27 SEPTEMBER 1997, Page 27

What corruption costs

ONE feature of the Asian model is that those in power can do very well out of it. That goes for the African model too, as the achievements of the late Mobutu Sese Seko can remind us, but there are no tigers in Africa. The story is that in Asia, nest- feathering and influence-peddling are all part of the model, doing no harm there and even helping things along. This specious argument has not withstood the scrutiny of Professor Shang-Jin Wei of Harvard, who has been doing some statistical work on what corruption actually costs. He looks at its effect on foreign investment and con- cludes that it works like a tax: 'An increase in the corruption level from that of Singa- pore to that of Mexico is equivalent to rais- ing the tax rate by over 20 per cent.' He finds no evidence for the belief that, so far as foreign investors are concerned, corrup- tion in Asia does not count. Indonesia, as he agrees, has until now been a magnet for investment even though it is governed by a family whose head is known as Mr Ten Per Cent. Professor Wei's figures suggest that though investors may still think the price worth paying, if the family were to reduce its tariff and take only 5 per cent, Indonesia would attract more investment. The laws of arithmetic have not been rewritten for Asia. So I would have thought, but it is nice to see the professor prove it with a Tobit model modified for heteroskedasticity. (No, don't ask.)