Objective. This manuscript tests competing explanations for the distri
bution of environmental risks produced by Toxics Release Inventory (TR
I) facilities. In particular, I evaluate claims that risks from TRI fa
cilities are concentrated in poor and minority areas, holding constant
multiple background characteristics that may also account for the dis
tribution of these risks. Methods. A series of probit, negative binomi
al event count, and weighted least squares models are used to account
for the distribution and density of TRI facilities and the concentrati
ons of TRI pollutants in all residential ZIP codes across the United S
tates. Results. Even when controlling for background factors, TRI faci
lities and pollutants are concentrated in residential ZIP codes with l
arge minority populations. Moreover, while other background characteri
stics provide the most powerful explanations for the distribution and
density of TRI facilities and the concentration of TRI pollutants, the
racial attributes of neighborhoods-more than their class attributes-b
est account for the remaining patterns in the distribution of environm
ental risk. Conclusions. These results provide additional evidence tha
t racial minorities experience disproportionate proximity to a wide va
riety of environmental risks. The results also suggest the racial comp
osition of residential neighborhoods as a relevant criterion when appr
oving the location of new TRI facilities.