Sammendrag:

Fiscal reforms in government are continually proposed, frequently adopted, and usually assessed as failing to achieve their desired outcomes. This paper develops a broad contingency framework that focuses on the third step—the assessment of fiscal reforms. The value of the framework is that it provides researchers as well as reformers a lens to assess the net benefits of reforms. Not only does it provide the lens, but it also suggests the accuracy or clarity of the lens in different circumstances. Our framework is contingent on two dimensions: type of fiscal reform (rules vs. judgment) and life cycle (innovators, early adopters, maturity). Our framework is broad in that it combines prior research from the budgeting and accounting literature. Recent budgeting research has suggested that an outcomes view of reform can be replaced with a process view. In addition to these viewpoints, we also analyze the monitoring and signaling viewpoints from accounting research. We believe the four viewpoints are complementary rather than mutually exclusive, and both accounting and budgeting researchers can learn from each other. We argue that signaling may be the best viewpoint for two reasons: (a) It provides insights into the net costs and benefits of adopting a reform, and (b) it provides insights into policy suggestions that are likely to increase the adoption of beneficial reforms. We then apply the framework to performance budgeting reforms of state governments in the United States. Our analysis suggests that performance measures might not be useful for the legislative decision process of allocating resources among disparate goals but appear to be useful for improving the quality and cost of providing services. Three policy suggestions are (a) legislators should focus on incentives for managers to report performance measures, (b) researchers and reformers should consider how performance budgeting could be done on a comparable basis across states, and (c) influential actors such as legislators and the media should be studied in greater detail.