might make some choices closer substitutes for certain investors. In this
paper we do not address this problem.
6For that reason, one would, for example, expect two adjacent counties
to be closer substitutes than two adjacent states.
7Usually, by including Census Divisions dummies in studies dealing with
choices across the U.S. states.
8Note that what were choices in the CLM are now observations.
9Note that the sector specific characteristics drop out of the next expres-
sion.
10 For example, in a state choice set analysis, introducing dummy variables
for the nine Census Divisions is equivalent to admitting that each investor
restricts his choice set to the particular Census Division where the invest-
ment was observed. The demonstration is provided in Appendix A. Note
also that, in light of this relation, introducing dummies variables for groups
of elemental alternatives is equivalent to estimating the lower levels of a
two-step limited information nested logit.
11This model is extensively reviewed in McFadden & Train (2000).
12See Chen & Kuo (2001)for a proof of this result.
31