Abstract
In this paper, we show that work schedule is a key factor driving investor inattention to Friday earnings announcements using the introduction of the five-day workweek system in the financial sector of Korea in 2002 as a natural experiment. Our analyses show a stronger immediate response and a weaker delayed response to Friday news under the six-day workweek system. The trend was, however, reversed under the five-day workweek system. We also find out consistent evidence from the movement of trading volume. These findings demonstrate that the trade-off between weekend distraction and additional working hours during the weekend determines investors’ attention to Friday earnings announcements.
Acknowledgments
We thank Jeremy Stein and Lawrence F. Katz for their excellent guidance. We would also like to express our gratitude to seminar participants at Harvard Graduate Student Seminar, the Eastern Economic Association and the Asian Econometric Society for their valuable comments.
Notes
* Accepted by Hong Hwang upon recommendation by Taychang Wang.
1. We found significant discrepancies in the early years of data. We assume that KIND data is more accurate and corrected all discrepancies in the data until 2002. After 2002, the rate of discrepancy fell significantly, and we assumed we could incorporate that amount of measurement error.
2. Before 2003, forecast data existed only for each company’s 4th quarter. Although estimates are for annual forecasts, because analysts made a forecast after companies announced their 3rd quarter earnings, we treated this information as equivalent to quarterly earnings forecasts.
3. Korean firms started to move from Generally Accepted Accounting Principles (GAAP) to International Financial Reporting Standards (IFRS) in 2009. We matched forecast and actual earnings according to each accounting standard.
4. Two thousand won is approximately equal to two US dollars. We lower the criterion to one thousand won when we examine delayed and long-term responses because these accumulated returns are less likely to be affected by temporary volatility in stock price.
5. We employed the Dickey-Fuller test and found a unit root in many individual volume series in our sample around the earnings announcement day.