Why letting off the bribe-giver will not reduce corruption?

The Chief Economic Advisor to Government of India, Prof Kaushik Basu has an interesting paper (via Mostly Economics) where he advocates that the act of giving harassment bribes (bribes that people often have to give to get what they are legally entitled to) be made legitimate. He writes,

"In other words the giver of a harassment bribe should have full immunity from any punitive action by the state. It is argued that this will cause a sharp decline in the incidence of bribery. The reasoning is that once the law is altered in this manner, after the act of bribery is committed, the interests of the bribe giver and the bribe taker will be at divergence. The bribe giver will be willing to co-operate in getting the bribe taker caught. Knowing that this will happen, the bribe taker will be deterred from taking a bribe... this entire punishment should be heaped on the bribe taker and the bribe giver should not be penalized at all, at least not for the act of offering or giving the bribe... such a change in the law will cause a dramatic drop in the incidence of bribery"

I could only say, "If only it were this simple"! Consider these,

1. Prof Basu's arguement makes the same mistake as other standard models that explain corruption. It overlooks the supply-side (bribe-giver) incentives and remains confined to the demand-side (bribe-taker). It needs to be borne in mind that apart from officials demanding bribes, harassment bribes are also paid because people are willing to pay a higher price (than the official fees) to access the service at their convenience.

As I have blogged earlier, this willingness to pay is manifested either in the higher direct bribes paid by those applying for government services or the payments made to middlemen who help procure the service. Merely addressing the demand-side incentives, while over-looking (and in this case, maybe, even furthering) the supply-side ones, is not likely to deter bribery.

2. Any amendment to the Prevention of Corruption Act (PCA) 1988 by repealing Section 12 (which defines the punishment to abettors) should have no effect on the incentives of those bribe-givers who bribe with an intention to trap the official. This category of bribe-givers are already protected by Section 24 read with the Supreme Court judgements (Bhupinder Singh Patel v. CBI, 2008 (3) CCR 247 at p. 261 (Del): 2008 Cri LJ 4396) indicated in Prof Basu's paper. This protection comes if the bribe-giver is able to establish that "the bribe was given unwillingly and in order to get the public servant trapped".

This is all the more easier to prove in case the bribe-givers are planning to trap the official with help of the local anti-corruption agency. In such cases, the entire bribe-giving process, though done secretly, is done with the assistance of the agency. And all this is in addition to the specific protections given under the respective state legislations to people approaching anti-corruption agencies to trap those demanding bribes.

Even without any amendment, the bribe-giver has the freedom to complain before the anti-corruption agency and lay a trap to catch the bribe-taker. In the circumstances, the amendment, atleast theoretically does not change the incentives of the bribe-givers.

3. The model suggested by Prof Basu gives excessive weight to incentives and assumes that changing those incentives will alter the bribe taking and giving environment. I am inclined to argue that this overlooks the critical role of enforcement.

There is nothing secretive about corruption in most government offices. It is part of the local gossip - who takes money, how much, through which channel, and so on. All anti-corruption agencies are well aware of this. It is only that they choose not to act unless prodded by a specific complaint.

In the circumstances, merely changing incentives will change nothing in the game. As already indicated, the incentives are already there (all the more so with sting operations by television channels etc). It is just that enforcement is abysmally weak.

4. There is the challenge of reforming a system where deviance is the norm. Standard models of addressing corruption works under the presumption that corrupt practices exist at the margins or are not widely prevalent. In such circumstances, the fear of being caught is large enough to deter bribe-taking at the margins.

However, when the practice is so widespread as to be a norm (as is the case with say getting a certificate from a Land Revenue official), conventional regulatory and incentives driven solutions become toothless. The deterrent effect of being caught is attenuated when the bribe-taker realizes that his actions are similar to that of the majority of his colleagues.

What will you do when the same corrupt practice bedevil most of the offices? How effective is the deterrent effect of being caught when the practice is widespread and the anti-corruption agencies are over-stretched (and in any case do not have the commitment to pursue such cases)?

In such a milieu, for "bribe-giver trapping bribe-taker" (and thereby deterring the "bribe-taker") strategy to be effective, there will have to be such incidents on an unimaginably massive scale. As we know, sending everyone to jail or suspending/trapping a considerable majority of officials is simply unrealistic.

5. Finally, there are perverse incentives associated with such changes, which are likely to harm the morale and fabric of the bureaucracy. It would come as a surprise to many readers to know that a not insubstantial number of such trappings across all states are motivated by work-place rivalries and social factors (caste politics). In fact, in many departments, threat of trappings are used as an instrument by local politicians to control officials. (None of this is to condone such corruption nor claim that such victims are innocent)

In fact, at the lower levels of the bureaucracy in rural areas, such threats are used to blackmail officials indulging in petty corruption. In fact, more often than not, those trapped are amongst the least venomous. The influential and the most corrupt have their strategies to manage the anti-corruption agencies and avoid being caught. This strategy also suits the officials of the anti-corruption departments who can meet their annual targets by targeting these lower-level officials.

In conclusion, for any anti-corruption (to contain harassment bribes) strategy to be effective, the bribe-taker and bribe-giver have to be respectively deterred from taking and giving bribes, and the enforcement machinery has to be responsive to those complaining to them and actively pursuing those suspected of taking bribes. Piece-meal administration of elements of this complex strategy is certain to fall flat, far from causing any "dramatic drop in the incidence of bribery".

PS: There are many real world issues I have not even mentioned - the difficulty of proving in a court of law that bribe was taken and the money has to be returned, the natural reluctance of citizens to be entangled in court litigation and being perceived as on the wrong side with government officialdom etc.

Update 1 (9.4.2011)

There is another reason why the bribe-giver-trapping-the-bribe-taker arrangment will not find too many takers. In a game theorey perspective, the bribe-giving citizen is not playing an one-off game. He plays repeated games with his bribe-taking interlocutors and their partners in the government. Accordingly, if a bribe-giver has a reputation for trapping officials, then he faces the strong possibility of being harassed (or atleast deeply inconvenienced) by the bribe-taking community. Therefore, when weighing the costs and benefits of trapping against not-trapping, the citizen feels it prudent to go along with the latter.