## wpiea2021117-print-pdf

## Source details

**Canonical URL:** [wpiea2021117-print-pdf](https://www.imf.org/-/media/files/publications/wp/2021/english/wpiea2021117-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2021/english/wpiea2021117-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2021/english/wpiea2021117-print-pdf.pdf.json)

---

### Financing options and overview
- Option 1: Tax financing through higher consumption tax (budget-neutral reallocation of indirect tax revenue—equivalent to a VAT). For this set of countries, a 1 percent of GDP increase in consumption tax would imply a hike in tax rate of about 1.4 to 1.7 percentage points.
- Option 2: Debt financing through financial markets (increase in public investment fully financed by market borrowing). Simulations assume a rise in the risk premium by 3 basis points per unit increase in the ratio of public debt to GDP. The change in the risk premium is assumed the same for all countries in the sample; impact on interest costs is larger for countries with a higher initial debt level.
- Option 3: Debt financing through concessional loans (rise in public investment fully financed by concessional loans granted with below-market interest rates). In Option 3, the FSGM simulations assume the risk premium does not respond to the rise in the debt-to-GDP ratio.
- Option 4: Private sector financing through public-private partnerships (PPPs) — user-funded PPPs where private concessionaires fully recover costs via user fees; model assumes PPPs have similar costs to traditional public investment and users are willing to pay fees.
- Option 4b: Variant of Option 4 where PPP public investment efficiency is reduced by 10 percent.

### Model and methodological notes
- Analysis uses the Asia and Pacific Department Model (APDMOD), a module of the Flexible System of Global Models (FSGM) with individual blocks for 15 Asian countries and 9 additional regions.
- Project costs are assumed the same across options. SOE financing not modeled due to inadequate cross-country SOE data.
- In the FSGM, market-financing is implemented by adjusting the fiscal deficit target to the additional discretionary spending; general lumpsum transfers adjust to cover increased debt-service costs.
- The assumption of 3 basis points per unit increase in debt-to-GDP for the risk premium is based on Kumar and Baldacci (2010); value chosen is somewhat conservative.

### How scaling up infrastructure investment bolsters output — key simulation findings
- General result: Increase in public investment lifts real GDP in the short run and long run across all four financing options; magnitude varies with financing option and initial conditions (especially government investment efficiency).
- Long-run (year 10) output gains from a 1 percent of GDP increase in public investment (selected ranges reported):
  - Option 1 (consumption tax-financed): cumulative increases in real GDP between around 1.6 percent and 2.7 percent in the long run. Output after twenty years: about 3.5 percent to 5.2 percent.
  - Option 2 (market-financed): real GDP gains comprised between 1.5 percent and 2.6 percent in the long run. Output after twenty years: about 3.2 percent to 5.0 percent.
  - Option 3 (concessional financing): overall real GDP is 2 percent to around 3 percent higher in the long run. Output after twenty years: about 3.9 percent to 5.6 percent.
  - Option 4 (PPPs baseline efficiency): level of GDP is between about 1.9 percent and 2.7 percent higher in the long run. Output after twenty years: about 3.8 percent to 5.4 percent.
- Short-term dynamics by option:
  - Option 1: Higher consumption taxes discourage private consumption in the short term, partially offsetting stimulative impact; long-run effects positive via higher public capital stock, higher marginal productivity, stimulated private investment, and raised labor demand.
  - Option 2: Market-financed option raises borrowing costs over time, crowds out private investment and depresses private capital stock; private consumption is higher and labor supply expands more than in tax-financed variant due to households working more to offset lower transfers.
  - Option 3: Concessional financing contains borrowing costs, does not significantly crowd out private investment; private consumption increases substantially in short to medium term as households work more to offset lower transfers.
  - Option 4: PPPs increase output through multiplier effects; private consumption initially somewhat dampened by lower targeted transfers (user fees); private investment and labor demand rise sharply, boosting medium- to long-run private consumption.
- Cross-country heterogeneity:
  - Countries with relatively lower initial public debt ratio and higher government investment efficiency record the largest cumulative real GDP increase.
  - Bangladesh displays the biggest output gains under all four options given higher calibrated efficiency and lower initial public debt-to-GDP.

### Growth-debt trade-offs (quantified)
- Option 1 (budget-neutral, consumption tax): general government balance unchanged by construction; debt-to-GDP ratios improve slightly in the short-to-medium term and converge to steady state in the long run.
- Option 2 (market financing): debt-to-GDP ratios increase by about 7-8 percentage points in most selected economies in the long run; higher public debt raises borrowing costs (higher risk premium).
- Option 3 (concessional financing): public debt ratios increase by about 6-7 percentage points in the long run (slightly more contained than Option 2).
- Option 4 (PPPs): government balance remains unchanged; debt-to-GDP ratios decline slightly in the short to medium term (reflecting higher output) and converge to steady state in the long run.

### Optimal financing choice — main conclusions and caveats
- Tax financing (Option 1), concessional lending (Option 3), or PPP financing (Option 4) are more advantageous than market financing (Option 2) in addressing the long-run growth-debt trade-off.
- Short-to-medium term: concessional resources (Option 3) yield higher output gains than market funding (Option 2); concessional lending also generates slightly higher output than consumption tax-financing (Option 1) and PPPs (Option 4). However, Option 3 increases public debt and can raise fiscal vulnerability; access to concessional financing may be limited in the wake of the pandemic.
- Long term: relative advantage of concessional financing over PPPs diminishes; Option 4 achieves a relatively higher output boost with no increase in public debt-to-GDP ratio, while Option 3 raises public debt. PPPs appear more appealing than VAT-financing for addressing the long-term growth-debt trade-off, especially where raising taxes is politically challenging post-Covid-19.
- Key caveat: PPPs entail implementation risks and potential higher costs than traditional public investment; larger growth payoffs under PPPs are feasible only if projects are well-designed and well-implemented and significantly more efficient. Lower PPP efficiency (e.g., 10 percent lower calibrated public investment efficiency) would dampen long-run GDP gains (Option 4b: about 1.7 to 2.4 percent in the long run versus 1.9 to 2.7 percent under baseline PPP efficiency).

### Infrastructure delivery institutions: SOEs, general government, and private participation
- Regional shares in 2017 (South Asia context):
  - SOEs carry out 44 percent of infrastructure investment, general government 40 percent, and private participation 15 percent.
- Public investment magnitudes (2017, South Asia averages reported in figures):
  - Annual general government public investment averaged 8 percent of GDP in 2017 across countries in South Asia.
  - Yearly PPP investments remained relatively small at 1 percent of GDP.
- Sectoral patterns: SOEs and private participation are prominent in transport and energy sectors; SOEs also invest in water infrastructure.

### PPPs: benefits, costs, and fiscal risks
- Potential benefits:
  - Attract private capital, supplement public resources, or release resources for other public needs; private sector compensated through fees.
  - Increase efficiency and use resources more effectively; reform sectors via reallocation of roles, incentives, and accountability.
  - Private-sector management and innovation may deliver better quality and lower cost services in some cases.
- Potential additional costs and risks:
  - Private sector borrowing costs often higher than public sector; PPPs can involve higher transaction costs (complex negotiation, implementation, and frequent renegotiation).
  - PPPs may be more expensive than traditional public investment in some studies; user fees needed to finance PPPs may have to be higher than taxes collected to finance traditional government investment.
  - Users’ willingness to pay varies; political feasibility of cost-based fees can be limited, leading to explicit (availability payments) or implicit (minimum revenue guarantees) government contributions.
  - Fiscal risks and contingent liabilities include capital subsidies, volume-based payments, tax incentives, and payments related to guarantees on revenue, exchange-rate, interest-rate or debt. Implicit liabilities from poor contract management or renegotiation add fiscal risk.

### Public-Private Partnerships (PPPs): definition, risks, and fiscal management
- Definition: A PPP, as defined by World Bank (2014a), is “a long-term contract between a private party and a government entity, for providing a public asset or service, in which the private party bears significant risk and management responsibility, and remuneration is linked to performance.”
- Risk characteristics:
  - Infrastructure projects entail risks related to land acquisition, environmental, and other clearances.
  - PPP contracts allocate these risks between the government and private sector; complexities arise in this allocation.
  - Unlike traditional government-financed projects that can be rolled back, PPPs create long-term binding commitments that cannot be scaled down once entered into.
- Fiscal integrity concerns:
  - If PPP costs and risks are not adequately disclosed, they could threaten the integrity of the budget and complicate fiscal discipline and good governance (IMF, 2004; Akitoby et al. 2007).
- Essential steps to manage PPP fiscal risks:
  - Sound PPP framework: policies, procedures, institutions, and rules that define how PPPs will be identified, planned, assessed, selected, budgeted, procured, monitored, and accounted.
  - Controlling costs (Irwin et al. 2018 detail these elements):
    - (1) establish a gateway process managed by the ministry of finance;
    - (2) develop a framework for risk sharing, where the government bears only those risks that it strongly influences;
    - (3) establish clear lines of accountability, combining central review of major commitments with decentralization of smaller decisions and contract monitoring;
    - (4) impose limits on the total sum of commitments on PPPs.
  - Disclosure of costs and risks:
    - Maintain an inventory of all PPP projects with proper fiscal accounting of their full lifetime costs.
    - Aim at full disclosure in the budget documents of current government commitments (including guarantees) and expected budgetary costs of existing PPP contracts.
    - Assess new PPPs within the budget process and medium-term fiscal framework to check for fiscal affordability.
    - Disclose fiscal risks created by PPPs in statements of fiscal risk (IMF, 2014; IMF 2018b).

### State-Owned Enterprises (SOEs): role, challenges, and governance reforms
- Rationale for SOE involvement:
  - Government infrastructure investment through SOEs is often justified to correct market failures (for example, natural monopoly characteristics in network industries such as energy, transportation, and water and sewer systems).
  - Governments may rely on SOEs when they can raise financing independently or to keep investment off the government’s balance sheet.
- Performance and governance issues:
  - Profits and labor productivity have been lower in SOEs than in private firms (IMF, 2020d); this can be partly attributed to weak governance and the cost of unfunded mandates (e.g., below-cost service provision or excess employment).
  - Inefficient and cash-strapped SOEs are unable to adequately maintain or expand infrastructure networks and may lack transparency and accountability.
- Fiscal costs:
  - Inefficient SOEs impose substantial fiscal costs via government guarantees, subsidies, soft loans, or capital injections.
  - IMF (2020d) finds that maximum annual support to nonfinancial SOEs reached 16 percent of GDP.
- Recommended reforms:
  - Improve financial oversight of SOEs through periodic monitoring of SOE financial performance.
  - Cost the delivery of quasi-fiscal activities in the annual budget and disclose them in financial reports (Allen and Alves, 2016; IMF Fiscal Transparency Code, 2014).
  - Promote objective pricing of infrastructure assets and services.
  - Implement effective supervision of public enterprises’ investment plans.
  - Reduce internal inefficiencies to improve profitability and the effectiveness of infrastructure spending.

### Improving Public Investment Efficiency (PIM)
- Rationale: Increasing the efficiency of public investment raises growth payoffs from public infrastructure spending.
- PIMA results for South Asia:
  - The estimated efficiency score for South Asia is around 50 percent of the best performing peer countries and is below other emerging and developing countries in Asia.
  - Reducing the public investment efficiency gap could increase long-run growth.
- Key PIM weaknesses identified in South Asia sample:
  - Project appraisal:
    - All sampled countries have weak mechanisms to effectively appraise investment project proposals before selection.
    - Lack of a standard appraisal methodology undermines project quality, causes delays, and raises the risk of selecting less worthy projects.
  - Project selection:
    - Lack of consistent criteria to identify and select investment projects and to quantify their costs.
    - None of the sampled countries have an effective review of major projects by a central ministry before inclusion in the budget.
    - Recommendation: create a pipeline of appraised, adequately costed projects for subsequent budget consideration and define clear, transparent criteria for project prioritization.
  - Maintenance funding:
    - Existing mechanisms for maintenance funding are not very effective; budgeted maintenance levels are low compared to the capital stock.
    - Most countries lack a standard methodology or sufficiently disaggregated data to estimate routine and major capital maintenance needs.
  - Multiyear budgeting:
    - Most sampled countries do not have a medium-term framework for capital spending to guide multiyear investment planning.
    - Countries do not publish projections of overall or disaggregated capital spending by ministry or sector over the medium term.
    - Financing of project outlays is frequently subject to budget cuts and cash rationing, leading to arrears and implementation delays.

### Overall conclusions for South Asia
- Greater infrastructure spending can support the Covid-19 recovery and raise growth over the medium term, but benefits depend on financing choices and management.
- Countries need to both extend infrastructure networks and improve their quality.
- Financing strategy (taxes, debt, or PPPs) will affect outcomes and is country specific—dependent on available fiscal space, taxation capacity, and public sector efficiency.
- Alternatives such as SOEs and PPPs can offer advantages in efficiency but bring implementation and fiscal risks that must be monitored, managed, and disclosed.
- Reforms to strengthen public investment management are necessary to maximize returns from additional infrastructure spending.

### Annex: Estimating the growth impact of infrastructure development — methodological highlights
- Empirical model (drawing on Calderón and Servén, 2004 and 2008):
  - Dependent variable: change in aggregate output per worker in logs over time for country i in period t.
  - Controls (X): human capital (secondary enrollment in percent of total with the age above 15, from Barro and Lee, 2013), financial development (domestic credit to private sector in percent of GDP), trade openness (trade in percent of GDP), inflation, government burden (government final consumption expenditure in percent of GDP), institutional quality (ICRG political Risk Index), terms of trade and their changes, and size of the modern (non-agricultural) sector in terms of value added; all expressed in logs unless otherwise specified.
  - Infrastructure variable (K) captures infrastructure measures.
  - Time effects and country fixed effects considered.
  - Estimation sample: panel of 80 countries for the period 1990−2017 using non-overlapping 5-year period averages.
- Construction of infrastructure indices:
  - Index of infrastructure stock:
    - First principal component of electricity generating capacity (in megawatts per 1000 workers), length of road network (in kilometers per a square meter of land area), and internet access (share of households with internet access), all in logs and normalized.
    - First PC weights: 0.61 (electricity), 0.54 (transport), 0.58 (telecommunication); captures 67 percent of overall variation; correlations with underlying variables range 0.75−0.83.
  - Index for quality:
    - First principal component of share of electricity transmitted and distributed to consumers (one minus losses), share of paved roads to total, and internet bandwidth per user rescaled between 0 and 1.
    - First PC weights: 0.59 (electricity quality), 0.61 (transport quality), 0.53 (telecommunication quality); captures 51 percent of overall variation; correlations with underlying variables range 0.7−0.8.
- Estimation approaches:
  - Various methods considered, including GMM estimation developed by Arellano and Bond (1991) to address endogeneity and bias from correlation between the lagged dependent variable and fixed effects.
  - Results confirm that infrastructure indices (both quantity and quality) have a positive and significant relationship with growth.

*Source: IMF staff analysis and FSGM/APDMOD simulations as presented in the provided content.*

### 1. Option 1: Tax financing through higher consumption tax. The f iscal cost is f inanced

### 1. Option 1: Tax financing through higher consumption tax. The f iscal cost is f inanced

### Financing options (overview)
- Option 1: Tax financing through higher consumption tax (budget-neutral reallocation of indirect tax revenue—equivalent to a VAT). For this set of countries, a 1 percent of GDP increase in consumption tax would imply a hike in tax rate of about 1.4 to 1.7 percentage points.
- Option 2: Debt financing through financial markets (increase in public investment fully financed by market borrowing). Simulations assume a rise in the risk premium by 3 basis points per unit increase in the ratio of public debt to GDP. The change in the risk premium is assumed the same for all countries in the sample; impact on interest costs is larger for countries with a higher initial debt level.
- Option 3: Debt financing through concessional loans (rise in public investment fully financed by concessional loans granted with below-market interest rates). In Option 3, the FSGM simulations assume the risk premium does not respond to the rise in the debt-to-GDP ratio.
- Option 4: Private sector financing through public-private partnerships (PPPs) — user-funded PPPs where private concessionaires fully recover costs via user fees; model assumes PPPs have similar costs to traditional public investment and users are willing to pay fees. A variant (Option 4b) assumes PPP public investment efficiency reduced by 10 percent.

### Model and methodological notes
- Analysis uses the Asia and Pacific Department Model (APDMOD), a module of the Flexible System of Global Models (FSGM) with individual blocks for 15 Asian countries and 9 additional regions.
- Project costs are assumed the same across options. SOE financing not modeled due to inadequate cross-country SOE data.
- In the FSGM, market-financing is implemented by adjusting the fiscal deficit target to the additional discretionary spending; general lumpsum transfers adjust to cover increased debt-service costs.
- The assumption of 3 basis points per unit increase in debt-to-GDP for the risk premium is based on Kumar and Baldacci (2010); value chosen is somewhat conservative.

### How scaling up infrastructure investment bolsters output — key simulation findings
- General result: Increase in public investment lifts real GDP in the short run and long run across all four financing options; magnitude varies with financing option and initial conditions (especially government investment efficiency).
- Long-run (year 10) output gains from a 1 percent of GDP increase in public investment (selected ranges reported):
  - Option 1 (consumption tax-financed): cumulative increases in real GDP between around 1.6 percent and 2.7 percent in the long run. Output after twenty years: about 3.5 percent to 5.2 percent (note: twenty-year figures provided separately).
  - Option 2 (market-financed): real GDP gains comprised between 1.5 percent and 2.6 percent in the long run. Output after twenty years: about 3.2 percent to 5.0 percent.
  - Option 3 (concessional financing): overall real GDP is 2 percent to around 3 percent higher in the long run. Output after twenty years: about 3.9 percent to 5.6 percent.
  - Option 4 (PPPs baseline efficiency): level of GDP is between about 1.9 percent and 2.7 percent higher in the long run. Output after twenty years: about 3.8 percent to 5.4 percent.
- Short-term dynamics:
  - Option 1: Higher consumption taxes discourage private consumption in the short term, partially offsetting stimulative impact; long-run effects positive via higher public capital stock, higher marginal productivity, stimulated private investment, and raised labor demand.
  - Option 2: Market-financed option raises borrowing costs over time, crowds out private investment and depresses private capital stock; private consumption is higher and labor supply expands more than in tax-financed variant due to households working more to offset lower transfers.
  - Option 3: Concessional financing contains borrowing costs, does not significantly crowd out private investment; private consumption increases substantially in short to medium term as households work more to offset lower transfers.
  - Option 4: PPPs increase output through multiplier effects; private consumption initially somewhat dampened by lower targeted transfers (user fees); private investment and labor demand rise sharply, boosting medium- to long-run private consumption.

- Cross-country heterogeneity: Countries with relatively lower initial public debt ratio and higher government investment efficiency record the largest cumulative real GDP increase. Bangladesh displays the biggest output gains under all four options given higher calibrated efficiency and lower initial public debt-to-GDP.

### Growth-debt trade-offs (quantified)
- Option 1 (budget-neutral, consumption tax): general government balance unchanged by construction; debt-to-GDP ratios improve slightly in the short-to-medium term and converge to steady state in the long run.
- Option 2 (market financing): debt-to-GDP ratios increase by about 7-8 percentage points in most selected economies in the long run; higher public debt raises borrowing costs (higher risk premium).
- Option 3 (concessional financing): public debt ratios increase by about 6-7 percentage points in the long run (slightly more contained than Option 2).
- Option 4 (PPPs): government balance remains unchanged; debt-to-GDP ratios decline slightly in the short to medium term (reflecting higher output) and converge to steady state in the long run.

### Optimal financing choice — main conclusions and caveats
- Tax financing (Option 1), concessional lending (Option 3), or PPP financing (Option 4) are more advantageous than market financing (Option 2) in addressing the long-run growth-debt trade-off.
- Short-to-medium term: concessional resources (Option 3) yield higher output gains than market funding (Option 2); concessional lending also generates slightly higher output than consumption tax-financing (Option 1) and PPPs (Option 4). However, Option 3 increases public debt and can raise fiscal vulnerability; access to concessional financing may be limited in the wake of the pandemic.
- Long term: relative advantage of concessional financing over PPPs diminishes; Option 4 achieves a relatively higher output boost with no increase in public debt-to-GDP ratio, while Option 3 raises public debt. PPPs appear more appealing than VAT-financing for addressing the long-term growth-debt trade-off, especially where raising taxes is politically challenging post-Covid-19.
- Key caveat: PPPs entail implementation risks and potential higher costs than traditional public investment; larger growth payoffs under PPPs are feasible only if projects are well-designed and well-implemented and significantly more efficient. PPPs pose fiscal management challenges and fiscal risks; lower PPP efficiency (e.g., 10 percent lower calibrated public investment efficiency) would dampen long-run GDP gains (Option 4b: about 1.7 to 2.4 percent in the long run versus 1.9 to 2.7 percent under baseline PPP efficiency).

### Infrastructure delivery institutions: SOEs, general government, and private participation
- Regional shares in 2017 (South Asia context):
  - SOEs carry out 44 percent of infrastructure investment, general government 40 percent, and private participation 15 percent (Figure 2.7 referenced).
- Public investment magnitudes (2017, South Asia averages reported in figures):
  - Annual general government public investment averaged 8 percent of GDP in 2017 across countries in South Asia.
  - Yearly PPP investments remained relatively small at 1 percent of GDP.
- Sectoral patterns: SOEs and private participation are prominent in transport and energy sectors; SOEs also invest in water infrastructure.

### PPPs: benefits, costs, and fiscal risks
- Potential benefits of PPPs:
  - Attract private capital, supplement public resources, or release resources for other public needs; private sector compensated through fees.
  - Increase efficiency and use resources more effectively; reform sectors via reallocation of roles, incentives, and accountability.
  - Private-sector management and innovation may deliver better quality and lower cost services in some cases.

- Potential additional costs and risks:
  - Private sector borrowing costs often higher than public sector; PPPs can involve higher transaction costs (complex negotiation, implementation, and frequent renegotiation).
  - PPPs may be more expensive than traditional public investment in some studies; user fees needed to finance PPPs may have to be higher than taxes collected to finance traditional government investment.
  - Users’ willingness to pay varies; political feasibility of cost-based fees can be limited, leading to explicit (availability payments) or implicit (minimum revenue guarantees) government contributions.
  - Fiscal risks and contingent liabilities include capital subsidies, volume-based payments, tax incentives, and payments related to guarantees on revenue, exchange-rate, interest-rate or debt. Implicit liabilities from poor contract management or renegotiation add fiscal risk.
  - Empirical evidence: Bova et al. (2016) find average fiscal cost of realized PPP contingencies in advanced and emerging market economies (reference to study; quantitative value not included in provided excerpt).

*Source: IMF staff analysis and FSGM/APDMOD simulations as presented in the provided content.*

### 1.2 percent of GDP. Moreover, if public investment through PPP projects are not adequately

### wpiea2021117-print-pdf - 1.2 percent of GDP. Moreover, if public investment through PPP projects are not adequately

### Public-Private Partnerships (PPPs): definition, risks, and fiscal management
- Definition: A PPP, as defined by World Bank (2014a), is “a long-term contract between a private party and a government entity, for providing a public asset or service, in which the private party bears significant risk and management responsibility, and remuneration is linked to performance.”
- Risk characteristics:
  - Infrastructure projects entail risks related to land acquisition, environmental, and other clearances.
  - PPP contracts allocate these risks between the government and private sector; complexities arise in this allocation.
  - Unlike traditional government-financed projects that can be rolled back, PPPs create long-term binding commitments that cannot be scaled down once entered into.
- Fiscal integrity concerns:
  - If PPP costs and risks are not adequately disclosed, they could threaten the integrity of the budget and complicate fiscal discipline and good governance (IMF, 2004; Akitoby et al. 2007).
- Essential steps to manage PPP fiscal risks:
  - Sound PPP framework:
    - Policies, procedures, institutions, and rules that define how PPPs will be identified, planned, assessed, selected, budgeted, procured, monitored, and accounted.
  - Controlling costs (Irwin et al. 2018 detail these elements):
    - (1) establish a gateway process managed by the ministry of finance;
    - (2) develop a framework for risk sharing, where the government bears only those risks that it strongly influences;
    - (3) establish clear lines of accountability, combining central review of major commitments with decentralization of smaller decisions and contract monitoring;
    - (4) impose limits on the total sum of commitments on PPPs.
  - Disclosure of costs and risks:
    - Maintain an inventory of all PPP projects with proper fiscal accounting of their full lifetime costs.
    - Aim at full disclosure in the budget documents of current government commitments (including guarantees) and expected budgetary costs of existing PPP contracts.
    - Assess new PPPs within the budget process and medium-term fiscal framework to check for fiscal affordability.
    - Disclose fiscal risks created by PPPs in statements of fiscal risk (IMF, 2014; IMF 2018b).

### State-Owned Enterprises (SOEs): role, challenges, and governance reforms
- Rationale for SOE involvement:
  - Government infrastructure investment through SOEs is often justified to correct market failures (for example, natural monopoly characteristics in network industries such as energy, transportation, and water and sewer systems).
  - Governments may rely on SOEs when they can raise financing independently or to keep investment off the government’s balance sheet.
- Performance and governance issues:
  - Profits and labor productivity have been lower in SOEs than in private firms (IMF, 2020d); this can be partly attributed to weak governance and the cost of unfunded mandates (e.g., below-cost service provision or excess employment).
  - Inefficient and cash-strapped SOEs are unable to adequately maintain or expand infrastructure networks and may lack transparency and accountability.
- Fiscal costs:
  - Inefficient SOEs impose substantial fiscal costs via government guarantees, subsidies, soft loans, or capital injections.
  - IMF (2020d) finds that maximum annual support to nonfinancial SOEs reached 16 percent of GDP.
- Recommended reforms to reduce fiscal risks and improve infrastructure outcomes:
  - Improve financial oversight of SOEs through periodic monitoring of SOE financial performance.
  - Cost the delivery of quasi-fiscal activities in the annual budget and disclose them in financial reports (Allen and Alves, 2016; IMF Fiscal Transparency Code, 2014).
  - Promote objective pricing of infrastructure assets and services.
  - Implement effective supervision of public enterprises’ investment plans.
  - Reduce internal inefficiencies to improve profitability and the effectiveness of infrastructure spending.

### Improving Public Investment Efficiency (PIM)
- Rationale:
  - Increasing the efficiency of public investment raises growth payoffs from public infrastructure spending.
- PIMA results for South Asia:
  - Public Investment Management Assessment (PIMA) results suggest South Asia has significant room to enhance public investment efficiency.
  - The estimated efficiency score for South Asia is around 50 percent of the best performing peer countries and is below other emerging and developing countries in Asia (Figure 2.10).
  - Reducing the public investment efficiency gap could increase long-run growth (as shown in model-based simulations and IMF (2015)).
- Key PIM weaknesses identified in South Asia sample (PIMA):
  - Project appraisal:
    - All sampled countries have weak mechanisms to effectively appraise investment project proposals before selection.
    - Lack of a standard appraisal methodology undermines project quality, causes delays, and raises the risk of selecting less worthy projects.
  - Project selection:
    - Major weakness: lack of consistent criteria to identify and select investment projects and to quantify their costs.
    - None of the sampled countries have an effective review of major projects by a central ministry before inclusion in the budget.
    - Recommendation: create a pipeline of appraised, adequately costed projects for subsequent budget consideration and define clear, transparent criteria for project prioritization.
  - Maintenance funding:
    - Existing mechanisms for maintenance funding are not very effective; budgeted maintenance levels are low compared to the capital stock.
    - Most countries lack a standard methodology or sufficiently disaggregated data to estimate routine and major capital maintenance needs.
  - Multiyear budgeting:
    - Most sampled countries do not have a medium-term framework for capital spending to guide multiyear investment planning.
    - Countries do not publish projections of overall or disaggregated capital spending by ministry or sector over the medium term.
    - Financing of project outlays is frequently subject to budget cuts and cash rationing, leading to arrears and implementation delays.

### Overall conclusions for South Asia
- Greater infrastructure spending can support the Covid-19 recovery and raise growth over the medium term, but benefits depend on financing choices and management.
- Countries need to both extend infrastructure networks and improve their quality.
- Financing strategy (taxes, debt, or PPPs) will affect outcomes and is country specific—dependent on available fiscal space, taxation capacity, and public sector efficiency.
- Alternatives such as SOEs and PPPs can offer advantages in efficiency but bring implementation and fiscal risks that must be monitored, managed, and disclosed.
- Reforms to strengthen public investment management are necessary to maximize returns from additional infrastructure spending.

### Annex: Estimating the growth impact of infrastructure development — methodological highlights
- Empirical model (drawing on Calderón and Servén, 2004 and 2008):
  - Dependent variable: change in aggregate output per worker in logs over time for country i in period t.
  - Controls (X): human capital (secondary enrollment in percent of total with the age above 15, from Barro and Lee, 2013), financial development (domestic credit to private sector in percent of GDP), trade openness (trade in percent of GDP), inflation, government burden (government final consumption expenditure in percent of GDP), institutional quality (ICRG political Risk Index), terms of trade and their changes, and size of the modern (non-agricultural) sector in terms of value added; all expressed in logs unless otherwise specified.
  - Infrastructure variable (K) captures infrastructure measures.
  - Time effects and country fixed effects considered.
  - Estimation sample: panel of 80 countries for the period 1990−2017 using non-overlapping 5-year period averages.
- Construction of infrastructure indices:
  - Use aggregate synthetic indices (first principal component) to capture quantity and quality of infrastructure due to high correlation among individual measures (bilateral correlations range between 0.5 and 0.9).
  - Index of infrastructure stock:
    - First principal component of electricity generating capacity (in megawatts per 1000 workers), length of road network (in kilometers per a square meter of land area), and internet access (share of households with internet access), all in logs and normalized.
    - First PC weights: 0.61 (electricity), 0.54 (transport), 0.58 (telecommunication); captures 67 percent of overall variation; correlations with underlying variables range 0.75−0.83.
  - Index for quality:
    - First principal component of share of electricity transmitted and distributed to consumers (one minus losses), share of paved roads to total, and internet bandwidth per user rescaled between 0 and 1.
    - First PC weights: 0.59 (electricity quality), 0.61 (transport quality), 0.53 (telecommunication quality); captures 51 percent of overall variation; correlations with underlying variables range 0.7−0.8.
- Estimation approaches:
  - Various methods considered, including GMM estimation developed by Arellano and Bond (1991) to address endogeneity and bias from correlation between the lagged dependent variable and fixed effects.
  - Results confirm that infrastructure indices (both quantity and quality) have a positive and significant relationship with growth (as reported in Table 2.1).

*Source: wpiea2021117-print-pdf - 1.2 percent of GDP. Moreover, if public investment through PPP projects are not adequately*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2021/english/wpiea2021117-print-pdf.pdf_