Yves here. This piece looks at events like Arab Spring and the revolutions of 1848 and concludes they were more successful than is commonly thought. When they did not succeed in overthrowing governments, they still led to reforms.
By Toke S. Aidt, Senior University Lecturer in the Faculty of Economics at University of Cambridge, Gabriel Leon, Lecturer at the Department of Economics, Bar Ilan University, Raphael Franck, Professor at the Department of Business and Economics, University of Southern Denmark, and Peter S. Jensen, Lecturer in Political Economy in the Department of Political Economy, King’s College London. Originally published at VoxEU
Some theories suggest that the threat of revolution plays a pivotal role in democratisation. This column provides new evidence in support of this hypothesis. The authors use democratic transitions from Europe in the 19th century, Africa at the turn at the 20th century, and the Great Reform Act of 1832 in Great Britain. They find that credible threats of revolution have systematically triggered pre-emptive democratic reforms throughout history.
The threat of Revolution Hypothesis
The wave of violent protests that swept across north Africa and parts of the Middle East during the Arab spring between 2010 and 2012 coincided with the fall of several long-established autocracies; in those that survived, policy reforms and redistributive policies aimed at calming the masses were hastily implemented. A century and a half before, something similar happened in western Europe. The revolutions in France and parts of Germany in 1848 were followed by democratic reforms in Denmark, Luxembourg, Belgium, and the Netherlands.
Episodes like these lend credence to the hypothesis that revolutions, riots, and other types of violent protest can trigger democratic change. The hypothesis is appealing because it resolves the franchise extension puzzle, namely why would incumbent autocrats with a monopoly on political power, and often on economic resources, agree to share their power with broader segments of the population whose goals they do not share? The threat of revolution hypothesis, developed in the work of Acemoglu and Robinson (2000, 2006) and Boix (2003) amongst others, suggests that autocrats might do so when they face a credible threat of revolution that, if successful, would eliminate their entire power base. Seen in this perspective, the reactions of autocrats in the Arab world today, and of monarchs in western Europe 150 years ago, are pre-emptive responses to a credible threat of revolution.
Not everyone agrees with this interpretation, however. In his discussion of democratic reforms between 1830 and 1930, Roger Congleton (2010, p. 15), for example, argues that: “In essentially all cases [countries], liberal reforms were adopted using pre-existing constitutional rules for amendment. In no case [country] is every liberal reform preceded by a large-scale revolt, and in most cases, there are examples of large-scale demonstrations that failed to produce obvious reform”.
Democratisation is a multi-faceted process and the challenge is to gauge the boundaries between the threat of revolution hypothesis and alternative explanations of the root causes of democracy. Establishing how far the threat of revolution hypothesis explains democratic transitions is, however, difficult for two reasons.
- Firstly, it is the threat of revolution that matters and that is hard to quantify.
- Secondly, if observed social protest is used as a proxy for the underlying threat, it is hard to establish if the threat causes democratisation or if both occur because of something else.
In a series of papers (Aidt and Jensen 2014, Aidt and Leon forthcoming, Aidt and Franck 2013, forthcoming), we use historical and recent data and a variety of identification strategies to engage with these issues. The picture that emerges from these studies, as well as from previous work by Przeworski (2009) and the recent study by Dorsch and Maarek (2015), is clear:
- The threat of revolution is one of the causal drivers of democratisation.
New Evidence of the Threat of Revolution
To answer the question posed in the title, it would be ideal to study the entire universe of democratic transitions. Since this is not possible, as researchers we must focus on particular time periods, countries, or even on specific reforms. By studying many of these, it is possible to learn a great deal about the link between violence, riots, and revolutions and democratisation. Our work has so far studied three ‘cases’: Europe during the long 19th century, sub-Saharan Africa at the turn of the 20th century, and Great Britain in the 1830s.
Each case identifies observable events which demonstrate to the relevant incumbent rulers that the collective action problem associated with launching an effective challenge against their hold on power has been temporarily resolved. We then track how they react to these events, and, in particular, if they adopt democratic reforms. Of course, it remains to establish a causal link between the two and each case needs a context-specific identification strategy.
International transmission of information about regime contention is one way to capture variation in threat perceptions. Kurt Weyland (2014, p. 120), writing about the European revolutions of 1848, uses the example of Denmark to illustrate this logic: “the Danish King had more time [than the Prussian King] to see the wave of contention coming and noticed the costly clashes in Vienna (March 13-15) and Berlin (March 18-19); on March 21, he therefore offered these changes [a liberal democratic constitution] to the crowds gathering outside his palace in order to pre-empt violence”. In Aidt and Jensen (2014), we study a panel of European countries during the first wave of democratisation (1820-1938). We explore the idea that actual revolutions in neighbouring countries served as signals to monarchs and potential revolutionaries in other countries about how threatening the situation was. We find a very robust relationship between these revolutionary events and suffrage reforms, which is stronger for countries that were linguistically or geographically close to the epicentres of the revolutions.
- Our estimates show that a revolution somewhere in Europe was associated with a 75% increase in the odds of a suffrage reform in the neighbouring countries.
Our work on democratic change in sub-Saharan Africa between 1990 and 2007 demonstrates that the link between the threat of revolution and democratisation is not unique to the first wave of democratisation in Europe (Aidt and Leon forthcoming). Here, we quantify the extent of regime contention with data on domestic riots. Since riots and political change are driven by multiple causes, we employ instrumental variable techniques to uncover the true effect of riots on democratic change. Inspired by Brückner and Ciccone (2011), Burke and Leigh (2010) and Franck (forthcoming), we use weather shocks (droughts) to instrument for political action. There are many reasons why droughts might lead to riots; for instance, the temporary reduction in income lowers the opportunity cost of contesting power, drought creates hardship in the countryside and leads to migration into cities that exacerbates existing tensions and worsens overcrowding, among others. We find that the probability of democratic change increases by 16.7 percentage points as a consequence of the impact drought has on riots.
While our studies of democratisation in Europe and Africa explore cross sections of countries over time, our study on the Great Reform Act of 1832 focuses on a major episode of democratic change in Great Britain (Aidt and Franck 2013, forthcoming). We explore the geographical dispersion of a rural uprising – the so-called Swing Riots – that took place in between the last two elections conducted under the rules of the ‘unreformed parliament’ in 1830 and 1831, respectively. We estimate the effect of riots that happened in the immediate neighbourhood of a constituency on the likelihood that it elected a reform-friendly politician to serve in the 1831 parliament which adopted the Great Reform Act. Of course, any correlation between local riots and the electoral success of reform-friendly politicians could be caused by many factors. We, therefore, explore that the riots spread through local social interaction effects along the pre-existing road network to isolate the exogenous variation in exposure to local riots. Our instrumental variable and matching estimates suggest that the two reform-friendly parties would not have obtained a majority in the House of Commons had it not been for the Swing Riots. Without such a majority, the reform process would have most likely been stopped.
The threat of a revolution plays a pivotal role in the theory of democratisation developed by Acemoglu and Robinson in a sequence of papers and in their book, The Economic Origins of Dictatorship and Democracy. The theory emphasises that democratisation happens at critical junctures in history. Our evidence supports this interpretation. However, our studies, and the threat of revolution theory itself, do not rule out that complex interactions between underlying, slow-moving economic processes, e.g., industrialisation, urbanisation, income growth, international trade, and inequality, and democratic triggers could be important. The ‘revolutionary shocks’, e.g., revolutions in other countries, exposure to local riots, or drought-induced protest, may push a country over a threshold and induce rulers to implement democratic reforms, but only if the underlying fundamentals of the economy are ‘close’ to the threshold to begin with.
See original post for references