Corruption is higher in countries that pursue active industrial policies. What''s more, the positive effects of industrial policy on investment can be reduced by corruption by as much as 44% and the positive effects on R&D spending by up to 50%. That is the ''unpleasant interventionist arithmetic'' of Alberto Ades and Rafael Di Tella, writing in the latest issue of the Economic Journal. Their work suggests that the optimal policy response to the existence of corruption may well imply lower subsidies. Certainly, a consideration of corruption should not be absent from cost-benefit analyses of market intervention.
The researchers'' key argument is that active industrial policy transfers profits to firms in favoured sectors. Bureaucrats with control rights over those firms can create mechanisms to extract some of those profits through bribes. Since corruption is known to have a negative effect on investment and growth, the total effect of industrial policy on investment can be broken down into two parts: a positive direct effect and a negative indirect effect through corruption. The same could be true for the impact of industrial policy on R&D spending.
Ades and Di Tella''s first finding is that the variables they use to measure how active industrial policy is – indicators of procurement preferences to ''national champions'' and unequal fiscal treatment of enterprises – are indeed associated with higher levels of corruption. They then estimate the determinants of investment. Using data that usually favours active industrial policies, they find that their indicators of active industrial policy increase investment. But they also find that corruption reduces investment.
Breaking down the effect of industrial policy on investment into a direct effect and an indirect corruption-induced effect, the researchers find that in the presence of corruption, the total effect of industrial policy on investment can be as low as only 56% of the direct impact. In other words, almost 44% of the benefits of industrial policies can be lost through corruption. Ades and Di Tella begin their article by noting the major debate over whether active industrial policy plays a substantial role in promoting economic growth. On one side, industrial policy activists use arguments that range from the traditional notions that free markets provide firms with insufficient incentives to invest in R&D to more colourful ideas about ''competitiveness'' and the benefits of supporting investment in high value-added sectors using sophisticated technologies. A popular argument is that investment can be heavily promoted through active industrial policies with the experience of Japan and Korea cited as evidence.
On the other side, the case against active industrial policy points out the lack of convincing empirical evidence on the benefits of industrial policy, and the lack of agreement on practical issues such as the criteria to use in picking the companies to be favoured as ''national champions''. Economists have shown that the benefits of industrial policy are at best either small or nonexistent and that often the net results are negative. But policy-makers and the general public find the rhetoric of industrial policy advocates quite compelling. These advocates dispute the validity
of the economists'' numbers, using the array of data in the Global Competitiveness Report to argue in favour of industrial policy. The result is a stalemate where neither side listens to what the other has to say.
Ades and Di Tella take a different approach. They do not question the idea of competitiveness nor the fact that it may be a good thing and that to achieve it the government needs to intervene through an active industrial policy. Instead, they take the need for active industrial policy as given and investigate whether its side effects compromise the achievement of its goals. In order to bridge the communication gap between the two sides of the debate, the researchers use the data produced by the industrial policy activists. Their aim is to examine whether the possible benefits of interventionist industrial policies, such as the promotion of investment or the support of R&D, may be reduced in the presence of corruption.
They conclude that if the only harmful effect of corruption operates through an adverse effect on investment, the optimal industrial policy is as follows: where in a corruption-free environment, industrial policy should be 1 peso, in a corrupt economy it should be between 1.19 and 1.79 pesos. In the more likely situation, where corruption has other deleterious effects (perhaps because of moral considerations or through its effect on fairness), it may lead to lower optimal industrial subsidies. Though these results should be treated carefully given the relative narrowness of the data set, the magnitude of these corrections suggests that the consideration of corruption should not be absent from cost-benefit analyses of industrial policies.
''National Champions and Corruption: Some Unpleasant Interventionist Arithmetic'' by Alberto Ades and Rafael Di Tella is published in the July 1997 issue of the Economic Journal. Ades is at Goldman, Sachs & Co and Di Tella is at Keble College, Oxford.
