3. Create a PA dummy for stores in Pennsylvania and regress emptot1 on the two state dummies
PA and NJ for each year to replicate rows (1), (2), and (3) of Table 3 in the paper and report
your results in a table following the format in the paper. Describe the differences or similarities
between your results and the ones reported in the paper.
4. Replicate figure 1 of the paper. Based on the figure, can we say that the fast food restaurants
in these two states indeed obey the minimum wage law? (You can try the user written command
byhist )
5. Create the GAP variable as defined in the paper (page 779)
GAPi = 0 for stores in PA
= 0 stores in New Jersey with W1i ≥ 5.05
= (5.05 − W1i)/5.05 other stores in New Jersey,
which is the proportional increase in wages at store i necessary to meet the new mini mum rate.
Replicate Table 4 using emptot2 as the measure of employment. As stated in the paper, please only
use stores with available data on employment and starting wages in waves 1 and 2. Your regression
should follow equations (1a) ∆Ei = a+b·Xj+c·NJi+ei and (1b) ∆Ei = a+b·Xj+c·GAPi+ei
from the paper. Report your results in Table 2 in the provided template. In the last column of
the table, you don’t need to control for pa1 and pa2. How to interpret the coefficient on GAP?
Describe the differences or similarities between your results and the ones reported in the paper.
6. Instead of using ∆E as the dependent variable, let’s try the standard DID specification.
1Due: Thur,
Jun 20, 2024
ECA5313
Please submit both your answer and Stata codes
emptot2it = α0 + α1Dt + α2NJi + δDt × NJi + ϵit
Report the estimation results in column (1) of Table 3 in the provided template. Can you still
claim minimum wage has positive impact on employment based on this regression.
7. Let’s try the standard DID specification with store fixed-effect.
emptot2it = α0 + α1Dt + α2NJi + δDt × NJi + ηi + ϵit
You can use areg ..., a(id)
Report your results in column (2) of Table 3 in the provided template. Why does the computer
drop the NJ dummy in the regression? Does control for store fixed effect improve the results?
Table 1: Table 1 template
Stores by state Stores in NJ Differences within NJ
PA NJ NJ - PA
FTE, before
FTE, after
Change in mean FTE
Table 2: Table 2 template
(i) (ii) (iii) (iv) (v)
NJ
GAP
Control for chain and ownership no yes no yes yes
Control for region no no no no yes
2Due: Thur,
Jun 20, 2024
ECA5313
Please submit both your answer and Stata codes
Table 3: Table 3 template
(1) (2)
NJ
Post policy
NJÖ(post policy)
Control for store fixed effect No Yes
infix ///
sheet 1-3 ///
chain 5-5 ///
co_owned 7-7 ///
state 9-9 ///
southj 11-11 ///
centralj 13-13 ///
northj 15-15 ///
pa1 17-17 ///
pa2 19-19 ///
using public.dat, clear;
#delimit cr
bys sheet: gen n=_N
tab n // Check the number of obs per sheet
gen id = _n // Since we know each row corresponding to one restaurant