Three experiments on corruption
thesisposted on 02.03.2017, 03:10 authored by Wu, Kevin Xin Chi
This thesis comprises three papers investigating the mechanisms surrounding the exchange of petty bribes prevalent in many parts of the developing world. The first two papers utilise laboratory experiments to investigate the efficacy of a self-reporting mechanism in deterring the exchange of a petty bribe. In the final paper I conduct a natural field experiment in New Delhi, India where we directly observe bribes being requested. In the first paper my co-author and I investigate whether offering a monetary reward to whistle-blowers who report their own involvement in petty bribery is able to destabilise trust and deter agents from offering or accepting a bribe. The mechanism is not dissimilar to laws instituted in the United States and EU that extend full or partial leniency to whistle-blowers who report on their organisation’s involvement in forming a cartel. We conduct a laboratory study to test this mechanism and find the mechanism approximately halves the number of bribes being exchanged. This outcome arises despite the participants engaging with the same partners over several periods and potential payouts being significantly higher for agents to exchange bribes and favours over several periods without self-reporting. In the second paper I investigate whether the monetary reward for whistle-blowing is as effective in scenarios where agents do not anticipate interacting with one another repeatedly. This is more reflective of many interactions where individuals pay bribes to public officials but even if bribery is routine they seldom interact with the same public official repeatedly. As trust is more difficult to establish in such scenarios the base level of bribery exchange was anticipated to be lower which allowed the possibility that a reward for whistle-blowing would not further erode trust. The results reveal the whistle-blowing mechanism again approximately halves the level of bribes being exchanged despite a lower baseline incidence of bribes. In the final paper my co-authors and I directly approach notaries public in New Delhi and request they attest a copy of a document as being genuine. In a control treatment we check whether prices above the legal rate are demanded for legitimate requests. In further treatments we request the attestation be backdated (mildly illegitimate request) and certified when the copy is shown to be augmented from the original (highly illegitimate request). We observe only a low degree of overcharging in the control treatment but substantial bribes being demanded in the backdating treatment. About half of the notaries refuse to attest the augmented document but those who consent demand significantly higher bribes. We also study the effect of competition by comparing notaries at court (high competition) and in isolated home offices (low competition). Surprisingly, notaries with high competition demand higher bribes. We attribute this effect to the crowding out of altruism by competition.