Doctoral Dissertation Research in Political Science: Policy Tool Bundling: Predicting the Selection of Economic Development Policy Instruments Using a Multivariate Probit Analysis

该奖项是根据2009年《美国复苏和再投资法》(公法111-5)提供资金的。经济发展激励工具的建模选择需要整合复杂的政治和经济考虑因素系统，并在一系列政策选择中考虑具体工具之间的关系。以往关于政策工具选择的研究几乎都是孤立地关注个别政策工具。这忽略了多个工具相互配合工作的现实。经济发展激励之间的关系可能源于地方官员追求效率，例如规模经济，或者他们通过满足意识形态或选民需求来追求政治利益。任何政策选项的选择都不是独立于其他政策选项的。就像经济产品一样，政策工具可以作为补充或替代。本文发展了一个政策工具捆绑理论来解释城市为促进其经济或政治利益而采取的发展政策的配置。政策捆绑方法建立在政府工具和政策扩散文献现有研究的基础上，通过调查地方政府内部因素和来自邻近司法管辖区的外部压力如何影响所选择的一套政策工具。使用汇总的纵向数据来确定佐治亚州城市激励政策之间的相互依存关系。这项研究的纵向性质允许预测以往和当前经济政策选择的影响，同时避免虚假的关系。专注于一个州有助于纳入扩散变量，以捕捉周围社区的行动。政治机构的影响，特别是职业经理人的存在和广大理事会的代表，预计将鼓励基于效率问题的政策捆绑。民选行政人员的存在和地区代表预计将鼓励基于政治利益的捆绑。贝叶斯多元概率模型被用于预测经济政策选择，因为它允许同时估计多个政策采用，并识别替代和互补效应。本文的理论价值来自于综合政策工具和扩散政策理论框架来解释政策捆绑。在这两个领域现有的经验研究中，缺乏对政策相互依存性的系统调查。本研究中提出的政策捆绑理论试图填补这一空白，并可以扩展到许多政策领域，包括经济发展、环境监管、税收政策和许多其他同时利用多种工具的政府举措。因此，这项研究有望使学者和实践者都受益，并扩大我们对政治制度如何影响政策结果的理解。

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).Modeling choices of economic development incentive tools requires integration of a complex system of political and economic considerations and accounting for relationships among specific instruments within a bundle of policy options. Previous studies of policy instrument choice focus almost exclusively on individual policy tools in isolation. This neglects the reality that multiple tools work in conjunction with one another. Relationships among economic development incentives can result from local officials' pursuit of efficiencies, such as scale economies, or their pursuit of political gains, from satisfying ideological or constituency demands. The selection of any policy alternative is not independent of the other policy options. Just as with economic goods, policy tools can operate as complements or substitutes. This dissertation develops a theory of policy tool bundling to explain the configurations of development policies cities adopt to promote their economic or political interests. The policy bundling approach builds on extant research from the tools of government and policy diffusion literatures by investigating how factors internal to local governments and external pressures from neighboring jurisdictions influence the set of policy instruments selected. Pooled, longitudinal data is employed to identify codependences among incentives policies for Georgia cities. The longitudinal nature of the study allows predictions of the influence of previous and current economic policies choices while avoiding spurious relationships. Focusing on one state facilitates the inclusion of diffusion variables capturing the actions of surrounding communities. The influences of political institutions, particularly the presence of professional managers and at-large council representation are expected to encourage policy bundling based on efficiency concerns. The presence of elected executives and district representation are expected to encourage bundling based on political interests. A Bayesian multivariate probit model is used to predict economic policy choice because it allows simultaneous estimation of multiple policy adoptions and identification substitutive and complementary effects. The intellectual merit of this dissertation is derived from the synthesizing of the policy tools and diffusion policy theoretical frameworks to explain policy bundling. Systematic investigation of policy interdependencies is absent from extant empirical research in both areas. The theory of policy bundling developed in this research seeks to fill this lacuna and can be extended to numerous policy areas including economic development, environmental regulation, tax policy, and numerous other government initiatives which utilize multiple instruments simultaneously. Thus, this research promises to benefit both academics and practitioners and expand our understanding of how political institutions influence policy outcomes.