Download Dia 1

POLITICAL ECONOMY OF THE
NATURAL RESOURCE CURSE
Rick van der Ploeg
OxCarre, University of Oxford
1
OVERVIEW OF TALK







Historical and case-study evidence of resource
curse
Four explanations of resource curse
Cross-country and panel evidence
Focus on role of volatility and financial development
Sustainable management of natural resources,
Hartwick rule and genuine saving
Policy proposals
Summing up
2
POOR DESPITE NATURAL
RESOURCE WEALTH


Nigeria: oil revenues per capita increased from $33
in 1965 to $325 in 2000 but income per capita
stagnated at about $1100 since its independence in
1960. Between 1970 and 2000 those on less than
$1/day increased from 26 to almost 70%. Top 2%
had as much as bottom 17 % in 1970 but staggering
bottom 55% in 2000. Declining TFP growth: -1.2%
per year. Only a third of capacity is utilized.
Successive military dictatorships have plundered oil
wealth.
Hopefully, the future will be brighter.
3
GDP /
capita
1970
GDP /
capita
2005
Corruption
Rank/141
Congo
221
85
1
Nigeria
358
373
11
Indonesia
515
849
Botswana
1,611
Russian F
% less
than
$1 /
day
Gini
Law and
Order
63.0
0.99
70.8
43.7
2.03
136
7.5
34.3
2.69
3,311
101
23.5
56.6
4.47
1,962
2,004
28
2.0
39.9
3.47
Venezuela
5,442
4,530
57
8.3
44.1
3.62
Libya
4,815
6,905
89
Norway
24,895
38,075
133
3.26
0
25.8
6.00
Source: World Bank Development Indicators 2006 (2000 US $) and
International Country Risk Guide, The PRS Group Inc.
4
DISAPPOINTING PERFORMANCE
DESPITE NATURAL RESOURCES





17th century Spain despite gold/silver from New
World. Resource Holland did much better.
Negative growth rates during past decades: e.g.,
Venezuela, Iran,Libya, Kuwait, Quatar.
Decline in OPEC GDP/ capita during last few
decades while other countries enjoyed growth.
Gold boom in 70’s did not help South Africa much
(Stokke, 2005).
Dutch economy and the Slochteren natural gas
reserves led to unsustainable welfare state.
5
NEGATIVE PARTIAL CORRELATION BETWEEN
ECONOMIC GROWTH AND RESOURCE
ABUNDANCE (not for food or agriculture exports)
5
Korea, Rep.
Taiwan, China
Singapore
Malta
Hong Cambodia
Kong, China
Thailand
Indonesia
Ireland
Malaysia
Sri Lanka
Tunisia
Egypt,
Arab Rep.
Norway
Portugal
Chile
Hungary
Japan
Spain
Austria Finland
United
Kingdom Belgium
Italy
United
States
Pakistan
Turkey
Greece
France
Israel
Brazil
Canada
Jordan
Australia Morocco
Netherlands
Costa Rica
SwedenColombia
Denmark
Mexico
Sudan
Ecuador New Zealand
Paraguay
Panama
Philippines
Cameroon
Uruguay
Burkina Faso
Algeria
Switzerland
Guatemala
Congo,
Rep.
Malawi
Honduras
Mali
Chad
Nigeria
Argentina
Solomon Islands
SaoEl
Tome
and Principe
Bolivia
Benin
Salvador
Peru
Senegal
Ghana
Burundi
Togo
0
India
Central African Republic
Niger
Madagascar
Trinidad and Tobago
Fiji
Gabon
Mauritania
Suriname
Guyana
Saudi Arabia
Venezuela, RB
Cote d'Ivoire
Nicaragua
Zambia
Congo, Dem. Rep.
Kuwait
-5
Libya
Liberia
0
20
40
60
Natural Resources exports in percent of GDP, 1970
80
Figure: Growth and Natural Resources Abundance
Data source: World Development Indicators, 2006
Source: World Bank Development Indicators 2006
6
POSITIVE EXPERIENCES





Botswana: 40% of GDP stems from diamonds but has second
highest education/GNP and highest growth rates since 1965.
GDP/capita is ten times that of Nigeria. Still, inequality is high as
it was during the colonial period.
Norway had huge growth in oil exports since 1971 and third
largest exporter after Saudi-Arabia, but has fared fairly well.
United Arab Emirates also turned curse into blessing by
investing in modernizing infrastructure and investing in welfare
state and free access to education.
Mineral abundance US mid 19th to mid 20th century explains
much of subsequent growth: driven by learning, IRTS and US
government claimed no ultimate title to mineral rights
(Habbakuk, 1962; David and Wright, 1977; Wright and Czelusta,
2004). Also, Germany and UK late 19th century.
Lessons: avoid corruption, diversify, education, and exploit
complementarities linkages of manufacturing with resource
sector.
7
RESOURCE ABUNDANCE ASSOCIATED WITH
(Gylfason and Zoega, 2002):
 Crowding out of non-resource exports and
foreign direct investment. Less openness.
 Elicits corruption and extreme rent seeking.
 Crowds out foreign capital, social capital,
human capital and financial capital.
 Erodes legal system.
 Bigger Gini index of inequality.
 Less school enrolment and expected years of
schooling (Botwana exception).
 Delays development of financial institutions.
 Armed conflicts and civil wars.
8
6
CORRUPTION AND RESOURCE ABUNDANCE
Finland
Denmark
Sweden
Netherlands
Canada
New Zealand
Norway
Switzerland
4
United Kingdom
Australia
Austria
United States
Costa Rica
Belgium
France
Portugal
Hong Kong, China
Greece
Japan Hungary
Israel
Nicaragua
Ireland
Spain
Madagascar
Singapore
Malaysia
Italy
Malta
Taiwan, China
Chile
Brazil
Jordan
2
Sri Lanka
Korea,Argentina
Rep.
Malawi
Congo, Rep.
Uruguay
Ecuador
Burkina Faso
Senegal Algeria
Peru
Tunisia
Morocco
Mexico
Turkey
El Salvador
Colombia
India
Ghana
Thailand
Guatemala
Niger Egypt, Philippines
Arab
Rep.
Bolivia
Mali
Pakistan
Panama
Libya
Cote d'Ivoire
Venezuela, RB
Trinidad Kuwait
and Tobago
Zambia
Suriname
Saudi Arabia
Guyana
Honduras
Togo
Nigeria
Sudan
Indonesia
Paraguay
Gabon
Liberia
0
Congo, Dem. Rep.
0
20
40
60
Natural Resources exports in percent of GDP, 1970
80
Figure: Corruption and Natural Resource Abundance
Data source: ICRG(PRS Group) & World Development Indicators, 2006
9
4
BUREAUCRATIC QUALITY AND
RESOURCE ABUNDANCE
United
Switzerland
United
States
Kingdom
Sweden
Australia
Canada
Belgium
Denmark Netherlands
New Zealand
Japan
Finland
Austria
Ireland
Norway
France
Singapore
Israel
Hungary
3
Spain
Korea, Rep. Taiwan, China
Italy
India
Hong Kong, China
Thailand
Greece
Portugal
Brazil Colombia
Malta
Chile
Argentina
Mexico
Turkey
2
Jordan
Morocco
Ecuador
Pakistan Egypt,Tunisia
Arab Rep.
Niger
Malaysia
Gabon
Saudi Arabia
Cote d'Ivoire
Ghana
Costa
Rica
Sri Lanka
Kuwait
Senegal
Philippines
Algeria
Venezuela, RB
Uruguay
Peru
Indonesia
Madagascar
Nigeria
Burkina Faso
Suriname
Malawi
Panama
Congo, Rep.
Bolivia
Guatemala
Paraguay
El Salvador
1
Trinidad and Tobago
Honduras
Nicaragua
Guyana
Libya
Zambia
0
Sudan
Togo
Congo, Dem. Rep.
Mali
0
Liberia
20
40
60
Natural Resources exports in percent of GDP, 1970
80
Figure: Bureaucracy and Natural Resource Abundance
Data source: ICRG(PRS Group) & World Development Indicators, 2006
10
6
RULE OF LAW AND RESOURCE ABUNDANCE
Austria
Sweden
Finland
Norway
Canada
Denmark Netherlands
Australia
New Zealand
United
States
Switzerland
Belgium
United Kingdom
Japan
Singapore
5
France
Italy Portugal
Hungary
Hong Kong, China
Ireland
Taiwan, China
Chile
Spain
Saudi Arabia
Malta
4
Morocco
Thailand
Israel
Greece
Tunisia
Korea, Rep.
Argentina
Ecuador
Turkey
Burkina Faso
Jordan
Egypt, Arab Rep.
Kuwait
Costa Rica
Malaysia
Trinidad and Tobago
Venezuela, RB
Libya
3
India
Cote d'Ivoire
Brazil
Mexico
Mali
Niger
Pakistan
Madagascar
Uruguay
Malawi
Paraguay
Indonesia
Panama
Togo
Senegal
Philippines Ghana
Algeria
Peru
El Salvador
Bolivia
Sri Lanka
Sudan
Congo,
Rep.
Nigeria
2
Zambia
Nicaragua
Gabon
Guyana
Honduras
Suriname
Guatemala
Liberia
1
Colombia
Congo, Dem. Rep.
0
20
40
60
Natural Resources exports in percent of GDP, 1970
80
Figure: Rule of Law and Natural Resource Abundance
Data source: ICRG(PRS Group) & World Development Indicators, 2006
11
10
GOVERNMENT STABILITY AND
RESOURCE ABUNDANCE
Switzerland
9
Singapore
Finland
Taiwan, China
Ireland
United States
Morocco
Malta
Netherlands
7
8
Austria
Australia
Egypt,
Arab Rep.
United
Spain
Kingdom
Tunisia
Jordan
Sweden
Algeria
Japan
Canada
France
Portugal
Belgium
Denmark Senegal
Norway
New Zealand
Ghana
Hungary
Thailand
Korea, Rep.
Congo, Rep.
Greece
Turkey
Mexico
Indonesia
Chile
Uruguay
Pakistan
Costa Rica
Italy
Hong Kong, China
6
Burkina Faso Madagascar
Israel
Colombia
Argentina
Brazil
Ecuador
Nigeria
India
Guatemala
Togo
Mali
Paraguay
Niger
Philippines
Bolivia
Panama El Salvador
Peru
Malaysia
Saudi Arabia
Libya
Kuwait
Gabon
Venezuela, RB
Trinidad and Tobago
Nicaragua
Honduras
Guyana
Cote d'Ivoire
Suriname
Sri Lanka
Malawi
Sudan
Zambia
5
Congo, Dem. Rep.
Liberia
0
20
40
60
Natural Resources exports in percent of GDP, 1970
80
Figure: Goverment stability and Natural Resource Abundance
Data source: ICRG(PRS Group) & World Development Indicators, 2006
12
150
DOMESTIC CREDIT AND
RESOURCE ABUNDANCE
Hong Kong, China
United States
100
Japan
Switzerland
Singapore
Sweden
France
Portugal
United
Kingdom
Spain
Austria
Malaysia
Netherlands
Thailand
50
Malta Canada
Panama
Norway
Italy Israel
Korea, Rep.
Finland
Jordan
Denmark Ireland
New Zealand
Tunisia
Brazil
Greece
Hungary
Australia
Chile
Belgium
Kuwait
Barbados
Saudi Arabia
0
Uruguay
Philippines
Egypt, ArabMorocco
Rep.
El SalvadorAlgeria Honduras
Venezuela, RB
Indonesia
Bolivia
Nicaragua
Cote d'Ivoire Mauritania
Colombia
Senegal
Fiji Suriname
IndiaPakistan
Ecuador
Mexico
Costa
Rica
Solomon
Islands
Argentina
TogoSri Lanka
Paraguay
Benin
Mali
Peru Cameroon
Guatemala
Madagascar
Gabon
Congo, Rep.
Malawi
Nigeria
Sao Tome and
Principe
Burkina
Faso
Turkey
Burundi
Niger
Republic
Chad Central African
Sudan Ghana
Cambodia
Congo, Dem. Rep.
0
Trinidad and Tobago
Guyana
Liberia
Libya
Zambia
20
40
60
Natural Resources exports in % of GDP, 1970
80
Figure: Financial Development and Natural Resource Abundance
Data source: World Development Indicators, 2006
13
12
External Conflict 1984-2005
Netherlands
FinlandBelgium
New Zealand
Portugal
Switzerland
Austria
Sweden
Denmark Ireland
Brazil
Italy
Canada
Norway
Japan
Uruguay
Philippines
Paraguay
Spain
Hungary
Malta
10
Australia
Madagascar
Chile
France
Mexico
Argentina Indonesia
Niger
Bolivia
United Kingdom
Trinidad and Tobago
Malaysia
Suriname
Gabon
Malawi
Algeria
Nigeria
Singapore
Venezuela, RB
Greece
Cote d'Ivoire
Taiwan, China
EXTERNAL
CONFLICT AND
RESOURCE
ABUNDANCE
United States
Turkey
Jordan
Ghana
Congo, Rep.
Egypt,Tunisia
Arab Rep.Senegal
Colombia
Thailand
Peru
El Salvador
Costa Rica
Ecuador
Guatemala
Panama
Morocco
Zambia
Sri Lanka
Hong Kong, China
Korea,
Rep. Faso
Burkina
Saudi Arabia
Togo
Mali
8
Guyana
India
Libya Kuwait
Nicaragua
Honduras
Pakistan
Congo, Dem. Rep.
Israel
Sudan
6
Liberia
0
20
40
60
Natural Resources exports in percent of GDP, 1970
80
80
Japan
Switzerland
Spain
Italy UnitedFrance
Kingdom
70
United States
Sweden
Norway
Canada
Greece Hong Kong, China Australia
Denmark
Finland
Belgium
Malta
Israel
Austria
Taiwan, China
Portugal
UruguayPanama
Argentina
Chile
Hungary
Mexico
Korea, Rep.
Jordan
Brazil
Turkey
60
Netherlands
New
Zealand
Ireland
Costa
Rica
Barbados
Sri Lanka
Pakistan
India
50
LIFE EXPECTANCY
AND RESOURCE
ABUNDANCE
Benin
Burkina FasoBurundi
Afghanistan
40
Life expectancy 1970-2004
90
Figure: External Conflict and Natural Resource Abundance
Data source: ICRG(PRS Group) & World Development Indicators, 2006
Chad
Singapore
Kuwait
Trinidad and Tobago
Venezuela, RB
Malaysia
Suriname
Paraguay
Colombia
Thailand
Fiji
Ecuador
Tunisia
Saudi Arabia
Philippines
Guyana
Libya
Peru
Algeria Honduras
Sao Tome and Principe
Solomon Islands
Nicaragua
Morocco
GuatemalaEl Salvador
Congo, Rep.
Egypt, Arab Rep.
Gabon
Indonesia
Togo
Cote d'Ivoire
Ghana
Bolivia
Zambia
Cameroon
Sudan
Madagascar
Congo,
Dem.
Rep.
Central
African
Republic
Senegal
Nigeria
Malawi
Mauritania
Liberia
Cambodia
Mali
Niger
0
1
2
3
4
Logarithm of Natural Resources exports in percent of GDP, 1970
Figure: Life expectancy and Natural Resource Abundance
Data source: World Development Indicators, 2006
14
FOUR EXPLANATIONS OF
RESOURCE CURSE




I. Old Dutch disease stories
II. Volatility
III. Bad policies
IV. Rent seeking, corruption and conflict
15
I. OLD EXPLANATIONS OF
RESOURCE CURSE





Windfall gain in demand for resources from abroad
induces an appreciation of the real exchange rate.
The non-resource export sectors go in decline.
The sheltered sector gets a boost as labour and
other factors move from traded to sheltered sectors.
Easy to extend to Heckser-Ohlin and factor use in
resource sector (Corden and Neary, EJ, 1982;
Corden, OEP, 1984)
Or to nominal wage rigidity in Dornbusch-style
models of the open economy (Eastwood and
Venables, EJ, 1982; Buiter and Purvis, 1983)
16
Is there a Dutch Disease?



‘It seems ungrateful to talk of a disease’ (The
Economist). Dutch Disease?
Decline of exposed sectors may just be the
efficient response to the resource boom.
However, if there is learning by doing in the
non-resource export sectors, there may well
be a loss in output and welfare (van
Wijnbergen, EJ, 1984; Krugman, JDE, 1987).
A lower growth rate may well result (Sachs
and Warner, 1997).
17
Worsening of competitiveness




P G(LN) = H F(1  LN) with H  HT /HN is LM
locus, which slopes upwards in P-LN space
Higher natural resource exports Q E boosts P
and induces more than proportionate income
in national income Y
Boost to output and consumption of NTsector
Consumption of T-goods rises despite
contraction of T-sector (supplied through
imports paid for by resource revenues)
18
Dynamic effects of a resource boom:
AAAB



On impact resource boom leads to real
appreciation (higher P), decline of exposed
sector and boom of sheltered sector
As relative productivity of labour in T-sector
gradually falls, the real exchange rate
depreciates (falling P) so labour shifts back
from sheltered to exposed sectors
In the long run there must be real
depreciation
19
Natural resource abundance reduces competitiveness
Relative price of
non-traded goods
LM
LM
A
A
A
NTGME
NTGME
B
NTGME
Fraction of labour in non-traded sector
Higher resource exports shifts A to A, so induces appreciation of real exchange rate.
With passing of time relative productivity of traded relative to that of non-traded sector declines if e. of s.
between traded and non-traded goods is less than unity. This shifts equilibrium from A to A and eventually all
the way to B. In the long run there is a real depreciation and allocation of labour returned to original level.
20
Extraction of natural resources requires
labour and capital
Resource movement as well as spending
effects of resource boom. Labour is drawn
both out NT and T to resource sectors.
 Within context of Heckscher-Ohlin the
Rybczynski theorem implies output of Kintensive non-resource sector expands.
 If T-sector is K-intensive, resource boom
induces pro-industrialisation if spending effect
is not too large.
21
II. VOLATILITY & RESOURCE CURSE




‘What commodity price lack in trend, they make up
for in variance’ (Deaton, JEcPersp, 1999).
Resource rich economies are extremely vulnerable
to the high volatility of resource prices, especially as
supply is fairly inelastic.
Particularly bad as many resource rich economies
are not much diversified: specialised in resources
and small sheltered sector. In fact, they specialise
away from non-resource traded goods which causes
even more volatility and interest rate rises! Traded
sector shrinks until it vanishes (Hausmann &
Rigobon, 2002).
Volatility also bad for growth, investment, income
distribution, poverty and educational attainment
(Ramey & Ramey, Aizenman & Marion, Flug et al)
22
23
.2
.4
.6
.8
1
Declining natural resource dependence
in the global economy
1970
1980
1990
Sub-Saharan Africa
Middle East & North Africa
Western Europe & North America
2000
2010
Latin America & Carib.
East Asia & Pacific
South Asia
Source: World Bank Development Indicators, 2005, World Bank
24
III. RESOURCES ENCOURAGES
UNSUSTAINABLE POLICIES





Erosion of critical faculties of politicians.
Netherlands in the seventies dressed up the welfare
state and governments since 1989 have been trying
to have a sustainable welfare state.
Induces excessive borrowing (Manzano and
Rigobon, 2002) & invest in ‘prestige’ projects.
Loose sight of growth-promoting policies and valuefor-money management.
State-led industrialisation through import substitution
and heavy subsidies for manufacturing.
25
IV. RESOURCES RENT SEEKING,
CORRUPTION & CONFLICT





Allocation of talent: countries with many rent seekers and
lawyers grow more slowly than countries with lots of engineers
(Murphy et al, JPE, 1989; AER, 1993)
Voracity effect: drag on economic growth (Tornell and Lane,
1999). Applies theory of common pool.
Corruption, political instability, bureaucratic inefficiency,
assassinations and conflict also hamper economic growth
(Mauro, 1995; Leite and Weidmann, 1999). Bad effects of
resource growth mainly operates via worsening of institutions,
rule of law, etc.
Increases civil strife and wars, especially in sub-Saharan Africa
thru’ weakening of state or finance of rebels (Collier and Hoeffler,
2004; Ross, 2004). War lord competition (Skaperdas, 2004).
Distinguish between grievance and greed (Ollson and Fors,
2004).
Especially bad for point-based rather than diffuse resources.
26
RENT GRABBING VERSUS PRODUCER
FRIENDLY INSTITUTIONS
(Mehlum, Moene and Torvik, EJ, 2005)
Key: A resource bonanza shifts equilibrium from A to A if there are
strong institutions, which means higher profits and more entrepreneurs.
In case of weak institutions the equilibrium shifts from A to A, so profits
decline and number of rent seekers increases.
27
CROSS-COUNTRY EVIDENCE
FOR RESOURCE CURSE




Sachs and Warner (1997)
Mehlum, Moene and Torvik (2005)
Boschini et al (2003)
Arezki and van der Ploeg (2007)
28
EFFECTS OF RESOURCE ABUNDANCE AND
INSTITUTIONAL QUALITY ON ECONOMIC GROWTH
Annual growth in
real GDP per
capita
Sachs and
Warner (1997a)
Based on data
in
Sachs and
Warner
(1997b)
Initial income
-1.76 (8.56)
-1.28 (6.65)
-1.26 (6.70)
Openness
1.33 (3.35)
1.45 (3.36)
1.66 (3.87)
Resource
abundance
-10.57 (7.01)
Rule of law
0.36 (3.54)
-
-
Institutional quality
-
0.6 (0.64)
-1.3 (1.13)
Investments
1.02 (3.45)
0.15 (6.73)
0.16 (7.15)
Interaction term
-
-
15.40 (2.40)
Number of
countries
71
Adjusted R2
0.72
-6.69 (5.43)
87
0.69
Mehlum, Moene
and
Torvik (2005a)
-14.34 (4.21)
87
0.71
29
CROSS-COUNTRY EVIDENCE
GDP growth
(1)
(2)
(3)
(4)
Initial log GDP
0.048
-0.051
-0.560
-0.407
(1.09)
(0.44)
(4.75)**
(3.03)**
2.410
1.985
1.511
1.403
(5.56)**
(3.35)**
(3.25)**
(3.09)**
0.009
0.009
-0.023
-0.092
(0.69)
(0.66)
(2.04)*
(2.78)**
0.098
0.150
0.031
(0.99)
(1.95)
(0.34)
0.188
0.192
(6.67)**
(7.01)**
Openness
Natural Resources
over GDP
Institution
GFCF over GDP
Interaction
0.007
(2.21)*
Observations
74
69
69
69
R-squared
0.64
0.65
0.80
0.81
Only countries with
poor institutions
suffer from resource
curse
Implies  = 0.32,
 = 0.0046 and
half-time 15 years
Source:
World Bank Development
Indicators 2006
International Country Risk
Guide, PRS Group plc.
Sachs and Warner (1997)
30
Marginal effects of different types of natural
resources on growth for
different levels of institutional quality
Primary
exports share
of GDP
Ores and
Mineral
Prod of gold,
metals
production as
silver and
exports as
share of GNP diamonds as
share of GDP
share of GDP
Worst
institutions
0.548
0.946
1.127
1.145
Average
institutions
0.378
0.425
0.304
0.279
Average +
one s.d.
institutions
0.288
1.152
1.062
1.183
Best
institutions
0.228
1.629
1.560
1.776
Source: Boschini, et. al. (2003)
31
VOLATILITY, FINANCIAL
DEVELOPMENT AND THE
NATURAL RESOURCE CURSE

Based on work with Steven Poelhekke
32
Motivation

Output volatility seems to matter for growth.
(e.g., Ramey & Ramey, AER, 1995)

What explains volatility? Or its absence?

Why do so many resource rich countries stay poor?

Is the resource curse cast in stone?

Can volatility explain the curse?
33
The Facts
10
Volatile countries have lower growth (Figure 1)
5
Equatorial Guinea
0
United Arab Emirates
Iraq
-5
1.
Liberia
0
10
20
Standard Deviation of Yearly GDP/Capita Growth
(1970-2003, %)
30
Fitted Values (slope = -.247 (.049); Adj. R2=.14)
34
The Facts: s.d. real GDP growth (Table 1)
2.
Developing countries are more volatile
Sub-Saharan Africa: 6.52 Western Europe: 2.33
Middle East/North Africa: 8.12! North America: 1.90
2.
Countries with poorly developed financial systems
are more volatile
1th Q (<16.2%): 6.40
4th Q (>52.9%): 4.40
3.
Remote (distance from waterway) countries are
more volatile
1th Q (<49km): 6.52
4th Q (>359km): 8.12
35
The Facts (Figure 2)
30
Resource dependent countries are more volatile
10
20
Liberia
United Arab Emirates
Bahamas, The
0
5.
0
20
40
60
Average Resource Share of GDP
(1970-2003, %)
80
100
Fitted Values (slope =.149 (.016); Adj. R2=.44)
36
The literature




Ramey & Ramey, AER, 1995: volatility affects growth
Koren, Tenreyro, QJE, 2007: volatility falls with
development as countries diversify away from
highly volatile sectors and improve macro-policy.
Blattman, Hwang, Williamson, JDE, 2007: resource
exporters have volatile terms of trade, less FDI,
less growth (1870-1939)
Sachs & Warner, 1997: windfall resource revenue ->
RER appreciation -> decline in non-resource
exports -> less learning by doing and lower TFP
growth
37
Key questions
1.
Does volatility affect growth negatively?
2.
Does resource dependence explain output volatility?
3.
A new explanation for the resource curse?
4.
Do financial development and openness mitigate any
adverse effects?

ML to simultaneously explain volatility of
unanticipated output shocks and its effect on growth.
38
How can volatility hamper growth?

Aghion at. al., CEPR 2006: RER volatility with credit
constraints stunts innovation

nominal exchange rate volatility: liquidity shock
Here:
shock = exogenous World commodity price
(Cashin et. al., IMF, 2002)
cash = profits + resource income

- nominal wage stickiness
39
Results:


If F(z) is concave so that E[F(·)]  F(E[·]),
then more volatility in natural resource revenues or
in nominal exchange rate lowers innovation and
growth, unless financial development is very large.
High and stable resource revenues and also a
stable nominal exchange rate ease constraints and
boost growth.
40

If F(z) is concave so that E[F(·)]  F(E[·]), more
volatility lowers innovation and growth unless
financial development is very large. Figure 3:
41
Other links between output volatility and
growth





Higher volatility means more uncertainty-induced errors and thus
less irreversible investment and lower growth (Bernanke, QJE,
1983; Pindyck, 1991; Aizenman and Marion, RIE, 1991)
Especially if it is costly to switch factors of production
(Bertola, JME, 1994; Dixit and Rob, JET, 1994)
Or: higher volatility leads to more precautionary saving and thus
more investment and growth (Mirman, Etrica, 1971)
Higher variance commands investments with higher return and
thus higher growth (Black, 1987)
Net effect of volatility on growth can in theory be negative or
positive, so needs to be settled empirically
42
Data






Heston, Summers, Aten, Penn World Tables 6.2
Human capital: Barro & Lee (average schooling
years in population age 25+)
Resources: World Development Indicators (export
revenue as % GDP)
Openness: Sachs & Warner dummy (Wacziarg,
Welch, 2003)
Financial development: WDI (domestic credit to
private sector)
Distance to coast/river: Center for International
Development (km)
43
Evidence: volatility and growth (Table 2)
Dependent Variable
Mean equation
Average investment share of GDP 1970-2003
Investment share of GDP 1970
Average population growth rate 1970-2003
log per capita GDP 1970
Human capital 1970
Volatility (σi)
Point-source resources 1970
Point-source rent share 1970
Diffuse resources 1970
Financial development 1970
Sachs Warner updated openness dummy 70
Point based resources * openness 70
Point-source rent share * openness 70
Point-source resources * Fin. Dev. 70
Point-source rent share * Fin. Dev. 70
Constant
st
1 Lag Error (ε)
2nd Lag Error (ε)
(1)
0.108***
-0.472***
-0.012***
0.001*
-0.110**
0.110***
yearly GDP growth per capita 1970-2003
(constant 2000 international dollars, PWT 6.2)
(4)
(5: Rents)
(0.012)
0.005
(0.013)
0.023
(0.118)
-0.654*** (0.204)
-0.730***
(0.001)
-0.016*** (0.003)
-0.013***
(0.000)
0.001**
(0.001)
-0.000
(0.049)
-0.324**
(0.135)
-0.410***
-0.054
(0.038)
-0.198**
0.012
(0.023)
0.052***
0.001
(0.005)
-0.004
0.007
(0.004)
0.006**
0.040
(0.066)
0.212***
0.327*
(0.181)
0.714**
(0.011)
0.167***
(0.020)
0.143***
0.252***
(0.018)
0.229***
-0.006
(0.020)
0.003
(0.014)
(0.220)
(0.002)
(0.001)
(0.138)
(0.087)
(0.017)
(0.006)
(0.003)
(0.065)
(0.335)
(0.018)
(0.018)
(0.020)
44
Evidence: underlying determinants of volatility
Dependent Variable
Mean equation
Investment share of GDP 1970
Average population growth rate 1970-2003
log per capita GDP 1970
Human capital 1970
Volatility (σi)
Point-source resources 1970
Point-source rent share 1970
Financial development 1970
Sachs Warner updated openness dummy 70
Point-source rent share * Fin. Dev. 70
Point-source rent share * openness 70
Constant
st
1 Lag Error (ε)
yearly GDP growth per capita 1970-2003
(6a)
(7a: Rents)
-0.005
(0.015)
0.022
(0.015)
-0.740***
(0.155)
-0.897***
(0.158)
-0.021***
(0.002)
-0.019***
(0.003)
0.003***
(0.001)
0.002*
(0.001)
-1.572***
(0.372)
-1.627***
(0.431)
0.088***
(0.024)
-0.089
(0.109)
-0.025***
(0.008)
-0.035***
(0.009)
-0.009
(0.007)
-0.013*
(0.008)
(0.308)
0.788**
(0.083)
0.233***
0.265***
(0.031)
0.259***
(0.033)
0.227***
(0.017)
0.222***
(0.018)
Variance equation
Point based resources 1970
Point based rent share 1970
Diffuse resources 1970
Financial development 1970
Sachs Warner updated openness dummy 70
Distance to nearest navigable river or coast
Constant
Observations
Log likelihood
1.551***
0.862**
-1.333***
-0.693***
0.001***
-6.073***
2084
3726.2
(0.205)
(0.359)
(0.089)
(0.048)
(0.000)
(0.064)
2.452***
0.215
-1.424***
-0.720***
0.001***
-5.953***
1980
3571.2
(0.624)
(0.354)
(0.092)
(0.048)
(0.000)
(0.065)
45
Evidence: volatility and growth (Table 2..)
Dependent Variable
Mean equation
Average investment share of GDP 1970-2003
Investment share of GDP 1970
Average population growth rate 1970-2003
log per capita GDP 1970
Human capital 1970
Volatility (σi)
(1)
0.108***
-0.472***
-0.012***
0.001*
-0.110**
yearly GDP growth per capita 1970-2003
(constant 2000 international dollars, PWT 6.2)
(4)
(5: Rents)
(0.012)
0.005
(0.013)
0.023
(0.118)
-0.654*** (0.204)
-0.730***
(0.001)
-0.016*** (0.003)
-0.013***
(0.000)
0.001**
(0.001)
-0.000
(0.049)
-0.324**
(0.135)
-0.410***
(other controls included)
(0.014)
(0.220)
(0.002)
(0.001)
(0.138)
Variance equation
Sub-Saharan Africa
Middle-East & North Africa
Latin America & Caribbean
Eastern Europe & Centra Asia
East Asia & Pacific
South Asia
Western Europe
North America
Constant
Country dummies in variance eq.
Observations
Log likelihood
(0.154)
2.653***
(0.160)
(0.160)
1.686***
(0.166)
(0.153)
1.571***
(0.158)
(0.274)
1.357***
(0.272)
(0.159)
0.868***
(0.162)
(0.194)
0.385*
(0.200)
(0.154)
0.274*
(0.158)
Reference region (least volatile)
-7.804*** (0.149)
-7.753*** (0.153)
no
no
2186
2014
4012.4
3748.2
2.588***
1.734***
1.604***
1.433***
1.027***
0.459**
0.211
-3.823***
yes
3448
5898.5
(0.118)
46
IV: Endogenous Investment



Unobserved country characteristics may determine
the investment share and growth simultaneously
(notably institutions)
Growing countries may attract more investment
IVs - ethnolinguistic fractionalization & polarization
(Montalvo, Reynal-Querol, JDE & AER 2005)
less trust, more corruption, less political rights
(Alesina et al, JEG, 2003)
- % pop. in temperate climate
(Gallup et al, 1999; CID)
47
IV: Endogenous Investment (cont.)
Dependent Variable
Mean equation
Investment share of GDP 1970
Average population growth rate 1970-2003
log per capita GDP 1970
Human capital 1970
Volatility (σi)
Point-source resources 1970
Financial development 1970
Sachs Warner updated openness dummy 70
Constant
1st Lag Error (ε)
Ethnic Polarization
% Population in Temperate Climate Zone
Ethnic Fractionalization Index
yearly GDP growth per
capita 1970-2003
(6a)
-0.005
(0.015)
-0.740***
(0.155)
-0.021***
(0.002)
0.003***
(0.001)
-1.572***
(0.372)
(0.024)
0.088***
-0.025***
(0.008)
-0.009
(0.007)
0.265***
(0.031)
0.227***
(0.017)
Initial investment share
GDP 1970
(6b: 1st stage)†
0.697
-0.097***
0.025**
(3.436)
(0.036)
(0.010)
0.899**
0.174**
0.029
0.822***
(0.406)
(0.083)
(0.036)
(0.221)
0.008
0.066
-0.133***
(0.057)
(0.062)
(0.044)
yearly GDP growth per
capita 1970-2003
(6c: 2nd stage)
0.065**
(0.029)
-0.581***
(0.148)
-0.015***
(0.003)
0.002*
(0.001)
-1.318***
(0.357)
(0.034)
0.029
-0.030***
(0.008)
-0.009
(0.006)
0.202***
(0.039)
0.223***
(0.017)
Variance equation
Point based resources 1970
1.551***
Diffuse resources 1970
0.862**
Financial development 1970
-1.333***
Sachs Warner updated openness dummy 70
-0.693***
Distance to nearest navigable river or coast
0.001***
Constant
-6.073***
F-stat. on excl. instruments
Hansen overidentification J-statistic
(p-value)
Observations
2084
R2
.
Log likelihood
3726.2
† Robust and clustered standard errors by country.
(0.205)
(0.359)
(0.089)
(0.048)
(0.000)
(0.064)
1.614***
0.900**
-1.316***
-0.679***
0.001***
-6.102***
(0.206)
(0.366)
(0.097)
(0.048)
(0.000)
(0.066)
3.63
0.127
2084
0.53
2084
.
3728.7
48
Direct Resource Effect: Curse or blessing?




Resource rents: -0.316**
Resource rents × financial development: +1.106***
Resource rents × openness: +0.243***
Continuous interactions: size and significance of
resource effect on growth depends on initial levels
of financial development and openness.
49
-.5
0
.5
1
1.5
Blessing for some
0
.2
.4
.6
.8
1
Finiancial Development in 1970
Marginal Effect of Rent Share on Growth
for Average 1970 Openness
95% Confidence Interval
50
Direct and Indirect Effect
-.06
-.04
-.02
0
.02
Direct resource effect on growth + indirectly through
volatility (abstracting from interactions)
Volatility of unanticipated output growth, estimated 1970-2003
Zambia
.09
.08
LL. Africa
Malawi
RR. Africa
.05
.04
Asian T.
OECD
.02
.01
-.08

51
Robustness: Revenue volatility & Policy
Dependent Variable
(base line mean equation with point-source
resources, fin. dev., openness)
(9b)
yearly GDP growth per capita 1970-2003
(9c)
(9d)
Variance equation
Initial point-source resources 1970
Initial diffuse resources 1970
Initial financial development 1970
Sachs Warner updated openness dummy 1970
Distance to nearest navigable river or coast
Point-source export share volatility 70-03
Diffuse export share volatility 70-03
Government share volatility 70-03
Agricultural R.M. resource share volatility 70-03
Foods resource share volatility 70-03
Ores & metals resource share volatility 70-03
Fuels resource share volatility 70-03
Financial development * point based volatility
Constant
Observations
Log likelihood
-0.803***
-0.314
-0.905***
-0.532***
0.001***
9.361***
4.703*
10.632***
-6.734***
2084
3806.0
(0.283)
(0.555)
(0.107)
(0.057)
(0.000)
(0.464)
(2.429)
(1.099)
(0.082)
-0.586*
-1.341**
-0.888***
-0.475***
0.001***
(0.324)
(0.583)
(0.108)
(0.066)
(0.000)
10.365***
0.699
12.691***
6.517***
9.369***
(1.225)
(2.117)
(3.453)
(2.255)
(0.478)
-6.815***
2084
3807.5
(0.086)
-0.620**
0.007
-0.800***
-0.540***
0.000***
15.410***
2.786
9.814***
(0.299)
(0.555)
(0.114)
(0.057)
(0.000)
(1.635)
(2.430)
(1.082)
-33.275***
-6.698***
2084
3810.4
(8.893)
(0.081)
52
Robustness: Economic restrictions

IMF’s Annual Report on Exchange Arrangement and Restrictions
yearly GDP growth
per capita 1970-2003
(11a)
Mean equation
Variance equation
Investment share of GDP 1970
0.016
(0.019)
Initial point-source resources 70
Average population growth rate ‘70-‘03 -0.770*** (0.202)
Initial diffuse resources 70
log per capita GDP 1970
-0.017*** (0.003)
Initial financial development 1970
Human capital 1970
0.003*** (0.001)
Distance to nearest navigable river or
coast
Volatility (σi)
-0.557*** (0.188)
Ethnic Polarization
Initial point-source resources 70
0.085*** (0.032)
Multiple Exchange Practices (yes=1)
Initial diffuse resources 70
-0.009
(0.030) Current Account Restrictions (yes=1)
Financial development 1970
-0.008
(0.006)
Capital Account restrictions (yes=1)
(0.003) Surrender of Export receipts (yes=1)
Current Account Restrictions (yes=1)
0.003
(0.004)
-0.003
Cur. Acc. Restrictions * Point
Capital Account restrictions (yes=1)
Resources 70
Cap. Acc. Restrictions * Point
Resources 70
Constant
0.178*** (0.028)
Constant
st
1 Lag Error (ε)
0.235*** (0.019)
nd
2 Lag Error (ε)
-0.001
(0.019)
Observations
2015
Log likelihood
3595.0
5.680***
1.949***
-1.726***
0.001***
(0.343)
(0.498)
(0.140)
(0.000)
-0.717***
0.446***
-0.442***
0.345***
4.485***
(0.061)
(0.071)
(0.126)
(0.112)
(0.945)
-1.988***
(0.468)
-6.573***
(0.112)
53
Robustness: Ethnic tensions

Polarization -> civil war, less investment, likely to
fight over resources
yearly GDP growth
per capita 1970-2003
(11a)
Mean equation
Variance equation
Investment share of GDP 1970
-0.005
(0.015)
Initial point-source resources 70
Average population growth rate 1970-2003 -0.569*** (0.215)
Initial diffuse resources 70
log per capita GDP 1970
-0.019*** (0.002)
Initial financial development 70
Human capital 1970
0.003*** (0.001)
Sachs Warner updated
openness dummy 70
Distance to nearest navigable
Volatility (σi)
-0.771*** (0.250)
river or coast
Initial point-source resources 70
-0.043
(0.053)
Financial development 1970
-0.041
(0.029)
Sachs Warner updated openness dummy 70 -0.011* (0.006)
Initial point-source resources * openness 70
-0.001
(0.005)
Initial point-source resources * Fin. Dev. 70
0.153
(0.096)
(0.231)
Ethnic Polarization
0.356
Ethnic Polarization
Point-source resources 70
* Ethnic Pololarization
Constant
0.211*** (0.025)
Constant
1st Lag Error (ε)
0.230*** (0.017)
Observations
2084
Log likelihood
3795.4
-1.143***
0.967**
-1.127***
-0.507***
(0.329)
(0.401)
(0.106)
(0.058)
0.001***
(0.000)
0.178**
13.914***
(0.073)
(0.779)
-6.475***
(0.081)
54
Robustness: Panel estimates
Dependent Variable
Mean equation
Investment share of GDP
Population growth rate
log per capita GDP
Human capital
Volatility (σit)
Point-source resource share
Diffuse resource share
Financial development
Sachs Warner updated openness dummy
Point-source resources*Financial
development.
Constant
1st Lag Error (ε)
2nd Lag Error (ε)
yearly GDP growth per capita 1970-2003
yearly GDP growth per capita 1990-2003
(13a)
0.034***
-0.808***
-0.019***
0.003***
-1.440***
0.157**
0.062**
-0.011**
-0.009
-0.095
(0.013)
(0.121)
(0.002)
(0.001)
(0.482)
(0.070)
(0.032)
(0.006)
(0.007)
(0.091)
(13b) Within
0.000
(0.018)
-0.217
(0.192)
-0.036*** (0.004)
0.000
(0.002)
0.130
(0.388)
-0.019
(0.024)
-0.031
(0.032)
0.001
(0.009)
0.008
(0.005)
0.014
(0.045)
(14a)
0.010
-0.584***
-0.011***
0.002*
-1.903**
0.224*
0.091
-0.019**
-0.022**
-0.157
(0.017)
(0.205)
(0.004)
(0.001)
(0.839)
(0.115)
(0.064)
(0.008)
(0.011)
(0.134)
(14b) Within
-0.119*** (0.030)
-0.218
(0.291)
-0.103*** (0.011)
-0.003
(0.004)
0.649
(0.500)
0.188*
(0.112)
-0.037
(0.059)
0.010
(0.009)
0.008
(0.006)
-0.206
(0.177)
0.245***
0.266***
(0.030)
(0.016)
-0.004
0.170***
(0.021)
(0.016)
0.183***
0.260***
0.027
(0.044)
(0.027)
(0.031)
-0.048**
-0.203***
0.001
(0.022)
(0.031)
(0.024)
4.692***
2.214***
-0.855***
-0.598***
0.000***
(0.315)
(0.383)
(0.070)
(0.055)
(0.000)
-0.097
-1.632**
-1.495***
-0.237***
-
(0.176)
(0.679)
(0.043)
(0.056)
5.677***
3.396***
-0.812***
-0.560***
-0.000
(0.513)
(0.672)
(0.086)
(0.091)
(0.000)
-10.771***
-3.563**
-1.302***
0.126
-
(1.748)
(1.683)
(0.059)
(0.177)
-4.138***
(0.953)
0.066
(0.823)
-5.926***
(1.498)
21.274***
(2.902)
-6.207***
2346
4342.4
(0.066)
-5.748***
2476
4492.1
(0.025)
-6.288***
1005
1990.5
(0.105)
-6.080***
1075
2148.6
(0.033)
Variance equation
Point-source resource share
Diffuse resource share
Financial development
Sachs Warner updated openness dummy
Distance to nearest navigable river or
coast
Point-source resources*Financial
development.
Constant
Observations
Log likelihood
55
The bottom line for some resource rich
countries: counterfactual exercise
Asian Tigers
Resource-Rich Africa versus the Asian Tigers
GDP per capita growth
Resourcerich Africa
on yearly
GDP/capita
growth rate
4.04%
0.25%
0.065 **
-0.581 ***
-0.015 ***
0.002 *
-1.318 ***
0.029
-0.030 ***
19.48%
1.86%
7.747
4.049
3.45%
4.32%
26.89%
30.42%
2.75%
7.129
1.476
6.04%
13.13%
14.43%
-0.71%
0.52%
-0.93%
0.51%
3.41%
-0.26%
-0.37%
1.614 ***
0.900 **
-1.316 ***
-0.679 ***
0.001 ***
4.32%
11.08%
26.89%
0.746
90.902
13.13%
10.52%
14.43%
0.000
552.571
0.57%
-0.02%
0.65%
1.85%
1.70%
Mean equation
Investment share of GDP 1970
Average population growth rate 1970-2003
Initial log per capita GDP 1970
Initial human capital 1970
Volatility (σi)
Initial point-source resources 1970
Initial financial development 1970
Variance equation
Initial point-source resources 1970
Initial diffuse resources 1970
Initial financial development 1970
Sachs Warner updated openness dummy 70
Distance to nearest navigable river or coast
Countries
4
6
Note: Resource-rich African counties are: Algeria, Congo, Rep., Ghana, Malawi, Togo, Zambia.
Asian Tigers are: South Korea, Malaysia, Philippines and Thailand.
56
Conclusions on volatility and finance




Volatility of unanticipated output growth is
quintessential feature of the natural
resource curse!
Positive direct effect of the level of natural
resource exports on growth is swamped
by negative indirect effect of volatility on
growth performance.
Countries with high degrees of financial
development can turn resource wealth into
blessing and boon for growth.
Point-base impact stronger than diffuse
resources.
57
Conclusions on volatility and finance ..



High levels of investment rates, human capital
and openness boost growth performance.
Countries with low initial GDP per capita catch
up, but countries with high population growth
rates grow more slowly.
Volatility increases with distance to waterways,
volatility of government share, ethnic
polarization, current account restrictions &
surrender of export receipts.
Volatility decreases with openness, multiple
exchange practices, capital account restrictions
financial development.
58
GENUINE SAVING AND
EXHAUSTIBLE RSOURCE RENTS
Source: World Bank (2006, Figure 3.4)
59
GENUINE SAVING RATES
AND GDP GROWTH, 2003
Source: World Bank (2006, Figure 3.6).
60
RESOURCE ABUNDANCE AND CAPITAL ACCUMULATION
(HARTWICK RULE)
Source: World Bank (2006, Figure 4.1).
61
HARTWICK RULE IN SMALL OPEN
ECONOMY: THE KUWAIT MODEL



No resources as factor input into production.
Country should save less on its current
account than the marginal Hotelling rents if
world resource prices are expected to
increase and exploration technology is
expected to advance.
It is then better to postpone extraction and
borrow and profit from future resource
scarcity and innovation.
62
HARTWICK RULE IN GLOBAL
ECONOMY?






Free trade in oil and goods & PCM & ZLM
Capital and resource intensities fixed by world
interest rate & world price of natural resources.
With zero technical progress and population growth
& identical technologies, maxi-min egalitarianism
can be characterized (cf., Asheim):
Hartwick rule for global economy as a whole.
Oil exporters run a deficit: Hotelling rule implies that
they expect capital gains and growing incomes.
Oil importers run a surplus to accumulate national
wealth by consuming only a fraction of the MPK to
compensate for decreasing return on capital.
63
WHY DO RESOURCE RICH
COUNTRIES HAVE NEGATIVE
GENUINE SAVING RATES?


Anticipation of better times (higher oil prices
in the future, improvements in future
extraction technology, etc.)?
Or rapacious rent seeking induced by
voracity effect and dynamic common pool
problem (building on Tornell and Lane)?
64
POLITICAL HOTELLING AND
HARTWICK RULES




In homogenous society s(t) =  (Hartwick
rule) and genuine saving rate is zero.
In fractionalized society with N>1, society
saves more than resource rents but still has
negative genuine saving.
With more rival fractions, genuine saving rate
is more negative and sustainable level of
consumption is lower.
Resource price appreciation exceeds the
interest rate and thus depletion is too fast.
65
POLICY PROPOSALS




Temporary subsidy/tax relief for non-resource
exposed sectors if learning by doing (van
Wijnbergen, QJE). Danger: policy addiction.
Staple trap view suggests gradual dual-track reform
by creating a dynamic market sector in early-reform
enclaves with post-reform benefits may work with
sustained rents from natural resources. Rapid
expansion of enclaves can pull the more backward
sectors up as well. Big push: works if IRTS in NTsector (Murphy et al, Sachs & Warner, 1997, JDE).
Put resource revenues into a fund to spread the
benefits to future generations by investing in
education, infrastructure, etc. Fund also helps to
cope with volatile resource prices.
If rapicious rent seeking, better to keep oil in ground
rather than deplete it and put the revenues in fund.
66
POLICY
PROPOSALS
ctd.
 Use revenues to reduce debt or invest in education





& infrastructure with market return.
Privatisation of state-owned oil and mining industries
& tendering exploitation rights to private companies.
Not clear that this works.
Improve institutions, rule of law, etc. Easier said
than done in presence of vested interests.
Exit caused by corruption does not necessarily
reduce welfare, so more competition not necessarily
good either (Bliss & Di Tella, JPE, 1997).
Distribute revenues as citizen dividends.
Government must then make better case for its pet
projects, since it has to tax its people: endowment
and an information effect (Sandbu, 2004).
Get exploitation companies at the peace negotiation
table to secure re-employment of ex-combatants.
67
Transparency is a must




Highest standards of public and corporate
accountability, PSR/CSR and transparency: publish
what you earn from exports and publish what you do
with the revenues.
Exploitation companies should publish their
payments to all governments and encourage
mandatory disclosure mechanism.
Make debt relief etc. contingent on transparency,
free press and anti-corruption efforts – role IMF,
World Bank and UNDP. Establish global information
office.
Western banks should be punished for allowing
tainted money to be deposited.
68