concerns through institutional and policy frameworks, by creating local constituencies, and by relying on NGOs for supervision. The emphasis is on gaining local support and commitment as well as increasing local capacities. The Bank tends to believe that design and the involvement of the borrower, more than supervision, improves projects. Yet the Wapenhans and Morse reports also underlined the importance of supervision. Accordingly, the Bank decided in 1993 to pay closer attention to the implementation of its environmental policy in the field.
This strategy has allowed the Bank to improve considerably its access to funds, projects, and political support. In that sense it has "learned." But it also faces a new dilemma because the new policy reduces the control that it has over its own performance. EA procedures, intended to protect the Bank from economic and political project failures, create new vulnerabilities as they throw the organization into the thicket of national politics. Other factors reinforce this trend: the pressure to include local participation, the role of broker among national administrations and social groups that the Bank plays through NEAPs and EAs, the use of environmental and human rights issues by ambitious local politicians, and the use of environmental issues by governments as a means of gaining access to international capital. The assumption that transparency and pluralism will contribute to consensus building behind a set of criteria governing development projects and process ( Goodland, Juras, and Pachauri, 1992) runs the risk that the outcome of environmental debates will rest not on sustainable development considerations but on internal political power struggles. Transparency identifies winners and losers, and pluralism ensures that the aggrieved parties will stay involved. Environmental learning, therefore, has reduced the autonomy of the organization by forcing it to assume a larger political role.