The relative efficiency of automatic and discretionary regional aid.

Discretionary subsidies are the primary instruments for regional policy in Europe. Such aid is typically targeted at `additional' projects: that is, projects that would not have been implemented without the subsidy, with the subsidy set at the minimum for the project to be implemented. This contrasts to automatic subsidies, where many of the aided projects are nonadditional and all eligible projects receive the same subsidy rate. The present paper builds on previous studies to compare three subsidy schemes: an automatic scheme and two types of discretionary scheme, one with accurate appraisal and the other with appraisal error. These schemes are assessed on their expected welfare impacts. The particular focus is the reduction in welfare gain imposed by the interaction of appraisal error and the requirements for accountability. This is potentially substantial and difficult to detect with conventional evaluation techniques. [ABSTRACT FROM AUTHOR]