Econometrics ECON 320
Final Research Project: Due11:59 pm Wednesday, May 8th
Step 1: Choose your Data Source: Either Time-Series Macro Data or Cross-Sectional World Bank Data
Email your choice to me no later than Wednesday, April 23rd. Include the names of all group members, if applicable.
Step 2: Choose your Dependent Variable
Step 3: Choose at least 5 explanatory variables Step 4: Gather and clean your data
Email your Stata .dta file to me as an attachment no later than Friday, April 26th. Make sure you have at least 40 observations for each variable included. Include in your email a description of your variables.
If you are using World Bank data, collapse the sample to a single year cross section [¡°keep if year==2015¡± for example]
Step 5: Descriptive Statistics
Create a table with basic descriptive statistics for each variable (using the ¡°sum¡± command). Construct and comment on relevant scatter plots, time series plots, and histograms.
Step 6: Initial Regression
Step 7: Generate at least one dummy variable Step 8: Complete Regression Model
Include all suggested transformation, your dummy variable, and at least one interaction term.
Write an interpretation of your key coefficients, goodness of fit, statistical significance, and variance inflation factors.
Step 9a: If using World Bank data: Test for Heteroskedasticity, use corrections if appropriate. Step 9b: if using Federal Reserve data: Test for Autocorrelation, use corrections if appropriate.
Step 10: Carefully consider the issue of endogeneity bias. Which of your RHS variables is likely to be correlated with the error term? What does that mean? For extra super bonus points, use a 2SLS/IV approach to address this issue.
Step 11: Write a conclusion that summarizes what we have learned about the data generating process for your dependent variable.
The end result will be a paper, not less than ten pages (including appropriately sized graphs and Stata tables) that includes your attempts to address each of the issues above. The paper should be organized as a single narrative, with Stata output and graphs embedded within your explanations and interpretations. No hand-written work allowed – use the equation editor in Word to express mathematical work if necessary.