Liberal democracy and sustainability

Abstract

Although theory strongly suggests that liberal democracies should perform better than autocracies on sustainability indicators, the empirical evidence is unclear. Using multivariate statistical techniques, political factors such as the stability of the system, public opinion, the nature of the party system and other institutional features such as presidentialism are considered. Liberal democracy is given qualified endorsement: it typically promotes weak sustainability, and stable core autocracies perform worse on strong sustainability measures than stable core democracies. Presidentialism generally is bad for sustainability. However, there is no compelling evidence that public opinion matters, even allowing for the intervening effects of the party system and institutional structure, which raises questions about the nature of the democratic process.

Keywords:

Notes

 1. An alternative is the World Economic Forum's sustainability index (GLTETF 2002), but Ward (2006) shows that it is inappropriate to add the components used as they do not satisfy the scalability criterion.

 2. Popgrow has a correlation of 0.42 (n = 125) with a nation's Gini coefficient, reported closest to the year 2000 in World Bank development tables. However the cross-national comparability of inequality measures is doubtful.

 3. If democracy is added alongside demsq to regression 1, the former variable is nowhere near significant whereas the latter is still highly significant. Various lagged value of a nation's democracy score were added to regression 1, but no significant effects showed up.

 4. Across the analysis here interactions with the continuous version of the democracy variable are generally weaker.

 5. Stabilility correlates highly with Kaufmann et al.' s (2004) composite measure of efforts to deal with corruption (0.79, n = 141) and moderately highly with the World Economic Forum's (GLTETF 2002) measure of capacity for environmental governance (0.57, n = 141). So part of what is being captured here might concern the effect of good governance on sustainability.

 6. Models for gensav generally suffered from heteroskedasticity, hence the reported significance scores are based on robust standard errors.

 7. I found no evidence for non-linear or more powerful lagged effects.

 8. The only comparable survey is the 2002 Gallup Voice of the People Poll carried out in 2002. The N is considerably lower at 31. Available from www.voice-of-the-people.net, accessed 24 April 2008.

 9. Envtaxsup and envincsup are based on thermometer scales where respondents were asked whether they strongly agreed (=1), agreed (=2), disagreed (=3) or strongly disagreed (=4) with a statement that taxes or higher prices should be used to fund environmental cleanup. After averaging I inverted the scales by subtracting the scores from 5.

10. I used Szajkowski (2005), the Psephos election archive and Electionworld. Results were cross-checked with calculations on the Global Greens website and in Müller-Rommel, 2002. Minor discrepancies were resolved by averaging. When it was clear that a green party had re-emerged or had a clear successor party, electoral performance of earlier incarnations was included.

11. Fredriksson et al. (2005) argue that the lack of competitiveness of a political system, measured by the vote share of the largest party, should matter; and they find some evidence that it does so in relation to one environmental indicator.

12. Presidential is only significant in democracies. Frac has a marginally significant positive effect in non-democracies, but this is hard to interpret.