BPE Senior Seminar
Fall 2024
Third assignment (100 points)
INSTRUCTIONS: This assignment is due on December 8th by 11.59 p.m. Please use the following guidelines to complete the assignment and submit our work through NYU Brightspace.
For this assignment, you are required to work in groups of three students. At the beginning of your write- up, clearly state the full names and net IDs of all three members of your group. It's essential that the members of your group remain the same throughout the semester.
You have the option to use either Stata or R software for this assignment. Once you've made your choice, please stick with that software for the entirety of the semester.
Your main write-up should be formatted using the Times New Roman font at a 12pt size with 1.5 spacing. If you're including any code in your assignment, place it at the end in a section titled "Appendix." This section should be formatted using the Courier New font, which is best suited for displaying code in documents.
It's important to note that each member of the group must upload the assignment separately. When you're ready to submit, go to the Assignments tab in BrightSpace. Make sure your document is saved and uploaded as a PDF file. An important detail to remember is to name your file using your network ID. For example, if your net ID is "rcd306", the file you upload should be named "rcd306.pdf". Although your code will not be graded, you are required to submit it. One or two groups will be randomly selected to present their code.
The files for completing this assignment are available on Brightspace and Dropbox
1. Plotting the Raw Data (20 points)
Download the Excel file announcement_replication.xlsx. In a do-file or R-file write down code that can perform. the following tasks. Remember that you need to use the foreign package to open the excel file.
We are going to replicate Figure 2A from “The Rise and Persistence of Illegal Crops: Evidence from a Naïve Policy Announcement” (Prem, Vargas, and Mejia, 2023) (page 348).
• Use the group_by and summarize commands to get an average of the suitability for coca and non-coca areas for each year. Don't forget to call library(dplyr) (5 points).
• Replicate the plot. You must distinguish between the two areas based on their suitability on the plot. Use a legend as in the paper (10 points).
• What is this plot suggesting? (5 points)
2. Replicating Table 1 (30 points)
Write down the code that replicates only columns 1, 2, and 3 of Table 1. The standard errors in this case will be the same as in the papers because the code automatically clusters at the municipality level, so don't worry about it.
• Replication of the main table (30 points) (Hint: You need to include different fixed effects and controls depending on the specification. Your outcome is coca_area_grid and your independent variable is suit_announce. Read what is included in the specification of each column. For example, column 2 includes not only Municipality FE and Year FE, but also Department Year FE. Thus, you need to interact -using *- coddepto with time FE. The specification of column 3 also includes the following controls: indrural_announcement, discapital_announcement, IPM_announcement, lpobl_tot_announcement; which are interactions of these with a cease fire dummy that identifies the period after 2013. Just include these you don't need to interact anything with the controls).
3. Interpretation (50 points)
• What is the control group? What is the treatment group? What is the pre-treatment period? What is the post-treatment period? (10 points)
• What is b capturing in all of these three cases? (5 points)
• Explain what does the suit_announce variable measures in this setup? (5 points)
• Why do the authors mention that the controls (municipality characteristics) are measured before the announcement? (10 points)
• Why do the authors include FE departments in column 2? If they include these FE, what are we comparing when we get our b? (10 points)
• We know that the parallel trends assumption is crucial in a diff-in-diff setup.
o What resources do the authors use in the paper to provide some suggestive evidence that low and high coca-suitable municipalities follow similar trends? Be specific (5 points).
o What resources do the authors use in the paper to provide formal evidence that low and high coca-suitable communities follow similar trends? Be specific (5 points).
4. Extra (25 points)
• Do a placebo regression in which you change the date of the announcement. Run the same regressions as you did in columns 1, 2, and 3. Show your results using a table (20 points).
• Are these result consistent with what we should have expected? (5 points)