Table 2: Summary Statistics for Defense-Welfare Model
Standard
Variables
Mean
Deviation
Minimum
Maximum
Dependent Variables
Defense
-2.17
0.83
-4.98
0.01
Health
-2.60
0.92
-6.50
-0.24
Education
-1.99
0.84
-6.11
-0.58
Independent Variables
Minor Conflict
0.14
0.35
0
1
Major Conflict
0.09
0.28
0
1
POLITY score
3.58
7.12
-10
10
Per Capita GDP
8.54
6.96
0.28
33.29
Note: N =1639
estimators if the equations have contemporaneously correlated error terms (Zellner 1962). A
SUR model is appropriate because, to the extent that the estimates are influenced by factors
not included in the model, it is important to allow the error terms to be correlated. Hence
I estimate the following three equations simultaneously:
Def ense = β
0
+ β
1
Democracy + β
2
M inor Conf lict + β
3
M ajor Conf lict + β
4
GDP + µ
d
Health = γ
0
+ γ
1
Democracy + γ
2
M inor Conf lict + γ
3
M ajor Conf lict + γ
4
GDP + µ
h
Education = δ
0
+ δ
1
Democracy + δ
2
M inor Conf lict + δ
3
M ajor Conf lict + δ
4
GDP + µ
e
In the next section, I present the results for this model, and discuss the implications of
budgetary trade-offs for public health in states that are involved in armed conflict.
4.2
Results
The results for the model evaluating budgetary trade-offs between defense and welfare ex-
penditures due to conflict are shown in Table 3. This analysis provides strong support for
18