19
ownership of the territory.
33
The key dependent variables are the relative salience score
and domestic vulnerability. For the latter variable, we use several measures that we
incorporate from the 1999 Reinhardt et al dataset on domestic vulnerability.
34
These
measures capture the loss of popular support for leaders, the existence of contested
institutions as well as domestic strife. For the purposes of this paper, we specifically
focus on leader popularity.
35
In line with Reinhardt’s model, we include the first order
lagged values of measures of domestic strife to address the possibility of reverse
causality.
36
Following on the work of Huth 2000
37
on territorial conflict, we include the
following control variables:
C= constant
X
1=
challenger democracy
X
2=
target democracy
X
3=
contiguity
X
5=
temporal count
µ
t=
error term
We have decided to include these control variables firstly because X
1,
X
3
are
conventional controls used in the interstate dispute literature. X
1
and X
2
capture dyadic
democracy
,
a concept that has its basis in the findings of the democratic peace literature;
namely that democracy-democracy dyads are less likely to resort to the use of military
33
ICOW codebook, 1-2; codebook available at http://garnet.acns.fsu.edu/~phensel/icow.html
34
Reinhardt et al. 1999. The dataset is publicly available at
http://userwww.service.emory.edu/erein/data/#Violence.
35
We limit our focus to leader popularity both for purposes of narrowing down the scope of study but also
because of missing data problems associated with the other domestic vulnerability measures.
36
Reinhardt et al. 1996; namely, reverse causality implies the possibility that a state’s pursuit of territorial
challenge may cause domestic strife.
37
Huth 2000, 70. See also Hensel, 2000.
Convention