## wp18133

## Source details

**Canonical URL:** [wp18133](https://www.imf.org/-/media/files/publications/wp/2018/wp18133.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2018/wp18133.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2018/wp18133.pdf.json)

---

### I. Introduction — context and scope
- Objective: estimate fiscal multipliers for GCC countries using expanded government spending data capturing developments since mid-2014; illustrate impact of fiscal adjustment on non-oil growth and derive policy implications.
- Countries: Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and UAE.
- Key background magnitudes:
  - During 2000–2014 GCC average oil revenue = 81 percent of total fiscal revenues.
  - Oil prices dropped starting mid-2014, bottoming at $30 per barrel in January 2016.
  - Fiscal positions shifted from an average fiscal surplus of 12.7 percent of non-oil GDP in 2014 to a deficit of 14.5 percent by 2016.
  - Between 2015–16 GCC countries on average consolidated by 13.8 percent.
- Main empirical summary:
  - Panel OLS with fixed effects over different periods indicates fiscal multipliers have declined in recent years.
  - Both capital and current multipliers declined; capital multipliers remain larger than current multipliers.
  - Implication: cuts in capital spending could hurt growth more than cuts in current spending.

### II. Data, structure, and stylized facts
- Data coverage and variables:
  - Annual data for six GCC countries, 1990–2016.
  - Variables: non-oil GDP (national accounts) and central government spending (current and capital). Global GDP growth included as control.
- Structure and averages (2011–16):
  - Oil revenues average 78 percent of government revenues.
  - Oil exports = 62 percent of exports.
  - Oil GDP = 42 percent of GDP.
  - Government spending ≈ 70 percent of non-oil GDP.
  - About 80 percent of government spending = current spending.
- Fiscal impact of oil price drop:
  - Region lost annual fiscal oil revenue of about $240 billion between 2014 and 2015.
  - Compared with 2013, the 2016 overall fiscal balance (as a share of non-oil GDP) deteriorated by 40 percentage points; the non-oil primary balance declined by 15 percentage points.
- Spending and non-oil growth (descriptives and correlations):
  - 1990–2016 average annual non-oil GDP growth = 6.6 percent; total government spending growth = 5.9 percent.
  - For 1990–2016, one percentage point growth in total spending associated with 1.1 percentage points of non-oil GDP growth (ratio GDP / Total spending = 1.1).
  - Ratio GDP / Total spending by subperiods:
    - 1990–2007 = 1.4
    - 2008–16 = 0.6
    - 2008–10 = 0.5
    - 2011–12 = 0.4
    - 2015–16 = ratio turned negative
  - Simple correlations (1990–2016): statistically significant associations between non-oil GDP growth and total, current, and capital spending; contemporaneous spending growth more strongly related than lagged values.
  - Cross-country heterogeneity noted; UAE correlations display opposite expected sign in some periods (1995–2005), possibly due to Government Related Entities (GREs) spending outside central budget.
- Leakages and changing propagation:
  - Imports of goods during 1990–16 averaged about 55 percent of non-oil GDP; during 2008–16 almost 80 percent of non-oil GDP.
  - Remittances in percent of non-oil GDP: about 11 percent (1990–16 average), 13 percent (2008–16).
  - Outbound tourism expenditure in percent of non-oil GDP: about 7 percent during 1995–2015; almost 8 percent during 2008–15.
  - Low efficiency of public investment, supply-side bottlenecks, and implementation constraints may have reduced the impact of spending on non-oil GDP.

### III. Literature context (selected prior estimates)
- Advanced and emerging market evidence:
  - Batini et al. (2014b): short-term spending multipliers: advanced economies 0.6–1.4; emerging markets 0.2–0.5.
- GCC-specific prior studies:
  - Espinoza and Senhadji (2011): long-run multipliers current spending 0.3–0.7; capital spending 0.6–1.1.
  - Cerisola and others (2015): long-run multipliers current spending 0.7–1.3; capital spending 1.4.
  - Country studies for Saudi Arabia find positive short-run effects mainly from current expenditures.

### IV. Methodology for re-estimating multipliers
- Two models estimated (OLS panel fixed effects, 1990–2016):
  - Linear model: regressions of real non-oil GDP (Y) on real government expenditures (G) with contemporaneous, first, and second lags; variables in first log difference; control for global GDP growth (Ggdp); country fixed effects φi.
  - Non-linear model: adds dummy D (value 1 for 2008–16 or 2011–16, 0 otherwise) and interactions between G and D to capture time variation.
- Endogeneity discussion:
  - Automatic stabilizers and endogenous fiscal policy may bias estimates downward; authors argue channels can be largely ruled out for GCC because spending driven by oil revenue (exogenous international oil prices) and fiscal policy is procyclical.
- Implied multiplier computation:
  - Short-term multiplier = contemporaneous impact (β_GY).
  - Long-term multiplier = cumulative impact over three years (contemporaneous + lags).

### V. Re-estimated multipliers — empirical results
- Regression findings:
  - Global GDP growth positively and significantly related with non-oil activity.
  - Linear model (full sample) shows persistence of capital spending impact (contemporaneous, first, second lags significant).
  - Current spending impact is shorter (contemporaneous and first lag significant).
  - Non-linear models:
    - Interaction coefficients not statistically different from zero for 2008–16 in some specifications.
    - Restricting sample to 2011–16: both current and capital spending positively and significantly correlated with non-oil growth but with lower cumulative impacts than full sample.
- Implied multipliers (Table 7) — preserve values exactly:
  - Full sample (1990 - 2016):
    - Current spending: Short-term 0.2; Long-term 0.4
    - Capital spending: Short-term 0.4; Long-term 1.3
  - Recent period (2011 - 2016):
    - Current spending: Short-term ...... (elasticities not statistically significant); Long-term 0.2
    - Capital spending: Short-term 0.4; Long-term 0.9
  - Note: "......" indicates elasticities not statistically significant.
- Interpretation:
  - Multipliers are lower in the more recent period compared to full sample.
  - Impact of government spending on non-oil growth weakened in recent years.
- Robust selected multiplier values from robustness checks (Table 8 examples):
  - Capital spending short-term 0.3; long-term 1.5 (column 1).
  - Capital spending short-term 0.3; long-term 0.9 (column 2).
  - Capital spending short-term 0.4; long-term 1.4 (column 3).
  - Capital spending short-term 0.4; long-term 0.8 (column 4).
  - Current spending short-term 0.2; long-term 0.4 (columns 5–8 variations).

### VI. Simulations and scenario outcomes
- Simulation setup:
  - Fiscal consolidation measured by annual changes in government expenditures in percentage of non-oil GDP.
  - Government spending shocks at time t affect activity over three periods (t, t+1, t+2).
  - Timing assumptions:
    - Current spending shocks: stronger short-term impact (first year).
    - Capital spending shocks: spread progressively, peak in third year.
  - Baseline: IMF (2014) projections of non-oil GDP growth.
- Simulations using existing multipliers (Cerisola et al. (2015) and Espinoza and Senhadji (2011)):
  - Using Cerisola et al. (2015) multipliers: non-oil growth falls into negative territory between 2015 and 2017; non-oil GDP growth could have fallen as low as -3 percent in 2016.
  - Using lowest multipliers from Espinoza and Senhadji (2011): predicted non-oil growth would be 0.7 percent in 2016 (compared to baseline 5.7 percent and an outcome of 1.6 percent).
- Simulations using updated (lower) multipliers from this paper:
  - Based on actual composition of fiscal consolidation, predicted trough of non-oil GDP growth in 2017 at 2.7 percent (or 4 percent when using multipliers for 2011–16), compared to baseline 5.7 percent.
  - If adjustment falls solely on current expenditure (2011–16 multipliers), non-oil growth would be at maximum 1.2 percentage points lower in 2017 (non-oil GDP growth would be 4.5).
  - If adjustment fully translates into cuts in capital spending, non-oil growth could be as low as -1 percent in 2017.
- Scenario implication:
  - Capital spending cuts imply a stronger adverse effect on non-oil growth than current spending cuts.

### VII. Robustness checks and controls for leakages
- Exercise 1: include inflation and oil prices (Table 8):
  - Neither inflation nor oil prices have statistically significant impact on non-oil growth.
  - Results consistent: fiscal multipliers much lower in recent years; current spending lower impact than investment.
- Exercise 2: control for leakages (imports, remittances, tourism) (Table 9):
  - Leakages likely positively correlated with government spending but negatively correlated with non-oil GDP growth, reducing observed impact of public expenditure.
  - Including leakages tends to increase estimated elasticities of government spending.
  - Estimated coefficients on leakages (percentage of non-oil GDP) reported (not statistically significant):
    - Imports: -0.005 and 0.008
    - Remittances: -0.022 and -0.026
    - Tourism expenditures: -0.052 and -0.046
  - Conclusion: controlling for leakages yields stronger cumulative impacts of capital and current spending.

### VIII. Conclusions and policy implications
- Two main findings:
  - Fiscal multipliers in the GCC have declined in recent years.
  - Capital multipliers remain larger than current multipliers, although both have declined.
- Policy recommendations:
  - Ongoing fiscal adjustment may be less costly than prior estimates suggested owing to lower multipliers.
  - Reducing (less productive) current spending can help limit adverse effects on economic activity during consolidation.
  - Given the larger and more persistent impact of capital spending on non-oil growth, (efficient) investment expenditures need to be protected.
  - Policies to improve local content, public investment efficiency, and absorptive capacity can raise the effectiveness of fiscal spending by reducing leakages (imports, remittances, tourism outflows).

*Italic: Source — IMF staff estimations and synthesis from wp18133 (PDF).*

### REFERENCES ___________________________________________________________________________________ 26

### wp18133 - REFERENCES ___________________________________________________________________________________ 26

### Figures and Tables (listed)
- Figures
  - 1. Real Government Spending and Non-oil GDP Growth in the GCC
  - 2. Total Spending Growth and Non-oil GDP Growth, 1990–2016
  - 3. GCC: Impact of Fiscal Adjustment on Economic Growth, Based on Existing Fiscal Multipliers
  - 4. GCC: Impact of Fiscal Adjustment on Economic Growth, Based on New Fiscal Multipliers
- Tables
  - 1. Structure of GCC Economies, Average 2011–16
  - 2. Simple Correlations: Government Spending Growth and Non-oil GDP Growth, 1990–2016
  - 3. Spending and Non-oil GDP, Growth Rates
  - 4. Simple Correlations: Government Spending Growth and Non-oil GDP Growth, 2008–16
  - 5. Existing Fiscal Multipliers for the GCC Countries
  - 6. Estimate Results: Government Spending and Economic Activity
  - 7. Implied Fiscal Multipliers
  - 8. Robustness to Inclusion of Additional Controls
  - 9. Controlling for the Leakages
- Annex
  - I. Empirical Regularities for Spending and Non-oil GDP Growth

### I. Introduction — context and scope
- Paper objective: estimate fiscal multipliers for GCC countries using expanded government spending data capturing developments since mid-2014; illustrate impact of fiscal adjustment on non-oil growth and derive policy implications.
- Countries included: Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and UAE (terms “GCC” and “Gulf” used interchangeably).
- Key background facts and magnitudes:
  - During 2000–2014 (commodity boom era) GCC average oil revenue was 81 percent of total fiscal revenues.
  - Oil prices dropped starting mid-2014, bottoming at $30 per barrel in January 2016.
  - Fiscal positions shifted from an average fiscal surplus of 12.7 percent of non-oil GDP in 2014 to a deficit of 14.5 percent by 2016.
  - Between 2015–16 GCC countries on average consolidated by 13.8 percent.
- Main empirical result summary:
  - Panel OLS with fixed effects over different periods indicates fiscal multipliers have declined in recent years.
  - Both capital and current multipliers declined; capital multipliers remain larger than current multipliers.
  - Implication: cuts in capital spending could hurt growth more than cuts in current spending.

### II. Literature — prior estimates and themes
- Literature status:
  - Extensive empirical evidence for advanced economies; limited consensus on estimation methods and multiplier sizes.
  - Emerging market evidence more limited; Batini et al. (2014b) finds short-term spending multipliers: advanced economies 0.6–1.4; emerging markets 0.2–0.5.
  - Studies document non-linearities, exchange rate interactions, cyclical state dependence, and time variation in multipliers.
- GCC-specific findings from prior studies:
  - Espinoza and Senhadji (2011): long-run multipliers between 0.3–0.7 for current spending and 0.6–1.1 for capital spending (panel linear regression).
  - Cerisola and others (2015): long-run multipliers between 0.7–1.3 for current spending and 1.4 for capital spending (VAR with sign restrictions).
  - Country-specific studies: for Saudi Arabia, Joharji and Starr (2010) and Alshahrani & Alsadiq (2014) find total expenditures have a positive short-run effect mainly originating from current expenditures.

### III. Data and preliminary discussion — dataset and stylized facts
- Data coverage:
  - Annual data for six GCC countries during 1990–2016.
  - Variables: non-oil GDP (national accounts) and central government spending (current and capital). Global GDP growth included as control in econometrics.
- Structure and averages (2011–16):
  - Oil revenues average 78 percent of government revenues.
  - Oil exports = 62 percent of exports.
  - Oil GDP = 42 percent of GDP.
  - Government spending = about 70 percent of non-oil GDP.
  - About 80 percent of government spending corresponds to current spending.
- Fiscal impact of oil price drop:
  - Region lost annual fiscal oil revenue of about $240 billion between 2014 and 2015.
  - Compared with 2013, the 2016 overall fiscal balance (as a share of non-oil GDP) deteriorated by 40 percentage points, while the non-oil primary balance declined by 15 percentage points.
- Spending and non-oil growth: descriptive statistics and correlations
  - For 1990–2016, average annual non-oil GDP growth = 6.6 percent; total government spending growth = 5.9 percent.
  - For 1990–2016, one percentage point growth in total spending associated with 1.1 percentage points of non-oil GDP growth (ratio GDP / Total spending = 1.1).
  - Ratio of GDP / Total spending fell from 1.4 during 1990–2007 to 0.6 during 2008–16; for 2008–10 and 2011–12 ratios were 0.5 and 0.4 respectively; ratio turned negative during 2015–16.
  - Simple correlations (1990–2016) show statistically significant associations between non-oil GDP growth and total, current, and capital spending; contemporaneous spending growth more strongly related than lagged values.
  - Correlations mask heterogeneity across countries: Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, UAE show distinct patterns; UAE correlations display opposite expected sign in some periods (1995–2005), possibly due to Government Related Entities (GREs) fiscal spending outside central budget.
- Leakages and changing propagation:
  - Imports of goods during 1990–16 averaged about 55 percent of non-oil GDP; during 2008–16 almost 80 percent of non-oil GDP.
  - Remittances in percent of non-oil GDP: about 11 percent (1990–16 average), 13 percent (2008–16).
  - Outbound tourism expenditure in percent of non-oil GDP increased from about 7 percent during 1995–2015 to almost 8 percent during 2008–15.
  - Low efficiency of public investment, supply-side bottlenecks, and implementation constraints may have reduced the impact of spending on non-oil GDP.

### IV. GCC fiscal multipliers — comparison and projections
- Existing multiplier estimates (summary of cited literature, Table 5 referenced):
  - Espinoza and Senhadji (2011): long-term current spending multipliers 0.3–0.7; long-term capital multipliers 0.6–1.1.
  - Cerisola and others (2015): long-run multipliers 0.7–1.3 for current spending and 1.4 for capital spending.
- Implication for growth projections (based on existing multipliers):
  - Prior estimates imply large negative impacts of fiscal consolidation on non-oil growth.
- New empirical findings (from this paper):
  - Estimated multipliers have declined in recent years relative to prior studies.
  - Capital multipliers remain larger than current multipliers despite declines.
  - Therefore, fiscal consolidation may be less costly for non-oil growth than previously suggested by existing multiplier estimates, but cuts to capital spending still pose larger growth risks.
- Policy-relevant inferences:
  - Because capital multipliers exceed current multipliers, priority should be given to protecting growth-enhancing capital spending where possible.
  - Increasing leakages (imports, remittances, outbound tourism) reduce domestic multipliers, implying policies to improve local content, public investment efficiency, and absorptive capacity can raise the effectiveness of fiscal spending.
  - Ongoing and planned fiscal consolidation since 2014 has contained overall fiscal deterioration; further consolidation is planned over the medium term and should account for the evolving, lower multiplier environment.

_Italic: Source — IMF staff estimations and synthesis from wp18133 REFERENCES (PDF). _

### 1.4 for current and capital spending, respectively.

### wp18133 - 1.4 for current and capital spending, respectively.

### Simulations: Impact of Ongoing Fiscal Consolidation on Non-oil Growth
- Fiscal consolidation measured by annual changes in government expenditures in percentage of non-oil GDP; government spending shocks at time t affect activity over three periods (t, t+1, and t+2).
- Assumptions on timing of effects:
  - Current spending shocks: stronger short-term impact (first year).
  - Capital spending shocks: spread out progressively, peak impact in the third year.
- Baseline for comparison: projections of non-oil GDP growth in IMF (2014).
- Simulations using existing multipliers (Table 5; Cerisola et al. (2015) and Espinoza and Senhadji (2011)) indicate:
  - Using Cerisola et al. (2015) multipliers, non-oil growth falls into negative territory between 2015 and 2017; non-oil GDP growth could have fallen as low as -3 percent in 2016.
  - Using lowest multipliers from Espinoza and Senhadji (2011), predicted non-oil growth would be 0.7 percent in 2016 (compared to baseline 5.7 percent and an outcome of 1.6 percent).
- Scenario analysis isolating spending types:
  - Capital spending cuts imply a stronger adverse effect on non-oil growth than current spending cuts.
  - Current spending adjustment would be relatively less costly compared to capital spending cuts, according to the two studies’ multipliers.

### Re-estimating the Multipliers: Methodology
- Empirical approach draws on Espinoza and Senhadji (2011); two models estimated:
  - Linear model (equation (1)): regressions of real non-oil GDP (Y) on real government expenditures (G) with contemporaneous, first, and second lags; variables in first log difference; control for global GDP growth (Ggdp); country fixed effects φi.
  - Non-linear model (equation (2)): includes dummy D (value 1 for period 2008–16 or 2011–16, and 0 otherwise) and interaction terms between G and D to capture potential changes over time.
- Estimation method: OLS panel fixed effects for period 1990–2016.
- Endogeneity concerns discussed: automatic stabilizers and endogenous fiscal policy may bias estimates downward; authors argue channels can be largely ruled out for GCC given spending driven by oil revenue (exogenous international oil prices) and procyclical fiscal policy evidence.
- Implied fiscal multipliers computed as X = [β_GY]. Short-term multiplier = contemporaneous impact; long-term multiplier = cumulative impact over three years.

### Re-estimating the Multipliers: Results (Table 6 and Table 7)
- Regression findings:
  - Global GDP growth is positively and significantly related with non-oil activity.
  - Linear model (full sample) shows persistence of capital spending impact: contemporaneous, first, and second lags statistically significant.
  - Current spending impact shorter: contemporaneous and first lag statistically significant.
  - Non-linear models:
    - Columns (2) and (5): interaction coefficients not statistically different from zero for period 2008–16.
    - Restricting sample to 2011–16 (columns 3 and 6): both current and capital spending positively and significantly correlated with non-oil growth but with lower cumulative impacts than full sample.
- Implied multipliers (Table 7):
  - Full sample (1990 - 2016):
    - Current spending: Short-term 0.2; Long-term 0.4
    - Capital spending: Short-term 0.4; Long-term 1.3
  - Recent period (2011 - 2016):
    - Current spending: Short-term ...... (elasticities not statistically significant); Long-term 0.2
    - Capital spending: Short-term 0.4; Long-term 0.9
  - Note: "..." indicates cases where elasticities are not statistically significant.
- Interpretation:
  - Multipliers are lower in the more recent period compared to full sample.
  - Impact of government spending on non-oil growth has weakened in recent years.
- Simulations using updated (lower) multipliers (Figure 4):
  - Based on actual composition of fiscal consolidation, simulations predict trough of non-oil GDP growth in 2017 at 2.7 percent (or 4 percent when using multipliers for 2011–16), compared to baseline 5.7 percent.
  - If adjustment falls solely on current expenditure (2011–16 multipliers), non-oil growth would be at maximum 1.2 percentage points lower in 2017 (non-oil GDP growth would be 4.5).
  - If adjustment fully translates into cuts in capital spending, non-oil growth could be as low as -1 percent in 2017.

### Robustness Checks
- Exercise 1: Include inflation and oil prices as additional controls (Table 8):
  - Neither inflation nor oil prices found to have a statistically significant impact on non-oil growth.
  - Results consistent with main findings:
    - Fiscal multipliers much lower in the most recent years.
    - Current spending has a much lower impact on economic activity compared to investment expenditures.
  - Selected multiplier values from robustness checks (Table 8):
    - Capital spending short-term 0.3; long-term 1.5 (column 1).
    - Capital spending short-term 0.3; long-term 0.9 (column 2).
    - Capital spending short-term 0.4; long-term 1.4 (column 3).
    - Capital spending short-term 0.4; long-term 0.8 (column 4).
    - Current spending short-term 0.2; long-term 0.4 (columns 5–8 variations, with some "..." where short-term insignificance).
- Exercise 2: Control for leakages (imports of goods and services, remittances outflows, tourism expenditures) (Table 9):
  - These leakages are likely positively correlated with government spending but negatively correlated with non-oil GDP growth, implying they reduce the observed impact of public expenditure on non-oil growth.
  - Including these controls tends to increase estimated elasticities of government spending.
  - Estimated coefficients on leakages (expressed in percentage of non-oil GDP):
    - Imports: coefficients -0.005 and 0.008 (not statistically significant).
    - Remittances: coefficients -0.022 and -0.026 (not statistically significant).
    - Tourism expenditures: coefficients -0.052 and -0.046 (not statistically significant).
  - Overall, controlling for leakages yields stronger cumulative impacts of capital and current spending, consistent with the interpretation that leakages weaken the relationship between government expenditure and non-oil activity.

### Conclusions and Policy Implications
- Two main findings:
  - Fiscal multipliers in the GCC have declined in recent years.
  - Capital multipliers remain larger than current multipliers, although both have declined.
- Policy implications:
  - Ongoing fiscal adjustment may be less costly than prior estimates suggested due to lower multipliers.
  - Reducing (less productive) current spending can help limit adverse effects on economic activity during consolidation.
  - Given the larger and more persistent impact of capital spending on non-oil growth, (efficient) investment expenditures need to be protected.

*Source: wp18133 - 1.4 for current and capital spending, respectively.*

### Annex I. Empirical Regularities for Spending and Non-oil GDP Growth

### Annex I. Empirical Regularities for Spending and Non-oil GDP Growth

### Overview
- Coverage: Spending Growth and Non-oil GDP Growth, 1990–2016.
- Source attribution on figures: IMF staff estimations.

### Figures and country panels (1990–2016)
- Figure 1A — Panels for: Bahrain, Kuwait, Oman, Qatar.
  - Common horizontal axis label present: "Non-oil GDP growth".
  - Vertical axis labels / tick values as shown in panels:
    - General total spending growth axes include tick values: -20, 0, 20, 40 (and variations such as -10, 0, 10, 20).
    - Capital spending growth axes include tick values: -50, 0, 50, 100 (and variations such as -500, 50, 100, 150 in some panels).
    - Current spending growth axes include tick values: -20, 0, 20, 40 (and variations such as -40, -20, 0, 20, 40).
  - Country-specific axis ranges and ticks visible:
    - Bahrain: total spending growth tick range includes -20 to 40; capital spending growth includes -200, 20, 40, 60 (as displayed); current spending growth includes -20 to 40.
    - Kuwait: total spending growth tick range includes -5, 0, 5, 10, 15; capital spending growth tick range includes -5, 0, 5, 10, 15; current spending growth tick range includes -20, 0, 20, 40.
    - Oman: non-oil GDP growth axis shows -10, 0, 10, 20, 30, 40; total spending growth axis shows -20, -10, 0, 10, 20, 30; capital spending growth axis shows 0, 50, 100; current spending growth axis shows -40, -20, 0, 20, 40.
    - Qatar: non-oil GDP growth axis shows -10, 0, 10, 20; total spending growth axis shows -200, 20, 40, 60 (as displayed); capital spending growth axis shows -500, 50, 100, 150 (as displayed); current spending growth axis shows -500, 50, 100.

- Figure 1B — Panels for: Oman, Qatar, Saudi Arabia, UAE.
  - Common horizontal axis label: "Non-oil GDP growth".
  - Vertical axis labels / tick values as shown:
    - Total spending growth axes include tick values: -10, 0, 10, 20, 30 (and -200, 20, 40 for some panels).
    - Capital spending growth axes include tick values: -10, 0, 10, 20, 30 (and -200, 20, 40, 60 for some panels; also -40, -20, 0, 20, 40, 60 for others).
    - Current spending growth axes include tick values: -20, 0, 20, 40 (and -40, -20, 0, 20, 40).
  - Saudi Arabia panels include non-oil GDP growth ticks: 0, 2, 4, 6, 8, 10 in some subplots; capital spending growth shows wide range ticks: -40, -20, 0, 20, 40, 60 in one panel.
  - UAE panels include non-oil GDP growth ticks: -10, 0, 10, 20, 30; capital spending growth ticks include -200, 20, 40, 60; current spending growth ticks include -20, -10, 0, 10, 20, 30.

- Figure 1C — Panels for: GCC aggregate, Saudi Arabia, UAE, and GCC (additional paneling).
  - Horizontal axis: "Non-oil GDP growth".
  - Vertical axis ticks and ranges displayed:
    - GCC total spending growth: -10, 0, 10, 20, 30, 40 (and for another panel 0, 2, 4, 6, 8, 10).
    - GCC capital spending growth: -20, 0, 20, 40, 60 (and -50, 0, 50, 100, 150 in another panel).
    - GCC current spending growth: -50, 0, 50, 100 (and -10, 0, 10, 20, 30 in another panel).
    - Saudi Arabia panels: non-oil GDP growth ticks include 0, 2, 4, 6, 8, 10; total spending growth ticks include -10, 0, 10, 20, 30; capital spending growth ticks include -40, -20, 0, 20, 40, 60.
    - UAE panels: non-oil GDP growth ticks include -10, 0, 10, 20, 30; current spending growth ticks include -20, -10, 0, 10, 20, 30.

- Figure 1D — Panels for: GCC (aggregate) with repeated presentation of spending vs non-oil GDP growth.
  - Horizontal axis: "Non-oil GDP growth" with tick values including -10, 0, 10, 20, 30, 40.
  - Vertical axis ticks include:
    - Total spending growth: -20, 0, 20, 40, 60 (in one panel).
    - Capital spending growth: -50, 0, 50, 100, 150 (in one panel).
    - Current spending growth: -50, 0, 50, 100 (in one panel).

### Empirical regularities visible (as presented in figures)
- The figures plot observations of spending growth (total, capital, current) against non-oil GDP growth for each country and for GCC aggregates over 1990–2016.
- Axis tick values indicate that:
  - Non-oil GDP growth observations span negative and positive values in all country panels, with non-oil GDP growth axis ticks shown at values such as -10, 0, 10, 20, 30, 40 and also at finer ticks 0, 2, 4, 6, 8, 10 in some panels.
  - Total spending growth, capital spending growth, and current spending growth observations exhibit wide ranges across countries, with capital spending growth axes in some panels extending to large positive values (e.g., 50, 100, 150) and large negative/large-range ticks present in some panels (e.g., -500, -200).
- All figures label the data source as "IMF staff estimations."

*Source: wp18133 - Annex I. Empirical Regularities for Spending and Non-oil GDP Growth*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2018/wp18133.pdf_