盈餘宣告、投資人情緒與股價報酬之關係

本篇論文研究主要目的是探討盈餘宣告對股價影響和投资人情緒的關性。因此分成兩個面向來探討，在盈餘宣告部份，如Ball and Brown (1968)首次結合會計資訊進行盈餘與股價之相關研究，實證結果顯示公司的預定盈餘宣告會產生股價的異常報酬。而在投資人情緒部分，因投資人直接情緒指標取得不易，故本篇研究參考Baker and Wurgler(2006)及其他文獻所提供的投資人間接情緒指標為代理變數做為本篇論文的依據，以盈餘宣告所存在的異常報酬做OLS迴歸分析，來探討投資人情緒對於盈餘宣告報酬的影響。本篇以台灣證券交易所(TWSE)上市公司為研究對象，並且參考沈中華與李建然(2000)事件研究法的市場模型來探討在IFRSs一致性的會計準則規範下，預定事件之盈餘宣告及投資人情緒是否會產生股價的異常報酬(Abnormal returns)，故取樣期間為2015年至2017年，共三年，探討是否有異常報酬，估期計為期239天，事件期以宣告日前3天至後3天，共7天。最後，由實證結果可知，盈餘宣告對股價所產生的異常報酬確實是存在的，而股價對盈餘宣告中所含的資訊反應是從宣告日才開始反應，並非在宣告日前就反應出其內含資訊。在盈餘宣告事件下之市場流動性，因投資人可從報章雜誌或新聞報導中取得資訊因此提前布局造成流動性增加，另外，投資人情緒代理變數的券資比與當日沖銷比均呈現負相關，表示投資人情緒傾向悲觀時，會出脫股票，後市股價異常報酬會增加。

關鍵字

盈餘宣告；投資人情緒；異常報酬

並列摘要

The main purpose of this research article is to explore the impact of earning announcement on stock price and investor sentiment, so the research article is divided into two aspects for discussion. In the aspect of early announcement, for example, Ball and Brown (1968) first combined accounting information to study earnings and stock prices, and empirical results showed that the company's scheduled early announcement would generate abnormal return in stock price. In the aspect of investor sentiment, it is not easy to obtain direct investor sentiment indicators, so this research refers to the direct investor sentiment indicators provided by Baker and Wurgler (2006) as well as other papers as proxy variables and the basis of this article and uses existing abnormal return of the earning announcement for the OLS regression analysis to explore the impact of investor sentiment on earning announcement and return. This article took the listed companies of Taiwan Stock Exchange Corporation (TWSE) as the research object and referred to the market model of the event study of Shen, Chung-Hua and Lee, Jan-Zan (2000) to study whether or not the early announcement and investor sentiment of scheduled events under the consistent accounting standards of IFRSs would generate abnormal returns in stock prices, so the sampling period was from 2015 to 2017, a total of three years, to explore whether there are abnormal returns, the estimation period was 239 days, and the event period was from 3 days before the announcement to 3 days after the announcement, a total of 7 days. Empirical results showed that earning announcement definitely had abnormal returns generated from stock price, and stock price only reacted to the information contained in the earning announcement on and after the announcement day, rather than before the announcement day. As to the market liquidity under the event of earning announcement, investors can obtain information from newspapers, magazines or news reports, so that the layout in advance causes an increase of liquidity. In addition, the short-sale/margin purchase ratio the proportion of day trades of the investor sentiment proxy variable are inversely correlated, indicating that if the investor sentiment tends to be pessimistic, then the stock will be sold out, and the abnormal return on the stock price will increase.