Four Projects about Macroeconomic Risk and Uncertainty: An Accomplishment Based Renewal

这个项目包括四个相关的研究方向。1.模型漂移和随机波动：该项目使用计算密集型贝叶斯方法来组织关于二战后美国总体时间序列向量自回归中漂移来源的证据。它试图将漂移分为两个部分，一方面是系统的或自回归的部分，另一方面是创新的条件方差。文献中有一些有影响力的研究声称肯定了向量自回归的系统部分的漂移，也有其他同样有影响力的研究否认了这一点，支持漂移波动性。对连续时间扩散的洞察使研究人员怀疑，检测系统部分的漂移比检测挥发部分的漂移困难得多。将漂移划分为这两个部分，对于定量评估货币和财政政策规则及其效果非常重要。该项目根据最近关注政府学习过程的漂移系数模型的理论公式来解释收集的证据。2.自我确认均衡与货币政策：自我确认均衡是理性预期的一种形式，它允许代理人的模型对在均衡状态下零概率发生的事件持不同意见。最近关于代理人学习的理论将这种形式的理性预期均衡作为一个极限点。这个项目将早期关于自我确认均衡的工作扩展到一个经验上更具雄心的工作，该工作基于一个版本的“新综合”宏观模型。它还研究了逃生路线，一种新形式的动态，以及它们所带来的教训，以应对对我们当前低通胀制度的可能威胁。3.宏观经济学中的鲁棒控制与滤波：本项目研究的是稳健滤波与控制的动态模型的应用。这项工作的目的是提供积极和规范的决策模型，其中决策者将其模型视为近似值。稳健决策理论提供了一种谨慎的模式。稳健控制理论被应用于资产定价难题以及货币和财政政策的制定。4.宏观经济历史：这个项目深入研究了两到三次重要的通胀稳定，并侧重于导致稳定的产出和失业成本高或低的条件。

This project consists of four related lines of research. 1. Model drift and stochastic volatility: The project uses computationally intensive Bayesian methods to organize evidence about the sources of drift in vector autoregressions of post WWII U.S. aggregate time series. It seeks to partition drift into parts due to the systematic or autoregressive part, on the one hand, and the conditional variances of innovations, on the other. There are some influential studies in the literature that purport to affirm drift in the systematic part of vector autoregressions, and other equally influential ones that deny it in favor of drifting volatilities. Insights about continuous time diffusions lead the investigator to suspect that it is much more difficult to detect drift in the systematic part than in the volatilities. Partitioning drift into these two parts is important for quantitatively evaluating monetary and fiscal policy rules and their effects. The project interprets the evidence assembled in the light of recent theoretical formulations of drifting coefficient models that focus on the government's learning process. 2. Self-confirming equilibria and monetary policy: A self-confirming equilibrium is a form of rational expectations that permits agents' models to disagree on events that occur with zero probability in equilibrium. Recent theories of agents' learning deliver this form of rational expectations equilibrium as a limit point. This project extends earlier work with self-confirming equilibria to an empirically more ambitious one based on a version of a 'new synthesis' macro model. It also studies escape routes, a new form of dynamics, and the lessons that they carry for a possible threat to our current low-inflation regime. 3. Robust control and filtering in macroeconomics: This project pursues applications of dynamic models of robust filtering and control. The aim of this work is to provide positive and normative models of decision making where the decision-maker regards his model as an approximation. Robust decision theory provides a model of caution. Robust control theory is applied to asset pricing puzzles and the formulation of monetary and fiscal policies. 4. Macroeconomic history: This project studies in depth two or three important stabilizations of inflation and focuses on the conditions that contributed to high or low output and unemployment costs of stabilization.