What can we learn from successful autocracies?

Tim Besley, Masa Kudamatsu 05 July 2007

a

A

One of the most striking economic phenomena of the past twenty years is growth and development in communist China.  Rates of growth in income per capita of around 10% per annum have led to one of the largest falls in absolute poverty that the world has ever seen.  But, while China has embraced many aspects of the market, it has resolutely opposed most aspects of democracy.

But there is another image of autocracy, typified by the recent experience of Zimbabwe, where a despotic leader (Robert Mugabe) seems content to preside over economic chaos, standing coolly by while millions of people descend into destitution.  There is no way of getting rid of a leader like Mugabe who rigs elections and represses the opposition in order to stay in power.  This is government failure on a vast scale.

A brief look at the data on growth in democratic and autocratic regimes shows that the stories of China and Zimbabwe typify the experience of economic development among autocracies in the post war period.  Economic growth rates differ more substantially among autocracies than among democracies. This is illustrated in Figure 1 which depicts the distribution of growth performance in autocracies and democracies that survive for five years or more.  Successful autocracies outperform democracies at the top of the distribution.  However, they perform worse at the bottom.

 

Figure 1: Economic Growth Distributions among Democracies and Autocracies

From a scientific point of view, this raises the question of whether we can develop a systematic account of why some autocracies perform better than others.  At a superficial level, it is clear that we are lumping together a wide variety of regime types under the rather blunt heading of “autocracy”, such as military dictators and former democrats who then cancel or subvert election. 

Our approach to answering this question focuses on how political institutions make political leaders accountable, or make their survival in office depend on their policy performance.  This is a long-standing theme in political economy.  Moreover, a central role of regularised democratic elections has been to create some feedback from good performance by government to continued hold on power, a link which is missing in autocratic government.

Autocrats typically rely on key groups to stay in power.  This could be a party structure or the military or a close group of allies.  Concerted action by this group may, in principle, depose the leader.  Following recent work, we refer to this group as the “selectorate” as opposed to the “electorate” in democracies. 

In the absence of elections, it is important to focus on what incentive this group has to support leaders that foster good policies.  As citizens, this group will tend to benefit from economic prosperity.  But, equally they are likely to enjoy some trappings of power by having allied themselves with the leader.  If deposing the leader threatens these benefits, they may be reluctant to do so.   This leads us to predict that secure selectorates will tend to create performance-related incentives for their leaders and would be willing to remove the leader from office if he does not perform well.  Selectorates whose own hold on power is closely tied to a specific leader are willing to preside over bad policy.  But to remove the leader from power, the selectorate has also to be strong enough to resist any attempts at repression.  The history of autocracy is replete with examples of despotic leaders purging their closest allies.  Given that his inner circle has the most credible threat against the leader, this is not surprising.

Extension of this logic also reveals why democracy is not always able to achieve the best policy outcome.  Effective democratic accountability requires that the salient issues on which elections are fought are those that promote general interests.  Democratic accountability may founder when factional rather than general interests influence election outcomes. Today’s Iraq may be an example of this.  Successful autocratic accountability can still work in such circumstances as long as one faction is able to hang on to power and discipline bad leaders.

The logic of this argument suggests three things that we should observe in reality. First, we should find that successful autocracies are those where the selectorate can and will exercise control to discipline poor performance.

In communist China, the selectorate are around 20 members of the Communist Party Politburo. When Hu Jintao succeeded Jiang Zemin as General Secretary in 2002, Jiang had reportedly tried to convince his colleagues in the Politburo to allow him to stay in power. But he failed, probably because the selectorate were concerned that rising income inequality in China under Jiang’s rule would undermine the overall economic success.

Another example comes from the Brazilian economic “miracle” in the late 1960s and the early 1970s under the military dictatorship. Members of the armed forces, the selectorate of the military regime, always picked as the next president the person behind whom the military could be united. The succession in 1967 resulted in a new president whom the predecessor tried to prevent from assuming office. One factor behind this succession struggle was that the armed forces were disgruntled with the incumbent’s economic policies. The Brazilian economy took off under the new president’s economic management.

Second, if our view of successful autocracy is correct, we should find that high growth autocracies have higher leadership turnover than low growth autocracies.  This turns out to be the case.  The probability of leadership change in a given year is 13% for high growth autocracies and 7% for low growth ones.

Finally, we can also exploit “natural experiments” when leaders die in office to see what happens.  We should find that successful autocracies are those where such random deaths do not undermine the power of the selectorate whereas poorly performing autocracies will experience change in the ruling elite.

The incapacitation of Portuguese dictator Oliveira Salazar in 1968 provides one such “experiment”. Under Salazar, Portugal experienced a rapid economic expansion. The succeeding dictator, Marcello Caetano, was appointed by the military and its civilian cooperators who were the supporters of Salazar. Caetano then managed to stay in power until 1974 without slowing down the economic growth rate. Similar regime stability was observed after the death of a Thai military dictator, Sarit Thanarat, in 1963 under whose rule Thailand’s economy took off.

On the other hand, the death of Ahmed Sekou Toure, a civilian dictator of Guinea in west Africa, in 1984 was immediately followed by a military coup toppling the regime. The Guinean economy had been shrinking under Sekou Toure’s rule. The Guinean selectorate apparently feared that getting rid of Sekou Toure would lead to the loss of their own power.  The selectorate in today’s Zimbabwe probably fear the same.

The aim of this research is not to argue for autocracy.  But there are lessons for institutionalising good government in a wider context.  To bring it closer to home, consider the recent experience of the UK where Gordon Brown has replaced leader Tony Blair without an election taking place.  The selectorate in this case was the ruling Labour party and their incentive was to pick the best leader they could find; they voted in Gordon Brown based on his record to date.  This serves as a reminder that, even in a mature democracy, accountability resides in more places than general elections.  In due course, there will, of course, be an election in the UK at which the new Prime Minister will have to win a fresh mandate.  However, the role of party leadership selection is an important supplementary force to achieve accountability and is a characteristic of Parliamentary systems.  Our analysis of successful autocracies shows that accountability by insiders becomes even more important in the absence of open elections.  This is also a challenge in one-party democracies like modern day South Africa.

The lesson from successful autocracies is that supporting good government with effective accountability requires institutional arrangements appropriate to that task, given the circumstances of the country in question.  Efforts to build effective governments that work in the interests of a broad range of citizens need to be cognisant of this and to identify the institutional features necessary for the task.

 


This commentary is based on “Making Autocracy Work” by the same authors available as CEPR Discussion Paper No. 6371 at http://www.cepr.org/pubs/new-dps/dplist.asp?dpno=6371

a

A

Topics:  Politics and economics

Tags:  China, autocracies, Zimbabwe, selectorate

Comments

"Extension of this logic also reveals why democracy is not always able to achieve the best policy outcome.  Effective democratic accountability requires that the salient issues on which elections are fought are those that promote general interests.  Democratic accountability may founder when factional rather than general interests influence election outcomes. Today’s Iraq may be an example of this.  Successful autocratic accountability can still work in such circumstances as long as one faction is able to hang on to power and discipline bad leaders".

Tim,

This is a bloody-minded argument. In the real world, the 'best policy outcome' is not necessarily the growth optimizing outcome. That's why we have democracies and would like to have—but often don't have much—parliamentary accountability for distributive economic policies like trade policies. Democracies exist to validate these compromise distributions; even 'bad' ones.

Alas, many democracies including Australia where I live, are really mixed systems in which parliament is so dominated by the executive—thanks to rigid party discipline—that they're similar in practice to the autocracy/selectorate model except in the immediate vicinity of an election (there are other important differences, of course, like a functional judicial system).

Which makes me wonder if, in fact, your two cases are well distinguished after all. You seem to say that the accountability function that differentiates democracies and autocracies takes place at 'turn-over' (your sang-froid is exemplary). But I say that in reality the 'accountability' function is constant throughout the life of the regime and is, perhaps, not as different between the two cases as your model implies.

Best wishes,

Peter
www.petergallagher.com.au

Along the lines of Bryan Caplan’s new book, "The Myth of the Rational Voter" I wonder if enlightened autocracies might be able to implement better policies than democracies? This could help explain the upper tail.