## _wp10118

## Source details

**Canonical URL:** [_wp10118](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp10118.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp10118.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp10118.pdf.json)

---

### Background and scope
- Focus: intersectoral linkages between oil and non-oil sectors in oil exporters using a Panel Vector Auto-Regression (PVAR) framework, accounting for the real exchange rate (REER) channel.
- Sample: 23 oil exporting countries (Algeria, Angola, Azerbaijan, Bahrain, Cameroon, Chad, Congo, Equatorial Guinea, Gabon, Indonesia, Iran, Kazakhstan, Kuwait, Libya, Nigeria, Oman, Qatar, Saudi Arabia, Syria, Turkmenistan, United Arab Emirates, Venezuela and Yemen).

### Key stylized facts
- Oil intensity: heterogeneous, ranging from an average of 6 percent (Chad) to nearly 63 percent (Libya).
- Oil intensity definition: average share of oil GDP in total GDP (constant prices) over 1985–2008.
- Oil-intensity groups:
  - Low oil-intensity: below sample median observation (11 countries).
  - High oil-intensity: above sample median observation (12 countries).
- Correlation patterns:
  - In low oil-intensity countries, oil and non-oil sectors growth correlation = 0.67 (Figure 3a).
  - In high oil-intensity countries, oil and non-oil sectors growth correlation = 0.21 (Figure 3b).
- Cross-country empirical relationship: negative slope between oil intensity and the correlation of oil and non-oil sector growth (Figure 2).

### Channels and mechanisms discussed
- Reviewed channels behind the “natural resource curse”:
  - Dutch Disease: REER appreciation and resource pull effects that crowd out the non-natural resource tradable sector.
  - Concentration and volatility: concentrated economies exposed to commodity price volatility, reducing investment and growth.
  - Policy and institutional effects: overconfidence in resource revenues weakening fiscal discipline; state ownership via parastatals that soften budget constraints and may manage resources inefficiently.
  - Political-economy channels: Rentier Effect, higher corruption, Delayed Modernization and Entrenched Inequality effects, and greater propensity for violent conflict.
- REER channel evidence in high oil-intensity countries:
  - REER cycles and oil sector growth cycles move in opposite directions: REER tends to appreciate when oil sector growth accelerates and depreciate when oil sector growth decelerates (Figure 4a).
  - REER cycles and non-oil sector growth cycles tend to move together in high oil-intensity countries (Figure 4b).

### Main empirical findings (PVAR results)
- High oil-intensity countries:
  - The estimation results largely support the “natural resource curse.”
  - In high oil-intensity countries, a one percent growth of the oil sector leads, on average, to an accumulated decline of 0.15 percent in the non-oil sector within two years.
- Low oil-intensity countries:
  - Oil sector has positive externalities on the non-oil sector despite adverse REER appreciation.
  - There is a positive and significant impact of the non-oil sector on the oil sector.
- Conclusion: adverse natural-resource effects operate mainly in high oil-intensity settings; positive externalities dominate in low oil-intensity settings.

### Fiscal channel and macroeconomic transmission
- Prudential use of oil revenue windfalls (investment in infrastructure and human capital consistent with absorption capacity) can yield positive externalities on the non-oil sector.
- Inefficient use of oil revenues can lead to:
  - increased inflationary pressures if output is close to potential,
  - sharp real exchange rate appreciation,
  - a decline in the marginal efficiency of capital,
  - a non-sustainable fiscal stance,
  - greater vulnerability to external shocks,
  - impaired external competitiveness,
  - discouraged foreign investment.

### Special Fiscal Institutions (SFIs), governance, and public spending quality
- SFIs established to cope with oil price volatility: oil funds, fiscal rules, fiscal responsibility legislation, budgetary oil prices.
- SFIs’ effectiveness varies; not a panacea.
- World Bank “government effectiveness” indicator (based on 2007 data; units range from about -2.5 to 2.5):
  - Controlling for income level, government effectiveness tends to be lower in countries with high oil-intensity.
  - Implies SFIs’ effectiveness and quality may be higher in low oil-intensity countries.
- Possible explanations for weaker fiscal institutions in high oil-intensity countries:
  - Government over-confidence that oil revenue windfalls will be long-lasting.
  - Where oil revenues account for the lion’s share of total revenues and income taxes are generally low, the public exerts less pressure to strengthen fiscal institutions.

### Budget transparency and investment climate
- Open Budget Index (OBI): negatively correlated with oil intensity in the sample of 12 countries with OBI data.
  - Lower transparency and poorer quality of public spending in high oil-intensity countries may adversely affect non-oil sector development.
- Doing Business ranking:
  - Positive link between oil intensity and worse Doing Business rankings in both higher and lower PPP per capita groups, suggesting weaker incentives to improve the business climate in high oil-intensity countries.

### Data, sample, and summary statistics
- Annual data for 1985–2008 for 23 developing oil-exporting countries.
- Averages (1985–2008):
  - non-oil sector growth: 5.6 percent,
  - oil sector growth: 5.8 percent,
  - REER depreciated by 2.2 percent on average.
- High standard deviation of oil sector growth indicates large variation across countries and time.
- Variables pass unit-root tests (alternative tests unanimously reject unit-root null).

### Oil-intensity groups and sample partition
- Sample split by median oil-intensity of 33 percent:
  - Low sub-sample: oil-intensity below 33 percent.
  - High sub-sample: oil-intensity above 33 percent.
- Composition changes over time as oil-intensity evolves:
  - Among 23 countries: nine constantly below 33 percent, four constantly above 33 percent, others shifted over time.

### Methodology: Panel VAR (PVAR)
- PVAR features:
  - Combines VAR (all variables endogenous) with panel data to capture unobserved heterogeneity.
  - Endogenous variable vector Yit = {oy, noy, rer} where:
    - oy = oil sector growth rate (real terms),
    - noy = non-oil sector growth rate (real terms),
    - rer = change in real effective exchange rate.
  - Fixed effects fi capture country specifics.
  - Forward mean-differencing (Helmert procedure) used to remove fixed effects correlation with regressors.
  - Estimation via GMM with lagged regressors as instruments.
  - Ordering for Cholesky orthogonalization: oy (most exogenous), noy, rer.
- Lag length selection:
  - Low group: one-lag specification slightly superior by AIC.
  - High group: three-lag specification has significantly lower AIC.

### Estimation results — Low oil-intensity countries (selected exact figures)
- Oil → non-oil:
  - A one percent growth in the oil sector leads, on average, to a 0.1 percent growth in the non-oil sector within one year.
- Non-oil → oil:
  - A one percent growth in the non-oil sector leads, on average, to a 0.26 percent growth in the oil sector in the subsequent year.
- Impulse Response Functions (IRFs):
  - Shocks in both sectors lead to REER appreciation.
  - No indications that REER movements have significant effects on oil and non-oil sector growth.
- Variance decomposition:
  - Non-oil sector variance attributes a greater share of oil sector forecast error variance than REER does.
  - REER has a higher share of attributed non-oil sector growth forecast compared to oil sector.

### Estimation results — High oil-intensity countries (selected exact figures)
- Oil → non-oil:
  - Estimated effect: a one percent growth in the oil sector leads, on average, to a decline of 0.15 percentage points in the non-oil sector’s growth within two years.
- REER effects:
  - Non-oil sector is adversely affected by REER appreciation within two to three years, but oil-sector shocks do not appear to trigger that appreciation.
  - REER appreciation also adversely affects the oil sector within three years.
- Non-oil → oil:
  - No significant impact of the non-oil sector on the oil sector in the High group.
- Interpretation:
  - Low share of non-oil sector forecast errors explained by REER variation suggests adverse oil-sector effects operate through channels not explicitly captured in the model (e.g., poor governance, weak fiscal institutions).

### Case study: Angola — overview and sector dynamics
- Civil war impact:
  - Non-oil sector share declined from 76 percent in 1991 to 32 percent in 1996.
- Post-1997 dynamics:
  - Non-oil sector average growth: 11.6 percent annually (1997–2008).
  - Oil sector expansion: 7.8 percent (1997–2008).
  - Non-oil share stable at around 40 percent in the last eight years.
- Oil exports account for 95 percent of total exports.
- Correlations:
  - 1992–1996: correlation between oil and non-oil growth = 0.26.
  - 1997–2008: correlation between oil and non-oil growth = 0.46.
- From 1996 onwards, oil sector growth appears to move in tandem with the REER; tradable non-oil sector is very small.

### Angola VAR estimation — caveats and key results
- Data limitations:
  - Exchange rate data available only from 1992; short sample; volatile conflict period weakens estimation power.
- VAR specification:
  - Three endogenous variables: oy, noy, rer.
  - Two dummies: PEACE for 2002–08; Dum_94 to capture the 1994 devaluation.
- Estimated short-run elasticities (selected exact figures):
  - A one percent growth in the oil sector leads to a 0.37 percent growth in the non-oil sector (NOY(t-1): 0.370*).
  - Accumulated IRF impact: a one percent oil shock leads to a 0.5 percent growth of the non-oil sector within 4 years.
- Additional Angola VAR table entries (selected exact figures):
  - OY(t-1) on RER: 1.294*
  - C (constants): OY 0.045, NOY 0.028, RER 0.065
  - PEACE: OY 0.117**, NOY 0.026, RER -0.124
  - Dum_94: OY 0.225, NOY -0.175*, RER -1.848*
  - Adj. R-squared: OY 0.506, NOY 0.928, RER 0.853
  - Number of Observations: 15
  - Significance notation: * significant at 1 percent; ** significant at 5 percent; *** significant at 10 percent
- Interpretation:
  - Oil-sector shocks lead to REER appreciation; sensitivity of the non-oil sector to REER movements is not significant in Angola’s estimates, so no adverse REER-channel impact on non-oil growth is detected.

### Constraints, impediments, and policy implications for Angola
- Identified hindrances to non-oil development:
  - Poor business climate
  - Poor infrastructure
  - Low effectiveness of SFI (reflected by low fiscal policy transparency and a wide gap between the non-oil primary fiscal balance and the sustainable level)
- Policy implications and recommendations (implied by analysis):
  - Substantial progress in institutional indicators (business climate, infrastructure, fiscal transparency/effectiveness) is necessary to:
    - Preserve current positive synergies between oil and non-oil sectors.
    - Avoid negative externalities observed in other high oil-intensity economies.
    - Allow greater positive spillovers between sectors and accelerate overall GDP growth.

### Comparative panel results and broader conclusions (selected exact figures)
- Panel VAR across 23 oil exporters (1985–2008):
  - “Natural resource curse” effect dominant only in high oil-intensity countries.
  - High oil-intensity estimated effect: a one percent growth of the oil sector leads to a decline of 0.15 percentage points in the non-oil sector’s growth in the subsequent two years.
- Low oil-intensity dynamics (selected exact figures):
  - One percent growth in the oil sector leads, on average, to about 0.1 percent growth in the non-oil sector in the subsequent year.
  - One percent growth in the non-oil sector leads to a 0.25 percent growth in the oil sector.
  - The non-oil sector contributes more to the oil sector than vice versa in low oil-intensity economies.

### Key statistics and diagnostics (selected exact figures)
- Non-oil sector share: 76 percent in 1991 → 32 percent in 1996.
- Non-oil sector average growth: 11.6 percent annually (1997–2008).
- Oil sector expansion: 7.8 percent (1997–2008).
- Correlations: 0.26 (1992–1996), 0.46 (1997–2008).
- Oil exports share of total exports: 95 percent.
- Angola VAR accumulated IRF impact: 0.5 percent non-oil growth within 4 years from a 1 percent oil shock.
- Panel result (high oil-intensity): 1 percent oil growth → decline of 0.15 percentage points in non-oil growth in subsequent two years.
- Low oil-intensity elasticities: oil → non-oil ≈ 0.1 percent (one year); non-oil → oil = 0.25 percent.
- Table A.1 (selected moments, full sample 1985–2008):
  - Non-oil sector growth: Mean 0.056408; Median 0.051853; Maximum 0.428917; Minimum -0.34444; Std. Dev. 0.075256; Observations 420.
  - Oil sector growth: Mean 0.058581; Median 0.023831; Maximum 2.38713; Minimum -1.20938; Std. Dev. 0.241096; Observations 407.
  - REER: Mean -0.02186; Median 0.000803; Maximum 0.76793; Minimum -1.89505; Std. Dev. 0.190623; Observations 491.
- Table A.7 (Selected governance/fiscal/business indicators, Angola exact entries):
  - Government Effectiveness Index*: -1.20
  - Open Budget Index 2008**: 3
  - 2009 Doing Business ranking***: 168

*Source: _wp10118 - 23. Angola is currently among the countries with highest oil-intensity in the sample.*

### Introduction ...........................................................................................................

### Introduction

### Background and scope
- Commodity exporters, particularly oil exporting countries, benefited from the pre-crisis boom in commodity prices; many recorded impressive economic growth and substantially strengthened their fiscal and external positions.
- This paper focuses on intersectoral linkages between oil and non-oil sectors in oil exporters using a Panel Vector Auto-Regression (PVAR) framework, examining to what extent growth rates of both sectors are linked while taking into account the real exchange rate (REER) channel.
- Sample: 23 oil exporting countries (Algeria, Angola, Azerbaijan, Bahrain, Cameroon, Chad, Congo, Equatorial Guinea, Gabon, Indonesia, Iran, Kazakhstan, Kuwait, Libya, Nigeria, Oman, Qatar, Saudi Arabia, Syria, Turkmenistan, United Arab Emirates, Venezuela and Yemen).

### Key stylized facts
- Oil intensity is heterogeneous across the sample, ranging from an average of 6 percent (Chad) to nearly 63 percent (Libya).
- Oil intensity is calculated as the average share of oil GDP in total GDP (constant prices) over 1985–2008.
- Low oil-intensity countries are defined as countries with an average (over the sample) weight of oil GDP in total GDP below the sample’s median observation (11 countries); high oil-intensity countries are those above the sample’s median observation (12 countries).
- Correlation patterns:
  - In low oil-intensity countries, oil and non-oil sectors growth are highly correlated: correlation = 0.67 (Figure 3a).
  - In high oil-intensity countries, the correlation between oil and non-oil sectors growth is much lower: correlation = 0.21 (Figure 3b).
- Empirical relationship across countries suggests a negative slope between oil intensity and the correlation of oil and non-oil sector growth (Figure 2).

### Channels and mechanisms discussed
- The paper reviews theoretical and empirical channels behind the “natural resource curse,” including:
  - Dutch Disease: REER appreciation and resource pull effects that crowd out the non-natural resource tradable sector.
  - Concentration and volatility: concentrated economies exposed to commodity price volatility, reducing investment and growth.
  - Policy and institutional effects: overconfidence in resource revenues leading to weaker fiscal discipline; state ownership through parastatals that soften budget constraints and may manage resources inefficiently.
  - Political-economy channels: rent-seeking behavior (Rentier Effect), higher corruption, Delayed Modernization and Entrenched Inequality effects, and greater propensity for violent conflict in resource-rich countries.
- Evidence on the REER channel in high oil-intensity countries:
  - REER cycles and oil sector growth cycles move in opposite directions: the REER tends to appreciate when oil sector growth accelerates and depreciate when oil sector growth decelerates (Figure 4a).
  - REER cycles and non-oil sector growth cycles tend to move together in high oil-intensity countries (Figure 4b).

### Main empirical findings (from PVAR estimation)
- The estimation results largely support the “natural resource curse,” but only in high oil-intensity countries:
  - In high oil-intensity countries, a one percent growth of the oil sector leads, on average, to an accumulated decline of 0.15 percent in the non-oil sector within two years.
- In low oil-intensity countries:
  - The oil sector has positive externalities on the non-oil sector, despite the adverse effect of REER appreciation caused by oil sector growth.
  - There is a positive and significant impact of the non-oil sector on the oil sector.
- The divergence of results between low and high oil-intensity countries supports the hypothesis that adverse natural-resource effects operate mainly in high oil-intensity settings, while positive externalities dominate in low oil-intensity settings.

### Case study: Angola
- Angola is among the highest oil-intensity countries in the sample but exhibits a positive link between the oil and non-oil sectors, perhaps reflecting the country’s unique circumstances.
- Figures and tables related to Angola include: Oil Intensity, 1991–2008; Oil Sector Growth and the Change in the REER, 1995–2008; The Oil and Non-Oil Sector Growth, 1992–2008; and Impulse Response for One-lag VAR of oy, noy and rer.

### Methodology and data notes
- The study uses a Panel Vector Auto-Regression (PVAR) framework to follow the dynamics of variables without imposing a specific structural economic model.
- Oil intensity and growth series rely on WEO and IMF staff calculations, with various moving averages used in graphical presentations (e.g., 2-year and 5-year moving averages, REER 4-year moving average with one year lag in some figures).

*Source: IMF staff calculations and analysis (from the paper’s Introduction and I. Some Stylized Facts).*

### 10. Given the public sector dominance in many of the selected oil exporters, an important

### 10. Given the public sector dominance in many of the selected oil exporters, an important

### Fiscal channel and macroeconomic transmission
- If oil revenue windfalls are used prudentially—directing public spending to investment in infrastructure and human capital in line with the economy’s absorption capacity—oil revenues can have positive externalities on the non-oil sector.
- Inefficient use of oil revenues can lead to:
  - increased inflationary pressures if output is close to potential,
  - sharp real exchange rate appreciation,
  - a decline in the marginal efficiency of capital,
  - a non-sustainable fiscal stance,
  - greater vulnerability to external shocks,
  - impaired external competitiveness,
  - discouraged foreign investment.

### Special Fiscal Institutions (SFIs), governance, and public spending quality
- Many oil exporters established SFIs to cope with oil price volatility and reduce vulnerability to external shocks; SFIs include oil funds, fiscal rules, fiscal responsibility legislation, and budgetary oil prices.
- SFIs are not a panacea; their quality and effectiveness vary across countries.
- World Bank “government effectiveness” indicator (based on 2007 data; units range from about -2.5 to 2.5):
  - Controlling for income level, government effectiveness tends to be lower in countries with high oil-intensity.
  - This implies SFIs’ effectiveness and quality may be higher in low oil-intensity countries.
- Possible explanations for weaker fiscal institutions in high oil-intensity countries:
  - Government over-confidence that oil revenue windfalls will be long-lasting.
  - Where oil revenues account for the lion’s share of total revenues and income taxes are generally low, the public exerts less pressure on governments to strengthen fiscal institutions.

### Budget transparency and investment climate
- Budget transparency (Open Budget Index, OBI) is negatively correlated with oil intensity in the sample of 12 countries with OBI data.
  - Lower transparency and poorer quality of public spending in high oil-intensity countries may adversely affect non-oil sector development.
- Business climate (proxied by the Doing Business ranking):
  - Positive link between oil intensity and worse Doing Business rankings in both higher and lower PPP per capita groups, suggesting weaker incentives to improve the business climate in high oil-intensity countries.

### Data, sample, and summary statistics
- Annual data for 1985–2008 for 23 developing oil-exporting countries (twenty of them where main export earning stems from oil).
- On average (1985–2008):
  - non-oil sector growth: 5.6 percent,
  - oil sector growth: 5.8 percent,
  - real effective exchange rate (REER) depreciated by 2.2 percent on average.
- High standard deviation of oil sector growth indicates large variation across countries and time.
- Variables pass unit-root tests (alternative tests unanimously reject unit-root null).

### Oil-intensity groups and sample partition
- Sample split by median oil-intensity of 33 percent:
  - Low sub-sample: oil-intensity below 33 percent.
  - High sub-sample: oil-intensity above 33 percent.
- Composition of sub-samples changes over time as oil-intensity evolves.
  - Among 23 countries: nine were constantly below 33 percent, four were constantly above 33 percent, others shifted over time.

### Methodology: Panel VAR (PVAR)
- PVAR combines VAR treating all variables as endogenous with panel data for unobserved heterogeneity.
- Endogenous variable vector Yit = {oy, noy, rer} where:
  - oy = oil sector growth rate (real terms),
  - noy = non-oil sector growth rate (real terms),
  - rer = change in real effective exchange rate.
- Fixed effects fi capture country specifics.
- Forward mean-differencing (Helmert procedure) used to remove fixed effects correlation with regressors.
- Estimation via GMM with lagged regressors as instruments.
- Ordering for Cholesky orthogonalization: oy (most exogenous), noy, rer.
- Lag length selection:
  - Low group: one-lag specification slightly superior by AIC.
  - High group: three-lag specification has significantly lower AIC.

### Estimation results — Low oil-intensity countries
- Contradicting the “natural resource curse,” oil sector growth has a significant and positive impact on non-oil sector growth.
  - Estimated coefficient: a one percent growth in the oil sector leads, on average, to a 0.1 percent growth in the non-oil sector within one year.
- Non-oil sector positively impacts oil sector:
  - Estimated coefficient: a one percent growth in the non-oil sector leads, on average, to a 0.26 percent growth in the oil sector in the subsequent year.
  - The non-oil sector’s contribution to the oil sector exceeds the reverse in this group.
- Impulse Response Functions (IRFs) show:
  - Shocks in both sectors lead to REER appreciation.
  - No indications that REER movements have significant effects on oil and non-oil sector growth.
- Variance decomposition:
  - Non-oil sector variance attributes a greater share of oil sector forecast error variance than REER does.
  - REER has a higher share of attributed non-oil sector growth forecast compared to oil sector.

### Estimation results — High oil-intensity countries
- Dynamics differ and align with the “natural resource curse”:
  - Estimated coefficients: a one percent growth in the oil sector leads, on average, to a decline of 0.15 percentage points in the non-oil sector’s growth within two years.
- REER effects:
  - Non-oil sector is adversely affected by REER appreciation within two to three years, but oil-sector shocks do not appear to trigger that appreciation.
  - REER appreciation also adversely affects the oil sector within three years (possible channels: profitability impacts, deterring foreign investment as domestic assets become more expensive for foreigners).
- Non-oil sector has no significant impact on the oil sector in the High group (consistent with a relatively small and undeveloped non-oil sector).
- Low share of non-oil sector forecast errors explained by REER variation supports the view that adverse oil-sector effects operate through channels not explicitly captured in the model (e.g., poor governance, weak fiscal institutions deterring investment and consumption).

### Illustrative country dynamics: Angola (figure references)
- Time-series figures for Angola illustrate:
  - Oil intensity, 1991–2008.
  - Oil sector growth and change in REER, 1995–2008.
  - Oil and non-oil sectors growth, 1992–2008.

*Source: IMF staff calculations and referenced data (WEO, World Bank governance indicators, International Budget Partnership, Doing Business Report).*

### 23. Angola is currently among the countries with highest oil-intensity in the sample.

### _wp10118 - 23. Angola is currently among the countries with highest oil-intensity in the sample.

### Overview
- The prolonged civil war led to the collapse of the non-oil sector: its share declined from 76 percent in 1991 to 32 percent in 1996.
- Since 1997 the non-oil sector registered average growth of 11.6 percent annually (1997–2008); the oil sector expanded in parallel by 7.8 percent over the same period, keeping the non-oil share stable at around 40 percent in the last eight years.
- Oil exports account for 95 percent of total exports.

### Sector dynamics and correlations
- Between 1992 and 1996, the correlation between oil and non-oil growth was 0.26; from 1997 to 2008 it rose to 0.46.
- From 1996 onwards, oil sector growth appears to move in tandem with the REER, implying past accelerations of oil-sector growth contributed to REER appreciation.
- The tradable non-oil sector in Angola is very small.

### VAR estimation for Angola (caveats)
- Data limitations: exchange rate data available only from 1992; limited data availability shortens the sample size significantly and the volatile conflict period weakens estimation power.
- VAR system contains three endogenous variables: oy (oil-sector growth), noy (non-oil sector growth), and rer (real effective exchange rate).
- Two dummies included in estimation: PEACE for 2002–08; Dum_94 to capture the 1994 devaluation.
- Results should be viewed with caution given short sample and volatile period.

### Angola VAR estimation results and impulse responses
- Estimated short-run elasticities:
  - A one percent growth in the oil sector leads to a 0.37 percent growth in the non-oil sector (estimated coefficient).
  - Impulse response functions (IRFs): the accumulated impact of a one percent oil shock leads to a 0.5 percent growth of the non-oil sector within 4 years.
- IRFs show an oil-sector shock leads to REER appreciation; the sensitivity of the non-oil sector to REER movements is not significant, so there is no adverse impact on non-oil growth via the REER channel.
- Table 2 (Angola VAR, 1994–2008) selected estimates and diagnostics:
  - NOY(t-1): 0.370* (impact on NOY)
  - OY(t-1) on RER: 1.294*
  - C (constants): OY 0.045, NOY 0.028, RER 0.065
  - PEACE: OY 0.117**, NOY 0.026, RER -0.124
  - Dum_94: OY 0.225, NOY -0.175*, RER -1.848*
  - Adj. R-squared: OY 0.506, NOY 0.928, RER 0.853
  - Number of Observations: 15
  - Significance level: * significant at 1 percent; ** significant at 5 percent; *** significant at 10 percent

### Constraints, impediments, and policy implications for Angola
- Identified hindrances to non-oil development:
  - Poor business climate
  - Poor infrastructure
  - Low effectiveness of SFI (as reflected by low fiscal policy transparency and a wide gap between the non-oil primary fiscal balance and the sustainable level)
- Policy implications and recommendations (implied by analysis):
  - Substantial progress in institutional indicators (business climate, infrastructure, fiscal transparency/effectiveness) is necessary to:
    - Preserve current positive synergies between oil and non-oil sectors.
    - Avoid negative externalities observed in other high oil-intensity economies.
    - Allow greater positive spillovers between sectors and accelerate overall GDP growth.

### Comparative panel results and broader conclusions
- Panel VAR across 23 oil exporters (1985–2008) finds:
  - The “natural resource curse” effect is dominant only in high oil-intensity countries.
  - Estimated effect in high oil-intensity countries: a one percent growth of the oil sector leads to a decline of 0.15 percentage points in the non-oil sector’s growth in the subsequent two years.
  - In high oil-intensity countries there are indications the non-oil sector negatively responds to REER appreciation, but the adverse effect of oil on non-oil does not appear to occur primarily through the REER channel.
- Low oil-intensity countries dynamics:
  - One percent growth in the oil sector leads, on average, to about 0.1 percent growth in the non-oil sector in the subsequent year.
  - One percent growth in the non-oil sector leads to a 0.25 percent growth in the oil sector.
  - The non-oil sector contributes more to the oil sector than vice versa in low oil-intensity economies.

### Key statistics and diagnostics (selected exact figures from the source)
- Non-oil sector share: 76 percent in 1991 → 32 percent in 1996.
- Non-oil sector average growth: 11.6 percent annually (1997–2008).
- Oil sector expansion: 7.8 percent (1997–2008).
- Correlations: 0.26 (1992–1996), 0.46 (1997–2008).
- Oil exports share of total exports: 95 percent.
- Angola VAR accumulated IRF impact: 0.5 percent non-oil growth within 4 years from a 1 percent oil shock.
- Panel result (high oil-intensity): 1 percent oil growth → decline of 0.15 percentage points in non-oil growth in subsequent two years.
- Low oil-intensity elasticities: oil → non-oil ≈ 0.1 percent (one year); non-oil → oil = 0.25 percent.
- Table A.1 (selected moments, full sample 1985–2008):
  - Non-oil sector growth: Mean 0.056408; Median 0.051853; Maximum 0.428917; Minimum -0.34444; Std. Dev. 0.075256; Observations 420.
  - Oil sector growth: Mean 0.058581; Median 0.023831; Maximum 2.38713; Minimum -1.20938; Std. Dev. 0.241096; Observations 407.
  - REER: Mean -0.02186; Median 0.000803; Maximum 0.76793; Minimum -1.89505; Std. Dev. 0.190623; Observations 491.
- Table A.7 (Selected governance/fiscal/business indicators, exact entries for Angola):
  - Government Effectiveness Index*: -1.20
  - Open Budget Index 2008**: 3
  - 2009 Doing Business ranking***: 168

*Italic source attribution line.*

*Source: _wp10118 - 23. Angola is currently among the countries with highest oil-intensity in the sample.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp10118.pdf_