station that disseminates the audit information intensifies the punishment of the voters: the probability of not being re-elected then rises to 30%.

Unfortunately, Pereira et al. (2009) showed that mayors exposed as corrupt are only less likely to be re-elected when the information of their misbehaviour is released in the same electoral year. Costas-Perez (2013), using the same database on Spain as in their previous article, seems to confirm that voters easily forget. The author finds that aggregate levels of corruption only lower the voter turnout if the scandal takes place in the year of election; corruption that occurs in years prior to an election has no effect. Citizens apparently suffer from amnesia when it comes to corruption.

A further sobering finding comes from Puerto Rico. Bobonis et al (2011) go beyond the question of whether corrupt politicians are sanctioned during elections and ask whether audit information about these corrupt activities induce a sustained reduction in corruption levels. The authors find that …municipal corruption levels are on average the same in municipalities audited preceding the previous election and those whose audits became publicly available afterwards. Furthermore, the authors find that mayors' re-election rates increase in future terms. Their findings thus shed doubts on whether the availability of information necessarily will change the behaviour of politicians. Just as in Brazil, contextual factors, such as low detection probability or impunity, may explain this pattern.

Beyond country-level evidence, there are only a few studies exploring the question through cross-country analysis. With data from 28 countries between 1996 and 2007, Krause and Méndez (2009), for instance, study the question whether voters are more likely to vote out a corrupt politician than to re-elect him. Instead of audit data, which is not available and/or comparable for different countries, the authors ask if citizens observe changes in the Corruption Perception Index by Transparency International between the current and the last elections. The results suggest that corruption is effectively punished by voters.

Crisp et al (2012) use data from the Global Corruption Barometer, a survey commissioned by Transparency International that includes, amongst other the following question: To what extent do you perceive the parliament/legislature in this country to be affected by corruption? With data drawn from 169 elections across 72 countries, the authors show that corruption does indeed provoke a sanction by the voters during elections. However, similar to the result in Puerto Rico, these sanctions are not followed by lower levels of perceived corruption suggesting that political accountability on its own may not be enough to cope with the complex phenomenon of corruption.

Finally, in one of the most recent contributions in this field of research, a working paper by Fernández-Vázquez and Gonzalo Rivero (2014) suggest to distinguish between two types of corruption, according to the consequences they have for the voters. The authors use data from the 2011 Spanish local elections and the corruption allegations associated with the Spanish housing boom. Again, the results are disappointing. The authors' findings seem to confirm that voters ignore corruption when they come along with benefits to them, and that politicians are only sanctioned when voters do not receive compensation.

4. Conclusion

This paper reviews the relationship between democracy and corruption. More specifically, it asks how and whether democracy may be able to control corruption. The rich and interesting literature dealing with the effects of corruption on democracy (see, for example, Seligson, 2002, or Morales Quiroga, 2009 - both for Latin America) has not been touched. This does not mean that I do not think that the causality runs both ways. Corruption in government undermines the economic development (Aidt, 2011) and the legitimacy of a state. If either elected or appointed public officials that have been entrusted to pursue the public interest are abusing their power acting based on narrower private interests, the citizens will lose trust in their government. This in turn may open the path for more corrupt behaviour in the society as a whole (Andvig and Moene, 1990). Investigating the potential vicious cycle between corruption and lower democracy is in itself an important and interesting avenue for future research.

The logic behind the hypothesis reviewed here, that democracy is able to control corruption, is straightforward in theory: Through their votes, citizens can sanctions corrupt politicians or politicians that do not invest sufficient effort in controlling corruption. The existence of these sanctions should influence the behaviour of politicians, and in the end the political competition for votes should lead to honest politics. A priori, the fact that the twenty least corrupt countries are all full democracies with exception of Singapore provides some first evidence for this logic.

However, many democracies do face significant corruption, and more importantly countries that have recently embarked on a path of democracy sometimes even seem to experience rising corruption levels. Therefore, it seems that the effect of democracy on corruption is contingent on some conditions; especially that voters know about the corrupt cases, that they are willing and able to react based on this information, and that horizontal control institutions (checks and balances) work effectively.

The empirical evidence reviewed yields various insights. First of all, what matters for the control of corruption is not so much the current state of democracy, but rather whether the country has a long standing tradition in democracy or not. This seems to show that democratic institutions take time to stabilize and that citizens must learn and develop a culture of democracy in order to make efficient and effective use of their political and civil rights. Second, the evidence contradicts clearly the often heard argument that it might be better to give up some political and civil rights so that a “strong man” can clean up the country: instable democracies are still less corrupt than stable autocracies. Third, cross-country evidence suggests that the relationship between democracy and corruption is not lineal, but quadratic or cubic, implying that countries may experience a rise in levels of corruption at the beginning of their democratization.

With respect to the question underlying the causal logic between democracy and corruption, whether citizens do actually sanction corrupt politicians, the evidence is mixed. Although most studies find that citizens do indeed sanction corrupt politicians, these sanctions do not seem to change political behaviour in a way that would lead to lower corruption levels overall. Even though this finding may be related to the difficulty to measure corruption, this finding definitely calls for more research.

Finally, a consistent finding throughout the empirical evidence reviewed is the complimentary relationship between democracy and freedom of the press. The media can play a very important role in uncovering and diffusing corrupt practices in a country giving the electorate the information basis they require. Let's therefore conclude this review with a quote from Freedomhouse regarding the latest state of freedom of the press in Colombia⁸:

The press in Colombia continued to work under difficult conditions in 2013. Although the legal environment improved slightly due to enhanced norms regarding access to information, the security of journalists remained a serious problem, with an increased number of threats, attacks, and killings recorded during the year.

From the perspective of the findings presented above and the levels of corruption at both national and subnational level, this situation should ring the alarm bells with respect to the path of Colombia towards a healthy and sustainable democracy.

Notas

³ See Transparency International (2004: 13) for a list of ten infamous kleptocrats and the amounts they stole from their countries.
⁴ Alternatively, the Corruption Perception Index (CPI) compiled yearly by Transparency International can be used obtaining the same result. See González-Espinosa and Boehm (2013) for caveats in the interpretation of these indices.
⁵ They also score a perfect 1 in the indicator measuring civil liberties. Note that Emerson (2006), using the Civil Liberties indicator from Freedomhouse instead of the Political Rights indicators as proxy for democracy, finds that the higher levels of democracy, measured this way, also come along with lower levels of corruption.
⁶ Such a dichotomy may be useful for certain types of analysis (see Acemoglu and Robinson, 2005: 16-19 and 51).
⁷ These are (Serra, 2006: 225-226): "...four economic variables (economic development, openness to international trade, state intervention in the economy, endowment of natural resources), five sociocultural variables (British legal system, British colonial Heritage, Protestant religion, ethnolinguistic fragmentation, education of the population) and seven political variables (base political rights, uninterrupted democracy, freedom of information, mass media diffusion, federalism, electoral system, political instability)."
⁸ https://freedomhouse.org/report/freedom-press/2014/colombia#.VUUi9fl_Oko.

References

Acemoglu, D., Robinson, J. (2005). Economic Origins of Dictatorship and Democracy. Cambridge: Cambridge University Press.         [ Links ]

Ades, A., DiTella, R. (1999). Rents, Competition, and Corruption. The American Economic Review 89(4), 982-993.         [ Links ]

Aidt, T. (2011): Corruption and Sustainable Development. En: Rose-Ackerman, S. & Søreide, T. [eds.]: International Handbook On The Economics Of Corruption, Volume Two. Cheltenham: Edward Elgar.         [ Links ]

Andvig, J.C., Moene K.O. (1990). How Corruption May Corrupt. Journal of Economic Behavior and Organization 13, 63-76.         [ Links ]

Bhattacharyya, S., Hodler, R. (2010). Natural resources, democracy and corruption. European Economic Review 54 (2010) 608-621.         [ Links ]

Bhattacharyya, S., Hodler, R. (2012). Media freedom and democracy in the fight against corruption. European Journal of Political Economy 39, 13-24.         [ Links ]

Billger, S.M., Goel, R.K. (2009). Do existing corruption levels matter in controlling corruption? Cross-country quantile regression estimates. Journal of Development Economics 90: 299-305.         [ Links ]

Chang, E.C., Golden, M.A., Hill, S.J. (2010). Legislative Malfeasance and Political Accountability. World Politics 62(2), 177-220.         [ Links ]

Chowdhury, S.K. (2004). The effects of democracy and press freedom on corruption: an empirical test. Economic Letters 85, 93-101.         [ Links ]

Costas-Pérez, E. (2013). The Effects of Political Corruption on Turnout: Mobilization or Disaffection? Universitat de Barcelona & Institut d'Economia de Barcelona (IEB).         [ Links ]

Costas-Pérez, E., Solé-Ollé, A., Sorribas-Navarro, P. (2012). Corruption scandals, voter information, and accountability. European Journal of Political Economy 28, 469-484.         [ Links ]

Crisp, B.F., Olivella, S., Potter, J.D., Mishler, W. (2014). Elections as instruments for punishing bad representatives and selecting good ones. Electoral Studies 34, 1-15.         [ Links ]

Drury, A.C., Krieckhaus, J., Lusztig, M. (2006). Corruption, Democracy, and Economic Growth. International Political Science Review 27(2), 121-136.         [ Links ]

Easterly, W. (2014). Economists, Dictators, and the Forgotten Rights of the Poor. Basic Books.         [ Links ]

Emerson, P.E. (2006). Corruption, competition and democracy. Journal of Development Economics 81, 193-212.         [ Links ]

Fernández-Vázquez, P., Gonzalo Rivero, P.B. (2014). Rooting Out Corruption or Rooting for Corruption? The Heterogeneous Electoral Consequences of Scandals. New York University.         [ Links ]

Golden, M.A., Chang E.C. (2001). Competitive Corruption. Factional Conflict and Political Malfeasance in Postwar Italian Christian Democracy. World Politics 53 (July), 588-622.         [ Links ]

Gardiner, J.A. (2005). Defining Corruption. In: Heidenheimer, A.J, and Johnston, M. [eds.]: Political Corruption. Concepts & Contexts. Third Edition. New Brunswick and New Jersey: Transaction Publishers.         [ Links ]

González-Espinosa, A.C., Boehm, F. (2013). Medir la Corrupción: De la generación de consciencia a los peligros de la mala interpretación. Revista OPERA 11, 259-277.         [ Links ]