Time-varying macroeconomic announcement risk
研究了公告波动率或事件风险随时间变化的问题，利用高频数据和灵活模型识别条件事件风险，以原油为例发现公告事件风险随时间变化幅度可达10倍。
This paper examines an issue overlooked in the finance and economics literature: time variation in announcement volatility or event risk. To identify this, we combine long spans of high-frequency data with a flexible model of returns. The model allows us to separately identify conditional event risk from other factors like time-varying volatility, jumps and intraday periodicity, and long time spans of data are needed given the infrequency of most announcements. We focus on crude oil due to its economic importance, high volatility and complex announcement structure. Results indicate strong evidence for time-varying announcement volatility as announcement event risk varies by as much as a factor of 10 over time.