## 1. Evolution of Debt in the Middle-East Since the Early 1990s by Country Groups

## Source details

**Canonical URL:** [1. Evolution of Debt in the Middle-East Since the Early 1990s by Country Groups](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp0712.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp0712.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp0712.pdf.json)

---

### I. Introduction — scope and key questions
- Sample: 19 Middle-Eastern countries (all Arab countries except Iraq, Sudan, West Bank and Gaza, and Somalia; plus Iran and Pakistan).
- Core questions:
  - Why have Middle Eastern countries had such high public debt levels?
  - Are the current levels sustainable?
- Approach:
  - Examine origins of public debt, decompose recent debt evolution into policy and non-policy factors, and explain the absence of defaults.
  - Assess sustainability by comparing current with sustainable primary balances, estimating policy reaction functions for country-groups, and discussing vulnerabilities.
- Note: 2003 chosen as a turning point before the recent oil price hikes.

### II. Country groups, public-debt definition, and data sources
- Country sub-groups:
  - High-income oil producers: Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and the UAE (Gulf Cooperation Council).
  - Middle-income oil producers: Algeria, Iran, Libya, Syria, and Yemen.
  - Middle-income non-oil producers: Egypt, Jordan, Lebanon, Morocco, and Tunisia.
  - Low-income non-oil producers: Djibouti, Pakistan, and Mauritania.
- Public debt definition and reporting differences:
  - “Public debt is defined as broadly as each country’s data allows.”
  - Data types by country: Egypt, Libya, Oman, Saudi Arabia, and UAE — public sector; Jordan, Pakistan, Tunisia — general government; remainder — central government.
  - Unless indicated, debt is gross debt without adjustments for bank deposits or other assets.
- Data sources: IMF staff reports and country desks; supplemented by World Bank, World Economic Outlook, and International Financial Corporation.

### III. Major empirical findings on debt levels, composition, and trends
- Relative positions (2003):
  - About half of Middle-Eastern countries had a debt level above the emerging markets’ mean of about 60 percent of GDP.
  - More than two-thirds (including all non-oil producing countries) were above the 50 percent benchmark (the median debt level of vulnerable emerging market economies).
- Regional evolution (1990–2005):
  - Over the whole period, debt declined by about 26 percentage points of GDP to around 55 percent.
  - The bulk of the reduction was from external public debt; domestic debt remained fairly constant at about 30 percent of GDP.
- Cross-group patterns (high-level):
  - High-income oil producers’ debt peaked in 1998; they relied mainly on domestic debt to finance deficits.
  - Middle-income non-oil countries: debt fell early 1990s partly due to debt relief, then increased as domestic debt accumulation outpaced external debt decline.
  - Middle-income oil producers: worse debt ratios in early 1990s driven by external debt accumulation; sharper debt reduction thereafter.
  - Low-income non-oil producers: very high debt ratios in the 1990s, declined somewhat since then.
- Composition highlights (2003 summary):
  - Debt highest in non-oil countries; some oil producers (Saudi Arabia and Qatar) also had high gross public debt.
  - Domestic debt share roughly constant at about 30 percent of GDP; external debt provided most of the decline.

### IV. Oil dependence, fiscal balances, and debt dynamics
- Regional dependence on oil:
  - Oil dependence (direct and indirect) influences debt dynamics across oil and non-oil countries via grants, remittances, regional trade, financial services, and tourism.
- Revenue and deficits:
  - Average revenue ratios in the region are well above the emerging market average.
  - High-income oil producers recorded an average revenue ratio as high as 40 percent of GDP, driven by oil proceeds.
  - The region has consistently recorded fiscal deficits — surpluses were only run by oil producers since the turn of the century.
- Volatility and expenditure rigidities:
  - Wide fluctuations in government deficits reflect volatility in oil revenue and slow adjustments in expenditures and other revenue.
  - Oil-boom spending in the 1970s increased structural expenditure levels; debt accumulation rose when oil prices declined.
- Structural contributors to debt accumulation:
  - Rigidities in expenditure, a volatile revenue base, and weak fiscal institutions hinder prompt fiscal responses, favor debt financing over adjustment.

### V. Quantitative markers and selected statistics
- Benchmarks and reported values:
  - Emerging markets mean public debt: about 60 percent of GDP.
  - Vulnerable emerging market median debt benchmark: 50 percent of GDP.
  - Regional debt decline over 1990–2005: about 26 percentage points of GDP to around 55 percent.
  - Average revenue for high-income oil producers: as high as 40 percent of GDP.
  - Domestic debt level roughly constant at about 30 percent of GDP.
  - High-income oil producers: debt peak in 1998 (group-level observation).

### VI. Box 2 — Regional characteristics that shaped debt dynamics
- Structural characteristics encouraging debt accumulation:
  - Wage bill averaged 15 percent of non-oil GDP (and was particularly high for oil producers), versus an average of 7 percent of GDP in 1997-2004 in emerging markets.
  - Non-oil tax revenue averaged 11 percent of non-oil GDP.
  - Grants in some non-oil producing countries averaged above 6 percentage points of GDP per year, with a standard deviation of 3 percentage points, and annual ranges from 2 percent of GDP to 14 percent.
  - Large budget fragmentation, extra-budgetary funds, pervasive off-budget spending, and deficiencies in budget execution and monitoring.
- Fiscal patterns in oil-producing countries:
  - Deficits as high as 60 percent of non-oil GDP on average in Oman (closely followed by Saudi Arabia).
  - Average non-oil primary deficits were high at 23 percent of GDP, improving for rich oil producers at the end of the sample period but deteriorating in middle-income oil countries partly because of higher oil prices.
- Debt decomposition (1999–2005; 1999–2003 vs 2004–05):
  - In oil-producing countries, the increase in oil revenue explains most of the reduction in debt in 2004–05.
  - In 1999–2003 oil revenues primarily financed the non-oil primary deficit; in 2004–05 higher oil revenue financed larger reductions in gross debt and asset accumulation (example: Algeria prepaid large part of its debt since 2004).
  - A large “other” category includes privatization revenue, debt relief, and asset accumulation and is significant for most countries.
- Non-oil country patterns:
  - Middle-income non-oil: debt declined modestly in 2004–05 after increasing in 1999–2003; growth reduced debt but high interest costs and real effective depreciation offset growth, increasing average debt ratios by 3 percent of GDP in 1999–2003.
  - Low-income non-oil: large debt reduction driven by strong growth and low/negative real interest rates; primary deficits initially added to debt but improved toward end of period.
- Pakistan case study (Box 3 summary):
  - Public debt peaked at almost 90 percent of GDP at end-June-2001, slightly over half external.
  - Turnaround drivers: Paris Club restructuring (December 2001), simultaneous restructuring of $600 million Eurobonds and $500 million short-term bank credits; fiscal tightening with primary surplus averaging over 2 percent of GDP between 2001/02 and 2003/04; lower domestic interest rates; accelerating economic activity, remittances, exports; 2004 tapping external market to retire high-interest external debt.
- Features that insulated the region from major debt crises:
  - Rich oil producers can quickly tap natural resources; region includes three of the ten top world producers of oil, five out of the ten top net exporters, and five (including Iraq) of the top ten countries with the highest proven reserves.
  - Oil producers often hold assets close to or exceeding debt levels.
  - Very little marketable debt; most external debt held by official creditors (bilateral and multilateral) except Lebanon and Qatar.
  - Favorable borrowing conditions: effective interest rate on external debt averaged 3 percent for countries with available data; Lebanon paid 17 percent.
  - Broad money-to-GDP ratio averaged 67 percent; domestic debt as ratio of M3 averaged around 55 percent.
  - Some countries replaced external nonconcessional debt with domestic local-currency debt to reduce debt service cost.

### VII. Box 4 — Lebanon: living with high public debt
- Debt levels and composition:
  - Lebanon’s debt-to-GDP ratio reached 176 percent of GDP in 2005, and is expected to increase further due to the conflict with Israel.
  - About half of the debt is denominated in foreign currency.
  - Gross public debt (excluding monetary liabilities) of over 175 percent of GDP and gross financing needs of over 65 percent of GDP in 2005.
- Risks in the public debt stock:
  - Exchange rate risk from foreign-currency debt.
  - Rollover risk: at least one quarter of debt has a residual maturity of a year or less.
  - Interest rate risk: short average maturity increases sensitivity to interest rates.
- Funding structure:
  - Domestic banking system is the main funding source; less than 20 percent of public sector debt is owed to nonresidents.
  - Banking sector deposit base grew to about 300 percent of GDP, enabling financing of the government.
- Factors supporting banking-sector resilience:
  - Large dedicated depositor base (Lebanese Diaspora and Arab investors).
  - Government’s ability to mobilize extraordinary official financing (e.g., Paris II).
- Fiscal-sustainability assessment:
  - Of eight non-oil countries in the sample, Lebanon is one of two that would have to adjust by 3 percentage points of GDP compared to the historical average to stabilize current debt levels (lower bound benchmark).
  - To reduce debt to 50 percent of GDP over an arbitrarily chosen period of ten years (upper bound benchmark), Lebanon would need to run a primary surplus about 17 percentage points of GDP per year higher than the historical average.
- Caveats:
  - The “safe” debt target is uncertain given risks from future interest and growth rates, contingent liabilities, and governments’ ability to adjust after shocks.
  - Historical policies used as indicators of politically and socially feasible future policies; econometric approach allows short-term deviations but relies on shorter time series and group averages.

### VIII. Fiscal reaction, regressions, and empirical markers
- Mean-reverting fiscal response:
  - Public debt dynamics appear mean-reversing: "0.8 percent of GDP for every 10 percentage point of GDP increase in public debt."
- Fiscal reaction via revenues and expenditures:
  - For every 10 percentage point increase in the debt to GDP ratio, the revenue ratio increases by 0.4 percent.
  - Expenditures react with the correct sign, but the effect is not statistically significant.
- Difficulty detecting cyclical primary balance responses:
  - Neither the output gap nor the oil-price gap turned out statistically significant in reported baseline specifications; output-gap coefficients were low and insignificant and dropped from reported specifications.
- Regression highlights (Non-Oil Countries, Table 2 excerpts):
  - Debt_1 / GDP_1 coefficients reported: 0.083** 0.048*** 0.061*** 0.039*** -0.022
  - Primary balance_1 / GDP_1: 0.492***
  - R2 values: 0.32 0.19 0.19 0.02
  - Sample sizes (N): 104 95 104 104 104
  - Note: * p<.1; ** p<.05; *** p<.01. All specifications include fixed effect, except dynamic baseline estimated using Arellano-Bond method.
- Oil-producing countries (POIM testing and results):
  - Empirical model: non-oil primary balance as a function of present value of oil revenue and past net debt (ratios to non-oil GDP).
  - Policies seem not to be driven by the POIM: coefficient of present value of oil revenue small and insignificant.
  - Link between non-oil primary balance and net debt is strong: an increase in the net debt ratio of 10 percent of GDP improves the primary balance by about 0.3 percent of GDP.
  - Cyclical responses to oil prices are weak but pro-cyclical: a decrease in the oil price gap by 10 percent increases the non-oil primary balance by 0.8 percent of non-oil GDP.
- Regression excerpts (Oil Producers, Table 3 reported coefficients and stats):
  - Net debt_1 / non-oil GDP coefficients: 0.031*** 0.033*** 0.030*** 0.029*** 0.022*** -0.010*
  - Assets_1 / non-oil GDP: -0.032***
  - Oil wealth / non-oil GDP: 0.000
  - Non-oil primary balance_1 / non-oil GDP_1: 0.442***
  - Oil price gap coefficients across specifications include: -0.079* -0.078** -0.058* -0.162** -0.077** 0.056** 0.152***
  - Constants and R2 range reported; N values: 108 108 99 108 108 108.
  - Note: * p<.1; ** p<.05; *** p<.01. All specifications include fixed effect, except dynamic baseline estimated using Arellano-Bond method.

### IX. Identified vulnerabilities and risks
- Volatility risks:
  - Volatility of oil revenue and grants is an important vulnerability, especially since non-oil countries rely heavily on revenue mobilization because of expenditure rigidities.
- Data and transparency risks:
  - Narrow coverage and large off-budget operations.
  - Lack of systematic data on contingent liabilities (pensions, bank restructuring, PPP guarantees) means government liabilities may arise without forewarning.
- Market access and financing cushion:
  - Gradual opening to international capital markets could remove stable regional financing cushions, though near-term liquidity may remain high while oil prices stay high.
- Rising domestic debt risks:
  - If not managed well, rising domestic debt can lead to inflation, low growth, and loss of confidence, and crowd out private investment.
- Persistent high debt:
  - Debt levels remain relatively high for non-oil producing countries and middle-income oil producers.

### X. Policy conclusions and recommendations
- Broad conclusions:
  - Public debt in the Middle East rose in the mid-nineties mainly due to large non-oil primary deficits in oil producing countries; recent decreases driven mainly by high oil revenue.
  - On average, fiscal policies have reacted to past high indebtedness and avoided default or debt crises.
  - Public debt levels remain uncomfortably high in many economies, particularly in non-oil producing countries and middle-income oil producers.
- Policy priorities implied by findings:
  - Non-oil producing countries: substantial adjustment of the primary balance would be needed to bring debt down to a safer level.
  - Increase expenditure flexibility in non-oil producing countries since adjustments rely mainly on increasing revenues.
  - Middle-income oil producing countries: may need to adjust policies to align with the Permanent Oil Income Model (POIM); specific recommendations should account for uncertainties and country-specific differences.
  - Strengthen data coverage, transparency, and reporting of contingent liabilities to reduce vulnerability.
  - Manage rising domestic debt carefully to avoid inflationary pressures, crowding out private investment, and loss of confidence.

### XI. Data and measurement issues (Appendix I summary) and methodological notes (Appendix II)
- Data limitations and availability problems:
  - Dataset: fiscal and public debt data for 19 countries from 1990 to 2005 compiled from IMF staff reports and country economists; some empirical work uses 1999–2004 because of gaps prior to 1999.
  - Examples of missing/broken series: UAE prior to 1999 (ADIA data), Egypt prior to 1999 (structural break), Pakistan (no breakdown of salaries/wages), Qatar (total financing 1990–92 missing), Kuwait (1990–93 Gulf War disruption).
  - Public debt reporting coverage varies: public sector (Egypt, Libya, Oman, Saudi Arabia, UAE), general government (Jordan, Pakistan, Tunisia), central government (remaining 11).
  - Detailed debt composition data (currency denomination, maturity, interest type) generally missing.
  - Government domestic and external assets data often unavailable, especially for GCC countries.
  - Contingent liabilities data largely unavailable.
- Appendix II — key methodological assumptions and specifications:
  - Non-oil producers baseline empirical specification: p_{i,t} = α_i + ρ b_{i,t-1} + β X_{i,t} + ε_{i,t}, where p is primary balance (% of GDP), b is debt stock (% of GDP), X includes output gap and oil-price gap.
  - For oil producers, empirical counterpart: p_{i,t} = α_i + ρ b_{i,t-1} + γ w_{i,t-1} + β X_{i,t} + ε_{i,t}, where w is oil wealth (% of non-oil GDP).
  - Consistency with theoretical optimality (POIM) if ρ = r/(1 + r) and γ = r/(1 + r).
  - Oil-wealth estimation assumptions:
    - Future production fixed at last observable level in energy-equivalent terms; gas-to-oil-equivalent conversion: 1 barrel oil = 5,487 cubic feet natural gas.
    - Future oil prices projected using a two-factor model (Schwartz and Smith (2000)), re-estimated with 1990–2005 data; conservative estimates use estimated drift close to zero; optimistic estimates replace drift with 2 percent US inflation.
    - Inflation, real interest rates, and real exchange rates assumed constant; PPP holds.
    - Real interest rate and discount rate: 3 percent.
  - Extensions tested: asymmetric effects for positive/negative output gaps and oil-price deviations, subgroup analyses for oil producers, threshold/spline tests for debt response (non-significant), separate revenue and expenditure functions.

*Source: _wp0712 - 1. Evolution of Debt in the Middle-East Since the Early 1990s by Country Groups; Box 2; Box 4; regression results and appendices (IMF PDF).*

### 1. Evolution of Debt in the Middle-East Since the Early 1990s by Country Groups ................7

### 1. Evolution of Debt in the Middle-East Since the Early 1990s by Country Groups ................7

### I. Introduction — key context and questions
- The paper studies 19 Middle-Eastern countries (all Arab countries except Iraq, Sudan, West Bank and Gaza, and Somalia; plus Iran and Pakistan).
- Two core questions:
  - Why have Middle Eastern countries had such high public debt levels?
  - Are the current levels sustainable?
- Approach:
  - Examine origins of public debt, decompose recent debt evolution into policy and non-policy factors, and explain the absence of defaults.
  - Assess sustainability by comparing current with sustainable primary balances, estimating policy reaction functions for country-groups, and discussing vulnerabilities.

### II. Country groups and data definitions
- Country sub-groups:
  - High-income oil producers: Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and the UAE (Gulf Cooperation Council).
  - Middle-income oil producers: Algeria, Iran, Libya, Syria, and Yemen.
  - Middle-income non-oil producers: Egypt, Jordan, Lebanon, Morocco, and Tunisia.
  - Low-income non-oil producers: Djibouti, Pakistan, and Mauritania.
- Public debt definition:
  - “Public debt is defined as broadly as each country’s data allows.”
  - Data types by country: Egypt, Libya, Oman, Saudi Arabia, and UAE — public sector; Jordan, Pakistan, Tunisia — general government; remainder — central government.
  - Unless indicated, debt is gross debt without adjustments for bank deposits or other assets.
- Data sources: IMF staff reports and country desks; supplemented by World Bank, World Economic Outlook, and International Financial Corporation.
- Note: 2003 chosen as a turning point before the recent oil price hikes.

### III. Major empirical findings on debt levels and composition
- Relative levels:
  - In 2003, about half of Middle-Eastern countries had a debt level above the emerging markets’ mean of about 60 percent of GDP.
  - In 2003, more than two-thirds (including all non-oil producing countries) were above the 50 percent benchmark (the median debt level of vulnerable emerging market economies).
- Overall regional evolution (1990–2005):
  - Over the whole period, debt declined by about 26 percentage points of GDP to around 55 percent.
  - The bulk of the reduction was from external public debt; domestic debt remained fairly constant at about 30 percent of GDP.
- Cross-group patterns (high-level):
  - High-income oil producers’ debt peaked in 1998; they relied mainly on domestic debt to finance deficits.
  - Middle-income non-oil countries’ debt fell at the beginning of the 1990s partly due to debt relief, but increased steadily thereafter as domestic debt accumulation outpaced the decline in external debt.
  - Middle-income oil producers had worse debt ratios than higher-income neighbors in the early 1990s; external debt accumulation drove rising ratios then, but debt reduction since then was much sharper.
  - Low-income non-oil producers’ debt ratios remained very high in the 1990s but have declined somewhat since then.
- Composition in 2003 (selected summary):
  - Debt was highest in non-oil countries; some oil producers (Saudi Arabia and Qatar) also had high gross public debt.
  - Domestic debt accounted for a roughly constant share (~30 percent of GDP) across the period, with external debt providing most of the decline.

### IV. Oil dependence, fiscal balances, and debt dynamics
- Oil dependence:
  - The Middle East is highly dependent on oil—directly or indirectly—affecting regional debt dynamics for both oil and non-oil countries (via grants, remittances, regional trade, financial services, and tourism).
- Revenue and deficits:
  - Average revenue ratios in the region are well above the emerging market average.
  - High-income oil producers recorded an average revenue ratio as high as 40 percent of GDP, driven by oil proceeds.
  - The region has consistently recorded fiscal deficits — surpluses were only run by oil producers since the turn of the century.
- Volatility and expenditure rigidities:
  - Wide fluctuations in government deficits reflect volatility in oil revenue and slow adjustments in expenditures and other revenue.
  - The oil boom in the 1970s increased spending to high levels; non-oil and middle-income oil producers financed higher expenditures by accumulating debt, accelerating when oil prices declined.
  - High-income oil producers built up assets despite higher spending, but some countries’ situations worsened due to large military and reconstruction expenditures during the first Gulf War (typically outside the budget).
- Structural contributors to debt accumulation:
  - Rigidities in expenditure, a volatile revenue base, and weak fiscal institutions hinder prompt fiscal responses to shocks, favoring debt financing over adjustment.

### V. Quantitative markers and graphical evidence (as reported)
- Figure and benchmark highlights:
  - Emerging markets mean public debt: about 60 percent of GDP (used as comparison).
  - Vulnerable emerging market median debt benchmark: 50 percent of GDP.
  - Regional debt decline over 1990–2005: about 26 percentage points of GDP to around 55 percent.
  - Average revenue for high-income oil producers: as high as 40 percent of GDP.
  - Domestic debt level roughly constant at about 30 percent of GDP.
  - High-income oil producers: debt peak in 1998 (group-level observation).

### VI. Analytical implications drawn in this section
- High debt ratios in the Middle East have multiple origins:
  - Direct dependence on volatile oil revenues and indirect regional transmission channels.
  - Persistent fiscal deficits outside oil-producer surpluses since the turn of the century.
  - Structural rigidities and weak fiscal institutions that impede timely fiscal adjustment.
- The recent decline in public debt has been driven mainly by reductions in external public debt rather than domestic debt contraction.
- Group heterogeneity matters: policy responses and debt trajectories differ substantially across high-income oil producers, middle-income oil producers, middle-income non-oil producers, and low-income non-oil producers.

*Source: _wp0712 - 1. Evolution of Debt in the Middle-East Since the Early 1990s by Country Groups ................7 (IMF PDF)*

### Box 2. The Role of Regional Characteristics in Shaping Debt Dynamics

### Box 2. The Role of Regional Characteristics in Shaping Debt Dynamics

### Structural characteristics that encouraged debt accumulation
- Rigidities in expenditure:
  - Wage bill averaged 15 percent of non-oil GDP (and was particularly high for oil producers), compared to an average of 7 percent of GDP in 1997-2004 in emerging markets.
- Uncertain and volatile revenue:
  - Non-oil tax revenue averaged 11 percent of non-oil GDP.
  - Grants in some non-oil producing countries averaged above 6 percentage points of GDP per year, with a standard deviation of 3 percentage points, and annual ranges from 2 percent of GDP to 14 percent.
- Weak fiscal institutions and public financial management:
  - Large budget fragmentation: investment budget separate from current budget; extra-budgetary funds widespread; pervasive off-budget spending.
  - Particular challenge: managing oil revenue to ensure appropriate spending within a medium-term sustainable context.
  - Deficiencies in budget execution and monitoring.
- Revenue administration improvements and remaining challenges:
  - Progress noted in VAT implementation, function-based tax administrations, establishment of Large Taxpayers Offices, and customs modernization.
  - Remaining challenges: low compliance with tax laws, insufficient political commitment, lack of capacity, expertise and resources, and ineffective customs administration.
  - Some countries (Egypt, Yemen) were until recently reluctant to implement a full-fledged modern VAT.
- Data and governance limitations:
  - Narrow coverage of the public sector; widespread off-budget operations.
  - Lack of data on contingent liabilities (future pension obligations, potential costs of bank restructuring, guarantees in PPPs) and public assets.
  - Relatively weak governance and transparency: seven countries in the sample in the top one-third of the Corruption Perception Index—five of which are rich oil producers.
  - In general, countries with large debt tend to have weaker institutional indicators.

### Debt and non-oil fiscal balances in oil-producing countries (key patterns)
- Oil producers ran large non-oil fiscal deficits throughout the period:
  - Deficits as high as 60 percent of non-oil GDP on average in Oman (closely followed by Saudi Arabia).
  - Average non-oil primary deficits were high at 23 percent of GDP, improving for rich oil producers at the end of the sample period, but deteriorating in middle-income oil countries partly because of higher oil prices.
- Decomposition of debt dynamics (1999–2005), split into 1999–2003 and 2004–05 to capture oil price effects:
  - In oil-producing countries, the increase in oil revenue explains most of the reduction in debt in 2004–05.
  - In 1999–2003 oil revenues were primarily used to finance the non-oil primary deficit; in 2004–05 higher oil revenue financed larger reductions in gross debt and the building up of financial assets (example: Algeria prepaid a large part of its debt to bilateral, multilateral and private creditors since 2004).
  - Middle-income oil producers in 2004–05 benefited from a higher contribution of oil revenue than high-income oil producers, but this was partly offset by a wider non-oil primary deficit for middle-income oil producers (while higher-income producers kept their non-oil primary balance roughly unchanged).
  - Note: a large “other” category in the decomposition includes privatization revenue, debt relief, and asset accumulation, and is significant for most countries, mainly reflecting asset accumulation.

### Debt evolution in non-oil producing countries
- Middle-income non-oil countries:
  - Debt declined modestly in 2004–05, while it increased in 1999–2003.
  - Economic growth was a strong force in reducing debt in both periods, but high interest cost and real effective depreciation more than offset growth, resulting in an increase in average debt ratios of 3 percent of GDP in 1999–2003.
  - The primary balance had a relatively small impact in both periods; the 1999–2003 result was dominated by poor performance in Lebanon and Egypt, while Jordan and Morocco reduced debt by 6–7 percentage points of GDP, with Tunisia essentially unchanged. In Jordan and Morocco, strong growth and privatization proceeds were main contributors to debt reduction.
- Low-income non-oil countries:
  - Achieved a large reduction in debt benefiting from good economic growth and low interest rates.
  - Reduction in both sub-periods mostly triggered by strong growth; negative real interest rates also contributed.
  - Primary deficits added to debt, but primary positions mildly improved towards the end of the period and in some countries played a significant role in reducing debt.
  - The impact of real effective depreciation turned from positive (increasing debt) in the first sub-period to negative in the second.

### Case study — Pakistan (Box 3): how public debt was reduced
- Pakistan’s public debt peaked at almost 90 percent of GDP at end-June-2001, slightly over half external.
- Turnaround drivers:
  - Paris Club restructuring agreement in December 2001 and simultaneous restructuring of $600 million in Eurobonds and $500 million in short-term credits held by commercial banks eased external debt service burden.
  - Fiscal policy tightened; primary surplus averaged over 2 percent of GDP between 2001/02 and 2003/04.
  - Domestic interest rates declined, lowering government cost of borrowing.
  - Accelerating economic activity, rising remittances and exports helped reduce public debt; in 2004 the government tapped the external debt market to retire high-interest external debt with new lower interest liabilities.

### Why the region avoided major debt crises
- Special financing features that partially insulated regional debt from typical market responses:
  - Rich oil producers can quickly tap vast natural resources, facilitating access to financing; the region includes three of the ten top world producers of oil, and five out of the ten top net exporters; five (including Iraq) of the top ten countries with the highest proven reserves are in the Middle East.
  - Oil producers generally hold assets close to, or exceeding, their debt levels, lowering vulnerability and improving market confidence, though raising questions on optimal fiscal management and reflecting asset management policies, rigidities in oil fund operations, or sterilization needs.
  - Very little marketable debt in the region—most external debt is held by official creditors (bilateral and multilateral), except Lebanon and Qatar; domestic debt may have non-market terms in a number of countries.
  - Favorable borrowing conditions:
    - Effective interest rate on external debt averaged 3 percent for countries with available data.
    - Lebanon paid a substantially higher interest rate of 17 percent.
  - Middle and low income non-oil producers benefited from a number of Paris Club reschedulings through the nineties and early 2000s.
  - Large and dedicated regional investor base enhanced market confidence (notably in Lebanon).
  - High monetization:
    - Broad money-to-GDP ratio averaged 67 percent during the period.
    - Domestic debt as a ratio of M3 averaged around 55 percent.
  - Some countries consciously replaced external nonconcessional debt with domestic (local currency) debt to reduce debt service cost and limit external vulnerabilities.

*Source: _wp0712 - Box 2. The Role of Regional Characteristics in Shaping Debt Dynamics*

### Box 4. Lebanon—Living with High Public Debt

### Box 4. Lebanon—Living with High Public Debt

### Debt dynamics and key statistics
- Lebanon’s debt-to-GDP ratio reached 176 percent of GDP in 2005, and is expected to increase further due to the conflict with Israel.  
- About half of the debt is denominated in foreign currency.  
- Up to 2002, the increase in debt was mostly explained through large primary deficits; primary deficits were turned into surpluses in 2003 through 2005.  
- The Paris II rescheduling in 2002 and generally lower interest rates since then moderated the negative impact from high real interest rates on debt accumulation.  
- The pick-up in economic growth between 2002 and 2004 contributed to stabilizing the debt-to-GDP ratio.  
- Gross public debt (excluding monetary liabilities) of over 175 percent of GDP and gross financing needs of over 65 percent of GDP in 2005 are far beyond ratios typically seen in emerging market countries.  
- Structure-related risks in the public debt stock:
  - Exchange rate risk: high share of foreign-currency denominated debt.
  - Rollover risk: at least one quarter of debt has a residual maturity of a year or less.
  - Interest rate risk: short average maturity increases sensitivity to interest rates.

### Funding structure and concentration of risk
- The domestic banking system is the public sector’s main funding source; the main rollover and interest rate risks are borne by the domestic banking system.  
- Less than 20 percent of the public sector’s debt is owed to nonresidents (who mostly hold foreign-currency debt).  
- The remainder of public debt is held by residents, mainly domestic banks; therefore, the banking sector’s willingness to roll over public debt holdings—without demanding a much larger risk premium—determines debt sustainability.  
- Banks’ ability to roll over public debt depends on their ability to renew their own monetary liabilities; the banking sector’s deposit base grew to about 300 percent of GDP, enabling its financing of the government.

### Factors supporting banking-sector resilience and continued funding
- Persistent funding of very high public financing needs through the domestic banking system is enabled by a large and dedicated depositor base (Lebanese Diaspora and Arab investors).  
- Indications that inflows from regional investors increased as a result of events post-September 11.  
- The government’s ability to mobilize extraordinary levels of official financing (such as under the donor conference of Paris II) may also play a role.  
- Maintaining confidence in the domestic banking system is key to Lebanon’s financial sustainability.  
- In the absence of substantial foreign assistance, the aftermath of the events of July 2006 is likely to put further strain on Lebanon’s public finances.

### Fiscal-sustainability assessment and required adjustments
- Using the paper’s methodology (primary gap and policy-reaction/solvency tests), Lebanon shows large fiscal adjustment needs relative to peers:
  - Of the eight non-oil countries in the sample, Lebanon is one of two that would have to adjust by 3 percentage points of GDP compared to the historical average to stabilize their current debt levels (lower bound benchmark).  
  - To reduce debt to 50 percent of GDP over an arbitrarily chosen period of ten years (upper bound benchmark), Lebanon would need to run a primary surplus about 17 percentage points of GDP per year higher than the historical average.  
- The significant adjustment requirement for Lebanon is due not only to its high debt level but also to particularly high effective interest rates.

### Caveats and contextual considerations
- The “safe” debt target is uncertain given risks from future interest and growth rates, contingent liabilities, and governments’ ability to adjust after shocks.  
- Multiple primary balance paths can be consistent with sustainability; some paths may be more back-loaded than assumed in the analysis.  
- Countries in the region may be able to finance larger debt ratios than other regions for reasons noted above (dedicated depositor base, official financing), but these factors do not eliminate vulnerabilities.  
- The analysis relies on historical policies as indicators of politically and socially feasible future policies; the second, econometric approach allows short-term deviations but relies on shorter time series and group averages.

*Source: IMF country reports, and staff estimates and calculations.*

### 0.8 percent of GDP for every 10 percentage point of GDP increase in public debt.

### _wp0712 - 0.8 percent of GDP for every 10 percentage point of GDP increase in public debt.

### Key empirical findings on debt dynamics and fiscal reaction
- Public debt dynamics appear mean-reversing: "0.8 percent of GDP for every 10 percentage point of GDP increase in public debt."
- Countries react primarily by adjusting revenues:
  - For every 10 percentage point increase in the debt to GDP ratio, the revenue ratio increases by 0.4 percent.
  - Expenditures react with the correct sign, but the effect is not statistically significant.
- A systematic relationship between the output gap and the primary balance is hard to detect:
  - Neither the output gap, nor the oil price gap turned out statistically significant in the reported specifications.
  - Coefficients of output gap were low and insignificant in both the revenue and expenditure equations and have been dropped from the reported specifications.
- Regression evidence (Table 2, Non-Oil Countries) highlights dynamic relationships (as reported):
  - Debt_1 / GDP_1: 0.083** 0.048*** 0.061*** 0.039*** -0.022
  - Output gap: -0.039 0.111
  - Oil prices gap: -0.022
  - Primary balance_1 / GDP_1: 0.492***
  - Output gap > 0: -0.344
  - Constant: -0.112*** 0.000 -0.096*** 0.189*** 0.289***
  - R2: 0.32 0.19 0.19 0.02
  - N: 104 95 104 104 104
  - Note: * p<.1; ** p<.05; *** p<.01. All specifications include fixed effect, except dynamic baseline estimated using Arellano-Bond method.

### Oil-producing countries: sustainability, POIM, and cyclical responses
- Modeling approach:
  - The non-oil primary balance is modeled as a function of the present value of oil revenue and past net debt, both as ratios to non-oil GDP.
  - Policy is consistent with the POIM if coefficients of the present value of oil revenue and of the net debt are both close to the real interest rate and the constant term is zero.
- Main results (Table 3, Oil-Producing Countries):
  - Policies seem not to be driven by the POIM: the coefficient of the present value of oil revenue has a small and insignificant sign.
  - The link between the non-oil primary balance and net debt is strong: the coefficient of net debt is positive and significantly different from zero.
    - An increase in the net debt ratio of 10 percent of GDP will improve the primary balance by about 0.3 percent of GDP.
  - The effect of cyclical changes in oil prices on the non-oil primary balance is weak but pro-cyclical:
    - A decrease in the oil price gap by 10 percent increases the non-oil primary balance by 0.8 percent of non-oil GDP.
  - Expenditures react more strongly to cyclical changes in oil prices than non-oil revenues, although both responses are weak.
- Regression excerpts (Table 3) as reported:
  - Non-oil rev. / non-oil GDP; Expenditure / non-oil GDP; Debt_1 / non-oil GDP: 0.027
  - Non-oil output gap: -0.125
  - Oil prices gap: -0.079* -0.078** -0.058* -0.162** -0.077** 0.056** 0.152***
  - Assets_1 / non-oil GDP: -0.032***
  - Oil wealth / non-oil GDP: 0.000
  - Net debt_1 / non-oil GDP: 0.031*** 0.033*** 0.030*** 0.029*** 0.022*** -0.010*
  - Non-oil primary balance_1 / non-oil GDP_1: 0.442***
  - Oil price gap > 0: 0.195
  - Net debt in low income countries: 0.021
  - Constant: -0.284*** -0.288*** -0.003 -0.303*** -0.295*** 0.219*** 0.517***
  - r2: 0.221 0.219 0.239 0.225 0.198 0.199
  - N: 108 108 99 108 108 108
  - Note: * p<.1; ** p<.05; *** p<.01. All specifications include fixed effect, except dynamic baseline estimated using Arellano-Bond method.

### Identified vulnerabilities
- Volatility of oil revenue and grants is an important vulnerability, especially since fiscal adjustments in non-oil producing countries rely heavily on revenue mobilization due to expenditure rigidities.
- Weakness in data coverage is a major vulnerability:
  - Narrow coverage and large off-budget operations.
  - Lack of systematic data on contingent liabilities means government liabilities may arise without forewarning.
- Gradual opening to international capital markets could remove the stable regional financing cushion, though near-term liquidity may remain high while oil prices stay high.
- Rising domestic debt, if not managed well, can lead to inflation, low growth, and loss of confidence:
  - Although debt ratios are on a declining trend, rising interest rates and crowding out of private sector investment could stifle growth and perpetuate debt.
  - Debt levels are still relatively high for non-oil producing countries.

### Conclusions and policy implications
- Historical context:
  - Public debt in the Middle East rose during the mid-nineties mainly due to large non-oil primary deficits in oil producing countries; debt has decreased in recent years mainly due to high oil revenue.
  - In oil-producing countries, fiscal policy has been procyclical with respect to both oil prices and growth.
  - In non-oil countries, growth, some primary fiscal adjustment, and debt relief contributed to a decline in debt levels.
- Overall assessment:
  - On average, fiscal policies have reacted to past high indebtedness and avoided default or debt crises.
  - Public debt levels remain uncomfortably high in many economies, particularly in non-oil producing countries and middle income oil producers.
- Policy recommendations and priorities implied by findings:
  - For non-oil producing countries: a substantial adjustment of the primary balance would be needed to bring debt down to a safer level.
  - Increase expenditure flexibility in non-oil producing countries since adjustments rely mainly on increasing revenues.
  - Middle-income oil producing countries may need to adjust policies if they want to align with the permanent oil income model (POIM); specific recommendations should account for uncertainties surrounding the permanent oil benchmark and country-specific differences.
  - Strengthen data coverage, transparency, and reporting of contingent liabilities to reduce vulnerability.
  - Manage rising domestic debt carefully to avoid inflationary pressures, crowding out private investment, and loss of confidence.

### Data and measurement issues (Appendix I summary)
- Dataset: fiscal and public debt data for 19 countries from 1990 to 2005 compiled from IMF staff reports and country economists; some empirical work uses 1999–2004 because of gaps prior to 1999.
- Data availability problems (examples):
  - No data for UAE prior to 1999 because of lack of revenue figures for Abu Dhabi Investment Authority (ADIA).
  - Egypt: data prior to 1999 unavailable due to a structural break.
  - Pakistan: no breakdown of fiscal expenditures related to salaries and wages.
  - Qatar: total financing data for 1990–92 missing.
  - Kuwait: data missing during 1990–93 due to Gulf War disruption.
- Public debt series:
  - Public debt narrowly defined as sum of government’s liabilities available for most countries from 1990; Iran, Egypt and UAE started producing this data in the late 1990s.
  - Detailed debt composition data (currency denomination, maturity, interest type) generally missing.
- Government assets and external assets:
  - Data on government domestic and external assets often unavailable, especially for GCC countries.
  - Stock of portfolio debt securities, bank deposits and other debt instruments from Lane and Milesi-Ferretti (2005) used as proxy where needed.
- Coverage of fiscal flows and public debt varies:
  - For fiscal flows: 2 countries report public sector (Iran and UAE (partially)), 2 report general government (Egypt and Pakistan), remainder central government only.
  - For public debt: 5 report public sector (Egypt, Libya, Oman, Saudi Arabia, and UAE), 3 report general government (Jordan, Pakistan, and Tunisia), remaining 11 central government.
- Contingent liabilities data (pensions, bank restructuring, PPP guarantees) largely unavailable.
- Long-term external debt trends analyzed using WB GDF data, which often has broader coverage and includes public sector. Figure A1 compares External debt (MCD data), External debt (AMF), External debt (WB), and External assets (LMF) vs External assets (BIS) as staff-estimated series.

*Source: _wp0712 - 0.8 percent of GDP for every 10 percentage point of GDP increase in public debt.*

### Appendix II

### Appendix II

### Specification for non-oil producers
- Estimation sample: period 1990–2005; panels are imbalanced due to incomplete country series.
- Baseline empirical specification (following Bohn (1998) and IMF (2003)):
  - p_{i,t} = α_i + ρ b_{i,t-1} + β X_{i,t} + ε_{i,t}
  - where:
    - p is the primary balance (percent of GDP),
    - b is debt stock (percent of GDP),
    - X is a vector of control variables,
    - i and t are country and time subscripts,
    - ρ captures the effect of the public debt stock on the primary balance.
- Interpretation:
  - A positive ρ guarantees fiscal sustainability as explained in the main text.
- Control variables for non-oil producers: output gap and oil-price gap as percentages of GDP.

### Permanent Oil Income Model (POIM) and specification for oil producers
- The POIM is derived by maximizing social welfare that depends on government expenditures other than those financed by non-oil revenue:
  - Max Σ_{t=1}^{∞} β^{t-1} U(pb_t), where pb stands for government non-oil primary balance deflated by non-GDP deflator.
- Government budget constraint:
  - b_{t} = (1 + r) b_{t-1} + pb_t − z_t
  - where b is government net debt (deflated by non-GDP deflator), z is oil revenue (deflated by non-GDP deflator), and r is real interest rate.
- With the transversality condition and assuming (1 + r) = 1/(1 − r)? [text gives condition as (1)1rb+=], the optimal non-oil primary balance expression is derived:
  - pb_t = r w_t − r_t? (expressed in text as t r w p b − = ) — the optimal non-oil primary balance can be expressed as a function of government wealth w.
- Government wealth w is defined as sum of net financial assets and oil wealth:
  - w_t = b_t + Σ_{s=t}^{T} z_s / Π_{j=t}^{s} (1 + r_j) (text expression preserved as in source).
- Caveats on optimal policy:
  - Uncertainty (especially regarding oil wealth) requires higher primary balances if governments are risk averse.
  - Large swings in primary balances due to variability in oil wealth may be suboptimal if consumers exhibit habit formation.
  - Alternative welfare function specifications (e.g., per capita or per non-oil GDP) change intergenerational distribution and may alter optimal policy.
  - Inclusion of other non-financial public sector balance-sheet elements may change sustainability assessment and treatment of consumption vs. investment.

### Empirical specification for oil producers
- Empirical counterpart estimated:
  - p_{i,t} = α_i + ρ b_{i,t-1} + γ w_{i,t-1} + β X_{i,t} + ε_{i,t}
  - where:
    - p is the primary balance,
    - b is debt stock,
    - w is oil wealth measured as percentages of non-oil GDP at time t,
    - X is a vector of control variables,
    - i and t are country and time subscripts,
    - ρ captures the effect of public debt stock on the primary balance,
    - γ captures the effect of oil wealth.
- Consistency with theoretical optimality:
  - Empirical specification is consistent with the derived optimal non-oil primary balance if ρ = r/(1 + r) and γ = r/(1 + r) (text expresses condition as ,1 r r rργ== + ).
  - If the sample mean of X_{i,t} is zero and country fixed effects α_i are zero on average, empirical results align with the theoretical optimal specification.
- Control variables for oil producers: non-oil output gap (relative deviation of real non-oil GDP from HP trend) and oil-price gap (relative deviation of oil price from HP trend).

### Extensions, non-linearities, and robustness tests
- Model extensions tested:
  - Asymmetric effects of positive and negative output gaps on the primary balance.
  - Asymmetric effects of above- and below-trend oil prices.
  - Debt and oil wealth effects in sub-groups (tested only for oil producers due to limited observations).
- Threshold/spline tests:
  - A fiscal response depending on whether public debt exceeds a threshold was tested (spline coefficient selected to maximize goodness of fit), but estimated effects were non-significant.
- Revenue and expenditure functions:
  - Specified analogously to the baseline fiscal policy reaction functions, estimated separately for oil and non-oil producers with revenue and total expenditure as dependent variables.

### Oil wealth estimation (local currency) — key assumptions and methods
- Oil wealth estimated sequentially using information available to policymakers at each year, under main assumptions:
  - Future production:
    - Fixed at the last observable level in energy-equivalent terms.
    - For countries not fully exploiting gas reserves, the future share of gas is assumed to increase while keeping energy content unchanged.
    - Data sources: Energy Information Administration for oil and gas production and oil reserves.
    - Gas transformed to oil equivalent assuming energy content of one barrel of oil = 5,487 cubic feet of natural gas.
  - Future oil prices (in USD):
    - Projected assuming a two-factor model as in Schwartz and Smith (2000).
    - Model re-estimated using 1990–2005 data for spot prices and 12 and 24 month-ahead WTI futures.
    - The model assumes long-run equilibrium prices follow a random walk with drift; short-term deviations follow an Ornstein-Uhlenbeck process.
    - Estimated drift rate is close to zero and used in “conservative” estimates of oil wealth.
    - Optimistic estimates replace the estimated value with the assumed long-term US inflation rate at 2 percent per year.
  - Other macro assumptions:
    - Inflation, real interest rates, and real exchange rates are constant, and PPP holds.
    - Real interest rate and discount rate is 3 percent.
- Figures (descriptive results in text):
  - Figure A2: presents spot and equilibrium prices and 12-month-ahead futures in the sample (log scale; sample period Jan-90 to Jan-05).
  - Figure A3: presents range of oil-wealth estimates (percent of non-oil GDP) showing conservative and optimistic oil wealth for listed oil-producing countries; optimistic series uses 2 percent US inflation assumption.

### Methodological notes and references used in appendix
- Empirical specifications follow Bohn (1998) and IMF (2003); related approaches noted in Kumar et al. (2005) and Abiad and Baig (2005).
- Oil-price dynamics modeled following Schwartz and Smith (2000); oil and gas data from Energy Information Administration.
- Key parameter values preserved from appendix assumptions:
  - Sample period: 1990–2005.
  - Energy conversion: 1 barrel oil = 5,487 cubic feet natural gas.
  - Optimistic oil-wealth inflation assumption: 2 percent per year.
  - Real interest rate and discount rate: 3 percent.

*Source: Appendix II, _wp0712 - Appendix II*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp0712.pdf_