38
Appendix
This section details the presidential election forecast procedure and its relation to the
impact of partisan shocks. In order to impute an election prediction I use a slightly modified
quarterly version of Fair’s (1998) election forecast model:
Vote = b
0
+ b
1
Inc + b
2
GDP
PER CAPITA
*Inc + b
3
GDP
DEFLATOR
*Inc + b
4
n*Inc
+ b
5
IncRun + b
6
IncDur + b
7
Trialheat,
(A1)
where Vote is the Democratic share of the two-party presidential vote; Inc is 1 if there is a
Democratic incumbent at the time of the election and -1 if there is a Republican incumbent;
IncRun
is 1 if a Democratic incumbent is running for election, -1 if a Republican incumbent is
running for election, and 0 otherwise; IncDur is 0 if the incumbent party has been in power for
one term, 1 for Democrats and -1 for Republicans if the incumbent party has been in power for
two consecutive terms, 1.25 for Democrats and -1.25 for Republicans if the incumbent party
has been in power for three consecutive terms, 1.50 and -1.50 for four consecutive terms, and
so forth; GDP
PER CAPITA
represents the growth rate of real per capita GDP during the three
quarters prior to the current quarter (t-3); GDP
DEFLATOR
is the absolute value of the annual
growth rate of the GDP deflator in the prior quarters of the administration (GDP data are from
the Bureau of Economic Analysis); n is the number of quarters during the current
administration in which the growth rate of real per capita GDP is greater than 3.2 percent at an
annual rate; and Trialheat is the Democrat’s trial-heat result which is entered for the 15
th
and
16
th
quarters of the election cycle, that is, the two quarters immediately prior to the election.
Gallup trial heat data collected prior to September of the election year have not proven to be of
serious value in election prediction efforts (Campbell and Garand 2000).
Convention