Testing the Tools Approach: Tax Expenditures versus Direct Expenditures

Abstract

Over the last 15 years, a number of scholars have called attention to the wide variety of policy used by governments. Led by Lester Salamon, they have tried to rewrite the usual equation of government equals direct expenditures plus laws and regulations by adding important but often overlooked policy such as loan guarantees and tax expenditures (policy are also known as policy instruments) (Fesler and Kettl, 1991; Hood, 1986; Leonard, 1986; Mosher, 1980; Salamon, 1981, 1989). The common thread linking these overlooked is that while they are publicly financed, the actual goods and services are delivered by private companies and nonprofits. One aim of this group has been to increase public awareness concerning the extent of third-party Recent demands for privatization, they point out, are a little misleading because the national government has always relied heavily upon third parties to build military weapons, issue home mortgages, provide employees with health insurance, and the like. A second aim has been to raise questions of accountability. While some of these receive so little official scrutiny that their impact and effectiveness are unknown, others allow third parties considerable latitude in administering public programs. Some of these scholars believe that a tools can significantly improve our understanding of policy making and public administration (e.g., the contributing authors in Salamon, 1989). In their view, previous research typically used the program or agency as the unit of analysis, focused on one or two cases in depth, and stressed the unique features of each case. Cumulative knowledge was slow to develop. By switching the unit of analysis to the policy tool, scholars will be able to generalize more broadly. They will recognize patterns across programs and across agencies that use similar tools, and draw lessons concerning the proper use of each tool. Such knowledge would not only be good in and of itself but would also improve the quality of policy making. The utility of the approach hinges on a key assumption: policy do, in fact, exhibit distinctive patterns. To date, advocates have developed a theoretical justification for the approach and described the politics of individual tools. They have predicted several ways in which policy should differ but offered little empirical evidence to demonstrate that important differences exist. This observation is not meant as a criticism; most of these advocates have rightly started by laying the groundwork for future research. One of the first tasks is determining whether policy differ consistently, and to do that, one must compare their actual performance. This article takes up that challenge. My purpose is to investigate the extent to which policy differ by comparing tax expenditures to direct expenditures. These two have been chosen because Salamon (1989) and others consider them to be distinct. They are also important because they entail massive commitments of public resources and span the full range of functions performed by the national government. Our confidence in the approach would increase if we discovered differences between along several different dimensions. Ideally, one would like to compare the performance of policy - how efficient and effective each is in achieving its primary objectives. There are, however, remarkably few evaluations of tax expenditures (which may help to explain their popularity). Accordingly, in this article, I compare the creation, growth, and administration of tax expenditures and direct expenditures. Is one tool easier to enact than another? Does real (inflation-adjusted) spending for one consistently exceed the other? Is one tool easier to implement than another? These are crucial questions for groups making demands on government, elected officials, and public bureaucrats to answer, particularly when faced with greater fiscal constraints on and public skepticism of government. …