Table 3. Price regression model (log price/mL) with date of implementation and date of announcement.

	All	High-Tax	Low-Tax	No-Tax
Point estimate (implement)	−0.010	−0.008	−0.017*	0.017
Standard error	0.01	0.01	0.01	0.02
Proportionate change (implement)	−1.0%	−0.8%	−1.7%*	1.7%
Point estimate (announcement)	0.016**	0.019*	0.013	0.001
Standard error	0.01	0.01	0.01	0.02
Proportionate change (announcement)	1.6%**	1.9%*	1.3%	0.1%
Mean outcome (log price/mL)	0.388	0.507	0.278	0.339
Number of Products	816	386	384	46
Number of Observations	37,699	19,693	15,632	2,374

Implement refers to September 2014 and Announcement refers to March 2014. Proportionate change = exp(point estimate) − 1.

*P < 0.05

**P < 0.01

***P < 0.001.