Unformatted text preview: d panel, 19602000, with data at 5year intervals, where the start date of the panel refers to the
dependent variable (i.e., t=1960, so t1=1955). Columns 5 and 10 use annual data from the same sample; Freedom House data begin in 1972, as we require 5 lags, this panel starts in 1977. In
columns 5 and 10, each right hand side variables has five annual lags; we report the pvalue from an Ftest for the joint significance of all 5 lags. For detailed data definitions and sources see
Table 1 and Appendix Table A1. Back in History
Of course if there is a contemporary relationship between the levels, if
you go back far enough in time, there has to be some sort of positive
correlation between the growth rates. Figure 5 shows this is true for
the change in income and the change in democracy since 1500.
Does this prove that the modernization hypothesis is true but with
“long and variable lags” to quote Milton Friedman out of context!
We argue no. Instead we propose that di¤erent societies move onto
di¤erent development paths where income and democracy coevolve.
Some circumstances lead a society...
View
Full
Document
This document was uploaded on 02/28/2014 for the course ECON 2328 at Harvard.
 Fall '09
 JamesRobinson

Click to edit the document details