## _wp1094 - References

## Source details

**Canonical URL:** [_wp1094 - References](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1094.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1094.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1094.pdf.json)

---

### Introduction — purpose and key empirical findings
- Purpose: investigate the behavior of expenditure policy during boom-bust commodity price cycles and its implication for real effective exchange rate (REER) movements in resource-rich (oil-producing) countries.
- Sample and scope:
  - Panel of 32 oil-producing countries over the period 1992 to 2009.
  - Government spending composition and country-specific oil price data from Villafuerte et al. (2009).
- Main empirical results:
  - Change in current spending has a bigger impact on the change in REER compared to capital spending.
  - Current spending is downwardly sticky, but increases in boom time; conversely capital spending behaves differently (more responsive to busts).
  - Limited evidence that fiscal rules help reduce responsiveness of current spending during booms; fiscal rules are associated with a significant reduction in capital expenditure during busts while responsiveness to boosts is more muted.

### Theoretical model — Dutch disease with fiscal stickiness
- Objective: illustrate how asymmetrical fiscal responses to revenue shocks (particularly downward stickiness in current spending) affect REER dynamics.
- Key modeling assumptions and notation (verbatim where provided):
  - Small open economy with two goods: tradable goods and a non-tradable government-provided service.
  - Government chooses current expenditure (non-tradable content) and capital expenditure (tradable goods).
  - Non-tradable government-provided services are subject to convex (quadratic) costs of adjustment with potential directional bias (downward stickiness).
  - REER is the shadow price of non-tradable government services relative to world price of tradables.
  - Stickiness modeled via quadratic costs of adjustment; parameter ω scales adjustment costs and parameter ν introduces persistent bias towards positive growth in government current expenditure when ν>0.
- Social planner problem (as presented, notation preserved exactly):
  - max ∫_0^∞ [C_t^{1−1/σ} + ψ G_t^{1−1/σ}]^{1−1/σ} exp(−ρt) dt subject to intertemporal budget constraint including ·F = rF_t + p_t G_t + ω(·G − ν)^2/2 + C_t − T_t − E_t N_t.
  - Notation preserved exactly: C and G are government consumption of tradable and non-tradable goods; ρ is discount rate; σ is elasticity of substitution; F is stock of debt; r is interest rate; p is relative price of non-tradable goods (i.e. the real exchange rate); T is production of tradables; N_t is value of natural resource wealth at time t.

### Model simulations — parameterization and qualitative outcomes
- Simulation parameter values (exact as in source):
  - α=0.6, β=0.5, ψ=0.5, r=ρ=0.04, ω=0.5, σ=2.
- Present value of resource expressed as:
  - E_t N^p_t = E_t ∫_t^∞ P_t N_t exp(−ρt) dt.
- Normalization and initial conditions:
  - ∫_t^∞ N_t exp(−ρt) dt = 10 (equal to 10 years of full employment in one sector).
  - Start at P^∗ = 1 with the economy at steady state.
  - Government started with zero net foreign debt (assumed without loss of generality).
- Shock specification and cases simulated:
  - Permanent shock in permanent price: price jumps from P^∗ to P^∗∗ at t = 0 with no uncertainty.
  - Increase in present value of resource: N^p_t = (P^∗∗ − P^∗) ∫_t^∞ N_t exp(−ρt) dt = ε_p ∫_t^∞ N_t exp(−ρt) dt.
  - Cases simulated for ν = 0 (no asymmetry) and ν = 0.2 (with asymmetry).
  - Positive (negative) resource price shocks considered with ε_p = 0.5(−0.5).
- Simulation outcomes (qualitative):
  - With asymmetry (ν > 0), steady state values for non-tradable spending and the REER are elevated relative to the no-asymmetry case.
  - Under asymmetry, the decline in spending on non-tradables and the REER following a resource price bust is smaller.
  - Capital (tradable) spending is more sensitive to revenue shocks; asymmetry shifts adjustment weight toward capital (tradable) goods leaving current (non-tradable) expenditure less affected.
  - When ν>0 and adjustments are smoother (larger ω), government current spending can permanently crowd out capital expenditure, leaving C lower in steady state.

### Data and empirical strategy — variables and identification
- Data sources and variable construction:
  - Government spending composition and country-specific oil price from Villafuerte et al. (2009) for 32 oil-producing countries, 1992–2009.
  - Oil export unit value: constructed using the international crude oil price interacted with a country-specific discount factor capturing crude quality.
  - Decomposition of oil export unit value into:
    - (i) international crude oil price (common across countries),
    - (ii) country-specific discount factor (captures quality).
  - Identification relies on interaction between international oil price and country-specific discount factor because common year fixed effects are used.
  - Government current and capital spending taken from Villafuerte et al. (2009); current spending proxied as having higher non-tradable content; capital spending proxied as having higher import/tradable content.
- Empirical strategy elements:
  - Distinguish booms and busts using a three-way price decomposition drawn from the energy demand literature to capture asymmetrical responses to price increases and decreases.
  - Empirical specification and estimators structured to isolate boom-bust cycles' effects on spending and on REER.

### Price decomposition — three-way decomposition of oil export unit value
- Decomposition (exact notation preserved):
  - oev_{i,t} = oevmax_{i,t} + oevmin_{i,t} + oevrec_{i,t}
- Component definitions (verbatim):
  - oevmax_{i,t}: cumulative increases in maximum historical oil unit export value for country i at time t; it is monotonically non-decreasing that is oevmax_{i,t} > 0.
  - oevmin_{i,t}: cumulative decreases in oil unit export value; it is monotonically non-increasing that is oevmin_{i,t} ≤ 0.
  - oevrec_{i,t}: cumulative sub-maximum increases in oil export value; it is monotonically non-decreasing that is oevrec_{i,t} > 0.
- Purpose: account for asymmetrical response of REER to oil price boom-bust cycles by separating maximum historical increases, cumulative decreases, and recoveries (sub-maximum increases).

### Specifications and Estimators (econometric model)
- Estimated model (as presented):
  - ∆ln(spending)it = αi + γt + βmax ∆ln(oevmax)it + βmin ∆ln(oevmin)it + βrec ∆ln(oevrec)it + γXit + uit
- Key estimation features:
  - αi are country fixed effects; γt are year fixed effects.
  - βj for j ∈ {max, min, rec} measure marginal effects of country-specific changes in logarithm of oil price components on country-specific changes in spending.
  - Controls Xit include first difference in logarithm of GDP, growthppp; and lagged difference in logarithm spending ∆ln(spending)t−1 as robustness checks.
  - Estimators: least squares and system-GMM (Blundell and Bond, 1998) to address dynamic panel bias with fixed effects.
  - Error term clustered at the country level; may be arbitrarily serially correlated within countries.
- Sign expectations:
  - βmax > 0, βmin > 0 (noting oevmin < 0), βrec > 0.
  - Expected ordering in absolute values: βrec < βmin < βmax.
  - Expectation that current spending responds less to new lows than capital spending.

### IV. Estimation Results — key empirical findings and numeric estimates
A. Composition of Spending and the Real Exchange Rate
- Point estimates (Table 1, column (1)):
  - D.lncepus (current spending): 0.242 (statistically significant at the 1 percent level).
  - D.lncaxus (capital spending): 0.055 (statistically significant at the 5 percent level).
- Interpretation in standardized terms:
  - A 1 standard deviation increase in current spending → increase in REER by about 0.33 standard deviations.
  - A 1 standard deviation increase in capital spending → increase in REER by about 0.13 standard deviations.
- Robustness and dynamics:
  - Current spending–REER link remains statistically significant when controlling for lagged current spending and within-country variation in income per capita.
  - Impact of capital spending on REER is much smaller; becomes statistically insignificant when controlling for lagged spending.
  - Dynamic specifications: lagged dependent variable enters as statistically significantly positive; implied persistence: an impact of a shock to REER at time t would take about 0.9 years to dissipate by one-half for oil exporters.
- Sensitivity:
  - Grubbs test applied; dropping outliers yielded statistically significant point estimates on current spending that were quantitatively larger than those reported in Table 1 (results not shown).
  - Authors can reject equality of current and capital expenditure coefficients at least at the 95 percent confidence level.
- Selected numeric table entries (Table 1, selected columns):
  - D2.lnereer: 0.471 ∗∗∗ (0.0206) in column (1); 0.461 ∗∗∗ (0.0196) in column (2).
  - D.lncepus: 0.242 ∗∗∗ (0.0703) in column (1); 0.278 ∗∗∗ (0.100) in column (2); 0.237 ∗∗∗ (0.0701) in column (3); 0.0909 ∗ (0.0527) in column (4); 0.0837 ∗ (0.0440) in column (5).
  - D.lncaxus: 0.0545 ∗∗ (0.0221) in column (1); 0.00464 (0.0308) in column (2); 0.0550 ∗∗ (0.0223) in column (3); 0.0212 (0.0164) in column (4); 0.0243 ∗ (0.0202) in column (5).
  - Observations: 449, 420, 421, 424, 424. Number of countries: 31.

B. Boom and Bust Cycle and Spending Responses
- Current spending (Table 2):
  - Point estimates on changes in new highs in oil price range from 0.277 (column (3)) to 0.376 (column (2)); statistically significant at the 10 percent and 5 percent levels respectively.
  - Other two oil price components have expected relative values but are not statistically significant.
  - Implication: current spending increases significantly when oil price reaches new highs but does not significantly decrease when oil price reaches new lows.
- Capital spending (Table 3):
  - Point estimates on change in new lows in oil price range from 0.840 (column (3)) to 1.507 (column (1)).
  - Other two components not statistically significant.
  - Implication: capital spending decreases when oil price reaches new lows, but does not significantly increase when oil price reaches new highs.
- Selected numeric table entries:
  - Table 2 (dependent variable D.lncepusppp): D2.lncepusppp: 0.465 ∗∗∗ (0.00877). D.lnoevmax: 0.338 ∗ (0.169) in column (1); 0.376 ∗∗ (0.176) in column (2); 0.277 ∗ (0.152) in column (3). Observations: 388, 388, 366. Number of countries: 31.
  - Table 3 (dependent variable D.lncaxus): D2.lncaxus: 0.430 ∗∗∗ (0.0163). D.lnoevmin: 1.306 ∗∗∗ (0.378) in column (1); 1.507 ∗∗∗ (0.427) in column (2); 0.840 ∗∗∗ (0.310) in column (3). growthppp: 1.677 ∗∗∗ (0.451). Observations: 452, 394, 430. Number of countries: 31.

C. Fiscal Rules and Government Spending
- Fiscal rules definition and sample split:
  - Dummy from Ossowski et al. (2007): value 1 if any fiscal rules are in place and 0 otherwise.
  - Average of dummy over sample period used to perform sample split with cut-off 0.5.
- Main findings (Tables 4–7):
  - Limited evidence that fiscal rules reduce increase in current spending during booms:
    - Coefficient for oil price new highs is positive and statistically significant for countries with no fiscal rules but not statistically significant for countries with fiscal rules (comparison of columns in Tables 4 and 5).
  - Fiscal rules and capital spending during busts:
    - Countries with fiscal rules significantly reduce capital spending during busts; coefficients on new lows are systematically significant at the 1 percent significance level (Table 7).
    - For countries without fiscal rules, capital spending reduction during busts is found but results are less robust.
  - Formal testing:
    - Generalized Chow test allowing for arbitrary within-country serial correlation indicates inability to reject that coefficients associated with oevmin in samples with and without fiscal rules are the same (tables 6 and 7).
- Selected numeric table entries by grouping:
  - Table 4 (No Fiscal Rules) — D.lncepusppp: D2.lncepusppp: 0.464 ∗∗∗ (0.00908). Observations: 189, 189, 179. Number of countries: 15.
  - Table 5 (With Fiscal Rules) — D.lncepusppp: D2.lncepusppp: 0.490 ∗∗∗ (0.00741). Observations: 192, 192, 180. Number of countries: 15.
  - Table 6 (No Fiscal Rules) — D.lncaxus: D2.lncaxus: 0.473 ∗∗∗ (0.0163). growthppp: 1.544 ∗∗ (0.604). Observations: 209, 193, 199. Number of countries: 15.
  - Table 7 (With Fiscal Rules) — D.lncaxus: D2.lncaxus: 0.436 ∗∗∗ (0.0278). D.lnoevmin: 1.112 ∗∗∗ (0.225) in column (1); 1.890 ∗∗∗ (0.578) in column (2); 1.018 ∗∗∗ (0.194) in column (3). growthppp: 2.078 ∗∗ (0.767). Observations: 235, 194, 223. Number of countries: 15.

D. Endogeneity Issues
- Potential simultaneity bias:
  - Appreciation in REER changes relative prices between tradables and non-tradables; substitution vs. income effect can bias least squares estimate on current spending downward.
- Instrumental variables strategy:
  - Use oevmax to instrument current spending; use oevmin to instrument capital spending.
  - Second-stage results do not provide evidence of a causal impact of spending on REER.
  - Various F-tests indicate instruments used in first stage are weak, casting doubt on validity of IV estimates.
- Note on GMM instruments:
  - Lagged values used as instruments in GMM may not satisfy exclusion restrictions, especially given potential rigidity in government spending.

### V. Conclusion and Policy Recommendations
- Empirical setting: Panel data for 32 oil producing countries over the period 1992 to 2009.
- Three main empirical results:
  - Within-country variation in current spending has a stronger impact on within-country variation in REER compared to capital spending.
  - Current spending is downwardly rigid but increases in booms; capital spending behaves conversely.
  - Fiscal rules show mixed results: limited reduction of current spending responsiveness during booms; associated with significant reduction in capital expenditure during busts while responsiveness to boosts is more muted.
- Economic implications:
  - Concern that restrictions that lead to capital spending cuts during busts may impede investment needed to diversify economies away from oil.
  - Lack of downward adjustment in REER during commodity busts may harm economic performance of resource-rich countries.
- Policy recommendations:
  - Limit increases in across-the-board spending during boom times to facilitate curbing spending during busts.
  - Tailor fiscal institutions to account for rigidity in current spending during busts to avoid crowding out capital spending.
  - Effectiveness of fiscal institutions depends on governments’ ability to design checks and balances to prevent rent seeking and limit creative accounting.

*Source: _wp1094 - References (PDF chapter/section).*

### References.  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  . 

### _wp1094 - References.

### Introduction — purpose and key empirical findings
- Purpose: investigate the behavior of expenditure policy during boom-bust commodity price cycles and its implication for real effective exchange rate (REER) movements in resource-rich (oil-producing) countries.
- Sample and scope:
  - Panel of 32 oil-producing countries over the period 1992 to 2009.
  - Government spending composition and country-specific oil price data from Villafuerte et al. (2009).
- Main empirical results (reported in the paper):
  - Change in current spending has a bigger impact on the change in REER compared to capital spending.
  - Current spending is downwardly sticky, but increases in boom time; conversely capital spending behaves differently (more responsive to busts).
  - Limited evidence that fiscal rules help reduce responsiveness of current spending during booms; fiscal rules are associated with a significant reduction in capital expenditure during busts while responsiveness to boosts is more muted.

### Theoretical model — Dutch disease with fiscal stickiness
- Objective: illustrate how asymmetrical fiscal responses to revenue shocks (particularly downward stickiness in current spending) affect REER dynamics.
- Key modeling assumptions:
  - Small open economy with two goods: tradable goods and a non-tradable government-provided service.
  - Government chooses current expenditure (non-tradable content: wage bill, subsidies, services) and capital expenditure (tradable goods).
  - Non-tradable government-provided services are subject to convex (quadratic) costs of adjustment with potential directional bias (downward stickiness).
  - REER is the shadow price of non-tradable government services relative to world price of tradables.
  - Stickiness modeled via quadratic costs of adjustment; parameter ω scales adjustment costs and parameter ν introduces persistent bias towards positive growth in government current expenditure when ν>0.
- Social planner problem (as presented):
  - max ∫_0^∞ [C_t^{1−1/σ} + ψ G_t^{1−1/σ}]^{1−1/σ} exp(−ρt) dt subject to intertemporal budget constraint including ·F = rF_t + p_t G_t + ω(·G − ν)^2/2 + C_t − T_t − E_t N_t.
  - Notation preserved exactly: C and G are government consumption of tradable and non-tradable goods; ρ is discount rate; σ is elasticity of substitution; F is stock of debt; r is interest rate; p is relative price of non-tradable goods (i.e. the real exchange rate); T is production of tradables; N_t is value of natural resource wealth at time t.

### Model simulations — parameterization and qualitative outcomes
- Simulation parameter values (exact as in source):
  - α=0.6, β=0.5, ψ=0.5, r=ρ=0.04, ω=0.5, σ=2.
- Present value of resource expressed as:
  - E_t N^p_t = E_t ∫_t^∞ P_t N_t exp(−ρt) dt.
- Normalization and initial conditions:
  - ∫_t^∞ N_t exp(−ρt) dt = 10 (equal to 10 years of full employment in one sector).
  - Start at P^∗ = 1 with the economy at steady state.
  - Government started with zero net foreign debt (assumed without loss of generality).
- Shock specification and cases simulated:
  - Permanent shock in permanent price: price jumps from P^∗ to P^∗∗ at t = 0 with no uncertainty.
  - Increase in present value of resource: N^p_t = (P^∗∗ − P^∗) ∫_t^∞ N_t exp(−ρt) dt = ε_p ∫_t^∞ N_t exp(−ρt) dt.
  - Cases simulated for ν = 0 (no asymmetry) and ν = 0.2 (with asymmetry).
  - Positive (negative) resource price shocks considered with ε_p = 0.5(−0.5).
- Simulation outcomes (qualitative, preserved from source):
  - With asymmetry between upward and downward adjustment (ν > 0), steady state values for non-tradable spending and the REER are elevated relative to the no-asymmetry case.
  - Under asymmetry, the decline in spending on non-tradables and the REER following a resource price bust is smaller.
  - Capital (tradable) spending is more sensitive to revenue shocks; asymmetry shifts adjustment weight toward capital (tradable) goods leaving current (non-tradable) expenditure less affected.
  - When ν>0 and adjustments are smoother (larger ω), government current spending can permanently crowd out capital expenditure, leaving C lower in steady state.

### Data and empirical strategy — variables and identification
- Data sources and variable construction (as described):
  - Government spending composition and country-specific oil price from Villafuerte et al. (2009) for 32 oil-producing countries, 1992–2009.
  - Oil export unit value: constructed using the international crude oil price interacted with a country-specific discount factor capturing crude quality (geological factors).
  - Decomposition of oil export unit value into:
    - (i) international crude oil price (common across countries),
    - (ii) country-specific discount factor (captures quality).
  - Because common year fixed effects are used in empirical analysis, identification of oil price impact comes from interaction between international oil price and country-specific discount factor.
  - Government current and capital spending taken from Villafuerte et al. (2009); current spending proxied as having higher non-tradable content (wage bill, subsidies, services); capital spending proxied as having higher import/tradable content.
- Empirical strategy elements:
  - Distinguish booms and busts using a three-way price decomposition drawn from the energy demand literature to capture asymmetrical responses to price increases and decreases.
  - Empirical specification and estimators are structured to isolate boom-bust cycles' effects on spending and on REER.

### Price decomposition — three-way decomposition of oil export unit value
- Decomposition (exact notation preserved):
  - oev_{i,t} = oevmax_{i,t} + oevmin_{i,t} + oevrec_{i,t}
- Component definitions (verbatim):
  - oevmax_{i,t}: cumulative increases in maximum historical oil unit export value for country i at time t; it is monotonically non-decreasing that is oevmax_{i,t} > 0.
  - oevmin_{i,t}: cumulative decreases in oil unit export value; it is monotonically non-increasing that is oevmin_{i,t} ≤ 0.
  - oevrec_{i,t}: cumulative sub-maximum increases in oil export value; it is monotonically non-decreasing that is oevrec_{i,t} > 0.
- Purpose: account for asymmetrical response of REER to oil price boom-bust cycles by separating maximum historical increases, cumulative decreases, and recoveries (sub-maximum increases).

*Content unit: _wp1094 - References. Publication PDF chapter/section provided by the source.*

### Appendix Edepicts the dynamic of the logarithm of the oil export unit valueoevand its decom-

### _wp1094 - Appendix Edepicts the dynamic of the logarithm of the oil export unit valueoevand its decom-

### D. Specifications and Estimators
- Model estimated (as presented in source):
  ∆ln(spending)it = αi + γt + βmax ∆ln(oevmax)it + βmin ∆ln(oevmin)it + βrec ∆ln(oevrec)it + γXit + uit
- Key elements of the estimation strategy:
  - αi are country fixed effects; γt are year fixed effects.
  - βj for j ∈ {max, min, rec} measure marginal effects of country-specific changes in logarithm of oil price components on country-specific changes in spending.
  - Control variables Xit (used as robustness checks) include first difference in logarithm of GDP, growthppp; and lagged difference in logarithm spending ∆ln(spending)t−1.
  - Estimators used: least squares and system-GMM (Blundell and Bond, 1998) to address dynamic panel bias with fixed effects.
  - Error term clustered at the country level; may be arbitrarily serially correlated within countries.
- Sign expectations for oil unit export value sub-components:
  - βmax > 0 (spending responds to change in oevmax).
  - βmin > 0 (spending responds to change in oevmin); note that oevmin < 0.
  - βrec > 0 (spending responds to change in oevrec).
  - Expected ordering in absolute values: βrec < βmin < βmax.
  - Expectation that current spending responds less to new lows than capital spending.

### IV. ESTIMATION RESULTS
A. Composition of Spending and the Real Exchange Rate
- Summary of key point estimates (Table 1, column (1)):
  - Point estimate on change in current spending: 0.242 (statistically significant at the 1 percent level).
  - Point estimate on change in capital spending: 0.055 (statistically significant at the 5 percent level).
  - Interpretation: A 1 standard deviation increase in current spending → increase in REER by about 0.33 standard deviations; a 1 standard deviation increase in capital spending → increase in REER by about 0.13 standard deviations.
- Robustness checks (columns (2) and (3)):
  - Current spending–REER link remains statistically significant when controlling for lagged current spending and within-country variation in income per capita.
  - Impact of capital spending on REER is much smaller; becomes statistically insignificant when controlling for lagged spending (column (2)).
- Dynamic specifications (columns (4) and (5); least squares and system-GMM):
  - Lagged dependent variable enters as statistically significantly positive.
  - Implied persistence: an impact of a shock to REER at time t would take about 0.9 years to dissipate by one-half for oil exporters.
  - Within-country increases in current spending continue to have statistically significant and larger effects than capital spending on REER.
- Sensitivity to outliers:
  - Grubbs test applied; dropping outliers yielded statistically significant point estimates on current spending that were quantitatively larger than those reported in Table 1 (results not shown).
- Statistical test:
  - For equations (1) to (3) authors can reject equality of current and capital expenditure coefficients at least at the 95 percent confidence level.

B. Boom and Bust Cycle and Spending Responses
- Framework: link within-country variations in three oil price components to within-country variations in current and capital spending; results summarized in Tables 2 and 3.
- Current spending (Table 2):
  - Point estimates on changes in new highs in oil price range from 0.277 (column (3)) to 0.376 (column (2)); statistically significant at the 10 percent and 5 percent levels respectively.
  - Other two oil price components have expected relative values but are not statistically significant.
  - Implication: current spending increases significantly when oil price reaches new highs but does not significantly decrease when oil price reaches new lows.
- Capital spending (Table 3):
  - Point estimates on change in new lows in oil price range from 0.840 (column (3)) to 1.507 (column (1)).
  - Other two components not statistically significant.
  - Implication: capital spending decreases when oil price reaches new lows, but does not significantly increase when oil price reaches new highs.

C. Fiscal Rules and Government Spending
- Fiscal rules definition and sample split:
  - Dummy from Ossowski et al. (2007): value 1 if any fiscal rules are in place and 0 otherwise.
  - Average of dummy over sample period used to perform sample split with cut-off 0.5.
- Main findings (Tables 4–7):
  - Limited evidence that fiscal rules reduce increase in current spending during booms.
    - Comparison of column (3) in Tables 4 and 5: coefficient for oil price new highs is positive and statistically significant for countries with no fiscal rules but not statistically significant for countries with fiscal rules.
  - Fiscal rules and capital spending during busts:
    - Countries with fiscal rules significantly reduce capital spending during busts; coefficients on new lows are systematically significant at the 1 percent significance level (Table 7).
    - For countries without fiscal rules, capital spending reduction during busts is found but results are less robust (column (1) of Table 6: coefficient on new highs not statistically significant).
  - Formal testing:
    - Generalized Chow test allowing for arbitrary within-country serial correlation indicates inability to reject that coefficients associated with oevmin in samples with and without fiscal rules are the same (tables 6 and 7).
- Interpretation and concern:
  - Fiscal rules have limited effectiveness in reducing current spending during booms.
  - Fiscal rules appear associated with systematic reduction of capital spending during busts — raising concerns about adverse long-term effects and the ability to diversify away from oil.

D. Endogeneity Issues
- Potential simultaneity bias in REER equation:
  - Appreciation in REER changes relative prices between tradables and non-tradables; substitution vs. income effect can bias least squares estimate on current spending downward.
- Instrumental variables strategy:
  - Use oevmax to instrument current spending; use oevmin to instrument capital spending.
  - Second-stage results do not provide evidence of a causal impact of spending on REER.
  - Various F-tests indicate instruments used in first stage are weak, casting doubt on validity of IV estimates.
- Note on GMM instruments:
  - Lagged values used as instruments in GMM may not satisfy exclusion restrictions, especially given potential rigidity in government spending.

### V. Conclusion and Policy Recommendations
- Empirical setting:
  - Panel data for 32 oil producing countries over the period 1992 to 2009.
- Three main empirical results:
  - Within-country variation in current spending has a stronger impact on within-country variation in REER compared to capital spending.
  - Current spending is downwardly rigid but increases in booms; capital spending behaves conversely.
  - Fiscal rules show mixed results: limited reduction of current spending responsiveness during booms; associated with significant reduction in capital expenditure during busts while responsiveness to boosts is more muted.
- Economic implications:
  - Concern that restrictions that lead to capital spending cuts during busts may impede investment needed to diversify economies away from oil.
  - Lack of downward adjustment in REER during commodity busts may harm economic performance of resource-rich countries.
- Policy recommendation:
  - Limit increases in across-the-board spending during boom times to facilitate curbing spending during busts.
  - Tailor fiscal institutions to account for rigidity in current spending during busts to avoid crowding out capital spending.
  - Effectiveness of fiscal institutions depends on governments’ ability to design checks and balances to prevent rent seeking and limit creative accounting.

*Source: _wp1094 - Appendix Edepicts the dynamic of the logarithm of the oil export unit valueoevand its decom-*

### REFERENCES

### REFERENCES

### Bibliographic sources cited
- Arezki, R. and M. Bruckner (2010). “International Commodity Price Shocks, Democracy and External Debt:”, IMF Working Paper No. 10/53, Washington DC
- Balassone, F. and M. Francese (2004). ”Cyclical asymmetry in fiscal policy, debt accumulation and the Treaty of Maastricht,” Temi di discussione (Economic working papers) 531, Bank of Italy, Economic Research Department.
- Corden W. M. and J. P. Neary (1982). “Booming Sector and De-industrialisation in a Small Open Economy.” The Economic Journal, 92 (December): pp. 825-848.
- Corden, W. M. (1984). “Boom Sector and Dutch Disease Economics: Survey and Consolidation.” Oxford Economic Papers 36: 362.
- Dargay, J. M. (1992). “The Irreversible Effects of High Oil Prices: Empirical Evidence for the Demand for Motor Fuels in France, Germany, and the UK, in Energy Demand: Evidence and Expectations, ed. D. Hawdon, London: Academic Press, 1992, pp. 165-82.
- Davis J. (2003). Fiscal policy formulation and implementation in Oil-Producing Countries, International Monetary Fund.
- Engel, E. and V. Vald ́e (2000). “Optimal Fiscal Strategy for Oil Exporting Countries”, IMF Working Paper No. 00/118
- Eslava, M. (2006). “The Political Economy of Fiscal Policy: Survey,” RES Working Papers 4487, Inter-American Development Bank, Research Department.
- Gately, D. (1992). “Imperfect Price-Reversibility of U.S. Gasoline Demand: Asymmetric Responses to Price Increases and Declines, Energy Journal, Vol. 13 (4), 1992, pp. 179-207.
- Gately, D. (1993). “The Imperfect Price Reversibility of World Oil Demand, Energy Journal, Vol. 14 (4), 1993, pp. 163-82.
- Gelb, A. and Associates (1988). Oil Windfalls: Blessing or Curse?, World Bank: Oxford University Press.
- Ismail, K. (2010a), “The Structural Manifestation of the ‘Dutch Disease’: The Case of Oil-Exporting Countries” IMF Working Paper No.10/ forthcoming.
- Ismail, K. (2010b), “Do Costs of Capital Adjustments Explain the Slow Dutch Disease?” , in ”Evidence of the ‘Dutch Disease’ and its Policy Implications.”, Dissertation, The Johns Hopkins University, 2010.
- Melitz, J. (1997). “Some cross-country evidence about debt, deficits and the behaviour of monetary and fiscal authorities”, CEPR Discussion Paper, 1653.
- Melitz, J. (2002). “Debt, Deficits and the Behaviour of Monetary and Fiscal Authorities”, in Buti, M., J. von Hagen and C. Martinez-Mongay, eds., The Behaviour of Fiscal Authorities, Palgrave.
- Milesi-Ferretti (2003). “Good, Bad or Ugly? On the Effects of Fiscal Rules with Creative Accounting,” Journal of Public Economics 88 no. 1-2, 2003, pp. 377-94.
- Ossowski, R., M. Villafuerte, P.A. Medas and T. Thomas (2008). “Managing the Oil Revenue Boom: The Role of Fiscal Institutions, Occasional Paper 260 (Washington, D.C.: International Monetary Fund).
- Rajan, R. G. and A. Subramanian (2005). “What Undermines Aid’s Impact on Growth?,” NBER Working Papers 11657, National Bureau of Economic Research, Inc.
- Spatafora, N. and A. Warner (2001). “Macroeconomic and Sectoral Effects of Terms -of-Trade Shocks -The Experience of the Oil-Exporting Developing Countries, IMF Staff Paper 134/99, October: 156.
- Villafuerte, M. and P. Lopez-Murphy (2009). “Fiscal Policy Responses of Oil Producing Countries to the Recent Oil Price Cycle”, IMF Working Paper No.10/28, Washington DC.
- Wijnbergen van, S. (1984). “The Dutch Disease: A Disease After All?” The Economic Journal, 94, pp. 373-41.

### Appendix A — Illustration of persistence of real exchange rate evolution during oil booms and busts
- Figures: 
  - Figure 1. Evolution of Nigeria’s REER and Oil Export Unit Value
  - Figure 2. Evolution of Venezuela’s REER and Oil Export Unit Value

### Appendix B — Model (summary of structure and key equations)
- Production:
  - Two goods: tradable (T) and non-tradable government provided services (G).
  - T_t ≡ L_T,t^β
  - G_t ≡ L_G,t^α
  - L_T,t + L_G,t ≡ 1
  - Tradable world price normalized to 1. Government non-tradable good cost p which equals the real exchange rate (p_t).
  - Labor market clearing condition: p_t α L_G,t^{α−1} = β L_T,t^{β−1}
- Fiscal policy (social planner problem):
  - Maximize ∫_0^∞ (C_t^{1−1/σ} + ψ G_t^{1−1/σ})/(1−1/σ) exp(−ρt) dt subject to intertemporal budget constraint
  - Budget constraint: ·F = rF_t + p_t G_t + ω(·G − ν)^2/2 + C_t − T_t − E_t N_t
  - Political cost term: ω(·G − ν)^2/2 with ν the downward bias parameter.
  - Hamiltonian: H_t(C_t,G_t,F_t) = C_t^{1−1/σ} + ψ G_t^{1−1/σ}/(1−1/σ) + λ_t [rF_t + p_t G_t + ω(·G_t − ν)^2/2 + C_t − T_t − E_t N_t]
- First-order conditions (selected):
  - G_t^{−1/σ} + λ_t (p_t + ω(·G_t − ν) + β(1 − G_t^{1/α})^{β−1} G_t^{1−α/α} α) = 0
  - C_t^{−1/σ} + λ_t = 0
  - ·λ = λ_t (ρ − r)
- Simplifications and characterization:
  - p_t = χ(G_t) = β(1 − G_t^{1/α})^{β−1} G_t^{1−α/α} α
  - G_t = (ψ/(2 p_t + ω(·G_t − ν)))^σ C_t
  - Assumption ρ = r (no consumption tilting, ·C = 0). Following a permanent resource price increase, present value of windfalls N_p^t rises, government increases G over time to limit adjustment costs, financing via debt accumulation then surpluses until an asset stock funds steady-state expenditure and import consumption.
- Long-run and steady-state relationships:
  - After resource depletes (N_t = 0): p^∗ G^∗ + ω ν^2/2 + C^∗ − T^∗ = − r F^∗
  - General path: p_t G_t + ω(·G − ν)^2/2 − (1 − G_t^{1/α})/β = − r F^∗ − C^∗
  - Fiscal adjustment path: p_t G_t + ω(·G − ν)^2/2 − (1 − G_t^{1/α})/β = r N_p^t − r F_t − C_t
  - N_p^t = ∫_t^∞ N_t exp(−ρ t) dt
  - Adjustment exhibits concave profile for resource boom and convex for resource bust; asymmetry when ν > 0 limits short-term downward adjustment.
- Steady-state when starting F = 0 and resource depletes:
  - r F^∗ = r N_p^0
  - G^∗ = [r N_p^t − ω ν^2/2 + (1 − G^{∗ 1/α})/β] ( (ψ/(2 p_t − ω ν))^σ + p^∗ )  (expression in source)
  - C^∗ = r N_p^t − ω ν^2/2 + (1 − ((ψ/(2 p^∗ − ω ν))^σ C^∗)^{1/α})/β (ψ/(2 p^∗ − ω ν))^σ + 1 = ( (2 p^∗ − ω ν)/ψ )^σ G^∗
  - p^∗ = β(1 − G_t^{1/α})^{β−1} G_t^{1−α/α} α
- Post-shock solution notation:
  - W^{∗∗} is post-shock amortized present value of country rent net of debt; C^{∗∗} is post-shock steady state tradable spending.

### Appendix C — Simulation results (summaries of scenarios)
- C.1 Positive Impulse to Natural Resource Price:
  - Figure 3. Adjustment to a 50 Percent Permanent Windfall Increase
  - Variables plotted over time t (axes labels reproduced from figures):
    - G_t: p_t with bias / without bias (values plotted in figure frame 0.19–0.22)
    - C_t: C with bias / C without bias (current and capital spending; frame 0.30–0.45)
    - p_t: REER with bias / REER without bias (frame 0.49–0.53)
    - T_t: T with bias / T without bias (Real Exchange Rate / Non-Mineral tradables production; frame 0.885–0.900)
    - M_t: M with bias / M without bias (Imports; frame −0.55 to −0.45)
- C.2 Negative Impulse to Natural Resource Price:
  - Figure 4. Adjustment to a 50 Percent Permanent Windfall Decrease
  - Variables plotted over time t (axes labels reproduced from figures):
    - G_t: p_t with bias / without bias (frame 0.16–0.19)
    - C_t: C with bias / C without bias (current and capital spending; frame 0.20–0.30)
    - p_t: REER with bias / REER without bias (frame 0.46–0.50)
    - T_t: T with bias / T without bias (Real Exchange Rate / Non-Mineral tradables production; frame 0.895–0.910)
    - M_t: M with bias / M without bias (Imports; frame −0.70 to −0.60)

### Appendix D — Data sources (variables and sources)
- oev: Oil export unit value (US$/barrel) — Villafuerte and Lopez-Murphy (2009)
- cepus: Central government current (primary) expenditures (US$) — Villafuerte and Lopez-Murphy (2009)
- cepusppp: Central government current (primary) expenditures (Constant International US$PPP adjusted) — Villafuerte and Lopez-Murphy (2009) and World Bank (2009)
- caxus: Central government capital expenditures (US$) — Villafuerte and Lopez-Murphy (2009)
- fiscalrule: Fiscal rule dummy; equals 0 if fiscal institution original dummy on average below 0.5, equals 1 if on average above 0.5 — Ossowski and al. (2008)

### Appendix E — Price decomposition
- Figure 5. Decomposition of the Oil Export Unit Value

### Appendix F — Empirical results (tables and key numeric findings)
- Table 1. The Effect of Government Current and Capital Expenditures on the Real Effective Exchange Rate (dependent variable D.lnereer)
  - Key coefficient examples (standard errors in parentheses):
    - D2.lnereer: 0.471 ∗∗∗ (0.0206) in column (1); 0.461 ∗∗∗ (0.0196) in column (2)
    - D.lncepus: 0.242 ∗∗∗ (0.0703) in column (1); 0.278 ∗∗∗ (0.100) in column (2); 0.237 ∗∗∗ (0.0701) in column (3); 0.0909 ∗ (0.0527) in column (4); 0.0837 ∗ (0.0440) in column (5)
    - D.lncaxus: 0.0545 ∗∗ (0.0221) in column (1); 0.00464 (0.0308) in column (2); 0.0550 ∗∗ (0.0223) in column (3); 0.0212 (0.0164) in column (4); 0.0243 ∗ (0.0202) in column (5)
  - Observations: 449, 420, 421, 424, 424 (across columns). R-squared examples: 0.220, 0.254, 0.231, 0.600 (selected).
  - Number of countries: 31 (all columns).
- Table 2. The Effect of Oil Unit Export Value’s Components on Government Current Spending (dependent variable D.lncepusppp)
  - D2.lncepusppp: 0.465 ∗∗∗ (0.00877)
  - D.lnoevmax: 0.338 ∗ (0.169) in column (1); 0.376 ∗∗ (0.176) in column (2); 0.277 ∗ (0.152) in column (3)
  - Observations: 388, 388, 366. R-squared: 0.100, 0.109. Number of countries: 31.
- Table 3. The Effect of Oil Unit Export Value’s Components on Government Capital Spending (dependent variable D.lncaxus)
  - D2.lncaxus: 0.430 ∗∗∗ (0.0163)
  - D.lnoevmin: 1.306 ∗∗∗ (0.378) in column (1); 1.507 ∗∗∗ (0.427) in column (2); 0.840 ∗∗∗ (0.310) in column (3)
  - growthppp: 1.677 ∗∗∗ (0.451)
  - Observations: 452, 394, 430. R-squared: 0.165, 0.186. Number of countries: 31.
- F.1 Results for Country Grouping by Fiscal Rule (selected tables):
  - Table 4 (No Fiscal Rules) — D.lncepusppp:
    - D2.lncepusppp: 0.464 ∗∗∗ (0.00908)
    - Observations: 189, 189, 179. R-squared examples: 0.081, 0.086. Number of countries: 15.
  - Table 5 (With Fiscal Rules) — D.lncepusppp:
    - D2.lncepusppp: 0.490 ∗∗∗ (0.00741)
    - Observations: 192, 192, 180. R-squared examples: 0.185, 0.220. Number of countries: 15.
  - Table 6 (No Fiscal Rules) — D.lncaxus:
    - D2.lncaxus: 0.473 ∗∗∗ (0.0163)
    - growthppp: 1.544 ∗∗ (0.604)
    - Observations: 209, 193, 199. R-squared examples: 0.185, 0.204. Number of countries: 15.
  - Table 7 (With Fiscal Rules) — D.lncaxus:
    - D2.lncaxus: 0.436 ∗∗∗ (0.0278)
    - D.lnoevmin: 1.112 ∗∗∗ (0.225) in column (1); 1.890 ∗∗∗ (0.578) in column (2); 1.018 ∗∗∗ (0.194) in column (3)
    - growthppp: 2.078 ∗∗ (0.767)
    - Observations: 235, 194, 223. R-squared examples: 0.217, 0.268. Number of countries: 15.

- Notes across tables:
  - Robust standard errors in parentheses.
  - Significance notation: ∗∗∗ p<0.01, ∗∗ p<0.05, ∗ p<0.1.

*Content derived from the document: _wp1094 - REFERENCES*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1094.pdf_