In the May/June 2011 issue of Foreign Affairs, my colleague and sometimes co-author Jack Goldstone assesses and draws some preliminary lessons from the revolutions roiling the Middle East and North Africa in 2011. Jack is widely recognized as a leading scholar of both revolutions and social movements, so his analysis of these events is surely attracting attention from both scholars and policy-makers, and deservedly so. I disagreed with important parts of that analysis, and I said so to Jack via email. He was kind enough to respond in detail to my critique, and to agree to have his reply posted here. Our exchange follows.
* * * * *
In your Foreign Affairs article, you claim (p.11) that “The revolutions unfolding across the Middle East represent the breakdown of increasingly corrupt sultanistic regimes.” As you see it (p.9), a sultanistic regime is one in which “a national leader expands his personal power at the expense of formal institutions. Sultanistic dictators appeal to no ideology and have no purpose other than maintaining their personal authority.” Your description and analysis of sultanistic regimes appears to draw on your own prior work on the topic, as well as work by Juan Linz with co-authors Alfred Stepan and H.E. Chehabi. All of these authors trace their ideas about sultanism back to Max Weber, who identifies these regimes as instutionalized patrimonialism in the extreme. According to Linz and Stepan (p.52), in sultanistic regimes, “The private and the public are fused, there is a strong tendency toward familial power and dynastic succession, there is no distinction between a state career and personal service to the ruler, there is a lack of rationalized impersonal ideology, economic success depends on a personal relationship to the ruler, and, most of all, the ruler acts only according to his own personal unchecked discretion, with no larger, impersonal goals.”
You argue that sultanistic regimes are more vulnerable to the occurrence and success of revolutionary uprisings than other forms of authoritarian rule and uses this claim to explain variation in the occurrence and outcome of uprisings in the Middle East in 2011. You write (p.8) that, “Although [sultanistic] regimes often appear unshakable, they are actually highly vulnerable, because the very strategies they use to stay in power make them brittle, not resilient. It is no coincidence that although popular protests have shaken much of the Middle East, the only revolutions to succeed so far–those in Tunisia and Egypt–have been against modern sultans.”
I gather that this claim is based on your own prior work on the subject and is consistent with claims made by Goodwin and Skocpol and Linz and Chehabi. The trail of citations gets a bit circular here, but Linz & Chehabi (p.41) cite Goodwin and Skocpol, who cite you and Eisenstat to argue that sultanistic regimes are more susceptible to revolution because their leaders “are more likely to generate elite and middle-class opposition from landlords, businessmen, clerics, and professionals, who resent their monopolization of key sectors of the economy, their heavy-handed control of the flow of ideas and information in schools and the press, their subservience to foreign powers, and the general climate of corruption.” In other words, the very corruption and repression that keep these regimes afloat when revolutionary movements are absent are thought to make critical constituencies more likely to swing against the sultan once those movements emerge.
I tripped over two parts of this analysis. First, my own empirical research on the effects of contentious collective action on the breakdown of authoritarian regimes directly contradicts the claim that sultanistic regimes are more likely than other forms of autocracy to collapse in the face of popular uprisings. In a paper published in 2005 in International Political Science Review, I report the results of a statistical analysis showing that personalist (a.k.a. sultanistic) regimes are least affected by anti-government demonstrations, strikes, and riots. Instead, it is single-party and military regimes that appear to become more vulnerable in the face of these kinds of contentious collective action. I reason (p.316) that sultanistic regimes are less likely to break down in response to contentious collective action because…
the absence of any social contract linking the despot’s right to rule to the will or well-being of the people reduces the impact of mass challenges…Absent any implicit link between the public interest and regime legitimacy, popular uprisings are more likely to play out as raw power struggles. Not surprisingly, regimes built explicitly on the use of coercion or terror usually perform well in such struggles, and bystanders, who generally recognize the regime’s extreme commitment to power maintenance, are less likely to [join the revolutionaries].
My analysis suffers its own flaws and is hardly definitive, but it’s noteworthy that it directly contradicts the theoretical claim at the heart of the prevailing narrative about sultanism. That narrative seems to draw heavily from a handful of high-profile cases where revolutions occurred and succeeded, including Cuba in 1959, Iran and Nicaragua in 1979, and the Philippines and Haiti in 1986. What may have been missing from those analyses are the large number of other cases in which personalist regimes successfully squashed popular uprisings, a nefarious club that includes Zaire for many years under Mobutu Sese Seko and contemporary Zimbabwe under Robert Mugabe.
My second stumbling point came with your application of the concept of sultanism to virtually every regime in the Middle East that is not a dynastic monarchy. In your Foreign Affairs piece, you state that the roster of sultans in the contemporary Middle East includes “Bashar al-Assad in Syria, Omar al-Bashir in Sudan, Zine el-Abidine Ben Ali in Tunisia, Hosni Mubarak in Egypt, Muammar al-Qaddafi in Libya, and Ali Abdullah Saleh in Yemen.” The regimes led by those rulers are contrasted to the monarchies in Jordan, Morocco, Oman, and the Persian Gulf region. You predict that these monarchies are more likely to survive the Arab Spring because dynastic succession and the presence of elected parliaments give them more room to maneuver in response to popular pressure. (The odd man out is this binary classification scheme is Iran, which you also see as more resilient than the purportedly sultanistic cases because of the popularity of its ideology, the division of power among several top leaders, and the strong support it enjoys from key elements of its security forces, the Basij and the Revolutionary Guards.)
Are all of those non-monarchies really “sultanistic,” though? Weber’s original ideal type and the recent derivations of it all describe an extreme concentration of power in a single individual and a concomitant lack of power in other individuals and organizations. A key feature of these regimes is that they “allow no stable group prerogatives in the polity–not even collective prerogatives for military officers or upper social or economic classes” (Goodwin and Skocpol as cited in O’Kane, p. 187).
Whether or not all of these regimes are “sultanistic” matters because it greatly affects the preliminary lessons we might draw from the “Arab Spring” about the conditions under which popular uprisings occur and, when they occur, succeed at toppling authoritarian rulers. You conclude that sultanism helps explain the outcome of the Arab revolutions because the only two cases where revolutions have succeeded so far–Egypt and Tunisia–were, by your reckoning, both sultanistic. Meanwhile, all of the monarchies that have faced uprisings have withstood those challenges. The distribution of uprisings and outcomes under this set of categorizations clearly supports your claims about the relative vulnerability of sultanistic regimes. By this reckoning, at least five of six sultanistic regimes have suffered popular uprisings (Egypt, Libya, Syria, Tunisia, Yemen, and maybe Sudan), compared with just one of seven monarchies (Bahrain), and the only two of the six uprisings that have toppled longtime rulers occurred in sultanistic regimes.
That pattern does not hold up, however, when we use other scholars’ categorizations of those same regimes. A cross-national data set compiled by Barbara Geddes, a leading scholar of comparative authoritarianism, identifies episodes of authoritarian rule worldwide since the middle of the 20th century and classifies them into one of five types: personalist (a.k.a. sultanistic), single-party, military, hybrids of those three forms, and monarchies. According to her classifications, the only purely personalist regimes in the region were Libya, Sudan, and Yemen. She categorizes Tunisia as a single-party regime and Egypt and Syria as single-party/personalist/military hybrids (a.k.a. “triple threats”). Using those categorizations, uprisings appear to be comparably likely in personalist, single-party, and triple-threat regimes, but the only two successes so far have come in single-party or hybrid cases, rather than the personalist ones. These results (so far) contradict your claim about the relative fragility of these regime types in the face of popular unrest, and they are consistent with the results of my own statistical analysis mentioned earlier.
Take Egypt, for example. In an award-winning book based on extensive field research, Jason Brownlee attributes the survival of the Egyptian regime to the strength of its ruling party, the National Democratic Party (NDP), rather than Mubarak’s personal authority. Even if we accept the idea that Egypt under Mubarak had elements of personalism (as Geddes’ hybrid classification suggests), Brownlee’s account puts it much closer to the single-party ideal type than the sultanistic one. Doubts about Mubarak’s personal power were also evident in pre-revolutionary accounts of the struggle over who would succeed him as president. In a sultantistic regime, there would have been no doubt that the president would choose his own successor, but Mubarak’s attempts to position his son as heir to the presidency were . Meanwhile, many of the post-revolutionary assessments I have read on Egypt emphasize the tremendous political and economic power wielded by that country’s military throughout Mubarak’s tenure. These accounts may be favoring the military because of the more prominent role it has played in post-revolutionary politics, but they also undercut the idea of Egypt under Mubarak as a sultanistic regime in which there are “no stable group prerogatives.”
The same points could be made about Tunisia, which Geddes identifies as a single-party regime without the hybrid aspects. From press accounts of cables obtained through Wikileaks, it sounds like the one-party system may have evolved towards sultanism in recent years as President Ben Ali and his family got greedier. Still, it’s hard to see the president as an all-powerful sultan when Tunisia’s military didn’t even bother to come to his defense once protests gained some momentum.
In sum, I don’t think the concept of sultanism is as important to explaining why uprisings are occurring in the Middle East in 2011 as you suggest, or in explaining why those uprisings are succeeding in some countries while failing in others.
A positive alternative analysis would require more time and thought than I can give to this topic right now, but I can say that I find Marc Beissinger’s ideas about the roles of observation, imitation, and learning in both the contagion and repression of popular uprisings across societies that share cultural and historical similarities to be a more useful lens through which to view these events. The structural conditions in many Middle Eastern autocracies have not changed much in recent years and are not dissimilar to those found in many currently surviving and previously failed authoritarian regimes in other parts of the world. What stands out to me about the uprisings of 2011 is the speed and breadth of their spread, and the speed with which surviving rulers have sought to learn from the failures of compatriots ousted by “early risers.”
Viewed from this perspective, the heart of explanation for the spread and outcomes of the “Arab spring” lies in its own dynamics, rather than structural preconditions. In his study of nationalist uprisings in the USSR, Beissinger writes (p.17) that, “Some level of influence from pre-existing structural conditions and institutional constraints almost always remains embedded in an event. But a reproducing chain of events can grow to the point that the initial structural influences that played a prominent role in unleashing the series seem buried in the distant past and relatively impotent within the ongoing production of events.” To my mind, this “eventful” perspective, as Beissinger calls it, stands to tell us more about the spread and outcomes of the Middle Eastern uprisings than approaches centered on comparisons of prior structure.
* * * * *
Our differences are real but reconcilable–because I think I can point to issues that did make the Egyptian and Tunisian regimes markedly more sultanistic in the last few years. And Algeria is another stable, non-sultanistic, non-monarchical Middle East state.
I am not contradicting your statistical analysis because I did not say that sultanistic regimes are always more vulnerable to revolt than other kinds of regimes. It is just that when revolutions against dictators have occurred in recent decades, they have generally been in those kinds of regimes. Personalistic or sultanistic regimes can be stable for decades if the ruler has the resources and can play the game of dividing his enemies and maintaining patronage. It is just that when patronage fails, sultanistic regimes are hard to reform and provoke widespread opposition, so they are more likely to end in revolution than other types. (I think even in the Geddes data personalistic regimes are more likely to end in big events than other types, even if they are not as vulnerable to coups or withdrawals.)
In Egypt and Tunisia–indeed, across the Middle East–rulers had deprived people of basic freedoms but in exchange had offered a wide range of subsidies and guaranteed jobs for college graduates. Both the subsidies and the job guarantees had been withdrawn in the last decade as unaffordable, just at the time when the youth cohorts and college enrollments soared. So the patronage game at the popular level was being played out, and unlike in the 1990s, people were feeling that they got few real benefits from the regimes.
At the same time, the concentration of crony capitalism in family and family-connected elites, who were the prime beneficiaries of foreign direct investment in modernizing services and finance, increasingly excluding ordinary businessmen and military officers, changed the patronage game among elites. This meant that what had been more hybrid personal/party regimes in both countries shifted to more personalist regimes. Indeed, the face of Egypt’s future was not the army or an army officer but Gamal Mubarak the banker; and in Libya it was Ben Ali’s wife and her family, known for demanding shares of any major new enterprise. This is what made these regimes much more personalist–and hated–as their wider patronage waned while the wealth of those in the inner circle rose to dizzying heights. I’m afraid the Geddes scoring does not capture this — but it is clear from comments of key actors that Ben Alis’ wife and Hosni Mubarak’s son were the main targets of elites–something that is inexplicable unless these had become highly personalist regimes.
My prediction that Syria would suffer similar fate–made in February, when most people felt that Syria was very different and would certainly remain stable–was based on seeing the very same process in that country. That is, under Bashar al-Assad, unlike his father, the regime went from a hybrid personalist/party regime, where the Ba’ath party was still meaningful, to a very personalist regime, where Bashar’s family and closest associates monopolized patronage and government contracts. Corruption was getting more exclusive and yet more brazen, just like in Tunisia and Egypt. And of course, the problems of declining subsidies, booming youth and college cohorts, and rising food prices obtained there as well.
So, to be clear–the theory of revolutions on which I am drawing does not claim that personalist regimes are more likely to end quickly. The theory recognizes that, if well-managed, they can last for decades. Rather, what the theory says is that if the patronage cannot be maintained, and the elites start to unify against the ruler, support for the government will collapse rather suddenly and completely. In explainng the Revolutions of 2011 in the Middle East, the point is not to argue they were especially likely to collapse because that regime type is inherently weak. Rather, the point is that under the stresses encountered all across the region in 2011–youth bulge, large college cohorts with high unemployment, rising food prices–then personalist regimes which had already been changing and limiting the terms of their patronage to the public and to elites were very likely to undergo a sudden collapse, with the personalist leader facing widespread defections. By contrast, other kinds of regimes, such as military/party regimes (Algeria and now Iran, given the role of the Revolutionary Guard) or monarchies, were more likely to retain enough elite loyalty to stay in power, while monarchies had more structural options for reform.