Confidence Intervals for Univariate Impulse Responses With a Near Unit Root
提出一种为近单位根单变量时间序列构建脉冲响应函数置信区间的方法，能控制覆盖概率，优于现有技术，并应用于美国总产出数据。
This article proposes a method for constructing confidence intervals for the impulse response function of a univariate time series with a near unit root. These confidence intervals control coverage, whereas the existing techniques can all have coverage far below the nominal level. I apply the proposed method to several measures of U.S. aggregate output.