The empirical results from this stu

The empirical results from this study suggest
that food price and availability measures play
a limited role in the weight outcomes of U.S.
adult men and women, in general, although
there are some significant effects for certain
subpopulations. Fast-food prices are not found
to be associated with male or female adult
weight in any of the models. This finding is different
from that of Chou,Grossman,and Saffer
(2004),who find adult obesity prevalence to be
sensitive to fast-food prices in a large sample of adults from a different data set; although they
do not separately control availability of fastfood
restaurants versus full-service restaurants,
their contextual factors are matched at the
broader state level, and they control for foodat-
home prices rather than fruit and vegetable
prices.
We find that higher fruit and vegetable
prices are related to higher adult female BMI
(elasticity of 0.03) but not statistically significantly
related to male BMI. In particular,
higher fruit and vegetable prices are related to
higher BMI among poor women (elasticity of
0.09) andwomenwith children (elasticity 0.04).
These findings among the subpopulations mirror
recent evidence of the sensitivity of children’s
weight to fruit and vegetable prices,
particularly among children in homes with low
socioeconomic status (Powell and Bao 2009;
Sturm and Datar 2005). The results suggest
that targeted subsidies to support fruit and vegetable
purchases, such as those recently introduced
through the Women, Infants, and Children
program (Oliveira and Frazao 2009) or
potential subsidies that could operate through
the Food Stamp program, may help to reduce
obesity and overweight prevalence among lowincome
women and those with children, who
have significantly higher obesity rates than
their respective nonpoor and without-children
counterparts,may face tighter constraints upon
food shopping, and may be likely to substitute
across food items.
No statistically significant effects of foodrelated
outlet availability on BMI are found
for men, and small significant effects are found
for women for convenience stores and fastfood
restaurants in the full sample.Among the
subpopulation estimates, the only statistically
significant effect of noteworthy magnitude
is found for the availability of supermarkets
among poor women. Previous research
has documented that low-income communities
have lower availability of supermarkets compared
with middle- and high-income neighborhoods
(Larson, Story, and Nelson 2009).
The results suggest that policies to improve
access to supermarkets in such underserved
poor neighborhoods would help to reduceBMI
among women living in poverty.
In building the evidence base on the potential
effectiveness of using fiscal pricing and
other policies to support access to healthy food
to help tackle the U.S. obesity epidemic, future
research should continue to exploit longitudinal
studies. Furthermore, evaluation of policy
changes that provide subsidies to certain food items (such as those recently implemented
through the Women’s, Infants, and Children
program),or other potential policies that might
subsidize the cost of healthy foods (e.g., the
Food Stamp Program), will provide helpful
empirical evidence.