An empirical study on the nature of corruption amongst Nigerian firms: causes, channels and detection

This dissertation discusses the economic issues surrounding corruption at the firm-level in Nigeria with a specific focus on bribery. This involves an analysis of the paying and reporting of bribes by firm managers in Nigeria. The first chapter uses data from two business surveys to explain the determinants of the incidence of bribery and the magnitude of bribes, respectively. A two-stage analysis is conducted to test for the independence of the processes determining the incidence and the magnitude of bribery. The results show that the propensity to bribe is determined by required meetings with public officials while the size of bribe is driven by firm profitability indicators. The second chapter tests the reliability of methods used to ask individuals sensitive questions on different forms of business malpractice. Indirect methods are tested against the randomised response method. The indirect method protects the managers from stigmatisation by asking them about the behaviour of an agent representative of themselves; the randomised response method asks the interviewee to base their response on the result of a private coin-toss. The results show that the indirect method produces higher and more plausible estimates of wrongdoing than the andomised response method. The third chapter investigates why the randomised response method sometimes fails in eliciting honest responses from sensitive questions despite assuring the managers of anonymity. The roles of trust in the interviewer and the probability of detection are considered along with other potential explanations. Results indicate that lack of trust and the fear of detection are associated with underreporting of sensitive acts. The final chapter examines the relationship between bribery and ethnic networks. The ethnicities of the managers and their local political representatives are used to measure ethnic networks. Results show that co-ethnic firm managers are less likely to pay a bribe than non co-ethnics. Also, there is a positive association between ethno-linguistic fractionalisation and bribery which, in fractionalised areas, eradicates the negative effect of co-ethnicity on bribery