Danila Serra

Combining Top-down and Bottom-up Accountability: Evidence from a Bribery Experiment

Monitoring corruption typically relies on top-down interventions aimed at increasing the probability of external controls and the severity of punishment. An alternative approach to fighting corruption is to induce bottom-up pressure for reform. Recent studies have shown that both top-down and bottom-up mechanisms are rarely able to keep service providers accountable. This paper investigates the effectiveness of an accountability system that combines bottom-up monitoring and top-down auditing using data from a specifically designed bribery lab experiment. We compare “public officials” tendency to ask for bribes under: 1) no monitoring; 2) conventional top-down auditing, and 3) an accountability system which gives citizens the possibility to report corrupt officials, knowing that reports lead to formal punishment with some low probability (the same as in 2). The experimental results suggest that “combined” accountability systems can be highly effectives in curbing corruption, even when citizens “voice” leads to formal punishment with a relatively low probability. In contrast, pure top-down auditing may prove ineffective, especially in a weak institutional environment.

Chapter 4 Anticorruption Policies: Lessons from the Lab

We review the existing laboratory experimental studies on corruption that have generated results with clear policy implications. We present and discuss experimental findings on the role that both monetary incentives and nonmonetary motivations may play in corruption decision-making, and, hence, in the fight against corruption.

Corruption, Social Judgment and Culture: An Experiment

We experimentally investigate the extent to which social observability of one’s actions and the possibility of social non-monetary judgment affect the decision to engage in rule breaking behavior. We consider three rule breaking scenarios — theft, bribery and embezzlement — in the absence of any formal enforcement mechanism. By involving a student sample characterized by cultural heterogeneity due to immigration of ancestors to the US, we are able to investigate whether the effectiveness of informal social enforcement mechanisms is conditional on the cultural background of the decision-maker. A total of 52 countries are represented in our sample, ranging from Low Rule of Law countries such as Liberia and Nigeria to High Rule of Law countries such as Sweden and Norway. Our data provide evidence that people with different cultural backgrounds do respond differently to increased social observability of their actions. In particular, while subjects that identify culturally with a High Rule of Law country respond to social obervability and judgment by lowering their propensities to engage in rule breaking, subjects that identify with Low Rule of Law countries do not. Our findings suggest that development policies that rely purely on social judgment to enforce behavior may not work with Low Rule of Law populations. JEL Codes: C90; D73; K42; Z10We experimentally investigate the extent to which social observability of one’s actions and the possibility of social non-monetary judgment affect the decision to engage in rule breaking behavior. We consider three rule breaking scenarios — theft, bribery and embezzlement — in the absence of any formal enforcement mechanism. By involving a student sample characterized by cultural heterogeneity due to immigration of ancestors to the US, we are able to investigate whether the effectiveness of informal social enforcement mechanisms is conditional on the cultural background of the decision-maker. A total of 52 countries are represented in our sample, ranging from Low Rule of Law countries such as Liberia and Nigeria to High Rule of Law countries such as Sweden and Norway. Our data provide evidence that people with different cultural backgrounds do respond differently to increased social observability of their actions. In particular, while subjects that identify culturally with a High Rule of Law country respond to social obervability and judgment by lowering their propensities to engage in rule breaking, subjects that identify with Low Rule of Law countries do not. Our findings suggest that development policies that rely purely on social judgment to enforce behavior may not work with Low Rule of Law populations.

Chapter 1 Experimental Research on Corruption: Introduction and Overview

In Chapter 2, Ananish Chaudhuri surveys the empirical evidence on the existence of gender differences in individuals’ propensity to engage in corruption. While the chapter begins with a review of the findings generated by cross-country studies, the main focus of the discussion is in the insights provided by laboratory experiments specifically designed to test for gender differentials in corrupt transactions. According to the carefully conducted survey of the literature, the existing experimental evidence suggests that females are either equally or less willing to engage in corruption than males; there is very little evidence that women behave more corruptly than men. The author discusses possible reasons for gender differentials in corrupt behavior, such as risk aversion and preferences for reciprocation. Finally, Chaudhuri emphasizes that gender effects are more likely to be observed in studies conducted in developed countries and calls for further research to be conducted in developing countries, with the aim of shedding light on the relationships between gender differences in corrupt behavior and the cultural background of the experimental participants.

Is more competition always better? An experimental study of extortionary corruption

Despite the abundance of theoretical and empirical studies on corruption, identifying suc- cessful anti-corruption strategies remains a challenge. This paper tests the eectiveness of an anti-corruption policy that is often discussed among practitioners: an increase in com- petition among o¢ cials providing the same good or service. In particular, we investigate whether overlapping jurisdictions reduce extortionary corruption, i.e., bribe demands for the provision of services that clients are entitled to receive. We overcome measurement and iden- ti…cation problems by addressing our research question in the laboratory. We conduct an extortionary bribery experiment where clients apply for a license from one of many available o¢ ces and o¢ cials can demand a bribe on top of the license fee. O¢ cials decide whether or not to demand a bribe and the size of the bribe simultaneously and clients engage in costly search. By manipulating the number of available o¢ ces and the size of search costs we are able to assess whether increasing competition reduces extortionary corruption. We …nd that, if search costs are unaected, increasing the number of providers may actually increase cor- ruption. In particular, our results show that increasing competition has either no eect (if search costs are high) or a positive eect (if search costs are low) on bribe demands. JEL classi…cation codes: D73, D49, C91We test the effectiveness of an anti-corruption policy that is often discussed among practitioners: an increase in competition among officials providing the same good or service. In particular, we investigate whether an increase in overlapping jurisdictions reduces extortionary corruption, i.e., bribe demands for the provision of services that clients are entitled to receive. We overcome measurement and identification problems by addressing our research question in the laboratory. We conduct an extortionary bribery experiment where clients apply for a license from one of many available offices and officials can demand a bribe on top of the license fee. By manipulating the number of available offices and the size of search costs we are able to assess whether increasing competition reduces extortionary corruption. We find that, if search costs are unaffected, increasing the number of providers may actually increase corruption. In particular, our results show that increasing competition has either no eeffect (if search costs are high) or a positive effect (if search costs are low) on bribe demands. We compare our findings to those obtained in a standard market environment and find evidence of different search behaviors in the two settings.