Until recently, an “oil democracy” seemed almost an oxymoron. The Middle East, where oil supplies are concentrated, was uniformly autocratic, and most other oil-rich countries were either autocracies like Indonesia or stunted democracies like Mexico. Things are changing, however, from two directions. Democracy is spreading to some oil economies, and oil is spreading to some democracies.
The spread of democracy to the oil economies in the Middle East is an explicit item on the U.S. foreign policy agenda. In other regions, the democratization of important oil economies has occurred in recent years without external pressure, notably in Indonesia, Mexico and Nigeria. The spread of oil to democracies, meanwhile, is in large part a side effect of the attempt to diversify U.S. and Western oil supplies from dependence on the Middle East. New discoveries have been made and are being exploited in a range of low-income democracies such as Gambia, São Tomé e Príncipe, Senegal and East Timor.
These complementary developments represent both a warning of danger and a herald of opportunity: a warning that oil riches might harm existing democracies, and the opportunity not only to prevent that from happening, but to learn how to reduce the perverse effects of oil rents on democratizing countries.
It is by now commonplace that corrupt governments have misused rents from oil and other natural resources, and that these rents have in turn fed corruption and undermined sound governance. The explanation for this proceeds in three phases. Undergraduate-level economics provides the first phase: Resource rents undermine the competitiveness of other exports by inflating the real exchange rate—a phenomenon termed “Dutch disease” because of the effect of Dutch gas exports on its manufacturing industry. In the second phase the explanation notches up to graduate-level economics: Worries about the level of the real exchange rate are replaced by worries about its volatility. This concern fits neatly into a larger debate among professional economists and finance ministers about how volatility might affect investment. In the third phase the explanation shifts to the recent fusion of economics and political science: Resource rents divert politics from the delivery of public goods to the delivery of private patronage. Yet patronage can be held in check by well-functioning, accountable institutions. This shift to a focus on governance implies that, since democracy enhances accountability, new oil democracies will handle resource rents better than autocracies. On the face of it, this ought to hold true in both directions: for oil-rich countries that manage to democratize, and for democracies that discover and exploit new fossil fuel resources.
Of course, the three phases describing how resource rents distort political processes are not mutually exclusive, nor does any one pattern among them fit all cases. Institutions in resource-rich societies vary significantly, and although resource riches have certainly affected them, most countries got their institutions before they discovered their most valuable tradable resources. Since the global variation in institutions is well reflected among those with plentiful resources, this enables us to tease out statistically how political institutions interact with resource wealth. This, in turn, allows us to predict how prospective oil democracies will fare.
The evidence so far suggests that democracy will not necessarily or easily resolve the problems posed by resource wealth. Indeed, democracy is far more likely to compound such problems if resource-wealthy democracies do not acquire a distinctive governance design. Specifically, the mechanisms that determine how power is achieved need to be rebalanced against those that determine how power is used. Put a little differently, electoral competition should be de-emphasized in oil democracies, and political restraints (checks and balances, as Americans like to say) should be stressed. Unfortunately, this is precisely the opposite of what we are likely to see in new oil democracies, and it is a trend that the international media’s fascination with telegenic elections is liable to inadvertently encourage.
The Awkward Facts
If one takes the trouble to do the statistical research on this question, one finds a consistent pattern: Rents from oil and other natural resources are particularly ill-suited to the pressures generated by electoral competition, absent checks and balances on the uses of political power. In the absence of natural resource rents, even electoral democracies with fewer checks and balances significantly outpace autocracies economically. This is an encouraging gloss on the economic consequences of democracy, for the usual social science assessment is that democracy has no net effect on economic success.
However, in the presence of large rents from natural resources, the economic performance vectors are reversed: Autocracies outperform democracies, and the difference is quite large. In the absence of natural resource rents, a fully democratic polity outperforms a despotic autocracy by about 1.5 percentage points of growth per year. By the time natural resource rents rise to about 8 percent of GDP in the absence of checks and balances, the growth advantage of democracy is eliminated. Beyond this the net economic effect of unrestrained democracy is plainly adverse. Take a country with no check and balances, and resource rents worth 20 percent of GDP (a pretty modest level of resource wealth): The switch from autocracy to intense electoral competition lowers the growth rate by about 2.5 percentage points.11.
For details on methodological issues, see Paul Collier and Anke Hoeffler, “Democracy and Resource Rents”, available at users.ox.ac.uk/~econpco/research/politicaleconomy.htm.
How does democracy undermine a country’s ability to harness its resource rents for economic growth? The answer obviously concerns economic policy choices. But which choices?
One possibility is that resource rents bloat the size of the state. Economists have established that higher public spending correlates with a lower growth rate. However, when we control for the level of public spending, we find that a smaller public expenditure does not reduce the adverse effect of democracy on the use of natural resource rents. Thus, government spending can’t explain why resource-rich democracies underperform.
What about the composition of expenditure? The most basic influence on economic growth is investment. If we control for the share of investment in GDP, the adverse effect of democracy is reduced. In other words, resource-rich democracies underinvest. Other researchers find that democracies in general underinvest. In resource-rich societies, investment is particularly important, since this is how resource rents could be transformed into sustained increases in income.
However, the real story turns out to be not the rate of investment, but the return on investment. Resource-rich democracies not only underinvest; they invest badly. But why?
Why the Politics Goes Wrong
To see how democratic politics goes wrong in the context of abundant resource rents, we must unbundle the concept of democracy. The design of any given democracy obviously varies considerably—presidential versus parliamentary systems, and proportional versus first-past-the-post voting, and the like. However, more important than these design aspects is the central distinction between mechanisms that regulate how power is achieved and mechanisms that limit how it is used. The former are concerned with electoral competition, the latter with restraints on governing, and democracies differ considerably as to which mechanisms are more prevalent. As it happens, the abundance of resource rents seems to alter how electoral competition is conducted far more than it alters how power is used. Essentially, resource rents let in the politics of patronage.
Electoral competition motivates political parties to attract votes in the most cost-effective manner possible. In normal circumstances, most parties in power do this by delivering public goods such as infrastructure and security better than their rivals. The extreme alternative to public-goods politics is the politics of private patronage: Select voters are bribed with public money. In its crudest form, candidates deliver cash in hand for ballots in the box, but there are far more potent means of vote-buying than that.
Obviously, if every voter has to be bribed, a party offering genuine public goods will out-compete a patronage party, because the blessing of public goods is that they can be used equally by all. Voters are far better off having public money distributed to them in the form of public goods than as private bribes. Patronage starts to look cost-effective for a political party, however, if votes can be bought wholesale by bribing just a few critically placed opinion leaders. Then the very universality of public goods starts to look like a disadvantage.
The fact is that people often rely on “opinion leaders” in making decisions rather than on the evidence of their own eyes. Such behavior is described in the theory of information cascades, which is a fancy social science way of describing some forms of herd behavior. A classic example of herd behavior may be found in the world of fashion. It is entirely rational for fashion companies to focus their marketing efforts on opinion leaders, knowing that consumers will imitate them. Voting sometimes resembles fashion in this regard. People often vote in blocs under the guidance of a local opinion leader, especially in places where ethnic or sectarian loyalties are strong, or where good information is scarce. These conditions obtain in many developing societies.
Suppose, then, we accept that in ethnically differentiated societies without a vibrant press, patronage politics is more cost-effective than the provision of public goods. This still leaves open the question of why this problem disproportionately affects resource-rich societies. After all, many societies are ethnically differentiated and have limited press freedoms.
The answer seems to be that patronage politics is a more cost-effective way of attracting votes than providing public goods only if the ruling party can subvert enough public funds. To finance large-scale patronage, a government first needs to embezzle public money out of the official state budget and move it into slush funds. If the restraints on embezzlement are sufficiently tight, then patronage politics is not feasible. Such restraints both influence and are reinforced by the purity of politicians’ motives for seeking power. A system in which patronage politics is not feasible is more likely to attract politicians genuinely motivated by a vision of the public good. In societies in which patronage is feasible, democratic politics tends instead to attract crooks in search of easy money.
Resource rents weaken political restraints for an obvious reason: They radically reduce the need to tax. Because resource-rich governments have less need to tax, they tend not to provoke citizens to scrutinize their dealings as closely.
While this process of accountability has been understood in its basic form for a long time, it has usually been proposed as an explanation for the autocratic tendencies of resource-rich societies—and rightly so. The key point here is that this same process of undermining accountability operates within electoral democracies, too. The effect does not damage electoral competition itself, but the political restraints on how power is used. As resource rents gradually erode checks and balances, assuming they were present to begin with, parties are freed up to compete for votes by means of patronage. As we have established, this is the most cost-effective way to attract votes in an ethically divided society with an unfree press. Parties that quixotically choose the public-goods route of appealing to voters simply lose the election: Political crooks will beat honest candidates every time in such situations. If you want a real-time measure of the problems faced by an honest and able candidate in an oil democracy, check how well Pat Utomi does as a presidential candidate in the Nigerian elections this April. The law of this particular political jungle is, unfortunately, the survival of the fattest.
Political scientists can measure a country’s political restraints on power. As with all quantification, this measure has its deficiencies: It simply counts how many of 17 possible checks and balances are incorporated into a political system. These restraints correlate with the real level of democracy. Again, the distinction between electoral competition and checks and balances is critical. When a nation achieves a high score on the democracy scale because of intense electoral competition and a low score for restraints, economic growth suffers severely. Conversely, with powerful political restraints, an oil democracy can be an economic success: Thus, Norway is fine, but Nigeria is in trouble.
What do “sufficient restraints” mean in practice? Let’s return to our example of a country with resource rents worth 20 percent of GDP and intense electoral competition. Recall that in the absence of checks and balances, such a society has on average a growth rate 2.5 percentage points lower than an autocracy. On the checks and balances measure we use, this society would need only four of the 17 restraints in order to eliminate this disadvantage. With eight restraints, it would outperform the autocracy by 2.8 percentage points.
Four restraints out of 17 doesn’t sound like a lot, but it is double what resource-rich societies typically have. Obviously, too, just adding any four checks from the 17 would not necessarily fix the problem. Indeed, if any single restraint is supremely important, it is a free press. Without it, citizens can readily be manipulated by government propaganda. I have used a quantitative measure developed by Freedom House to investigate whether press freedom matters in resource-rich societies. I found that not only is a free press generally associated with a faster growth rate, but also that the effect is significantly larger in the context of resource riches.
I can also show why political restraints are so important for resource-rich democracies. The answer refers us back to the investment decision: Restraints raise the return on investment for public expenditures, as Nigeria’s experience exemplifies. Nigeria’s first transition from military government to democracy took place in 1978. The new democratic government of President Shehu Shagari provided a classic example of patronage-driven electoral competition and political power unrestrained by checks and balances. One of the Shagari Administration’s first acts was to recall a massive public investment project for a dam awarded under the military government. The project was re-awarded, but its cost rose from $120 million to $600 million. How did this happen? Politicians had spent a fortune buying the votes that got them elected, and the dam was a means of recouping their investments. The additional $480 million in “costs” were essentially additional kickbacks to those who had just become politically well-connected.
The beginning of President Olusegun Obasanjo’s second elected term in 2003 marked Nigeria’s second transitional experience. Obasanjo’s first elected term had been, like the Shagari regime, a transition from military rule to democracy. Indeed, it was a kind of re-run of the Shagari regime. Electoral competition was intense: At the end of the Obasanjo’s first term, 80 percent of incumbent senators were defeated. Checks and balances, however, were virtually non-existent. There had been no time to put them in place, and all the powerful sectional interests had opposed them. At the start of his second and final term, Obasanjo began to introduce the missing checks and balances—a determination that took considerable courage. One of the first restraints, whose prior absence was astonishing, was a requirement for competitive bidding in all public investment projects. When this requirement was first introduced, it was made slightly retroactive: Some previously approved projects were recalled. The introduction of competitive bidding reduced the cost of these recalled projects by an average of 40 percent. These two examples—the huge increase in costs generated by the transition to unrestrained electoral competition, and the equally huge reduction in costs generated by basic restraints—illustrate the sheer scale of what is at stake in oil democracies.
These limited results do not demonstrate the benefits of political restraints in all circumstances, but they do support the value of checks and balances in the context of resource-rich democracies. In other very different contexts, restraints on the use of democratically acquired political power might have offsetting effects, for example making it easier to block economic reform, as may have been the case recently in Germany.
The Resource Trap
Economic growth is not only good in itself; it reinforces democracy, too. In particular, higher levels of per capita income appear to prompt the development of political restraints. For example, in their 2004 NBER working paper, “Do Institutions Cause Growth?” Edward Glaeser and his associates have demonstrated that economic development gradually induces healthy institutional change. So political institutions both reflect past growth and influence future growth.
This insight should warn us about the possibility of a political development trap. A low-income, resource-rich society like Nigeria or Gambia that acquires an instant but lopsided democracy, with vigorous electoral competition but without checks and balances, is likely to so misuse its opportunities that it fails to grow. This in turn closes off the route that most societies have taken to achieve a balanced form of democracy—namely, a path of gradually strengthening political institutions sustained by broad economic development. This, for example, is how many west European countries devolved their monarchies and oligarchies gradually into democracies in the course of the 19th and early 20th centuries.
Conversely, if such a society could only introduce effective checks and balances and somehow sustain them for a while—such as through a phase of unusually far-sighted political leadership—the growth this would generate could underwrite the maintenance of the checks and balances and thus make the country safe for no-holds-barred electoral politics. This is the hope in Nigeria at present. It needs to inform our development-assistance strategy in all oil democracies.
This rather depressing evidence might appear to suggest that the resource-rich societies should stick with autocracy—at least if economic growth is the main objective of public policy. Yet quite aside from the fact that democracy is good in itself, there is reason to doubt this cynical conclusion for at least one important sub-group of resource-rich countries.
While autocracy sometimes appears better from the limited perspective of generating economic growth, in one important context—ethnic diversity—autocracy is very bad news. What may have worked for critical periods in Turkey, Chile, Taiwan and South Korea—all fairly homogeneous societies—is unlikely to work for Indonesia, Nigeria and Trinidad & Tobago. The most likely reason autocracy reduces growth in the context of ethnic diversity is that it tends to narrow the support base of the autocrat. Typically, autocrats in ethnically diverse societies depend on the support of their own ethnic group. The more diverse the society, the smaller this group is liable to be. The narrower the base of social support, the stronger the incentive for economic policy to sacrifice growth in favor of redistributing income to the favored group. Typical examples include Burundi under Tutsi minority rule and the pro-Alawite policies in Syria.
Hence, in the context of ethnically diverse countries such as Nigeria and, especially, pre-war Iraq, autocracy failed to generate conditions conducive to growth, and its return would not improve matters. The uncomfortable implication is that electoral competition is not enough to overcome the blockage to growth generated by autocracy; it merely shifts the blockage to a more diffuse waste of resources through patronage. In the context of ethnic diversity and resource rents, therefore, electoral competition is necessary but not sufficient to produce development.
This combination of a resource-rich, ethnically diverse society with strong political restraints is rare, but it does exist: Botswana. Botswana has diamond mines rather than oil wells, but this difference is immaterial: Big money still flows out of the ground into government coffers. Although Botswana has remained a democracy since its independence, none of its elections has actually changed the governing party. Electoral competition therefore could not reasonably be described as intense. It has managed, however, to preserve strong checks and balances. For example, all public investment projects have been required by law to meet a minimum rate of return. The clear evidence that this requirement has been enforced is that Botswana’s government has accumulated a large surplus of foreign assets. Botswana’s democracy thus distinguishes itself from Nigeria’s: The two countries have carved out a radically different balance between electoral competition and checks and balances. They have also had radically different growth outcomes: Botswana has transformed itself into a middle-income country, achieving for a long time the world’s fastest growth rate.
Putting Democracy Right
Since the fall of the Soviet Union, democracy has spread rapidly across the developing world. However, this transition has thus far been defined overwhelmingly in terms of elections. We tend to focus on electoral competition because it can be introduced with great speed even in unpromising conditions, such as those Afghanistan. By contrast, checks and balances require real institutions and a modicum of social trust. They take time to introduce, and they are often the equivalent of political orphans: Parties that expect to rule resist the introduction of political restraints, and the entire political class stands to lose if patronage politics becomes impractical.
The elections-institutions dichotomy, on the other hand, has not been a problem for developed democracies: They are either not resource-rich, or, as with Norway, their resource wealth is moderate and recent, secured by the full range of democratic institutions and an already high level of income. It is the set of low-income, resource-rich developing societies that are rapidly becoming democracies and the group of fragile, low-income democracies now discovering resource wealth that are all beginning to confront an historically novel range of problems. As a result, developed democracies have so far not shown much understanding of these new circumstances.
Because of the different time scales for the development of elections and checks and balances, “instant democracies” almost inevitably pass through a phase in which the electoral victor reaps all of the political spoils. The real issue is whether this becomes a permanent feature of the polity, entangling the country in the political development trap, or whether institutions catch up in due course. Here is where the international media plays an outsized role.
The recent and rapid spread of contested elections, as in Ukraine from November 2004 to January 2005, demonstrates the power of international influence, especially that of the international media. Being relatively simple and photography-friendly events, elections receive intense media focus. Citizens in the developing world take their cue from this focus, inevitably leading them to see elections as the defining feature of democratic legitimacy. And the focus on elections not only influences citizens’ views, it also allows them to harness the power of international pressure. This is reflected in the fact that many of the banners at political demonstrations demanding fair elections—as in Egypt, for example—are in English rather than the local language. We are part of the intended audience for these protests. Hence, our message back to them matters a great deal. To date, Americans on the whole and the Bush Administration in particular have stressed elections. Checks and balances, on the other hand, are ongoing and complex and so have been much less newsworthy. The mature democracies now need to counteract this tendency by emphasizing the less visible characteristics of democracy, especially in resource-rich young democracies.
Since economic growth itself gradually increases income to the level at which a population demands secure checks and balances, improvement in political restraints eventually becomes self-sustaining. An international effort to promote checks and balances would therefore need to be only temporary. The wave of electoral competition that swept the developing world in the 1990s, and that for all anyone knows may still sweep the Middle East, thus needs to be complemented by a wave of enthusiasm for political restraints. Not only prosperity hangs in the balance, but freedom as well.