The untold story of the modernization thesis: Urbanization and corruption in developing countries

Abstract

How does urbanization affect corruption? Modernization theory suggests a negative relationship. Existing empirical studies tend to confirm this hypothesis, showing that urbanization is inversely associated with corruption. In this paper, I provide an alternative perspective on corruption, focusing especially on Sub-Saharan Africa. I argue that the modernization process, ironically, can exacerbate rather than ameliorate corruption. Urbanization is conducive to corruption in an African context because urbanization is characterized by a more individualistic lifestyle reducing thus the cost of being corrupt as there are fewer social sanctions and peer pressure compared to community-based traditional life. A time-series cross-sectional analysis (1972−2015) shows across several regression models and estimators that urbanization is directly associated with corruption in a sample of African countries, but it has mixed effects on a global sample. Re-conceptualizing how urbanization shapes political culture in Africa is important for a continent that is rapidly urbanizing, working to control corruption, and in need of further development.

Keywords

Modernization urbanization corruption development Africa

Introduction

Very few theories in the social sciences have amassed scholarly attention as has modernization theory. Considering the amount and quality of work building on it (Acemoglu et al., 2009; Boix and Stokes, 2003; Inglehart and Welzel, 2010; Kennedy, 2010; Lipset, 1959; Przeworski and Limongi, 1997), modernization theory is arguably one of the most influential theories in political economy. Part of the influence is that its central claim is simple: economic development is associated with mostly predictable changes in political, social, and cultural life (Lipset, 1959; Inglehart and Baker, 2000). Being one of the most prominent signs of modernization (Huntington, 2006: 32), urbanization is praised and encouraged by development experts in developing countries who see it as a path to economic development as it encourages and supports the transition from a primitive agricultural economy to a more sophisticated economy that relies on industries and services (Brett, 2009; Davis and Henderson, 2003; Njoh, 2003; OECD, 2015; Quigley, 2009).

However, instead of promoting economic development, developing countries in Africa are the most corrupt in the world and happen to be urbanizing at the fastest and higher rates (Fox, 2013; Svensson, 2005). Is there a connection between urbanization and corruption? Despite the extensive literature that exists on the determinants of corruption (Serra, 2006; Treisman, 2000, 2007), we have a limited and unclear understanding of how urbanization influences corruption. A recent review of the determinants of corruption by Dimant and Tosato (2018) shows that we have contradictory findings. Theoretically, scholars have paid little attention to the relationship between the two variables, albeit the modernization story suggests that the relationship should be negative because urbanization is part of the social changes engendered by modernization, which taken together are conducive to democratization and democratic values (Lipset, 1959; Przeworski and Limongi, 1997) and by extension less corruption.

The empirical research, however, largely is made up of country-specific studies that mostly address the question using corruption conviction rates in the United States (Goel and Nelson, 2011; Meier and Holbrook, 1992). Important country-specific works exist outside of the United States as well but with more oriented analysis in a domain such as urban planning (Server, 1996; Quesada and Sanchez, 2017). The few cross-country empirical findings show that urbanization is indeed associated with lower levels of corruption (Billger and Goel, 2009; Goel and Nelson, 2010). This conclusion is questionable because the existing studies assume unit homogeneity between the developed and developing countries and offer limited theoretical expectations as to why urbanization should reduce corruption, especially in a developing country context.

In Africa and arguably across the developing world, the inverse relationship between urbanization and corruption is unlikely to hold, and instead, urbanization is likely to increase corruption in these countries. With relatively weak political institutions, and most undergoing a massive urbanization process, although still about half urban and half rural (World Bank, 2017), the process of urbanization in Africa is much different than it is in the world's developed countries especially the close connection between the lifestyle and urbanization. As such, one should not expect the impact to be the same in these two groups of countries especially given that developed countries have a longer urbanization history and stronger political institutions. Urbanization, as I explain, will cause more corruption in the developing world.

To test these arguments, I use a time-series cross-sectional analysis of a broad number of countries and a sub-sample of 51 African countries over a period of 44 years. The results show that urbanization is negatively associated with corruption as suggested by modernization theory when unit homogeneity is assumed in the analysis, although this effect disappears with fixed effects estimates. However, the findings also suggest that urbanization is positively associated with corruption with a sample based on African countries. This theoretically and empirically contradicts existing cross-country studies.

This paper makes at least two contributions to the literature on development broadly defined. First, it shows that our understanding of corruption and development is still victim to simplistic modernization mythology, by showing that modernization does not always create a virtuous circle as suggested by existing findings. By encouraging urbanization as part of the development process and weakening social bonds embedded in traditional life, modernization contributes to more corruption in the developing world and is in that sense an impediment to the objective it promotes in the first place. It, therefore, calls policymakers to pay greater attention to the phenomenon of urbanization and its effects on development outcomes.

Second, this study furthers our understanding of the determinants of corruption in the developing world, which is especially important as one of the main causes of economic stagnation in these countries is corruption (Gyimah-Brempong, 2002; Mauro, 1995). To address this problem, it is imperative to first understand the causes of corruption, and here this analysis enhances our understanding of corruption showing that urbanization is an important part of how we think about and understand corruption in Africa and other parts of the developing world.

The rest of this paper is structured around four sections. First, I review the literature tying modernization and urbanization to corruption. Second, I present a theoretical discussion of the social, institutional, and political conditions under which urbanization leads to more corruption in the developing world, with a focus on Africa. Next, I introduce and describe the data I use to test my hypothesis—a panel data from 1972 to 2015—considering a number of control variables both descriptively and in multivariate models. Finally, the last section concludes the study, discusses the implications of the findings, and proposes directions for future research.

Urbanization and corruption

It is difficult to find studies with fully fleshed-out theories explaining the impact of urbanization on corruption. This is surprising because following modernization theory's central claim, one assumes the relationship to be negative (Inglehart and Baker, 2000; Inglehart and Welzel, 2010; Lipset, 1959), even when there are not many direct empirical findings to corroborate this relationship. Urbanization is generally seen as desirable because it is part of the modernization process that brings mainly good political, social, and economic outcomes (Boone, 2012; Brett, 2009; Davis and Henderson, 2003; Fox, 2013; Lewis, 1954; Njoh, 2003; Pye, 1969: 401; Quigley, 2009). Huntington (2006: 59) is one of the first scholars to explicitly address the possible connection between the two variables. Specifically, he argues that the social changes that come with the urbanization process, may weaken social norms, create new sources of wealth, and lead therefore to more corruption. However, Huntington addresses the issue with a broad conceptualization of modernization and without providing empirical evidence to support his claim.

Aside from Huntington, it is hard to find explicit studies theorizing about the impact of urbanization on corruption with a few exceptions including Billger and Goel (2009), Goel and Nelson (2010), and Meier and Holbrook (1992). Of these existing studies, only a few mention urbanization and do so mainly as a control variable providing limited if any discussion about the relationship (Alt and Lassen, 2003; Benito et al., 2015; Glaeser and Saks, 2006). Among the studies theorizing about urbanization mentioned above, only Meier and Holbrook (1992)¹ uses urbanization as a central variable to explain corruption. But this study—as many others²—is country-specific, addressing features of political corruption that are specific to the American political context (Meier and Holbrook, 1992: 138). For a more thorough assessment of the state of our knowledge on the question, however, it seems more logical to look at the cross-country studies that take a more systematic approach.³

The most empirically rigorous studies analyzing the impact of urbanization on corruption reach the same conclusion, such that it is reasonable to call this conclusion conventional wisdom. In fact, the two studies taking a systematic approach in analyzing the impact of urbanization on corruption, Billger and Goel (2009) and Goel and Nelson (2010), find that urbanization decreases corruption. In a cross-sectional OLS regression (n = 98) supplemented by a quantile regression, Billger and Goel (2009) find that urbanization is negatively related to corruption. They argue that this finding can be explained by the fact that urban concentration deters corruption via government oversight or social stigma attached to corruption practices. Their sample is very limited (2001 − 2003) with no explanation why they chose this period. Further, the mean of urbanization in their sample is about twice the mean in Africa (i.e. 59.98% vs. 34.47%) signaling that more developed countries dominate their sample.

Using a random-effects model, Goel and Nelson (2010) offer a similar finding with a greater emphasis that urbanization “strongly” reduces corruption. Urban concentration, they argue, makes corrupt practices easier to detect, which deters potential corrupt acts. Yet again their sample is similar to the one of Billger and Goel (2009), with a slightly different period under consideration. With these two empirically rigorous studies reaching the same conclusion, it is thus not an exaggeration to say we are left with conventional wisdom that urbanization reduces corruption.

Nevertheless, there are several caveats in the existing arguments and assumptions that may make them irrelevant in a developing country context. To this point, consider Goel and Nelson’s (2010) claim that corrupt acts would be easier to stigmatize in urban settings and Billger and Goel’s (2009) conclusion that urban concentration deters corruption via government oversight. These seem to be unrealistic in contexts such as those found in African countries. In fact, in an African context corruption will be relatively easier to stigmatize in a rural setting because community life is inherently more accountable due to its small size and ethnic connections, and more difficult in the city because of the impersonal and individualistic lifestyle (Ikuenobe, 2006). Furthermore, government oversight over corrupt practices in African countries is non-existent because government institutions and democracy are not strong enough for effective control. It is clear that these existing explanations fail to account for the urbanization-corruption dynamics in developing countries.

Building on these points, this study represents a significant contribution to filling the gap in our understanding of the impact of urbanization on corruption by showing that the unit homogeneity assumption is not reasonable, and by providing new explanations and expectations about the relationship in a developing country context. Doing so also contributes to the broader literature on modernization and its effects, and the determinants of corruption. In the next section, I theorize about the effects of urbanization on corruption in the developing world, focusing on Africa.

Urbanization and corruption in Africa

Corruption is commonly defined as the misuse of a public position for private gain (Bardhan, 1997; Rose-Ackerman, 1975). Yet as Helman (2007) notes, culture is the lens through which we understand and interpret events, and therefore corruption should take contextual differences into account, especially factors that influence the costs and benefits associated with it. Following Johnston (1996: 331 − 334), corruption here is “the abuse, according to the legal or social standards constituting a society's system of public order, of a public role or resource for private benefit.” This definition accounts for the fact that informal rules, social norms, and preferences may be different from one context to the other (Bontis et al., 2009; Bukuluki, 2013; Torsello and Venard, 2016).

This is especially important because urbanization and corruption in Africa are certainly different processes when compared to what happens elsewhere, specifically in the developed world. In addition to the high executive embezzlement and bribery, corruption in Africa tends to be public and petty such that it is easily seen and usually a low level in terms of the amount exchanged. Many first-time visitors to Africa are often surprised to see their taxi or car pulled over by a police officer in which a quick handshake takes place, a process that is unusual if not never seen in developed countries.

Likewise, it is important to understand how urbanization and life in an urban area differ from life in a rural area. In developed nations, material success may come with moving away from the city while in developing countries, modernization and success mean moving to the city. This reality leads individuals to associate each setting with specific expectations. Ekeh’s (1975) work on this point is especially important and helps us to better understand the dynamics of corruption in Africa in the era of modernization. Individuals, according to Ekeh (1975), establish a clear distinction between two political systems, the primordial public, life in the community, in which morality is expected, and the civic public in which the individual is incentivized to be amoral with financial gain being the prime motive. Within each sphere, there are costs and benefits associated with behavior among peers, and as I explain, this makes corruption more attractive in one space than the other. This distinction is closely intertwined with the process of urbanization as it significantly impacts the costs and benefits associated with corruption. I turn next to explain how each setting affects incentives to be corrupt.

Corruption in rural areas

Many scholars believe that corruption is part of the African culture, with gift-giving being the example used most often to justify this allegation (De Sardan, 1999; Egbue, 2006). However, as Ayittey (2018) notes, this is a “confusion and bastardization of the traditional practice.” Life in rural Africa, away from modernization, has strong mechanisms of social control for the leaders and populations against deviant practices, including corruption. Norms of social exclusion and other measures of punishment are easy to enforce and make corruption a costly activity. Ayittey (2018: 195) supports that historical evidence suggests that Africans make the difference between gift-giving and corruption even in the modern days. Examples of rebellions that destroyed dynasties and removed chiefs from office for corruption exist (Ayittey, 2018: 195; Diop, 1988: 65). Africa is therefore not intrinsically corrupt as usually claimed.

Life in rural Africa is lived under well-defined traditions in geographically delimited clans (Ayittey, 2006; Ekeh, 1980: 15; Epstein, 1967). In a rural community, the individual has a duty to make traditions endure because they represent their life, and therefore a powerful constraint for individualistic endeavors either good or bad. Obligations associated with community life require one to act in the best interest of their family and village more generally. This mode of life has a deep sense of community and reciprocity as captured by Mbiti (1990) when he states “I am because we are. And since we are, therefore I am.” In this environment, individuals strive to be exemplary people who do not deceive others since deception can be consequential for their own well-being and members of their own group. Individuals who are corrupt are easily seen and punished, and they will not be taking bribes from people of their own ethnic group as it is immoral.

Moreover, even if one was successful and able to bribe people, the funds from any corrupt act would not be for the individual's sole benefit. Rather, the expectation is that funds generated from corruption must be used for the community-at-large because each member is expected to contribute to the group's survival (Bukuluki, 2013; Ikuenobe, 2006; Ekeh, 1975). Thus, the ability to use funds from corrupt activities for personal purposes is severely circumscribed in rural areas. In the end, this lowers the benefits of corruption as one must dedicate their funds collected from corruption to others rather than themselves alone.

In rural areas, then, the costs of corruption are quite high as one must either figure out how to extract bribes from their own community members or somehow extract them from outsiders. Collecting from outsiders is the most likely source, but it is likely to be costly as it is unpredictable and inconsistent since outsiders are not that frequent in rural areas. Additionally, on the benefit side, one is going to be forced to share the proceeds from corruption with the community. The costs of corruption are thus high while the benefits are low making it unlikely that people will engage in corruption in rural areas.

Corruption in urban areas

In contrast to the rural setting, urban life alters the costs and benefits of corruption thus making it more appealing. In contrast to rural life, urban life is individualistic such that the group loses its importance in favor of the individual. Albeit, there is a tendency now to see a blending of social norms in rural and urban areas, the factors determining life, such as “self-identification,” are still fundamentally different in the two settings (Epstein, 1967; Stephens, 2015). Interactions in urban areas are usually open, secular, and farther from traditions and religion (Huntington, 2006: 72; Meier and Holbrook, 1992). In such conditions, the cost of corruption goes down in considerable ways. When engaging in corrupt activities, people will see this illicit behavior happening, but the costs are different because in an urban setting life is busy, crowded, and mixed with people from different parts of the country and even the region (Epstein, 1967). African modern states do not have strong formal institutions that can provide an alternative to community checks. Indeed, in an urban setting, individuals are less likely to interact with the same people several times, and the chances that they are from the same ethnic group are considerably lower than in rural areas. Thus, the ex-communication threat is much lower if not nonexistent. Relatively, then, the costs of being corrupt in an urban setting are much lower than they are in a rural setting.

Alternatively, urbanization raises the benefits of corruption. Anonymity in urban areas allows one to freely spend revenues from corruption on oneself without the risk of being stigmatized as in a rural setting. While collecting bribes in an urban area may be easily seen, it will be hard for people seeing the corrupt individual to tell other people so as to force that person to use those funds for a collective group. It is quite likely that one taking bribes in a certain area of the city lives in a distant part making it basically impossible for those seeing the corruption to identify that person as one who has gained money from being corrupt. Without the ability to identify the corrupt individual to the larger community, it will be difficult if not impossible to condition that person to use the funds for the community thus freeing them up to dedicate them to their own personal needs and wants. Urbanization thus raises the benefits of corruption.

In sum, two broad mechanisms explain the positive impact of urbanization on corruption in Africa and by extension developing countries. Urbanization increases corruption by reducing the costs and increasing the benefits associated with bribe-taking and other forms of corruption. Losing the eye of the community allows people to easily engage in corrupt practices without the fear of being punished. Further, individuals can dedicate the resources generated from corruption for their own personal use. Based on this argument, I propose the following hypothesis:

Hypothesis: In a comparison of African countries, those with higher levels of urbanization will have more corruption than those with lower levels of urbanization.

Empirical analysis

Data and sample

To test my argument about the relationship between corruption and urbanization, I use large-n multivariate regression analysis on a large sample of countries. The data are organized as a time-series cross-section with time measured on a yearly basis from 1972 to 2015. I run the analysis with two samples. First, I estimate the models with a global sample of 119 countries. This sample is used to test the conventional wisdom that assumes unit homogeneity. It includes all the countries that have data available for the variables under consideration. Second, I estimate the urbanization–corruption relationship with a sample of 51 African countries⁴to test my theory.

Summary statistics for the developing countries sample is presented in Table 1. The unit of analysis is country-year with observations ranging from a high of 4159 to a low of 742 depending on the control variables and the corruption index used as the dependent variable. A data description in the appendix gives further information about the data sources and other transformations made to the variables. The two samples have important differences in means for both the independent and dependent variables. Urbanization has a mean of 48.10 in the global sample and 34.47 in the African sample (p = 0.000). For corruption, however, Africa takes the lead with a mean of 63.41 while the global sample has a mean of 47.40 (p = 0.000).

Table 1.

Summary statistics (African countries).

	N	Mean	SD	Min.	Max.
Corruption (Vdem)	2186	63.41	20.67	15.22	96.90
Urban percentage	2182	34.47	17.31	3.10	87.15
Log aid	2160	19.02	1.43	9.90	23.16
Log GDP	2077	6.44	1.08	4.12	10.03
Resources % of GDP	2073	11.97	12.54	0	89.17
Trade % of GDP	1961	72.32	44.07	6.32	531.73
British colony	2186	0.38	0.48	0	1
Ethnic fractionalization	2186	0.63	0.26	0	0.95
Democracy (polity)	2012	−2.05	6.11	−10	10
Corruption (ICRG)	1140	3.62	0.99	0	6
Control of corruption (World Bank)	867	0.068	0.61	−1.77	1.32
Population growth	2182	2.54	1.15	−6.76	11.01

Estimation equation

Following Treisman’s (2000) seminal analysis of the determinants of corruption, the core estimable equation for corruption includes democracy, colonizer identity, ethnic fractionalization, trade, GDP per capita, resources export, and multiple additional controls. The general statistical model used in this analysis is presented in the following equation:

\begin{aligned} C o r r u p t i o n_{i t} = & α + β_{1} U r b a n i z a t i o n_{i t} + β_{2} l n (A i d)_{i t} \\ + β_{3} l n (G D P)_{i t} + β_{4} R e s o u r c e s_{i t} + β_{5} T r a d e_{i t} \\ + β_{6} C o l o n i z e r + β_{7} E t h n i c F r a c t i o n a l i z a t i o n \\ + β_{8} D e m o c r a c y_{i t} + ε_{i t} \end{aligned}

The dependent variable—corruption—is captured by the Varieties of Democracy's (V-Dem) political corruption index measured on a scale from 0–1⁵ with lower values meaning less corruption (McMann et al., 2016). The index is based on expert surveys with at least three-fifths of the coders native of the countries they are assigned to code. An item response theory measurement model is afterwards used to aggregate the experts’ responses. In comparison to other indices, especially, the Transparency International's Corruption Perception Index (CPI), the V-Dem political corruption index has several advantages that make it the best index for this study and arguably for academic studies. First, it has an extensive geographical and temporal coverage. The CPI and World Bank control of corruption are only available from 1996 for most African countries, while the V-Dem data starts from 1900. Using the CPI, for instance, makes the analysis lose at least two decades of data.

Second, the V-Dem index has a stronger internal validity in its broader conceptualization of corruption. While most indices measure only bureaucratic and/or public-sector corruption, the V-Dem index goes several steps further to include six indicators resulting from distinct types of corruption either petty or grand that cover several areas of a given country. Third, the V-Dem data minimizes bias. By producing its own original data, the V-Dem reduces the chances for bias in comparison to the CPI or World Bank indices that are “polls of polls,” thus, multiplying sources of bias resulting but not limited to each original source⁶. For these reasons, recent studies tend to use V-Dem data to capture corruption (Carbone and Pellegata, 2020; McMann et al., 2020).

To operationalize urbanization, the main independent variable, I use the World Bank's World Development Indicators measure of the percent of urban population (World Bank, 2017). Based on the theoretical discussion, I expect the estimate to be positive indicating a positive association with corruption.

Building on the estimation equation, I isolate the effect of urbanization on corruption from other confounding factors. Thus, I include several control variables⁷ broadly classified into economics variables, and institutional and social variables.

Economic control variables

To isolate the effect of urbanization from wealth, I include a control variable for GDP per capita as wealthier countries have higher rates of urbanization and lower levels of corruption. I therefore anticipate wealth to have a negative estimate. It represents by far one of the most consistent predictors of corruption and is consistently found to be inversely associated with corruption (Ades and Di Tella, 1999; Serra, 2006; Svensson, 2005; Treisman, 2000).

In addition to GDP per capita, I include the availability of natural resources, especially oil rents as a percentage of GDP (β₄ in the equation). This is suggested by Ades and Di Tella (1999) who argue that natural resources give greater access to rents for bureaucrats to capture. The expectation is that it has a positive effect on corruption.

I also include trade and aid as part of the economic control variables. The trade variable emerges from the neo-liberalism literature which suggests that it positively influences growth. Treisman (2000) finds a negative correlation between GDP share of imports and corruption. Accordingly, I expect it to have a negative relationship with corruption. In contrast, I expect aid to have a positive effect on corruption. African countries receive large inflows of international aid, and extensive evidence supports that unearned cash helps non-democratic regimes maintain their power and fosters corruption (Ahmed, 2012; Asongu, 2012).

Institutional and social control variables

Following Treisman (2007) and the role of political institutions in the control of corruption, I include democracy and colonial legacy in the regression model. Democracy is measured by the polity IV score, and this variable ranges from a negative 10 to a positive 10 with lower values indicating more autocratic regimes and higher values more democratic regimes. The Polity score is one of the most widely used democracy indices in the literature (Marshall et al., 2013). This is an important control as democracy is thought to have a positive influence on accountability through the mechanisms of checks and balances. Drury et al. (2006), for instance, find that corruption has a large negative effect on non-democracies’ economic growth, while it has no effect on democracies because electoral competition deters political leaders from engaging in damaging corrupt activities. Moreover, democracy is also correlated with urbanization (Lipset, 1959), although some studies find that dictatorships are associated with urbanization as they promote large dominant single cities (Ades and Glaeser, 1995). The political regime is thus an important confounding factor that should be controlled for in the regression analysis.

Colonial legacy is thought to be an important determinant of institutional quality. Specifically, Britain is thought to have a legacy of more transparent and democratic institutions as compared to other colonizers (Acemoglu et al., 2001; Serra, 2006; Treisman, 2000), albeit empirical evidence shows that the effect is not particularly large. Based on the literature, one shall expect less corruption in ancient British colonies in comparison to French and other European colonies.

Finally, I include ethnic fractionalization as a control variable based on the coding of Fearon and Laitin (2003). In fact, it is established theoretically, and empirically that a country that has a high ethnic fractionalization is subject to higher levels of corruption because the cost of engaging in corruption and for enforcing corrupt contracts is reduced (Easterly and Levine, 1997). It, therefore, represents an important control variable for this analysis.

Estimator

To determine the appropriate methods to estimate my statistical models, I conducted several preliminary statistical tests. A unit root test using the Augmented Dickey–Fuller test shows no sign of unit roots in my main variables of interest, which means the series is stationary and there is thus no need to take first differences. Further tests on the data suggest that there is autocorrelation, heteroskedasticity, and group-wise heteroskedasticity. These problems relate to the fact that errors for units at different times are not independent of one another, errors in panels do not have the same variance, and errors are not independent across units. Ordinary Least Squares (OLS) under these circumstances would likely produce biased results (Beck and Katz, 1995). The most appropriate models for my empirical test are therefore panel-corrected standard errors (hereafter PCSE) (Beck and Katz, 1995), fixed effects, Driscoll and Kraay estimator (Hoechle, 2007), and generalized least squares (Greene, 2003). Addressing the issues mentioned earlier, I use PCSE with the options to fix autocorrelation and panel-level heteroskedasticity as my main estimation technique. Nevertheless, other estimation techniques mentioned are also used as estimators for robustness checks.

Results

To begin, I consider the relationship between urbanization and corruption in a global sample. Here, the objective is to test modernization theory's prediction that the relationship between urbanization and corruption should be negative. Table 2 presents the global sample that considers all the countries that have data, including Africa. The results show mixed effects for the modernization hypothesis. In the models using PCSE and Driscoll and Kraay estimators, the results are consistent with the existing literature, namely that urbanization decreases corruption (Billger and Goel, 2009; Goel and Nelson, 2010). The relationship is statistically significant at the 95% confidence level in both models. However, once I control for country-specific effects, the relationship disappears and the sign turns out to be positive, although it is not significant. This suggests that the existing findings warrant further investigations.

Table 2.

Effect of urbanization on corruption (global sample).

	(1)	(2)	(3)
	Panel-corrected	Drisc/Kraay	FE robust stand errors
Urban percentage	−0.190***	−0.0927**	0.209
	(0.0335)	(0.0416)	(0.136)
Log aid	0.440***	3.602***	2.106***
	(0.118)	(0.268)	(0.540)
Log GDP	−0.624*	−2.882**	−1.629
	(0.355)	(1.240)	(1.214)
Resources % of GDP	0.0321***	0.455***	0.0907
	(0.0114)	(0.0410)	(0.0703)
Trade % of GDP	−0.00463	−0.0252	−0.0230
	(0.00590)	(0.0155)	(0.0252)
British colony	−13.85***	−15.07***
	(1.413)	(0.723)
Ethnic fractionalization	16.87***	8.221***
	(2.313)	(2.059)
Democracy	−0.167***	−0.314***	−0.238
	(0.0320)	(0.0587)	(0.187)
Constant	60.41***	14.85*	23.22**
	(3.252)	(8.590)	(11.04)
Observations	4159	4159	4159
R-squared	0.487	0.298	0.060
Number of countries	119	119	119
Country FE	No	No	Yes

Notes: Standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1.

Perhaps including the African countries in the sample biases the results against the modernization hypothesis. To address this point, I estimate the same models without African countries in the global sample (Supplemental Table 8 in the appendix). The effect is even stronger with a 1% increase in urbanization causing nearly 1/2 of a percent decrease in corruption. Again, once I control for country-specific effects, the relationship disappears.

Based on the results displayed in Supplemental Tables 2 and 8, one may conclude that urbanization has a negative effect or no effect at all on corruption, depending on the model specification or estimator. However, this conclusion assumes unit homogeneity which is unrealistic especially in Africa. Addressing this point, I now test my theory that ceteris paribus African countries with higher levels of urbanization will experience higher levels of corruption. I start my analysis by looking at the models displayed in Table 3 that includes all countries of the African continent except South Sudan and Somalia that do not have enough data points as mentioned earlier. The urban population percentage coefficients point to the direction of my expectations, namely, positive coefficients that are statistically significant at the 99% confidence level in models 4 and 5, and at the 95% confidence level in model 6.

Table 3.

Effect of urbanization on corruption in African countries (including North Africa).

	(4)	(5)	(6)
	Panel-corrected	Drisc/Kraay	FE robust stand errors
Urban percentage	0.252***	0.378***	0.444**
	(0.0489)	(0.107)	(0.195)
Log aid	−0.0386	0.868***	1.530
	(0.188)	(0.242)	(1.067)
Log GDP	−0.808	−3.803***	−2.508**
	(0.495)	(1.111)	(1.239)
Resources % of GDP	0.0645***	0.618***	0.114
	(0.0156)	(0.0465)	(0.0885)
Trade % of GDP	0.000592	−0.0678***	−0.0497**
	(0.00518)	(0.0124)	(0.0233)
British colony	−5.544***	−1.789
	(1.733)	(1.141)
Ethnic fractionalization	19.67***	8.623**
	(3.061)	(3.340)
Democracy	−0.182***	−0.572***	0.139
	(0.0587)	(0.0635)	(0.309)
Constant	51.00***	51.40***	38.88**
	(4.879)	(9.658)	(16.45)
Observations	1799	1799	1799
R-squared	0.585	0.284	0.142
Number of countries	49	49	49
Country FE	No	No	Yes

Notes: Standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1.

On average, a 1% increase in urbanization is associated with at least a 1/4 to a nearly 1/2 of a percent increase in corruption depending on the estimator. The PCSE model shows a 0.25% increase in corruption for every 1% increase in urbanization. Figure 1 graphs this effect with a margins plot. The Driscoll and Kraay estimator in model 5 suggests that for every 1% increase in urbanization, corruption will increase by nearly 0.38%. The coefficients on the fixed effects model (model 6) are the largest in effect with every 1% increase in urbanization leading to 0.44% increase in corruption. Model 6 is probably less reliable for the coefficients since sometimes invariant variables are omitted from it. However, it does show that the effect of urbanization on corruption is still positive and robust even when one controls for country-specific effects.

Figure 1.

Predicted impact of urbanization on corruption.

The control variables included in the models behave as predicted for the most part with the PCSE, and Driscoll and Kraay estimators showing the most consistent results. Looking first at the economic control variables, GDP per capita shows a negative effect in models 4, 5, and 6. Albeit statistically insignificant in model 4, it shows a strong effect in the other two models thus confirming previous findings (Serra, 2006; Svensson, 2005; Treisman, 2000). International aid shows a mixed effect with only model 5 being consistent with the prediction that it positively affects corruption. This is not surprising especially that there is a renewed debate about the direction of the effect (Asongu, 2012; Okada and Samreth, 2012). Exports of natural resources as a percentage of GDP positively influence corruption. The effect is statistically significant at the 99% confidence level in models 4 and 5 and insignificant in model 6. Finally, trade has a negative effect and significant at the 99 and 95% confidence level in models 5 and 6, respectively.

Considering the institutional and social variables, the results show that democracy has a strong negative effect in models 4 and 5. The sign turns positive in the fixed effects estimation, however (model 6). This mixed finding is somehow consistent with the contemporary debate around the effect of democracy on corruption (McMann et al., 2020; Rock, 2009; Treisman, 2000).

Being a former British colony negatively affects corruption, which is consistent with the findings of Serra (2006) who presents this variable as one of the most stable predictors of corruption. Ethnic fractionalization shows a consistently strong positive effect statistically significant at least at the 95% confidence level in models 4 and 5 as previous studies find (Serra, 2006).

In Table 4, I estimate the same models excluding the 5 countries of North Africa (Egypt, Morocco, Tunisia, Algeria, and Libya). The results are unchanged, with urbanization being statistically significant at the 99% confidence level for the Panel-corrected model, at the 95% confidence level for the Driscoll and Kraay estimator, and the fixed effect robust standard errors model. The goodness of fit is much better in this restricted model, and it appears that there is not a significant difference between North Africa and sub-Saharan Africa when it comes to the effect of urbanization on corruption. The control variables included behave the same way as in the main models. International aid and resource rents have a positive and statistically significant effect on corruption as predicted by the literature. GDP and trade have the opposite effect on corruption and are statistically significant, which is consistent with my expectations. Being a British colony seems to have a strong negative effect on corruption, which counters Treisman's (2000, 2007) findings suggesting a weak relationship. Ethnic fractionalization has the sign predicted in the literature (positive), and the estimate is statistically significant.

Table 4.

Effect of urbanization on corruption in African countries (excluding North Africa).

	(7)	(8)	(9)
	Panel-corrected	Drisc/Kraay	FE Robust Stand Errors
Urban percentage	0.217***	0.300**	0.436**
	(0.0521)	(0.113)	(0.198)
Log aid	−0.125	0.0109	1.923
	(0.212)	(0.407)	(1.154)
Log GDP	−0.991*	−3.420***	−3.221***
	(0.527)	(1.113)	(1.191)
Resources % of GDP	0.0690***	0.639***	0.137
	(0.0172)	(0.0519)	(0.0965)
Trade % of GDP	0.00157	−0.0638***	−0.0560**
	(0.00555)	(0.0109)	(0.0221)
British colony	−7.605***	−3.337***
	(1.701)	(1.140)
Ethnic fractionalization	24.40***	14.84***
	(3.536)	(1.178)
Democracy	−0.157**	−0.453***	0.145
	(0.0614)	(0.0873)	(0.313)
Constant	51.92***	63.56***	36.49**
	(5.253)	(13.71)	(17.70)
Observations	1606	1606	1606
R-squared	0.604	0.302	0.159
Number of countries	44	44	44
Country FE	No	No	Yes

Notes: Standard errors in parentheses ***p < 0.01, **p < 0.05, *p < 0.1.

Robustness checks

To check the robustness of the findings, I estimate my main models using an alternative measure of corruption, the International Country Risk Guide (ICRG) corruption index. Compiled by the Political Risk Services Group, it is one of the most used indices in the empirical study of corruption. The results (displayed in Table 5) are unchanged and thus robust to using this alternative measure of corruption. Specifically, the urbanization estimates are statistically significant across several specifications including the PCSE and fixed effects robust standard errors.

Table 5.

Effect of urbanization on corruption in African countries (ICRG corruption Index).

	(10)	(11)
	Panel-corrected	FE robust stand errors
Urban percentage	0.0132**	0.0394***
	(0.00564)	(0.0139)
Log aid	0.0272	−0.100
	(0.0261)	(0.0594)
Log GDP	−0.105	0.112
	(0.0646)	(0.144)
Resources % of GDP	0.00280	0.00225
	(0.00224)	(0.0106)
Trade % of GDP	−0.00184	0.00537
	(0.00119)	(0.00357)
British colony	0.0962
	(0.157)
Ethnic fractionalization	0.621*
	(0.332)
Democracy	−0.00621	0.0138
	(0.00640)	(0.0158)
Constant	2.941***	2.914**
	(0.710)	(1.289)
Observations	986	986
R-squared	0.434	0.146
Number of countries	35	35
Country FE	No	Yes

Notes: Standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1.

Following Asongu (2013), there are reasons to believe that the extent to which different governments fight corruption (existing levels of corruption) and population growth may represent confounding factors. Indeed, governments with good corruption control mechanisms may mitigate the negative externality of urbanization underlined in this article. Also, he finds that population growth positively influences corruption in low-income countries. To take these findings into account, I estimate the models controlling for government control of corruption using the World Bank's World Governance Indicators index (Kaufmann et al., 2009), and population growth. The results are reported in Supplemental Tables 6 and 7 respectively.

Including the control of corruption makes the estimates smaller but the effect is much stronger with the estimates being statistically significant at the 99% confidence level in both models (12 and 13). This suggests that the existing control of corruption in a given country is important (Asongu, 2013). Controlling for population growth makes my estimates slightly stronger with bigger coefficients significant at the 99% confidence level.

There is also a possibility that the effect of urbanization on corruption is not immediate and even not captured in the corruption index for that given year. To test this possibility, I allow urbanization to affect corruption with a lag of up to 3 years. The results presented in Supplemental Table 9 (panel corrected) and 10 (fixed effects with robust standard errors) of the appendix confirm that this intuition may have some truth to it. In fact, the further away urbanization is measured, the stronger the relationship between urbanization and corruption is, albeit the change is not that big. If anything, these results corroborate my initial findings. The findings in this paper are therefore robust to alternative model specifications and further theoretical considerations.

Conclusion

In this study, I challenged the conventional wisdom that modernization creates a virtuous circle that ultimately lowers corruption. For a typical African country, urbanization—the most prominent sign of modernization—lowers the costs and raises the benefits of corruption such that it is a contributing factor to corruption. Unlike in most developed countries, the transition from the community-based life in the village to an individualistic lifestyle in the city decreases the cost and increases the benefits of being corrupt because of the significant reduction in peer pressure and associated community checks. This process is especially true in Africa where modern political institutions are fragile and unable to provide a viable alternative to community checks. Using a time-series cross-section analysis across several samples, models, and estimators, I find that urbanization is associated with higher levels of corruption in Africa. These results contradict the existing systematic empirical studies and re-conceptualizes the impact of urbanization on corruption.

While my developing countries sample was only made up of African countries, it is reasonable to argue that these findings can also apply to developing countries in other parts of the world albeit some exceptions may exist given the mechanisms explored here. The collectivist life in African societies is not necessarily the same in other parts of the world. In addition, major events such as slavery and colonization have transformed the continent in significant ways that affect how the modern state is perceived (Ekeh, 1975).

This study makes a number of contributions to further our understanding of modernization outcomes and the determinants of corruption. The traditional rhetoric espoused by development experts has been to advocate for economic prosperity, often praising urbanization (OECD, 2015) as a process that causes economic growth and promotes well-being (Davis and Henderson, 2003; Njoh, 2003; Quigley, 2009). Yet, after decades of incredibly high rates of urbanization, many developing countries, especially those in Africa are still very poor (Beegle and Christiaensen, 2019). Economic growth did not follow urbanization trends as many development experts predicted (Fox, 2013; OECD, 2015). Instead, urbanization seems to work against the major objective it was supposed to promote. This analysis shows that it does so because urbanization increases corruption, which as we know is a major determinant of economic growth and social well-being (Doces, 2020; Gyimah-Brempong, 2002; Mauro, 1995). This finding is consistent with Henderson (2003) who posits that urbanization is not a synonym for development, and also Fox (2013) who shows that economic growth does not follow urbanization in Sub-Saharan Africa as it does in other parts of the world.

The implications of this analysis are especially important for the development policies of western countries and international institutions like the International Monetary Fund, United Nations, and World Bank. Efforts to promote development by these external actors have often looked to increase growth and improve health indicators, as outlined in the UN's Millennium Development Goals, with an emphasis on improved governance as a means to achieving these objectives (Lopez-Calva et al., 2017). However, in encouraging African countries to modernize, these policies might in the process be sowing the seeds of their own demise. By pushing urbanization through growth, they might be unleashing the forces that make corruption more likely and thus undermine the development process they want to promote in the first place. This could be why many African countries have experienced rapid growth over the last twenty years yet extreme poverty remains a serious problem across the region. Understanding the possible side effects of these development programs is especially important if they are to accomplish their goals.

This paper raises the question of whether urbanization causes corruption indefinitely or if it is just a transitional stage. It seems likely that it is more transitional, and as the urban population continues to grow, it will eventually reach the level where many OECD countries are at today. In their early development stages, countries such as the United States have experienced higher corruption levels with urbanization (Meier and Holbrook, 1992). But some of the peculiarities of African countries are the economic conditions, the rate of urbanization, and the institutional setting. In a recent study, Feruni et al. (2020) find that urbanization is conducive to the expected human development in the Western Balkans. While the countries they include in the sample are somehow comparable to those in Africa, they have a meager, stable, and sometimes negative urbanization growth rate, which is not the case for many African countries yet. This study suggests that African countries may be able to take full advantage of the development benefits modernization brings once their urbanization rate stabilizes. Therefore, future research may shed some light on the critical point where urbanization may help curb corruption and improve development outcomes, as in the developed world and the Western Balkans.

Supplemental Material

sj-docx-1-ias-10.1177_22338659221112992 - Supplemental material for The untold story of the modernization thesis: Urbanization and corruption in developing countries

Supplemental material, sj-docx-1-ias-10.1177_22338659221112992 for The untold story of the modernization thesis: Urbanization and corruption in developing countries by Kouakou Donatien Adou in International Area Studies Review

Footnotes

Acknowledgments

I would like to thank my dissertation advisor Dr. Jonathan Krieckhaus, Dr. Jake Haselswerdt, and Dr. John Doces for thorough comments and guidance in writing this article. I would also like to thank all my committee members and anonymous reviewers for their helpful comments.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Kouakou Donatien Adou

Supplemental material

Supplemental material for this article is available online.

Notes

References

Acemoglu

Johnson

Robinson

(2001) The colonial origins of comparative development: an empirical investigation. American Economic Review 91: 1369–1401.

Acemoglu

Johnson

Robinson

, et al. (2009) Reevaluating the modernization hypothesis. Journal of Monetary Economics 56: 1043–1058.

Ades

Glaeser

(1995) Trade and circuses: explaining urban giants. The Quarterly Journal of Economics 110: 195–227.

Ades

Di Tella

(1999) Rents. Competition, and Corruption. American Economic Review 89: 982–993.

Ahmed

(2012) The perils of unearned foreign income: aid, remittances, and government survival. American Political Science Review 106: 146–165.

Alt

Lassen

(2003) The political economy of institutions and corruption in American states. Journal of Theoretical Politics 15: 341–365.

Asongu

(2012) On the effect of foreign aid on corruption. Economics Bulletin 32: 2174–2180.

Asongu

(2013) Fighting corruption when existing corruption-control levels count: what do wealth-effects tell us in Africa? Institutions and Economies 5: 53–74.

Ayittey

(2006) Indigenous African Institutions. Leiden, Netherlands: Brill.

10.

Ayittey

(2018) Applied Economics for Africa. Arlignton, VA: Atlas Network.

11.

Bardhan

(1997) Corruption and development: a review of issues. Journal of Economic Literature 35: 1320–1346.

12.

Beck

Katz

(1995) What to do (and not to do) with time-series cross-section data. American Political Science Review 89: 634–647.

13.

Beegle

Christiaensen

(Eds.). (2019). Accelerating poverty reduction in Africa. Washington, DC: World Bank Publications.

14.

Benito

Guillamón

Bastida

(2015) Determinants of urban political corruption in local governments. Crime. Law and Social Change 63: 191–210.

15.

Billger

Goel

(2009) Do existing corruption levels matter in controlling corruption?. Cross-country quantile regression estimates. Journal of Development Economics 90: 299–305.

16.

Brett

(2009). Reconstructing development theory: International inequality, institutional reform and social emancipation. London, UK: Bloomsbury Publishing.

17.

Boix

Stokes

(2003) Endogenous democratization. World Politics 55(4): 517–549.

18.

Bontis

Bart

Seleim

(2009) The relationship between culture and corruption: a cross-national study. Journal of Intellectual Capital 10(1): 165–184.

19.

Boone

(2012) Territorial politics and the reach of the state: unevenness by design. Revista de Ciencia Política 32: 623–641.

20.

Brett

(2009) Reconstructing Development Theory: International Inequality, Institutional Reform and Social Emancipation. Macmillan International Higher Education.

21.

Bukuluki

(2013) “When I steal, it is for the benefit of me and you”: is collectivism engendering corruption in Uganda? International Letters of Social and Humanistic Sciences 5: 27–44.

22.

Carbone

Pellegata

(2020) Political Leadership in Africa: Leaders and Development South of the Sahara. Cambridge, UK: Cambridge University Press.

23.

Davis

Henderson

(2003) Evidence on the political economy of the urbanization process. Journal of Urban Economics 53: 98–125.

24.

De Sardan

OJ-P

(1999) A moral economy of corruption in Africa? Journal of Modern African Studies 37: 25–52.

25.

Dimant

Tosato

(2018) Causes and effects of corruption: what has past decade’s empirical research taught us? A survey. Journal of Economic Surveys 32(2): 335–356.

26.

Diop

(1988) Precolonial Black Africa. Chicago, IL: Chicago Review Press.

27.

Doces

(2020) Democracy, consumption, and growth in sub-Saharan Africa. International Area Studies Review 23: 28–48.

28.

Drury

Krieckhaus

Lusztig

(2006) Corruption, democracy, and economic growth. International Political Science Review 27: 121–136.

29.

Easterly

Levine

(1997) Africa’s growth tragedy: policies and ethnic divisions. Quarterly Journal of Economics 112: 1203–1250.

30.

Egbue

(2006) Africa: cultural dimensions of corruption and possibilities for change. Journal of Social Sciences 12: 83–91.

31.

Ekeh

(1975) Colonialism and the two publics in Africa: a theoretical statement. Comparative Studies in Society and History 17: 91–112.

32.

Ekeh

(1980) Colonialism and social structure. An Inaugural Lecture .

33.

Epstein

(1967) Urbanization and social change in Africa. Current Anthropology 8: 275–284.

34.

Fearon

Laitin

(2003) Ethnicity, insurgency, and civil war. American Political Science Review 97: 75–90.

35.

Feruni

Hysa

Panait

, et al. (2020) The impact of corruption, economic freedom and urbanization on economic development: Western Balkans versus EU-27. Sustainability, 12(22), 9743.

36.

Fox, S. (2013). The political economy of urbanisation and development in sub-Saharan Africa (Doctoral dissertation, London School of Economics and Political Science).

37.

Glaeser

Saks

(2006) Corruption in America. Journal of Public Economics 90: 1053–1072.

38.

Goel

Nelson

(2010) Causes of corruption: history, geography and government. Journal of Policy Modeling 32: 433–447.

39.

Goel

Nelson

(2011) Measures of corruption and determinants of US corruption. Economics of Governance 12: 155–176.

40.

Greene

(2003) Econometric Analysis. Noida: Pearson Education India.

41.

Gyimah-Brempong

(2002) Corruption, economic growth, and income inequality in Africa. Economics of Governance 3: 183–209.

42.

Helman

(2007) Culture, Health and Illness. Boca Raton, FL: CRC press.

43.

Henderson

(2003) Urbanization and economic development. Annals of Economics and Finance 4: 275–342.

44.

Hoechle

(2007) Robust standard errors for panel regressions with cross-sectional dependence. The Stata Journal 7: 281–312.

45.

Huntington

(2006) Political Order in Changing Societies. New Haven, CT: Yale University Press.

46.

Ikuenobe

(2006) Philosophical Perspectives on Communalism and Morality in African Traditions. Lanham, MD: Lexington Books.

47.

Inglehart

Baker

(2000) Modernization. Cultural change, and the persistence of traditional values Ronald Inglehart; Wayne E. Baker American sociological review 65: 1. Looking forward, looking back: continuity and change at the turn of the millenium. (Feb. American Sociological Review 65: 19–51.

48.

Inglehart

Welzel

(2010) Changing mass priorities: the link between modernization and democracy. Perspectives on Politics 8: 551–567.

49.

Johnston

(1996) The search for definitions: the vitality of politics and the issue of corruption. International Social Science Journal 48: 321–335.

50.

Kaufmann

Kraay

Mastruzzi

(2009) Governance matters VII: aggregate and individual governance indicators 1996–2007. In Peters

Koechlin

Forster

Fenner

(Eds.) Non-State Actors as Standard Setters. Cambridge: Cambridge University Press, 146–188.

51.

Kennedy

(2010) The contradiction of modernization: a conditional model of endogenous democratization. The Journal of Politics 72: 785–798.

52.

Lewis

(1954) Economic development with unlimited supplies of labour. The Manchester School 22: 139–191.

53.

Lipset

(1959) Some social requisites of democracy: economic development and political legitimacy. American Political Science Review 53, 69–105. Available from Jstor as of July 5, 2022 at this link: https://www-jstor-org-443.web.bisu.edu.cn/stable/pdf/1951731.pdf

54.

Lopez-Calva

L.-F

Zhou

Al-Dahdah

, et al. (2017) World Development Report 2017: Governance and the Law. Washigton, DC: The World Bank.

55.

Marshall

Gurr

Jaggers

(2013) Polity IV project: political regime characteristics and transitions, 1800–2012. Center for Systemic Peace .

56.

Mauro

(1995) Corruption and growth. The Quarterly Journal of Economics 110: 681–712.

57.

Mbiti

(1990) African Religions & Philosophy. Halley Court, Oxford: Heinemann.

58.

McMann

Pemstein

Seim

, et al. (2016) Strategies of validation: assessing the varieties of democracy corruption data. V-Dem Working Paper 23.

59.

McMann, Kelly M. and Pemstein, Daniel and Seim, Brigitte and Teorell, Jan and Lindberg, Staffan I., Strategies of Validation: Assessing the Varieties of Democracy Corruption Data (February 1, 2016). V-Dem Working Paper 2016:23, Available at SSRN: https://ssrn.com/abstract=2727595 or http://doi.org/10.2139/ssrn.2727595

60.

McMann

Seim

Teorell

, et al. (2020) Why low levels of democracy promote corruption and high levels diminish it. Political Research Quarterly 73(4): 893–907.

61.

Meier

Holbrook

(1992) 2). “I seen my opportunities and I took ‘em:” political corruption in the American states. The Journal of Politics 54: 135–155.

62.

Njoh

(2003) Urbanization and development in sub-Saharan Africa. Cities (London, England) 20: 167–174.

63.

OECD

(2015) The Metropolitan Century: Understanding Urbanisation and its Consequences. Paris: OECD Publishing.

64.

Okada

Samreth

(2012) The effect of foreign aid on corruption: a quantile regression approach. Economics Letters 115: 240–243.

65.

Przeworski

Limongi

(1997) Theories and facts. World Politics 49: 155–183.

66.

Pye

(1969) The political implications of urbanization and the development process. In Breese

(Ed), The City in Newly Developing Countries: Readings on Urbanism and Urbanization. Englewood Cliffs, New Jersey: Prentice Hall, 401–406.

67.

Quesada

Sánchez

(2017) Corrupt Cities: The Illicit in Local Urban. Development: The Spanish Case. In The Illicit and Illegal in Regional and Urban Governance and Development: Corrupt Places. Milton Park, Oxfordshire: Routledge, 221 − 2237.

68.

Quigley

(2009) Urbanization, agglomeration, and economic development. Urbanization and Growth 115.

69.

Rock

(2009) Corruption and democracy. The Journal of Development Studies 45: 55–75.

70.

Rose-Ackerman

(1975) The economics of corruption. Journal of Public Economics 4: 187–203.

71.

Serra

(2006) Empirical determinants of corruption: a sensitivity analysis. Public Choice 126: 225–256.

72.

Server

(1996) Corruption: a major problem for urban management: some evidence from Indonesia. Habitat International 20(1): 23 − 241.

73.

Spence

Annez

P. C.

Buckley

R. M

. (Eds.). (2008). Chapter 4: Urbanization and growth. Washington, DC: World Bank Publications, 115–132.

74.

Stephens

(2015) The indigenous experience of urbanization. State of the World’s Minorities and Indigenous Peoples 2015: 54–61.

75.

Svensson

(2005) Eight questions about corruption. Journal of Economic Perspectives 19: 19–42.

76.

Torsello

Venard

(2016) The anthropology of corruption. Journal of Management Inquiry 25: 34–54.

77.

Treisman

(2000) 6). The causes of corruption: a cross-national study. Journal of Public Economics 76: 399–457.

78.

Treisman

(2007) What have we learned about the causes of corruption from ten years of cross-national empirical research? Annual Review of Political Science 10: 211–244.

79.

World Bank. (2017, June). World Development Indicators . Retrieved from World Bank Open Data: https://data.worldbank.org/

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.03 MB