Experts routinely tout democratization as the key to promoting freedom and prosperity in underdeveloped nations. They argue that making leaders accountable to their citizens would promote good governance and remove the institutional barriers to economic development. Adherents of this position cite a large number of empirical studies, which show that democratic countries tend to perform better than autocracies across a variety of well-being indicators.
Development agencies and scholars therefore give democratization high priority relative to other anti-poverty programs. But these same experts completely disregard alternative governance models, such as radical decentralization. Thus, it seems everybody knows democracy is the best way to promote robust economic development, so the challenge is in finding the best way to promote democracy.
The way scholars define and measure democracy, however, includes a bias. This bias prevents a fair evaluation of the alternatives. That is, if we want to know whether promoting democracy in failed or authoritarian states is a good idea, we need to treat democracy as a set of institutional inputs analytically distinct from the effects of those institutions. Even critics offering a minimalist definition of democracy include outcomes in their definitions.
Both therefore stack the deck in democracy’s favor. Let me explain.
Under some definitions, democracy requires not only certain mechanisms for collective decision-making, but also liberal policy outcomes in various areas. While some conceptions of democracy are even more restrictive, the broadest commonly accepted definition in contemporary political science has four criteria:
- free and fair elections;
- close to universal adult suffrage;
- freedom of speech, association, and press; and
- elected officials not unduly influenced by unelected groups such as the military or religious leaders.
If any one of these conditions is routinely violated, the country is deemed undemocratic, or at least less democratic than countries that do meet the criteria.
Political Systems: Inputs, Dynamics, and Outputs
Broadly speaking, political systems can be defined at one or more of three levels: (1) institutional inputs, (2) political dynamics, or (3) policy outcomes. At the first level, we have the basic rules of the political game such as the electoral system and constitution. At the second, we have the interaction of political players within those rules—how voters vote and how parties and candidates compete with one another. At the third level, we have the policy decisions that emerge from this interaction.
“Thick” definitions of democracy reference all three levels: A democracy needs particular institutional inputs as well as certain patterns of electoral competition and policy outcomes. Minimalists insist the third level has no place in the definition of democracy, because it is an output. Such minimalists fail to recognize, however, that their own second-level definition is also an output rather than an input. The degree of competition in an electoral system cannot be directly controlled. Rather, it emerges from the interaction of politicians, voters, and special interests given the rules defined at the first level.
When we define democracy in terms of competition we make an implicit assumption that democracies are necessarily competitive. Since political competition is an intermediate goal of democracy, the minimalist conception of democracy picks out democracies that are at least moderately successful in a particular way.
More on Thick and Thin Conceptions
The Economist Intelligence Unit’s Democracy Index uses such a “thick” conception, which contains categories not only for the functioning of the electoral system but also for civil liberties and effective governance. A definition like this clearly allows us to say very little about the desirability of democratic institutions as a set of inputs. Countries with relatively little corruption and repression are likely to perform well on a number of other dimensions, but such a lack might have nothing to do with democracy’s institutional machinery.
Recognizing this problem, many political scientists have followed Joseph Schumpeter in defining democracy as a system in which collective decisions are made through a competitive struggle for votes. In Adam Przeworski’s words, a democracy is “a system in which parties lose elections.” Countries are deemed democratic if and only if there are somewhat competitive elections. Nominally democratic countries with rigged or otherwise uncompetitive elections are excluded, but there is no requirement of a free press or an autonomous legislature.
The widely used Polity IV database of regime type takes such a minimalist approach, considering only “key qualities of executive recruitment, constraints on executive authority, and political competition.” Here, there is no automatic assumption that democracies respect civil liberties or operate effectively.
While a minimalist—or "thin"—definition of democracy is far better than the alternative, it does not go far enough in defining democracy in terms of inputs rather than outputs. A competitive electoral system is not an institutional input, but one possible intermediate effect of democratic institutions. The mistake here is not quite as obvious as defining democracy in terms of policy outcomes, but it is hugely important for the practice of so-called “comparative institutional analysis.” In other words, if we’re going to compare sets of institutions in terms of their ability to improve overall peace and well-being, shouldn’t we exclude output biases altogether?
Consider what this means for comparative institutional analysis.
We want to know whether some failed state would be “better off” embracing democracy, autocracy, or anarchy. We look around at the performance of democracies as conventionally defined and see that they perform well on a variety of economic and social measures. The problem is that our very definition of democracy excludes many of democracy’s failures. It is not uncommon, for example, for a country with democratic institutions to become dominated by a minority faction able to prevent meaningful competition while retaining the institutions of democracy.
Venezuela has had democratic institutions since 1958. The Chavez regime limited political competition through force and fraud, making the country less democratic by conventional standards. Such was reflected in a decline in Venezuela’s Polity score, for example, as Chavez took power and Venezuela eventually shifted out of the democratic category altogether. Venezuela is no doubt a failed democracy when judged in terms of its political dynamics—the Fifth Republic and United Socialist parties had virtually no chance of losing their rigged elections—but this doesn’t make the country any less of a democracy in terms of institutional inputs. Democratic institutions allowed Chavez to gain power and limit political competition. That this can happen is an important fact to consider when thinking about whether democracy is a good idea for other countries.
If we want a fair comparison among systems, we need to define political systems in terms of their institutional inputs and nothing more. Some anarchists might claim, for example, that Somalia is not really anarchic because tribal groups have gained some territorial power; and some communists might claim that the USSR was not really communist because its rulers were insufficiently committed to the communist vision. These arguments commit the “no true Scotsman” fallacy in that they arbitrarily narrow the definition of a term in order to preserve a hypothesis in the face of conflicting evidence.
The claim that Venezuela is not really democratic commits the same fallacy in a subtler way, which generally goes unnoticed. Democracy cannot reasonably be defined as a system with genuinely competitive elections any more than anarchy can reasonably be defined as a system in which there is no coercion, or autocracy as a system with a wise and benevolent despot. An institution defined by its goals is virtually guaranteed to be successful under such a construal.
For democracy, the defining feature is an electoral system in which elected officials have the power to make laws and policies. Particular democracies will have additional rules designed to improve democratic performance, but these will always be formal rules with the potential to fail.
It seems unlikely that autocracy has a general advantage over democracy in poor countries, but the process of democratic transition is itself costly and should only be undertaken if the expected benefits outweigh these costs. Moreover, the evidence from Somalia suggests that statelessness might sometimes be a viable alternative to democracy. As Benjamin Powell writes in a recent Freeman article, Somalia’s “imperfect anarchy seems to be doing better than the very imperfect state that preceded it and many of those states it shares a continent with.”
Chris Coyne has called into question liberal democracy’s viability in failing states on the grounds that this political form depends on a number of informal institutions that cannot be designed from on high. I think the confusion regarding democratization runs even deeper.
Scholars and state-builders do not simply neglect the possibility that democratic institutions will not stick; they work with a definition of democracy that allows the most complete failures of democracy to be blamed on autocracy. This confusion gives the impression that democracy would promote freedom and development if only we could make it stick. In reality, democratic institutions simply produce poor outcomes, which sometimes don’t look particularly democratic. Recognition of this fact should force a re-evaluation of the humanitarian project of democratization and the desirability of institutional alternatives such as anarchy.