代写Empirical Finance Spring II 2024 Assignment 3代做留学生Python程序

Empirical Finance Spring II 2024

Assignment 3

Q1. (90pts) Event study

Please open the file “assignment3.xlsx”, where you’ll discover the stock returns for Com- panies A and B from Day 1 toDay 70. Under “Text 1”, you’ll find headlines related to the IT sector, while “Text 2” contains headlines for the Biopharma sector. Notably, the CEO of Company A underwent a change from “John” to “Susan” on Day 51.

1.  (20pts) Conduct sentiment analysis on “Text 1” and “Text 2”. For each case, you will create a vector (length 70) with sentiment scores (sentiment score is set to zero if there is no news released on that day).  Let’s call this S1 and S2.  For this, you should rely on Python codes that I uploaded in canvas. Note that when there is no news, it will read as NaN value.  You need to adjust the code to handle the NaN values.  Report the sample averages of the two sentiment score vectors.

(Grading rule: There is no partial credit for this question.)

2.  (20pts) With the two sentiment score vectors in hand, our objective is to explain the returns for B. Regress B returns on each of sentiment vector (S1, S2) and report the coefficient estimates.  So, two regressions (i) regress B returns on constant and S1 and (ii) regress B returns on constant and S2.  Based on the regression results, infer whether B is affiliated with the IT sector or the Biopharma sector.

(Grading rule: You are required to report the coefficient estimates on the sentiment vector and the R2 value from each regression.  Failure to do so will result in deduc- tions in increments of 10 points.)

3.  (20 pts) Begin by reporting the full sample correlation of returns for A with B. Next, provide the sample correlation of returns for A and B during periods when John was the CEO of A. Finally, report the sample correlation of returns for A and B during periods when Susan was the CEO of A. What conclusions can be drawn regarding the characteristics (IT versus Biopharma) of company A following Susan’s appointment as CEO?

(Grading rule: You should discuss any significant changes in correlation patterns. Any deficiencies will result in deductions in increments of 10 points.)

4.  (30 pts) From the previous question, we can infer that company A went through structural change after Susan was appointed.  Let’s accommodate this feature into the regression model. Your goal is to reach the highest adjusted R2 value as possible.

(Grading rule: You will receive 20 points for explaining how the adjusted R2 value can exceed 0.90.  An additional 10 points will be awarded if you can demonstrate how to achieve a value higher than 0.97. Any deficiencies will result in deductions in increments of 10 points.)

Q2. (110pts) Conducting your own research

Based on our class, you are to demonstrate your research skill:

S1: Create your own research question (explain why it’s an important question);

S2: Collect relevant data and discuss summary statistics for your study;

S3: Apply the text-analysis technique you learned in class;

S4: Explain the results you find.

Let me provide you with an example based on what we covered in class.  Suppose your task is to analyze the impact of earnings announcements on Carey stock prices.  Explain why this question is important.  This represents S1.  You collect daily Carey stock prices and compile earnings reports.  This constitutes S2.  Compute the sentiment score evalua- tion via text analysis and regress daily returns on the sentiment score. This corresponds to S3. Finally, explain the estimation results (e.g., discuss the significance of estimated coeffi- cients and provide interpretation). This is S4.

Grading instruction: Scores will be assigned in the following order of decreasing excel- lence: 110 (excellent), 80 (good), 50 (okay), and 20 (poor), emphasizing the significance of this research project. Ten percent (10%) of the entire groups will receive an excellent (poor) grade, while the remaining 80% will receive okay or poor grades. You will get zero points if I find that your submissions are similar to those from other groups.