Comparative studies of corruption focus on the selection and incentives of policymakers. With few exceptions, actors who are in charge of implementing policies have been neglected. This article analyzes an original data set on the bureaucratic features and its effects on corruption in fifty-two countries. Two empirical findings challenge the conventional wisdom in literature. First, certain bureaucratic factors, particularly meritocratic recruitment, reduce corruption, even when controlling for a large set of alternative explanations. Second, the analysis shows that other allegedly relevant bureaucratic factors, such as public employees' competitive salaries, career stability, or internal promotion, do not have a significant impact.
Corruption, Bureaucracy, Meritocratic recruitment, Public administration
Jesus Gil, the mayor of Marbella, Spain, from 1991 to 2003, replaced a professional bureaucracy with political appointees. The result of this absence of bureaucratic checks of elected officials' activities was a corruption network in which some individuals accumulated hundreds of millions of euro. By contrast, during the summer of 2009 the politically appointed county governor in Gotland, Marianne Samuelsson, granted permission to an important businessman to build by the seashore, which is normally protected by law in Sweden. Samuelsson justified her decision by referring to the businessman's importance. Her actions were exposed by a professional bureaucrat in Samuelsson's staff, and media accused her of setting aside the fundamental legal principle of impartiality. As a consequence, Samuelsson was forced to resign.
These examples illustrate the important role professional bureaucrats may play in curbing corruption. The actions and choices of public employees have nevertheless often been overlooked in the comparative literature on corruption. This article argues that the employment terms of public employees and, in particular, the extent to which they are dependent on their political masters are essential for understanding why some states have been able to establish noncorrupt institutions while others are stuck with corruption and bad government.
The current literature on deterrents of corruption focus mainly on the political side of the state, for example, the effect of democracy, electoral systems, or veto players (Andrews and Montinola 2004; Keefer 2007; Montinola and Jackman 2002; Persson, Tabellini, and Trebbi 2003; Treisman 2000). There are however only a few examples of studies that consider the bureaucratic side of the state and its role for controlling corruption (e.g., Rauch and Evans 2000).1 Moreover, despite indications that both political and bureaucratic factors have an impact on controlling corruption, they have rarely been tested together.
This article first develops three hypotheses taking the bureaucratic side of the state into account and second, aims at bridging the gap between the two institutional approaches by testing which political and bureaucratic factors do matter for curbing corruption. To fulfill the second goal, this article uses a unique data set based on a survey completed by 520 experts from fifty-two countries, which, to the best of our knowledge, represents the most encompassing data set on bureaucratic structures at the cross-country level.
While the previous literature on bureaucracies relies heavily on socialization of bureaucrats into well-functioning units (e.g., Rauch and Evans 2000), this article suggests a more institutional mechanism, namely, the separation of careers between politicians and bureaucrats. In short, we argue that whereas Marianne Samuelson in Gotland was checked by a professional body of public servants, Jesus Gil in Marbella was able to coordinate his corrupt intentions with the bureaucrats he had himself selected. This mechanism does not require any assumptions about higher competence, higher morals, or in any way a "better" nature of professional bureaucrats as compared to political appointees. …