Volatility And The Natural Resource Curse.

Abstract

We provide cross-country evidence that rejects the traditional interpretation of the natural resource curse. First, growth depends negatively on volatility of unanticipated output growth independent of initial income, investment, human capital, trade openness, natural resource dependence, and population growth. Second, the direct positive effect of resources on growth is swamped by the indirect negative effect through volatility. Third, with well developed financial sectors, the resource curse is less pronounced. Fourth, landlocked countries with ethnic tensions have higher volatility and lower growth. Fifth, restrictions on the current account raise volatility and depress growth whereas capital account restrictions lower volatility and boost growth. Our key message is thus that volatility is a quintessential feature of the resource curse. Copyright 2009 Oxford University Press 2009 All rights reserved, Oxford University Press.

Full text

DEPARTMENT OF ECONOMICS
OxCarre
Oxford Centre for the Analysis of Resource Rich Economies
Manor Road Building, Manor Road, Oxford OX1 3UQ
Tel: +44(0)1865 271089 Fax: +44(0)1865 271094
oxcarre@economics.ox.ac.uk www.economics.ox.ac.uk
Direct tel: +44(0) 1865 281281 E-mail: oxcarre@economics.ox.ac.uk
_
OxCarre Research Paper
No. 2008-03
Volatility and the Natural Resource Curse
Frederick van der Ploeg
(OxCarre, Department of Economics, University of Oxford)
Steven Poelhekke
(European University Institute)

VOLATILITY AND THE NATURAL RESOURCE CURSE
Frederick van der Ploeg* and Steven Poelhekke** §
Abstract
We provide cross-country evidence that rejects the traditional interpretation of the natural
resource curse. First, growth depends negatively on volatility of unanticipated output growth
independent of initial income, investment, human capital, trade openness, natural resource
dependence and population growth. Second, the direct positive effect of resources on growth is
swamped by the indirect negative effect through volatility. Third, with well developed financial
sectors, the resource curse is less pronounced. Fourth, landlocked countries with ethnic tensions
have higher volatility and lower growth. Fifth, restrictions on the current account raise volatility
and depress growth whereas capital account restrictions lower volatility and boost growth. Our
key message is thus that volatility is a quintessential feature of the resource curse.
Keywords: volatility, growth, resource curse, financial development, openness, landlocked,
ethnic tensions, restrictions on current and capital account
JEL code: C12, C21, C23, F43, G20, O11, O41, Q32
Revised 28 February 2009
Address for correspondence:
Department of Economics, Manor Road Building, Manor Road
Oxford OX1 3UQ, United Kingdom
rick.vanderploeg@economics.ox.ac.uk; steven.poelhekke@eui.eu
* University of Oxford. Also affiliated with University of Amsterdam, CEPR and CESifo. This
work is supported by the BP funded Oxford Centre for the Analysis of Resource Rich Economies
and is a revised version of CEPR Discussion Paper 6513.
** De Nederlandsche Bank, Amsterdam and European University Institute, Florence.
§ We thank the referees, Rob Alessie, Anindya Banerjee, Benedikt Goderis and Tony Venables
for helpful comments on an earlier version. We also benefited from comments received during
presentations at Glasgow University, Napels University, the 2007 OxCarre Launch Conference,
the Royal Economic Society 2008 Annual Meeting and De Nederlandsche Bank.

1
1. Introduction
The key determinants of economic growth highlighted in the empirical literature − institutions,
geography and culture – show far more persistence than the growth rates they are supposed to
explain (Easterly, et al., 1993). One candidate to explain the volatility of growth in income per
capita is the volatility of commodity prices. This includes not only oil, but also for example grain
and coffee prices. What commodity prices lack in trend, they make up for in volatility (Deaton,
1999). A recent detailed examination of the growth performance of 35 countries during the
historical period 1870-1939 led to the following conclusions (Blattman, Hwang and Williamson,
2007). Countries that specialize in commodities with substantial price volatility have more
volatility in their terms of trade, enjoy less foreign direct investment and experience lower growth
rates than countries that specialize in commodities with more stable prices or countries that are
industrial leaders. Countries in the periphery with volatile commodity prices and undiversified
economies fall behind in economic development. Also, the long-run volatility of the real
exchange rate of developing countries is approximately three times greater than that of
industrialized countries (Hausmann, et al., 2004). Another study employs data for 83 countries
over the period 1960-2000 and also finds robust evidence for a strong and negative link between
real exchange rate volatility and growth performance after correcting for initial output per worker,
enrolment in secondary education, trade openness, government consumption, inflation and even
banking or currency crises (Aghion, et al., 2006). Furthermore, the adverse effect of exchange
rate volatility on growth is weaker for countries with well developed financial systems.
The pioneering work of Ramey and Ramey (1995) takes a different tack. It investigates
the link between volatility of unanticipated output growth (rather than volatility of the terms of
trade) and growth performance. It uses the Heston-Summers data to provide cross-country
evidence for a negative link between volatility and mean growth rates controlling for initial
income, population growth, human capital and physical capital. Interestingly, this study finds
evidence for this negative link regardless of whether one includes the share of investment in
national income or not. It also estimates the relationship between volatility and growth in a panel
model that controls for both time and country fixed effects. To allow for the time-varying nature
of volatility, a measure of government spending volatility is used that is correlated with volatility
of output across both time and countries. The negative link between volatility and growth seems
robust to a large set of conceivable controls that vary with time period or country.1 In a cross-
1 However, Imbs (2007) shows that growth and volatility correlate positively across sectors. Within the
context of a mean-variance portfolio setup, it is understandable that volatile sectors command higher
investment rates and thus higher growth rates. A critique of Ramey and Ramey (1995) may be that the

2
section of 91 countries policy variability in inflation and government spending exerts a strong and
negative impact on growth (Fatás and Mihov, 2005).
Our main objective is to extend Ramey and Ramey (1995) by allowing for the direct
effect of natural resource dependence on growth and, more importantly, the indirect effect of
natural resources on growth performance via volatility. We thus follow Blattman, Hwang and
Williamson (2007) and allow for the role of natural resources in macroeconomic volatility. We
allow natural resources, financial development, openness and distance from waterways to be the
underlying determinants of volatility. These variables affect the volatility of the real exchange
rate and thus also GDP growth.
Another objective is to give evidence against the conventional interpretation of the
natural resource curse following from Sachs and Warner (1997ab, 2001) and many others.2
Brunnschweiler and Bulte (2008) found that, using resource abundance (i.e., stocks of natural
resource wealth) rather than resource dependence (i.e., natural resource exports as a percentage of
GDP) as an explanatory variable, leads to a positive rather than a negative effect of resources on
growth. In a similar vein, we find that the direct effect of natural resources on growth
performance may well be positive. However, we take the argument further and establish that the
indirect effect of natural resources on growth via the volatility channel is negative. We thus test
whether any adverse indirect effect of natural resources on growth performance via volatility of
unanticipated output growth dominates any direct effect of natural resource abundance on
economic growth. Inspired by Aghion, et al. (2006), we test whether the adverse effect of natural
resources on volatility and growth is weakened if there are well developed financial institutions.
We also test whether being landlocked, ethnic tensions and restrictions on the current account
boost volatility and curb growth and whether restrictions on the capital account and exchange
controls reduce volatility and boost growth. To avoid omitted variable bias, we control for initial
income per capita, population growth, investment rates and primary schooling.
To motivate our multivariate econometric tests for the importance of volatility for the
resource curse, we first present some telling stylized facts and partial correlations for the period
1970-2003 in Figures 1–3 and Table 1:
observed negative effect of volatility on growth may be driven by the stark contrast between developed
countries with low volatility and developing countries with high volatility. To the extent that our results
allow for a richer set of controls in the growth equation and, more importantly, try to simultaneously
explain the volatility of unanticipated output growth with institutional, geographical and economic
variables, our results are less susceptible to this critique.
2 The windfall resource revenues lead to appreciation of the real exchange rate and decline of the non-
resource export sectors. If there is substantial loss in learning by doing in the non-resource export sectors,
there will be a fall in total factor productivity growth as in Sachs and Warner (1995). Natural resources may
also invite rapacious rent seeking and thus hamper growth.

3
• First, volatile countries with a high standard deviation of yearly growth in GDP per capita
have on average lower growth in GDP per capita. Figure 1 illustrates this simple correlation
while Ramey and Ramey (1995) show that this also holds after controlling for initial
income per capita, population growth, human capital and physical capital.
• Second, developing countries have more volatile output growth than developed countries.
Whereas Western Europe and North America have a standard deviation of, respectively,
2.33 and 1.90 %-points of yearly growth in GDP per capita, the figures for Asia are 4.4 to 5
%-points and for Latin America & Caribbean 4.54%-points. Most striking is that Sub-
Saharan Africa and the Middle East & North Africa have highest volatility. Their standard
deviations of average growth in GDP per capita are, respectively, 6.52 and 8.12 %-points.
• Third, countries with poorly developed financial systems are more volatile. Countries in the
bottom quartile of financial development have a standard deviation of annual growth in
GDP per capita 2 %-points higher than those in the top quartile. North America and
Western Europe have well developed financial systems while Eastern Europe & Central
Asia and especially South Asia and Middle East & North Africa have poor functioning
financial systems. Resource-rich and landlocked economies have less developed financial
systems than resource-poor countries.
• Fourth, countries that depend a lot on natural resources are much more volatile. Countries
with a share of natural resource exports in GDP greater than 19% (the top quartile) have a
staggeringly high standard deviation of output growth of 7.37 %-points. For countries with
a natural resource exports share of less than 5 per cent of GDP (the bottom quartile), the
figure is only 2.83 %-points. Figure 2 also indicates that resource-rich countries have
greater macroeconomic volatility. Figure 3 shows that world commodity prices are
extremely volatile and are the main reason why natural resource export revenues are so
volatile. Crude petroleum prices are more volatile than food prices and ores & metals
prices. Volatility of agricultural raw material prices is less, but still substantial. Monthly
price deviations of 10%-points from their base level (year 2000) are quite normal.
• Fifth, landlocked countries suffer much more from volatility. Indeed, countries that are less
than 49 kilometres from the nearest waterway have a standard deviation of growth in GDP
per capita that is 1.6 %-points lower than countries that are more than 359 kilometres from
the nearest waterway. Empirical work also finds that remote countries are more likely to
have undiversified exports and to experience greater volatility in output growth (Malik and
Temple, 2006). Since Figure 1 indicates that the negative correlation between volatility and
growth in income per capita is not much different for landlocked countries, the

4
disappointing growth performance of landlocked countries may be due to their higher
volatility rather than being landlocked.
Although these stylized facts are suggestive, we perform a proper multivariate econometric
analysis and control for all potential factors affecting the rate of economic growth.
Several papers have looked closer at the sources of volatility. The sophisticated statistical
decomposition analysis performed in Koren and Tenreyro (2007) sheds light on why poor
economies are more volatile than rich economies. They suggest four reasons why poor countries
are much more volatile than rich countries: they specialize in more volatile sectors; specialize in
fewer sectors; experience more frequent and more severe aggregate shocks (e.g., from
macroeconomic policy); and their macroeconomic fluctuations are more highly correlated with
the shocks of the sectors they specialize in. The evidence suggests that, as countries develop their
economies, their productive structure shifts from more to less volatile sectors. Also, the degree of
specialization declines in early stages of development and increases a little in later stages of
development. Furthermore, the volatility of country-specific macroeconomic shocks falls with
development.
Our multivariate econometric analysis provides complementary evidence on the factors
affecting volatility by focusing on one of the most volatile sectors: natural resources. We argue
that crucial and strongly related sources of macroeconomic volatility and poor growth
performance are natural resource dependence, but also lack of a sophisticated financial system
and whether a country is landlocked or not. We also provide evidence that economic restrictions
and ethnic tensions play a role. Landlocked countries with a large dependence on natural
resources are typically less diversified and vulnerable to volatile world commodity prices. Natural
resource revenues tend to be volatile (much more so than GDP), because the supply of natural
resources exhibits low price elasticities of supply. Furthermore, as documented in Bloom and
Sachs (1998) and indicated by Figure 4, Sub-Saharan Africa is most vulnerable to volatility of
commodity prices as it depends so much on natural resources. Dutch Disease effects may also
induce real exchange rate volatility and thus a fall in investment in physical capital and learning,
and further contraction of the traded sector and lower productivity growth (e.g., Gylfason, et al.,
1999; Herbertsson, et al., 2000). Volatile resource revenues are disliked by risk-averse
households. The welfare losses induced by consumption risk are tiny compared with those
resulting from imperfect financial markets. However, a recent dynamic stochastic general
equilibrium study of Zimbabwe highlights the incompleteness of financial markets and suggests
that the observed volatility in commodity prices depresses capital accumulation and output by
about 40 percent (Elbers, et al., 2007). Furthermore, the relatively high macroeconomic volatility

5
in developing countries induces relatively high welfare cost of consumption volatility and the
welfare cost of removing this volatility may exceed the welfare gain from a permanent additional
percentage point of growth (Pallage and Robe, 2003).
Our paper gives a prominent role to the quality of financial markets in understanding how
the volatility of commodity prices and natural resource export revenues might depress growth.
We adapt the liquidity shock arguments put forward by Aghion, et al. (2006). Effectively, larger
natural resource revenues make it easier to overcome negative liquidity shocks. We thus show
that more volatile commodity prices will harm innovation and growth.
Section 2 discusses why volatility may harm output growth, especially in countries with
poor financial systems. Since there are also theoretical reasons for volatility to boost growth, the
issue needs to be settled empirically. Section 3 gives cross-country evidence which shows that the
traditional estimates of the natural resource curse are not robust, where Appendices 1 and 2
describe the data that we have used in our estimates. Section 4 presents our cross-country
estimates on the determinants of volatility and the effect of volatility on economic growth where
our econometric methodology is set out in Appendix 3. Section 5 uses our core estimates to
compare resource-rich and landlocked Africa with a sample of South-East Asian countries.
Section 6 concludes.
2. Why Might the Volatility of Natural Resource Revenues Hamper Growth?
2.1. Economic arguments
Aghion, et al. (2006) show that with macroeconomic volatility driven by nominal exchange rate
movements, firms are more likely to hit liquidity constraints and thus cannot afford to innovate
which depresses growth, especially in economies with poorly developed financial institutions.3
We adopt this argument to show that volatility in natural resource revenues, induced by volatility
in primary commodity prices, curbs growth in economies with badly functioning financial
systems. A high and stable level of resource revenues eases liquidity constraints and thus boosts
innovations and economic growth. However, for a given expected level of natural resource
revenues, more volatility in commodity prices and resource revenues harms innovation and
growth, especially if financial development is weak.
3 It is assumed that the price level is determined by the nominal exchange rate (the law of one price),
nominal wages are pre-set not knowing the realization of the price level, the production function is of the
Cobb-Douglas variety, the cumulative density function of liquidity shocks is concave, and firms maximize
profits and can only innovate if they have enough cash (profits plus resource revenue) to cope with adverse
liquidity shocks. With higher profits or resource revenues and a more developed financial system, more
firms are more able to overcome liquidity shocks and thus the probability of innovation is higher. It can
then be shown that moving from a peg to a float curbs innovation and growth.

6
IMF data on 44 commodities and national commodity export shares and monthly indices
on national commodity export prices for 58 countries during 1980-2002 indicate that real
commodity prices affect real exchange rate volatility (Cashin, et al., 2002). Since we have seen
that real exchange rate uncertainty exacerbates the negative effects of domestic credit market
constraints, this gives another reason why volatility of commodity prices curbs economic growth.
Also, many resource-rich countries suffer from poorly developed financial systems and financial
remoteness and thus suffer from greater macroeconomic volatility (Aghion, et al., 2006; Rose and
Spiegel, 2007). Given the high volatility of primary commodity prices and resource revenues and
thus of the real exchange rate of many resource-rich countries, we expect resource-rich countries
with poorly developed financial systems to have poor growth performance.
With complete financial markets, long-term investment is counter-cyclical and mitigates
volatility. However, if firms face tight credit constraints, investment is pro-cyclical and amplifies
volatility. Of course, there may be other reasons why volatility may depress economic growth
(Aghion, et al., 2005). Learning by doing and human capital accumulation is increasing and
concave in the cyclical component of production (Martin and Rogers, 2000). In that case, long-
run growth should be negatively related to the amplitude of the business cycle.4 This explanation
does not require uncertainty and holds for predictable shocks as well. With irreversible
investment, increased volatility holds back investment and thus depresses growth (Bernanke,
1983; Pindyck, 1991; Aizenman and Marion, 1991). The costs of volatility come from firms
making uncertainty-induced planning errors (Ramey and Ramey, 1991). These costs arise if it is
costly to switch factors of production between sectors (Bertola, 1994; Dixit and Rob, 1994).
However, if firms choose to use technologies with a higher variance and a higher expected return
(Black, 1987) or if higher volatility induces more precautionary saving and thus more investment
(Mirman, 1971), there may be a positive link between volatility and growth. If the activity that
generates productivity growth is a substitute to production, the opportunity cost of productivity
enhancing activities is lower in recessions and thus volatility may boost growth (Aghion and
Saint Paul, 1998). Ultimately, the question of whether anticipated or unanticipated volatility
harms or boosts growth thus needs to be settled empirically.
In economies where only debt contracts are available and bankruptcy is costly, the real
exchange rate becomes much more volatile if the traded sector is heavily dependent on natural
resources and not very diversified (Hausmann and Ribogon, 2002). Shocks to the demand for
non-traded goods and services – associated with shocks to natural resource income – are then not
4 They find that for industrialized countries and European regions a higher standard deviation of growth
and of unemployment tends to depress growth rates.

7
accommodated by movements in the allocation of labour but by expenditure switching. This
demands much higher relative price movements. Due to bankruptcy costs, interest rates increase
with relative price volatility. This causes the economy to specialize away from non-resource
traded goods and services, which is inefficient. The less it produces of these goods and services,
the more volatile the economy becomes and the higher the interest rate has to be. This causes the
sector to shrink further until it vanishes. Others stress that resource revenues are used as collateral
and encourage countries to engage in ‘excessive’ borrowing at the expense of future generations,
which can harm the economy in both the short and the long run (Mansoorian, 1991).
Volatility is bad for growth, investment, income distribution, poverty and educational
attainment (e.g., Ramey and Ramey, 1995; Aizenman and Marion, 1999; Flug et al., 1999). To
get round such natural resource curses, the government could resort to stabilization and saving
policies and improve the efficiency of financial markets. It also helps to have a fully diversified
economy, since then shocks to non-traded demand can be accommodated through changes in the
structure of production rather than expenditure switching. This is relevant for inefficiently
specialized countries such as Nigeria and Venezuela, but less for diversified countries like
Mexico or Indonesia or naturally specialized countries such as some Gulf States. Unfortunately,
resource-rich economies are often specialized in production of natural resources and thus tend to
be more volatile.
2.2. Political arguments
Natural resource bonanzas reduce critical faculties of politicians and induce a false sense of
security. This can lead to investment in ‘white elephant’ projects, bad policies (e.g., import
substitution or unsustainable budgetary policies), and favours to political clientele, which cannot
be financed once resource revenues dry up. Politicians lose sight of growth-promoting policies,
free trade and ‘value for money’ management. During commodity booms countries often engage
in exuberant public spending as if resource revenues last forever. This carries the danger of
unsustainable spending programmes, which need to be reversed when global commodity prices
collapse and revenues dry up. Perhaps encouraged by the Prebisch hypothesis (the secular decline
of world prices of primary exports), some developing countries have promoted state-led
industrialization through prolonged import substitution to avoid resource dependency. These
policies may also have been a reaction to the appreciation of the real exchange rate and decline of
the traded manufacturing sectors caused by natural resource dependence. Once natural resource
income has ceased, policies often had to be reversed. The resulting policy-induced volatility
harms growth and welfare. Table 1 indicates that resource-rich countries indeed have a relatively

8
high volatility in the national income share of government. Case studies suggest that during the
1970s and 1980s many oil windfalls could have been put to better use (Gelb, 1988).
Political scientists have also argued that states adopt and maintain sub-optimal policies,
and have studied the resource curse in great detail (e.g., Ross, 1999). Cognitive theories blame
policy failures on short-sightedness of state actors, who ignore the adverse effects of their actions
on the generations that come after the natural resource is exhausted, thus leading to myopic sloth
and exuberance. These cognitive theories highlight a get-quick-rich mentality among
businessmen, a boom-and-bust psychology among policy makers, and abuse of resource wealth
by privileged classes, sectors, client networks and interest groups. Of course, some of these
choices may well be rational when leaders have short-term horizons due to political instability or
other reasons (Caselli and Cunningham, 2009).
3. Is the Traditional Natural Resource Curse a Red Herring?
Ding and Field (2005), Alexeev and Conrad (2005) and Brunnschweiler and Bulte (2008)
demonstrate that the natural resource curse as estimated by Sachs and Warner (1995, 1997ab,
2001) is not robust. They show that, once resource abundance (proxied by a measure of natural
resource wealth) rather than resource dependence (the average 1970-80 national income share of
natural resource exports) is used, the effect of natural resources on growth performance is
positive and thus the resource curse disappears. Resource abundance is measured as the net-
present value of natural capital in USD per capita in 1994, including subsoil assets, forest
resources, protected areas, and agricultural land, with a constant discount factor of 4%. Since this
measure makes several assumptions about the valuation of resources and has a limited time span,
we feel that our approach is better served by the actual export revenue received on world markets.
The latter is reported since 1970. The drawback on the other hand is that it does not measure how
many resources a country has, but only its dependence on resources. Dependence is however the
main channel through which our story runs. The literature uses the words dependence and
abundance interchangeably, but we will refer to dependence as the export value of resources as a
share of GDP, and abundance as the net present value of natural capital.
Sachs and Warner find for a wide range of control variables that natural resource
dependence harms growth during the years 1970-90 even after allowing for the effects of
geography and quality of institutions. Table 2 re-estimates the Sachs and Warner regressions with
more recent data for the period 1970-2003. Instead of the average budget balance as in Sachs and
Warner (1997b), we use the average investment share which captures both public and private
investment. The first regression indicates that growth performance is better in countries that are

9
poor, open to international trade, have small population growth rates and a long life expectancy
(as a proxy for human capital). Furthermore, growth seems to be higher in countries with a
superior rule of law. Countries that are poor grow faster than rich countries (i.e., there is
conditional convergence), especially if they are open to international trade. Investment does not
seem to be a statistically significant determinant of economic growth. The main point is that the
first regression also indicates that, even allowing for all of these determinants of growth, there is a
strong negative effect of resources on the average annual growth in income per capita. A country
with a ratio of natural resource exports to GDP of 40 percent seems to enjoy 1%-point growth per
annum less than a country which does not export natural resources. This is a substantial effect and
has been coined the natural resource curse. However, we find that this type of evidence for the
natural resource curse is not robust to including other important determinants of economic
growth. For example, after adding the standard deviation of actual annual growth in GDP per
capita for the 33 year period as an additional explanatory variable, the effect of natural resources
on growth performance vanishes. In this sense, the natural resource curse is indeed a red herring.
In the remainder of this paper we provide our empirical evidence on volatility and the
resource curse. Table 3 in section 4.1 then yields our core estimates of volatility on growth and
we establish that the direct effect of natural resources on growth is positive while the indirect
effect through volatility is negative. Section 4.1 also probes deeper into the causes of
macroeconomic volatility. We also show that the negative indirect effect of volatility is especially
strong in highly volatile countries. Section 4.2 discusses the empirical effects of the volatility of
various commodity resource shares on volatility on unanticipated growth and thus on growth
performance. Table 4 in section 4.3 extends the analysis further and estimates the effects of ethnic
fractions and of various current account and capital restrictions on volatility and thus growth.
4. Is Volatility the Quintessential Feature of the Natural Resource Curse?
Having rejected the traditional resource curse and the implied negative effect of resource
dependence on economic growth, it could be that resource dependence affects growth through
other channels. For example, resource dependence may erode the quality of institutions or the
legal system and thus hamper growth. Or resource dependence may lower human capital
formation or physical investment and thus dampen growth prospects. However, the stylized facts
discussed in the introduction and the second regression in Table 2 suggest that natural resources
must be given a key role in understanding macroeconomic volatility and growth prospects. We
therefore estimate growth regressions simultaneously with regressions explaining volatility of
unanticipated growth in income per capita (see Appendix 3). Once account is taken of the

10
negative effect of cross-country variations in volatility on the rate of economic growth, the level
of resource dependence may exert a positive effect on growth.5 From a policy perspective, it is
important to know whether any negative indirect effect of natural resources on growth
performance via volatility of unanticipated output growth dominates any positive direct effect of
resource dependence on growth, and whether the adverse effects are weakened if there are well
developed financial institutions. Furthermore, we test whether landlocked countries experience
higher volatility and lower growth. To get meaningful results, we control for initial income per
capita, population growth, investment rates and primary schooling on growth. We have also
estimated all our regressions with year dummies included in the annual growth equation, but this
does not yield substantially different results. The countries used in our sample for the core
regressions are reported in Appendix 1 while the data are described in Appendix 2.
4.1. Volatility is the key channel for the resource curse
To better understand the effects of natural resource dependence on growth, we need to dig deeper
into the determinants of volatility. Regression 6a in Table 3 does exactly that. It still finds that
investment in physical and human capital boost economic growth while population growth
depresses growth in income per capita. There is also again evidence for poor countries catching
up. Interestingly, there is now evidence of a significant positive direct effect of point-source
natural resource export revenue on economic growth. There is no evidence for a significant effect
of openness on growth. There is evidence for a significant direct effect of financial development
on economic growth, but unfortunately it is negative. More important, volatility of unanticipated
growth exerts a powerful and negative effect on growth in GDP per head. As expected, volatility
itself increases with the GDP share of point-source resources but not significantly with the GDP
share of diffuse resources. Volatility also decreases with the degree of financial development and
openness of a country to international trade, which supports the hypothesis put forward by
Aghion et al. (2006) and Rose and Spiegel (2007). In line with Malik and Temple (2006), we find
that volatility increases with the distance from navigable coast or rivers, which is their strongest
geographical predictor of output volatility.
Figure 5 calculates on the basis of regression 6a the marginal effect of resource
dependence on growth. This effect depends on volatility of unanticipated output growth, because
5 As already mentioned in section 3, If the explanatory variable is natural resource abundance (proxied by
natural resource wealth per capita) rather than natural resource dependence, there appears to be a positive
effect on growth performance. From our point of view, this does not seem surprising as natural resource
wealth is much less volatile than natural resource export revenues and more likely to boost the rate of
economic growth.

11
resource dependence enters the volatility equation in a non-linear way as described in equation
(A2).6 We thus see that the total effect of resource dependence on growth is given by a direct
effect (measured by the relevant parameter in θ and an indirect effect through volatility, measured
by the relevant term in 1
2i
λ
σ
γ
). Natural resource dependence is thus, due to the indirect effect, a
curse for volatile countries, but a boon for countries with relatively stable unanticipated output
growth. In fact, if σ exceeds 0.064 (i.e. 2*0.05/(1.621*0.971)), resource dependence curbs growth
and otherwise it boosts growth. More open and financially developed countries are expected to be
more stable and grow faster even if they export many resources. We see from Figure 5 that for the
less volatile OECD (including Norway) and South-East Asia, resource dependence is a boon for
growth, while for volatile landlocked Africa (especially Zambia) a curse. For resource-rich Africa
the positive direct effect of resource dependence is more or less cancelled out by the indirect
effect through volatility. However, this is a best-case scenario based on a weakly significant
direct positive effect. In later regressions we find a negative direct effect of resource dependence
on growth, in which case the line in Figure 5 shifts down such that the curse is apparent for more
regions and countries. The resource curse is always more severe for more volatile countries.
Growing countries attract more investment, so the direction of causality may go either
way. Even though we control for openness and financial development, we probably do not
capture enough of the institutional effects on growth and investment. We therefore looked for an
exogenous variable that strongly predicts the investment share, but does not affect growth or
correlate with other important unobserved characteristics. We instrumented the investment share
with an index of ethno-linguistic fractionalization. This index measures the probability that two
randomly selected individuals from a given country will not belong to the same ethnic group
(Montalvo and Reynal-Querol, 2005a).7 The rationale is that trust, ability to communicate and
social cohesion are essential prerequisites for successful investment. Fractionalized countries
have lower levels of trust, more corruption, less transfers, subsidies and political rights (Alesina
et al., 2003). These factors should lower the investment rate, since they increase uncertainty about
returns and expropriation.8 We also include two geographical variables: whether a country is
6 Ramey and Ramey (1995) have used the same specification. We also tried the logarithm of the variance in
the mean equation, but this gave a much worse fit. The exponential specification forces volatility to take on
positive values only.
7 They base their data on the World Christian Encyclopedia. They argue that fractionalization is a poor
predictor of civil war compared to ethnic polarization. We are therefore more confident that there is no
effect of fractionalization on growth via the link of conflicts.
8 Montalvo and Reynal-Querol (2005b) argue that ethnic polarization affects investment but not growth,
while fractionalization affects growth directly as in Easterly and Levine (1997), but not investment.

12
landlocked or not, and a climate variable. Our results (available on request) show that all our
effects are qualitatively robust to this IV strategy, also if we additionally control for the
possibility that fractionalization affects growth directly through the general quality of
bureaucracy and corruption.9 .
Summing up and probing deeper into the determinants of volatility, we find that countries
that are closed to international trade, have badly functioning financial markets, are landlocked
and have a high share of natural resource exports have higher volatility in unanticipated growth in
output per capita and therefore worse growth prospects. These results suggest, in contrast to the
previous literature, that volatility of commodity prices is a key feature of the resource curse.
4.2. Volatility of commodity export shares and macroeconomic volatility
With regression 6a as the benchmark, regression 6b in Table 3 tries to see if the marginal effect of
initial resource dependence on volatility is weaker if a country starts off from a higher level of
financial development as well. This seems not to be the case. However, it is more likely that
financial services give countries the means to deal with large world price shocks and will reduce
the effect of resource wealth fluctuations on output volatility. Financial development may limit
the pass-through of volatile resource income into general output volatility through insurance and
easing of borrowing constraints. The second half of Table 3 therefore focuses on ML estimates of
regressions with fluctuations in the GDP shares of resource exports as an additional explanatory
variable in the variance equation. Since resource quantities are relatively inelastic, most of the
revenue movement will originate in world prices. Regression 7a indicates that adding the
volatility of the GDP share of both point-source and diffuse resources to the variance equation
significantly helps to explain the volatility of unanticipated output growth. Regression 7b
indicates that, inspired by Fatás and Mihov (2005), adding the volatility of the GDP share of
government spending (capturing policy shocks and spending bonanzas following windfall
revenues) also significantly improves our estimate of the volatility of unanticipated growth.
Furthermore, regression 7c shows that especially the volatility of the food export share, the
volatility of the fuel export share and the volatility of the ores & metals export share contribute to
the volatility of unanticipated output growth. The volatility effect of natural resources is thus not
However, these growth regressions do not control for population growth or volatility. If we run regression
6a with ethnic fractionalization and polarization using their ethnicity data, we find no growth effects of
these two variables. Adding polarization to the first stage yields no effect of polarization, but still gives a
significant negative effect of fractionalization on investment. Taking the effect of volatility into account
seems to have important effects on the link between ethnicity and growth, and should be seen as
complementary.
9 As measured in the International Country Risk Guide (PRS Group, 2006).

13
limited to oil-producing countries, but also includes for example copper, coffee, banana and
tobacco exporters. The qualitative results of the estimated equation for annual growth are not
much affected, except that the estimated negative effect of volatility of unanticipated output
growth on mean annual growth is almost three times smaller and closer to the black-box estimate
with individual country dummies (despite being much more parsimonious).
Although we did not find evidence for significant interaction between financial
development and initial point-source resource dependence in the variance equation, regression 7d
suggests that well-functioning capital markets reduce the effect that shocks in the resource share
have on volatility. Consistent with the model of section 2, a stable share of natural resources in
GDP does not increase volatility by itself, but rapid fluctuations in the share through prices create
liquidity constraints and harm growth. Financial development gives a country the means to deal
with sudden changes in resource revenues even when controlling for terms-of-trade shocks.
4.3. Impact of ethnic tensions and economic restrictions on volatility and growth
Table 4 presents some further refinements and robustness tests of our results. Since ethnic
polarization as defined by Montalvo and Reynal-Querol (2005b) is a good predictor of civil
conflict, it may also be a good predictor of volatility. We want to check whether resources still
have an independent effect on volatility when we allow for an effect of ethnic polarization.
Furthermore, this measure takes its highest value if a country is equally divided into two groups.
Such a situation may increase instability if natural resources are present as well. Regression 8a
indicates indeed that ethnic polarization significantly improves the estimate of the volatility of
unanticipated output growth, but does not have an independent direct effect on growth. The
interaction term shows that there is also significant positive interaction with resource dependence:
the more polarized a country, the more resources lead to volatile economies through conflict and
rent-seeking government policy. Regression 8b shows that ethnic polarization is no longer
important once volatility of export shares of point-source and diffuse resources, and volatility of
GDP share of government spending are used as explanatory variables of volatility. Resources are
not necessarily bad, but anything that magnifies already volatile prices, such as public spending
booms and busts (possibly related to civil strife), seems to harm long-run growth prospects.
Table 4 also tests for the impact of economic restrictions to examine whether financial
and trade liberalization boosts or depresses growth. Grilli and Milesi-Ferretti (1995) found no
significant effect on growth but discuss their benefits for reducing capital flow volatility and
facilitating stabilization policies. They may be important omitted variables in our variance
equation. We therefore replace the single openness dummy with four dichotomous measures of

14
restrictions from the Annual Report on Exchange Arrangements and Restrictions (IMF, 2006).
These include for example limits on repatriation of profits, exchange rate controls and restrictions
on international payment. Regression 9a indicates that capital account restrictions have a
somewhat negative direct impact on growth. However, this is swamped by the negative effect of
capital account restrictions on volatility and thus the positive effect on growth, especially for
countries with high resource dependence. Capital account restrictions may thus curb volatility and
boost growth, especially in resource-rich countries.10 Access to international capital markets may
be pro-cyclical, which may generate higher output volatility especially in resource-rich,
developing economies. Current account restrictions have no significant direct effect on growth,
but do contribute to volatility especially in resource-rich countries and thus hamper growth.
Regression 9a also indicates that the surrender of export receipts is associated with higher
volatility and lower growth. Multiple exchange practices lower volatility and increase growth,
since these exchange controls curb volatile capital in- and outflows. Regression 9b drops the
interaction terms but includes the volatility of revenues, government spending and terms-of-trade
shocks as explanatory variables in the variance equation, but this does not change the effect of the
four restrictions.
5. Accounting for Growth Performance: Africa versus South-East Asian Countries
To get a feeling for what our estimates of the determinants of growth in GDP per capita imply in
practice, it is interesting to perform some counterfactual exercises. We perform these exercises
based on our core equation 6a of Table 3. It is insightful to compare the African countries with
some fast-growing resource-poor South-East Asian countries (including some South-East Asian
countries)11, since they have similar starting positions (in 1970). We therefore compare in Table 5
resource-rich and landlocked Africa with the South-East Asian Sample. Resource-rich countries
are those in the global top 25 and natural resource exports valuing on average more than 17.31%
of GDP during 1970-2003. Since the resource-rich countries of Africa were poorer in 1970 than
the South-East Asian countries, they grow faster and catch up, everything else equal. We see from
the top panel of Table 5 that this growth differential amounts to 0.87%-point per year (the
difference in initial GDP per capita times the coefficient). Allowing for the positive direct growth
effects of higher natural resource dependence in Africa, we see that the growth differential with
the South-East Asian countries becomes 1.31%-point. Now if those African countries had
10 Kose et al. (2003) find that increased gross financial flows and the absence of capital account restrictions
lead to an increase in the relative volatility of consumption.
11 The South-East Asian countries in our sample are South Korea, Malaysia, Philippines and Thailand

15
invested as much in physical and human capital as their South-East Asian counterparts, they
would add a further 0.65%- and 0.46%-points, respectively to their annual growth rate. If
resource-rich Africa’s population growth rate were to be reduced in line with the South-East
Asian sample, Africa would gain yet another 0.43%-point annual growth. These three factors
combined yield an extra growth bonus of 1.54%-point. However, the key message is how much
potential growth is lost due to the high volatility of unanticipated output growth in resource-rich
Africa compared with their South-East Asian counterparts: 2.98%-point extra growth per annum!
The main reasons for the high volatility of resource-rich Africa are their heavy dependence on
resources (increasing volatility by 0.41%-point, translating into a 0.52%-points loss in growth),
lack of openness (1.71%-point), under-developed financial markets (0.58%-points) and distance
from waterways (1.07%-point).12
The bottom panel of Table 5 compares landlocked Africa with the South-East Asian
countries in our sample. The results are similar, although the prospects of these countries are
perhaps even more miserable. Still, as landlocked Africa starts off from a worse starting position
than resource-rich Africa, it catches up more quickly and thus grows 1.41%-point faster than the
South-East Asian countries. Accounting for landlocked Africa being more dependent on
resources than the South-East Asian countries, would raise this growth differential to 1.74%-
point. Now bringing mainly investment in physical and human capital but also population growth
in line with the South-East Asian countries would add an extra 1.47%-point growth per annum.
This offers some hope. However, if landlocked Africa were to be able to bring down its volatility
of unanticipated output in line with that of the South-East Asian countriesa it would boost growth
by a further 1.97%-point per annum. The potential growth bonus is thus 3.44%-point. If this were
feasible, landlocked Africa’s negative growth differential with the South-East Asian countries of -
3.82%-points could have been reduced to a little as -0.38%-points. The countries Malawi and
Zambia are resource rich and landlocked. They also have relatively high volatility and poorly
developed financial systems. Not surprisingly, they have a lot to gain.
We conclude that a big push to economic growth occurs if the volatility of unanticipated
output growth in Africa is brought down to the level of the South-East Asian countries. The big
contributing factors to Africa’s volatility are its volatile stream of mainly point-source natural
resource revenues, its lack of fully developed financial markets and openness to international
trade, and its disadvantages of being relatively more landlocked.
12 Each number is obtained by keeping all other variables constant and using the country’s value for the
respective variable. The effect on growth is then calculated using the coefficient on volatility in the mean
equation. They therefore reflect the growth effect of changing only one variable to the country’s 1970 level.

16
6. Concluding Remarks
We have shown that the resource curse is foremost a problem of volatility. The high volatility of
world prices of natural resources causes severe volatility of output per capita growth in countries
that depend heavily on them. The resulting volatility of unanticipated output growth has a robust
negative effect on long-run growth itself and is a curse. This is not limited to oil-exporters, but
also applies to exporters of copper, coffee, foods, etc. which include many of the world’s worst
performing countries. Also, ethnic tensions, which are often fuelled by resource wealth, and
current account restrictions increase volatility. The latter effect is especially strong in resource-
rich countries. Government spending bonanzas after windfall resource revenues also increase
volatility to the detriment of growth, because revenue drops inevitably follow.
Volatility can fortunately be reduced provided that countries have a sound financial
system to cope with large and sudden fluctuations in resource income. Fewer capital account
restrictions, openness and physical access to world trade also lower volatility. Countries can turn
the curse even into a blessing, because we find evidence for a positive direct effect of natural
resource dependence on growth after controlling for volatility. The key to a turn-around for many
resource-rich countries is financial development, ensuring openness and mitigating the effect of
being landlocked, because the indirect negative effect of resource dependence on growth, via
volatility, is much larger than any direct positive effect. While it may be difficult to lower price
volatility of resources themselves, it should be feasible to deal with volatility in a more efficient
way. It is increasingly realised that large external shocks, volatile macroeconomic policies,
microeconomic rigidities and weak institutions induce substantial income volatility in many
developing countries, which imposes significant welfare losses for risk-averse individuals (e.g.,
Loayza et al., 2007). Future research should thus be focused on ways on how to cope with such
volatility and manage the associated risks. Future work might investigate ways to overcome the
political temptations of short-run resource wealth to create the financial and political institutions
needed to reduce volatility, soften the impact of volatility on growth and prevent poverty.
References
Aghion, P. and G. Saint Paul (1998). On the virtues of bad times, Macroeconomic Dynamics, 2,
3, 322-344.
Aghion, P., G.-M. Angeletos, A. Banerjee and K. Manova (2005). Volatility and growth: credit
constraints and productivity-enhancing investments, Working Paper 11349, NBER,
Cambridge, Ma.

17
Aghion, P., P. Bacchetta, R. Rancière and K. Rogoff (2006). Exchange rate volatility and
productivity growth: the role of financial development, Discussion Paper No. 5629,
CEPR, London.
Aizenman, J. and N. Marion (1991). Policy uncertainty, persistence and growth, Review
ofInternational Economics, 1, 2, 145-163.
Aizenman, J. and N. Marion (1999). Volatility and investment: interpreting evidence from
developing countries, Economica, 66, 157-179.
Alesina, A., A. Devleeschauwer, W. Easterly, S. Kurlat and R. Wacziarg (2003).
Fractionalization, Journal of Economic Growth, 8, 155-194.
Alexeev, M. and R. Conrad (2005). The elusive curse of oil, Working Paper SAN05-07, Terry
Sandford Institute of Public Policy, Duke University, Durham.
Barro, R.J. and J.-W. Lee (2000). International data on educational attainment: updates and
implications, Working Paper No. 42, Centre for International Development, Harvard
University.
Bernanke, B.S. (1983). Irreversibility, uncertainty and cyclical investment, Quarterly Journal of
Economics, 98, 1, 85-106.
Bertola, G. (1994). Flexibility, investment, and growth, Journal of Monetary Economics, 34, 2,
215-238.
Black, F. (1987). Business Cycles and Equilibrium, Basil Blackwell, New York.
Blattman, C., J. Hwang and J.G. Williamson (2007). Winners and losers in the commodity
lottery: The impact of terms of trade growth and volatility in the Periphery 1870-1939,
Journal of Development Economics, 82, 156-179.
Bloom, D.E. and J. Sachs (1998). Geography, demography, and economic growth in Africa,
Brookings Papers on Economic Activity, 2, 207-295.
Brunnschweiler, C. and E.H. Bulte (2008). The natural resource curse revised and revisited: A
tale of paradoxes and red herrings, Journal of Environmental Economics and
Management, 55, 3, 248-264.
Cameron, A.C. and P.K. Trivedi (2005). Microeconometrics: Methods and Applications,
Cambridge University Press, New York.
Caselli, F. and T. Cunningham (2009). Leader behavior and the natural resource curse, Oxford
Economic Papers, this issue.
Céspedes and R. Sahay (2002). Keyens, cocoa, and copper: In search of commodity
currencies, Working Paper WP/02/223, International Monetary Fund, Washington, D.C.

18
Deaton, A. (1999). Commodity prices and growth in Africa, Journal of Economic Perspectives,
13, 3, 23-40.
Ding, N. and B.C. Field (2005). Natural resource abundance and economic growth, Land
Economics, 81, 3, 496-502.
Dixit, A. and R. Rob (1994). Switching costs and sectoral adjustments in general equilibrium with
uninsured risk, Journal of Economic Theory, 62, 1, 48-69.
Easterly, W., M. Kremer, L. Pritchett and L.H. Summers (1993). Good policy or good luck?
Country growth performance and temporary shocks, Journal of Monetary Economics, 32,
459-483.
Easterly, W. and R. Levine (1997). Africa’s growth tragedy: policies and ethnic divisions,
Quarterly Journal of Economics, 12, 1203-1250.
Edwards, J.A. (2007). Dynamics of the cross-country growth/volatility relationship, Global
Economy Journal, 7, 2, 4.
Elbers, C., J.W. Gunning and B. Kinsey (2007). Growth and risk: Methodology and Micro
Evidence, The World Bank Economic Review, 21, 1, 1-20.
Fatás, A. and I. Mihov (2005). Policy volatility, institutions and economic growth, Discussion
Paper No. 5388, CEPR, London.
Flug, K., A. Splilimbergo and E. Wachtenheim (1999). Investment in education: do economic
volatility and credit constraints matter?, Journal of Development Economics, 55, 2, 465-
481.
Gallup, J.L, J.D. Sachs and A. Mellinger (1999). Geography and economic development,
Working Paper No. 1, Centre for International Development, Harvard University.
Gelb, A. (1988) (ed.), Oil Windfalls: Blessing or Curse?, Oxford University Press, Oxford.
Grilli, V. and G.M. Milesi-Ferretti, (1995). Economic effects and structural determinants of
capital controls, IMF Staff Papers 42, 3, 517–551.
Gylfason, T., T.T. Herbertsson and G. Zoega (1999). A mixed blessing: natural resources and
economic growth, Macroeconomic Dynamics, 3, 204-225.
Hausmann, R. and R. Ribogon (2002). An alternative interpretation of the ‘resource curse’:
theory and policy implications, Working Paper No. 9424, NBER, Cambridge, MA.
Hausmann, R., U. Panizza and R. Rigobon (2004). The long-run volatility puzzle of the real
exchange rate, Working Paper No. 10751, NBER, Cambridge, MA.
Herbertsson, T.T., M. Skuladottir and G. Zoega (2000). Three symptoms and a cure: A
contribution to the economics of the Dutch Disease, Discussion Paper No. 2364, CEPR,
London.

19
Heston, A., R. Summers and B. Aten (2006). Penn World Table 6.2, Center for International
Comparisons of Production, Income and Prices, University of Pennsylvania.
Imbs, J. (2006). Growth and volatility, Journal of Monetary Economics, 54, 7, 1848-1862.
International Monetary Fund (2006, and earlier). Annual Report on Exchange Arrangements and
Restrictions, International Monetary Fund, Washington, D.C.
Koren, M. and S. Tenreyro (2007). Volatility and Development, Quarterly Journal of Economics,
122, 1, 243-287.
Kose, M.A., E.S. Prasad and M.E. Torrones (2003). Financial integration and macroeconomic
volatility, IMF Staff Papers, 50, 119-142.
Kose, M.A., E.S. Prasad and M.E. Torrones (2006). How do trade and financial integration affect
the relationship between growth and volatility?, Journal of International Economics, 69,
176-202.
Loayza, N.V., R. Rancière, L. Servén and J. Ventura (2007). Macroeconomic volatility and
welfare in developing countries: an introduction, World Bank Economic Review, 21, 3,
343-357.
Malik, A. and J. Temple (2006). The geography of output volatility, Journal of Development
Economics, forthcoming.
Mansoorian, A. (1991). Resource discoveries and ‘excessive’ external borrowing, Economic
Journal, 101, 1497-1509.
Martin, P. and C.A. Rogers (2000). Long-term growth and short-term economic instability,
European Economic Review, 44, 359-381.
Mirman, L.J. (1971). Uncertainty and optimal consumption decisions, Econometrica, 39, 179-
185.
Montalvo, J.G. and M. Reynal-Querol (2005a). Ethnic diversity and economic development,
Journal of Development Economics, 76, 293-323.
Montalvo, J.G. and M. Reynal-Querol (2005b). Ethnic polarization, potential conflict, and civil
wars, American Economic Review, 95, 797-816.
Pallage, S. and M.A. Robe (2003). On the welfare cost of economic fluctuations in developing
countries, International Economic Review, 44, 2, 677-698.
Pagan, A.R. (1984). Econometric Issues in the Analysis of Regressions with Generated
Regressors, International Economic Review, 25(1), 221-247.
Pindyck, R.S. (1991). Irreversibility, uncertainty, and investment, Journal of Economic Literature,
129, 3, 1110-1148.
Political Risk Services Group, (2006). International Country Risk Guide, East Syracuse, NY.

20
Ramey, G. and V.A. Ramey (1991). Technology commitment and the cost of economic
fluctuations, Working Paper No. 3755, NBER, Cambridge, MA.
Ramey, G. and V.A. Ramey (1995). Cross-country evidence on the link between volatility and
growth, American Economic Review, 85, 5, 1138-1151.
Rose, A.K. and M.M. Spiegel (2007). International financial remoteness and macroeconomic
volatility, Journal of Development Economics, forthcoming.
Ross, M.L. (1999). The political economy of the resource curse, World Politics, 51, 297-322.
Sachs, J.D. and A.M. Warner (1995 and revised 1997a). Natural resource abundance and
economic growth, in G. Meier and J. Rauch (eds.), Leading Issues in Economic
Development, Oxford University Press, Oxford.
Sachs, J.D. and A.M. Warner (1997b). Fundamental sources of long-run growth, American
Economic Review, Papers and Proceedings, 87, 2, 184-188.
Sachs, J.D. and A.M. Warner (2001). The curse of natural resources, European Economic
Review, 45, 827-838.
Stock, J.H. and M. Yogo (2002). Testing for weak instruments in linear IV regression, Technical
Working Paper 284, NBER, Cambridge, MA.
UNCTAD (2007). Commodity price statistics, United Nations Conference on Trade and
Development, New York, NY.
Wacziarg, R. and K.H. Welch (2008). Trade liberalization and growth: new evidence, World
Bank Economic Review, 22, 2, 187-231.
World Bank (2006). World Development Indicators, World Bank, Washington, D.C.
World Bank (2007). Adjusted Net Savings Data, Environmental Economics and Indicators,
World Bank, Washington, D.C.

21
Appendix 1: Countries Included in our Core Estimates
Algeria France Pakistan
Argentina Ghana Panama
Australia Greece Peru
Austria Guatemala Philippines
Belgium Honduras Portugal
Benin India Senegal
Bolivia Ireland Spain
Brazil Israel Sri Lanka
Cameroon Italy Sweden
Canada Japan Switzerland
Central African Republic Jordan Thailand
Chile Korea, Rep. Togo
Colombia Malawi Trinidad and Tobago
Congo, Dem. Rep. Malaysia Tunisia
Congo, Rep. Mali Turkey
Costa Rica Mexico United Kingdom
Denmark Netherlands United States
Ecuador New Zealand Uruguay
Egypt, Arab Rep. Nicaragua Venezuela, RB
El Salvador Niger Zambia
Finland Norway

22
Appendix 2: Description of the Cross-Country Data
VARIABLE NAME DEFINITION SOURCE
GDP/capita growth rate Annual ln difference in real GDP per capita, Laspeyres; Averages are taken by
country across the given period and annual growth rates. PWT 6.2 from Heston et al.
(2006)
Investment share of GDP Gross fixed capital formation as % of GDP PWT 6.2 from Heston et al.
(2006)
Average population growth
rate Ln difference in total population PWT 6.2 from Heston et al.
(2006)
log per capita GDP Ln real GDP per capita PWT 6.2 from Heston et al.
(2006)
Human capital Average schooling years in the population (age 25+) Barro & Lee (2000)
Rule of Law 1984 A country’s score on the law and order index in 1984 (first year available). PRS Group (2006)
Total resources The sum of point-source resources and diffuse resources. WDI (2006)
Point-source resources F.o.b. value of exported fuels + ores & metals as a percentage of GDP WDI (2006)
Diffuse resources F.o.b. value of exported foods and agricultural raw materials as a percentage of
GDP WDI (2006)
Fuels F.o.b. value of exports as a percentage of GDP. Corresponds to SITC section 3
(mineral fuels). WDI (2006)
Ores & Metals F.o.b. value of exports as a percentage of GDP. Commodities in SITC divisions 27,
28, and 68 (nonferrous metals). WDI (2006)
Agricultural Raw Materials F.o.b. value of exports as a percentage of GDP. Corresponds to SITC section: 2
(crude materials except fuels) excluding divisions 22, 27 (crude fertilizers and
minerals excluding coal, petroleum, and precious stones), and 28 (metalliferous
ores and scrap).
WDI (2006)
Foods F.o.b. value of exports as a percentage of GDP. Commodities in SITC sections: 0
(food and live animals), 1 (beverages and tobacco), and 4 (animal and vegetable
oils and fats) and SITC division 22 (oil seeds, oil nuts, and oil kernels).
WDI (2006)
Monthly world commodity
prices Monthly averages of free-market price indices for all food, agricultural raw
materials, minerals, ores & metals, crude petroleum (average of Dubai/Brent/Texas
equally weighted). Base year 2000 = 100.
UNCTAD, 2007
Financial development Domestic credit to private sector (% of GDP) WDI (2006)
Sachs Warner updated
openness dummy open to trade = 1 Wacziarg & Welch (2003)
Fraction of years open to
trade number of total years open to trade divided by years in sample Wacziarg & Welch (2003)
Landlocked dummy =1 if a country has no access to sea Gallup et al. (1999)
% population in temperate
climate zone % 1995 pop in Koeppen-Geiger temperate zones (Cf+Cs+Df+DW) CID, General Measures of
Geography, 2007
Distance to nearest
navigable river or coast minimum distance in km, fixed effect CID, General Measures of
Geography, 2007
Life expectancy 1970 Life expectancy at birth WDI (2006)
Ethnic Polarization Index of ethno-linguistic polarization (0: many small groups, to 1: two large
groups) Montalvo & Reynal-Querol
(2005)
Ethnic Fractionalization Index of ethno-linguistic fractionalization (0 to 1), the probability that two
randomly selected individuals from a given country will not belong to the same
ethnic group.
Montalvo & Reynal-Querol
(2005)
Multiple Exchange
Practices dummy, yes = 1 IMF (2006)
Current Account
Restrictions dummy, yes = 1 IMF (2006)
Capital Account
Restrictions dummy, yes = 1 IMF (2006)
Surrender of Export receipts dummy, yes = 1 IMF (2006)
Government spending
volatility standard deviation of yearly share of government expenditure of GDP PWT 6.2 from Heston et al.
(2006)
sd ToT index growth standard deviation of yearly terms-of-trade index growth rate, where the terms-of-
trade index is defined as the value of total exports over total imports PWT 6.2 from Heston et al.
(2006)

23
Appendix 3: Econometric methodology
We use a dataset with N countries and a sample period of T years. Extending Ramey and Ramey
(1995) we specify the following econometric model for growth in GDP per capita:
70 70
log( ) ,
it i i i it
y
λ
σε
Δ=+++XθZβ 270
exp( )
ii
c
σ
=
+Zγ and
2
(0, ), 1,.., , 1,.., .
it i
NiNt
εσ
==T
where yit is GDP per capita in country i for year t,
σ
i is the standard deviation for country i of the
error term
ε
it, Xi70 is a vector of control variables for country i and year 1970, and θ is a vector of
coefficients assumed to be constant across countries. The errors
ε
it are the deviations of growth
from the predicted values based on the controls. Average volatility σi is assumed constant over
time, but differs for each country depending on the initial country characteristics captured in Zi70.
We also allow for direct effects of these variables on growth (β). We estimate parameters {
λ
, θ,
γ, c and β} by maximizing the log-likelihood function. The error terms are assumed to be
uncorrelated across countries.
Table 1: Growth, Volatility, Financial Development and Resources in the World Economy
Regional Characteristics (%, 1970-2003, at least 10 observations per country)
Export Value Share of GDP
Region
Yearly real GDP
per capita growth
rate
Fuels, Ores
& Metals
Agricultural
Raw Materials,
Foods All Resources Government
Share Financial
Development
mean sd mean sd mean sd mean sd sd mean
Middle East & North Africa (MENA) 1.18 8.12 22.24 9.30 2.51 1.52 24.75 9.07 5.82 41.41
Sub-Saharan Africa (SSA) 0.47 6.52 9.60 3.97 10.24 3.60 19.65 5.66 4.76 17.44
East Asia & Pacific (EAP) 2.47 5.00 6.81 3.45 10.04 3.11 16.71 5.49 2.72 51.77
Latin America & Caribbean (LAC) 1.47 4.54 4.99 2.64 9.66 3.70 14.59 5.34 3.98 34.87
South Asia (SA) 2.41 4.41 0.52 0.42 4.25 1.55 4.77 1.83 2.98 17.33
Eastern Europe & Central Asia (ECA) 2.56 4.34 2.07 0.66 3.50 1.03 5.57 1.54 2.52 22.70
Western Europe (WE) 2.35 2.33 2.71 1.00 5.20 0.95 7.86 1.60 1.53 76.08
North America (NA) 2.09 1.90 2.90 0.52 2.99 0.45 5.88 0.85 1.60 109.36
1st q. Av. Fin. Development (<=16.2) 0.70 6.40 9.71 4.23 7.64 3.00 17.06 5.52 4.64 10.38
4th q. Av. Fin. Development (>=52.9) 2.32 4.40 4.68 2.29 5.28 1.78 9.89 3.45 3.03 80.92
1st q. Av. Resource Dep. (<=6.1) 2.73 2.83 1.17 0.48 2.23 0.64 3.41 0.93 2.38 64.96
4th q. Av. Resource Dep. (>=19.3) 1.08 7.37 23.22 10.00 11.62 3.59 34.67 10.85 4.72 25.47
1st q. Distance to waterway (<=49km) 1.76 8.12 6.72 3.41 8.22 2.65 24.75 9.07 5.82 41.41
4th q. Distance to waterway (>=359km) 1.46 6.52 8.22 3.68 8.59 3.43 19.65 5.66 4.76 17.44
Note: Means are cross-country averages of country average growth rates or variable shares between 1970 and 2003. Standard deviations (sd) are the
average cross-country standard deviations of country yearly growth rates or variable shares over the corresponding period.

24
Table 2: Does the Traditional Natural Resource Curse Really Exist?
Dependent Variable
(constant 2000 international dollars, PWT 6.2) average GDP growth per
capita 1970-2003
(1) (2)
Annual growth Equation
Total resources 1970 -0.027** -0.012
(0.013) (0.013)
Initial log per capita GDP -0.012*** -0.010**
(0.004) (0.004)
Fraction of years open to trade 0.151*** 0.177***
(0.040) (0.038)
Initial GDP/capita * fraction years open -0.017*** -0.021***
(0.005) (0.005)
Average investment share of GDP 0.041 0.041
(0.030) (0.028)
Rule of Law 1984 0.004** 0.004**
(0.001) (0.001)
Average yearly growth GDP per capita 60-70 -0.013 0.008
(0.072) (0.067)
Distance to nearest navigable river or coast -0.000 -0.000
(0.000) (0.000)
Fraction of population in temperate climate 0.003 0.006
(0.005) (0.005)
Life expectancy 1970 0.001** 0.001*
(0.000) (0.000)
Average population growth rate 1970-2003 -0.524** -0.470**
(0.251) (0.233)
Human capital 1970 0.001 0.001
(0.001) (0.001)
Standard deviation of GDP/capita growth -0.292***
(0.099)
Constant 0.061** 0.057**
(0.026) (0.024)
Observations 58 58
R-squared 0.75 0.79
Countries 58 58
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1.

25
Table 3: Effects of Various Commodity Exports on Volatility and Growth
Dependent Variable yearly GDP growth per
capita 1970-2003 yearly GDP growth per capita 1970-2003
(constant 2000 international dollars, PWT 6.2) (6a) (6b) (7a) (7b) (7c) (7d) (7e)
Annual growth equation
1st lag GDP per capita growth 0.221*** 0.220*** 0.232*** 0.230*** 0.230*** 0.226*** 0.226***
(0.025) (0.025) (0.026) (0.027) (0.027) (0.027) (0.028)
Average investment share of GDP ‘70-‘03 0.045* 0.045* 0.063** 0.065** 0.063** 0.065** 0.074***
(0.025) (0.025) (0.025) (0.026) (0.026) (0.026) (0.026)
Average population growth rate 1970-2003 -0.478*** -0.478*** -0.461*** -0.346** -0.343** -0.358** -0.307**
(0.144) (0.145) (0.133) (0.152) (0.149) (0.139) (0.147)
log per capita GDP 1970 -0.014*** -0.014*** -0.012*** -0.011*** -0.010*** -0.011*** -0.010***
(0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002)
Human capital 1970 0.002** 0.002** 0.001* 0.001 0.001 0.001 0.001
(0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001)
Volatility (σi) -0.971** -1.022*** -0.427*** -0.350** -0.334** -0.426*** -0.388***
(0.378) (0.297) (0.129) (0.141) (0.148) (0.148) (0.127)
Point based resources 1970 0.050* 0.054* 0.014 0.008 0.005 0.018 0.016
(0.030) (0.028) (0.023) (0.023) (0.029) (0.023) (0.022)
Financial development 1970 -0.018** -0.018*** -0.010* -0.008 -0.008 -0.009 -0.007
(0.007) (0.006) (0.005) (0.005) (0.005) (0.005) (0.005)
Sachs Warner updated openness dummy 70 -0.006 -0.007* 0.001 0.002 0.003 0.001 0.001
(0.005) (0.004) (0.003) (0.003) (0.003) (0.003) (0.003)
Constant 0.170*** 0.174*** 0.121*** 0.107*** 0.106*** 0.115*** 0.104***
(0.030) (0.027) (0.018) (0.018) (0.019) (0.019) (0.015)
Variance equation
Initial point-source resources 1970 1.621*** 2.125*** -0.426 -0.720 -0.493 -0.563 -1.247***
(0.589) (0.596) (0.488) (0.634) (0.645) (0.862) (0.337)
Initial diffuse resources 1970 0.801 0.807 -0.897*** -0.133 -1.076 0.167 0.483
(0.514) (0.497) (0.323) (0.638) (0.974) (0.430) (0.378)
Initial financial development 1970 -1.290*** -1.266*** -1.063*** -0.858*** -0.842*** -0.754*** -0.594***
(0.072) (0.121) (0.136) (0.096) (0.226) (0.166) (0.153)
Sachs Warner updated openness dummy 1970 -0.693*** -0.700*** -0.467*** -0.536*** -0.487** -0.545*** -0.215**
(0.160) (0.160) (0.180) (0.174) (0.207) (0.164) (0.095)
Distance to nearest navigable river or coast 0.001*** 0.001*** 0.001*** 0.001*** 0.001*** 0.000** 0.000*
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Financial development * point based share -2.780
(2.049)
Point-source export share volatility 70-03 9.303*** 9.528*** 15.837*** 14.491***
(0.774) (1.286) (1.141) (0.588)
Diffuse export share volatility 70-03 10.907*** 3.899* 1.841* 3.737***
(1.491) (2.004) (1.047) (1.377)
Government share volatility 70-03 10.525*** 10.406*** 9.786*** 8.372***
(1.179) (3.260) (2.709) (1.510)
Agricultural R.M. resource share volatility 70-03 0.631
(3.023)
Foods resource share volatility 70-03 10.916***
(1.690)
Ores & metals resource share volatility 70-03 6.626***
(2.543)
Fuels resource share volatility 70-03 9.513***
(1.719)
Financial development * point based volatility -34.343*** -29.620***
(6.542) (3.295)
sd ToT index growth 4.321***
(0.181)
Constant -6.100*** -6.093*** -6.517*** -6.751*** -6.826*** -6.711*** -7.401***
(0.062) (0.067) (0.030) (0.035) (0.057) (0.075) (0.020)
Observations 2084 2084 2084 2084 2084 2084 2084
Log likelihood 3732.3 3732.5 3792.2 3814.4 3815.2 3819.0 3842.8
Countries 62 62 62 62 62 62 62
Robust and clustered (by country) standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

26
Table 4: Ethnic Tensions, Economic Restrictions and the Resource Curse
Ethnic tensions (8a) (8b) Economic Restrictions (9a) (9b)
Annual growth equation Annual growth equation
1st lag GDP per capita growth 0.219*** 0.229*** 1st lag GDP per capita growth 0.220*** 0.228***
(0.028) (0.027) (0.028) (0.029)
Average investment share of GDP 70-03 0.053** 0.069*** Average investment share of GDP 70-03 0.077*** 0.081***
(0.026) (0.026) (0.018) (0.019)
Average population growth rate 1970-2003 -0.451*** -0.244 Average population growth rate 70-03 -0.426*** -0.394***
(0.153) (0.160) (0.130) (0.114)
log per capita GDP 1970 -0.013*** -0.010*** log per capita GDP 1970 -0.012*** -0.010***
(0.002) (0.002) (0.002) (0.002)
Human capital 1970 0.001** 0.001 Human capital 1970 0.002*** 0.001*
(0.001) (0.001) (0.001) (0.001)
Volatility (σi) -0.686*** -0.320*** Volatility (σi) -0.490*** -0.337***
(0.212) (0.121) (0.153) (0.114)
Initial point-source resources 70 0.019 0.008 Initial point-source resources 70 0.064*** 0.037**
(0.037) (0.028) (0.021) (0.019)
Financial development 1970 -0.014** -0.007 Financial development 1970 -0.012** -0.006
(0.005) (0.005) (0.005) (0.005)
Sachs Warner updated openness dummy 70 -0.002 0.002 Current Account Restrictions (yes=1) 0.004* 0.002
(0.004) (0.003) (0.002) (0.002)
Ethnic Polarization 0.002 -0.003 Capital Account restrictions (yes=1) -0.005* -0.003
(0.004) (0.004) (0.003) (0.003)
Constant 0.143*** 0.098*** Constant 0.125*** 0.105***
(0.023) (0.014) (0.017) (0.012)
Variance equation
Variance equation
Initial point based resources 70 -4.785*** -1.348** Initial point based resources 70 5.314*** 0.016
(0.395) (0.542) (0.340) (0.455)
Initial diffuse resources 70 0.863 0.213 Initial diffuse resources 70 2.082** 0.121
(0.611) (0.523) (0.862) (0.878)
Initial financial development 1970 -1.140*** -0.683*** Initial financial development 1970 -1.232*** -0.582***
(0.113) (0.095) (0.197) (0.132)
Sachs Warner updated openness dummy 70 -0.624*** -0.187
(0.097) (0.139)
Distance to nearest navigable river or coast 0.001*** 0.000** Distance to nearest navigable river or
coast 0.001*** 0.000
(0.000) (0.000) (0.000) (0.000)
Ethnic Polarization 0.402*** 0.056 Ethnic Polarization 0.965*** 0.296***
(0.088) (0.127) (0.122) (0.058)
Point-source export share volatility 70-03 8.834*** Point-source export share volatility 70-
03 7.811***
(1.327) (0.542)
Diffuse export share volatility 70-03 5.827** Diffuse export share volatility 70-03 10.504***
(2.601) (1.118)
Government share volatility 70-03 8.725*** Government share volatility 70-03 3.752*
(2.072) (2.179)
sd TOT index growth 4.537*** sd TOT index growth 4.166***
(0.483) (0.322)
Constant -6.406*** -7.491*** Constant -7.303*** -7.813***
(0.043) (0.034) (0.027) (0.048)
Point-source resources 70 * Eth. Pol. 8.536*** Multiple Exchange Practices (yes=1) -0.759*** -0.438**
(0.369) (0.186) (0.178)
Current Account Restrictions (yes=1) 0.426*** 0.511***
(0.105) (0.069)
Capital Account restrictions (yes=1) -0.294*** -0.311***
(0.045) (0.070)
Surrender of Export receipts (yes=1) 0.384*** 0.242***
(0.081) (0.082)
Cur. Acc. Restrictions * Point Resources
70 4.877***
(1.080)
Cap. Acc. Restrictions * Point
Resources 70 -2.508***
(0.634)
Observations 2084 2084 Observations 2013 2013
Log likelihood 3748.2 3840.1 Log likelihood 3622.6 3707.2
Countries 62 62 Countries 60 60
Robust and clustered (by country) standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1.

27
Table 5: Counterfactual Experiments for Resource-Rich and Landlocked Africa
Resource-Rich Africa versus South-East Asian sample
mean South-
East Asia Resource-
rich Africa Difference on volatility
on yearly
GDP/capita
growth rate
GDP per capita growth 1.49% 4.04% 0.25% -3.79%
Mean equation
1st lag GDP per capita growth 0.221*** 1.48% 4.00% 1.07% -2.94% 0.65%
Average investment share of GDP 1970-2003 0.045* 17.26% 24.45% 14.96% -9.50% 0.43%
Average population growth rate 1970-2003 -0.478*** 1.72% 1.86% 2.75% 0.89% 0.43%
Initial log per capita GDP 1970 -0.014*** 8.362 7.747 7.129 -0.619 -0.87%
Initial human capital 1970 0.002** 4.140 4.049 1.476 -2.574 0.46%
Volatility (σi) -0.971** 4.04% 3.43% 6.02% 2.59% 2.98%
Initial point-source resources 1970 0.050* 4.35% 4.32% 13.13% 8.80% -0.44%
Initial financial development 1970 -0.018** 29.07% 26.89% 14.43% -12.47% -0.22%
Variance equation
Initial point-source resources 1970 1.621*** 4.35% 4.32% 13.13% 8.80% -0.41% 0.52%
Initial diffuse resources 1970 0.801 7.27% 11.08% 10.52% -0.56% 0.01% -0.02%
Initial financial development 1970 -1.290*** 29.07% 26.89% 14.43% -12.47% -0.47% 0.58%
Sachs Warner updated openness dummy 70 -0.693*** 0.374 0.746 0 -0.746 -1.37% 1.71%
Distance to nearest navigable river or coast 0.001*** 277.763 90.902 552.571 461.669 -0.86% 1.07%
Estimated volatility 4.04% 3.43% 6.02% 2.59%
Countries 62 4 6
Note: Resource-rich African counties are Algeria, Congo, Rep.. Ghana, Malawi, Togo, and Zambia. The South-East Asian countries in our
sample are South Korea, Malaysia, Philippines and Thailand. The calculations are based on regression (6a). The effect of each variable on the
growth rate (or on volatility) is measured as the effect of changing the respective variable to the sample mean level of the South-East Asian
countries, while keeping all other variables constant.
Landlocked Africa versus South-East Asia sample
mean South-
East Asia Landlocked
Africa Difference on volatility
on yearly
GDP/capita
growth rate
GDP per capita growth 1.49% 4.04% 0.22% -3.82%
Mean equation
1st lag GDP per capita growth 0.221*** 1.48% 4.00% 0.50% -3.51% 0.78%
Average investment share of GDP 1970-2003 0.045* 17.26% 24.45% 12.13% -12.32% 0.56%
Average population growth rate 1970-2003 -0.478*** 1.72% 1.86% 2.57% 0.71% 0.34%
Initial log per capita GDP 1970 -0.014*** 8.362 7.747 6.744 -1.004 -1.41%
Initial human capital 1970 0.002** 4.140 4.049 0.874 -3.176 0.57%
Volatility (σi) -0.971** 4.04% 3.43% 6.88% 3.45% 1.97%
Initial point-source resources 1970 0.050* 4.35% 4.32% 10.97% 6.65% -0.33%
Initial financial development 1970 -0.018** 29.07% 26.89% 12.05% -14.84% -0.27%
Variance equation
Initial point-source resources 1970 1.621*** 4.35% 4.32% 10.97% 6.65% -0.36% 0.45%
Initial diffuse resources 1970 0.801 7.27% 11.08% 7.99% -3.09% 0.09% -0.11%
Initial financial development 1970 -1.290*** 29.07% 26.89% 12.05% -14.84% -0.63% 0.78%
Sachs Warner updated openness dummy 70 -0.693*** 0.374 0.746 0 -0.746 -1.56% 1.95%
Distance to nearest navigable river or coast 0.001*** 277.763 90.902 979.419 888.516 -1.76% 2.19%
Estimated volatility 4.04% 3.43% 6.88% 3.45%
Countries 62 4 5
Note: Landlocked Africa are Central African Republic, Malawi, Mali, Niger and Zambia. South-East Asian countries are South Korea, Malaysia,
Philippines and Thailand. The calculations are based on regression (6a). The effect of each variable on the growth rate (or on volatility) is
measured as the effect of changing the respective variable to the sample mean level of South-EastAsia, while keeping all other variables constant.

28
Figure 1: Volatility Correlated with Low Growth in GDP per Capita
Equatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial GuineaEquatorial Guinea
IraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraqIraq
LiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberia
United Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab Emirates
-5 0 5 10
Average Yearly GDP/Capita Growth
(1970-2003, %)
0 10 20 30
Standard Deviation of Yearly GDP/Capita Growth
(1970-2003, %)
Fitted Values (slope = -.247***; Adj. R2=.14; n=150)
Landlocked countries
Figure 2: Resource-Rich Economies Are More Volatile
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
Baham
a
LiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberiaLiberia
United Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab EmiratesUnited Arab Emirates
010 20 30
Standard Deviation of Yearly
GDP/Capita Growth (1970-2003, %)
0 20 40 60 80 100
Average Resource Share of GDP
(1970-2003, %)
Fitted Values (slope =.149 (.016); Adj. R2=.44)
Note: Resource share measures the total of food, agricultural raw materials, mining and fuel export revenue, as a
percentage of GDP, average over the period 1970-2003.

29
Figure 3: Cumulative density function of volatility of commodity prices
Note: The x-axis measures the yearly standard deviation of the monthly price index levels
Figure 4: Declining Natural Resource Dependence in the Global Economy
.2 .4 .6 .8 1
Ratio of Values of Primary Commodity Exports
to Total Exports
1970 1980 1990 2000 2010
Sub-Saharan Africa Latin America & Carib.
Middle East & North Africa East Asia & Pacific
Western Europe & North America South Asia

30
Figure 5: Marginal Effect of Point-Source Resource Dependence on Growth
-.04 -.02 0.02 .04
Marginal effect on growth of
point-source dependence
.01
.02
OECD
SEA
.04
LAC
.05
Malawi
RR. Africa
.0637
LL. Africa
.08
.09
Zambia
Volatility of unanticipated output growth, estimated 1970-2003
Note: RR. Africa = resource-rich Africa; LL. Africa = Landlocked Africa; SEA = South-East Asia,
corresponding to Table 7. Based on regression 6a of Table 4.