Cross-National Study of Opposition Media in Arab Authoritarian Regimes
5.3 Limitations of this Study and Suggestions for Future ResearchFuture Research
As any model must, the one I have presented makes many simplifying assumptions that limit its realism. Extended versions of the model can determine whether the results depend on such strong assumptions, as well as yielding additional and more nuanced empirical predictions.
The following are some extensions worth pursuing:
• Instead of assuming that the regime leader and opposition leader have the same information, assume that they see private but correlated signals of the probability of protest success. Such a model would yield insight into how the regime and the opposition might behave when the opposition believes protest is likely to succeed but the regime does not, and vice versa.
• In constructing the dataset, I came across many cases in which there were multiple significant opposition groups in a single regime-year. Often these groups differed significantly from each other in terms of their ideologies or their willingness to cooperate with the regime. An extended model could include multiple opposition leaders, each of whom has some chance of coming to power if the regime is overthrown. Some opposition players may prefer the continuation of the regime to a transition to rule by certain other opposition groups and their groups may have varying degrees of public support. In which cases will each opposition leader want to encourage or discourage protest, and when will the regime allow allow some of them to have media but not others? What inferences do citizens make if they see that some opposition groups have media and others do not?
• Lastly, the game could be repeated over multiple periods. We could assume that
whether the opposition has the resources to produce media in a given period is positively correlated with whether they had the resources in the prior period. In such a model, if the opposition have media in one period and not the next, citizens would put more weight in the later period on the probability that the government actively prevented the opposition from producing media, as opposed to the opposition lacking the resources, compared to a situation in which the opposition did not have media in either period. This effect may reduce the dictator’s willingness to allow opposition media in the first place. Alternatively, allowing an opposition media outlet in one period might enable the regime leader to more credibly commit to allowing it in the future.
Further research is needed to address some limitations of my case study of Tunisia under Ben Ali. I was not able to obtain interviews with as many individuals who held high-ranking positions in the regime during 1987-1992 as I was with leaders of opposition groups from that period. More interviews with former regime officials could reveal whether their beliefs and goals at the time were consistent with those in my model. Content analysis of opposition media from this period would also be illuminating. When the government was tolerating certain opposition media outlets, did those outlets call only for limited policy reforms in the short-run (even if the groups’ long-term goals included a regime transition or other sweeping change), as my theory predicts? Did they discourage or at least not actively call for anti-government protests? On occasions when opposition media did call for protests or immediate regime change, did Ben Ali’s government invariably shut them down, as he did with El-Fajr in June 1990?
In order to estimate causal effects, the cross-national analysis made use of regime fixed effects and period dummies. However, this is not a perfect identification strategy, as it does not control for possible omitted variables that vary over time within regimes. Future
research should look for exogenous shocks to unemployment, growth, fuel revenue, and other explanatory variables in order to identify their causal effects on toleration of opposition media with greater confidence. By obviating the need for regime fixed effects, this approach would also make it possible to test the effects of time-invariant regime characteristics, such as how the ruling group came to power.
Since the unit of observation in the cross-national analysis was the regime-year, it did not measure variation in governments’ treatment of media across different opposition groups in a given year, and how that such differential treatment is affected by group-specific variables, such as how recently the group was established, its ideological orientation (e.g. Islamist or secular), or whether the group advocated an eventual regime transition or only reform within the existing regime. Because I collected detailed information on the media of each of the opposition groups on which the dataset was based, it would be straightforward in future to expand the analysis to the regime-group-year level.
Again, content analysis is another promising avenue for research on this topic. For contemporary regimes, it would be instructive to analyze opposition groups’ accounts on international social media platforms like Facebook and Twitter, which are difficult for regimes to control. A testable implication of my theory is that, if we compare groups that are allowed to have newspapers, unblocked websites, or other media that governments can cut off, with groups that are not allowed any such media, we would find the former posting relatively mild criticism on their social media accounts and not calling for protest, while the latter groups post more critical content and are more likely to call for protest. We can also compare the same group over time when it was and was not allowed to have media outlets. Automated methods would enable content analysis of large volumes of posts across many nations and years.
To further test the robustness of my findings from the cross-national chapter, it would also be useful to produce alternative versions of my dataset based on different coding rules. For instance, do the results change if we include all opposition groups that meet the electoral or protest criterion for significance, even in regime-years when there is also an identifiable
“main” opposition group? Does it matter if we use lower thresholds to code groups as significant (e.g. winning 5% of votes rather than 10% or organizing protests with over 100 participants rather than 1,000)?
The cross-national evidence was limited to Arab countries. Do the findings from that region generalize to other regions? Eventually the dataset I collected for Arab countries should be expanded to include authoritarian regimes around the world.
With the theory, the case study, and the cross-national analysis, I deliberately focused on regimes in which there were opposition groups with non-negligible ability to mobilize citizens.
I suspect that regime leaders’ incentives to allow or prohibit opposition media might differ when they face the smallest opposition groups, with almost no mobilization capacity. The potential benefit of giving such a group a public platform to discourage protest by its few followers may be small compared to the risk that allowing its media would make more citizens aware of the group’s existence, helping it to grow. Thus, when we consider the entire universe of authoritarian regimes and opposition groups, the relationship between an opposition group’s mobilization capability (relative to the regime’s repression capability) and the likelihood that the regime will let them have a media outlet may be an inverted U curve. This argument might explain the apparent absence of opposition media in certain stable states with only minuscule opposition groups, such as the United Arab Emirates.
Social scientists must remain humble about the longevity of the patterns they uncover.
Technological changes — particularly in how people communicate and process information
— may have profound and rapid effects on the nature of opposition media in authoritarian regimes. As these changes unfold, they will raise new questions for political scientists on that topic and possibly change the answers to current questions. In the long-run, who will be most empowered by technological developments: opposition leaders, authoritarian regimes, or ordinary citizens? Will opposition groups gain increasing ability to communicate with citizens in ways that regimes cannot stop? Will citizens gain greater access to alternative sources of information — beyond those produced by governments or opposition groups — as well as becoming more able to communicate directly with each other? Or will regimes become increasingly adept at filtering their citizens’ access to information from all sources?
How will these changes in the distribution of technical capacities affect governments’ policies regarding opposition media, opposition groups’ strategies for promoting political change, and citizens’ decisions to oppose, acquiesce to, or actively support their rulers?
5.4 Conclusion
From the perspective of those who value freedom of expression and democracy, my conclusions are in some ways pessimistic. They suggest that the existence of opposition media under an authoritarian regime does not imply a lively public sphere in which opposition groups fiercely challenge the regime. Nor is the toleration of opposition media necessarily a harbinger of further political liberalization: in fact, it can be a means for preventing political change. But what I have found also has at least one hopeful implication. Censorship, specifically governments’ efforts to stop opposition media outlets from functioning or to block the public’s access to them, is not completely effective at hiding a regime’s weakness. The very absence of opposition media can provide a strong clue to citizens that their rulers fear their collective power.