"Pareto-Optimal" Bribery?

In many parts of India, there exists a de-facto system for resolving traffic tickets (among other things), with which I am sure many of you are familiar. Typically, when a person is pulled over for some traffic violation, they have the option of paying roughly half the official ticket amount as a bribe directly to the policeman instead of receiving the official challan (fine). Having directly witnessed this system in action countless times, I often find myself wondering whether this is actually such a bad thing.

There are three major reasons why bribery is considered a bad thing in situations like this. First, the government loses a source of legitimate revenue. Instead of going into government coffers, the fine money is going into the pocket of the policeman. Second, by taking bribes the police undermine their legitimacy as those who uphold the law. Widespread knowledge of police corruption is not conducive to maintenance of law and order. Third, the option of bribery reduces the incentive for people not to break laws in the first place (assuming that the bribe is less than the expected punishment conditional on being caught).

In the Indian context, however, not all of these concerns are necessarily applicable. The first issue of depriving the government of revenue is not particularly notable when one considers that corruption is endemic to almost every administrative branch of the Indian government, and it is therefore difficult to determine whether the money is better off in the policeman’s pocket. If it were a safe assumption that the government would use the fines collected for some socially useful purpose, it would be easier to condemn such instances of bribery. Unfortunately, it is rare that one can accurately make such an assumption in the Indian context.

The issue of police undermining their own legitimacy also seems to be a much more muted concern India, primarily because the police themselves do not enjoy a great deal of social legitimacy to begin with. Bribery is to a large degree institutionalized and often considered simply a cost of doing business with the police. The vast majority of the police force in India is also severely underpaid – taking bribes like this allows them to supplement their meager income. As long as the police are severely underpaid and have the bargaining power to take bribes, one can only expect that most will. Assuming that the police will seek to supplement their earnings with bribes, we can only hope that they engage in types of bribery that have a relatively low social cost. The social cost of accepting bribes instead of traffic tickets seems relatively low compared with say, demanding bribes to file police reports or to have criminal cases dropped.

Bribery as practiced in this context also maintains some incentive to follow traffic laws. Since the police will generally only pull over someone who has actually broken a law, the bribe replaces the fine as a financial incentive to follow the rules. Since the bribe will always be smaller than the fine itself, its possible that the incentive is watered down a bit, but the fact that bribes are only taken from those who have broken the rules in the first place maintains some degree of incentive to follow the law. This is in sharp contrast to hafta bribes (weekly payments collected from local businesses) which are more indiscriminate in nature.

The inherent opaqueness of bribery makes it extremely difficult to rigorously evaluate its social costs and benefits. I would be very interested in hearing people’s opinions or observations, especially those specific to the Indian context.