Getting down to the business of corruption

Fighting bribery and corruption is one of the toughest challenges facing governments in many developing countries, but CEPR Research Affiliate Benjamin Olken and his co-author Patrick Barron believe economics can help. They have conducted an extraordinary experiment in Indonesia to prove that although corrupt officials may be outside the law, they are not beyond the reach of market forces. In fact, Barron and Olken found that individual bribe-takers seem to behave rather like firms operating in a market, trying their best - sometimes in quite sophisticated ways - to maximize their 'profits'.

In the experiment, volunteer surveyors accompanied truck-drivers on more than 300 journeys along two routes through the Indonesian provinces of Aceh and North Sumatra, recording the level of bribes they paid at checkpoints and weigh-stations along the way. On average, drivers were forced to pay $40 - around 13% of the total cost of the trip. The results make clear there is some bargaining process involved: the 'price' of passing through a checkpoint or weigh station is not simply set in advance. Where, for example, the corrupt official involved has a gun - a very obvious demonstration of his bargaining power - the value of the average bribe increases by 17%. Having another person manning the checkpoint also helps the bribe-takers to make their case: one official can spend time haggling with a truck driver, without fearing that another driver will slip past.

Accordingly, the average bribe increases by 5% for each extra member of staff present. But bribe-takers - and payers - are also responding to much more subtle signals than who is waving a gun. During the period covered by the experiments - November 2005 to July 2006 - the Indonesian government withdrew 30,000 police officers and troops from the province of Aceh, as part of the peace agreement ending a 30-year long civil war. On the section of the truckers' route passing through Aceh, the number of checkpoints dropped dramatically as a result, so that the average journey involved 15, instead of as many as 40 before the withdrawal.

Corrupt officials along the rest of the route, through North Sumatra, responded just like good profit-maximisers: knowing that truck drivers were facing lower bribes elsewhere, they demanded more cash for themselves. However, the total amount paid in bribes did fall after the troops were withdrawn. The authors say that this suggests 'pricing' decisions by corrupt officials are decentralised: if an organised criminal gang were coordinating the collection of bribes, they would probably simply have made up every dollar they lost in Aceh, on the North Sumatra side of the border.

It also appears to matter at what stage on a truck-driver's journey a particular checkpoint occurs. At least on one of the two routes examined in the study, the later in the journey an official tackles the driver, the larger a bribe he is able to exact.

This is an example of what economists call 'double marginalisation': firms at different positions in a supply-chain are able to exert different levels of market power. Earlier in the journey, both the bribe-taker and the truck driver know that he is facing further checkpoints later in the route. Again, corrupt officials, working outside the law, appear to be mimicking the behaviour of profit-maximising firms.

As well as taking into account prevailing market conditions - such as the number of other checkpoints faced by the drivers, and the stage of their journey they have reached - corrupt officials showed evidence of using another, more sophisticated tool of the successful firm: differential pricing.

Just as airlines charge business passengers many times the price tourists are willing to pay for a flight, the corrupt officials at checkpoints seemed to be trying to adjust the size of bribes to fit truck drivers' ability to pay up. Hence having a new, shiny truck or carrying a particularly expensive cargo were both likely to incur a heavier bribe.

In one case, officials at a weigh-station at Gebang, in North Sumatra, had even gone to the lengths of offering two alternative tariffs: for a fixed fee, drivers could purchase a date-stamped coupon from a criminal organisation in Medan that would limit the total price they were forced to pay for having an overloaded lorry once they arrived at the weigh-station.

In fact, drivers didn't seem to be particularly accurate at choosing the cheapest tariff to suit their particular load; but the existence of the complex pricing system suggests officials are at least trying to differentiate between different types of 'customer'.

One result of this meticulous examination of the behaviour of corrupt officials is to show that intuitions and models usually applied to firms also have relevance in much less conventional spheres, and even beyond the law.

The findings also have important practical implications for developing countries struggling to tackle corruption. Given that, in this case at least, the 'pricing' behaviour of corrupt officials is decentralised, for example, fighting corruption by attacking it 'at the top' may actually fail to reduce the total amount taken in bribes.

Second, although the North Sumatran officials did increase their takings when the Aceh checkpoints were reduced, total bribes paid declined. That suggests that simply cutting the number of potentially corrupt officials the public have to encounter is a good idea. Barron and Olken's analysis shows that when governments of developing countries are designing anti-corruption crackdowns, they need to think carefully about the market for bribes.

CEPR DP No. 6332 The Simple Economics of Extortion: Evidence from Trucking in Aceh
Patrick Barron and Benjamin Olken