Empirical Finance Spring II 2025
Assignment 1 - 200 pts
Here are the assignment objectives:
• Source earnings event data for publicly traded stocks.
• Evaluate whether the “Sell the News” phenomenon is evident.
• Develop an investment strategy, analyze its performance using summary statistics.
For example, take a look at the Yahoo Finance earnings calendar:
It provides a list of companies announcing earnings on March 11, 2025. While not all com- panies have this information, for some you can find whether their earnings beat market expectations. Based on that, you can manually assign positive or negative scores. You can download individual historical stock price series https://www.investing.com/ for free. You can also use the Bloomberg Terminal we have access to at Hopkins for this purpose.
Specifically, you can view Netflix’s historical earnings announcements here:
Here are the instructions:
(1) Pick at least 20 of your favorite companies. Ideally, these should be well-known firms.
Collect at least 10 cases of earnings for each company, including historical stock prices, earnings announcement dates, and assigned scores (positive or negative). Ideally, try to obtain roughly half with positive scores and the remaining with negative scores.
Be careful when measuring the impact of an earnings announcement on stock prices. If the announcement was released after hours, it's more accurate to measure the change from the closing price on the announcement day to the opening price the following day, rather than relying on close-to-close price changes. Obviously, if the announcement was released pre-market, you should still compute the close-to-open change—but using the closing price of the day before the announcement and the opening price on the announcement day.
To be concrete, the above Netflix example includes at least 10 cases, with at least 5 assigned negative scores. You can download the corresponding historical stock (open/close) prices from https://www.investing.com/ and compute the impact by (i) taking the log transformation of prices and (ii) calculating changes in close-to-open prices around the announcement.
(2) Provide summary statistics—that is, the average changes in log prices on earnings announcement days—using the full sample as well as subsets based on positive and negative scores. Discuss your findings in the context of the \sell the news" phenomenon. Do you find any evidence supporting or contradicting it?
(3) Now, expand the holding period beyond just intraday changes to include multiple days. See if any patterns emerge when positions are held over a longer horizon. For example, you can calculate the return from buying three days before the announcement and selling on the announcement day (assuming the announcement is made after hours). This strategy avoids being directly affected by the announcement itself, as the posi- tion is closed beforehand. Similarly, you can examine returns from buying after the announcement and selling a few days later.
Use your creativity to explore and report any meaningful findings—for example, av- erage returns from specific holding period strategies (e.g., holding for xx days). If no clear pattern emerges, that in itself is still a valuable finding.