There is No Evidence that Minimum Prices in British Columbia Had an Impact on Total Alcohol Consumption
6 Pages Posted: 9 Jun 2013 Last revised: 31 Jul 2013
Date Written: July 31, 2013
Stockwell, Auld, Zhao, and Martin (2012) claim that minimum alcohol prices had a signicant impact on total alcohol consumption in British Columbia, Canada. To show this they perform a regression with the log of total alcohol consumption as the dependent variable and the log of the average of minimum alcohol prices as one of the independent variables using quarterly data from 1989 to 2010. We show that these two variables are integrated processes. A KPSS-test (Kwiatkowski, Phillips, Schmidt, and Shin (1992)) leads to the rejection of the null of stationarity at the 5% level of signicance for both time-series. It is well known that regressing one integrated time-series against another integrated time-series can produce spurious correlation (see e.g. Wooldridge (2009, Chapter 11.3)). We then perform the same regression as in Stockwell, Auld, Zhao, and Martin (2012) but in annual dierences (i.e. in yearly changes). In this regression (changes in) minimum prices have no signicant eect on (changes in) total alcohol consumption for the given data (p-value 0.2166). Therefore, there is no evidence that the way minimum prices were set had an impact on total alcohol consumption in British Columbia.
Keywords: total alcohol consumption, minimum prices, British Columbia
JEL Classification: I18, C20
Suggested Citation: Suggested Citation