Researchers (Kim and Margalit) are interested in understanding the political impact of one foreign policy used to wage war through non-violent means: tariffs that raise the price of purchasing goods made in a particular country. They hypothesize that one country may seek to ‘eliminate’ another country’s leader by hurting the economic fortunes of workers whose support is pivotal for the given leader, making them consider voting the leader out in subsequent elections. If workers tend to vote similarly within each industry, and industries are geographically concentrated, it may be possible to levy tariffs that “punish” supporters of the country’s leader in specific industries, located in politically important districts. To test this, the researchers study the case of the tariffs China imposed on the US in the trade war while former President Trump was in office. The researchers create a measure of the US county-level share of employment targeted in each round of tariffs imposed by China. They next examine the relationship between the measure of pain inflicted on the local workforce and the political relevance of the locality based on the county’s voting in past (pre-tariff) elections. The model they estimate is given by:
πi = πΌ + π½1πΊππ ππππΈ ππ»π΄π πΈi + π½2πππΌππΊ π·πΌπππ πΌπΆπi
+π½$πΊππ ππππΈ ππ»π΄π πΈi ∗ πππΌππΊ π·πΌπππ πΌπΆπi + πi + πi
where the dependent variable Yi is the share of workers employed in targeted industries in county i, GOP VOTE SHARE is the average Republican two-party vote share in the 2014 and 2016 House elections, and SWING DISTRICT is a binary indicator denoting whether at least 20 percent of a given county i’s land is located within any swing district as we defined. We include an interaction term between GOP VOTE SHARE and SWING DISTRICT as well as state fixed effects µ in all models (which will be explained below
– you can ignore this for now). Observations are weighted by counties’ total voting age population as of 2016.
1 a) (15 Points) The authors start out with a simple OLS model, where they regress the county-level share of workers employed in industries targeted by tariffs from April 2018 on the GOP vote share and whether or not it the county is located in a Swing district. (Ignore the discussion of state-level fixed effects). The results of this estimation are shown in Column 1 of Table 1 below, with standard errors for each coefficient shown in parentheses below the coefficient. Interpret the estimated coefficient on GOP Vote Share. Is this statistically significant? Is this a causal relationship? Why/why not?
1 b) (5 Points) Using the results for GOP Vote Share in Column 1, calculate the 95% confidence interval around the estimated effect for GOP Vote Share.
1 c) (10 Points) Column 2 presents a model that is identical to the model run in Column 1 except that the interaction of GOP Vote Share and Swing District status is included. How do the results in Column 2 differ from the results in Column 1? Why is this the case?
1 d) (10 Points) Column 6 runs the same specification as Column 2 for a later round of tariffs (in August). How similar are the results in Column 6 and 2? Do you note differences? If so, explain what they are and whether or not they change the conclusions you reach about the author’s main hypothesis.
1 e) (5 Points) Columns 1-8 all include state fixed effects, that is, they include a control variable for each of the US States in the data. Read the explanation of fixed effects below, and then explain in your own words what the benefit of including state fixed effects is for this study.
Fixed effects are often used to reduce potential sources of endogeneity due to omitted variables. They capture differences in the dependent variable (outcome of interest) associated with each relevant unit/period of time. This is part of the unobserved error that has the same value for every observation.
for a given unit/time period. For example, if a researcher is interested in whether district-level rainfall (in the prior month) predicts conflict, (s)he may decide to study monthly rainfall and conflict levels for 100 countries from 1950-2000. Here, the researcher might choose to include fixed effects for each country (that is, including 99 control variables, one for each country except a reference country), to control for the ‘average’ rate of conflict in each of the 100 countries studied between 1950-2000. The regression results will identify the impact of recent rainfall on local conflict, controlling for average levels of conflict in each country. So the regression will tell us: how much variation in local conflict within each country (that is, deviation from the country-specific baseline level of conflict) is explained by (district-level recent) rainfall levels?
1 f) (5 Points) Do you think multicollinearity would be a relevant problem for the main regression model, as implemented in Column 2? How would you know? What would you do to resolve this if it is indeed a problem?
1 g) (15 Points) Based on the results in Table 1, do you think that Chinese tariffs are likely to have affected the 2020 US Presidential elections? Explain why (or why not). Finally, write out a regression model that you would use to assess the effectiveness of Chinese tariffs in affecting the electoral outcome: explain what measure you would use for your dependent variable, your main independent variable, and any other factors you consider important to include as controls along with an explanation of why.
To answer this question, first download the data set “final_IR602_Spring2021_data.dta” from the Final Exam folder on our course’s Blackboard site.
To what extent does one’s status as the main income earner (the “breadwinner”) limit political ambition for American women? To answer this question, take a look at the data set provided, which studies whether or not women decide to run for office out of a large number who completed a training program to help them compete in and win local electoral races in the US.
2.a) (5 Points) Calculate summary statistics on how much the decision to run for office (“ran”, where those who decided to run are coded as 1) varies by breadwinner status (“breadwinner”, where those contributing over 50% of the household income are coded as “breadwinners”, e.g. 1). To do this, utilize the tabstat command, where you can run the command: tabstat y, by(x) to see how an outcome variable
(y) varies by level of an independent variable (x). Remember, if you want to look at specific statistics, such as mean, standard deviation, and number of observations, you could specify: tabstat y, by (x) stats(mean sd n).
2.b) (5 Points) Write a bivariate regression model you would use to study the relationship between breadwinner status and the decision to run for office, using the basic format: π¦i = π½% + π½1πi + πi.
2.c) (5 Points) Run the regression in Stata. Interpret the main coefficient of interest.
2.d) (15 Points) The authors consider a more nuanced measure of women’s economic status in the household based on a finer-grained coding of their economic contribution as either 0-25% of household income, 25-50%, 50-75%, or 75-100% of the household income (“hhexpenses”). Do you think this is likely a better or worse measure of the relationship between women’s economic status within the household and political ambition? Why? Answer, then run this bivariate regression and interpret the coefficient on the result. Is this what you expected? Why/not?
2.e) (15 Points) What is a likely source of omitted variable bias? Why? What would be the direction of bias you would predict? Take a look at the variables within the dataset and figure out the closest version of this variable available. Write out the equation for the multivariate regression that may solve this form of omitted variable bias. Run the regression. What do you find? Does it present evidence in favor of or contrary to your hypothesis?
2.f) (15 Points) Consider the two key measures of “goodness of fit” for the regression models in parts 2.c), 2.d), and 2.e) (that is, the R-squared and Root MSE). What do these look like for each specification? How do the three specifications compare, specifically which model provides the “best” fit?
2.g) (10 Points) Considering the regression models you ran in parts 2.c), 2.d), and 2.e), what would you conclude about the relationship between women’s economic status within the household and their political ambition? Why?
CS 340 Milestone One Guidelines and Rubric Overview: For this assignment, you will implement the fundamental operations of create, read, update,
Retail Transaction Programming Project Project Requirements: Develop a program to emulate a purchase transaction at a retail store. This
7COM1028 Secure Systems Programming Referral Coursework: Secure
Create a GUI program that:Accepts the following from a user:Item NameItem QuantityItem PriceAllows the user to create a file to store the sales receip
CS 340 Final Project Guidelines and Rubric Overview The final project will encompass developing a web service using a software stack and impleme
