The differences in the political systems scales reported in this website primarily reflect the degree of democracy within each country. This dimension is measured using four items:
D1ij is the difference between countries i & j using the POLCON V ‘political constraints’ scale
D2ij is the difference between countries i & j using the Modified POLITY IV democracy-autocracy scale
D3ij is the difference between countries i & j using the Freedom House Political Rights scale , and
D4ij is the difference between countries i & j using the Freedom House Civil Liberties scale .
In the previous version of our scales, we did include a second aspect of differences in political systems – the political ideology of the group in power. Unfortunately we only have one single indicator for this dimension, and the scale in general performed poorly in terms of predicting international business outcomes.
The scores for each of the four democracy indicators and the resultant democracy factor (see below concerning the confirmatory factor analysis), can be found in an Excel spreadsheet attached at the bottom of this page. This spreadsheet contains the values for 22,350 country pairs (i.e. n x n-1 for 150 countries) for three specific time periods (1995, 2005 and 2015). The precise coding for these variables is explained below. Please note that in accordance with the arguments and analyses presented in our JIBS paper, the absolute value of this dimension is the appropriate form if you are using them as indicators of psychic distance.
POLITICAL SYSTEM INDICATOR CALCULATIONS
D1ij = (D1i – D1j)
Where, D1i = Henisz’s (2000) POLCON V measure (0 to 1, 3 year average) for the exporting country (i)
And, D1j = Henisz’s (2000) POLCON V measure (0 to 1, 3 year average) for the importing country (j)
D2ij = (D2i – D2j)
Where, D2i = Bollen’s Polity IV – Polity measure (-10 to +10 scale, 3 year average) for the exporting country (i)
And, D2j = Bollen’s Polity IV – Polity measure (-10 to +10 scale, 3 year average) for the importing country (j)
D3ij = (D3i – D3j)
Where, D3i = the Freedom House Political Rights scale (1-7; reverse coded, 3 year average) for the exporting country (i)
And, D3j = the Freedom House Political Rights scale (1-7; reverse coded, 3 year average) for the importing country (j)
D4ij = (D4i – D4j)
Where, D4i = the Freedom House Civil Liberties scale (1-7; reverse coded, 3 year average) for the exporting country (i)
And, D4j = the Freedom House Civil Liberties scale (1-7; reverse coded, 3 year average) for the importing country (j)
Dem F – is the single-factor solution, using principal component analysis, for D1ij, D2ij, D3ij and D4ij
Dem F (abs) = absolute value of Dem F
Dem Dist – is Dem F (abs) re-scaled such that ‘no differences’ equals zero (0), and the maximum equals ten (10)
Dem f– Difference in Democracy Factor:
The four democracy indicators have be reduced to a single factor using confirmatory factor analysis (cfa). This factor score has been estimated using the full set of country pairs (22,350) and time periods (3). The individual factor loadings and the Cronbach alpha are reported below.
Cronbach Alpha |
Factor Loading |
|
Dem f – 4 item factor score for differences in the degree of democracy | 0.956 | |
D1ij – Difference in the POLCON scale | 0.890 | |
D2ij – Difference in the Modif POLITY IV scale | 0.945 | |
D3ij – Difference in the FH Political Rights scale | 0.970 | |
D4ij – Difference in the FH Civil Liberties scale | 0.955 |
SOURCES
The primary sources for these estimates are:
-
-
- Henisz, W. J. (2000) ‘The Institutional environment for Economic Growth’, Economics and Politics 12(1): 1-31. [www document] http://weber.ucsd.edu/~kgledits/Polity.html (accessed 4 April 2018).
- Gleditsch, K. S. (2003) ‘Modified Polity P4 and P4D Data’, [www document] http://weber.ucsd.edu/~kgledits/Polity.html (accessed 4 April 2018).
- Freedom_House (2000) ‘Freedom in the World’, [www document] http://www.freedomhouse.org (accessed 5 April 2018).
-
MISSING DATA
For the four democracy metrics, missing data represents 3.2% of the cells. In all cases, given the confirmatory nature of the items, the missing item was estimated via a multivariate regression of the other democracy items in the same time period.