1
A Theory ofSultanism 1
A Type of Nondemocratk Rule H. E. Chehabi and Juan J. Linz
The concept of "sultanistic regime" emerged from Juan Linz's comparative analysis of nondemocratic regimes, which systematically developed the differences between totalitarian and authoritarian forms of rule.1 At the time Linz developed this distinction in the early 1970s, political democracies were few, whereas the range of nondemocratic regimes was enormous. Since totalitarian regimes were rare and limited to the communist world,2 it was tempting to group all other nondemocratic regimes in the residual category "authoritarian," which denotes a wide range of governments with distinctive characteristics that allow their societies a limited pluralism short of genuine democracy. Authoritarian regimes can thus be mostly civilian one-party states, ranging from a highly institutionalized authoritarian regime such as Franco's Spain to the ephemeral single-party regimes of Africa in the very early days of independence; nonhierarchical or hierarchical military regimes like the bureaucratic-authoritarian regimes of the Southern Cone of South America;3 or even royal dictatorships like those in the Balkans in the interwar years (as opposed to traditionally legitimated monarchies).
The structural differences among them are therefore vast.4 But as vast as they are, some regimes appeared distinct on all the major dimensions used in the conceptualization of nondemocratic rule; Linz called them "sultanistic." The differencesjpetween these and authoritarian or totalitarian regimes are not Tnerely a matter of degree but He in-their, rulers'over airconceptfonoT politics,
4     "     COMPARATIVE STUDIES
A Theory ofSuhanism 1   • 5
the structure of power, and the relation to the social structure, the economy, and ultimately the subjects of such rule. But before analyzing "sultanistic" regimes, the use of the term should be explained.
Excursus on "Sultanism"
The term "sultanism" was<>riginaJlygoinedby Max Weber, whe-nsed-iHo-rgfef •• Jo an extreme case of patrimonialism, which in his tripartite classification of j the forms of legitimate authority is a form of traditional authority:
Patrimonialism and, in the extreme case, sultanism tend to arise whenever traditional domination develops an administration and a military force which are purely personal instruments of the master. . . . Where domination is primarily ■■ traditional, even though it is exercised by virtue of the ruler's personal autonomy, it will be called patrimonial authority; where indeed it operates primarily on the basis of discretion, it will be called sultanism. . . . Sometimes it appears that sultanism is completely unrestrained by tradition, but this is never in fact the case. The non-traditional element is not, however, rationalized in impersonal terms, but consists only in an extreme development of the rulers discretion. It is this which distinguishes it from every form of rational authority.5 I
Weber's notion of patrimonialism caught on and was used extensively, and in an influential article Guenther Roth applied it to Third World politics.6 Given the differences between the modern states in which the patrimonial logic \ operates and the traditional patrimonial systems Weber had in mind, the term "neopatrimonialism" came to be widely used.7 But Weber's formulation of j| patrimonialism's extreme form, "sultanism," was neglected by scholars.8
The term was adopted by Juan Linz in his classification of nondemocratic governments. When he began working on Spain's Franco regime in the late 1950s, he soon realized that the model of "totalitarianism" then current, based on the Stalinist and Nazi experience, did not fit. The result was the conceptualization of the "authoritarian" regime, whose different aspects he explored in a series of essays.9 A meeting with a Spanish exile was to convince him that the [ authoritarian/totalitarian dichotomy did not exhaust the range of nondemo- j cratic regimes either. In the early 1950s Linz met his Spanish compatriot Jesus J de Galindez, a representative of the exiled Basque government, who had taught \ international law in the Dominican Republic. Although a republican emigre, \ Galindez was treated courteously by Spanish diplomats in New York. But when | he wrote a doctoral dissertation at Columbia University revealing some of the ; inside workings of the Trujillo regime, he confided to Juan Linz in 1955 that he feared for his life, and that he had deposited his manuscript in a safe place in 1
case something happened to him. Soon afterward the dictator had Galindez abducted in New York and taken to the Dominican Republic, where he was tortured to death.10 The contrast between Franco's nondemocratic regime and Trujillo's rule led to the conceptualization of a regime type for which Linz borrowed Weber's term "sultanism," since it too rested on the extreme development of the ruler's discretion. Moreover, Weber, although constructing the ideal type of sultanism in a section on traditional authority, explicitly contrasted the traditional basis of patrimonialism with the discretionary aspect of sultanism (see quotation above), implying that tradition played little role in the latter. Just as Franco's rule became the archetype of an authoritarian regime, Trujillo's became that of a sultanistic regime in Linz's 1973 classification of nondemocratic regimes."
Subsequently a number of scholars confirmed the applicability of Linz's paradigm to a number of regimes not mentioned in the 1975 article. Crawford Young and Thomas Turner wrote about Mobutu's regime in Zaire that "in the personalist patrimonial state fashioned by Mobutu, we may discern much that resembles what Linz, borrowing a Weberian term, has labelled 'sultanism.' "12 Terry Karl characterized the regimes of Juan Vicente Gomez and Marcos Perez Jimenez in Venezuela as sultanistic, and H. E. Chehabi explored how the Shah of Iran's sultanism contributed to the Islamic revolution. For the Philippines, John Thayer Sidel applied it to local politics, and Mark Thompson to Marcos's rule. Finally, Samuel Huntington..wro^ such as those of Marcos and Ceausescu, like.those.of.Somoza, Duvalier,,Mobutu, and the Shah, exemplified Weber's model of sultanistic regime characterized by patronage, nepotíšmrcfonýism, and-corruption."13 The criticism so often leveled against inductively derived theoretical concepts—that since, in Pareto's words, they lead "from facts to concepts, and from concepts back to facts,"14 they produce circular argument—does not apply to sultanism as a concept, since it has fruitfully been applied to a different set of facts.
Independently of Linz's revival of the term, Richard Sandbrook called Weber's notion of sultanism "more relevant to the circumstances of contemporary Africa." He added that it nourished "under a number of guises: civilian, quasi-military or military forms of government, one-party or competitive-party systems or even under the socialist veneer of Guinea, Benin, and the Republic of Congo."15
At the 1990 conference on which this book is based, the late David Nicholls pointed out that whereas Weber's sultanism was a subtype of traditional authority, our cases were characterized by the decay or incomplete development of modern legal-rational authority rather than by the disappearance of all
6     "     COMPARATIVE STUDIES
1
A Theory ofSultanism 1   • 7
TABLE 1.1
Types of Patrimonial Rule
	Type of Authority	Extreme Form
Traditional form Modern form	Patrimonialism Ne op atrimonialism	Sultanism Neosultanism
remnants of traditional •authority^It4s4ndeed-lMe-lbat-fee^epraes-we-stu<iied '-relate to Weber's sultanistic regimes just as neopatrimonial regimes relate to Weber's patrimonial type of traditional authority; therefore to be precise one should refer to them as "neosultanistic" regimes. This usage would have the advantage of not only distinguishing them from the Weberian use of the term "sultanism," but also maintaining the logic of Weber's terminology; as we shall r-see, just as for Weber the transition between patrimonialism and sultanism is [, "definitely continuous,"16 neosultanistic regimes are an extreme version of neopatrimonial forms of governance. The scheme shown in table 1.1 obtains.
Also, the etymology of the term "sultanism" strikes some as "orientalist." At r. the conference, Jonathan Hartlyn suggested the term "discretionary neopatri- j monialism" to replace "sultanism." For Weber, the Near East was.the. "classical ! location of 'sultanism,'"" which is presumably why he chose a term derived ; from the Arabic language. Weber's study of the Islamic world was less thorough than his systematic explorations of Christianity, ancient Judaism, and even Confucianism, and it is not for us to pass judgment on the accuracy of his I analyses,18 more particularly on whether traditional regimes headed by a sultan j, were in fact "sultanistic" in the sense defined by Weber: recent scholarship [• tends to refute this view in the case of the Ottoman Empire.19 Our distinction between the theoretical concept and the empirical reality that led to its initial formulation is congruent with Weber's own methodology, since he wrote elsewhere that "the concept of 'Kadi-justice' has [little] to do with the actual legal L principles whereby the kadi renders legal decisions."20 Furthermore, Weber does apply the term to other societies, such as China,21 and the doyen of ^' Ottoman historians, Halil Inalcik, has applied Weber's concepts to the Ottoman | Empire without ascribing negative connotations to the term "sultanism."22 Nor does the term have an anti-Islamic tinge, since, unlike the caliphate, the sultan- |. ate was a secular office.23 Moreover, even the original pre-Ottoman caliphate, j: though endowed with a religious aura, derived in its actual organization from t nonreligious models. As Inalcik puts it: [
What Max Weber meant by sultanism was originally derived not from Islamic i precepts but from the caliphal state organization, which owed its basic philoso- | phy and structure to the Byzantine and Sassanian heritage. This Iranian state 1
tradition was transmitted to the Ottomans through native bureaucrats and the literary activity of the Iranian converts who translated Sassanian advice literature into Arabic.24
To summarize, although the etymological objection to "sultanism" seems untenable, that the term belongs to the world of traditional forms of authority poses a real problem. Our approach to the social sciences is nominalist rather than essentialist, however, and since "sultanism" has already achieved some recognition among political scientists, we retain the term even though "neosultanism" would be more accurate, so as not to add to the terminological confusion that is already too prevalent in the social sciences. Let us now turn to the definition of sultanistic rule.
Definition and Prevalence of Sultanistic Regimes
No king was despotic of right, not even in Persia; but every bold and dissembling prince who amasses money, becomes despotic in little time, voltaire, Philosophical Dictionary
The. ideaLtype of a contemporary sultanistic regime can be constructed as follows: It is based on personal rulership, but loyalty to the ruler is motivated not by his embodying or articulating an ideology, nor by a unique personal mission, nor by any charismatic qualities, but by a mixture of fear and rewards to his collaborators. The rulerexercises. his power without restraint, at his own discretion and above all unencumbered by rules or by any commitment to an ideology or value system. The finding norms and relations of bureaucratic administration are constantly subverted by arbitrary personal decisions of the ruler, which he does not feel constrained to justify m ideological terms. As a result corruption reigns supreme at all levels of society. The staff of such a ruler is constituted not by an establishment with distinctive career lines, like a bureaucratic army or a civil service, recruited based on more or less universal criteria, but largelv bv people chosen directly by the •tiler. Among them we very often find members of his family, friends, business associates, or individuals directly involved in using violence'to sustain the regime. Their position derives from their purely personal submission to the ruler, and their position of authority in society derives merely from this relation. The rulexand-his associates do not represent any class or corporate interests. Although such regimes can in many ways be modern, what characterizes them,is the weakness of traditional and legal-rational legitimation and the lack of ideological justification.-21
8     '     COMPARATIVE STUDIES
No regime fits this ideal type perfectly. Paraphrasing Weber, one might even say that although (neo)sultanism may appear to be completely unrestrained by legal-rational norms, this is never in fact the case. Regimes approximating this ideal type can be found all over the world. The regime of Rafael Leönidas Trujillo Molina in the Dominican Republic, lean-Claude Duvalier's rule in Haiti, Fulgencio Batista's dictatorship in Cuba, the rule of the Somoza family in Nicaragua, the later stages of both -Pahkvi-shahs -reign-in-Jran-,- Ferdinand Marcos's presidency after his declaration of martial law in 1972, and Manuel Noriega's dictatorship in Panama come closest, as do many of the personalistic dictatorships in sub-Saharan Africa.26
As in any typology, there are of course borderline cases that the comparativ-ist has difficulty assigning to one category or another. Alfredo Stroessner's rule in Paraguay is one such case. Although many comparative studies of Latin American military dictatorships include his regime as a case of sultanism, we decided against doing so, for our analysis convinced us that, as Alain Rouquie put it,
Paraguay has not been transformed into a Nicaraguan-style Stroessner fief-dom. . .. The Paraguayan state has a tangible existence, and the army is not the personal property of the president. It is unlikely that a Stroessner dynasty will be established. The oldest son, an aviation officer, has no military base, and the marriage of the youngest son to the daughter of a powerful general, Andres Rodriguez, did not produce the results that had been anticipated.27
But a few years after this was written, "the powerful general" overthrew his daughter's father-in-law and became president himself before handing power to an elected civilian head of state. To complicate matters further, a Paraguayan political scientist's systematic attempt to characterize "Stronismo" typologically yielded the result that Juan Linz's notion of "sultanism" fit it best, with the proviso that it should be called "neosultanism."28 We have decided to stick by our original choice, however, a decision that was confirmed by the Colorado Party's retention of power after the first free elections in Paraguay, which is unique in Latin American transitions and shows to what extent Paraguay's ruling party had an identity independent of Stroessner. A similar outcome obtained in Bulgaria, where the Bulgarian Socialist Party in 1990, alone among Eastern Europe's postcommunist parties, won the first free elections, in contrast to the muddle that resulted from the overthrow of Ceausescu in neighboring Romania.29
The African states present a similar dilemma. Personalism is present in most of them, but it comes in several varieties that Africanists designate in different ways. Whereas Richard Sandbrook found "sultanism" almost everywhere, Robert H. Jackson and Carl G. Rösberg divided personalist rulers into "prince,"
A Theory of Sultanism 1   • 9
"autocrat," "prophet," and "tyrant."30 Another team of Africanists classified Africa's regimes as administrative-hegemonial, pluralist, party-mobilizing, party-centralist, personal coercive, populist, and ambiguous.31 Michael Bratton f and Nicolas van de Willie designated almost all regimes neopatrimonial but
further divided them into personal dictatorships, plebiscitary one-party sys-terns, military oligarchies, and competitive one-party systems.32 ~f Although the "tyrants," "personal-coercive" rulers, and "neopatrimonial
personal dictatorships" do not totally coincide, by and large they correspond to what Linz called "sultanism" two decades ago. And here the clearest cases are Jean Bedel Bokassa in the Central African Republic, Mobutu Sese Seko in Zaire, Francisco Macias Nguema in Equatorial Guinea, and Idi Amin in Uganda.33 In the last two cases, however, the typical enrichment of the ruler and his family |! was not so significant, while arbitrary rule and destruction of society were.34
|- Under communism both Nicolae Ceausescu's rule in Romania and the re-
gime of Kim U-sung and his son, Kim Jong-il, in North Korea show clear stil-s tanistic tendencies. And among democracies, the rule of Eric Gairy in Grenada
■ (1974-79) and that of the Bird family in Antigua and Barbuda evince some
\ sultanistic traits.35
Yet even in the regimes we call sultanistic, elements of a legal-rational order j or of a legitimizing ideology are not totally absent. The concept of "sultanistic
■h regime" is not a genetic but an evolutionary one, in the sense that most such
[ regimes develop out of other forms of rule. Jean-Claude Duvalier owed his
p presidency to his being the son of the champipn of noirisme. Francois Duvalier
initially came to power through democratic elections in 1957, as did Ferdinand Marcos in 1965, Eric Gairy in 1967, Macias Nguema in 1968, and V. C. ("Papa") Bird Sr. in 1976. The Shah never abrogated the 1906 constitution. Batista in J the 1950s ruled under an amended constitution and tolerated a Congress and
1 courts. Ceausescu came to power through the normal functioning of the Ro-
I manian Communist Party. And yet other political regimes that also had des-
| potic and arbitrary dimensions did not develop into sultanism.
I As is the case with all ideal-typical concepts, no empirical reality fully
I matches all characteristics of a sultanistic regime. It could be argued that it
would be preferable to talk about "sjtkanjsiictandend^ ► ment in the.direction Qj._t.he ideal^typical sultanisticregime.36 We could then
speak of sultanistic tendencies that can occur in different kinds of regimes, can be stronger or weaker at different moments, and can coexist with other aspects J and dimensions of governance. For simplicity, however, we will use the term
I "sultanistic regime" most of the time.
I A regime in which some sultanistic tendencies are present, but where the
1 circle of clients is wider and the discretion of the ruler less extensive, should be
10     •     COMPARATIVE STUDIES
A Theory ofSultanism i   • n
called neopatrimonial. Personalist rulers whose regimes penetrate society by means of a political party, such as the Colorado Party in General Alfredo Stroessner's Paraguay, or who represent a certain segment in society, like Francois Duvalier, who based his rule on the black middle classes in opposition to Haiti's mulatto elite, do not fit the pure type of sultanistic regime. Of course if the circle of people included in the rulers patron-client relationships narrows, the regime can hecome.sultanistk:-such-was-tfte-trans!tion-frorn-Francois Duvalier s rule in Haiti to Jean-Claude Duvalier's. Let us now turn to a detailed analysis of this type of regime.
Characteristics of Sultanistic Regimes
In the course of this analysis we will illuminate these regimes' distinctiveness by constantly referring to other types of nondemocratic regimes around the following themes: the state over which they preside, their personalism, their constitutional hypocrisy, their social base, and their political economy.
The Blurring of the Line between Regime and State
By "regime" we mean the patterns of allocation, use, and abuse of power in a polity. This encompasses more than the political institutions in a democracy and less than the comprehensive structures of domination in totalitarian systems. Robert Fishman has noted that for understanding authoritarian regimes (and more specifically transitions to democracy) it is useful to distinguish between regime and state.37 Under authoritarian regimes the more limited politicization of society makes it possible to discriminate between those who hold political power in the government, in the party, or in the military and those who exercise functions normally associated with a modern state in a ; professional bureaucracy, in the armed forces, or in the judiciary. Therefore the state apparatus is likely to persist with only limited modification in many £ regime changes, except when they are revolutionary.38
Under sultanistic rule the distinction between regime and state is much more blurred, and in very advanced cases of sultanism one can even speak of a fusion between regime.and state.39 Bureaucratic structures persist and may even be streamlined and rendered more efficient, especially at the local level where extractive opportunities are not always present, but the ruler and his associates directly intervene in the structures of governance, disregarding their, internal norms, professional standards, and ethos. As we shall see, this renders the "state structures" less serviceable after a regime change.
Trujillo, Reza Shah, Somoza Garcia, and Mobutu all began by strengthening
the state after a period of institutional decay; furthermore, all subdued regional challenges to the hegemony of the center over the periphery in the early phases of their rule. It is this initial centralization of posver, accompanied as it is by a strengthening of the military and other state institutions, that makes sultanistic control over society possible.
As the founders' power increases, sultanistic tendencies appear. Beyond the ruler himself, his immediate relatives, and his close associates or "cronies," the sultanistic state is characterized by an absence or perversion of legal-rational norms that is unrelated to an ideological project, and by rampant corruption and venality. In the state administration and the army such features of modern bureaucracies as areas of official jurisdiction, an office hierarchy with channels of appeal that stipulate a clearly established system of super- and subordination, the separation of official activity from the sphere of private life, and finally the management of offices according to general rules may exist on paper and, depending on the country's level of development, even in reality, but they are under constant attack by the sultanistic practice of regulating all relationships through individual privileges and bestowals of favor.40
At some point in their autocratic rule, the dictators often bring technocrats into the regime—both to please their American patrons and to propose to the country an apolitical and competent alternative to the political chaos that would ensue in the absence of the ruler. In Iran the last shah brought in a number of well-trained bureaucrats after the White Revolution of 1963, in the course of which he eliminated big landowners as the politically dominant social class. In what was then Congo-Kinshasa, Mobutu very early "established numerous public and semipublic agencies headed by young technocrats and former politicians concerned with the Congo's social and economic development" and relied increasingly "on the talent and expertise of such social groups as university and technical school students and former civil servants."41 In the Philippines Marcos "brought a corps of technocrats into his government," announced in mid-1981 that he would "sit back and let the technocrats run things," and favorably impressed both U.S. officials and International Monetary Fund-World Bank officials.'12 In fact, however, the
technocrats were given the prerogative to formulate and rhetorize the public agenda in the form of economic and development plans which formed the basis for foreign loans. The political leadership then allowed the unconstrained introduction of exceptions that made complete mockery of the spirit and letter of the plans.43
In Haiti, Jean-Claude Duvalier "fired his ministers and brought in a number of well-regarded young technocrats to clean up the government and impress the United States," but within "a short time ... he had dismissed a number of
12    •     COMPARATIVE STUDIES
A Theory of Sultanism i   • 13
them, some due to his wife's objections, and others because they actually intended to fulfill the assignment."44
Clearly, sultanistic rule is not incompatible with a certain rationalization of .the,administration, as long as this rationalization enhances the ability of the ruler and his cronies to extract resources from society. If the technocrats try to resist the regime's dominant ethos, however, they are marginalized, for even when an official is M0t'a,pgrs0"nal,,dep"eirdentn3ftfre^uieiTlhirTuler demands compliance. In his view the official's loyalty is not an impersonal commitment to impersonal tasks that define the extent and content of his office, but rather a servant's loyalty based on a strictly personal relationship to the ruler and an obligation that in principle permits no limitation.
These officials enjoy little security: they are promoted and dismissed at will and enjoy no independent status. In Haiti, President Jean-Claude Duvalier "often dismissed and replaced cabinet members on the advice of his wife or counsel of some trusted minister. These shake-ups occurred twice a year on average ... and ministers operated in a climate of insecurity and paranoia."45 In extreme cases, ministers may even be subject to dishonor and persecution one day and return to the graces of the ruler next, as exemplified by the Zairian politician Nguza Karl-I-Bond, who was foreign minister, broke with Mobutu, went into Belgian exile, was condemned to death, and later repented, returned, and became prime minister. It bears emphasizing again that most modern sultanistic states do have pockets of bureaucratic organization, but the more f these come under attack, the more we can say that the sultanization of the { regime is advanced.
Of particular importance is the organization of the armed forces and their relation to the rest of the state apparatus. Sultanistic rulers typically maintain • an atmosphere of distrust among various branches of the military and encour- | age mutual espionage to protect themselves from a military coup; some even i use private militias to sustain their control of society, such as the Duvaliers' [ Tonton Macoutes or Eric Gairy's "mongoose men." Intricate structures are | maintained to make each branch help control the others. Often there is no ; unified command, and the commanders of each force report directly to the i rulers. The normal criteria for promotion are disregarded or subverted, creat- ' ing resentment. The armed forces are thus deprofessionalized as the rulers aim \ at converting them into their private instruments for power.46 In the Philippines Marcos made his driver, Fabian Ver, chief of staff, and in the Central African Republic Bokassa named Sublieutenant Bozize air force general after I the man slapped a Frenchman who had treated Bokassa disrespectfully.47 As the rulers' men in the military gain access to patronage, the gap between them and I
their less fortunate colleagues grows. In Iran, the Shah's multibillion-dollar arms purchases in the 1970s enriched those generals who dealt with procurements and thereby alienated many others.48
In discussing the armed forces of sultanisticregimes, itjsjjseful to distinguish analytically between their autonomy and their professionalism. The Iranian armv was highly professional, but at the same time,.closely related-to the Shah. Likewise, Nicaragua's National Guard was loyal to Somoza and fought well in the civil war. In a country like Paraguay, it is perhaps the armed forces' relative professionalism, the outcome of the many wars the country was forced to fight against its neighbors, that has prevented the state's neopatrimonialism from degenerating into outright sultanism. The position of the armed forces vis-a-vis the ruler is crucial when the regime undergoes a crisis and may indeed determine the outcome, a theme analyzed in detail in Richard Snyder's contribution to this volume.
Sd^isdgjrjJ^jqmetinies create single parties: When they do, these par-tieslyerxnames bespeak the rulers' efforts to make them look like "revolutionary" movements, destined to be the link between the leader and the people he "guides toward new horizons: Kilusang Bagong Lipunan (New Society Movement) in the Philippines, Mouvement Populaire de la Revolution in Zaire, Comite National d'Action Jean-Claudiste in Haiti, and Rastakhiz (Resurgence) in Iran. In the Central African Republic Bokassa used the preexisting Mouvement pour FEvolution Sociale en Afrique Noire (MESAN) for this purpose. Yet in reality the revolutionary quality of these parties is like the charisma of its leaders—mosdy declarative. Nor do they fulfill the functions assigned to parties in established single-party states: the legitimation of the political system, political recruitment, and policy formulation.49 The nature of the sultanistic regime militates against such a role.
Personalism
Paradoxically enough, whether or not their official position in the state corresponds to their actual power, sultanistic leaders do not conceal the highly personalistic.nature.oftheirriile. Outwardly this personalismhas-two facets: a pronounced cult., of personality around the leader and, a tendency, toward dynasticism; ■■■
Sultanistic leaders crave charisma and surround themselves with the trappings of charismatic leadership precisely because they know they lack it.50 They invent new titles for themselves: Trujilio called himself generalisimo and benefactor de la patria; the Shah became Aryamehr ("Light of the Aryans"). Mobutu
14     '     COMPARATIVE STUDIES
changed his name from Joseph-Desire Mobutu to Mobutu Sese Seko Kuku
Ngbendu Wa Za Banga, meaning "The all-powerful warrior who will go from f~
conquest to conquest, leaving fire in his wake," and he was called, among other
things, "savior."5' The dictator of the Central African Republic assumed the title
of emperor, and Ceaujescu revived General Ion Antonescu's title of conducator
in addition to being described as "architect of world peace" and "hero among
the nation's heroes."52 Amin styled himself-field-marshal-and-eBE(conqueror
of the British Empire), and Macias Nguema topped all others by claiming f
divinity. f
Sultanistic rulers also like to be thought of as great thinkers and fill many f beautifully bound volumes with their speeches, declarations, and proclamations in addition to their numerous (mostjy ghostwritten) books. Frequently, \' feeling the need to-legitimize their.regime ideologically, they come up with an i ideology that, reflecting the regime's personalism, often bears the ruler's name. In Haiti "Jean-Claudism" became official state ideology in 1978, Ferdinand : Marcos wrote a book called An Ideology for Filipinos,™ the Shah published a [■ book under his name titled The White Revolution, which was taught as a subject \ in Iranian high schools in ninth and eleventh grades, and Mobutu's works were |_ collected in Les grands textes du Mobutisme,'* Mobutuism having become offi- ^ cial state ideology in 1974. The implementation of this ideology is sometimes ' accompanied by a "revolution," such as the Shah's "White Revolution" of 1963 [ (later renamed "Revolution of the Shah and the People") or Marcos's "Revolu- f tion from the Center,"55 but the main function of these revolutions seems to be r eliminating political opponents.
The sultanist ideology often exalts the nation's, ancient glories and draws, on t an "invented tradition"56 to demarcate the nation from its neighbors ethnically and even racially. The Pahlavi shahs' implicitly anti-Arab emphasis on Iran's an- r cient Persian heritage, discussed by Homa Katouzian, Trujillo's anti-Haitianism, discussed by Jonathan Hartlyn, Francois Duvalier's noirisme, Ceau§escu's insis- ■ tence that Romanians are descendants of the ancient Dacians,57 and Mobutu's | cult of African "authenticity" are cases in point. One who truly believed in this return to the roots was Macias Nguema, who purged his country of all Western influence, even modern medicine.58
However, the function of this ideology is different from.that.of .totalitarian , systems, where "leaders, individuals or groups... derive much of their sense of mission, their legitimation, and often very specific policies from their commitment to some holistic conception of man and society."55 The purer a regime's sultanism, the more its ideology is likely to be mere window dressing,.elaborated after the onset of the ruler's regime to justify it. In Iran, Reza Shah was to fa
A Theory of Sultanism 1   ■ 15
some extent motivated by an ideology of Persian nationalism, and in Haiti Francois Duvalier was inspired by the old tradition of noirisme and used it to his advantage.60 And in Romania, which, unlike all other East European states except Albania, never experienced a transition to posttotalitarianism, Ceau^es-cu's rule was an aberration of the Marxism-Leninism prevalent in that part of the world at the time and thus a survival of Stalinism. It is reminiscent of Stalin's personality cult that the entry on Ceau^escu in the Romanian encyclopedia was longer than those on Marx, Engels, and Lenin combined.61
The cult of the.,pfirsonality.alsg Jeayes. .gepgraphictraces, since sultanistic rulers like to name cities and even najural feature^
and AfKalibecame Reza'ieh and Bandar Pahlaviin honor of Reza Shah, Santo Domingo was renamed Ciudad Trujillo, and in Haiti the town of Cabaret became Duvalierville. In Africa Lake Albert and Lake Edward were renamed Lake Mobutu Sese Seko and Lake Idi Amin, and the island of Fernando Poo became Macias Nguema Island. Only in Nicaragua did the dictator have the modesty to give his name just to a new port established on the Atlantic coast: it became Puerto Somoza (and was renamed Puerto Cabezas after his ouster).
The second aspect of personalism is the prominent role of family members in these regimes,, .which has led many analysts to speak of "dynasties."62 In Haiti and Nicaragua sons "inherited" the presidency from their fathers, and in Nicaragua and the Dominican Republic family members occupied important commands in the armed forces: in the latter the dictator's son Ramfis was a brigadier general at the tender age of nine. Wives play an important role too. In Haiti both first ladies of the Duvalier regime wielded considerable influence and were involved in corruption.63 In the Philippines, Imelda Marcos held a number of cabinet-rank positions and was named mayor of Metro-Manila, while her husband's sister was governor of the Marcoses' native Ilocos Norte province, with her nephew Ferdinand Marcos Jr. as vice-governor at age twenty-one.64 In Romania, Ceaus;escu's accession to the newly created post of "president of the republic" in 1974 was marked by ceremonies closely patterned after coronations.65 His wife was the Communist Party's second in command by the early 1980s, his four brothers all held key levers of power, and their son Nicu was groomed to succeed his father until his constant brawling, gambling, and philandering turned party leaders (including other relatives) against the succession.60 In Iran, of course, the old monarchical tradition was there for the Shah and his relatives to use. When he celebrated twenty-five hundred years of that tradition in.1971, one of his guests, Imelda Marcos, was so impressed that she is said to have suggested to her husband that they become emperor and empress of the Philippines (see chapter 9 below). In Equatorial Guinea, the
COMPARATIVE STUDIES
dictator filled all major positions with members of his Esengui clan. Although he himself was ousted and executed by his nephew, his dynasty endures to this day and governs in a manner only slightly less repressive than Macias Nguema's rule.67 In Antigua and Barbuda, Vere Bird's two eldest sons, Vere Jr. and Lester, were cabinet members, and in March 1994 the younger of the two, Lester Bird, succeeded his father by leading the Antigua Labor Party to yet another election victory. He did not include hisolder-brether-in-the-cabmetrperhaps because a 1990 inquiry into his alleged involvement in transporting Israeli arms to Colombian drug traffickers had recommended that he be barred from public office.68 The only ruler who went so far as to abolish the republic and found a monarchy was Jean Bedel Bokassa, who proclaimed himself emperor in 1976 and proceeded to crown himself in 1977. The idea originated with Mobutu, who had planned to declare himself "emperor of the Bantus" but was beaten to it by Bokassa.69
It is the combination of personalism and dynasticism that is specific to sultanism. Other personalistic regimes, such as those of Stalin and Tito, did not elevate the rulers' relatives to political prominence, and dynasties can also appear in nonsultanistic settings. In such relatively stable democracies as Sri Lanka, India, and Jamaica, successive members of the Bandaranaike-Kumara-tunga, Nehru-Gandhi, and Manley families have been democratically elected heads of government; in Pakistan, Bangladesh, and Burma daughters or widows of politicians have been the people's choice to replace authoritarian regimes. The difference is that in most cases they were not designated by their predecessors but inherited their charisma.70 But in Texas and Alabama state governors have had their wives run for the governorship when they could not, practicing in effect the "politics of understudy" described below.71
The reason for the dynasticism of sultanistic regimes is perhaps that the rulers feel that they can trust only their kith and kin. Most often the sultanistic rulers come from humble origins and are looked down upon by the traditional elite. The first Somoza, Anastasio Somoza Garcia, had spent some time in the United States as a used-car salesman.72 The first Pahlavi shah began his military career as a soldier in the Russian-officered Cossack Brigade. Trujillo grew up in very poor circumstances and as a young man was employed as security chief for a sugar mill.73 Batista was a mulatto who had worked as a sergeant stenographer in the Cuban army, after stints as a cane cutter, carpenter, and railway worker.74 Ceausescu was of peasant origin, which, though not astonishing for a communist leader, contrasted with other Romanian communist leaders like Ana Pauker, who belonged to the intelligentsia. Finally, Amin, Bokassa, and Macias
A Theory of Sultanism 1   • 17
Nguema "all possessed low traditional status," and "their ethnic and socioeconomic backgrounds directly conditioned their behavior while in power."75
Given their lowly origins and the tenuous social base of their regimes, sultanistic rulers often attempt to create family alliances with the old elite, so as to co-opt at least part of it. Reza Shah thus took a Qajar princess as his second wife, •fhe first Somoza married into the powerful Debayle-Sacasa family, Batista divorced his first wife and married "up," and Jean-Claude Duvalier chose his wife from the light-skinned Bennett family, which, though not part of the old mulatto elite his father had eliminated from political power, was nonetheless rich.
The importance of family members is not limited to the political realm and often carries over into economic life, as different family members carve out bailiwicks for themselves. It is even possible that pressures toward sultanization of the regime come from the family of an authoritarian ruler rather than from the ruler himself.
Constitutional Hypocrisy
Rulers who destroy men's freedom commonly begin by trying to retain its forms. alexisde tocqueville,
Ttte Old Regime and the French Revolution
Patrimonial rulers who claim traditional authority preside over political systems based on historically enacted rules, above all customary law, but they generally have ignored modern constitutions Sultanistic jegimes lacking in ideological basis for their institutions, often govern with, -constitutions mher-ife3Tr^"a^re^ns'aemxs^2tic regime or enacted to give a legitimate appearance to their rule. However, the sultanistic ruler does not necessarily oc-cupy""the"pbsifidfi that is constitutionally the most powerful, a policy called politique de doublure, or "politics of understudy" in Haiti. The Somozas occasionally put a puppet in charge of the presidency; in the course of his long rule Trujillo left it to four different men, one of whom was his brother; in Iran the Pahlavi shahs always retained prime ministers who, according to the constitution, were in charge of governing the country; and in Panama Manuel Noriega continued the habit of his predecessor, General Omar Torrijos Herrera, of ruling as head of the military with a nominal president as head of state. In the African cases this practice does not obtain, except to some extent in the "imperial" phase of Bokassa's rule, when the constitution designated the prime minister as the effective ruler of the Central African Empire.
18     *     COMPARATIVE STUDIES
The constitutional facade of sultanistic regimes means that they pay hp service to constitutions that provide for elected chief executives and parliaments, and in some cases even multiparty systems. The leaders often make a point of extolling democracy in their country while redefining it.77 This made it easier for conservative anticommunists in the United States and other Western countries to back them. The sultanistic rulers therefore organized elections and maintained a nominally multiparty polity,-but-somehow the government always won.78 Furthermore, the defeated "minority" was not granted a secure status or an institutional role of any sort. The civil liberties of the opposition were severely restricted and arbitrarily violated.
The manipulation of the electoral process can take various forms. Trujillo held elections throughout his rule. In Nicaragua, the Somoza family gained control of the Liberal Part)' in the 1930s and from then on would split the opposition Conservative Party before each scheduled election by making deals with some of its leaders, then prevent the electoral success of the others bv fraud.79 Nonetheless, new opposition parties were allowed to operate in the 1970s. In Cuba, Batista held presidential elections in 1954, but under conditions that, in the end, left him as the only candidate. In the Philippines, Marcos encouraged (and perhaps paid) a candidate to ran against him in the 1981 presidential elections but let him carry only his hometown. Even about Romania, which was by definition a "people's democracy" and therefore did not pretend to have competitive elections, one observer wrote that "in no other East European regime [was] the chasm between rhetoric and reality so painfully experienced as in Ceau§escu's Romania."80
The most extreme.situation obtained in Iran, where until 1975 the Shah maintained the fiction of a two-party system. The opposition Mardom, or "People's" Party existed on paper only;' one can callTtT"pseudo-opposition."81 In its decorative function this pseudo-opposition bore some superficial resemblance to the bloc parties of communist Eastern Europe,82 in that it too lacked any social base independent of the regime and was controlled by it. In both cases party leaders were in and of the regime rather than oppositionists who had been co-opted. However, the Shah's loyal pseudo-opposition differed from the bloc parties in that unlike these it officially competed with the Qilingjran-e Novin, or New Iran-Party:83 In 1974 it briefly seemed as though the Mardom Party could play an autonomous role.when it.was allowed to campaign freely in a by-election, but although it won handily, the results were rigged to deny the party the seat. Soon afterward the Shah instituted a one-party system.
Sultanistic rulers also often turn to plebiscites to prove their democratic legitimacy; needless to say, they never lose one. Plebiscites are part of the
A Theory of Sultanism 1   • 19
democratic facade that sultanistic rulers like to erect, but they also create the image of a charismatic leader who rules by popular acclamation. In the communist world plebiscites were not customary, but Ceau^escu held one, on the pretext of wanting to legitimize a reduction in his country's military spending.84 As Tocqueville put it, these leaders "cherish the illusion that they can combine the prerogatives of absolute power with the moral authority that comes from popular assent. Almost all have failed in this endeavor and learned to their cost that it is impossible to keep up such appearances for long when there is no reality behind them."85 The longtime manipulation of the constitutional arrangements and the cynical acting out of electoral procedures have nefarious consequences for a hypothetical transition to democracy, as we shall see later,
The Narrow Social Bases of Sultanism
Very <^fteri_$.uljumi^cj.nler^r4gndly come to power withjhs .support of clearly recqgnizaljla.gro.ups—sometimes even, as in the case of Marcos, through fair elections. In Ira^IJ^.SlJ^^^d^sg^eofpcnKej.in the.1920s.had the support of a sizable segment of politically articulate-Iranians, including the Left.86 In Romania, Ceausescu came to power through the workings of the Communist Party and enjoyed genuine popularity in the 1960s as a de-Stalinizer who repudiated the 1968 Soviet invasion of Czechoslovakia and championed Romanian independence. In the Philippines, finally, Marcos's imposition of martial law in 1972 was greeted with relief by significant sectors of the Filipino population. In Central America regimes that later degenerate into sultanism often start with a certain antioligarchic dimension. The Somozas and Batista had some backing from the hitherto somewhat excluded mestizos and mulattoes, respectively, while Francois Duvalier and Eric Gairy in the beginning championed the cause of the blacks in relation to their countries' socially dominant mulattoes. In Panama, finally, Omar Torrijos's dictatorship, the forerunner of Noriega's sultanistic rule, improved the lot of the blacks and mestizos.
As the regimes become s^anistic, however, they lose much of their initial sodal^suppijrt and begin to relyincreasingly-onamixture of fear and. rewards. The beginning of the sultanistic phase can usually be dated .w.ith..the.benefit of hindsight. In Iran, Reza.§halis dismissal of mo.stpf.his.early modernist collaborators in 1933, and MohammadJteza.Shah's "White Revolution"_Qf, 19.63 heralded the two shahs' suhanistic,perlod,,In the Philippines the declaration of martial law in 1972 was a crucial turning point.
The ability of sultanistic rulers to stay in power depends on their freedom
20     •     COMPARATIVE STUDIES
from the need to forge alliances with civil society and to build coalitions. This freedom increases, as we will see later, if they can monopolize certain economic resources. Given the weakness of the regime's links with civil society, crucial social strata that might support a capitalist authoritarian regime because of its pro-business and anticommunist policies are alienated from the sultanistic regime. In Romania, by subverting the hierarchy of the Communist Party Ceaus^scu alienated the'OTMErfWmwa.'m^
party apparatus that ousted him: Ceau§escu's successor, President Ion Uiescu, had once been the old dictator's protege but was later purged for "liberalism" and "intellectualism." In the end the social bases of a sultanistic regime are restricted to its clients: family members of the rulers and their cronies. For these, however, loyalty to the ruler derives only from their own interests rather than from any impersonal principles, for which reason they do not constitute a distinct stratum, group, or social class. Nonetheless they do exist, and through their own clients they create a multiplier effect that links enough people to the regime so that it can function. Although a lot of people may benefit from a sultanistic regime, therefore, they do not really constitute a broad social base.
There are, of course, exceptions. A dictator may enjoy genuine support in his home region or from his own ethnic segment because he favors it. Marcos favored the Ilocos Norte province in the Philippines, and in Uganda Idi Amin's fellow northerners benefited, relatively speaking, from his rule.87 Occasionally the ruler will combine his rule through patrimonial officials with populist gestures: this seems to have been the pattern of the first Duvalier, who exploited the tensions between Haiti's blacks and mulattoes.
A puzzling feature of some sultanistic regimes is the success of public figures identified with them in free elections after the overthrow of the regime. Joaquin Balaguer's repeated victories in the Dominican Republic's presidential elections, even though not always obtained through honest means, testify to a certain popularity. In Nicaragua, Arnoldo Aleman handsomely, won the presidential election of 1996, beating Daniel Ortega and relegating the candidate of the center to a distant third place. It is important to remember that these victories are not belated vindications of the sultanistic rules of Trujillo and Somoza: they reflect the candidates' personal popularity and do not mean that most of the population wants to return to the old days, just as the victories of former communists in Eastern Europe do not mean that Eastern Europeans yearn for a return to pre-Gorbachev communism. In all these cases public figures associated with the old regime run for office under the rules of the new regime, and their electoral platforms do not include the reestablishment of the old regime.
J A Theory of Sultanism 1   • 21
i It is often alleged that sultanistic Leaders substitute superpower patronage
I for a dome.stic power, base. The close association of most sultanistic leaders
l with the United. States .is beyond doubt, but one should not deduce from this
{ that they are American puppets: at times, during more populist interludes, they
I may even temporarily assume anti-American positions. These dictators are
I adroit at making friends in the United States, and they lavish important sums
} on lobbying. Often congressmen and even senators become their defenders in
f the United States; it is therefore by no means clear who manipulates whom to
I whose greater advantage. The recently published diaries of the Shah's longtime
^ minister of court and boon companion, Amir Asadollah Alam, make it quite
' clear that not only did the Shah not take any orders from the United States in
* his heyday, he even maintained a network of powerful supporters in Wash-
■ ington's inner circle to influence U.S. policy.88 Likewise, Marcos "knew that ! neocolonial manipulation can be a two-way street," and both he and his wife ; maneuvered adroitly in Washington circles.89 Somoza too was able to mobilize
influential friends in Washington,90 and before them Trujillo had retained the services of lobbyists in the U.S. capital.
Distorted Capitalism
I A prince must abstain from the property of others; because
i men sooner forget the death of their father than the loss of
ri their patrimony. niccol6 machiavelli,
1 The Prince
< Although some sultanistic regimes have presided over considerable economic
grawtht.in.theJong run the personalism and corruption of the political system i negatively affects economic .development. To some extent these distortions are
i adumbrated by Weber's analysis of the economics of patrimonialism. Under
1 patrimonialism, he wrote, "there is wide scope for actual arbitrariness and the
1 expression of purely personal whims on the part of the ruler and the members
of his administrative staff. The opening for bribery and corruption ... tends to be a matter which is settled from case to case with every individual official and I thus highly variable." This means that "two bases of the rationalization of
■ activity are entirely lacking; namely, a basis for the calculability of obligations i          and of the extent of freedom which will be allowed to private enterprise." Given ;          the absence of procedural predictability, "under the dominance of a patrimonial regime only certain types of capitalism are able to develop," and he lists capitalist trading, capitalist tax farming, capitalist provision of supplies for the
22     a     COMPARATIVE STUDIES
A Theory of Sultanism 1   • 23
state, and capitalist plantations and other colonial enterprises. But a patrimonial system can also be conducive to the establishment of monopolies," an option contemporary sultanist rulers have often taken.
True patrimonial systems are limited by tradition, whereas in modern sultanistic ones the scope for arbitrariness is much wider, while the greater development of the state apparatus means that bureaucratization is more advanced. The personalistic use of power fer-the-essentially private ends of the ruler and his collaborators means that the country is run like a huge domain The boundaries between the public treasury and the private wealth of the ruler become blurred. He and his collaborators, with his consent, freely appropriate public funds, establish profit-oriented monopolies, and demand gifts and payoffs from business for which no public accounting is given. Often the privileged are thus exploited, as landowners, merchants, and foreign capitalists buy their peace by making payments to the regime.
TJiejulers enterprises contract with the state, and the ruler often shows generosity to his followers and to his subjects in a particularistic way. His family most often shares in the spoils. The economy is subject to consiSeratiKg^rn-ment interference, but this interference is rarely tied to any grand project of sustainable accumulation: the main aim is to extract resources. This "klepto-,cratic.stat?"!i°P?Htes bycontrol over essentk monopoly ownership
of critical commodities, kickbacks on contracl^^dplainiranfeMtionSpfqj-£rty..This monopoly ownership can take the form of state-owned industries (such as oil in Iran, copper in Zaire, or diamonds in me Central African Republic/Empire) whose revenues can be to a smaller (Iran) or larger (Africa) extent appropriated by the rulers. Or rulers can set up monopolies as private persons. Examples of the latter include Batistas lottery scheme or Empress Catherine's monopoly on the sale of school uniforms in the Central African Empire, uniforms whose manufacture was a monopoly of one of Bokassa's mistresses.'3
The constant patrimonial interference in economic life leads to distorted market economies that, though embedded in the sphere of capitalism, cannot truly be called capitalistic. Whereas capitalism is based on the "sacredness of private property" and on the working of the market in allocating resources, opportunities, and profits, sultanism often involves arbitrary takeover of property, directly or indirectly, by coercion and without compensation. Also, individual entrepreneurs' opportunities in the market are distorted by the ruler's intervention in economic life: he has the power to deny access, he can allocate public funds to private enterprises linked with himself or with his cronies, and he can eliminate from the market competitors not ready to abide by his de-
i rnands. Mafia economics is not market economics. Nonetheless, to maximize
I the benefits the regime derives from the economy, a certain rationalization of
f activity can take place in the more economically advanced countries, leading to
the appearance of pockets of rationality. But the absence of secure property rights inhibits long-term productive investment, since sultanistic regimes are I unable to develop the institutions necessary for a dynamic and complex econ-
r omy. Given the personalism that pervades institutions, the institutional and
legal framework necessary for resolving conflicts and enforcing contracts cannot be set up.
! This said, there are periods in the life span of a sultanistic regime when it
■ may preside over quite respectable growth rates. Given how it is managed, 1 however, this growth tends not to add to the ruler's legitimacy. Moreover, given > that sultanism subverts accountability and predictability, it thwarts sustainable '-, accumulation.
The characteristics of sultanistic regimes can further be grasped if we com-| pare them with the other types of nondemocratic systems.
; Sultanism and Other Types of Nondemocratic Rule
Since so many of our cases are in Latin America, it is important to define the difference between sultanistic rulers and the traditional caudillos of that continent. Caudillos were a product of the nineteenth century. In the wake of independence, as state authority disintegrated, military leaders based in the rural areas seized power at the head of armed bands of loyal followers and thus filled the vacuum of authority.94 Some became state builders: Uruguay owes its existence to General Jose Gervasio Artigas. Unlike modern sultanistic rulers, whose
1 power derives from existing structures at the center of the polity, the caudillos
had a more local base, possessed genuine leadership qualities, and at least initially, inspired personal loyalty.95 Although there is some resemblance between caudillism and sultanism, and though such sultanistic dictators as Rafael
[ Trujillo and the first Somoza perhaps evinced some continuity with the caudi-
i llos of an earlier age, the typical sultanistic ruler does not partake of the heroic
world of the traditional caudillo; the sort of mythology and folklore that caudillism begot is unlikely to be engendered by sultanism.
Sultanism differs from totalitarianism in that, like authoritarianism, it lacks
S a genuine ideology, articulated by pro-regime intellectuals, to legitimizeit and
guide its policy formulation, Romania and North Korea, where Ceausescu and
■ Kim Il-sung did take Marxism-Leninism seriously, are obvious exceptions here, but in the Dominican Republic too, ideologues such as Joaquin Balaguer elabo-
24     •     COMPARATIVE STUDIES
A Theory ofSultanism 1   ' 25
rated a detailed ideology that claimed to guide Dominicans toward democracy but was honest enough not to call the regime democratic.96 As we have seen, some sultanistic rulers produce (or have ghostwriters produce) something they like to call "ideology." But we have only to think of the appeal totalitarian ideologies have for intellectuals not subject to their rule, for young people, and for students to perceive the absence of anything remotely similar under sultan-ism. No one not subject to-their rule, not-even most-of-thcirsupportcrs and probably not even the rulers themselves, takes these ideological efforts seriously: they are pseudo-ideologies.
Another major difference from totalitarian regimes is the fusion between the private and the public roles of the ruler and the lack of commitment to impersonal purposes. Totalitarian dictators such as Stalin, Hitler, Mao, and Castro believe in meir.qwil mission, and so do their followers. ThereJs a purpose to :heir ru e other than personal enr\hr:c!it. and for this cause they succeed in mobilizing intellectuals not only in their own societies but also outside. In addition, they cultivate an ascetic image (which may or may not reflect their true nature) that is quite at variance with the undisguised hedonism of most (but not all) sultanistic rulers and especially their relatives and cronies.
„A Ihipl.differfince with. totalitarian regimes, and a consequence of the previous two, is the absence not only of a single party but also of the ancillary organizations such.as women's groups and youth groups that were so essential toj^zisrn andcommunism. Related to this is the absence of continuous political mobilization for a variety of tasks that provide a sense of participation in social and political life. As inautfioritarian systems, passivity and apathy characterize sultanistic regimes, which offer few, if any, channels for participation, even to their supporters.
TV final rfiffftiyn^fairn^atitarianiCTn is that sultanistic regimes penetrate their societies very unevenly. Those areas ofpubliC Mfe'lKaTtESten the ruling group's extraction ofresoufeeX such as the press, may be cqntroUedto a degree approximating totalitarianism, while others are left alone. Somoza's acquiescence to de facto autonomy for the Indians of the Atlantic coast is an example,97 Even in tightly centralized Pahlavi Iran, the state's penetration of distant Baluchistan province was less thorough than elsewhere.
As in all regimes, the specific traits of a sultanistic regime are most developed at the top, elite level. In a modern society, the post office and certain administrations like the judiciary (political cases excepted) will work in similar ways for most people no matter what the political regime. In rural areas village life may go on relatively untouched by the changes at the center.
Sultanistic regimes differ from authoritarian regimes in a number of ways.
* Authoritarian regimes are more institutionalized, and the limited political,, and
I "even more important, social pluralism they tolerate creates a variety of struc-
| tures that support the regime, which recruits its elites from them. The much
I simpler clientelist structures, the absence of predictable paths of career ad-
f vancement in the bureaucracy and the army, the arbitrary recruitment by the
; ruler of his lieutenants thus contrast with authoritarian regimes, although in
'■ this case the contrast is less sharp than with totalitarianism.
I The second,difference is the absence of the rule of law, be it a repressive one:
'• sultanistic regimes constantly violate their own norms. Moreover, in line with
I the constitutional facade they maintain, sulLanistic,regimes,.often privatize re-
; pression, in the sense that it is carried out by informal groups in the service of
the ruler as well as by formally, constituted state agencies implementing explicitly repressive legislation.98 Ujjd.ex^Jtam^ ofpeople's basic human rights, is arbitrary rand may strike supporters and opponents : S          of the regime alike, giving rise to pervasivej£ar_.and. suspicion in society.99 In !          any event, sultanistic regimes can exercise degrees of repression ranging from ;«          the relatively benign reign of Batista100 to the demented paranoia of Macias Nguema, who is estimated to have been responsible for a minimum of fifty ■ l          thousand deaths, drove a third of his country's population into exile, and eliminated its intelligentsia on a scale comparable only to the Khmer Rouge.101 I          Regimes that act illegally according to their own laws create uncertainty and : I           unpredictability in public life.This capriciousness has an adverse effect on eco-:          nomic development, among other things, since entrepreneurs who lack con-t          nections to the ruler cannot plan ahead rationally, a point we will come back to.
Oppositional, activity: against .sultanistic regimes otter, coikvir.r.uos abmac. 5 as sultanistic domination drives many citizens, especially intellectuals, into
exile. These exiles can be a fertile base for oppositional undertakings, since they i deeply resent not only the corruption and repression at home, but also the
ideological vacuum behind it. Intellectuals are more likely to excuse repression when it is carried out in the name of a transformational ideology ("You can't make an omelette without breaking eggs") than when it is exercised for private gain. This resentment is heightened by the tendency in many sultanistic regimes to expand the repressive activities of the state abroad: opponents in exile : are often kidnapped or killed. The existence of an opposition in exile compli-
cates political life in the event of an overthrow of the sultanistic regime.102
f
2
A Theory ofSultanism 2
Genesis and Demise of Sultanistic Regimes
H. E. Chehabi and Juan J. Linz
The Genesis ofSultanism
Sultanistic regimes have been relatively few in number, and almost all have by now disappeared. However, the appearance of new sultanism cannot be excluded. Sultanistic regimes are unlikely to be established in advanced industrial soaetiesjjyet an underdeveloped economy is not a sufficient precondition for . the emergence of sultanism, The factors favoring the emergence of sultanism are both macrostructural and institutional, but these variables do not explain everything: the presence or absence in a given society of individuals who are willing to become sultanistic rulers matters as well.1
Macrostructural Factors
Two sets of factors stand out: economic conditions and the international environment.
Socioeconomic Conditions
About patrimonial structures Weber wrote that "little can be said about purely economic preconditions for [their] rise.'" Matters are slightly different for modern sultanistic regimes: although there are no necessary and sufficient
A Theory of Sultanism 2   ' 27
conditions for their emergence, it is nonetheless possible to identify certain factors that help bring them about.
The stabilization and continuity of sultanistic regimes require a certain modernization of transportation and communications as well as of the military and police organizations and some civilian administrations, to provide funds to sustain the rule and prevent threats to it. Thej.soJatical.of the rural masses, their lackuit^ejiucatjon, and their poverty are probably necessary to ensure their passive_£ubmijsxon^out of fear combined with gratitude for occasional pafer-l nalistic welfare measures made possible by a modicum of development.
|~ Since some of the classic sultanistic regimes of the Caribbean and Central
j America appeared in small countries, one might speculate whether size was a
p factor. Small internal markets mean that where customs fees are an important
f- source of funds, political control can be extended through customs, whose
revenues are easily appropriated. In more complex societies, sultanistic regimes have a chance of survival only when they can dispose of considerable resources produced by sectors that do not require a large modern industrial labor force and entrepreneurial class, a modern administration, urbanization, expansion of education, and so forth. Rentier states in which the regime is not bound by tradition (unlike the oil monarchies of the Arabian peninsula) ^uttJluj&iRSre, vulnerable to sultanization. Easily exploitable natural resources whose production is in the hands of one or only a few enterprises with high profits can provide the resources for such a regime, especially when elites are weak. Sugar, ; oil, and copper exemplify this, each in a different way.
; Sugarcane cannot be sold without industrial processing. Even independent
farmers are thus dependent on the sugar mills, which are capital intensive. And since the market for sugar is mostly international, the product has to go through customs. This constellation of factors allows for the political-economic symbiosis that lends itself to sultanism, as Cuba and the Dominican Republic
1 illustrate.
In Venezuela the discovery and initial exploitation of oil by foreign companies coincided with the rule of Juan Vicente Gomez (1908-35), and an oil-led boom also provided the basis for the rule of Marcos Perez Jimenez (1948-58); both dictatorships displayed marked sultanistic traits.3 Oil also made possible Iran's "petrolic despotism" in the latter years of the last shah's reign, from 1963 to 1979.4 And in Zaire "the copper industry [was] the declining treasure trove of Mobutu's corrupted state; Mobutu survive[d] politically and enriche[d] himself primarily by taking a percentage off the top of copper export sales."5
Finally^ massive doses of for,eign..aid or-loans can. encourage corruption, especially if the aid is unconditional. In the Philippines the enormous borrow-
f
28     •     COMPARATIVE STUDIES
ing of the years 1979-84 made it possible for crony capitalism to thrive.6 In Haiti, U.S. aid grew tremendously after Papa Doc's death,7 but much of the money was diverted into the pockets of both Haitian and American officials. As one observer put it, "the principal recipient of foreign aid was Jean-Claude Duvalier."8 After the 1972 Managua earthquake, Somoza bought cheap land ■outside the city, then sold it at a hundred times the original price to the National Housing-Jnsátete^^uiid-iiCTHtousiiig. A'U.S. idiefyrarpaid'for the purchase, but not a single house was built.9 Foreign aid also helped enrich Mobutu;10 in Equatorial Guinea "the Nguema dictatorship in its last years was largely sustained by the fiscal infusions of the United Nations Development Program into the country";11 and Antigua and Barbuda, whose population is about seventy thousand, has received about $200 million in various kinds of U.S. aid since the late 1970s.12
It would be a mistake to consider sultanistic regimes an inevitable outcome of socioeconomic structures, however, although there are certain structures that facilitate the emergence of sultanistic tendencies. One cannot ignore many other factors contributing to the emergence and stability of such regimes-including the interest in "order" of foreign investors who have established stable "business relations" with the ruler. In many cases the rulers owe their rule to foreign intervention in the first place.
Crisis of Sovereignty
I will take advantage of your generosity to express the doubts which ... we moderns have about such things as... the Nicaraguan independence.    c. K. Chesterton, The Napoleon ofNottingHill
One feature most of our cases have in common is that throughout their contemporary history their independence was ambiguous and often not respected by more powerful neighbors. In terms of political development, therefore, these polities have suffered persistent crises of sovereignty.13 This pattern is most striking for Central America and the Caribbean.14 It can be no coincidence that an international law textbook published in 1922 lists in its chapter on "dependent states" Cuba, Panama, the Dominican Republic, Haiti, and Nica ragua—all nominally sovereign at the time but precisely those states that later fell victim to sultanism.15
Perhaps the most striking example is the Dominican Republic. After gaining independence from Spain in 1821, the country was annexed by neighboring Haiti in 1822, only to recover its independence in 1844. By 1861 the ongoing civil,
A Theory of Sultanism 2   • 29
political, and international strife impelled the authorities to request the return of Spanish colonial rule. The Spaniards stayed until 1865. Once it regained its independence, the country's president sold his country to the United States, but the U.S. Senate refused to give its consent to the ratification of the treaty. In 1903 the United States took over the customs receivership and occupied the country from 1916 to 1924, actually administering it between 1916 and 1922. Thrjillo came to power in 1930, and four years after his overthrow the United States again invaded the country in 1965.16
In Nicaragua foreign powers constantly intervened in civil wars of the nineteenth century, until in 1855 an American adventurer, William Walker, invaded the country with fifty-eight fellow Americans and declared himself president. From 1912 to 1933 Nicaragua was constantly occupied by U.S. troops, against which Augusto Sandino waged the struggle that made him the lodestar of Nicaraguan nationalism.17 This era ended only when Somoza Garcia took power in 1934. After his family's overthrow in 1979, the United States again meddled heavily in Nicaraguan affairs by arming rebels and mining the country's harbors.
The annexation of Cuba either by purchase or by outright seizure had been openly discussed in the United States during the nineteenth century.18 When the country finally became independent in 1902, U.S. influence and interference were legally guaranteed by an amendment to the Cuban constitution that Cuba's constitutional assembly had approved at the insistence of the United States. The "Piatt Amendment" granted the United States access to bases and coaling stations and constrained Cuba's sovereignty in financial and territorial matters. Its crucial third article stated that the United States could "exercise the right to intervene for the preservation of Cuban independence [and] for the maintenance of a government adequate for the protection of life, property, and individual liberty."'5 Although the clause was abrogated in 1934 by mutual agreement, the United States remained an interested actor in Cuban politics until 1959,20 occupied the island between 1906 and 1909, and sponsored an invasion of it 1961. To this day the United States leases Cuban territory for a military base at Guantanamo Bay.
Panama owed its independence from Colombia in 1903 to U.S. intervention, motivated by interest in constructing a canal linking the Pacific and Atlantic Oceans. By a treaty signed in 1903 the young republic granted the United States the use, occupation, and control of parts of its territory (what became the Canal Zone) and the right to deploy military forces to defend the canal and shipping. Of all the countries of Central America, Panama remained most closely tied to the United States. It never had its own currency, and its use of the U.S. dollar led
30     *     COMPARATIVE STUDIES
to the development of a sophisticated banking sector that was used for both legitimate and criminal purposes (laundering drug money).21 In 1941 the United States promoted the ouster of elected president Arnulfo Arias, who was suspected of fascist sympathies. A U.S. exclave cut Panama in half until President Carter signed a treaty in 1977 gradually turning over administration of the Canal to the Panamanians. And in the end it was by an invasion that Manuel Noriega was toppled in 1989.22.......—.....-...........-......----------------------------------------.............- -
Haiti, the second oldest independent republic of the Western Hemisphere, suffered American intervention several times in the nineteenth century and was occupied by American troops from 1915 to 1934. Francois Duvalier's rule (1957-71) contained a strong element of black nationalism and was at times tinged with anti-American rhetoric. But in 1986 Americans played a decisive role in organizing his son's departure, and they invaded again in 1994, this time to help reestablish democracy.
The crises of sovereignty of the five nations are well described in the aforementioned textbook of international law, which concludes that "in proportion as the United States by virtue of these conventions exercises rights which they confer as a privilege peculiarly its own, and in which no foreign State is permitted to participate, it appears to assume internationally a certain responsibility for conditions of government within the territories concerned."23
In the Philippines the American withdrawal after World War II was fraught with ambiguities, as David Steinberg explains: "Monetary arrangements, trade issues, access to natural resources, and land ownership were all decided in ways that were favorable to American interests, and the decision to retain vast territories as military bases shattered the illusion that independence would usher in a golden era of Philippine-American partnership based on mutual respect."24 The infatuation of many Filipinos with the United States was such that at times there have been movements advocating U.S. statehood for the islands.
Iran was never a colony in the formal sense of the word, but its independence was heavily impinged upon in the heyday of European imperialism, a status Lenin termed "semi-colony."25 By 1907 the United Kingdom and Russia had divided the country into zones of influence; during World War I the belligerent states did not respect Iran's proclaimed neutrality; in 1919 Britain tried to force a protectorate on it; in 1941 the Allies again invaded Iran in disregard of its neutrality; and in 1953 the United States and Great Britain engineered the overthrow of the Nationalist government and enabled the late Shah to rule as an absolute monarch until his overthrow in 1979.26
In other cases foreign intervention is more circumscribed. An example of
A Theory of Sultanism 2   • 31
more limited intervention is Venezuela, where Gomez's coup of 1908 was directly aided by the United States.
In this context one might also note that when the provinces that later became Romania received autonomy under the Ottoman Empire, they were ruled until 1877 by Greek phanariots rather than by indigenous elites concerned with emancipating their country, as was the case, for instance, when Bulgaria was an Ottoman vassal state from 1878 to 1908. Ottoman suzerainty over Romania ended as a result of the Russo-Turkish War of 1877; the country's "independence and existence as a national state were thus externally determined."27
The newly independent Third World states whose political regimes are more or less sultanistic also show strong foreign impact. In Zaire the country's very independence in i960 sparked an international crisis that led to United Nations intervention, and when Mobutu staged a coup in 1965, he was helped by the U.S. Central Intelligence Agency.28 Among those British colonies that hosted U.S. bases as a result of the 1940 agreement between Churchill and Roosevelt, Antigua was the only island that kept its American bases after independence,29 and the agreement granting the U.S. military access to V. C. Bird International Airport is described by the U.S. State Department as "one of [the most] if not the most, generous arrangements with a foreign government anywhere in the world."30
What accounts for these infringements on the sovereignty of these countries? One factor that immediately comes to mind is that all lie in strategically sensitive areas. Hispaniola and Cuba are very close to the United States, Nicaragua and Panama were possible sites for a canal linking the Pacific and Atlantic Oceans, the Philippines hosted America's most important bases in Asia, and Iran occupies the land between Russia on the one side and India and the Near East on the other. Zaire is in the heart of Africa, and Antigua was home to important U.S. military installations during the Cold War.31 The domestic situation in these countries has therefore been of interest to foreign powers, chiefly the United States.32
A second factor at work in the Caribbean was that chronically unstable governments were unable to repay their debts, which would then lead to foreign intervention to secure payments. In Nicaragua, Haiti, and the Dominican Republic this would be done by taking over customs.
The fact remains that not all countries that have experienced long crises of sovereignty have developed along sultanistic lines: Egypt and China come readily to mind. Prolonged crises therefore seem to be a favorable but not sufficient precondition for the appearance of sultanism. How, then, can we elucidate the
32     ■     COMPARATIVE STUDIES
A Theory of Sultanism 2   • 33
link between foreign intervention and sultanism? It seems that sultanistic ten-dendejsari^ewhen a foreign power seeks some control over a country but wants to retreat from direct occupation. The sultanistic ruler then becomes the guarantor of the external power's interests. Therefore it is in some sense the very "retreat of the foreign power that favors sultanism, since the foreign power leaves behind a partially modernized administrative and especially military apparatus that enables the ruler to c&BCjentatej»«er in .his handsJus^itaatfon-applks-to the Nicaraguan and Dominican National Guards, Haiti's Garde d'Ha'iti, and the Panamanian Defense Forces, which were strengthened by the United States after the 1977 treaty so as to provide security for the Canal and which after Omar Torrijos's death in 1981 enabled his eventual successor, Manuel Noriega, to rule in a more and more sultanistic way after he took power in 1984.33 In the African cases the rise of the intellectually mediocre Macias Nguema was engineered by the Spanish so as to neutralize more educated nationalists, and the British in Uganda constantly promoted Idi Amin in disregard of all criteria for advancement, so that a man who had enlisted in the King's African Rifles as a cook and remained functionally illiterate all his life was rewarded for his services (including those against the Mau Mau rebellion in neighboring Kenya) by being given the highest rank open to an African in i960, just before independence.34
In the American sphere of influence, the sultanistic rulers have U.S. support for most of their tenure. Trujillo enjoyed it until Batista was overthrown, and when Francois Duvalier died in 1971 the United States deployed two warships between Haiti and the American mainland to prevent the return of exiles, thus aiding the transfer of power from father to son. The Shah was called by Henry Kissinger "that rarest of leaders, an unconditional all/';35 Ferdmand'Marcos enjoyed the support of five U.S. presidents; and Mobutu's rule was twice rescued by Western intervention when rebel forces invaded Shaba province in 1977 and 1978. But we should also point out that America was interested not in perpetuating sultanistic rulers per se, but in maintaining stability and a general pro-American stance under strong states. Democratic regimes fulfilling these conditions were also supported,36 as was communist Yugoslavia.
Thus domestic instability first creates the conditions that make it easier for foreign powers to intervene, then foreign intervention leads to regimes that substitute foreign support for domestic coalition building and may become sultanistic. The foreign aid that flows into these countries can further deepen sultanism. Finally, note that the very fact that all these regimes (with the exception of Romania, which is a borderline case) have appeared in the Western sphere of interest is perhaps no coincidence, since their lack of identification
with a genuine ideology is more compatible with dominant Western views of politics. In those Third World regimes that were in the communist sphere of influence, such as certain sotialisant regimes in Africa, or influenced by fascism, such as Getiilio Vargas's in Brazil or Juan Domingo Peron's in Argentina, the self-conscious imitation of the ideological models in the developed countries probably acted as a brake on sultanistic tendencies. Romania of course does not fit this pattern, but even here the appearance of sultanistic traits in Ceausescu's regime more or less coincided with his rapprochement with the West, where he always had a better press than many less brutal communist dictators.
Political Institutional Factors
As we saw earlier, sultanistic regimes do not have a specific set of institutions; sultanistic tendencies appear under various constitutional arrangements. Two sets of paths to sultanism can be empirically observed: the breakdown of clientelist democracy and the decay of nondemocratic regimes.
Breakdown of Clientelistic Democracy
Some sultanistic regimes have come to power in countries with at least some previous history of democratic politics. We know that certain features of a democratic regime can survive the breakdown of democracy and resurface under the changed conditions of the successor dictatorship. Similarly, it seems that dictatorships arising after the breakdown of clientelistic democracies will sometimes display sultanistic tendencies: one could say that the number of patrons has been reduced to one." In a clientelistic democracy, patronage, and favors have to be widely distributed to win elections, but the alternation of parties in power keeps one party from monopolizing the pork barrel indefinitely. The electoral competition thus favors a wider distribution in society of the benefits of patron-client relationships. Batista's regime after his coup of 1952 and Marcos's after 1972 fit this pattern most clearly, but elements of it are present also in Nicaragua, where, as John Booth shows in his chapter, the Somozas' rule began with the first Somoza's taking over the Liberal Party, one of the country's two traditional clientelism-based parties. But Nicaragua is different in that the Liberals and Conservatives had deep class and regional roots in the country and in fact represented two camps in Nicaraguan politics, somewhat as happened in Colombia.58
Cuba enjoyed a constitutional government between 1939 and 1952, but "clientelistic politics that revolved around personal attachments rather than doctrinal commitments made the system vulnerable to shifting partisan arrangements
4
34     '     COMPARATIVE STUDIES
and manipulation from above."3' The widespread corruption that accompanied democratic government led to considerable initial support for Batista's coup of 1952 and little overt opposition to it.
In the Philippines, pre-1972 democracy had been marked by competition between the Liberal and Nationalist Parties.-10 Party labels did not denote ideological differences, however, and candidates would often switch affiliation before elections. Pork-barrel was the--pr-in6ipal-"method-ef-gaiiMng-votesv The — system remained competitive because those who lost, or believed they did not benefit sufficiently, switched to the other party. After his election in 1965, Ferdinand Marcos used this system to a greater extent than all his predecessors—so much so that in 1969 he became the first sitting president to be elected to a second term.4'
Decay of Authoritarian and Totalitarian Regimes
In a traditional polity like premodern China, the harem and the eunuchs associated with it would often wield power after an emperor died, while his underage successor was under the tutelage of women.42 In the West, the principle of primogeniture prevented the emergence of court parties that put rulers on thrones, whose occupants would then trust nobody and perpetuate a climate of insecurity. In Iran, as Homa Katouzian shows in his chapter, the sultanism of the two Pahlavi shahs derived in part from the old patrimonial monarchy that Had been in place until the constitution^ revolution of 1906-7. But that revolution, although ultimately unsuccessful in installing a stable .liberal regime, did prepare for the creation of a centralized state with a modern bureaucracy, a state Reza Shah inherited and developed further. He and his son first established authoritarian regimes (1921-33 and 1953-63, respectively) and then proceeded to govern sultanistically after eliminating their erstwhile allies from all positions of influence (1933-41 and 1963-79). The development of sultanism in Iran therefore combines elements of both its traditional and modern varieties.
Modern authoritarian, regimes, .sometimes become deinstitutionalized as time passes and the rulers become more mercurial and corrupt. Batista's second presidency (1952-58) had strong sultanistic tendencies, whereas in his first term (1940-44) he had ruled constitutionally.43 Jean-Claude Duvalier's rule (1971-86) was more sultanistic than his father's (1957-71), since he neglected to maintain the patronage network his father had built. Noriega's sultanism followed the mild and rather progressive authoritarianism of Omar Torrijos, who had made great efforts to extend the social bases of his regime by reaching beyond Panama's traditional white elite.
A Theory of Sultanism 2   • 35
African sultanism also results from a degeneration of authoritarianism, but of a different kind. The colonial legacy in the sub-Saharan states had two consequences that favored personalism. First, these states' arbitrary borders meant that each state's population was a culturally heterogeneous mix of peoples with no common precolonial traditions, and so at independence the young states started out with no remnant of traditional legitimacy. Second, colonial government was by definition authoritarian government, and it was under this form of rule that the new states' elites were socialized into politics.44 When, after independence, the new elites took over the administrative bureaucracies bequeathed by the Europeans, these bureaucracies were transformed from instruments of policy to "patrimonial administrations in which staff were less agents of state policy (civil servants) than proprietors, distributors, and even major consumers of the authority and resources of government."45 This development is in contrast to the experience in the other two main areas of post-World War II decolonization—the Caribbean, where responsible self-government preceded independence by many years, and Asia, where precolonial state traditions survived European imperialism.46 Personalism is thus rampant in Africa, although as we saw earlier it has many guises, and only a few regimes in fact fit our model of sultanism.
Totelitorjaji_rjegime^can also at times be vulnerable to institutional decay leading to sultanization. Ceau§escu's regmfe is IfhiOwsMx^pie of this, al-"though strong sultanistic tendencies can also be observed in Saddam Hussein's Iraq. Here the Baath Party's 1968 coup d'etat created a one-party regime endowed with a distinct ideology that can in many ways be called totalitarian.47 Yet by the 1990s Saddam Hussein's relatives occupied many vital positions in the state, some of them being heavily involved in corruption; a personality cult had developed around Hussein that even grew after his defeat in the Gulf War; and a cult of ancient Babylonia appeared alongside the Baath's Arab nationalism.45
One pattern of arrested sultanization deserves particular notice. In modem authoritarian regimes, sultanistic tendencies, especially those associated with family-related corruption, sometimes appear toward the end of a ruler's tenure in office. One.of. the-biggest-problems-faeing -any nondemocratic regime is succession, since very few have any fixed and accepted rules to regulate the passing of power from one ruler to another.45 Authoritarian leadersthus often stay in office well beyond die point where, they can effectively exercise power. Within the regime the lack of a widely accepted successor can then lead to an inertia where all involved agree to postpone the inevitable as long as possible. At this point those who enjoy the closest personal access to the aging ruler, often family members, can wield great influence behind the scenes, since they are in a
36     *     COMPARATIVE STUDIES
A Theory of Sultanism 2   * 37
position to manipulate him in ways that further their own interests. One might call this phenomenon fin-de-regne sultanism. Some of the hierarchical military regimes in Latin America—Argentina, Brazil, and Uruguay—avoided this eventuality by insisting on collective rather than personal leadership and on limited mandates with presidential elections among the military.
In Spain, the aging Franco was subject to considerable pressures from family members (most nota'v. hi* -■..i!-ii«-!-i-.v (he mart:i.i» «>f Villawrdf: tv-A pc'on* in his entourage to tamper with the rules of succession so as to replace Prince Juan Carlos with his cousin Alfonso de Borbon-Dampierre, who was married to Franco's granddaughter. Don Alfonso was made duke of Cadiz, but Franco ultimately resisted these pressures. The emerging camarilla was also suspected of corrupt dealings.
In the Soviet Union, the last years of Leonid Brezhnev's tenure as secretary-general of the Communist Party were marked by increased political clout and self-enrichment by his family members.50 In Tunisia sultanistic tendencies went beyond family corruption in the last years of President Habib Bourguiba's long rule. Tunisia's postindependence political system was in many ways a classic authoritarian regime in which a single party, the Parti Socialiste Destourien, provided the backbone of the institutional structure.51 In 1974 Bourguiba, who had reached the end of his constitutionally allotted three terms, was "elected" president for life. At the same time, seven high-ranking signatories of a declaration deploring arbitrary decision making were expelled from the party.52 By June 1986 institutional decay had reached the point where the president could suspend the election of the members of the PSD's central committee by a party congress and instead name them himself. His niece, who had managed to drive the president's wife into American exile, was now his closest aide and screened all his visitors. She, his chef de cabinet, and the president's personal physician became Bourguiba's closest advisers and opposed the prime minister, Mohammed Mzali, who had tried to reach an accommodation with the opposition.53
The state of affairs in Spain in the very last stage of Franco's rule, the Soviet Union in the last years of Brezhnev, and Tunisia in the last years of Bourguiba were a departure from previous norms and tended toward what might be called a "sultanistic situation,"54 In Spain and the USSR the death of the ruler set in motion institutional mechanisms that in the first case led to a legally regulated establishment of the monarchy that later eased the transition to democracy,55 and in the second resulted in the reformist rule of Yuri Andropov, whose brief tenure in office in 1982-83 prepared the ground for Mikhail Gorbachev's reforms. In Tunisia Bourguiba's hand-picked security chief Zine El Abidine Ben
Ali ousted the president for life in a palace coup and inaugurated a period of liberalization that for a while seemed to lead to democratization but instead ended in a reconsolidation of the authoritarian regime.55 In authoritarian regimes, therefore, sultanistic pressures face obstacles that diminish the likelihood of a sultanistic situation's leading to a sultanistic regime. However, the possibility always exists.
Leadership Factors: The Ruler's Personality
There is no simple biographical and psychological portrait of the ideal-typical sultanistic ruler, but certain traits are found with some frequency in these dictators, although of course none can be found in all. Founders.oT sultanistic regimes tend to have limited education and come from socially margmartacEgrounds, and their upward mobility tends to have gone come through 'accidental channels. They tend to be shrewd but morally unscrupulous, distrustful, and vindictive, and they often show an amazing capacity to lie and womanize. In addition to their hedonism, they often betray a streak of personal cruelty.57 Some of them, like Francois Duvalier, Macfas Nguema, Gairy, Bokassa, and Amin, dabble in the occult.58
Given that sultanism is an extreme form of personalism, it is almost tautological to point out that the personality of the ruler is a key element in understanding a sultanistic regime. But whether a leader with a good number of these traits appears on the scene is not structurally determined, and thus one cannot rale out an element of chance. An interesting feature of some sultanistic regimes is mat .the founder is often more politically .savvy than his son (and somtime heir): examples are Reza Shah.and his son.Mohammad Reza Shah, Trujillo and his son Ramfis, Papa Doc and Baby Doc, Somoza Garcia and Somoza Debayle, Nicolae and Nicu Ceausescu, and perhaps Kim Il-sung and Kim Jong-il.
The Transformation, Breakdown, and Legacy of Sultanism
A sultanistic regime can endure a long time, but experience shows that most end in a. more or less chaotic way. When they do come to an end, they are less likely than other types of nondemocratic regimes to be replaced by democracies.55 Even in Africa, where personalism is rampant, the record of the early 1990s shows that democracy was least likely in those regimes we would call sultanistic."0
38     '     COMPARATIVE STUDIES
The specific patterns of regime breakdown and transition are dealt with by Richard Snyder in the next chapter. In the following sections we will merely discuss some general themes that often inform the end of these regimes.
The Difficulty of Political Liberalization
The most perilous moment for a bad government is one when it seeks to mend its ways. alexis de tocqueviile,
The Old Regime
The weakness of institutions and the manipulation of democratic procedures make the establishment and consolidation of democratic politics very difficult. If the sultanistic leader or his successors in an interim regime decide to begin to liberalize, the chances that the attempt might succeed and pave the way for a genuine democracy are slim.
An established one-party state can democratize itself by allowing .one..or, more opposition parties to contest elections: Turkey in 1945, Senegal in 1976, South Korea in 1988, Eastern Europe after 1989, the two island states of Cape Verde and Sao Tome' e Principe in 1991, and Taiwan in 1989 are examples of this scenario.61 In bureaucratic-authoritarian regimes, which more often than not present themselves as temporary solutions with no claim to permanent legitimacy, the ruling military can decide to extricate itself from the exercise of power by holding elections and handing over control to the winners of these elections. That the military ruled as an institution usually means that the extrication of the military is not contested.62
In either case, the "first free elections," often "stunning elections" in which the opposition wins unexpectedly for the rulers,63 are a clear break with the past and are likely to provide subsequent politics with considerable legitimacy. At the same time, the institutional continuity with the predemocratic regime means that elements more or less connected with the authoritarian regime have a role to play in the new democracy, which weakens opposition to it: examples are Hernan Buchi in Chile, Manuel Fraga Iribarne in Spain, and the renamed "socialist" parties in Eastern Europe. The institutional continuity also means that former single parties can come to power again, as happened in Turkey, Lithuania, Hungary, Bulgaria, and Albania. The relative strength of the state of law under the predemocratic regime also allowed the constitution of opposition groups in the crucial period before and during the transition.
Another factor favoring the establishment of democracy is that injiooi sultanistic regimes pockets of legal-rational authority remain, such as capitalist
A Theory of Sultanism 2   • 39
enterprises, the judiciary, and the armed forces. Entrepreneurs are interested in predictability for their planning and thus prefer democracy to political upheavals,64 judges jealous of their independence can provide legal protection to critics of the regime, and the military is interested in maintaining its institutional integrity.65 The spread of legal-rational authority in society is crucial, for without it there can be no stable liberal politics.
Under sultanistic regimes the.situatiorLis.diSerent. If, for whatever reason (foreign pressure, attempt to defuse mounting opposition by providing a safety valve), the sultanistic ruler decides to liberalize hisjegime,the:chances that this might lead to democracy are limited. Given the prior history of fear and suspicion, the democratic opposition is unlikely to trust the change. In addition, various actors are likely to accuse each other of past collaboration with the regime or of readiness to be co-opted by it. Since the.regime,,is.exttemely. personalistic, th&,.r.ukr.-himself is the target of the opposition:, compromise between the regime and the opposition becomes well-nigh impossible, since there is no neutral force to which both could appeal as an arbiter. Xheuop.po.si-tion demands nothing less than the ouster of the ruler and often his trial, for it cannot lend credence to his promises to lead democratically henceforth or to abandon power peacefully, given his record of deceit and manipulation. Although there may be groups within the regime willing to defect from it, there are no trusted moderates in the regime to negotiate with, since everybody is tainted by association with the ruler. All of this means that the ruler cannot look forward to a peaceful life after his ouster.66 The cost of tolerating an opposition is therefore high, and this makes liberalization problematic.
In the period after he promises liberalization, the emperor's nakedness becomes apparent to all: as the crisis drags on, the probability of the regime's weathering it diminishes. Given its lack of links with civil society and its narrow social base, the regime's capacity for countermobilization is severely limited, which encourages maximalist tendencies in the opposition.
It is during the crisis following the initial promise of liberalization that the regime loses whatever vestiges of legitimacy it may have retained with at least some citizens. Natural calamities, which are "unique historical events" and therefore not integrated in theoretical constructs,67 seem to come into play with striking regularity in sultanistic regimes. By demonstrating the regimes' inability to provide adequate relief and giving them an opportunity to pocket aid coming from abroad, natural disasters delegitimize them. In Iran, the two earthquakes of 1962 and 1978 (claiming twenty-five thousand and twelve thousand victims, respectively) preceded the uprising of June 1963 and the 1979 revolution. In Nicaragua, the Managua earthquake of 1972 (five thousand vie-
40     *     COMPARATIVE STUDIES
A Theory of Sultanism 2-41
tims) led to greater corruption and deepened the citizenry's outrage. In the Philippines, the 1976 earthquake (eight thousand victims) destroyed the church in which President Marcos's daughter had been married amid great pomp and circumstance, a symbol not lost on the population. Finally, a series of natural disasters characterized the early 1980s: in Haiti swine fever broke out in 1981 and led to a government-ordered slaughter of pigs, the peasants' traditional source of wealth; Hurricane AllendestroyedjfHsh.ofilie.m&eJiiltur^heidMtifiGa---tion of AIDS and its discovery there hurt the tourist industry; and in 1985 a drought led to a 20 percent fall in food production and dried up the hydroelectric capacity of Lake Peligre, leading to widespread rural flight, transportation difficulties, and hunger. The blatant inability of Duvalier to respond worsened the situation and helped undermined his regime.68 The special vulnerability of sultanistic regimes to natural disasters is demonstrated by the relative ease with which the Mexican regime coped with the 1985 earthquake in that country: in spite of Mexico's endemic corruption the political system was institutionalized enough to cope with the crisis.6*
The basic illegitimacy of the sultanistic regime has yet another dimension in the case of the pseudodemocracies, since here "free elections" are not a clear break with the past, because the government claims that elections have always been free. This is precisely what arouses the suspicion of the opposition, which denies that the government has the will and the moral caliber to organize genuine elections. It also strengthens the hand of the nondemocratic opposition in its competition with the democratic opposition, for it can accuse the latter of being what we would call a pseudo-opposition, opportunistic or, at best, hopelessly naive, The manipulation of democratic procedures by the sultanistic regime thus weakens the appeal of democracy itself. A corollary of this is that sultanism begets a political atmosphere in which scheming, conspiracies, betrayals, secretiveness, and rumor mongering flourish. And these are not ingredients of a political culture conducive to democracy.
The contrast with the people's democracies of Eastern Europe is telling: the bloc parties never pretended to rival the ruling Communist parties and consequently were able to play a role during the transition to democracy, either as autonomous parties or, in the case of East Germany, as regional organizations of the established West German parties. It seems that only if a sultanistically ruled country has a history of democratic politics can democracy be revived in the course of an election: the Philippines is the pertinent case,
Faced with a serious challenge, the sultanistic regime disintegrates rapidly. When Batista was faced with Castro VrebeTs,"the Shah with essentiallyuharmed demonstrators, Idi Amin with an invasion force consisting of exiles and Tan-
zanian troops, Jean-Claude Duvalier with widespread demonstrations, and Mobutu with a strong rebel force, their regimes collapsed. It is a sign of these regimes' close links with foreign powers that the latter are closely involved in the final transfer of power. Both the dictators and their opponents believe that foreigners make and unmake governments, and this itself can draw in outside powers. Given mounting opposition, the rulers' foreign allies abandon them and often try to look for a "third force" that can stave off a revolution; occasionally, as in the Philippines and Haiti, they even find one and arrange for the dictator to leave. If no third force is found, the ruler and his cronies flee abroad at the last moment, as they are abandoned by everybody.70 The least lucky ones are ousted by their former patrons once they are no longer useful: the U.S. invasion of Panama and the subsequent arrest of Manuel Noriega, which cost the lives of five thousand Panamanians, is the most glaring example, although France also aided militarily in toppling Emperor Bokassa in 1979—after it had paid for his coronation.
Sultanism and Mass Movements
Men trust a great man because they do not trust themselves. And hence the worship of great men always appears in time of weakness and cowardice; we never hear of great men until the time when all other men are small.
g.  k. chesterton,
Heretics
As Jeff Goodwin and Theda Skocpol have argued, sultanistic regimes are more vulnerable to revolutionary overthrow than liberal democracies or inclusionary authoritarian regimes:71 recent examples include Cuba, Iran, Nicaragua, and Grenada.73 Sultanistic dictators are more likely to generate elite and middle-class opposition from landlords, businessmen, clerics, and professionals, who resent their monopolization of key sectors of the economy, their heavy-handed control of the flow of ideas and information in schools and the press, their subservience to foreign powers, and the general climate of corruption.
In certain circumstances, analyzed by Richard Snyder in the next chapter, su|tanisticrulers come to face.a genuine revolutionary movement. In a vacuum of.authority and. with increasing delegitimation of existing institutions, there arises the need for someone to assume leadership, A full-blown societal crisis favors the emergence of charismatic authority or, as Weber put it, "the 'natural' leaders in moments of distress—whether psychic, physical, economic, ethical,
42     •     COMPARATIVE STUDIES
A Theory of Sultanism 2-43
religious, or political—[are] neither appointed officeholders nor . . . persons performing against compensation a 'profession' based on training and special expertise ... but rather the bearers of specific gifts of body and mind," that is, charismatic leaders.73 Of course one must concede an element of chance here, since the presence or absence of a leader able to catch the imagination of the masses is not determined structurally: Cuba, Iran, and the Philippines found such leaders in Fidel Castro, Ayatollah Ruhollah-Khomeinirand Corazon Aquino (none of them officeholders or endowed with special training or expertise in statecraft); in Nicaragua someone like Eden Pastora, who had marched into the capital after Somoza's flight, might have emerged as a charismatic leader, but he decided to break with the Sandinistas; in the end the comandantes collectively derived some charismatic authority from the myth of Sandino. In Haiti it took a few years for the charismatic leadership of Father Jean-Bertrand Aristide to emerge.
Religion plays an important part in the opposition to sultanistic rulers, at least in the last stage of the regime. To some extent this is due to the weakness of civil society under sultanism, which, almost by default, confers great importance on those few institutions that maintain a presence in society. Religious organizations and in the end organized religion become a major locus of oppositional activity as they provide support, resources, and leadership.
But there is more to the role of religion than its organizational resources. The ubiquity of sin and the general degeneration of mores under high sultanism (the last years of Somoza's, the Shah's and Jean-Claude Duvalier's regimes), combined with the unrestrained nouveau-riche lifestyle of the elites, breed resentment. This resentment leads to a "religious ethic of the disprivi-leged which .. . teaches that the unequal distribution of goods is caused by the sinfulness and the illegality of the privileged, and that sooner or later God's wrath will overtake them."74 This explains why religious opposition typically begins not among the hierarchy, but among the laity or lower-ranking priesthood. In the days before Vatican II and the Latin American bishops' conference of Medellin in 1968, Catholic bishops could still support Trujillo during most of his rule,75 or be divided in their attitude toward them (Batista),76 but in the 1970s Latin America's Catholic Church became more critical. In Nicaragua, Christian base communities that started in the mid-1960s and increased in numbers after the 1972 earthquake opposed Somoza's dictatorship.77 In Haiti the Catholic base communities (ti legliz) organized opposition among the peasantry and were initially regarded with suspicion by the hierarchy.78 In the Philippines a number of priests were active in the poorest regions of the country at a time when a considerable sector of the bishops was still apolitical or discreetly
pro-Marcos.79 In Iran, finally, religious agitation in the 1970s was mainly inspired by Ali Shariati, whose teachings were frowned upon by the ulema.80
This antidictatorial stance of lower-ranking clerics or even laypeople initially contrasts with that of the hierarchies, and it is only gradually that the latter come to join the active opposition.81 Nicaragua's bishops, all of whom had been critical of Somoza, and the ayatollahs residing in Iran, who had spoken against the Shah, joined the struggle openly only a few months before the overthrow of the dictators. In the Philippines and Haiti it was visits by the pope himself, in 1981 and in 19S3, that put the church on record being critical of those regimes.82 In Nicaragua and the Philippines the two events that gave the conservative hierarchy the final push were the assassinations of Pedro Joaquin Chamorro in 1978 and Benigno Aquino in 1983. Both men were moderate opponents of the sultanistic dictators, and their deaths signaled to the clergy that repression had reached an intolerable level—especially since the Chamorros were friends of Archbishop Miguel Obando y Bravo and the Aquinos of Archbishop Jaime Sin.
In the Catholic countries the switch in the attitudes of the clergy stripped the dictators, who liked to present themselves as champions in the struggle against atheistic communism, of any shred of legitimacy. An almost mystical atmosphere pervades the change of power, which has aspects of what Victor Turner called "iiminality":83 In Cuba a dove's landing on Castro's shoulder at a mass rally on 8 January 1959 was interpreted by some Catholics as a sign of the Holy Spirit,84 and in the Philippines on Easter Sunday 1986 Cardinal Sin drew a parallel between the resurrection of Christ and the redemption of the Philippine nation.85
Contemporary Catholic doctrine frowns on the clergy's direct participation in partisan politics, and so relations between the hierarchy and such activist priests as Miguel d'Escoto and Ernesto Cardenal in Nicaragua, or Jean-Bertrand Aristide in Haiti, soon cool down, to the point where by now the last two have left the priesthood. The politicized part of the Shi'ite ulema in Iran, unbridled by the principle of rendering to Caesar the things that are Caesar's and to God the things that are God's,86 took power soon after the Shah's overthrow and neutralized both the politicized laity and the quietist ulema opposed to the exercise of direct power.87
The sultanistic regime's inability to find a political solution to the political crisis has already been shown, yet when confronted with a military challenge, it does not fare much better. The armed forces are particularly corrupt and at times even inefficient, in part because the dictator has been more concerned with preventing his own overthrow by military coup than with establishing an effective fighting force. The very fact that they are a praetorian guard in the end
44
COMPARATIVE STUDIES
promotes the defection of many members from the sultanistic ruler. The Cuban military's dealings with Fidel Castro, discussed by Jorge Dominguez in his chapter, are the best example of the unreliability of a sultanistic army, a pattern confirmed by the performance of Mobutu's troops in 1997, but such a fighting force can also be quite willing to fight, as was the case in Nicaragua, where Somoza's National Guard fought the revolutionaries and was defeated in a civil war. When the rulers, finally, decide to-step-down,-their-armies,rgiven-then-organizational incoherence, often disintegrate, opening the way for a guerrilla army (Cuba) or countergovernment supported by mass movements (Iran, the Philippines) to take power.
After the installation of a revolutionary government, prospects for democracy are bleak. If a charismatic leader emerges who is committed to democracy, it has a chance, as the Philippines show; but even here that nation's long history of democratic governance before 1972 helped too. Where the charismatic leaders are not democrats, the mass movements they lead will drown the democratic forces, which are weak for reasons already discussed. Even if the postsultanistic regime lacks a charismatic leader, the absence of strong and independent state institutions makes the functioning of democracy very difficult: Nicaragua, Haiti, and Romania illustrate this problem. The flight of the ruler and the delegitimation of his regime lead to a genuine break with the past and the installation of a provisional government, which is composed of ideologically heterogeneous people who have no electoral mandate but who face enormous societal expectations.88 The moderates in such governments are sooner or later pushed aside either by radicals (Cuba, Nicaragua, Iran) or by remnants of the old regime (Dominican Republic, Haiti, Romania). If the regimes thus constituted are more authoritarian than sultanistic, a transition to democracy may still come about, but not without international pressure: in the Dominican Republic the United States was involved in the assassination of Trujillo and insisted in 1978 that the incumbent regime accept the electoral victory of the opposition. In Haiti and the Philippines the United States actively sought to remove Jean-Claude Duvalier and Ferdinand Marcos from office after widespread demonstrations led the Reagan administration to fear another "Iran." In Panama Americans removed the dictator themselves. In Nicaragua American assistance to the contras resulted in a stalemated civil war, then international pressures and assistance led to an agreement that made free and fair elections possible. Thus Violeta Barrios de Chainorro acceded to the presidency in i99o.89 Finally, outside assistance was also required in Haiti to reinstate the democratically elected president, who had been ousted by a coup.90
In the revolutionary cases, the crises of sovereignty that the sultanistic rulers
A Theory of Sultanism 2
45
never resolved come to haunt their former foreign patrons after the revolution. The postsultanistic regimes are more likely than not to blame their countries' travails on foreign powers, and since the new powerholders share the view that their predecessors were foreign puppets, they interpret their victory as a victory over imperialism, leading to hubris that results in confrontational policies toward the United States. The Cuban, Iranian, and Nicaraguan revolutions were all anti-American.91
After Sultanism
Liberty is a food that is good to taste but hard to digest it sets well only on a good stomach. 1 laugh at those debased peoples that let themselves be stirred up by agitators and dare to speak of liberty without so much as having the idea of it; with their hearts still heavy with the vices of slaves, they imagine that they have only to be mutinous in order to be free. iean-jacques rousseau,
The Government of Poland
Whether a sultanistic regime's overthrow results in another dictatorship or a democracy, traces of it can often be found in the successor regime; the scars it leaves in a nation's polity are deep. If the sultanistic regime is replaced by a democracy, chances arejthis new democracy will display strong clientelist tendencies, with the .democratically elected leaders using the resources of their office to build nationwide patron-client relationships. Joaquin Balaguer is the best example; his last reelection in 1994 was riddled with irregularities. Even Corazon Aquino was accused of favoring her relatives after she took office.92
If the sultanistic regime is replaced by a revolutionary one, institution building in the new regime can be very slow, as is shown by the initial difficulties in Rornaniarcbfif rasting with more successful democratization elsewhere in Eastern Europe (Yugoslavia excepted).9' Moreover, the new system tends to pivot around the founder as long as he is alive, which means that personalism remains a feature of the system.. Castro's Cuba, and Khomeini's Iran illustrate Weber s insight that 'charisma knows no formal and regulated appointment or dismissal, no career, advancement or salary, no supervisory or appeals body, no local or purely technical jurisdiction, and no permanent institutions in the manner ot oureaucratic agencies, which are independent 01 tne incumbents and their personal charisma."91 In Cuba, after a period lasting roughly from the mid-1970s to the mid-1980s in which a halfhearted attempt was made to copy Eastern Europe's institutional structure, Castro once again rules in a manner
46     •     COMPARATIVE STUDIES
that one observer has called "socialist caudillism."95 Iran's political institutional structure is riddled with inconsistencies, and corruption and arbitrary government prevail.96 In Cuba as in Nicaragua, the Castro and Ortega brothers divided the presidency and the leadership of the armed forces between them, just as the Somozas and Trujillos had done before. And when the Sandinistas transferred the nationalized industries to the army after their election loss in 1990, the Ortega brothers, by maintaining their control ofthearmyrmfactlined their pockets handsomely. Under their successors, the politics of understudy continued, with President Chamorro allowing her son-in-law, Antonio Lacayo, to run the country.97
In terms of the rationalization and institutionalization of government, therefore, the situation in Cuba, Iran, and Nicaragua is quite reminiscent of the sultanistic regimes that were to be replaced. This is not to suggest that the revolutions made no difference: the masses probably benefited from the overthrow of the dictators, but we must also remember that Cuba's and Iran's revolutionary regimes have jailed, killed, and driven into exile far more citizens than Batista or the Shah.*8 Like the regimes they replaced, these revolutionary regimes present unfavorable preconditions for a transition to democracy," as is shown by the fact that Cuba followed neither other Latin American states nor the Soviet bloc on the path to democracy.
The "loss" of American clients in revolutions that ousted sultanistic rulers has not been easily digested by the U.S. government. Repeatedly the new regimes had to face American interference in their internal affairs, as old habits died hard: the Bay of Pigs invasion in Cuba, the Dominican intervention in 1965, and the support for the contras in Nicaragua have been cases in point, as was, in a different way, the invasion of Panama in 1989 after the local client was no longer acceptable.
The main conclusion to be drawn from a comparative analysis of sultanistic regimes is that, if overthrown, they are more likely to be replaced by a revolutionary or an authoritarian regime than by a democracy. To the extent that some of the sultanistic rulers could play on the West's interest in stability and on Western support during the Cold War era, they might become more vulnerable now that global competition between the superpowers has ended. However, a greater U.S. desire to avoid any foreign intervention and similar reluctance by other industrial democracies to get involved might allow the survival of the few sultanistic rulers that remain or even the establishment of new sultanistic regimes.
Most of the regimes analyzed in these introductory chapters belong to the past. This poses the question whether sultanism is still relevant to an understanding of today s world. Sadly, the answer is yes. Some sultanistic regimes in
A Theory of Sultanism 2   • 47
Africa, such as those ruling Zaire and Equatorial Guinea, have shown considerable resilience, and in those unfortunate countries where the state has ceased to exist, such as Somalia and Liberia, the various petty warlords that have replaced the state display many sultanistic traits.100 But subnational sultanisms are not confined to territories where the central state authority has broken down; they can also come about and persist in countries where the state is not strong enough to penetrate the nation. John Thayer Sidel has argued that although pre-Marcos national Philippine politics was dominated by bifactional competition, at the regional level many "local kingpins" monopolized political power in a quintessentially sultanistic way through "guns, goons, and gold."
What is more, this pattern of "petty or local sultanism" persisted under Marcos, even surviving the demise of his regime.101 Similarly, Guillermo O'Donnell has warned that in the aftermath of democratization and economic restructuring in such countries as Argentina, Brazil, and Peru, attempts to reduce the size and deficits of the state bureaucracy have weakened the government's ideological legitimation and its ability to uphold legality equally in different parts of the national territory. "Provinces or districts peripheral to the national center (which are usually hardest hit by economic crises and are already endowed with weaker bureaucracies than the center) create (or reinforce) systems of local power which tend to reach extremes of violent, personalistic rule—patrimonial, even sultanistic—open to all sorts of violent and arbitrary practices."102 Similar developments may be taking place in some Mexican states. Since it is at the local level that state policies are carried out, contested, reshaped, resisted, or revised,103 the spreading of sultanistic practices at the local level bodes ill for the emerging democracies. Parts of the former Soviet Union could conceivably fall victim to sultanistic tendencies. The persistence of nomenklatura control, coupled with the absence of an entrepreneurial class, has engendered an intertwining of political and economic power that could well lead to sultanism if a leader emerges who has the requisite qualities to raise himself above the party apparatus. Belarus's president Alexander Lukashenko seems to be on the way to becoming such a dictator, as is Turmenistan's Saparmurad Niyazov.'04 On the other hand, the higher density of international exchanges, the emergence of a transnational civil society, and the end of the Cold War, which has left some sort of democracy as the only internationally legitimate form of government, may signify that the sort of sultanistic regime we have described is a thing of the past, at best viable in a few out of the way places, like Equatorial Guinea, that the world overlooks.
We close by stressing again the destructive legacy of sultanism at the national level. The corruption of society and the illegitimacy of individuals and institu-
48     •     COMPARATIVE STUDIES
tions mean that these countries lack the bureaucracies, police, and armies that a state of law and a democracy need. Because habits of violence, distrust, and lack of social solidarity pervade these unhappy nations, the transition to democracy or even to stable authoritarianism will not be easy. There lies the biggest challenge both to democrats in these countries and to outside forces that want to encourage democracy in the Third World.
3
Paths out of Sultanistic Regimes
Combining Structural and Voluntarist Perspectives
Richard Snyder
Sultanistic regimes command great interest both in the literature on social revolutions and in the literature on regime transitions. Students of revolutionary change are interested in sultanistic regimes because so many are linked to the rare phenomenon of social revolution. For example, the sultanistic regimes of Shah Mohammad Reza Pahlavi in Iran, Anastasio Somoza Debayle in Nicaragua, and Fulgencio Batista in Cuba, all of which were toppled by revolutionaries, have been analyzed comparatively by numerous scholars seeking to specify the causes of revolution. These three cases have led students of revolution to identify sultanistic dictatorships as one of the regime types most vulnerable to revolution.'
Students of regime transitions, on the other hand, have been drawn to the set of sultanistic regimes because it contains numerous nondemocratic holdouts—regimes that have resisted the wave of democratization that has swept the globe during the past two decades.2 Cases such as Haiti and Zaire, where democratization has stalled or been stillborn, are of interest because studying them can shed light on factors at work in failed transitions.
Successful transitions to democracy such as occurred in the Philippines after Marcos pose puzzles for students of both revolutions and transitions. The combined presence of a sultanistic regime and a powerful revolutionary opposition in the Philippines under Marcos make it a "most likely" case of revolu-