Computational Finance Extra-credit Project: How do Elections Move Markets?

Instructions

Total points: 6 (directly add to your final grade without discounting) Please read this entire file carefully before you start working on the project.

1. This project is optional.

2. Students will collect data, perform data analysis and write a report.

3. If you choose to work on this project, please email me your report in PDF file, your data in excel (or .csv) and your R code by 11:59PM Nov. 22. No late submission will be accepted or graded.

4. Let me know if you have any question before the due date.

1 1 Project Guideline

We are experiencing 2020 U.S. election, which was the 59th quadrennial presidential election, in the midst of pandemic. For this project, you will analyze the impact of U.S. elections on stock markets. This is an open-ended project with little instruction. Please be creative. At the same time, this is designed to be a simple and fun project and you do not need to over complicate it, unless you have enough time.

1.1 Data collection

Begin your project by collecting data. We will be analyzing the impact of presidential election on stock market. Therefore, we need to collect date on the market and on the election results. For simplicity, you can just collect daily closing index level of one market index, for example, S&P 500 (from 1928) from Yahoo!Finance or Bloomberg terminal. Feel free to include data from more than one index. Collect data on presidential election. Please explore data resources online yourself. You want to collect data on - the election date or the actual date the elections were called, - which party won (democratic or republican), - and whether the incumbent president was reelected. - You may want to collect more data based on your analysis, for example, poll results, number of votes, etc. This is not necessary. You do not need to collect the results on all the president elections. The election data before the start of your market index data are not necessary. For example, if you only use S&P 500 index, you just need to collect data on presidential election from 1928.

1.2 Analysis

You will use R to perform the analysis. You can then begin your analysis by simply asking yourself a few questions. How did the market react (positively or negatively in terms of return) on the second day, the week, and the month after the election was called? Did the market return differ by which party got elected? Did the market return differ by if the incumbent party got reelected? Plotting is a virtual exploration of the data and can be very informative. Regression is the statistical way to analyze the data. What plots to include and what tests and regressions should you run? 2 You can draw a time-series daily return of the index and labeling all the election dates (or the actual date when the election was called) while using different colors to denote the election results. For example, to distinguish the market reaction by which party got elected, you can draw a time-series daily return plot of the index and labeling all the election dates while using blue and red to denote democratic and republican party. You can create a new dataset, with variables as below. Election dates party got elected incumbent got reelected? r1-Day r1-Week r1-Month where r1-Day, r1-Week and r1-Month the second day, the week, and the month after the election was called The 2016 election was called on Nov. 8th. Tuesday. To calculate the 1-week return after the election, you want to take the close index level on Nov. 8, 2016. You want to extract the close index level on Nov. 15, 2016 (exactly 1-week after). r1-Week = Closing IndexNov. 15, 2016 Closing IndexNov. 8, 2016 . You can calculate the average 1-day, 1-week, and 1-month after return by parties being elected, by if the incumbent got reelected and make several bar plot. For regression, you may want to regress the market returns (try one day, one week and one month) on a dummy variable indicating which party get elected or a dummy variable indicating whether the incumbent part gets reelected or both dummies together. Typically, we should add control variables in the regression. For simplicity, you can ignore controls and only regress the market returns on the variables of interest.

2 Report Guidelines

2.1 Formatting The report will be at least 8 pages long (NOT including the cover page, the table of contents page, or the reference page), with font size 11-12, single line spacing, 0 pt before paragraph, and 6 pt after paragraph.

2.2 Contents Your report needs to include, but NOT restricted to those part: - Cover page: this should include the title of your project, your name, and date. - Table of Contents: the table of contents should direct the reader to each section and subsection of the report. 3 - Main report part: please separate your main report into numbered secsions, subsection, etc. to make it easier to follow. Motivation of the project, background of the election, etc. Brief summary of the analysis results.

Data introduction

Empirical Analysis You want to elaborate on the plots you made, empirical test your performed, and what your inference is. All supporting charts, and graphs should be included where it was referred in the report. Please do NOT include any data or R code in the report.

Conclusion. - References: A separate reference page should be included that cites references. Note: reference should start on a new page.