VoxEU Column Politics and economics

What determines democracy? And what helps to maintain it?

Will democracy establish itself in the Middle East? This column looks at what is needed to start democracies are what is needed to keep them going. It argues that that it is the level of economic development – not short-run economic growth – that is needed for democracy to survive.

Democracy is on the move in the Arab world. Whether democratic regimes will emerge and survive remains an open question, and the intense media coverage of the Arab Spring has revived public interest in the determinants of democracy. The quest to understand why democracy emerges and survives, however, has long been on the agenda of economists and political scientists.

Scientific findings suggest that life is better with more democracies in the world. People living under democracy have higher incomes and tend to enjoy more human rights than people living under authoritarian regimes (Hathaway 2003).1 Democracies also don’t go to war against each other, and make better trade partners (Russett and Oneal 2001).

So the promotion of democracy is laudable, but do we really know how to ‘cause’ democracy? Scholars have proposed and empirically tested many theories. Yet, no conclusive picture has emerged. Moreover, it is not clear whether the same set of variables that drives the democratisation process also guarantee its sustainability.

Economic explanations of democracy date back to Lipset (1959) who is often cited as father of 'modernisation theory'. The theory contends that as countries develop economically, social structures become too complex for authoritarian regimes to manage. At some point, dictatorship collapses and democracy emerges as the alternative. Przeworski et al. (2000) contend, however, that the emergence of democracy is random with respect to economic development. The correlation between development and democracy is driven instead by the survival of democracy. Przeworski (2005, 253) argues that "democracy prevails in developed societies because too much is at stake in turning against it." Conversely, in poor democracies, "the value of becoming a dictator is greater and the accumulated cost of destroying capital stock is lower" (Przeworski and Limongi 1997, 166). So democracy emerges idiosyncratically, but it survives in countries with high levels of economic development.

If democracy emerges at random, however, why has it so far failed to appear throughout most of the Middle East? Confronting this observation, some people retreat to culturalist explanations, claiming that Islam is incompatible with democracy. Others argue that the presence of oil is the problem. Ross (2012) argues that there is a political “resource curse" whereby the rents enable dictatorial regimes to use low taxes and high spending to maintain power.

Another approach that has implications for the Middle East considers the effects of ‘diffusion.’ Consider the work of Gleditsch (2002), summarised by the title of his book: All (International) Politics are Local. The political regime of one country may have a connection with regimes of neighbouring countries, through various political, cultural, and economic forces that spill over national borders. Thus we have solidly democratic regions, such as Europe, dictatorial regions, such as the Middle East, and regions where countries transition together in waves, such as Latin America. Pevehouse (2005) suggests that the key facilitating these spill overs is participation in regional international organisations.

So, we have many stories of democracy. This highly stylised and brief sketch of the literature indicates a vigorous debate. In research forthcoming in Journal of Conflict Resolution, we cite many more studies of democracy and identify 59 factors that have been proposed to cause democracy (Gassebner et al. 2012). We then pit these factors up against each other in a series of logistical regressions. Our approach is extreme: We evaluate over 1.7 million regressions of the emergence of democracy, and over 1.4 million regressions for the survival of democracy. Our method, Extreme Bound Analysis (EBA), involves the analysis of various combinations of control variables. The basic idea is to consider many regressions, continuously permutating through combinations of explanatory variables, focusing on one particular independent variable at a time, testing how its statistical significance conditional on other variables 'behaves'.

The most striking finding of our analysis is that most of the variables suggested in the literature do not survive EBA. While many of these factors are shown elsewhere to have statistically significant effects in plausible and well-specified models, when put to the rigors of being tested alongside many other variables, they fail our EBA test. We do not suggest that this implies these factors are unimportant. Many of the findings that fail our EBA test are valid within the confines of the original statistical model proposed in the literature, and our approach focuses on reduced forms of a specific model. Moreover, to the extent that some variables fail our test, this could be because they are poor proxies for otherwise strong theories of democracy. The standard of surviving our particular EBA is just a high one, and only the strongest of statistical relationships survive it.

Some variables do indeed survive. Regarding transitions to democracy, we find that economic growth has a robust negative effect. This finding, standing in stark contrast to modernization theory, suggests that autocracies with strong economic performance are unlikely to see democracy emerge. Instead, economic contraction causes dictatorships to break down. Also in contrast to modernization theory, but consistent with the argument of Przeworski et al. (2000), the level of GDP per capita does not have a robust relationship with the emergence of democracy. Next, we find evidence that membership in the OECD has a positive effect, but this connection could be endogenous or due to reverse causality. Finally, previous regime transitions also increase the likelihood of the emergence of democracy. The only other variables for which we find any evidence of a robustly significant effect are fuel exports and the share of the population that is Muslim. Both lose some statistical significance when included exclusively with the other highly robust variables, however, and we suspect that this may be a Middle East fixed effect. Tests certainly show that the effect of Islam is driven by fuel exports; Islam has no effect among non-fuel exporting Muslim countries.

Regarding the survival of democracy, the most robust determinants are level of economic development (a positive effect) and the number of past transitions (a negative effect). There is also some evidence that having a former military leader as the chief executive has a negative effect while having other democracies as neighbours has a positive effect. But these last two findings lose significance in the presence of the number of past transitions.

We conclude that there are many plausible theories of democracy but few robust predictors. For scholars, we suggest that these specific variables should be included in their models of regime transitions. As for policymakers, they should note that there is little that one can do to promote democracy. One thing is clear: promoting investment in dictatorships – encouraging vigorous economic growth – is not likely to cause a dictatorial regime to fall. When it comes to promoting the survival of nascent democracies, raising the level of development helps.

Applied to the newly developing democracies in the Maghreb region, our results suggest that the international community should help to ensure the economic success in these states. Note, however, that level of economic development – not short-run economic growth – causes democracy to survive. This finding suggests the importance of a long-run commitment to engaging these nascent democracies economically.


Gassebner, Martin, Michael J Lamla, James R Vreeland (2012), “Extreme Bounds of Democracy”, Journal of Conflict Resolution.

Gleditsch, Kristian Skrede (2002), All International Politics is Local: The Diffusion of Conflict, Integration, and Democratization, University of Michigan Press, Ann Arbor.

Hathaway, Oona (2003), “The Cost of Commitment”, Stanford Law Review, 55:1821–62.

Lipset, Seymour Martin (1959), “Some Social Requisites of Democracy: Economic Development and Political Legitimacy”, American Political Science Review, 53:69-105.

Pevehouse, Jon C (2005), Democracy from Above, Cambridge University Press.

Przeworski, Adam (2005), “Democracy as an equilibrium”, Public Choice, 123:253-273.

Przeworski, Adam, Michael Alvarez, José Antonio Cheibub, and Fernando Limongi (2000), Democracy and Development, Cambridge University Press.

Przeworski, Adam and Fernando Limongi (1997), “Modernization: Theories and facts”, World Politics, 49:155-183.

Ross, Michael L (2012), The Oil Curse, Princeton University Press.

Russett, Bruce and John R Oneal (2001), Triangulating Peace, WW Norton.

Papaioannou, Elias and Gregorios Siourounis (2012), “Democratisation and growth”, VoxEU.org, 25 October.

on “Democratisation and growth” (link: http://www.voxeu.org/article/democratisation-and-growth-within-country-c...)

1 See also the Vox column by Papaioannou and Siourounis (2008).

1,890 Reads