The Influence of Estimation Period News Events on Standardized Market Model Prediction Errors.
发现《华尔街日报》新闻事件会影响股票风险调整后的残差收益分布，并提出一种新闻条件模型作为传统事件研究方法的替代方案，用于更准确地衡量预测误差。
Abstract In many accounting and finance research studies it is hypothesized that the news release under study has valuation implications. Results often indicate that the distribution of risk adjusted residual common stock returns, conditional on the occurrence of a wide variety of specific news event types, differs in one or more moments from the distribution of returns when such events are absent. This paper demonstrates that the distribution of Wall Street Journal news-conditional residual returns differs from the distribution of returns when such news is absent. A "news-conditional" model of the process generating security returns is proposed as an alternative to models typically used in previous event studies. Standardized prediction errors and squared standardized prediction errors from the news-conditional model are compared with those generated by conventional procedures.