Risk, Robustness and Volatility

9709876 Hansen这个项目有助于我们理解风险、稳健性和波动性。有三个不同的部分：1。分散经济中的风险敏感性和稳健性。动态经济的模型是建立在消费者/投资者可以部分基于稳健性考虑做出决策的基础上的。研究借鉴了递归效用理论、鲁棒控制理论和奈特不确定性下的去中心化理论。特别令人感兴趣的是，对稳健性的偏好如何反映在均衡价格（包括风险的市场价格）中。2. 校准的经验基础。研究了将微观数据集的估计集成到动态随机一般均衡模型中的替代方法。该项目评估并扩展了当前的微观计量经济学方法，以解释代理异质性和信贷约束。它获得了围绕微观经济决策的不确定性的新估计。它研究了新的微观计量投入的一般均衡含义。3. 扩散模型的柔性估计。扩散模型作为非线性时间序列的模型，由于其编码信息（如命中时间概率和长期平稳分布）的简单方式而具有吸引力。特别注意的是描述扩散系数或局部方差依赖于马尔可夫状态的方式。在非线性扩散模型中，这种依赖足以将过程拉回到其分布的中心。采用时间序列版本的筛法来提供感兴趣的挥发性表征。？?

9709876 Hansen This project contributes to our understanding of risk, robustness and volatility. There are three distinct parts: 1. Risk Sensitivity and Robustness in Decentralized Economies. Models of dynamic economies are built in which consumers/investors can make decisions based in part upon robustness consideration. The investigation draws upon ideas from recursive utility theory, robust control theory and decentralization under Knightian uncertainty. Of particular interest is how a preference for robustness becomes reflected in equilibrium prices including the market price of risk. 2. Empirical Foundations of Calibration. Alternative ways of integrating estimates from micro data sets into dynamic stochastic general equilibrium models are studied. The project assesses and extends current microeconometric methods to account for agent heterogeneity and credit constraints. It obtains new estimates of the uncertainty surround microeconomic decision making. It investigates the general equilibrium implications of the new microeconometric inputs. 3. Flexible Estimation of Diffusion Models. Diffusion models as models of nonlinear time series are attractive because of the simple manner in which they encode information such as hitting time probabilities and long run stationary distributions. Particular attention is devoted to characterizing the manner in which the diffusion coefficient or local variance depends on the Markov state. In nonlinear diffusion models, this dependence alone can be sufficient to pull the process back to the center of its distribution. Time series versions of sieve methods are employed to provide the volatility characterization of interest. ??