Saturday 22 August 2015

The West, Black Knights and the Political Economy of Regime Persistence

“You know… we've got a much easier job now than we should have had fifty years ago. If we'd had to modernise a country then it would have meant constitutional monarchy, bicameral legislature, proportional representation, women's suffrage, independent judicature, freedom of the press, referendums…"
"What is all that?" asked the Emperor.
"Just a few ideas that have ceased to be modern.”
- Evelyn Waugh, Black Mischief
For those who like to spend their time criticising the West, our relationship with Saudi Arabia and other authoritarian countries provides them with endless ammunition. It’s not an unreasonable question: how can we claim to be for human rights, whilst at the same time providing countries with terrible human rights records financial, military and rhetorical support? Whilst not unreasonable, I think it’s overstated. This post is a purely consequentialist account of our relationship. If you believe in deontological or moralistic arguments for not allying with authoritarians, you may be right but I am not seeking to dispute that argument here. Note the following questions are different:  
  1. Does the West’s relationship with authoritarian countries create better outcomes than would exist without the West’s relationship?
  2. In the recent past, has the West’s relationship been a force for good?
The first question is directly toward policy prescriptions without necessarily implying good outcomes (just better ones than the null condition). The second question would be harder to prove and is not my focus. My view is that the reason for human rights abuses and authoritarianism is predominantly an institutional issue that has domestic origins. The role that this leaves for Western machinations, accordingly, declines significantly. More interestingly, most of the literature (by no means all) suggests that it may have a neutral or positive effect which should veer us toward maintaining our links. The alternative to not being allies with Gulf states is far worse.
This is an extremely long post, if you aren’t interested in the literature on institutional quirks in the Middle East, oil rents, foreign aid and democratisation then this is not a post for you. As usual, bibliography is below the endnotes. In this post, I have tried my best to draw attention to studies that disagree with my view.  I am aware that this may make my argument appear weak. I also know that arguing that we should have strong relationships with authoritarian countries because we care about human rights is controversial and a difficult argument to make. And that’s a good thing, I’m not really sure how far my argument goes. In the last post I referred to Cowen’s Second Law, but Cowen’s First Law is just as important:
There is something wrong with everything (by which [Cowen] mean[s] there are few decisive or knockdown articles or arguments, and furthermore until you have found the major flaws in an argument, you do not understand it).
Just as a side note before this post starts: I will be addressing some criticisms of my last post in the near future. I consider most of them to be weak hence the weak incentive of replying but, that being said, I have been surprised and pleased with the quality of response. I was expecting to rile up some of my favourite bloggers and they duly obliged so please do read their responses (Jamie’s here, David’s here, and some critical comments from Anthony Cooper here). Who says people on the internet can’t be nice when disagreeing?
1. The Limited Role of the West
1.1 The Institutional Origins of Authoritarianism
To explain authoritarianism, human rights abuses and the lack of democracy in the Middle East, one must start with an institutional framework that explains weak civil society. As I have detailed before, the people across the Middle East want an illiberal form of democracy – why have they not achieved their aims? Bear in mind that the success of civil society in non-democratic countries is generally high, from The Empirics of Free Speech, Part 1:
Chenoweth and Stephan (2008) look at ‘data on 323 violent and nonviolent resistance campaigns from 1900 to 2006.’ They find that ‘in the face of regime crackdowns, nonviolent campaigns are more than six times likelier to achieve full success than violent campaigns that also faced regime repression.’ Their data show that no non-violent campaign.. that garnered the participation of 3.5% of the population failed to reach its goal.  
Accordingly, an explanation that accounts for the failure to achieve democratic status must necessarily relate to the weakness of civil society. It is strong civil society that reaches its goals. And it is here that we find not only the origins of lack of democracy but its persistence. Kuran (2013) builds on his work on Islamic economic institutions and shows how Islamic institutions (not people or culture) have inhibited the rise of a civil society. The most significant of these is the waqf – this is an institution which provides public services (like schools and water). This was a form of trust which was funded by a private individual’s assets. It allowed the individual to protect their assets without giving it to the government (see Table 1 in Kuran (2015)).
As Kuran states, ‘the waqf served as the delivery vehicle for functions met in the West generally through corporations’. The reason the waqf matters is because a waqf could not participate in politics, could not align with other waqfs and was not accountable to its users. As Kuran (2015) notes in a more recent article: ‘the Islamic waqf promoted neither broad political participation nor transparency in governance’:
[Complainants] had to prove that the deed [setting the waqf up] was being violated. Because information concerning the waqf’s finances and actions were not public knowledge, beneficiary-launched lawsuits against caretakers were rare. Out of 1544 waqf-related lawsuits in a seventeenth-century Istanbul sample, only six entailed an accusation of caretaker mismanagement.
Every waqf defined a group of beneficiaries who formed a potential political entity. Whether they developed a group identity was not assured. The users of an inn on a trade route could remain mutual strangers for years. The rules of the waqf did nothing to enhance group consciousness among co-beneficiaries without reason to interact anyway.
One indication of the powerlessness of waqf beneficiaries lies in the tenure of caretakers. In the Anatolian town of Sivas, 1902 waqf caretakers were replaced between 1700 and 1850. No fewer than 74 percent of the replacements followed a death in office. In the remaining cases, the successor was typically the retiring caretaker’s son...
Leaving waqf beneficiaries ignorant about management promoted stability by keeping waqfs from becoming foci of discontent. The passivity expected of beneficiaries limited mass political activity. It also reduced the capacity of the masses to react effectively to economic downturns, military defeats, and other such misfortunes. In the absence of pre-existing consumer organizations and lacking experience in organized collective action, disgruntled subjects had to start organizing from scratch.
Waqfs did not form political coalitions: where Europe had guilds across cities campaigning against taxes, the waqfs in the Middle East were impotent. Even at the very establishment of the waqf, there had to be a single property owner – and this meant that individuals could not pool their resources together. Compare this with corporations, trade unions, universities and guilds which can take part in politics and were accountable to its customers and its stakeholders. And the absence of these bodies matters:
Democratic rights got established because of epic struggles driven by groups organized, usually as corporations, within universities, as cities, as religious orders, as unions, or as merchant associations. Such groups demanded rights. They articulated requests. They developed blueprints for alternative orders. They stimulated intellectual life (Kuran 2013)
The horizontal axis represents the organization’s discretion regarding asset management and service delivery. The vertical axis represents the share of its beneficiaries and officials who participate in decisions (Kuran 2015).
Hence this absence of the corporation as an institution left ‘the Islamic world without politically influential social structures situated between the individual and the state’. There are more flexible waqf structures now (see Fig. 3 in Kuran (2015)) but as Kuran notes, the legacy endures. One mechanism is kinship ties. Because organisations could not form, people relied on their blood ties. This was part of the reason that commercial transaction were and still are (relatively) personal rather than the impersonal model generally followed in developed states. In any event, I think there’s a lot to be said for Kuran’s bottom line conclusion:
...the proximate factors that have made authoritarianism the Middle Eastern political norm rest on longstanding historical patterns. In the modern era oppressive coalitions have been able to establish entrenched autocracies because the region’s masses entered it with stunted political capabilities. These capabilities depend on the organizational skills, civic concerns, and expressive capabilities that individuals acquire as part of their socialization. They depend also on precedents regarding civic engagement. In both these respects, the Middle East has faced handicaps that have constrained, and still constrain, its political development.
Tyler Cowen has said that Timur Kuran is one of the most important thinkers of our age. And I agree with him, Kuran is really underappreciated. The issue with Kuran, however, is his reliance on qualitative elements for his biggest quantitative claims. I find his arguments about waqfs very persuasive but it’s limited because of the dearth of data. That said, there is some research from Blaydes and Chaney (2013) that we can draw on in showing that Kuran is quantitatively correct.  
Blaydes and Chaney are trying to explain the divergence between the length of time that a ruler was in power (see above). Their model is, crudely, something like this: in Europe, especially following the fall of the Roman Empire, the feudal system that existed allowed for constraints to be put on the ruling monarch by barons. It’s not exactly the same as corporations and guilds etc., but it is similar to the idea that non-state groups are vital to the control of the executive.
Charlamagne was unable to extract rents from his population to fund an army and so he set up a feudal system in which he gained troops from barons. This has to major effects: first, some landowners pooled their land (to avoid the need for themselves to fight) and second, the need for mounted troops meant wealthy individuals were required to ensure horses could be funded. These wealth individuals were compensated with land grants and other privileges (i.e., the start of constraints on power).
You can clearly see the mechanisms by which this led to executive constraints: first, because the monarch would have to cede privileges and secondly, because in allowing for barons to set up their own armies, men and weapons, he was creating a formidable enemy against himself (should he ever go against the barons). In the Middle East, however, there wasn’t such a decentralisation of power which meant that there wasn’t a need to negotiate (and hence short leadership duration as a result of power plays). This was because of the use of slave armies (mamluks) in the Islamic world which creates a significant difference:
…while European rulers were negotiating with the local gentry to raise armies for matters of defence, Islamic ruler bypassed local elites by creating highly skilled armies of foreigners who had no ties to the existing gentry and swore allegiance directly to the sultan.
The theoretical logic behind our historical narrative is straightforward: decentralising power increases the cost of an unsuccessful revolt for the monarch’s rivals. In other words, armed local elites were able to extract a better “soft contract” from their monarch. [This has the implication that] the introduction of the feudal institution led to increase in ruler stability.
Blaydes and Chaney have the empirical support for what they’re saying. They look at the Carolingian dynasty. As the Carolingian dynasty came into power, the feudal institutions did too. Accordingly, there should be a break in the trend of decreased executive constraints (and hence less negotiation and hence more revolts). In the graph below, the vertical line represents the year 790 AD (the mid-point of Charlamagne’s reign).
You can see that leadership duration trends clearly changed after that mid-point. To make it clear why all of this relevant: Kuran is suggesting that the institutional set up did not allow for a vibrant civil society, a strong opposition force to absolutist power. This has importance even today and Blaydes and Chaney’s paper support that broad argument through the use of different institutional factors: in the Middle East there was less negotiation, less executive constraints and hence short ruler duration. Given its importance, I will return to the issue of civil society below but the issue with Kuran’s view is that doesn’t necessarily take into the full set of variation in the Islamic world in terms of authoritarianism or human rights abuses. Turkey is not the same as Saudi Arabia which is not the same as Indonesia.
The institution of slave armies, already eluded to, does however also provide another forceful mechanism which, along with waqfs show why the origins of authoritarianism are so deep and entrenched that it leaves less need to require another force (Western support) to be used as an explanation. Chaney (2012) is a fascinating piece of research which finds that the way an area was incorporated into the Islamic empire has an impact on democracy today.
For those countries which were incorporated in the Ottoman period (where slave armies were not used) or by natural conversion, there was a need to acquiesce to local elites and not crush civil society. This opposite was the case for Arab conquest and the use of slave armies: smash civil society and dominate. Chaney finds that ‘the percentage of a country’s landmass that was conquered by Arab armies following the death of the prophet Muhammad statistically accounts for this deficit.’
Hence, Chaney’s conclusion that ‘these historic control structures have left a legacy of weak civil societies where political power is concentrated today in the hands of military and religious leaders that work to perpetuate the status quo.’ The final study I would refer to is the Sarkissian (2012) who finds that ‘Once GRI [government regulation of religion] is added in model 2 [testing the relationship between Muslim-majority countries and democracy], the coefficient is reduced to −2.39 and loses its significance.’ I wrote about the implications of this in The Empirics of Free Speech, Part 1:
You can explain the democracy gap between the Middle East and the West by accounting for the extent to which a state has free speech curtailed. The point that Sarkissian is making is that laws which curtail free speech are used against political dissenters. Its why, for example, Muslims are more likely to be targeted rather than non-Muslims in laws against proselytizing. For reasons that I have gone into elsewhere, Islamic institutions are often the only vestige of civil society left in the Middle East. This leads to a fear of ‘competition and threats to [autocrat’s] power, [and then] repressive governments restrict the religious arena in order to forestall the rise of political opposition.’  
1.2 Dame Mas Gasolina
If these authoritarian states can fund their oppressive security apparatus without us, then it further limits the role of the West. There is some literature that would make this line of argument persuasive. For example Anders and Alasken (2013) find that ‘a 1 percentage point increase in the value of oil production in GDP is associated with an average increase in the duration of the current political leadership of 1.1 percent.’ Ulfelder (2007) also finds the same relationship:
… autocracies that derive a larger share of their gross national income from mineral depletion are substantially less likely to transition to democracy in any given year [and] during the past three decades at least, resource wealth has played a significant role in helping to sustain authoritarian rule in numerous countries that have never attempted democracy and in delaying a return to democracy in some countries that have
…. an autocracy that obtains just 15% of its gross national income from energy and mineral depletion is only about half as likely to transition to democracy, and a regime at my sample maximum on this measure—91.8% is only about one tenth as likely to transition as the median autocracy.
The literature doesn’t always necessarily find the same results. Haber and Menaldo (2011) who find that there is no link between oil wealth and reduced democracy. Their dataset is the longest I’ve seen and they provide the following graphs (and many more for other countries) which are supposed to show there’s no relationship. What they look at (at least in the first half of the study) is oil revenues over time against polity scores.
The issue with this study, however, is that dataset is too long. This is usually a good thing, but the dataset starts in 1800 when no country produce economically significant amount of oil until the end of the Great War. Moreover, the rents were not always used by the states themselves. As Anderson and Ross (2012) note:
Until the late 1960s, most of the rents generated by oil production in non-Western countries were captured by a handful of large, vertically-integrated international oil companies – sometimes called “the Seven Sisters.” But in the 1970s, the industry was transformed by a wave of nationalizations and contract revisions that enabled the governments of host countries to seize control of these rents.
Another issue they note with Haber and Menaldo is that for their 53 ‘resource reliant’ countries, they apply an ECM cointegration test, which involves a within country analysis over time. I literally have no idea what they were thinking:
Haber and Menaldo compare countries ‘treated’ with natural resource wealth to themselves over time, instead of comparing them to countries without resource wealth. Since they find that when countries are ‘treated’ with resource wealth, they do not become less democratic, they mistakenly infer that the treatment had no effect.
And when they look at the results from 1970 onwards Haber and Menaldo’s with this corrected, they find:
Notice that there are two differences between the two lines: the level of democracy is much lower in the oil states; and since around 1980, the gap between the two groups has widened…. Haber and Menaldo’s findings are significantly altered once we account for the historic changes of the 1970s.
Beginning in 1976, the interaction term has a negative coefficient [i.e., oil revenue is negatively associated with democracy]; beginning in 1980, this coefficient becomes statistically significant; and beginning in 1982, the net effect of oil becomes statistically significant and negative, and remains so for all subsequent break points in the table
Wright et al (2015) find that ‘higher oil wealth increases autocratic regime survival’ but this not through deterring democratic forces but other autocratic groups. This negates the argument that the West has a limited role because these states often have their own oil revenues they can utilise. I would, however, simply note that within countries differences don’t seem to me to indicate much. Levels will likely never go below an amount that fund an oppressive state and keep members loyal. Accordingly, I think the comparison levels between non-oil producing states and oil producing states (between country) is more relevant and hence Wright et al and Haber and Menaldo should not be given as much weight (Smith (2004)).
Moreover, we have some fairly robust studies on specific mechanisms. Egorov et al (2009) find that oil-rich non-democracies have less free media than non-oil rich democracies (and oil rich democracies). The mechanism is important:
In the absence of free media, incentives for lower-tier bureaucrats to provide sufficient effort and transmit necessary information to higher levels prove[s] inadequate… To induce high effort, the dictator needs some verifiable information on the bureaucrats ‘performance. The dictator can rely on special monitoring agencies, but these are vulnerable to collusion with the bureaucrats they monitor; preventing such collusion is costly. In the case of uncensored media, this collusion is ruled out by the free-rider problem and competition between decentralized media…
In a resource-rich country, the dictator is more interested in remaining in office (as rents are higher); moreover, he cares less for bureaucratic incentives because the resource rents can compensate for poor economic policies…
I quite like this study because it doesn’t just look at oil prices, but oil reserves. It also tries to disentangle the effects of oil, democracy and media effects. The find that a oil does have a negative relationship with democracy, but they also provide evidence for their mechanism showing a link between oil reserves and media freedom in the next year:
Whether we use oil reserves (regressions (1)–(4)) or oil production (regression (5)) or whether we use the full sample (regressions (1) and (5)) or split it into autocracies (regression (2), Polity2 is not positive in 1992), imperfect democracies (regression (3), Polity2 positive but not higher than 8), or democracies (regression (4), Polity2 above 8 in 1992), oil abundance has a negative and significant effect on democracy..
…in columns (6) and (7) of Table 3, we disentangle the effect of oil on democracy from the effect of oil on media freedom through the following two-stage procedure. In the first stage, we run media freedom on democracy, country fixed effects, and other control variables except for oil. Then we take the residuals, and estimate a regression for the residuals on oil, country fixed effects, and its other determinants and controls except for democracy. the second-stage regressions for the media freedom residuals capture the relationship between oil and media freedom. The results are also consistent with our theoretical predictions: both oil reserves (regression (6)) and oil production (regression (7)) affect media freedom residuals negatively and significantly.
It’s also noteworthy that in a meta-analysis of 120 studies, Ahmadov (2015) finds that ‘all means indicate that there is a negative effect of oil on political regime’ and ‘confidence and credibility intervals confirm this nontrivial, negative partial correlation and also rule out the possibility of a positive and no association.’ Interestingly Ahmadov then divides up all the studies into regions being assessed and finds:
The coefficient of MENA is negative, indicating that when estimates are derived using data that cover this region, the results tend to be more negative… The coefficient of Latin America, on the other hand, is positive.
Why would there be a regional difference? Thad Dunning (2008) in his book, Crude Democracy: Natural Resource Wealth and Political Regimes argues that natural resources lead to democracy in some states (particularly in South American countries) but not for the rest of the world. Crude Democracy is a pretty fun application of typical game theory. Here’s how he starts the model:
In the first move of this game, a relatively poor democratic majority sets economic policy. In the second move of the game, a set of relatively rich elites decides whether to stage a coup. I assume that there is some exogenous cost (φH) to staging a coup that is independent of economic interests (such as the risk that a coup fails, as mentioned above). In deciding whether to stage a coup, elites weigh the payoff to authoritarianism and the utility (or disutility) of democracy against this coup cost
Note, that as Dunning goes on to explain: it doesn’t matter if there isn’t actually an existing democracy for the model to apply. The point is that oil resources have conflicting effects on the decision of these elites. Argument for elites acting in favour of a coup would be (i) increased rents from taking over the oil wealth themselves an (ii) stopping the poor majority from taxing the elites. Against the coup there are converse factors (e.g. the rents mean that the state can pay off the poor majority without an effect on the elites). Which pressures ultimately win, Dunning says, depend on the wider context:
Imagine first country A, where the resource sector is a predominant force in the total economy: the ratio of resource rents to gross domestic product (GDP) is extremely high, so resource rents are very nearly the only economic “game in town.” Moreover, to the extent there is any economic activity in the private sector, income and assets (such as access to capital, land, and so on) in this sector are distributed relatively equally. This country is “resource-dependent” as well as resource-abundant... Moreover, inequality of private wealth or income provides little in the way of a potentially salient dimension of political conflict. consider country B. Here, the resource sector may be an important part of the total economy, but it is not preponderant: the ratio of resource rents to GDP is substantially lower than in country A, as there is important private economic activity outside of the resource sector. Furthermore, suppose that in this country there is also substantial inequality of assets or income in the private sector
Dunning’s model would suggest that there would be a stronger democratic impulse in country B rather than country A. Looking at his Figure 1.2 (below), you can see that resource dependency leads to broadly net authoritarian effects and he, accordingly, puts Gulf states in this category. The final thing worth exploring is his distinction between “equal / unequal private sector”.
It not until Chapter 3 that he tells you what he’s on about: he believes inequality increases the chance of social conflict because ‘inequality gives elites more to lose from redistributive taxation and therefore heightens the future cost of living under democracy.’ Hence, where there is inequality, even in non-resource dependent states, there may be “intermediate cases”. And this is where Dunning loses me: the idea that inequality and “poor masses” desire for distribution is a cause of democratic breakdown. Those familiar with Acemoglu and Robinson’s work can’t help but see this as a throwback to their Economic Origins of Democracy and Dictatorship:
[Acemoglu and Robinson’s] framework predicts that in highly unequal societies, democratic policies should be highly redistributive but then abruptly come to an end with a coup that reverts back to much less redistributive policies… In democracy, the elites are unhappy because of the high degree of redistribution and, in consequence, may undertake coups against the democratic regime (p.38, 222)
But the idea that (economic) inequality leads to democratic breakdown, to my mind, seems incredibly weak as per Slater et al (2014):
Cross-sectional time series data from 139 countries between 1972 and 2007 show that the taxation of income, profits, and capital gains—which effectively proxies for both wealth redistribution and state capacity—is if anything negatively correlated with military coups against democratic regimes. Our quantitative analysis also shows that the democratic breakdowns that follow such coups do not seem to result in any systematic reduction of redistribution.
…democratic breakdown in the post-World War II era is best understood as the product of postcolonial state weakness. On the militarized side of the state apparatus, officers typically overthrow democracy for reasons of their own, not in support of particular economic classes. On the postcolonial state’s civilian side, administrative incapacity means that recurrent crises of governability will repeatedly tempt and enable military intervention to restore political stability

The coefficient on direct taxes implies that increasing direct tax collection by one percentage point would be expected to decrease the rate of successful coups by a factor of 0.85 (p=0.01). To clarify the result further, if the same specification is used in an OLS model with robust standard errors clustered by country, a one-percentage-point increase in direct taxes is associated with a 4.4 percent decrease in the probability of a regime-changing coup (p=0.08)… successful coups (that do not necessarily change the regime) and attempted coups are negatively correlated with direct taxes. Increasing tax collection by one point decreases the rate of successful coups by a factor of 0.88 and attempted coups by a factor of 0.90.
Dunning’s model, then, is flawed. But I think his idea that it has different effects over different regions is right. We have already discussed the role of civil society – which could be relevant to this question. Interestingly Ahmadov himself says the institutional framework matters. Significantly, for our purposes, institutions set up the British Empire are not related:
… institutional contexts alter the oil–democracy link, thus circumscribing the negative effect of oil on democracy to developing countries. These results speak to an emerging debate that puts institutions at the forefront of explaining why resource curse takes place in some countries but not others… While our results are largely suggestive because of data and methodological limitations, they nonetheless indicate that a conditioning effect of institutions may lie more in a broader set of political and economic institutions associated with OECD membership, less so in previous political regime, but not institutions associated with British colonial past
Pseduo doesn’t like the generality of the above: he thinks the ‘institutional origins of authoritarianism’ and oil rents is broad. I agree that it is broad, but I’m not putting forward a general theory: just one that shows prevalence of authoritarianism across the Middle East. Maybe I’m being too broad about the Middle East? Maybe. But I have tried to distinguish between specific types (non-slave vs. slave). Nonetheless, when Pseduo says something he’s probably right so just to add a few more distinctions, I’ve drafted the few paragraphs below.
We have a series of studies which tries to explain the differences between Middle Eastern countries. Loyalist regimes (commonly in the form of hereditary regimes) compared with republican regimes provide one such important difference. Brownlee et al (2013) try to explain why there was such a ‘modest harvest’ in the Arab Spring. By which they mean, only some countries appear to have had regime change in this period. They conclude:
…the success of a domestic campaign to oust the ruler was structurally preconditioned by two variables: oil wealth (which endows the ruler with enough material resources to forestall or contain challenges) and the precedent of hereditary succession (which indicates the heightened loyalty of coercive agents to the executive). We find that only regimes that lacked major oil revenue and had not established hereditary succession succumbed relatively quickly and nonviolently to domestic uprisings.

One of the better studies is Bank, Richter and Sunik (2013). They use fsQCA (fuzzy set Qualitative Comparative Analysis) to analyse all monarchies existent in the Middle East between 1945 and 2012 which is unlike any other study referenced in this post. If you do not know how fsQCA works, I’d recommend reading about it before reading these paragraphs (see here for a short primer). They look at the following 5 factors that could lead to regime persistence:
(A) Strategic external support by the leading Western powers, the US, the UK and France (esupport), relates to the geostrategic dimension of monarchical survival.
(B) Rent revenues (rents) flowing into state coffers due to the export of natural resources such as oil and natural gas are at the core of political economic explanations for monarchical survival.
(C) Family participation (family) in political decision-making.
(D) The monarch’s claim to legitimate rule (leg-claim) – based on historical and/or religious premises – is a central aspect of the political legitimation from above.
(E) Hard repression, which is sometimes, but not regularly, used to quell opposition within and outside of the monarchical regime.
Their first finding is that none of these factors is necessary for survival, nor is anyone of them sufficient. Rather there are seven pathways for survival:
….the first covers the classical linchpin monarchies of Jordan and Morocco. It highlights the importance of the historical-religious claim to legitimate rule despite the absence of high rents and family participation. The last two pathways (6 and 7) explain the survival of all of the Gulf monarchies (with the exception of Oman). A combination of high rent revenues and the participation of family members in political decision-making without using hard repression are of key importance in Bahrain, Kuwait, Qatar, Saudi Arabia, and the UAE. A distinct pathway (7) exclusively for Saudi Arabia illustrates the specificity of this kingdom: Rent revenues and family participation are accompanied with a strong claim to legitimate rule.
My takeaway from the study is that external support just really isn’t important in most cases but it also goes some way to address Pseudo’s criticism that my analysis is too broad and doesn’t tell us about differences between Middle Eastern states: there are different pathways to regime persistence. I would emphasise not straying too far from the institutional analysis for the region as the region as a whole is backward on most measures of human rights and polity. Accordingly, whilst specificity can tell us more about the range, it doesn’t explain why it’s such they all have such low scores – which is why I’m more interested in the overall, broader claims.
Where does all this leave us? As stated, we should veer away from studies like Wright et al and Haber and Menaldo. There does seem to be a resource curse. The mechanism is through reinforcing the oppressive state and loyalty of key individuals. I believe the literature is not uniformly consistent because it’s comparing countries with different institutional frameworks and degrees of loyalty. I don’t buy Dunning’s idea about inequality but can see merit in the idea that resource dependency is correlated with authoritarianism – but for different reasons. I have outlined above that civil society is historically and continues to be weak in the Middle East. This dependence on oil means that economic survival is not premised on a vibrant private sector, and there is therefore less of a need for it. And, accordingly, the legacy of weak civil society continues.
It would be interesting to test this idea by measuring oil wealth/dependency against measures of civil society to see whether this idea holds out. The closest I’ve come across is Gandhi and Przeworski (2006) who argue the need for cooperation (to avoid rebellion) with non-governmental groups is lower when oil rents are high. They find that the presence of oil lead to fewer political parties and organisations (see Table 3, the relevant coefficient is -0.5045).
There is a mesh of politics and economics here: if oil was not there, there would be a need for decentralised groups, corporations etc. Because oil is there, there is a strong state. Historical Islamic institutions have laid the ground work for weak civil society. All of these factors interact: Kuran / Chaney et al + Ufelder + Brownlee et al + Salter et al. When these factors are taken together, the role of the West in propping up Middle Eastern regimes or being at all relevant in their operations starts to fade. Crudely, it should make you feel less guilty about our relationship with authoritarians: the institutional framework along with oil rents means their heinous acts would occur without us.
2. The Relationship
You can account for the extent to which these countries are authoritarian in ways that don’t rely on Western influence. Nonetheless is might be argued that whilst the West does not cause the authoritarianism, it “helps” them stay propped up with the provision of arms, financial aid and the like. I am sceptical of this claim for a number of reasons leaving aside the slipperiness of “helps” but it’s worth exploring the role of our actions, in a bit more detail.
2.1 Benefits
Levitsky and Way (2010) in their book Competitive Authoritarianism: Hybrid Regimes after the Cold War provide a pretty good starting point for discussion of the benefits of being friends with authoritarian regimes. Their book is a rich qualitative study of 35 countries which became, what they call, ‘competitive authoritarian’ states in the early 1990s. These regimes are not fully authoritarian but take the form of a state like Russia where there is a guide of democratic behaviour. These regimes typically have clear political inequality and mass human rights violations.
The aim of their study is to study what determined whether these states became democratic / broadly liberal in 2008. This is their argument in brief:
... [authoritarian] incumbents’ capacity to hold onto power – and the fate of competitive authoritarian regimes more generally – hinges primarily on two factors: (1) linkage to the West, or the density of ties (economic, political, diplomatic, social, and organizational) and cross-border flows (of capital, goods and services, people, and information) between particular countries and the United States and the EU; and (2) incumbents’ organizational power, or the scope and cohesion of state and governing-party structures (p.23)
More particularly, Levitsky and Way talk about Western linkage and leverage. Western leverage is high where a state ‘lacks bargaining power and are heavily affected by Western punitive action’ (p.41). Western linkage is defined as economic (trade or credit), intergovernmental (bilateral economic and military ties), technocratic (country’s elite being educated in the West), social (flow of people between borders), information (flows of Western ideas) and civil society linkage (Western NGOs) (p.43-4). Linkage works via three mechanisms:
  1. it [i.e., linkage] heightened the international reverberation caused by autocratic abuse [because linkage] increases the probability that... Western governments will take action in response to reported abuse;
  2. it created domestic constituencies for democratic norm-abiding behaviour [in the form of personal, financial and professional ties to the West and] international isolation triggered by flawed elections, human rights abuses [etc.] would put these ties and consequently, valued markets, investment flows, grants, jobs prospects and reputations – at risk; and
  3. it reshaped the domestic distribution of power and resource strengthening democratic and opposition forces and weakening and isolating autocrats [because] ties to Westerns governments... may provide critical resources to opposition and prodemocracy movements helping to level the playing field against autocratic governments (p.44-48, 70).
For countries where linkage and leverage is high, their model predict ‘external democratizing pressure is consistent and intense’ and ‘autocracies are less likely to survive’ (p.53). You cans ee what their model predictions in Table 2.1 above and Figure 2.2 below. Organisational power is essentially ‘state coercive capacity’ – the ability of the state to crush dissent. Levitsky and Way at times downplay the role of civil society but it seems to me their arguments (see point iii above) rely on it heavily. More importantly, state coercive capacity is simply the opposite side of the conflict between civil society and oppressive state structures.
As you can see, even where organisation power is high, linkage can still have significant democratising effects. Western leverage matters less than linkage but it’s still significant. Importantly, ‘only where ties to the West are extensive on all (or nearly all) dimensions – as opposed to being concentrated in one or two dimensions’ do these effects really come out (p.50). How does their model perform for their 35 countries? Remarkably well:
…[their] theory correctly predicts regime outcomes in 28 of 35 cases. Among the seven nonmatches, six are “near misses,” in that our theory accounts for key aspects of regime evolution and outcomes are close to those predicted by our theory. Only one case – Ghana – falls entirely outside our theoretical framework (p.340; note I’ve cut the table significantly because its so long, Table 8.1 is on p.342)
This gives us a reason to pursue further ties with authoritarians. I know what you’re thinking: the issue with Levitsky and Way is that they do not consider any purely Middle Eastern authoritarian countries. But I think what they have to say applies to them as well. In 2011, the New York Times noticed something that not many others did:
Even as the United States poured billions of dollars into foreign military programs and anti-terrorism campaigns, a small core of American government-financed organizations were promoting democracy in authoritarian Arab states… But as American officials and others look back at the uprisings of the Arab Spring, they are seeing that the United States’ democracy-building campaigns played a bigger role in fomenting protests than was previously known, with key leaders of the movements having been trained by the Americans in campaigning, organizing through new media tools and monitoring elections.
For someone who believes that weak civil societies perpetuate authoritarianism, U.S. and other Western programmes that promote democracy seem to be getting at the heart of the issue. This is directly relevant to Levitsky and Way’s third limb quoted above. And fortunately, we have a couple of studies. First, Scott and Steele (2011) find that
…countries receiving democracy aid experienced greater improvement in their Polity scores from 1988 to 2003 (5.54 points) than those who did not (2.40 points). This difference is statistically significant at the .01 level. Hence, the descriptive data provide initial—if limited—evidence that democracy aid is associated with progress toward democracy, with recipients outpacing non-recipients by more than twice the rate of improvement.
…even after controlling for a general trend (time, or year, which is statistically significant and positively related to democracy aid allocations), democratization is significantly and positively related to the allocation of democracy aid… In substantive terms, a change of 1 in the Polity score is related to a small, but statistically significant increment in democracy aid (about $50,000).
One of their mechanisms is what they call ‘agent empowerment’ which entails helping set up a vibrant civil society. I should note that Scott and Steele acknowledge that ‘studies examining specific democracy aid… offer mixed results’ but that ‘in general, [they are] more supportive of the fundamental aid-democratization linkage’ (see in particular Finkel et al (2006), Schmitter (2008) and Finkel et al (2008)). Bush’s ‘Freedom Agenda’ wasn’t just about democracy aid, it was broader. Gilley (2013) essay in Political Science Quarterly is instructive:
From 2001 to 2008, the policies of the Bush administrations in support of democracy in the Middle East involved four dimensions: rhetorical, diplomatic, material, and structural [e.g. Iraq and Afghanistan regime change]… [In particular] Bush increased the dollar amount per year spent on democracy in the Middle East by a factor of four and doubled its proportion of overall U.S. democracy assistance spending… Structurally, the Freedom Agenda included attempts to support local actors in building functioning democracies in Afghanistan and Iraq as part of the U.S. occupations of those countries.
…the countries most vulnerable to Freedom Agenda effects did indeed experience greater democratic gains in this period. The main outliers from this pattern are Jordan (below predicted) and Libya (above predicted).
In relation to the material and structural support, Gilley finds that ‘for the most part, the spending seems to have had the expected small but positive effects.’ More importantly, Gilley provides qualitative evidence for the other mechanisms. Some of them are dubious (for example, the U.S. was instrumental in the National Democratic Initiative in Bahrain but its probably a stretch to say they are linked to any reforms) and some seem sensible (for example, the U.S. support for the Damascus Declaration for Democratic National Change in Syria). More interestingly, Gilley argues that:
… it is probably fair to say that the diffusion effects of democracy building in Iraq and Afghanistan on the movements were negative overall. The manner of the Iraq invasion in particular and the missteps in the reconstruction effort, according to Gamal Selim’s careful analysis, “resulted in the discrediting of the Iraqi political experience among mainstream Arab political and civil society forces as a potential model to be emulated or diffused.”
I disagree: I think its clear that the regimes that exist in those countries are infinitely better than the previous regimes. I’m not going to address this because that goes more toward hard power than soft power analysis that I’m focussing on here. That said, the mechanism which Gilley states argues that there was an unexpected democratic effect isn’t limited to just Afghanistan and Iraq:
…vocal regime critics gained a new and unexpected voice. In particular, by sparking nationalist sentiments, the wars, like the rhetoric, discredited authoritarian regimes that had long claimed to defend the Arab world against U.S. imperialism, creating a new discourse of strength through democracy.
It’s for these reason that he claims the structural changes had a number of effect: some positive and some negative (see Table 3 above). I think Gilley’s study has a number of limitations (the lack of quantitative data being one) and I don’t really buy the ‘deligitimation’ mechanism described above. Briefly, I think the evidence for it is quite weak (see, for example, Bush and Jamal (2014) who find that American endorsement of women in politics has no average effect on popular support for women's representation) and doesn’t seem to correspond with support for democracy in the Arab world (see my review of The Missing Martyrs). That said, it does show how Levitsky and Way-type scenarios arises in the Middle East.
But it’s not just programmes like this that lead to a positive effect. Scott Ritter (2015) in his book The Iron Cage of Liberalism provides evidence for the first Levitsky and Way limb described above. Ritter’s argument is that ‘friendly international relations between the authoritarian regimes in Iran, Tunisia, and Egypt and their Western patrons’ leads to a insincere use of liberal ideals which then ‘constrains authoritarian regimes and their Western allies by holding them accountable to liberal expectations’ (p.18-19).
This, in turn, means that non-violent revolutions work because the regimes cannot suppress the people. The idea is that there is an ‘iterative, relation process’ by which (1) the authoritarian regimes, (2) the West and (3) the opposition forces in a society coalesce to create an ‘iron cage of liberalism.’ The West pressures its allies to reform any authoritarian bents that the regime has. The authoritarian regimes respond by espousing democratic ideals without making substantive changes. Ritter refers to this as a ‘façade democracy’ which is a weak attempt to accept liberal, democratic norms.
Opposition forces are strengthened directly by the West funding civil society programmes and the regimes allowing for modicums of reform which allow the emergence of a civil society. By espousing such ideas, citizens also judged their rulers by standards which they obviously fell short of and then accordingly had to offer more. It is these reinforcing mechanisms that lead to the iron cage, and therefore successful non-violent revolutions:
As a result of their close ties to the West, the shah, Ben Ali, and Mubarak eventually found themselves trapped by the liberal discourse they had adopted precisely in order to promulgate their relationships with the democratic world. When large crowds of nonviolent protesters materialized [the regimes were] unable to resort to violence [and] they were swept away by the empty hands of unarmed revolutionaries (p.214).
So far all I have talked about are effects on the overall status of the regime (i.e., our links help foster liberal / democratic norms). But there’s also other ways our links help. Swed and Weinreb (2015) argue that arms sales -> less repression. This is through two mechanisms: first, it creates a dependency on us and second (and more persuasive) ‘the acquisition of complex military hardware facilitates the diffusion of western ideologies regarding the role of the military in the state, and regarding when, how and against whom state-sponsored military activity can legitimately be used.’
[There is a] relationship between increasingly complex weapon systems (right-side axis), opportunities for collaboration (left-side axis), and overall levels of interpersonal interaction. That is, even the sale of small arms may generate some military coordination. But sophisticated weapons create a wider array of opportunities for interaction. With every purchase of a complex weapons system, instructors, operators, advisers, and other experts are employed to closely chaperone and guide recipients…
… even where narrow technical training has nothing to with the military’s political role or responses to civil unrest, it is hard to imagine that a wider array of ideas about professional standards, from matters of general military doctrine and strategy to professionalization and ideal civilian-military relationship, do not also diffuse.
Swed and Weinreb go through both qualitative analysis and quantitative analysis. Their former shows that the ‘behavior of the military elite [i.e., repression] in Tunisia, Egypt, Libya and Syria after Arab Spring demonstrations began is strongly associated with two empirical indicators of military westernization [i.e, the two mechanisms following Westernisation of the military described above].’ This is pretty self-explanatory and I wont be elaborating on it.
Their quantitative analysis, however, is more interesting. They look at 2,523 cases of political unrest between 1996 and 2005. Their bottom line results show that ‘the number of deaths that result from government-initiated response to political unrest is significantly lower where the military is westernized. These differences are particularly notable in poorer settings, and were particularly strong in the pre-GWOT (1996–2000) era.’
Model 3 shows that among poor countries whose main combat aircraft, tanks and APCs in the pre-1991 era did not come from western countries, we see e1.85 (6.4) more deaths than in wealthier countries. But among poor countries whose main weapon systems in the pre-1991 era did come from western countries, we see e-3.03 (3.7%) the number of deaths.
2.2 First Aid: Two Steps Forward, One Step Back
The literature on foreign aid is a mine field. It’s utter chaos. Not only are there a lot of contradictory findings but there’s just so much of it, I will never comfortable with or attempt to read all of it. Hence, what follows are summaries I’ve taken from studies rather than actual studies. My main concern here is to find out whether aid we give has a negative, neutral or positive effect in terms of democratisation and human rights. But I just want to re-emphasise: democratisation is predominantly a domestic issue.
A wide consensus of the literature holds that criticality generally rests with internal rather than external factors, especially in regard to democratic consolidation (Gilley (2013))
Above I discussed the effect of democracy aid, this subsection is about foreign aid generally. If I’m saying that oil rents are relevant to regime survival, surely foreign aid and assistance would work in the same way? This is plausible and, indeed, there are some studies to support the idea (see in particular Morrission (2009) who finds revenues, no matter where they come from, increase regime durability and Djankov et al (2008) which finds that foreign aid has a worse impact than oil rents). But, as always, we should follow literatures not individual studies. Scott and Steele say

…most studies of broad foreign aid or official development assistance (ODA) conclude that such assistance does not affect democratization (see Collins 2009)… For example, Knack’s (2004) extensive study of aid and democratization from 1975 to 2000 concluded that foreign aid was not a significant factor in democratic change…
The aforementioned Ahmadov meta-analysis finds that foreign aid is not a significant cofounder on the oil rents variable. I would argue that given the domestic causes of authoritarianism, an unstable variable like foreign aid just doesn’t seem to have a positive or negative effect. But, Bader and Faust (2014) review all of the literature on the subject suggest that more recent studies find that it has a negative effect:

…. more recent and more nuanced studies of the relation between foreign aid, political structures, and regime survival suggest that foreign aid has an amplification effect (Dutta, Leeson, and Williamson 2013). Accordingly, aid tends to stabilize and consolidate the type of regime it encounters. Foreign aid stabilizes autocratic structures in autocracies while helping to consolidate democratic governance in democracies. Moreover, foreign aid feeds patronage politics in autocracies, but not in democracies (Hodler and Raschky 2014). These findings suggest that foreign aid is politically captured by authoritarian governments, and indeed, makes autocratic survival more likely (Bueno de Mesquita and Smith 2010; Ahmed 2012).
Like I said, the literature on foreign aid is a mess. There are three studies that are referenced frequently to show that aid can entrench regimes: Bueno de Mesquita and Smith (2010); Ahmed (2012) (see above extract) and Morrisson (2009) (which is quoted above). I think there are two fundamental flaws with this line of research. First, Bermeo (2011) shows that there is a distinction based on who is giving the aid. Many of the larger cross-sectional studies look at all foreign aid and look for democratisation. Bermeo shows that this is not the case:
Model 2 separates aid from democratic and authoritarian sources as described above. The results provide strong support for H[ypothesis] 1: while increases in aid from democratic donors are associated with a greater likelihood of a democratic transition, the opposite is true for aid from authoritarian sources. Monte Carlo simulations show that, holding all other variables at their median, a change in the amount of democratic aid from its value at the 25th percentile to its value at the 75th percentile is associated with an increase in the likelihood of a democratic transition during a country year from a baseline of 2.15 to 3.12%, a 45% increase
Second, is another study by Bermeo (2014). She notes that the relationship of aid and democratisation is heavily dependent on the type of aid. Fungible aid is much like oil in helping regime durability but non-fungible aid is not. Whether to give fungible or non-fungible aid is a decision made on the basis of strategic importance (high strategic importance -> fungible aid). In this study she is only looking at democratic donors and she does find there’s a difference:
The coefficient on aid is negative but not significant for the full period (Model 1) As Models 2 and 3 show, this masks heterogeneity across periods. During the cold war, more aid is associated with a decreased likelihood that a country experiences democratic change. The same is not true in the post-cold war, where the coefficient on aid is positive but not significant.
The idea is that in the pre-1991 period, aid decisions would have given more weight to strategic rather than purely democratic concerns. Interestingly, you’d think that these strategic concerns might arise again in the post-2001 period because of the War on Terror. Apparently not: ‘If the period is further restricted to the post-2001 “war on terror” time period… there is still no negative effect of foreign aid.’
But Bermeo asks something interesting: can states which have strategic importance to us, manipulate the composition of the aid? She looks at the top 5 states which obtain military aid (not included in foreign aid, fyi) and gives them an indicator score of 1:
The coefficients on the interaction term between aid per capita and strategic importance, Top 5*Aid are negative in both models and significant in the post-cold war period. The results are consistent with strategically important recipient governments using aid to prevent democratic change even in the post-cold war period.
Bermeo then does something even more interesting: she re-runs the analysis on Morrison (2009), Bueno de Mesquita and Smith (2010) and Ahmed (2012). I wont repeat each result but here’s her conclusion:

The articles [i.e., the three listed above] revisited in this section claim that foreign aid operates in a similar manner to oil revenue or remittances and that it suppresses political change in authoritarian countries. In no case does a further examination of the results support the conclusion that aid prevents political change in the post-cold war, or that aid operates in a similar way to other forms of non-tax revenue.
Gibson et al (2015) find that increased technical assistance from the West led to less patronage which, in turn, led to more liberal political concessions. As they show:
Embedded image permalink
[an] increase in technical assistance is associated with an increase incumbents’ level of concessions. Technical assistance as a share of GDP is a statistically significant predictor of political concessions across all the models... these results are consistent with the theory that technical assistance played a role in leaders’ decisions to liberalize politics in their country... an increase of 10 percentage points (0.1) in TA/GDP [technical assistance] would increase political concessions by nearly a full point (0.8).
Embedded image permalink
... an increase in technical assistance from its minimum to its maximum level decreases government expenditures from 19% of GDP to about 10% of GDP. Wages likewise decrease from about 7% of GDP to 1.5% of GDP [these are used as proxies for patronage]/
All the above means I believe that Western foreign aid has either neutral or positive effects in most cases. There are, however, cases where it is negative and represents one step back.  That said, per Bermeo, foreign aid from the West does not usually entrench states. But I would be open to reductions of fungible foreign aid to authoritarian regimes (but I’m not overly insistent on it). More importantly, this must all be weighed against the effect of what happens with the alternative - i.e., not only the loss of benefits above but also the effect of authoritarian donors (see section 3 below).
2.3 A Plea
Every time I’ve ever read a book about international affairs, even books not focussed on the issue of regime persistence or authoritarianism, there will be a reference to Eva Bellin’s essay ‘The Robustness of Authoritarianism in the Middle East: Exceptionalism in Comparative Perspective’. It currently has 502 citations on Google Scholar. In the essay she starts by citing the work of the Skocpol who argues that the key to successful revolutions was state coercive capacity. This is a fair point, one to which I have alluded to above. Bellin then asks what factors influence state coercive capacity?
She lists four factors: fiscal health, institutionalisation of the security apparatus, the extent to which the regime faces high mobilisations of the population against it and:
…the robustness of the coercive apparatus is also shaped by successful maintenance of international support networks.
I know a few international affairs academics read this blog so I have a question: why are you all quoting this? Why is this your main reference? What is your obsession with Bellin? It barely qualifies as a qualitative study, it’s almost entirely pure assertion. In her essay, she repeats the same assertions for the Middle East in two paragraphs with no real qualitative let alone quantitative evidence. Here are a few extracts:
With regard to international support, the region is exceptional for the unique position it enjoys in the international arena… the region is exceptional in that the cold war's end has not signaled great power retreat from patronage of authoritarianism, as in Latin America, Africa, and elsewhere. Playing on the West's multiple security concerns has allowed authoritarian regimes in the region to retain international support.
No attempt is made to explain the mechanism, no attempt is made to provide an example of this international support. I make a plea: stop quoting Bellin (2004). There are actual studies, there are actual qualitative studies you can use rather than this unsubstantiated essay. Bellin (2012) gave an update to her oddly famous 2004 essay where she tried to see how well her theory stood up after the Arab Spring. This does contain qualitative elements worth quoting but even then, it’s really not that persuasive:
The first two factors—fiscal health and maintenance of international support—are crucial to determining the coercive apparatus’ capacity, that is, the physical wherewithal to muster the men and materiel necessary to repress. 17
Footnote 17 states:
Though at times international support can prove important to sustaining the coercive apparatus’ will to repress as well. For example, during the Tahrir uprising in Egypt in January and February of 2011, clear signaling on the part of the United States that it would not look kindly on a decision by the Egyptian military to massacre civilians certainly chipped away at the Egyptian military’s will to repress.
Oh, right. The next mention of international support networks:
No doubt international factors came into play in the military’s calculations as well. 36
Footnote 36 states:
Throughout this period, the United States was communicating to the Egyptian military leadership, primarily via mil-to-mil contacts, that shooting on civilians would not be in the military’s institutional interests. The United States possessed significant persuasive powers, not only thanks to the $2 billion in military aid that it provided Egypt each year, but also thanks to the fact that the Egyptian military’s arsenal was entirely U.S. made. Consequently, the Egyptian military was entirely dependent on the United States for upgrades, spare parts, and training. To remake the Egyptian military would take years. In the meantime it had to listen carefully to the counsel of its primary supplier.
That is it. That is the extent of what she says on international support networks. You can guess what my second plea is: if your argument is that the West is propping up these regimes, don’t quote Bellin (2012) either. Which brings me to Jason Brownlee. I have quoted his studies a number of times in this post. Brownlee (2002) is better study to use (it only has 120 citations on Google Scholar). His is a rich qualitative study in which he looks, in depth, at Syria, Libya, Iraq and Tunisia and finds:
In Syria, Iraq, Libya, and Tunisia, independence from a superpower patron left the regimes free to suppress domestic insurgencies. Of course, the converse does not always obtain. Dependence on external support does not necessarily constrain an authoritarian regime. Among Snyder’s cases of high dependence, Somoza García’s experience illustrates this point… [Samoza Garcia] received large amounts of support from the U.S. government, this aid helped him repress challengers, whereas in the latter cases external dependence impaired the leaders’ responses to crises
By contrast, he goes on to say ‘the Iranian, Filipino, Cuban, Haitian, and later Nicaraguan regimes found themselves hampered by their reliance on American support.’ This support my general argument but Brownlee then says something counter to it as well:
The Somoza García regime is not the only instance of a government enabled by the U.S. to suppress movements for political change. In the Middle East, for example, the current Saudi Arabian and Egyptian regimes receive large amounts of military and, in the case of Egypt, economic aid, with few political strings attached
Brownlee doesn’t provide anything substantive for this – but then, in 2012 he released Democracy Prevention: The Politics of the U.S.-Egyptian Alliance. I don’t consider the book particularly persuasive (Scott Ritter’s analysis is far more in line with what actually happened) but people should be quoting Brownlee (2002, 2007 and 2012), not Bellin. One of the fundamental problems with Brownlee’s analysis in the book is the failure to consider the counter-factual: what if the West didn’t have close relationships? And that brings us to the final section of this post.
3. Black and White Knights
3.1 The Russo-Chinese Devil You Don’t Know: The Alternative
It should be clear that from the sections above that without our assistance, these authoritarian regimes would still exist and they would be significantly more repressive and illiberal. But the argument can go further: the alternative isn’t likely just a withdrawal of benefits but the entrance of what political scientists call ‘Black Knights’ (other authoritarian countries that assist them) with significant negative effects.
I know I’m probably boring you but I want to be clear and consistent: these Black Knights are not a cause of regime persistence. Regime persistence is, again, rooted in domestic factors – but in the same way the West can create positive impacts, Black Knights can create negative impacts. The above study from Bermeo has already shown that aid from authoritarian countries leads to less likelihood of democratisation. But there is some literature to suggest it isn’t as bad: Bader (2015) looks at the effect of aid from China on regime persistence. The first point is:  
As regimes differ in their domestic strategies of political survival, they vary in their ability to translate this type of assistance from China into their specific survival strategy.. party-based regimes are more vulnerable to domestic mass protest than other regimes: as party-based regimes often claim to serve the broader interest..
…the type of external assistance that China provides is more useful to some types of regimes than others. More specifically, existing literature would suggest that party-based regimes have higher incentives to engage in performance-based legitimation strategies and are more able to translate China’s economic engagement into power preserving societal support than other types of regimes
Given that Chinese assistance is usually strongly tied, personal regimes, for example, which are strongly reliant on clientelist networks, may find it difficult to channel Chinese assistance selectively to their particular support group defined by familial, ethnical or clan ties.
Bader finds:
Figure 1 illustrates the effect of Chinese Economic cooperation for model 2 by plotting how the probability of collapse (Y-axis) changes with increasing amounts of Economic cooperation (X-axis) for party and non-party regimes. For non-party regimes (left graph), the slope of the function is increasing, while it is slightly decreasing for party-based regimes (right graph).
…column 3 [of table not shown] estimates how Economic cooperation affects regime durability if only transitions to democracy are considered (as opposed to no regime transition). Interestingly, as column 3 indicates, the direction of the unconditioned effect of Economic cooperation on transition to democracy is positive – that is, increasing the risk for democratic transition. However, as Figure 3 shows, this effect is significant only for non-party regimes and only for Economic cooperation between 0.2 and 0.7 per cent of GDP.

So maybe it’s not certain for all regime types that Black Knights will disrupt democratisation. The argument I would put forward is that where the West is not operating, and Black Knights are, you don’t have the benefits described above. But I’m not inclined to accept there isn’t a negative effect (contra Bader). Elsewhere, Bader (2013) finds that a ‘1% increase in export dependence on China lowers a leader's hazard rate of losing power by 5% (1 − exp(−0.06)).’ Or take Russia’s behaviour: Babayan (2015) outlines a comprehensive review of Russia’s actions in its area:
…through economic sanctions, military threats, and even through such formal institutions as the Eurasian Union, Russia has contributed to the stagnation of democratization in its near abroad. It counteracted democracy promotion or, for that matter, any other Western policies, which it considered a threat to its geostrategic interests and ambitions for restoring its great power status.
As Levitsky and Way state:
[Western] leverage may be reduced by the existence of what Hufbauer et al. call “black knights,” or counter-hegemonic powers whose economic, military, and/or diplomatic support helps blunt the impact of U.S. or EU democratizing pressure. Russia, China, Japan, and France played this role at various times during the post–Cold War period, using economic, diplomatic, and other assistance to shore up authoritarian governments in neighboring (or, in the case of France, former colonial) states. Examples include Russia’s support for autocrats in Belarus and France’s support for autocrats in former colonies such as Cameroon and Gabon (p.41)

Better is Vanderhill’s book Promoting Authoritarianism Abroad (2013) gives detailed qualitative analysis of how authoritarians can act as Black Knights. Her approach, fortunately, acknowledges local forces and shows international effects are not determinative. Rather, as per Western influence, can change incentives to behave in certain ways.
There is a lot of fairly persuasive literature which argues that Black Knights have a neutral effect. The latest issue of the journal Democratization is a special issue filled with essays about the effects of Black Knights. Here is an extract from the introductory essay:
Western democracy promoters are likely to empower liberal groups in the target countries, while countervailing efforts by non-democratic regional powers will empower illiberal groups. The differential empowerment of domestic forces depends in turn on the leverage of the EU and the US powers as compared to that of illiberal regional powers in terms of credibility of commitment, legitimacy, and resources... the decisive struggle is being fought in the target country between pro-democratic and anti-democratic forces – and external actors cannot do much more than try to affect this domestic balance of power.
Way (2015) (the same Way of Levitsky and Way) recently published ‘The Limits of Autocracy Promotion’ in the European Journal of Political Research in which he concludes:
Russia has played a role in promoting secessionist conflicts in Georgia, Moldova and Ukraine. At the same time, a closer examination of Russian foreign policy and its impacts raises doubts about just how much the failure of democracy in the region can b blamed on Putin or previous Russian governments
Way makes a few good points that are worth repeating. Black Knights do not promote authoritarianism in the way that the West attempts to promote democracy. It is not an ideological end, rather, Black Knights pursue their own interests which may not necessarily tack with regime type. I take this all as supporting the idea that democratisation is primarily a domestic issue: ‘much of the pressure and assistance has been directed at countries that already have quite weak democratic pre-requisites.’
There is an interesting point that Russian activity may empower democratisation because it riles the population up (see Borzel (2015)). Nonetheless, counter-factuals in which Black Knights swoop in are vital: these authoritarians will simply get their arms from someone else, they will get their aid from someone else and they will continue to survive.
We have already seen that the West constrains its allies. And before you all give me examples of Western allies committing human rights abuses, you’re missing the point: think of the alternative. I have provided evidence of a constraining effect. It might not be good enough to stop all abuses, but its better than not having that effect. I’m open to accepting that the role of Russia and China is similarly neutral (primarily because its a domestic issue).  To repeat: my primary argument is that Black Knights are bad because they don’t bring the benefits the West does. My secondary argument is that Black Knights lead to negative outcomes (which is not as well supported).
3.2 The Liberal, Democratic Devil You Know and Balance
I have outlined benefits of the relationship to the countries and their citizens already but engaging with these regimes has other benefits to us: (i) terrorism related; and (ii)) wider geopolitical concerns. To take the first, we have extensive security and intelligence cooperation with our Gulf allies. Our Prime Minister has openly said:
… you can be prime minister and say exactly what you think about every regime in the world and make great headlines, and give great speeches. But I think my first job is to try and keep this country safe from terrorism and if that means you have to build strong relationships sometimes with regimes you don't always agree with, that I think is part of the job and that is the way I do it… I can tell you one time since I've been prime minister, a piece of information that we have been given by that country has saved potentially hundreds of lives here in Britain.
In a report prepared by the House of Commons Foreign Affairs Committee, we find this:
British-Saudi collaboration has resulted in the foiling of AQAP terrorist attacks, which would have caused substantial destruction and loss of life, including the provision of information to protect British interests… The initial alert came from the Saudi authorities, who have been quick to provide information to protect British interests on many other occasions
This is the part of the essay where I, unfortunately, cannot provide a study but ask you to use your own judgment. Do you think these are lies? Do you think an ally is more or less likely than a non-ally to share information with us about terrorism? You have no idea how frustrating I find it that we don’t have a study for this: I don’t trust my instinct (and without a study, I don’t know the extent), but you might trust yours.
By the way, it is no use say that that Saudi also funds terrorism. In respect of ISIS, al Qaeda and al Qaeda in Iraq, these are insignificant amounts. But even accepting the argument, you are not choosing between an alternative of (a) no terrorist support and (b) terrorist support. You are choosing between (a) terrorist support, and hindrance in form of intelligence and (b) terrorist support with no hindrance (i.e., no assistance to the West).
Second, we have vital geopolitical interests that are served by these alliances. Unfortunately this will require a post in and of itself. Briefly though, we can summarise the arguments from John Ikenberry. His book Liberal Leviathan is required reading for everyone, the salient parts of which are summarised in ‘Don’t Come Home America’:
Historically, as Gilpin and other theorists of hegemonic order have shown, the background security and stability that the United States provided facilitated the creation of multilateral institutions for ongoing cooperation across policy  areas... The United States’ extended system of security commitments creates a set of institutional relationships that foster political communication. Alliance institutions are in the first instance about security protection, but they are also mechanisms that provide a kind of “political architecture” that is useful beyond narrow issues of military affairs.

I want to make it very clear: I am not an ardent realist. Our geopolitical interest should be those which promote overall liberal democracy all over the world. Our alliances help preserve peace in the region (e.g. ensuring the Egypt-Israel peace treaty sticks or working to undermine Iran’s menacing activities across the region) an help foster a liberal world order. Even the spread of broadly ‘neo-liberal’ norms helps bring peace (see The Empirics of Free Spreech, Part 1). The benefits – to both us, the region, the people of these authoritarian countries – therefore demand that we maintain relationships.
3.3 Limitations of the Argument
Here’s the Argument that should be reasonably well supported by the above:
  1. These regimes will exists without us. They will commit human rights abuses without us. We do not cause or contribute in any critical way.
  2. Having a relationship with them allows us to tame them – even when we aren’t intending to – and leads to the following better outcomes: slightly more democratic, slightly more liberal, slightly less likely to kill protestors, significantly more helpful for our vital geopolitical ends.
  3. Having no relationship with them not only leads to a withdrawal of benefits obtained from the relationship but opens space for Black Knights which have worse outcomes.

The West is simultaneously powerful and impotent: unless you have a good plan for regime change, your best bet is engagement. Obviously this argument has limits: when a regime is about to slaughter tens of thousands or people or commit a genocide, if you have a viable plan for humanitarian intervention, you should eliminate the regime. I supported breaking off our ties with Libya and destroying the regime in 2011, but I also supported engagement prior to that. There is no contradiction, only a balance. Make no mistake, I am not apologising for these regimes. They are authoritarian cess pits that must be eradicated. The only point, though, is about how to reach this goal.
“How many people need to die before you change your mind?” I’m really sorry, but I don’t know. I want maximum liberal democratic outcomes, sometimes that means war and sometimes that means reluctance peace. “ISIS and Saudi Arabia are pretty much the same.” Well, no not really but the most important difference between the two is that we have a viable military option to destroy ISIS. Present me a plan to destroy Saudi and I’ll support you. But right now, we have limited resources and we’re not even pursuing the right strategy for ISIS, so Saudi can wait.
A second limitation the argument has is that I am not arguing for us to treat Germany and Egypt in the same way. We should raise human rights concerns, we should condemn them and we should even think about short term sanctions or withdrawals of arms licenses to get the message through. I do not think this will make a difference materially (i.e., those arms licences withdrawals are irrelevant to guns being used to kill people), but over the long term may help shape behaviours. None of this, however, should be prioritised over maintaining relationships which, as shown, leads to positive outcomes.

The above is a tweet from Rory following someone ridiculously, and quite frankly pathetically, saying that the British Empire was as bad as the Nazis. Before you all start talking to me about flying the flags at half-mast for Saudi royals and examples of callous disregard for human life on the part of the West, what Rory says about Empire, I believe about our relationship with these authoritarians.
I think the West usually gets the balance mostly right. It condemns abuses of its allies, it tries to push for good outcomes and engages in the withdrawal of licenses. I often see people – Daniel Wickham comes to mind – who will never be happy with the level of condemnation or action (see this exchange). I usually avoid talking about people’s motivations because I do think its mostly irrelevant and most people have good intentions. But  I can’t shake the feeling that there’s an anti-Western zeal, the conclusions of which aren’t necessarily unreasonable, behind a lot of condemnation. Maybe I’m wrong and this an expression of one my biases. I haven’t provided any evidence so that’s plausible. In any event, for the purposes of this post, I’m not interested in individual countries or examples: I am interested in a general strategy and general outcomes.
There is a third limitation in that there is a massive, gaping hole in the argument above: I am primarily a consequentialist. I care about outcomes. But what if promoting democratic, liberal civil society leads to bad (non-liberal, non-democratic) results? Take Egypt: let’s say Western support was part of the reason why Mubarak fell (for all the reasons laid out above). Is the situation better today than under Mubarak? It would be a difficult argument to make. And I don’t have an answer to this criticism. This post is simply to show that the same short-term argument that tries to make us feel guilty in the West is wrong in that same short term period.
The final point I want to make is a point of regret. I have tried to emphasise that these authoritarianism is primarily a domestic issue. But by talking about the West so much I have likely given the impression that we matter a lot. We do not. And that is why our relationship has to be this way. It’s far too easy to say our support for bad regimes is bad. It should be apparent that I support something close to the status quo because a lack of support for bad regimes may lead to worse regimes and worse outcomes. You can be a liberal, without being an ardent realist, and still not feel guilty or ashamed of what we do but still demand less dickish-ness.
P.S. If you managed to read all of this, please can I ask that if you have something to say, leave comments below or write a post. Drafting responses to tweets is infinitely more fruitless, marginally more inconvenient and significantly reduces the likelihood that people will see it. And I hate it.
[1]  As an interesting counterpoint to the idea that civil society matters for democratisation, consider Sahoo (2014). He documents how the Vanvasi Kalyan Parishad, a Hindutva voluntary organization, was bad for democratisation. The Hindutva ideology is essentially a nationalist ideology that often acts at the expense of religious minorities. Sahoo looks at the voluntary organisation’s activities which it used to set up a civil society grouping. 
The VKP ‘utilized its cultural programmes, such as Bhajan Mandlis (evening religious gatherings) and Bal Sanskar Kendras (childcare centres), to socialize the tribal community into Hindutva ideology.’ These programmes have contributed to the ‘spread of violence and demonization of religious minorities but also a serious undermining of cultural pluralism and democratic values of Indian society.’ It’s clear, then, that civil society does not cause democracy. Rather, what matters is the type of group. The point about the Middle East is that there is no civil society, let alone a good or bad one. But as Sahoo states, we should not forget this vital factor:

… there are “uncivil” society groups that could undermine democratization and the functioning of democracy. The essential question is what type of civil society positively contributes to the democratization process. The choice is not between civil society and state, but rather between different types of civil society, their nature, and the interests that preside over them. If civil society is to contribute positively towards democratization it needs to be dominated by groups that have an interest in democratic civility and in pushing the state in a liberal democratic direction.

[2] Blaydes and Chaney are not using leadership duration as a proxy for executive constraints, they provide evidence for the association between leadership duration and executive constraints. Parliamenti is an indicator based on a dataset showing how often parliament would meet in a polity. You can see from the following table, Column 2 and 3, the results clearly show a positive relationship between Parliamenti and ruler stability:
[3] I sent a draft of this to Pseudoerasmus and he has some misgivings about the Chaney and Blaydes study. He said that whilst he agrees that slave armies ‘could have had something to do w[ith] the institutional evolution of core [I]slamic societies’ that ‘even if that's true, it's not exactly clear why modern Islamic societies suffer from this institutional persistence.’ It’s a good point which I’ve only partially explained above (i.e., that there’s a lower starting point for civil society). But, I think when you add in the factors below, you can account for the persistence.  Pseduo also has some good arguments about decentralisation being bad for economic development in Europe. My only response is to say that Blaydes and Chaney probably have the broad-brush correct and that’s all that matters for my argument.
[4] The first distinction that may explain different outcomes between Middle Eastern states I came across years ago was Spinks et al (2008). They were looking at whether there were difference in liberalisation between republics and monarchies in the Middle East. The find that monarchies commit less personal integrity and civil liberties violations, have more press freedom and HDI scores than republics in the Middle East (see Table 2).
The reason? Spink et al say it’s because ‘monarchical rule is based fundamentally on legitimacy and tribal tradition, many of the incumbent republican political systems were the result of coups.’ Accordingly, ‘monarchies base their legitimacy and power on tribal tradition, which appears to structure monarchical politics in a manner more conducive to liberalization.’

I don’t buy it. In particular, I don’t buy the mechanism. Look at Saudi Arabia’s score in Table 2 compared with any other republic. The authors say that Saudi Arabia’s outlier status ‘does not necessarily disconfirm the differences that seem to set monarchies apart’ – which is plausible, until you read Josua and Edel (2015) who find that the distinction doesn’t hold up for regime repressive response to the Arab Spring. Rather, as per Brownlee et al (discussed above), hereditary regimes may have more loyalty when coupled with oil rents

[5] See also Fleck and Kelby (2010). They find ‘U.S. aid flows–for the poorest as well as other developing countries–increased with the War on Terror’ but that ‘  after rising for 35 years, the emphasis placed on need has been falling steadily for core aid recipients during the War on Terror.’
[6]  Of course, you can say they wouldn’t withdrawal all assistance but even people who make this argument accept that ‘there would be damage to the quality and the timeliness of information provided to us’ (see House of Commons Foreign Affairs Committee Report).
Bibliography [To Be Updated]
Anders and Alaksen (2013), ‘Oil and Political Survival’, Journal of Development Economics (2013), Vol. 100, Issue 1, 89
Bank et al, ‘Long-term monarchical survival in the Middle East: a configurational comparison, 1945–2012’, Democraization (2013), Vol. 22, Issue 1, 179
Blaydes and Chaney, ‘The Feudal Revolution and Europe’s Rise: Political Divergence of the Christian West and the Muslim World before 1500 CE’, American Political Science Review (2013), 1
Chaney, ‘Democratic Change in the Arab World, Past and Present’, Brookings Papers on Economic Activity (2013), 1
Djankov et al, ‘The Curse of Aid’, Journal of Economic Growth (2008), Vol.13, 169
Dunning, Crude Democracy, Cambridge University Press (2008)
Egorov et al, Why Resource-Poor Dictators Allow Freer Media: A Theory and Evidence from Panel Data, American Political Science Review (2009), Vol. 103, Issue No. 4, 645
Fleck and Kilby, 'Changing aid regimes? U.S. foreign aid from the Cold War to the War on Terror', Journal of Development Economics (2010), Vol. 91, Issue 2, 185
Gibson et al, 'Did Aid Promote Democracy in Africa? The Role of Technical Assistance in Africa’s Transitions', World Development (2015), Vol. 68, 323
Haber and Menaldo, Do Natural Resources Fuel Authoritarianism? A Reappraisal of the Resource Curse, American Political Science Review (2011), 1
Kuran, ‘Legal Roots of Authoritarian Rule in the Middle East: Civic Legacies of the Islamic Waqf’ American Journal of Comparative Law (2015), Vol. 63
‘The Political Consequences of Islam's Economic Legacy’ Philosophy & Social Criticism (2013), Vol. 29, 345
Levitsky and Way, Competitive Authoritarianism: Hybrid Regimes After the Cold War, Cambridge University Press (2012)
Sahoo, ‘Civil society and democratization: a counter-case from India’, Democraization (2014), Vol. 21, Issue 1, 480
Salter et al, ‘Economic Origins of Democratic Breakdown?’, Perspective on Politics (2014), Vol.12, Issue 2, 1
Swed and Weinreb, Military westernization and state repression in the post-Cold War era, Social Science Research (2015), Vol. 53, 270
Ulfelder, ‘Natural-Resource Wealth and the Survival of Autocracy’, Comparative Political Studies (2007), Vol. 40, Issue No. 8, 995
Wright et al, ‘Oil and Autocratic Regime Survival’, British Journal of Political Science (2015), Vol. 45, Issue 2, 287


  1. I found the above piece informative and illuminating. It gave me a much clearer understanding of the issues and I agree with your conclusions.

    I have only been following your work for a few months, and subscribe to your RSS feed so that I don't miss a post. It is always thought provoking, even when I disagree with the conclusions as with your previous post. Thank you for the link to the counterviews. I had read Jamie Palmer's before, but not the other two.

    I am a great fan of Timur Kuran's work, and found "The Long Divergence" very insightful.

  2. Thanks for this - I found it very interesting although (particularly without sufficient knowledge of the context) I didn't always feel I had internalised each stage of the argument. The larger points towards the end came over very clearly and forcefully though.

  3. Interesting post (just like many of the previous ones)! One small edit though: It was not the former UN observer Scott Ritter who wrote the Iron Cage of Liberalism, but rather Daniel Ritter.

  4. Hai, If your career will be in politics or you want to grow your skills in political path, I think this website will give the right way.
    Online masters in political science india
    Career as a politician in india
    Election management courses in india

  5. There were Islamic guilds, although they differed a bit from European ones.

  6. The West, Black Knights, and the Political Economy of Regime Persistence explores the intricate dynamics of power and governance. Through meticulous analysis, it delves into how historical legacies and economic factors shape the endurance of regimes. This topic could offer insightful parallels for understanding classroom dynamics in my PGCE behavior management assignment.