Skip to main content
. Author manuscript; available in PMC: 2022 Dec 1.
Published in final edited form as: J Public Econ. 2021 Nov 13;204:104520. doi: 10.1016/j.jpubeco.2021.104520

Table A14:

Log Ethanol Sales (Not seasonally adjusted)

(1)
Main
(2)
Donut
(3)
Synth
(4)
Synth-Donut
Treatjt 0.004 0.005 −0.017*** −0.018***
Cluster SEs (0.010) (0.010) - -
Donald-Lang SEs (0.011) (0.011) (0.006) (0.005)
DL - Newey - West SEs (0.017) (0.017) (0.008) (0.007)
Observations 879,270 845,774 210 202
R-squared 0.934 0.934 0.043 0.056
Store-Fixed-Effects Yes Yes No No
Year-Month-Fixed-Effects Yes Yes No No
Seasonal Adj. No No No No

Notes: The dependent variable is the natural logarithm of total ethanol sales (in gallons) in the store-week cell. Treatjt equals one for Illinois (treatment) stores from Sep 2009 on, and zero otherwise. Cluster SEs account for clustering at the state-level. Donald-Lang and DL-Newey-West: The dependent variable is the difference in the Illinois and the control states’ store-average, by week. Sample size is 210 weeks. Newey-West method adjusts for first-order autocorrelation with maximum lag set at 3 weeks. Donut-specifications leave out 8 weeks around policy change. In columns (3) and (4), the dependent variable is the difference between Illinois state aggregates and a synthetic control state aggregate.

***/**/*

indicate statistical significance at the 1%/5%/10%-level.