the assumption of responsibility for the welfare of the world











2010 April 12


The existence of high-repression, high-loyalty régimes in the work of Ronald Wintrobe


Though totalitarian régimes may have existed in prior times, the term ‘totalitarian’ and its usual applications belong to and help define the twentieth century — Nazi Germany, Fascist Italy, Soviet Russia, Maoist China. A political theory of the present day must account for them, and those who believe totalitarianism is a distinct phenomenon must mark off its boundaries. Ronald Wintrobe seems convinced, and adopts the distinction for his own account of non-democratic régimes.1 Totalitarian régimes are those that (as is usual in the understanding of the term) involve themselves in the totality of their subjects’ lives; in Wintrobe’s theory, these are also régimes that simultaneously exert great repression and secure great loyalty from their subjects. How is it possible to achieve this congruence? What is the nature of this loyalty?


Wintrobe’s model is explicitly for unelected governments, and considers that each ruler provides for his continued rule through a mix of repression and loyalty. Though he calls them “instruments”, on the surface they are better characterized as conditions that a non-democratic government is attempting to create. He makes clear what he has in mind for the first:

By repression I refer to restrictions on the rights of citizens to criticize the government, restrictions on the freedom of the press, restrictions on the rights of opposition parties to campaign against the government, or, as is common under totalitarian dictatorship, the outright prohibition of groups, associations, or political parties opposed to the government. To be effective, these restrictions must be accompanied by the monitoring of the population and by sanctions for disobedience. [p33-4]
The last sentence is particularly worth emphasis. The restrictions themselves, while illiberal, might seem innocuous — the citizen is not, prima facie, deprived of food or shelter, nor placed in any other life-threatening situation, nor confined to prison. But the necessary sanctions can very well do all of these things. The sum of effective repression is that citizens are forbidden to oppose the government, by word, association, or other deed, the government will be monitoring to ensure that they do not, and will punish those who do. How they will punish the disobedient Wintrobe leaves at this point to the reader’s imagination.

Loyalty is not so directly defined. But after offering a primary example, the distribution of rents, a practice which he says is much more prevalent in dictatorships than democracies, he proceeds to say that “there are other, more subtle strategies available to a dictator to gain the trust of his or her subjects” [p36]. Wintrobe discusses the strategies of Augustus to build loyalty — “Although he was not averse to creating fear” [p37, footnote]. Grammatically this implies a contrast, and must be taken as evidence that loyalty is not the result of fear. In any case, the most important evidence as to the nature of loyalty is that it is not repression. Wintrobe chooses them as different parameters.


This brings us to Wintrobe’s graph of ruling strategies, with a horizontal axis of loyalty, and a vertical axis of repression, and the problem of its two intersecting curves. The magnitude of each parameter must be taken, over the entirety of the theory, as an abstract amount of repression or loyalty. In earlier parts of the theory, though, this amount can be taken as a matter of emphasis; though it is not stated explicitly, it is possible to interpret the axes as denominated in some form of currency, such as purchasing-power parity. The more resources the dictator places into repression, the higher its value on the graph. The more resources the dictator places into loyalty, the higher its value on the graph.2

In the first of Wintrobe’s general curves, power or π, loyalty decreases as repression increases. This represents the power state of the dictator; he has the combined power resulting from the two values, from the particular combination of repression and loyalty indicated by any point on the curve. Any attempt to increase loyalty will necessitate a decrease in repression, and any attempt to increase repression will result in a decrease in loyalty. It is as fair to say, from the dictator’s point of view, that any increase in repression will allow a decrease in loyalty, will allow the dictator to care less about and spend less on loyalty.

A special case of the curve, minimum power or πmin, represents the minimum of all combinations of repression and loyalty that will keep the dictator in power. The space underneath the curve represents the totality of repression-loyalty combinations that will be insufficient to keep the dictator in power. Thus, any decrease in one parameter must be compensated for with an increase in the other, in order for the dictator to maintain power. This is consistent with the previous definitions of repression and loyalty, as well as with common sense: the reaction of the population to repression is not what is commonly called “loyalty”, not the equivalent of the reaction to payoffs or attempts to earn trust. Repression does not earn trust; it breeds fear.


In the other of Wintrobe’s two general curves, supply of loyalty or Ls, loyalty and repression increase together, in direct proportion.3 He gets to this point by what can only be described as a sleight of hand.

Whichever form it [an increase in repression] takes, the risks of disloyalty among the citizens are increased, and the expected rate of return is diminished. Consequently, from the point of view of citizens and interest groups, the attractiveness of dealing with the opposition decreases, and the relative attractiveness of exchanges with the dictator, or with his or her representatives, increases. This substitution effect implies that a typical citizen’s loyalty — and hence the aggregate supply of loyalty to the dictator — will be positively related to the level of repression. [p48]
This suggests that loyalty is simply the willingness to do as the dictator wishes. The earlier definition of loyalty is the rough equivalent of love, this second of obedience, even fear. Fear does not decrease as repression increases; love does not increase as repression increases. We might ask why, if increasing repression increases loyalty, any dictator would ever bother to spend on loyalty alone, or why the power curve π looks as it does. Do citizens of those repressive states not realize the costs of disloyalty?

One result of the two intersecting curves is that the entire repression-loyalty space can be usefully identified, to Wintrobe at least, with a state of reality. The π curve transverses the states where repression is high and loyalty is low, where repression and loyalty are both low, and where repression is low and loyalty is high. The Ls curve transverses the states where both parameters are low and where both are high. Wintrobe refers to the dictators in each state by separate mnemonics: a tyrant rules through high repression and low loyalty, a tinpot rules through low repression and low loyalty, a timocrat rules through low repression and high loyalty, and a totalitarian rules through high repression and high loyalty.


The π curve is a reasonable summation of possibilities, and therefore timocracy4, however it is to be labelled, does have a theoretical space to occupy. But it is, even to Wintrobe, an exceedingly-rare occurrence. He supposes that perhaps only the “good emperors” of Rome were timocrats, and in this he relies heavily on the history of Edward Gibbon. If we reached for examples in the modern world, we would be forced to look to the seemingly-popular régimes of “managed democracy”, to use the Russian term of art; Russia itself would be an example, and Singapore5 would be another. The personal corruption of Vladimir Putin or Lee Kuan Yew and his associates6 is not generally an issue, and both Putin and the People’s Action Party (PAP) of Singapore are probably genuinely popular with a segment of the population. But this is itself affected by repression. The government of Russia or the siloviki (“powerists”) loyal to Putin control television in Russia, and critics, including notably Mikhail Khodorkovsky, have been prosecuted. Critics of the PAP find themselves technically in civil court being sued for libel (consider for example the account of J.B. Jeyaretnam), but the result (including the effect of self-censorship) is the same. Citizens of Russia and Singapore are exposed to a highly-biased appraisal of their governments, and the space for opposition is heavily circumscribed.

The tinpot is defined by his pecuniary motive, and might thus better be called a kleptocrat. He seeks wealth (for himself and his protégés), and thus wishes to spend as little as possible on repression and loyalty, so as to keep any remaining government resources for himself. He keeps himself on the πmin curve to ensure his continuation in office, and at the lowest valence level on that curve (closest to the origin), to spend as little as possible on non-consumption. While the notion of the tinpot being minimally repressive is a relative and debatable assessment, the pecuniary motive is easily enough identified, with Suharto and Sani Abacha being good recent examples.

Tyrants are equally easy to imagine and identify. There are many highly-repressive states in the world, and surely always have been. Why would the tyrant exhaust resources on repression if there is a lower-repression point on the πmin curve which is therefore cheaper, and leaves more state resources for consumption? One possibility is that suggested by Mark Bowden in the case of Saddam Hussein, namely vanity.7 Saddam’s grandiose vision of his role in Iraqi and greater Arab society and history required the complete absence of apparent opposition; in some ways, the presence of fear (“respect”) was as good to him as love, and lack of open opposition could certainly be interpreted in some circumstances as love, whether it is or not.

But Saddam, though a tyrant, would fit the traditional definition of a totalitarian ruler. The degree of control of the Iraqi state exercised over individual lives was as near to total as the term can imply (consider the descriptions provided by Bowden, Kanan Makiya8, and to some extent Marion Farouk-Sluglett and Peter Sluglett). Had Saddam, through his dramatic repression, convinced the Iraqis to be loyal? To parse Wintrobe’s position on this, we can look at his own discussion of perhaps the exemplary totalitarian state, Soviet Russia. Wintrobe turns to Jeane Kirkpatrick, who famously9 uses ‘totalitarian’ for a state along the Stalinist model, and asks whether totalitarianism works as intended. Here I quote Wintrobe’s own quote, as well as his reaction:

She writes, “But have they [the Politburo leaders] managed to educate Soviet citizens so that they would freely choose to behave according to the norms of Soviet culture if the constraints of coercion were removed? The answer of course is that we don’t know” (emphasis in original, p. 123).
      After the fall of communism in the Soviet Union and Eastern Europe, it appears that we do know the answer to this question, and it is “no.” [p45]
Wintrobe immediately claims that this answer does not mean that totalitarian rulers cannot obtain their subjects’ “loyal support, which is quite a different matter” [p45]. But this is not convincing.


The only way to make sense of Wintrobe’s various arguments is to accept that he is using ‘loyalty’ in two different ways. In one, loyalty is somewhat akin to love, and can be bought, or earned through trustworthy acts. In the other, loyalty is obedience, and can be obtained through fear of repression. While the first is perhaps more like the common understanding, a case could be made for the second in the vernacular. But Wintrobe’s theory and scheme fall apart unless these two “loyalties” are the same thing. He needs a substitution effect and the loyalty-supply curve, Ls, to create a totalitarian space in his scheme. He needs loyalty to be both the love that is inherently lost when the dictator becomes more oppressive, and the obedience that is inherently gained when the dictator becomes more oppressive. The fact is that if loyalty is a voluntary attachment that is separate from repression (and thus merits inclusion as an axis separate from and perpendicular to repression), if obedience is the result of fear of sanctions which are in Wintrobe’s own words a tool of repression, then there can be no direct proportionality between repression and loyalty. In other words, there is no loyalty-supply curve, Ls. Hence, finally, there is no separate category of totalitarianism, with high repression and high loyalty.


With the practical absence of the timocrat, it becomes possible to examine non-democracies primarily along the axis of repression, but following the power curve, π. Some autocrats, such as Augusto Pinochet, can provide or be perceived to be providing sufficient public goods — in his case, economic growth, political stability, traditional values and nationalism, and resistance to Marxism — to please a segment of the population, earn limited but genuine loyalty, and reduce the need for repression.10 Though it is not necessarily related to the power curve, some dictators expend the least resources possible to stay in power, minimizing expenditures on repression and loyalty alike, in order to retain maximum resources for personal use; Wintrobe’s tinpot is a genuine archetype of the political world.

There are some régimes instead whose practice is of high repression. This may be power for power’s sake — call it totalitarian, perhaps — or it may be vanity, as in Saddam’s case. It may result from poor economic management or even bad economic luck, reducing the possibility of loyalty, raising the likelihood of dissatisfaction (as in present-day Zimbabwe) and giving the régime no other option to maintain power. It may simply result from a slippery slope: one moment of overuse of repression yields increasing dissatisfaction, requiring more repression, yielding more dissatisfaction. It is along this path that numerous régimes have fallen, such as those in Eastern Europe and arguably the Soviet Union itself. Given the thaw of the Gorbachev years, the sudden application of repression to prevent the fall of various Communist Parties or the independence of various Soviet republics may have shocked society to the breaking point, with a populace unwilling to accept the sudden repression, or security forces unwilling to apply it.

The balance of repression and loyalty, then, is a useful way of examining the means by which non-democracies retain power. This is so, though, only if obedience from fear is understood as a consequence of repression, and loyalty is understood as something else, the voluntary devotion to a régime that is seen to provide public or private goods. We can see in this balance the choices régimes make, the outcomes they value. But there is no practical symmetry of the repression-loyalty space. Non-democracies do not generate high loyalty in real-world examples, and they cannot create a state of high repression and high loyalty even in theory.


1. It may also be that his desire for alliteration in category names constrained him to the word; but his reference to other uses, as I will discuss, suggests that ‘totalitarian’, at least, was chosen for reasons beyond poetry.

2. When the loyalty-supply curve, Ls, is introduced, this interpretation is somewhat more complicated, as discussed in the relevant footnote.

3. Here the amounts of repression and loyalty cannot be a matter of emphasis, though imagining them as denominated in purchasing-power parity is still reasonable: at a given level of repression (judged by expenditure), a given value of loyalty (judged by the expenditure it would otherwise cost), is freely available to the dictator.

4. Wintrobe: “the Greek root of the word “timocracy” is Thymos — to love.” [p80] The spelling alone should indicate that this is false. ‘ΘΥΜΟΣ’ (t‛umos) is a noun, and means “soul” or “heart” (identical for the Ancient Greeks), and by extension, “passion”. ‘Timocracy’ is actually from ‘ΤΙΜΗ’ (timē), meaning “honor”.

5. I treat Singapore here as a collective régime led by Lee Kuan Yew. That was certainly the case during his premiership; his role during the official premierships of Goh Chok Tong and Lee’s son Lee Hsien Loong is opaque, but his current official role of “Minister Mentor” sounds not unlike an important role he played while prime minister, according to a sympathetic account by Raj Vasil. Whether Goh or the younger Lee would ever contradict the private urging of the elder Lee is to me rather doubtful, though each (and the other ministers throughout independence) surely had significant responsibilities.

6. Putin’s associates are not included in this specific statement.

7. Saddam and his family and associates were rich in absolute terms, but not so relative to the natural wealth of Iraq.

8. ‘Republic of fear’ was published, in its own words, pseudonymously; I accept that the proper attribution is to Makiya.

9. Kirkpatrick’s ideological motivations are equally famous; there is an obvious political convenience to possessing a theoretical distinction between the dictatorships that the US allied with in the Cold War and those that it fought.

10. This is not to deny, as his actual supporters did, that Pinochet was repressive. His opponents certainly spent time in jail — those who were not “disappeared” into the sea.


Mark Bowden. ‘Tales of the tyrant: the private life and inner world of Saddam Hussein’, p35+ in the Atlantic Monthly, v289 n5 (2002 May).
Edward Gibbon. ‘The history of the decline and fall of the Roman Empire.’ Strahan and Cadell, 1776-89.
J.B. Jeyaretnam. ‘The hatchet man of Singapore’. Jeya, 2003.
Jeane Kirkpatrick. ‘Dictatorship and double standards: rationalism and reason in politics’. Simon and Schuster, 1982.
Kanan Makiya (Samir al-Khalil). ‘Republic of fear: the politics of modern Iraq’. University of California, 1989.
Marion Farouk-Sluglett and Peter Sluglett. ‘Iraq since 1958: from revolution to dictatorship’. KPI, 1987.
Raj Vasil. ‘Governing Singapore: a history of national development and democracy’. Allen and Unwin, 2000.
Ronald Wintrobe. ‘The political economy of dictatorship’. Cambridge University, 1998.



Home of the Stewardship Project
and O.T. Ford