Homework Assignment #5
We are going to test how the strictness of employment protection affects the level of unemployment in
a panel of OECD countries.
1. Read in OECD data for real (1) GDP growth, (2) strictness of employment protection (individual
and collective), and (3) the unemployment rate. All at yearly frequency.
Hint: The package ‘OECD’ lets you do this from within R.
2. Create a panel data set from these variables. Create a nice plot/plots of the data. Nicest plot
gest one bonus point!
3. Run the following models
a. (pooled 1) 𝑢𝑖𝑡 = 𝛽0 + 𝛽1𝐸𝑃𝑖𝑡 + 𝜀𝑖𝑡
b. (pooled 2) 𝑢𝑖𝑡 = 𝛽0 + 𝛽1𝐸𝑃𝑖𝑡 + 𝛽2𝑌𝑖𝑡 + 𝜀𝑖𝑡
c. (FE 1) 𝑢𝑖𝑡 = 𝛽0 + 𝛽1𝐸𝑃𝑖𝑡 + 𝛼𝑖 + 𝜀𝑖𝑡
d. (FE 2) 𝑢𝑖𝑡 = 𝛽0 + 𝛽1𝐸𝑃𝑖𝑡 + 𝛽2𝑌𝑖𝑡 + 𝛼𝑖 + 𝜀𝑖𝑡
e. (twoway 1) 𝑢𝑖𝑡 = 𝛽0 + 𝛽1𝐸𝑃𝑖𝑡 + 𝛼𝑖 + 𝜏𝑡 + 𝜀𝑖𝑡
f. (twoway 2) 𝑢𝑖𝑡 = 𝛽0 + 𝛽1𝐸𝑃𝑖𝑡 + 𝛽2𝑌𝑖𝑡 + 𝛼𝑖 + 𝜏𝑡 + 𝜀𝑖𝑡
4. Briefly explain each regression. What are the strengths weaknesses of each regression, are the
results as expected? How do you explain the difference in sign between pooled and fixed
effects? Does model f. make sense? Which is the best model?
5. Formally test if
a. you should include country level fixed effects.
b. How about two-way fixed effects?
c. Should GDP be included in the regression?
d. How about lagged GDP 𝑢𝑖𝑡 = 𝛽0 + 𝛽1𝐸𝑃𝑖𝑡 + 𝛽2𝑌𝑖𝑡−1 + 𝛼𝑖 + 𝜀𝑖𝑡