Available now: USAvailable now: UK

More on Natural Resources and Democracy  

In our last post, we discussed the relationship between natural resource wealth and democracy. We mentioned earlier research by Michael Ross, which suggested that natural resources retarded or perhaps prevented the emergence of democracies. We also showcased more recent research by Stephen Haber and Victor Menaldo who focused on within-country, over-time variation, instead of the cross-country focus of Michael Ross’s early research. These authors challenge the earlier findings and report results suggesting that there is no negative effect of natural resource and oil wealth on democracy.

Michael Ross, however, has pointed us to his new work with Jorgen Andersen, challenging Haber and Menaldo’s findings, which those interested in this area, and particularly on factors retarding the emergence of democracy in the Middle East, should look at. Andersen and Ross argue that even with country fixed effects, there is a negative effect of oil wealth on democracy, but to see this one needs to interact oil wealth with the post-1980 dummy (partly because the role of oil has changed quite a bit after to oil price hikes of this era). So the debate continues, and there probably is more to uncover in this area — we suspect especially in the more detailed interaction between specific aspects of institutions and the role of natural resource wealth as we will discuss next.


Natural Resources and Political Institutions: Democracy  

The discussion of the curse of natural resources in the Cameroon raises the issue of whether or not natural resources wealth, specifically oil, might influence political institutions as well as economics. The mechanisms we proposed for the resource curse, not surprisingly, worked through politics. But it did take political institutions as given.

Might oil or more generally natural resource wealth lead to institutional deterioration in the political sphere? Might the oil wealth have helped, or even induced, President Biya in Cameroon to subvert the hoped-for transition to democracy in the 1990s?

These questions were first articulated in two articles by political scientists. Their answer was a resounding yes.

Michael Ross did this at the world level in his 2001 paper “Does Oil Hinder Democracy?”. Nathan Jensen and Leonard Wantchekon did it for Africa in “Resource Wealth and Political Regimes in Africa”. Both papers showed that different measures of natural resource abundance or importance in the economy were negatively correlated with how democratic a country was. Greater resource wealth, less democratic was the resounding message of these papers that became conventional wisdom in political science.

One can think of many mechanisms via which this might work. For instance when resource wealth goes up, as in the theoretical papers we discussed, it becomes more valuable to be in power. Thus autocratic leaders are more prepared to use repression or other means to avoid having to democratize or to avoid losing power if they have to hold elections (which is exactly what Biya did).

Yet there is a gnawing issue with this literature. The empirical evidence used the cross-sectional variation to estimate the effect of resource wealth on democracy. Yet almost by definition poor countries are resource dependent and natural resource wealth is large in the total size of the economy, exports and as a source of finance for the government. For example, poor countries do not have good fiscal and tax systems and thus tend to rely on natural resource rents. But poor countries also tend to be much less democratic. Hence this literature left the concern that the correlation between resource wealth, however measured, and autocracy might not represent a causal relationship at all.

This issue was tackled by Stephen Haber and Victor Menaldo who switched attention to the “within variation,” i.e. what happened within a country to the level of democracy when natural resource abundance or dependence changes? In their paper “Do Natural Resources Fuel Authoritarianism? A Reappraisal of the Resource Curse”, they find there is no negative effect of natural resource wealth (oil in fact) on democracy at all. They even find some evidence for positive effects.

Such a finding is not a complete surprise. In fact there is case study literature for Venezuela arguing that oil wealth was crucial for sustaining democracy because it meant that democratic governments could fund public expenditure without taxing rich people, reducing the threat of a coup d’etat. This argument was developed in Thad Dunning’s Crude Democracy, which argued that resource abundance had democracy promoting and democracy retarding effects, the strength of which are conditional on inequality. In his model and empirical results, oil wealth only retards democracy when inequality is low.

Dunning’s work suggests that Haber and Menaldo’s findings may not be the last word. Just as resource wealth may have heterogeneous effects on economic growth depending on institutions, it probably also has heterogeneous effects on democracy.

But what institutions and historical factors are important in these heterogeneous effects? This is a research question that nobody has attacked yet to our knowledge.


The Economic Nature of the Resource Curse: Mechanisms  

We saw in our last post that cross-national evidence suggests countries with poor institutions, such as lack of checks and balances or high levels of corruption, experience economic contractions when they discover natural resources? The question is why.

There is no consensus amongst academics on this. A very nice overview of many arguments is provided in a recent 2011 survey by Rick van der Ploeg, an expert on resource economics, “Natural resources: curse or blessing?”.

One set of hypotheses is developed in our work “Economic Backwardness in Political Perspective” and in Daron’s survey article “Modeling Inefficient Institutions”. The idea is simple: if a dictator or a group of elites is willing to do inefficient things in order to cling to power — in the way that the Russian and Austria-Hungarian elites in early 19th century were willing to block railways and industrialization as we discussed in Why Nations Fail — then greater natural resource rents will make matters worse. In particular, natural resources will raise the “political stakes” and increase the desire of such elites to cling to power and expand the range of inefficient policies, institutions and repressive strategies they would be willing to utilize in order to achieve this objective.

Another related idea that greater political stakes will not only make the elites more likely to pursue inefficient policies or repression to cling to power, but also encourage would-be elites to contest power in order to take control of natural resources rents — a possibility only made more plausible by the all too frequent civil wars in resource-rich countries. (Theoretically, this result follows from standard contest models applied to politics, for example, as in the work of Stergios Skaperdas, and also has a bit of the flavor of the “war of attrition” games, though these may not be as natural in this context).

A related theory, perhaps better suited to some of the dynamics of Cameroon, is the one developed by James in collaboration with Ragnar Torvik and Thierry Verdier (“The Political Economy of the Resource Curse”). In this model an incumbent politician is trying to stay in power when faced with an election, or more generally a contest, by engaging in patronage. In the model patronage takes the form of granting employment in the public sector to some selected groups. This is socially inefficient because people are more productive in the private sector but the role of public sector employment is to tie people’s future incomes to the incumbent staying in power, thus giving them an incentive to support him in the election/contest. If there is a boom in natural resources, it becomes much more desirable for the incumbent to stay in power and so he engages much more aggressively in patronage, expanding the size of the public sector. This reduces national income. Of course the increase in natural resource wealth mechanically increases national income.

The paper shows however that the indirect negative effect can actually be so large as to dominate the direct positive effect so that a resource windfall can lead to a fall in national income. This happens when patronage is a very effective way of staying in power that the paper interprets in terms of institutions. When there are few checks and balances for example or when the state is weak so it is easy to over-ride meritocratic criteria for employment in the public sector as was the case in Cameroon at the end of the 1970s, then a boom in natural resource wealth can reduce income per-capita.


The Economic Nature of the Resource Curse: Evidence  

So oil has been a curse for Cameroon.

Is it true more generally that natural resource wealth, or perhaps more specifically oil wealth, has a negative effect on economic growth? If it does, via what mechanisms?

Before we start thinking about mechanisms let’s focus on the cross-national evidence. In fact, the more careful evidence does not suggest that there is such an unconditional effect.

Early work by Jeffrey Sachs and his collaborators suggested that this was the case. And it is true that the discovery and exploitation of oil in the Cameroon coincided with a massive economic decline and deterioration in human development.

But as we discuss in Chapter 14 of Why Nations Fail, this was not the case in Botswana where the diamond wealth has been a key part of the economic and human development success of the country. Other obvious examples of resource wealth helping economic development include Australia, Chile, Norway and the United States.

So the claim that the average effect of natural resource wealth on economic growth is negative must either be wrong or uninteresting – meaning that the heterogeneous effect of resource wealth in different contexts are what is really interesting to study. What context? As we point out in Why Nations Fail, Botswana’s distinctive characteristic was its institutional development prior to the discovery of diamonds.

It is also obvious that while Cameroon had poor institutions in 1977, Australia, Chile, Norway and the US all had relatively good institutions when they benefitted from resource discoveries.

The idea that the economic impact of natural resources is conditional on the quality of institutions was brought home vividly in a paper by Karl Moene, Halvor Mehlum and Ragnar Torvik, “Institutions and the Resource Curse,”. (See also this related paper). The three Norwegians showed that there is only a “conditional resource curse” in the sense that there is a negative correlation between resource abundance (as measured by the ratio of primary exports to GDP in 1970) and economic growth for countries with low institutional quality. But the same correlation is positive for countries, such as Norway, with stronger institutions (or what we would call “inclusive institutions”).

There are of course many ways to measure institutional quality, and several of these measures are correlated. The three Norwegians created an institutional quality index as an un-weighted average of five indexes based on data from Political Risk Services: a rule of law index, a bureaucratic quality index, a corruption in government index, a risk of expropriation index, and a government repudiation of contracts index. Since many of the measures of institutions used in this literature are the outcomes of political processes, they are also closely related to policy outcomes. Thus this conditional resource curse bundles quite a large number of factors into what conditions the impact of resources which stretch all the way to fundamental aspects of the political institutions of a country (such as the nature of the constitution) to basic economic institutions (security of property rights), the nature of the state (bureaucratic quality) all the way to government policy (repudiation of contracts).

All the same, the paper produces a very important bottom line: to the extent that Cameroon did experience a resource curse after 1977 this was because some key facets of its institutions were initially poor.

What about the mechanisms? We will turn to this in our next post.


Is There a Curse of Resources? The Case of the Cameroon  

Are natural resources really a curse? Before discussing cross-country regressions analysis let’s begin with a case study that appears to illustrate exactly what people have in mind when they talk about the resource curse. In their recent edited volume “Plundered Nations? Successes and Failures in Natural Resource Extraction” Paul Collier and Anthony Venables have a very interesting chapter by Bernard Gauthier and Albert Zeufack called “Governance and Oil Revenues in Cameroon”. There is no better place to start to study the resource curse.

Oil was discovered in Cameroon in 1977. At the time the economy had been growing well based on coffee and cocoa exports. The arrival of oil triggered a boom with the economy growing at an average rate of 9.4% between 1977 and 1986, but then the slide began. Positive turned negative, and by 1993 income per capita was half of the 1986 level. Though growth then resumed, Cameroon is still about 1/3 poorer than it was in 1986. This poor growth experience went along with deteriorating human development. Life expectancy fell from 56 to 50 years between 1995 and 2006 and infant mortality increased by almost 30% over the same period. Primary and secondary school enrollments declined by 10%. A lot of these declines are due to a collapse of public investment.

On the face of it this is odd given that over the period since 1977 Gauthier and Zeufack estimate around US$20 billion has accrued as oil rents to the Cameroonian government. This represents around 67% of the total oil rents, so the lions share of the rent went to the government, not to big bad oil companies.

What did the government do with this windfall?

Gauthier and Zeufack don’t know the answer to this because for a start only 54% could be properly accounted for in the sense that they appeared somewhere in the government budget. They rest simply vanished and “may have been looted”.

Just getting to this first number takes a lot of work and the use of many sources because “The oil sector in Cameroon has been surrounded by official secrecy for most of the last 30 years … very little is known about the level of resources accruing to the country and the use of these resources.’’

Interestingly in 1977 President Ahmadou Ahidjo of the Cameroon decided to create an extra-budgetary account abroad to “manage” the oil revenues. The size of this account was never published, and the president released no information about it. This was not a sovereign wealth fund of the type run by Norway or Chile. Sadly, some erroneous ex post justifications were provided for this lack of transparency even by the international community which should have known better.

Gauthier and Zeufack quote a 1988 World Bank report as noting: 

While this secrecy has potentially dubious effects on the responsibility and accountability for public revenues, it does presumably have the benefit of reducing various pressures to increase government spending, which emerge once it becomes clear that the government is flush with funds

Meanwhile, between 1978 and 1986 government expenditures went from 17% of GDP to 26% of a much larger number. Public sector wages increased as did subsidies. There was also a boom in government capital formation that tripled. The government, based on a temporary increase in oil revenues, embarked on an unsustainable boom in consumption and investment, much of which was in “white elephant” projects with little social value. The crash came in 1986 and in 1988 Cameroon entered into a structural adjustment program with the IMF.

Cameroon has limped along since then. The President Paul Biya, who replaced Ahidjo in 1982, is still in power having survived the democratization of the country in 1992. In October 2011 he won his sixth term in office with 77.9% of the votes. In the meantime he had to amend the 1996 Constitution to remove the two-term presidential term limit.

Oil seems to have brought few development benefits to Cameroon, indeed quite the opposite. This looks like the resource curse doesn’t it? So is resource wealth always a curse?