Monday, January 1, 2007

Dictatorships Often Survive with Local Support

STANFORD GRADUATE SCHOOL OF BUSINESS—All dictatorships are cruel and wasteful. They deprive the populace of basic rights while enriching a small minority at the expense of rational development.

Some maintain power through terror and the naked use of force applied by huge standing armies and police forces. Paradoxically though, many dictatorships survive with the support of a significant swath of the populace. In sub-Saharan Africa, for example, rulers such as Kenya's former President Daniel arap Moi maintain power by exploiting ethnic and regional differences via a policy of selective economic rewards and privileges.

In effect, the deep ethnic divisions found in much of Africa are life insurance policies for dictators.

African dictators not only deprive their subjects of human rights, they loot their countries so openly that they have become known as "kleptocracies." Efforts by the developed world to aid these economies are often sabotaged by corrupt bureaucracies that siphon off a huge percentage of external development money. Uganda, for example, at one point derived roughly 20 percent of its budget from foreign aid, without noticeable benefits to the population at large.

In a groundbreaking study, Gerard Padro i Miquel, lecturer at the London School of Economics, demonstrated that African dictators distort their economies and steal foreign aid as the means to buy support from selected segments of the populace. He is a former faculty member at the Stanford Graduate School of Business.

Most academic work on dictatorships has focused on the use of violence and repression. But Padro argues that leaders of countries with sharply defined ethnic differences, such as Kenya, could probably not stay in power without the support of their own ethnic group. Although dictators like Moi may not be terribly popular, given the corrupt and repressive nature of their regimes, members of their ethnic group fear that a new leader from a different ethnic group would be much worse for them. "The fear of falling under an equally inefficient and venal ruler that favors another group is enough to discipline supporters," Padro says.

"If they [members of the dictator's ethnic group] decide to oust him, they face a chaotic succession process in which they cannot guarantee the next leader will belong to their group." Thus the dictator can exploit his own people with little fear that they will move to depose him.

Even so, members of the dictator's group do receive certain privileges, often in the form of patronage or public works projects. This is one reason why countries run by dictatorships often labor under the weight of a bloated public payroll and why so many development projects never succeed.

"The use of public money in the form of bureaucratic posts, infrastructure or even access to schools as a form of patronage, as well as the ethnic bias in the allocation of these goods, has been widely documented in Africa. The Gikuyus and later Kalenjin groups in Kenya, northern groups in both Nigeria and Uganda, or Tutsis in Burundi are just salient examples that have reproduced across the continent," says Padro.

Privileges handed out to favored groups must be paid for, of course. You might think that a dictator would settle the bill by taxing the out-of-power ethnic group and not his own. In fact, says Padro, the cost of supporting a dictatorship generally leads to across-the-board tax increases, since the dictator's group has gotten something for its money and won't protest, and the other ethnic groups have no choice in the matter. This strategy allows the dictator to extract a maximum amount of money from the whole country while spending money on just one group.

Not surprisingly, taxes, both direct and indirect, are levied unevenly and used as a tool to placate supporters. Uneven spending damages the economy since the government chooses which sectors to aid on the basis of political favoritism, not sound policy. "In Ghana and Uganda, among other examples, the coalition that supported the leader extracted resources from the coffee and cocoa planters—both crops that involve substantial specific long-term investment. On the contrary, in Kenya the Gikuyu controlled the coffee-growing parts of the country, and hence the discrimination against these crops was much less evident," writes Padro.

It's worth noting, though, that the actual benefits of ethnic favoritism are not spread evenly. Although Padro does not use the word class (he prefers the terms elites and non-elites), it's clear that the wealthy and upper-class members of the dominant ethnic group receive so many privileges that part of the money to pay for them is extracted from the pockets of the poor and working-class members of the same group.

Padro says the methodology of "The Politics of Fear" is fairly standard; he built a mathematical model based on a wide reading of case studies and other secondary sources. But because of the continued instability in East Africa it was not possible to visit at the time he was conducting his research. He points out that mathematical models in the social sciences test ideas for logical consistency, but do not necessarily prove the truth of a particular conclusion. Moreover, few dictatorships keep accurate records of illicit revenue and expenditures.

[Interestingly, the government of Peru did keep detailed records of bribes to politicians, an error that was instrumental in bringing down the government of Alberto Fujimori, as detailed by Professor John McMillan and Pablo Zoido of the Graduate School of Business. 

For those who think that democracy would solve the problems of ethnically based dictatorship, Padro has this somewhat discouraging conclusion: "Finally, note that no difference is made between democracy and dictatorship in the model. The evidence from Africa shows that democracies have not behaved differently than dictatorships when supporting kleptocracies and corruption."

Related Information

"The Control of Politicians in Divided Societies: The Politics of Fear," Gerard Padro i Miguel, NBER Working Paper 12573, October 2006.