Identification in Macroeconomics

我提出的论文侧重于估计宏观经济政策，尤其是货币政策的因果效应。这个问题是现代动态宏观经济学研究的核心，与政策制定有直接关系。到目前为止，文学一直受到特殊概念框架的严重影响，在该框架下，“因果效应”的概念被明确定义为“冲击”的影响--即与特定来源有关的意想不到的、自主的变化。我的论文试图挑战这一概念框架，取而代之的是从经济学和社会科学的其他领域中通常使用的框架中获得灵感，在这些领域中，因果关系是用反事实来定义的。我计划通过结合统计学理论来证明“基于震惊”的因果关系概念中的缺陷，以及通过各种有效的经验应用来强调通过将因果关系假设为反事实可以获得的洞察力。我的研究工作到目前为止已经试图做到这一点。特别是，作为一个正在进行的项目的一部分，我已经开发了一些证据来强调与具体应用“基于冲击”的方法相关的问题，在这种方法中，人们试图通过控制同时驱动政策决策和利益变量的混杂因素来估计政策的因果效应。我计划在DPhil.这项工作的基础上再接再厉。首先，我计划开发类似的证据，以强调在其他情况下(即从业者试图控制混杂因素的情况除外)“基于休克”的方法会出现类似的问题。其次，我计划展示可以通过应用工作的替代方法获得的见解，重点是估计货币政策的效果(这是迄今文献中的一个定义挑战)。例如，我希望通过含蓄地透露自己的预测，修改对央行通过其政策决定影响宏观经济程度的估计。我的假设是，文献夸大了这一“信息渠道”，这是由于基本的“选择偏差”问题--例如，央行传达经济前景正在恶化的信息往往与糟糕的经济结果联系在一起，仅仅是因为央行有能力有效地预测未来的衰退，而不是因为它们的传达方式与宏观经济之间存在任何因果关系。这对货币政策的实施具有重要意义，尤其是在经济衰退期间。

My proposed thesis focuses on estimating the causal effects ofmacroeconomic policies, particularly monetary policy. This question is centralto the study of modern dynamic macroeconomics and has direct relevance forpolicy making. The literature to-date has been heavily influenced by aparticular conceptual framework, under which the notion of a "causal effect" isexplicitly defined in terms of the effect of a "shock" - that is, an unanticipated,autonomous change linked to some specific source. My thesis would seek tochallenge this conceptual framework, taking inspiration instead fromframeworks typically used in other areas of economics and social scienceswhere causality is defined in terms of counterfactuals. I plan to do this througha combination of statistical theory to prove the deficiencies in the"shock-based" conception of causality, as well as through various effectiveempirical applications highlighting the insights that can be gained throughconceiving of causality as counterfactuals.My research work to-date has already sought to do just that. In particular, aspart of an ongoing project, I have developed a number of proofs to highlight theproblems associated with a specific application of the "shock-based" approachin which one attempts to estimate the causal effect of policy by controlling forconfounding factors that simultaneously drive policy decisions and outcomevariables of interest. I plan to build on this work for the DPhil.First, I plan to develop similar proofs to highlight that similar issues arise withthe "shock-based" approach in other settings (i.e. other than ones in whichpractitioners attempt to control for confounding factors). Second, I plan todemonstrate insights that can be garnered through alternate approachesthrough applied work with a focus on estimating the effect of monetary policy (adefining challenge in the literature to-date). For example, I hope to revisitestimates of the extent to which central banks influence the macroeconomythrough their policy decisions by implicitly revealing their own forecasts. Myhypothesis is that the literature over-states this "information channel" due to afundamental issue of "selection bias" - for example, central bankcommunications that the economic outlook is worsening tend to be associatedwith bad economic outcomes simply because the central bank has the ability toeffectively predict future recessions rather than due to any causal relationshipbetween their communications and the macroeconomy. This has significantimplications for the conduct of monetary policy, especially around recessions