## wpiea2019062

## Source details

**Canonical URL:** [wpiea2019062](https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019062.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2019/wpiea2019062.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2019/wpiea2019062.pdf.json)

---

### I. Model framework and methodological features
- Model class and solution:
  - Real dynamic open-economy DSGE models with traded and non-traded sectors; DIGNAR adds a natural resource sector.
  - Neoclassical assumptions: flexible prices, perfectly competitive markets; solved with a fully non-linear perfect-foresight method.
- Three integrated building blocks:
  - Investment–growth nexus:
    - Production uses labor, private capital, and public capital in Cobb-Douglas form.
    - Public investment inefficiency captured by ε with 0≤ε≤1, so public investment i_t^g maps into public capital k_t^g = (1−δ)k_{t−1}^g + ε i_t^g.
    - DIGNAR allows ε to decline when growth rate of public investment exceeds historical levels to capture absorptive capacity constraints.
  - Fiscal adjustment:
    - Government budget constraint includes ∆a_t (external commercial borrowing), ∆d_t (external concessional borrowing), ∆b_t (domestic borrowing), ∆f_t (resource fund accumulation), interest payments, i_t^g, g_t (consumption/transfers), grants A_t, royalties ℛ_t, and taxes τ_{jt} n_{jt}.
    - Fiscal instruments follow non-linear reaction rules with ceilings/floors: τ_{jt} = Min{τ_j^u, τ_{jt}^r}, where τ_{jt}^r = τ_{jt−1} + λ_τ(τ_{jt}^{target} − τ_{jt−1}) + λ_x(x_t − x_t^{target}), with λ_τ, λ_x > 0 and x_t a debt-to-GDP ratio.
  - Private sector response:
    - Private investment faces adjustment costs and limited international market access.
    - Heterogeneous consumers (asset-holders and hand-to-mouth) break Ricardian equivalence under distortionary taxes.
    - Public investment can crowd in private investment in the long run if projects are productive; fiscal financing can induce crowding-out, especially during the transition and with limited foreign financing.

### II. Empirical scope and calibration practices
- Scope:
  - Analysis draws on over 65 DIG and DIGNAR country applications.
  - Models calibrated per country to quantify output multipliers, growth gains from efficiency improvements, costs of absorptive constraints, crowding-out magnitudes, and mitigation effects of aid and revenue mobilization.
- Calibration inputs and indicators:
  - Public Investment Management Assessment (PIMA) scores used in some applications to inform ε calibration.
  - Country-specific extensions include energy sectors, operations & maintenance, climate-resilience, human capital vs economic infrastructure trade-offs, and governance channels.

### III. Key quantitative findings and scenario outcomes
- Improving public investment efficiency / returns:
  - DIG simulation (average LIC, Buffie et al. (2012) calibration):
    - Scaling-up: 10-year average scaling-up of 4.5 percentage points of public investment to GDP.
    - Efficiency improvement: from 0.5 to 0.75.
    - Result: gain of 1.4 percentage points in per-capita GDP growth over a decade.
    - Counterfactual: efficiency decline from 0.5 to 0.25 yields much smaller growth dividends and can make public debt unsustainable if financed with more non-concessional borrowing.
  - DIGNAR — Botswana (IMF, 2016b):
    - Efficiency: improves from 60 to 80 percent over ten years.
    - Annual net return on public capital: increases from 15 to 20 percent.
    - Investment path: gradual scaling up of infrastructure up to 12 percent of GDP over the medium term.
    - Result: increase in growth rate of non-mineral real GDP by an additional 1.2 percentage points per year vs baseline; private investment and consumption significantly higher; debt burden smaller.
- Prudent versus aggressive scaling-ups:
  - General result: prudent and gradual scaling-ups reduce short-run crowding-out, respect absorptive limits, and lower debt sustainability risks compared with aggressive front-loaded approaches.
  - DIGNAR — Kazakhstan (IMF, 2013a):
    - Prudent: public investment raised to 6.4 percent of GDP on average.
    - Aggressive: public investment raised to 12 percent of GDP on average.
    - Result: aggressive scenario could raise non-oil growth by more than 0.2 percentage points on average in the next decade, but consumption could fall by up to 5 percent over the next decade; stronger crowding-out of private consumption and tradable-sector investment.
  - DIGNAR — Botswana absorptive illustration:
    - Gradual: growth-enhancing investment increases from roughly 8 percent to 12 percent of GDP.
    - Aggressive: public investment reaches about 14 percent of GDP in the short run.
    - Efficiency implication: aggressive approach reduced efficiency from 0.6 to 0.55; delivered only slightly larger public capital build-up but had higher financing needs and larger decline in private consumption.
  - CEMAC quantified cost of constraints (Berg et al., 2013):
    - When expenditure levels exceed 60 percent of historical level, efficiency falls from 0.5 to 0.35.
    - Present-value cumulative non-oil output multiplier: 0.41 with constraints vs 0.64 without constraints.
- Resource-rich country scenarios and Mozambique case:
  - Mozambique public investment scenarios (conservative, gradual, aggressive):
    - Exports expected to peak at 30 percent of non-oil GDP by 2028.
    - Fiscal revenue from the sector could bring in as much as 9 percent of non-oil GDP.
    - Aggressive scaling-up could push public debt to levels above 70 percent of GDP.
    - Adverse scenario with 80 percent lower LNG output and decline in prices: public debt becomes unsustainable (explosive) under aggressive approach.
    - Mechanism: aggressive approach can produce real exchange rate appreciation (by year 4 appreciation could be twice that of gradual approach), reducing traded-sector output.
  - External savings / resource funds:
    - Saving part of resource revenues abroad mitigates Dutch disease and builds fiscal buffers; under conservative approach Mozambique could accumulate fund savings usable under adverse LNG scenarios; under aggressive approach and adverse scenario there would be no fund savings and public debt is unsustainable.
- Domestic revenue mobilization, aid, and concessional financing:
  - CEMAC and commodity-exporting LICs:
    - Mobilizing domestic revenues contains public debt surges and repayment burdens.
    - Example: improvements in tax collection of 20 percent combined with additional concessional financing of 5 percent of GDP could significantly slow public debt accumulation and stabilize debt-to-GDP ratios at around 45 percent in the medium term.
    - VAT increases may induce non-oil GDP fall on impact but can lead to debt tapering off and non-oil GDP recovery in the medium run; regressive impacts on HTM consumers can be mitigated via targeted cash transfers or increased public investment.
  - Serbia DIG simulations:
    - Aid financing (50 percent grants / 50 percent concessional loans) eliminates need for painful fiscal adjustments and raises real GDP growth.
    - In the first decade, relative to non-concessional borrowing, aid could imply lower consumption tax rates and bigger real GDP growth rates by about 1 and 0.2 percentage points, respectively.
  - Commodity-exporting LICs (2016 DIGNAR): declines in oil-related revenues can generate rapid public debt surges—about 28 percentage points of GDP in 3 years; concessional financing helps address fiscal gaps and contain interest burden increases.
- Sector-specific and other quantitative results:
  - Ethiopia energy calibration:
    - Returns: infrastructure 20 percent, energy 30 percent.
    - Energy investment and Renaissance Dam efficiency gains could increase total energy supply four-fold (300 percent) at 10-year horizon; infrastructure stock would double in long run.
    - Financing trade-offs: domestic bank borrowing — first 5 years require transfers cut by more than 5 percent of GDP; private investment falls by more than 10 percent; inflation could rise above 10 percent. External commercial borrowing allows higher public investment without cuts in recurrent expenditures but external public debt could reach 50 percent of GDP.
  - Senegal energy application: coordinated low-cost hydro and phased contraction of oil sector can raise energy supply by 70 percent and increase real wages and real GDP by more than 10 percent medium term; short-term fiscal deficit could widen by 1 percent of GDP, medium-term fiscal gains about 4 percent of GDP.
  - Governance reforms (EMDE calibration): improving governance from first to second quartile over 10 years yields private investment 70 percent higher after 10 years, output and private consumption increase by about 30 percent, and public debt falls by almost 15 percent of GDP after 10 years.
  - Operations & Maintenance (O&M): deficient O&M raises public capital depreciation and reduces output from existing public capital; investment surges without matching O&M produce substantially lower growth rates.
  - Vanuatu disaster resilience:
    - Two public capital types (standard and adaptation); adaptation raises marginal product of private capital and crowds in private investment.
    - Tax-financed reconstruction can reduce private consumption and investment by as much as 20 and 40 percent, respectively.
    - Debt-financed reconstruction leads to unsustainable debt paths; viable option requires average additional grants equal to 50 percent of pre-cyclone grants levels over 15 years after the disaster.

### IV. Recurrent policy lessons (empirical regularities)
- Efficiency and returns:
  - Improving public investment efficiency and/or raising project rates of return boosts growth and lowers debt sustainability risks; reforms raising both magnify macro gains and reduce debt risks.
  - Theoretical invariance (Cobb-Douglas): high- and low-efficiency countries reap the same growth gain from the same additional public investment spending; what matters is increase in efficiency (ε) rather than its initial level.
- Pace and sequencing:
  - Prudent, gradual, and well-sequenced scaling-ups preferable to aggressive front-loaded plans to avoid large short-run crowding-out and binding absorptive constraints.
- Financing mix:
  - Leverage concessional financing and grants to smooth fiscal adjustment and stabilize debt.
  - Mobilize domestic revenues to reduce reliance on commercial borrowing and contain repayment burdens.
  - Use external savings/resource funds to mitigate Dutch-disease effects and create fiscal buffers.
- Fiscal reaction design:
  - Calibrate fiscal reaction functions (ceiling/floor constraints and speed parameters λ_τ, λ_x) to realistic political and administrative capacities to avoid explosive debt dynamics.
- Private sector and distributional effects:
  - Account for heterogeneous households (HTM vs asset-holders); short-term crowding-out may dominate even if long-run crowding-in is expected.
  - Mitigate regressive impacts of distortionary tax increases (e.g., VAT) with targeted cash transfers or compensatory public investment.
- Sectoral, institutional, and operational considerations:
  - Incorporate sectoral extensions (energy, human capital, resilience, security, governance) to capture project-specific trade-offs and distributional consequences.
  - Ensure adequate O&M funding alongside new investment to realize projected growth gains.
- Use of model-based tools:
  - Employ DIG and DIGNAR for transparent, consistent quantification of trade-offs; use them to provide consistency checks on narrative projections (e.g., 20-year DSAs), test alternative financing scenarios, and inform mission-team policy discussions.

### V. Specific numerical highlights and scenario thresholds (select)
- DIG LIC scaling-up example: 4.5 percentage points of public investment to GDP over 10 years → 1.4 percentage points per-capita GDP growth gain over a decade when ε improves from 0.5 to 0.75.
- Botswana DIGNAR: efficiency 60→80 percent and returns 15→20 percent with infrastructure up to 12 percent of GDP → additional 1.2 percentage points annual non-mineral growth vs baseline.
- Kazakhstan scenarios: prudent 6.4 percent of GDP vs aggressive 12 percent of GDP public investment; aggressive can reduce consumption by up to 5 percent over a decade.
- CEMAC absorptive threshold: exceeding 60 percent of historical expenditure levels → ε falls from 0.5 to 0.35; PV cumulative non-oil output multiplier 0.41 with constraints vs 0.64 without.
- Mozambique: exports peak at 30 percent of non-oil GDP by 2028; sector fiscal revenue up to 9 percent of non-oil GDP; aggressive scaling-up could push public debt above 70 percent of GDP; adverse 80 percent LNG output collapse → unsustainable debt under aggressive approach.
- Ethiopia energy: potential 300 percent increase in total energy supply at 10 years with high-return projects; domestic bank financing consequences — transfers cut by more than 5 percent of GDP in first 5 years, private investment falls by more than 10 percent, inflation could rise above 10 percent; external commercial borrowing could raise external public debt to 50 percent of GDP.
- Tax/revenue illustrative policy: improving tax collection by 20 percent + concessional financing 5 percent of GDP → stabilize debt-to-GDP at around 45 percent in medium term.
- Aid/Serbia: aid financing could imply lower consumption tax rates and higher real GDP growth by about 1 and 0.2 percentage points, respectively, in the first decade vs non-concessional borrowing.
- Commodity export shock: revenue declines can generate public debt surges of about 28 percentage points of GDP in 3 years (commodity-exporting LICs DIGNAR application).
- Governance reform gains: private investment 70 percent higher; output and private consumption about 30 percent higher; public debt falls almost 15 percent of GDP after 10 years.
- Vanuatu reconstruction: tax-financed reconstruction can reduce private consumption and investment by as much as 20 and 40 percent; viable post-disaster path requires additional grants averaging 50 percent of pre-cyclone grants levels over 15 years.

### VI. Policy recommendations (operational guidance)
- Sequence and pace:
  - Favor gradual, well-sequenced public investment scaling-ups to avoid short-run crowding-out and respect absorptive limits.
- Improve efficiency and project selection:
  - Prioritize investment appraisal, selection, implementation, and evaluation to raise ε and the average real return on public capital.
- Optimize financing mix:
  - Use concessional financing and grants where available; mobilize domestic revenues to reduce commercial borrowing; save resource revenues in sovereign wealth funds.
- Design fiscal reaction rules carefully:
  - Implement non-linear reaction functions with realistic ceilings/floors and calibrate λ_τ, λ_x to political/administrative capacities.
- Account for private sector dynamics and equity:
  - Anticipate short-term crowding-out and heterogeneous household impacts; accompany tax increases with targeted compensatory measures.
- Ensure complementary policies:
  - Maintain adequate O&M budgets; pursue governance reforms and tax-administration improvements to raise investment effectiveness.
- Utilize model-based analysis:
  - Use DIG and DIGNAR as transparent tools for scenario analysis, consistency checks on DSAs and long-run projections, and structured policy discussions; perform continuous sensitivity analysis.

*Italic: Source: wpiea2019062 — IMF staff paper content (selected sections: Figures list, I. Introduction, II. Model description, III. Policy lessons).*

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

### wpiea2019062 - References

### I. Introduction — framing the public investment challenge
- Infrastructure constraints limit growth in developing countries; scaling up public investment is a central policy objective and a component of the 2030 Development Agenda.
- Key policy trade-offs: choice and mix of projects, pace of scaling-up, effects on growth, absorptive capacity constraints, investment efficiency and returns, and public debt sustainability.
- With rising debt vulnerabilities, countries must balance ambitious public investment plans with macroeconomic and debt sustainability objectives.
- Criticisms of existing DSF approaches include lack of explicit macroeconomic linkages between public investment, borrowing, and growth, and assumptions that governments do not react to shocks.
- To address these gaps, Fund staff developed two structural DSGE-based frameworks: the Debt, Investment, and Growth (DIG) model and its natural-resource extension DIGNAR; both emphasize explicit linkages between public investment, growth, and debt, and the private sector and fiscal policy reaction functions.

### II. Model features and methodological approach
- Model class and solution:
  - Both models are real dynamic open economy models with traded and non-traded sectors and separate export and import prices; DIGNAR adds a natural resource sector.
  - Neoclassical assumptions: flexible prices, perfectly competitive markets, focus on medium- to long-term outcomes, solved with a fully non-linear perfect foresight solution method.
- Three integrated pieces:
  - Investment-growth nexus:
    - Production uses labor, private capital, and public capital (Cobb-Douglas form).
    - Public capital raises private marginal productivity and can crowd in private investment.
    - Public investment inefficiencies captured via an efficiency parameter 0≤ε≤1 so that public investment i_t^g translates into public capital k_t^g = (1−δ)k_{t−1}^g + ε i_t^g.
    - DIGNAR captures absorptive capacity constraints by allowing efficiency ε to decline when the growth rate of public investment exceeds historical levels.
    - Public Investment Management Assessment (PIMA) scores have been used to inform calibration of efficiency in some applications.
  - Fiscal adjustment:
    - Government budget constraint explicitly includes external commercial borrowing (∆a_t), external concessional borrowing (∆d_t), domestic borrowing (∆b_t), resource fund accumulation (∆f_t), interest payments, public investment i_t^g, consumption/transfers g_t, grants A_t, resource royalties ℛ_t, and taxes represented in general form τ_{jt} n_{jt}.
    - Grants and concessional borrowing are taken as exogenous; fiscal gaps from scaling-up are typically financed by external commercial and domestic borrowing, with taxes and transfers responding to stabilize debt.
    - DIGNAR features a resource fund that acts as a fiscal buffer subject to a minimum savings bound; withdrawals are an additional financing option.
    - Fiscal instruments follow non-linear reaction rules with feasibility ceilings/floors: τ_{jt} = Min{τ_j^u, τ_{jt}^r}, where τ_{jt}^r = τ_{jt−1} + λ_τ(τ_{jt}^{target} − τ_{jt−1}) + λ_x(x_t − x_t^{target}), with λ_τ, λ_x > 0 and x_t a debt-to-GDP ratio.
    - Slow fiscal adjustment (low λ_τ and λ_x) or binding ceilings τ_j^u can lead to explosive debt dynamics; policy must "win the race against time" to reach debt-stabilizing outcomes.
  - Private sector response:
    - Private investment subject to adjustment costs and limited international capital market access.
    - Heterogeneous consumers: asset-holders who can smooth consumption and hand-to-mouth consumers — breaking Ricardian equivalence when distortionary taxes are present.
    - Public investment generates crowding-in via higher private marginal product in the long run if projects are productive; fiscal financing (tax increases, transfer cuts, domestic resource use) can generate crowding-out, especially during the transition and with limited foreign financing. Transition dominance depends on financing mix and absorptive constraints.

### III. Empirical and policy-relevant findings from applications
- Scope of applications:
  - The analysis draws on over 65 DIG and DIGNAR country applications.
  - Models have been calibrated per country to quantify effects such as output multipliers of public investment, growth gains from improved investment efficiency, costs of absorptive capacity constraints, magnitude of crowding-out on private demand, and mitigation effects of aid and revenue mobilization.
- Recurrent policy lessons across applications:
  - Improving public investment efficiency and/or raising project rates of return boosts growth and lowers debt sustainability risks.
  - Prudent and gradual public investment scaling-ups are generally preferable to aggressive, front-loaded scaling-ups because they reduce private sector crowding-out, ease absorptive capacity constraints, and lower debt sustainability risks.
  - Domestic revenue mobilization helps create fiscal space for investment scaling-ups by containing public debt surges and repayment burdens.
  - Aid smooths fiscal adjustments associated with public investment increases and may lower the risk of unsustainable debt.
  - External savings (including resource funds) mitigate Dutch-disease macroeconomic effects and serve as fiscal buffers.
- Model extensions and use cases:
  - DIG extensions: explicit energy sector (applications: Ethiopia, Senegal, Lesotho), operations and maintenance costs, climate-resilient infrastructure, trade-offs between economic and social infrastructure, trade-offs with security spending.
  - DIGNAR extensions: poverty reduction via direct cash transfers, design of growth-friendly fiscal adjustments.
  - Models have been used to provide consistency checks on 20-year projections used in DSF-style debt analyses and to generate internally consistent, model-based policy scenarios incorporating fiscal reaction functions and non-linear adjustment dynamics.

### IV. Policy implications and recommended strategies
- Sequence and pacing:
  - Favor gradual, well-sequenced public investment scaling-ups to avoid large short-run crowding-out and to respect absorptive capacity limits.
- Efficiency and project selection:
  - Prioritize improving public investment efficiency (project selection, management, implementation) to increase the conversion rate ε of investment into productive public capital and improve growth payoffs.
- Financing mix:
  - Leverage concessional financing and grants where available to smooth fiscal adjustment and stabilize debt.
  - Mobilize domestic revenues to reduce reliance on commercial borrowing and to contain later repayment burdens.
  - Use resource funds and external savings as fiscal buffers to mitigate macroeconomic and Dutch-disease effects.
- Fiscal reaction design:
  - Design non-linear fiscal reaction functions with feasible ceilings/floors that avoid binding constraints that could precipitate explosive debt dynamics.
  - Calibrate the speed parameters (λ_τ, λ_x) to realistic political and administrative capacities to ensure timely adjustment.
- Consider private sector dynamics:
  - Account for heterogeneous household behavior and limited market access; recognize that short-term crowding-out may dominate even when long-run crowding-in is expected.
- Use of DSGE-based tools:
  - Employ DIG and DIGNAR as transparent, consistent frameworks to quantify trade-offs, provide consistency checks on narrative projections, and compare alternative financing and policy scenarios.

*Italic: Source: wpiea2019062 — IMF staff paper content (selected sections: Figures list, I. Introduction, II. Model description, III. Policy lessons).*

### 1.      Improving public investment efficiency and/or raising the rate of return of public

### 1.      Improving public investment efficiency and/or raising the rate of return of public

### A. Improving public investment efficiency and/or raising the rate of return of public projects
- Key conceptual finding:
  - Improving investment appraisal, selection, implementation, and evaluation can increase efficiency and translate public investment into more public capital and higher growth.
  - DIG and DIGNAR models used to quantify macroeconomic effects of changes in investment efficiency and project returns.
- DIG simulation calibrated to an average LIC (following Buffie et al. (2012)):
  - Scaling-up: 10-year average scaling-up of 4.5 percentage points of public investment to GDP.
  - Efficiency improvement: from 0.5 to 0.75.
  - Result: could deliver a gain of 1.4 percentage points in per-capita GDP growth over a decade.
  - Counterfactual: decline of efficiency from 0.5 to 0.25 → much smaller growth dividends; if financed with more non-concessional borrowing, public debt could become unsustainable.
- Theoretical/invariance result (Cobb-Douglas production functions):
  - High- and low-efficiency countries reap the same growth gain from the same additional public investment spending (invariance result of Berg et al. (2018)).
  - Explanation: marginal contribution = (amount of capital installed) × (marginal productivity). Lower installed capital raises marginal productivity; effects offset under Cobb-Douglas. What matters is the increase in efficiency, not its level.
- DIGNAR application to Botswana (IMF, 2016b) — combining efficiency and returns:
  - Assumed reforms: efficiency improves from 60 to 80 percent over ten years; annual net return on public capital increases from 15 to 20 percent.
  - Investment path: gradual scaling up of infrastructure up to 12 percent of GDP over the medium-term.
  - Result: increase in growth rate of non-mineral real GDP, on average, by an additional 1.2 percentage points per year compared to the baseline without reforms.
  - Additional outcomes: private investment and consumption significantly higher; debt burden smaller.
- Policy implication:
  - Reforms that raise both efficiency and the average real return on public capital magnify macroeconomic gains and reduce debt risks.

### B. Prudent and gradual public investment scaling-ups are preferable to aggressive front-loaded ones
- Overall finding:
  - Large, front-loaded scaling-ups can exacerbate private-sector crowding-out, reveal absorptive capacity constraints, and raise fiscal and debt sustainability risks.
  - Prudent and gradual scaling-ups are generally preferable when considering crowding-out, absorptive capacity, and debt sustainability.
- Crowding-out versus crowding-in dynamics:
  - With high returns and improved efficiency, public investment can crowd in private investment and consumption in the medium to long run.
  - Short run: fiscal adjustments (higher taxes or cuts) and higher interest rates can crowd out private demand.
  - Domestic borrowing intensifies crowding-out; external commercial borrowing can smooth private-sector adjustment but increases public debt-to-GDP ratios.
- DIGNAR application to Kazakhstan (IMF, 2013a) — prudent vs aggressive:
  - Prudent scenario: public investment raised, on average, to 6.4 percent of GDP.
  - Aggressive scenario: public investment raised, on average, to 12 percent of GDP.
  - Result: aggressive scenario could boost public capital and non-oil growth relative to the prudent scenario (non-oil growth could be higher by more than 0.2 percentage points, on average, in the next decade).
  - Cost: aggressive scenario could lead to stronger crowding-out of private consumption and tradable-sector investment; consumption could fall by up to 5 percent in the aggressive scenario over the next decade, contrasted with persistent increase in the prudent scenario.
- Absorptive capacity constraints (DIGNAR endogenization):
  - Efficiency is modeled as a decreasing function of the pace of the scaling-up; fast pace → binding absorptive constraints → lower efficiency → more wasted spending.
- DIGNAR application to Botswana (IMF, 2016b) — absorptive capacity illustration:
  - Gradual approach: spending on growth-enhancing investment projects increased from roughly 8 percent to 12 percent of GDP over the short to medium run.
  - Aggressive approach: public investment increased faster, reaching about 14 percent of GDP in the short run.
  - Efficiency implication: aggressive approach led to a decrease in public investment efficiency from 0.6 to 0.55.
  - Outcome: aggressive approach delivered only a slightly larger build-up of public capital and non-mineral output relative to the gradual approach, but implied higher financing needs (tax rate hikes and more borrowing) and a more pronounced decline of private consumption.
- Quantified cost of absorptive capacity constraints (CEMAC application, Berg et al., 2013):
  - When expenditure levels surpassed 60 percent of their historical level, absorptive capacity constraints were binding and investment efficiency fell from 0.5 to 0.35.
  - Present-value cumulative non-oil output multiplier:
    - All-investing approach with constraints: 0.41.
    - In absence of constraints: 0.64.
  - Implication: substantial output loss because of absorptive capacity constraints inducing investment inefficiencies.
- Public debt sustainability risk:
  - Sharp investment increases financed by external commercial borrowing can push public debt to dangerous levels, especially under negative shocks, potentially necessitating painful fiscal adjustments in the medium term.

*Source: wpiea2019062 (IMF).*

### 2020. By 2028 exports were expected to peak at 30 percent of non-oil GDP, while fiscal

### wpiea2019062 - 2020. By 2028 exports were expected to peak at 30 percent of non-oil GDP, while fiscal

### Public investment scaling-up scenarios (Mozambique)
- Three approaches analyzed: conservative, gradual, aggressive.
- Key quantitative outcomes:
  - Exports expected to peak at 30 percent of non-oil GDP by 2028.
  - Fiscal revenue from the sector could bring in as much as 9 percent of non-oil GDP.
  - Aggressive scaling-up could push public debt to levels above 70 percent of GDP.
  - Under an adverse scenario of 80 percent lower LNG output and decline in prices, public debt becomes unsustainable (explosive) with the aggressive approach.
- Mechanism and risks:
  - Conservative approach: investment increases only when LNG production starts; implies neither fiscal adjustment nor debt accumulation.
  - Gradual approach: anticipates some LNG revenue and covers remainder with borrowing; public debt rises to sustainable levels in the medium term, requiring a small fiscal adjustment.
  - Aggressive approach: fully anticipates future LNG revenues and front-loads investment; leads to large debt build-up and more painful fiscal adjustment, greater vulnerability.
- Dutch disease and traded sector effects:
  - Aggressive approach induces larger real exchange rate appreciation (e.g., by year 4 appreciation could be twice that of the gradual approach), reducing traded sector output and exacerbating productivity losses.

### Domestic revenue mobilization (CEMAC and resource-rich LICs)
- Role and effects:
  - Mobilizing domestic revenues helps contain public debt surges and the burden of principal and interest repayments.
  - Average revenue-to-GDP in SSA remains lower than other regions despite gradual improvement.
  - For resource-rich countries, improved tax administration, broader tax base, prioritizing current expenditures, and reducing subsidies (e.g., fuel) can mitigate commodity price volatility effects.
- Simulation results (CEMAC):
  - Increases in non-oil tax revenues create fiscal space for public investment while reducing debt pressures.
  - Raising the VAT rate induces a non-oil GDP fall on impact but leads to debt tapering off; non-oil GDP recovers in the medium run driven by higher private investment and net exports.
  - VAT increases can negatively affect poor consumers (HTM) due to larger marginal propensity to consume; mitigation via targeted cash transfers or increased public investment can offset adverse effects.
- Quantitative example for commodity-exporting LICs:
  - Improvements in tax collection of 20 percent combined with additional concessional financing of 5 percent of GDP could significantly slow public debt accumulation and stabilize debt-to-GDP ratios at around 45 percent in the medium term.

### Aid and concessional financing (Serbia; commodity-exporting LICs)
- Financing options compared: domestic borrowing, external commercial borrowing, aid (50 percent grants / 50 percent concessional loans).
- Serbia DIG simulations:
  - Aid financing eliminates the need for painful fiscal adjustments, leading to higher real GDP growth.
  - In the first decade, relative to non-concessional borrowing, aid could imply lower consumption tax rates and bigger real GDP growth rates by about 1 and 0.2 percentage points, respectively.
  - Grant/concessional financing has almost negligible impact on debt-to-GDP ratios compared with non-concessional financing.
- Commodity-exporting LICs (2016 WEO DIGNAR application):
  - Declines in government oil-related revenues could generate rapid public debt surges—about 28 percentage points of GDP in 3 years.
  - Concessional financing helps address fiscal gaps and contain increases in interest burden and sovereign risk premia.

### External savings and sovereign wealth funds
- Rationale:
  - Saving part of natural resource revenues abroad can mitigate Dutch disease (real appreciation, productivity losses in tradable sector) and build fiscal buffers for adverse commodity shocks.
- Mozambique application:
  - Under conservative approach, government savings to a resource fund create substantial fiscal buffers usable under adverse LNG scenarios; public debt trajectories and required tax increases under adverse scenarios similar in magnitude to baseline when savings occur.
  - Under aggressive approach and adverse scenario, there would be no fund savings and public debt follows an unsustainable path.

### Extensions and sector-specific applications
- Energy sector (Ethiopia, Senegal, Lesotho):
  - Ethiopia calibration: high returns—infrastructure 20 percent, energy 30 percent; energy investment and Renaissance Dam efficiency gains could increase total energy supply four-fold (300 percent) at 10-year horizon; infrastructure stock would double in long run.
  - Financing trade-offs:
    - Reliance on domestic bank borrowing: first 5 years require transfers cut by more than 5 percent of GDP; private investment falls by more than 10 percent; inflation could rise above 10 percent.
    - External commercial borrowing: allows higher public investment without cuts in recurrent expenditures or temporary crowding-out, but external public debt could reach 50 percent of GDP.
  - Senegal: coordinated low-cost hydroelectric and phased contraction of oil-based sector could raise energy supply by 70 percent and increase real wages and real GDP by more than 10 percent medium term; short-term fiscal deficit could widen by 1 percent of GDP, medium-term fiscal gains about 4 percent of GDP.
- Energy price and revenue measures:
  - Doubling energy prices charged to firms or increasing tax revenues by about 1 percent of GDP could induce large reductions in growth of external debt.
- Counterfactuals and DSA interpretation (Togo):
  - DIG used to invert model dynamics and retrieve implied structural shocks; revealed IMF-WB DSA forecasts implicitly assumed a long-run fall in share of exports to GDP potentially reflecting negative productivity shocks in export sector; projections were corrected accordingly.
- Human capital vs economic infrastructure (Atolia et al.):
  - With a 15 percent return differential favoring schools, optimal allocation under a “big push” still limits the share of investment in schools to about a half.
  - Political myopia: planners with horizons under 30 years would not invest in schools at all; tied concessional financing and grants can counteract debt aversion and political myopia.
- Social transfers and poverty (Mano River Union; Liberia):
  - Diverting a fraction of resource revenue to transfers toward poor financially constrained consumers can substantially raise private consumption, sometimes more than public investment alone raises non-resource GDP.
- Growth-friendly fiscal consolidation (Namibia):
  - Revenue-and-expenditure-based consolidation (using consumption taxes and expenditure measures) lowers negative growth effects relative to pure expenditure-based adjustment.
  - Further gains if combined with: (i) expenditure composition shift toward public investment; (ii) public financial management reforms raising investment efficiency; (iii) improved tax collection efficiency.
- Natural disasters and resilience (Vanuatu):
  - Two public capital types introduced: standard and adaptation infrastructure; adaptation can raise marginal product of private capital and crowd in private investment.
  - Tax-financed reconstruction can reduce private consumption and investment by as much as 20 and 40 percent, respectively.
  - Debt-financed reconstruction leads to unsustainable debt paths; viable option requires a “big push” in grants—on average an additional 50 percent of pre-cyclone grants levels over 15 years after the disaster.
- Security vs infrastructure (Afghanistan):
  - Cuts in foreign aid reduce public investment and security spending, compounding negative output effects; issuing domestic debt protects investment but worsens private investment and consumption and can make public finances unsustainable; domestic revenue mobilization reforms needed to secure outcomes.
- Governance reforms (EMDE calibration):
  - Weak governance modeled via: output losses proportional to production, reduced public investment efficiency, and tax revenue leakage.
  - Improving governance from first to second quartile over 10 years:
    - Private investment 70 percent higher after 10 years.
    - Output and private consumption increase by about 30 percent.
    - Public debt falls by almost 15 percent of GDP after 10 years, allowing tax rate reductions and further private demand stimulus.
- Operations & Maintenance (O&M):
  - Deficient O&M causes higher public capital depreciation and lower output from existing public capital.
  - Public investment surges not matched by commensurate O&M produce substantially lower growth rates.

### Policy recommendations and lessons
- Design investment scaling-ups mindful of pace and absorptive capacity; very ambitious, front-loaded plans can worsen medium-term growth and debt sustainability.
- Prioritize reforms to raise public investment efficiency and rates of return on projects to maximize growth pay-offs and lower debt risks.
- Strengthen domestic revenue mobilization (tax administration, broaden tax base, prioritize current expenditures, reduce subsidies) to create fiscal space and contain debt surges.
- Use targeted compensatory measures (cash transfers or public investment) when raising distortionary taxes (e.g., VAT) to protect poor consumers and support non-oil GDP.
- Favor concessional aid/grants and concessional loans where available to smooth fiscal adjustments, reduce need for painful domestic fiscal consolidation, and stabilize debt trajectories.
- Save a portion of resource revenues in sovereign wealth funds to mitigate Dutch disease and build fiscal buffers for adverse commodity shocks.
- Carefully weigh financing choices: domestic bank borrowing can crowd out private investment and raise inflation; external commercial borrowing can avoid short-term macro adjustments but raises external public debt.
- Incorporate sectoral extensions (energy, human capital, resilience, governance) in macro frameworks to capture project-specific trade-offs and distributional consequences.
- Ensure adequate operations and maintenance funding alongside new investment to realize projected growth gains.
- Use model-based frameworks (DIG, DIGNAR) to quantify impacts, test scenarios, inform DSA adjustments, and structure policy discussions; continuous sensitivity analysis and mission-team engagement are critical for credible applications.

*Source: DIGNAR and DIG country applications and model-based analyses as presented in the supplied text.*

### REFERENCES

### wpiea2019062 - REFERENCES

### Content overview
- Contains a bibliography of works cited in the study, including IMF Working Papers, IMF Country Reports, IMF Policy Papers, journal articles, NBER and HEI working papers, and other academic and policy literature.
- Bibliographic material spans multiple topics central to the main study: public investment, debt sustainability, resource wealth management, infrastructure and growth, political economy of investment, DSGE modeling for developing economies, and debt-relief initiatives (e.g., HIPC).

### Thematic coverage in the referenced literature
- Public investment, growth, and debt sustainability:
  - Multiple IMF Working Papers and articles addressing public investment surges, public investment efficiency, and the interaction between public investment and debt dynamics.
  - Modeling approaches cited include the DIGNAR model and DIG/DIGNA(R) frameworks applied to developing and resource-rich countries.
- Infrastructure and productivity:
  - References discuss the role of infrastructure capital, its utilization, and impacts on economic growth.
- Political economy and governance:
  - Literature on political myopia, electoral incentives, and governance guidance for Fund engagement.
- Debt sustainability and policy frameworks:
  - Works on debt sustainability assessment methodologies, HIPC Initiative evaluation, and fiscal space in advanced and low-income economies.
- Modeling methods and DSGE approaches:
  - Citations include discussions of DSGE modeling relevance and applications to African and other developing economies.
- Country-focused analyses:
  - Numerous IMF country-specific studies and selected issues papers informing the broader analysis.

### Appendix: country applications (summary)
- Includes a table of country model applications by Year, Country, Model, Publication Type, and Web Link.
- Time span covered: 2011 through 2018.
- Models and publication types recorded include DIG, DIGNAR, DI, Staff Report Annex, IMF Working Paper, Selected Issues Paper, WEO chapter, and departmental papers.
- The appendix documents country applications across a broad set of low-income and developing countries and regional groupings.

*Source: wpiea2019062 - REFERENCES (IMF).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019062.pdf_