Competition in Bureaucracy and Corruption

1. Competition in Bureaucracy andCorruption Mikhail Drugov Department of Economics and Nuffield College University of Oxford EXLEGI 17 January 2007

2. Example and motivation • In India, the driving licence must be obtained at the police office of the local district where the applicant lives • In Russia, it can be obtained at any road police office of the region where the applicant lives • In both countries there is a lot of corruption in obtaining the driving licence

3. Example and motivation cont. • Indian case is monopoly: applicants do not have any choice of where to apply • Russian case is competition: applicants can choose where to apply What are the consequences of this institutional difference for applicants, policemen and society?

4. Other examples • Notaries (don’t) have to be local • Passports can be obtained (not) only in the local administration • Firms can(not) choose where to pay taxes • Firms can(not) choose the court • Many others…

5. What do we know? “A citizen can obtain a U.S. passport without paying a bribe. The likely reason for this is that if an official asks him for a bribe, he will go to another window or another city. Because collusion between several agents is difficult, bribe competition between the providers will drive the level of bribes down to zero.” Shleifer and Vishny (1993, p. 607) Competition reduces the size of bribes

6. Two kinds of corruption Changing a bit Shleifer and Vishny (1993) Extortion (=corruption without theft): bribe for serving a qualified applicant. For example, a policeman extorts a bribe from an applicant who knows how to drive. Collusion (=corruption with theft): bribe for serving an unqualified applicant. For example, a policeman asks for a bribe from an applicant who does not know how to drive. The applicant and the policeman collude against the society.

7. What do we know? cont. Informal analysis of Rose-Ackerman (1978, 1999), Shleifer and Vishny (1993) • Competition in bureaucracy reduces the size of bribes • This is good when the corruption is without theft • But not so good when the corruption is with theft as more unqualified applicants are served • In dynamics, competition, but not monopoly, may yield an intermediate stable equilibrium (if there are more bribe takers, bribes fall and some of bribe takers stop accepting bribes) (picture here) There is no model that compares the two regimes

8. This paper • The two types of corruption coexist, that is, there are both applicants who know how to drive and who do not know • Moreover, the type of corruption is endogenous, that is, applicants can invest and learn how to drive • Compare the two regimes along two dimensions: • ex post allocation: who (and how) obtains the licence given the investment decision • ex ante incentives: the incentives of applicants to become qualified

9. This paper cont. • A number of extensions • Different information structures (intermediaries, rotation, etc): How does the information about bureaucrats’ honesty/dishonesty affect results? • Punishments: In which regime are punishments more effective? What are the optimal punishments? • Dynamic setting: what is long-term composition of bureaucracy?

10. Modelling issues • Cadot (1987): competition, exogenous qualification, partial equilibrium, correction of ex post allocation • Acemoglu and Verdier (2000): monopoly, endogenous qualification, general equilibrium, correction of ex ante incentives • Our model: compare monopoly and competition, endogenous qualification, partial equilibrium (some results on the general), bureaucrats correct both ex ante incentives and ex post allocation • Red tape exists to screen the applicants: it improves ex post allocation (Banerjee (1997), Saha (2001), Guriev (2004)) • Our model: screening is useless as the bureaucrats observe the type of applicants at no cost. The role of the red tape is in providing of ex ante incentives

11. The model: firms There are firms (applicants) and bureaucrats (policemen) Firms need a licence to produce (for example, a licence certifies that the technology is clean) • Initially firms are unqualified (they have an obsolete technology) • They might invest to become qualified but this is costly • If an unqualified firm produces, it generates a negative externality (pollution)

12. The model: bureaucrats Bureaucrats: some are honest, others are dishonest • Honest ones give licence only to qualified firms • Dishonest ones give the licence to any firm in exchange for a bribe (to be determined) • Firms do not know the type of the bureaucrat before they apply • Bureaucrats know the type of the firm at no cost • Bureaucrats have unlimited capacity of serving firms

13. The model: timing • Many periods (infinite horizon) • Reapplication is costly because future is discounted and/or there is a (re)application fee • In each period a firm • Decides whether to invest (if it has not invested before) • Applies to a bureaucrat • Bargains about the bribe if the bureaucrat is dishonest • If obtains the licence, produces and the game ends • If not, the next period starts

14. Solving the model: monopoly • The bureaucrat is honest: • If the firm is qualified, he gives the licence • If not, he rejects. The firm invests, comes back in the next period and obtains the licence • The bureaucrat is dishonest: • He asks for a bribe and the firm pays as it does not have any other choice (denunciation is not possible). The type of the firm (qualified or not) does not matter • Conclusion: • Unqualified firms do not obtain the licence if the bureaucrat in the district is honest • The incentives to invest exist for one single reason: to avoid reapplication when the bureaucrat is honest

15. Solving the model: competition • The bureaucrat is honest: • If the firm is qualified, he gives the licence • If not, he rejects. The firm decides whether to invest and reapplies in the next period • The bureaucrat is dishonest: • He asks for a bribe. Now the firm has an outside option to reapply to another bureaucrat. Its value depends on whether the firm is qualified or not. A qualified firm has a higher outside option and therefore pays a lower bribe than an unqualified firm • Conclusion: • Unqualified firms may obtain the licence by reapplying until they meet a dishonest bureaucrat • The incentives to invest exist for two reasons: (1) to avoid reapplication when the bureaucrat is honest, (2) to pay a lower bribe when the bureaucrat is dishonest

16. Comparison of the two regimes Main result: The monopoly regime is better at implementing ex post allocation but the competition regime is better at providing ex ante incentives Which regime is better overall, depends on the relative importance of the two effects (note also that the bribes are lower in the competition regime for any firm)

17. Applications • Providing driving licences Learning how to drive is not very costly for any applicant. Ex ante incentives are very important → competition regime is probably better • Providing passports Investing to become a citizen is not really possible in most cases. Ex post allocation is crucial → monopoly regime is probably better

18. Extensions • Rotation of bureaucrats and intermediaries • Punishments for corruption • Dynamic setting (endogenous composition of bureaucracy)

19. Informational assumptions So far, firms did not know the type of the bureaucrat before they applied, they applied completely at random • Often, however, firms do have some information about bureaucrats’ honesty • Intermediaries (“agents”, “consultants”) provide this information • Both investing and non-investing are more attractive; ex post allocation will likely worsen; bribes fall in competition • In reality dishonest bureaucrats are better known (and they have incentives to advertise themselves!). Then, this information is rather bad • Rotation of bureaucrats may be a solution

20. Punishments • Punishment is an expected fine for a corrupt deal • It depends whether the firm is qualified or not • Only the sum of punishments for the firm and the bureaucrat matter • There is some maximum punishment smaller than the value of the service How do punishments affect ex ante incentives and ex post allocation (when bureaucrats stop asking bribes)? What are the optimal punishments?

21. Punishments cont. • Incentives to invest depend on the (weighted) difference between punishments of unqualified (collustion) and qualified firms (extortion) • If corruption is not prevented, punishment for collusion should be as high as possible and for extortion it should as low as possible • Incentives to invest are more sensitive to punishments in the competition regime (effect through the outside option that is absent in the monopoly regime) • Punishments are more effective in the competition regime

22. Punishments cont. The surplus that the firm and bureaucrat share is lower in the competition regime since the firm’s outside option is positive A higher punishment has two effects on the surplus: • A direct negative effect • An indirect positive through a lower outside option • The second effect is weaker when the firm is qualified →Extortion from qualified firms can be prevented by a lower punishment • If this punishment is feasible, then punishments for extortion should be set high (for collusion it should always be high) • Punishments for extortion exhibit very high variation in the competition regime (do we observe this?)

23. Dynamic (general equilibrium) analysis • Agents choose to become entrepreneurs or bureaucrats • Agents can be honest or dishonest • The type does not matter for entrepreneurship but does for bureaucracy • Honest agents enter bureaucracy at some exogenous rate (“enthusiasts”) • Decision of dishonest ones depends on the relative income of the two occupations

24. Equilibria in the two regimes

25. Conclusions • Competition is better in providing ex ante incentives • But monopoly is better in correcting ex post allocation • Rotation of bureaucrats may be beneficial • Punishments are (weakly) more effective in the competition regime • The two regimes may have the same (qualitatively) dynamic equilibria • Obviously, for different ranges of parameter values • The competition regime seems better in this sense (preliminary)

26. Future work • Optimal policy of the government w.r.t. red tape (application fee and period length) • TESTING!!! Thank you!