It used to be trendy to talk about privatization as the solution for corruption. The World Bank, for example, declared back in 1997 that “any reform that increases the competitiveness of the economy will reduce incentives for corrupt behavior. Thus policies that lower controls on foreign trade, remove entry barriers to private industry, and privatize state firms in a way that ensures competition will all support the fight [against corruption].” (See also here, here, and here.) Although this theory declined rapidly after its peak in the 1990s, anticorruption policy ideas, like fashion, seem to be cyclical. Even as the privatization dogma has become démodé in Western anticorruption circles, it has gained new life elsewhere. As “privatization as a solution to corruption” debates reemerge in India and the Philippines, it’s worth reexamining the flaws in such policy proposals that made them fall out of favor twenty years ago.
The logic behind the idea that privatization inherently(or at least usually)decreases corruption is the notion that private shareholders are more interested than government bureaucrats in the efficient usage of whatever resources they control, and are therefore more likely to crack down on corruption. Relatedly, competition in the private market should favor those entities that can provide a service most efficiently—and if graft is inefficient, as many believe, market competition should drive corruption down. On top of this, private organizations also reduce corruption by offering more competitive wages, which means that employees aren’t forced to turn to corrupt means to supplement their incomes.
That’s the theory. The problem is that it isn’t supported by empirical evidence. Starting in the early 2000s and continuing well into the present, scholarship examining the aftermath of the privatization wave of the 1990s has repeatedly found that privatization has been largely unhelpful, and in some cases outright detrimental, to efforts to bring corruption under control (see here, here, here, here, here and here, to cite but a few sources). Why is this? Three main problems stand out:


