FDI as the independent variable). If the latter model proves to be a better predictor
than the former, it will suggest that rational observation of the attractiveness of flat tax
countries does contribute to the explanation of a flat tax revolution.
We use these variables to estimate two different classes of models. The first (Table
1) models the overall reform environment, as described in Hypothesis Six. For this,
we use measures of total revenue from privatization (Privatization), denominated in US
dollars (Brune, 2005); European Bank for Reconstruction and Development measures
of current-account liberalization (Current-acct liberalization) and the removal of price
controls (Price liberalization) (EBRD, 1990-2005), and measures of institutional reform
(Kostadinova, 2005). We also control for the ability to enact policy reform by including
a measure of veto players (Political Constraints) and for capital-account openness. We
also model the pull from Europe in terms of geography. We use the standard measure of
kilometers from the Rhein valley (Km from Rhein) as a proxy for the pull of European
commerce and culture, as is common in the literature (Sachs, 1999). The second (Table 2)
uses many of the standard control variables that test arguments about taxation levels —
but again, not adoption of particular taxation regimes. Both models include all measures
of diffusion as well as economically liberal government leanings (Ideology).
Results in Table One show the results of a logit model on the adoption of the flat tax
in a given country year (coded 0 if the tax was not adopted and 1 if it was).
[TABLE ONE ABOUT HERE]
As suspected, none of the reform measures here have consistently significant effects on
the probability of flat tax adoption, though capital-account openness does. An economi-
cally liberal ideology, however, is associated with statistically significant increases in the
probability of flat tax adoption. That the coefficient on ideology proves strong indicates
21