All five studies that assessed the association
between outlet density and population-level alcohol
consumption found that they were positively associated;
increased density was associated with increased
consumption, and vice versa. Three studies examined
the relationship between outlet density and the consumption
of spirits in the U.S. The first study estimated
that, from 1955 to 1980, for each additional outlet
license per 1000 population, there was an increase of
0.027 gallons in per capita consumption of spirits
ethanol (p0.01).28 The second study reported an
elasticity of 0.14 (p0.01) for outlet density and spirits
for the period 1970–1975.31 The third study examined
the association of outlet density and the sale of spirits
and wine in 38 states over a period of 18 years; the
effects of consumption on density were separated out
by use of two-stage least squares regression. The elasticity
for spirits and wine was found to be 0.033 (NS) and
0.015 (NS), respectively.26
A study assessing trends from 1952 to 1992 in the
United Kingdom25 reported an elasticity of 2.43 (p
0.05) for off-premises density and beer consumption
but no significant association for other beverages (except
hard cider). Finally, a study33 examining data from
1968 to 1986 in Canada reported a significant association
between reductions in off-premises density and
reductions in alcohol consumption. This study also
found an association between changes in outlet
density and cirrhosis mortality, which was mediated
by changes in alcohol consumption. When the alcohol
consumption variable was added to the analytic
model, the coefficient for cirrhosis mortality was no
longer significant.