Decentralization and Corruption

A Cross-Country Analysis

Scientific Essay 2010 21 Pages

Politics - Political Systems - General and Comparisons



1. Introduction

2. Theoretical background
2.1. Defining the key concepts
2.1.1. Defining decentralization
2.1.2. Defining corruption
2.2. Literature review
2.2.1. Lines of argumentation
2.2.2. Empirical results

3. Theory building
3.1. The accountability-argument
3.2. The competition-argument

4. Theory testing
4.1. Capturing the key concepts
4.1.1. Measuring corruption
4.1.2. Measuring decentralization
4.2. Empirical analysis
4.2.1. Procedure
4.2.2. Controls
4.2.3. Findings

5. Conclusion

List of tables

Table 1: Corruption and decentralization scores per country

Table 2: Summary statistics Corruption Perception Index scores

Table 3: Control variables: Description and sources

Table 4: Summary statistics Control variables

Table 5: Correlations

1. Introduction

Corruption is one of the most widespread political problems throughout the world. Decentralization, in contrast, is considered by many to be one of the most helpful ways to improve a country’s economic performance. Nevertheless, only recently have scholars began to look at a possible linkage between the level of decentralization in a country and its level of corruption. Up to now, theoretical insights and empirical evidence in this emerging debate remain both rare and highly contested.

This paper aims at contributing to the debate in two ways. On the one hand through challenging and developing several arguments speaking for and against a possible causal relationship between decentralization and corruption; and on the other hand through testing the decentralization-corruption-hypothesis empirically using four different measurements of decentralization. Both parts aim at answering the following question: Is there a causal relationship between a country’s level of decentralization and its level of corruption?

The rest of this paper is structured as follows: The first part establishes the theoretical foundations through reviewing the existing literature, defining the key concepts and rethinking the main arguments. The second part aims tests the research question empirically. The reader finds a description of the variables and the controls we include. Furthermore, the results of the regression analysis will be presented as well as its limitations. The paper closes with a conclusion about the insights this study could provide and some thoughts about what future research projects could try to investigate in more detail.

2. Theoretical background

2.1. Defining the key concepts

2.1.1. Defining decentralization

to define decentralization have been numerous. Beginning with the static definition of decentralization as the “total expenditure of sub-national [...] governments over total spending by all levels [...] of government” (Fisman and Gatti 2002: 341) to dynamic definitions like Arikan’s (2004) “increase in the number of competing jurisdictions” (Arikan 2004: 176), a multitude of meanings have been attributed to the term decentralization.

Due to the inconsistent terminology among scholars, this paper is based on a very broad definition of decentralization, which is compatible with most previous definitions and furthermore enables us to accurately operationalize the term for the empirical analysis. Under decentralization we furthermore understand the devolution of power from higher to lower levels of government.

2.1.2. Defining corruption

Corruption might seem to be as hard to define as decentralization, but fortunately, the literature on corruption has made enormous steps towards finding and agreeing on a common terminology (see Redlawsk and McCann 2005: 263).[1] Two frequently cited definitions are: corruption as the “behavior which deviates from the formal duties of a public role because of private regarding [...] pecuniary or status gains; or violates rules against the exercise of certain types of private regarding influence” (Nye 1989: 966), and the „standard definition’ of corruption[2] as the abuse of public office for private gain. Both definitions are nowadays cited with slight modifications by most scholars, by the World Bank and by Transparency International (see Bukovansky 2006: 191). In order to allow for theoretical and empirical accuracy and consistency, this study, too, defines corruption as the abuse of public office for private gain.

2.2. Literature review

2.2.1. Lines of argumentation

In the debate about decentralization and corruption one can observe a basically outweighed number of scholars arguing for a negative, for a positive, and for no causal relationship between decentralization and corruption. Those who regard decentralization as a remedy for corruption use the key words interjurisdictional competition and accountability (see e.g. Fisman and Gatti 2002: 327f.)[3], while the opposite camp argues for example that there are more coordination problems and less bureaucratic competence at local levels (see e.g. Fan et al. 2009: 29-32)[4].

2.2.2. Empirical results

Empirical analysis within the research about decentralization and corruption is commonly based on data from perceived-corruption-indices (see Tavits 2007: 222). The use of subjective evaluation data is often criticized (see Lederman et al. 2005: 8f.), but this is not the only problem scholars have to deal with when examining the relationship between decentralization and corruption. Instead, there seems to be a general difficulty to develop convincing arguments or theories concerning this possible relationship and to get valid data in support for any of them. The empirical results have been “contradictory” (Fan et al. 2006: 14), “incomplete [...], mixed and anecdotal” (Faguet 2007: 1101).

3. Theory building

3.1. The accountability-argument

The issue of accountability seems to be the most crucial point within the whole decentralization-corruption-debate. Even if other factors have an impact on the level of corruption in a country, too, there is no denying that the effects that increased accountability can have on the level of corruption outweighs all of them (see Tambulasi and Kayuni 2007: 168). More controversial is the issue of whether it is true that there is a positive correlation between decentralization and accountability (as assumed for example by Lederman et al. 2005).

As Grant and Keohane (2005) put it, accountability “implies that some actors have the right to hold other actors to a set of standards, to judge whether they have fulfilled their responsibilities in light of these standards, and to impose sanctions if they determine that these responsibilities have not been met” (Grant and Keohane 2005: 29).

Accountability can take many forms, internal and external accountability, vertical and horizontal accountability, non-democratic and democratic accountability and so forth (see Tambulasi and Kayuni 2007: 167). Within the decentralization-corruption-debate, however, the focus almost exclusively lies on horizontal accountability, sometimes also called downward accountability (see Vérón et al. 2006: 1928).

As jurisdictions get smaller, some argue (see e.g. Tommasi and Weinschelbaum 2007: 380), the ability of citizens to control their officials increases, resulting eventually in enhanced accountability. But what does it mean for a citizen to control an official? It is indeed easier to go to the official's bureau if you live closer to him, and if you want to follow him wherever he goes and observe all his steps, you can fulfill your task the easier the smaller the jurisdiction you live in is.

However, one very important tool to control officials almost completely lacks at local levels: the press (see Gunther and Mughan 2000: 418-420). While politics at higher levels of government are thoroughly scrutinized and followed by journalists of all kinds, regional and local issues get only scant attention by the press (see Asthana 2008: 188). Without the controlling and sanctioning role the press plays at higher levels of government, local officials can exert their functions with a freedom officials at higher levels of government could only dream of.

3.2. The competition-argument

The second line of argumentation that scholars often use to sustain their argument that decentralization reduces corruption builds on the idea of interjurisdictional competition. From this point of view, jurisdictions compete for capital (investors) and labor (citizens), putting "subnational authorities in direct competition with one another" (Hankla 2008: 636). In its attempt to survive under such circumstances, each jurisdiction (according to the theory) seeks to provide the "maximum quality of government services at the lowest cost" (Hankla 2008: 636), thus reduces corruption.

However, according to Cai and Treisman (2005), an increase in interjurisdictional competition due to an increase in the level of decentralization occurs only in homogeneous countries. In geographically diverse ones, they show, one might even observe decentralization to lead to a reduction of interjurisdictional competition (see Cai and Treisman 2005: 20).


[1] For a more pessimistic point of view see Bukovansky 2006.

[2] No initial author could be found.

[3] A decrease of corruption due to decentralization is also stated by e.g. Arikan (2004), Falleti (2005), Bardhan and Mookherjee (2006), and Vérón et al. (2006).

[4] An increase of corruption due to decentralization is also stated by e.g. Azfar et al. (2001), Lederman et al. (2005) and Kunicová and Rose-Ackermann (2005).


ISBN (eBook)
ISBN (Book)
File size
561 KB
Catalog Number
Institution / College
Johannes Gutenberg University Mainz – Institut für Politikwissenschaft
decentralization corruption comparative politics empirical analysis



Title: Decentralization and Corruption