Corruption and Legitimacy: The Problem of Causality

The purpose of this paper is to critically examine the causal relationship between legitimacy and corruption. More specifically, to focus on the under-researched issue of whether acceptance of legitimacy has an impact on the tendency to take part in corrupt activity. The correlation between the absence of corruption and legitimacy is well established. However, causality can be debated. The standard interpretation is that corruption undermines legitimacy. Nonetheless and importantly, theories of legitimacy assume that legitimacy is crucial because of its ability to persuade citizens to behave in a law-abiding manner, suggesting that legitimacy may curb corruption. Thus, corruption is often conceived of as a circular problem, operating in either vicious or virtuous circles. Corruption may undermine legitimacy which in turn aggravates corruption; and conversely, successful anti-corruption reforms may boost legitimacy and trust in government, which in turn reduces corruption. For anti-corruption reformers this poses the problem of where the circle begins. Can we start just about anywhere, or must we go beyond the circle allegory? Rothstein suggests that a ‘big bang’ approach may be appropriate to understand the formation of virtuous circles: anti-corruption is unlikely to succeed in a piecemeal manner but needs to proceed on a massive scale in order to win credibility and acceptance. However, this still leaves scope for ambiguity concerning the causal relationship between legitimacy and corruption: Is legitimacy a prerequisite for successful anti-corruption reform or can reforms succeed even in the absence of legitimacy? In this paper, we aim to establish the importance of legitimacy in accounting for compliance (here: the decision not to take part in corruption). To do so, we move beyond simple correlations and address the causal relationship. We do so by employing survey experiments. By treating participants to different manipulations, we examine the effects of variable conditions and hence conclude how such variations affect our dependent variables. Our main dependent variable is intended compliance, i.e. the decision to take part or not to take part in corrupt activity. Respondents are treated to different versions of a vignette dealing with the allocation of service flats which affects a close relative. If legitimacy theory accounts for willingness to abstain from corruption, we expect features of the decision-making process on allocations to play a role (e.g. input legitimacy, procedural legitimacy etc.). However, we will also consider how perceptions of compliance by others affects the dependent variable, irrespective of acceptance of legitimacy claims. The research is conducted among a panel of random participants in Iceland, an affluent country generally characterized by relatively little corruption and high legitimacy. This is helpful to establish a context where corruption and legitimacy are not basic survival issues and can be considered from the vantage point of incremental improvements or regression.