## wpiea2023031-print-pdf

## Source details

**Canonical URL:** [wpiea2023031-print-pdf](https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023031-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2023/english/wpiea2023031-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2023/english/wpiea2023031-print-pdf.pdf.json)

---

### Key findings and context
- Public infrastructure quality (2019 Global Competitiveness Index): Western Balkans ranked 78 (out of 141) versus EU-15 at 17 and CESEE at 51.
- Country variation (2019 infrastructure quality): Serbia ranked 51; Albania ranked 98 (out of 141).
- Public debt levels as of 2021: range between 36 percent of GDP in Bosnia and Herzegovina and 87 percent of GDP in Montenegro.
- Regional average public debt: increased from 54 percent in 2019 to a projected 62 percent in 2022.
- Historical PPP investments in Western Balkans since 2012: averaged around 0.5 percent of GDP—higher than the EU and global averages for the same period.

### PPP rationale, benefits, and risks
- Definition: PPPs are long-term, risk-sharing contracts where the private party typically invests in an asset and provides services subject to performance criteria.
- PPP types by funding source:
  - Government-funded or availability-based PPPs: government makes predetermined payments for services over the contract duration or payments per volume of service.
  - User-funded PPPs (concessions): users pay fees for services; sizeable government spending may still be needed to subsidize investment or guarantee private debt/revenue.
  - Mixed PPPs combine elements of both.
- Potential benefits:
  - Increase short-term output via aggregate demand and longer-term output by expanding productive capacity.
  - Contribute to job creation.
  - Help meet energy transition goals by pooling capital and private-sector technological expertise.
- Fiscal and other risks:
  - PPPs can introduce large fiscal risks and increased costs if poorly designed or managed.
  - PPPs may create “fiscal illusion” because long-term cashflow profiles can mask fiscal commitments.
  - Risks can materialize during shocks (e.g., COVID-19 pandemic, war in Ukraine).
  - Historical disruptions: EU PPP transaction values declined by 50 percent from 2007 to 2009 during the GFC; PPP project counts dropped around 40–50 percent during the GFC and Euro Area crisis.

### Stylized facts and global/regional trends
- Global median PPP capital stock:
  - 0.6 percent of GDP in 2001
  - 1.8 percent of GDP in 2011
  - 2.5 percent of GDP in 2019
- At end-2019:
  - Emerging markets PPP capital stock as share of GDP: about 3.7 percent (driven by Asian emerging markets).
  - Advanced economies: largest PPP capital stock in per capita terms.
  - Ten European economies among top 20 countries by per capita PPP capital stock.
- Median annual PPP investments:
  - Peaked at 0.2 percent of GDP in 2010; declined to about 0.1 percent of GDP thereafter.
  - Low-income countries: median PPP investments around 0.2 percent of GDP over the past decade.
- Sectoral patterns:
  - Europe: PPPs concentrated in transportation by project value; education and healthcare dominated by number of projects in the mid-2000s.
  - Emerging market and developing economies: PPP investment concentrated in energy, followed by transportation.
  - Since 2020: number and value of user-funded PPP projects in Europe increased.

### PPP prevalence and scale in the Western Balkans
- PPP capital stock as share of GDP and total capital stock: rising since 2005 with sharp acceleration in 2012/13.
- As of 2019: Western Balkans share higher than EU-15 or CESEE.
- Annual PPP investment as share of GDP: volatile but grown since 2005, particularly in Albania, Bosnia, and Serbia.
- Implication: individual PPP project failure could create a larger burden for these countries.
- Data caveat: coverage of PPP projects drawn from different sources might vary across countries.

### Conceptual arguments: efficiency, fiscal space, and fiscal risks
- Efficiency:
  - Pro-PPP: PPPs can complement public provision and generate efficiency gains that can offset higher private financing costs.
  - Anti-PPP: Complexity increases costs; procurement costs can reach 5-10 percent of capital cost for large projects and do not decline proportionately for smaller projects.
- Fiscal space:
  - Claimed benefit: private financing presents as creating fiscal space by avoiding upfront government debt.
  - Counterargument: this can be a “fiscal illusion”; upfront savings offset by future payments; PPPs may merely reallocate revenues/expenditures intertemporally.
- Fiscal risks:
  - PPPs create direct liabilities (timing and amounts known) and contingent liabilities (depend on uncertain events, natural source of large contingent liabilities).
  - Long-term contracts often involve repeated renegotiations; bargaining power can shift to private parties.
  - Recommendation: include provisions for potential defaults in the medium-term budget and annual budget; monitor and report contingent liabilities.

### Empirical evidence on contingent liabilities and payouts
- Official data on contingent liabilities from off-balance-sheet PPPs available for a limited number of European countries, with ratios lower than 2.5 percent of GDP.
- Example data points:
  - Open Data Albania (September 2020): PPP-related contingent liabilities about 1.5 billion euro (about [10] percent of GDP).
  - Bova et al. (2019) dataset (1990–2014, 80 countries): excludes realizations with gross payout < 0.2 percent of GDP; covers 192 contingent liability realizations with only 8 related to PPPs; 4 of them from Europe.
  - Selected realized payouts:
    - Hungary: payout around USD 1.7 billion (1.5 percent of GDP) — nationalization of motorways PPP in 2005.
    - United Kingdom: payout USD 6.1 billion USD (0.2 percent of GDP) — nationalization of London Underground PPP in 2010.
    - Portugal: payout around USD 1.4 billion (0.6 percent of GDP) — reclassification of PPP stock into central government debt in 2014.

### Empirical evidence on PPP performance and failures
- Literature limitations: scope, coverage, point-in-time analyses, anecdotal evidence, and attribution challenges.
- Findings:
  - Efficiency gains exist but are context-dependent (type/size of projects, regulatory environment, governance).
  - Cost-savings do not always translate into lower consumer prices.
  - Value for money (VfM) studies: mixed results.
  - Many documented failed PPPs with poor design, excessive costs, cancellations, and renegotiations—often linked to increased costs and delays.
- Safeguards to mitigate renegotiation incentives:
  - Limit present value of private sector compensation to original bid amount and require auctioning additional work to the lowest bidder.
  - Example: India reserves renegotiations for very exceptional circumstances.

### Governance and policy implications — summary
- Strengthen monitoring and reporting frameworks for contingent and direct PPP-related liabilities.
- Develop techniques to consolidate PPP-related contingent and direct liabilities and calculate probability of realization.
- Ensure medium-term and annual budgets include provisions for potential defaults and PPP-related shocks.
- Apply contractual safeguards to limit renegotiation incentives and protect public finances.
- Treat PPP suitability as project- and context-specific; integrate PPPs within broader investment strategy.

### Institutional control of PPPs — findings and standards
- Legal framework scope:
  - Western Balkans: most countries have established legal frameworks; not all have PPP-specific frameworks.
  - Albania, Kosovo, Montenegro, North Macedonia, Belarus, and Moldova: single PPP-specific primary legislation.
  - Other countries (Serbia, Bosnia and Herzegovina, Turkey): PPP governance spread across multiple laws.
  - Some PPP frameworks exclude sectors (examples include air transport, roads, ICT, certain renewable energy, water and irrigation, natural resources, large energy PPPs).
- Integration with investment strategy and budget:
  - Best practice: PPPs should originate from national investment strategies; all public investment projects assessed through the same planning process.
  - Observation: PPPs in Western Balkans often planned through parallel processes; projects typically appraised via separate PPP pipelines.
  - Budget approval: In all Western Balkan countries PPPs are not approved through the budget. Albania’s law requires future PPP obligations to be covered in approved annual budget and MTBF, but practice keeps PPPs outside PIM and budget.
- Ministry of Finance veto power:
  - Best practice: Minister of Finance should have authority to veto projects throughout lifecycle.
  - Observations:
    - Albania, Bosnia and Herzegovina, Kosovo, and Montenegro: Finance Minister approval required before procurement launch and before contract signature.
    - North Macedonia and Ukraine: approval before procurement launch but not before contract signature.
    - Serbia (projects exceeding EUR 50 million), Turkey and Ukraine: Finance Ministers may issue only non-binding opinions.
    - North Macedonia and Moldova: ministerial assessment required only if direct budget participation is needed.
- PPP units:
  - Roles: technical support, project assessment, procurement standardization, capacity building.
  - Conflict of interest: promotion and gatekeeping functions should be separated.
  - Observations in Western Balkans:
    - Albania, Bosnia and Herzegovina, Kosovo, Serbia: PPP units assist and promote PPPs; Albania’s unit dependent on Ministry of Finance but not mandated with fiscal management functions.
    - North Macedonia: PPP units limited/advisory.
    - Montenegro: new Montenegrin Investment Agency tasked with promotion and support; role in procurement and implementation unclear.
    - Kosovo and Serbia: PPP units carry both promotion and fiscal management functions (conflict of interest concern).
    - Turkey: no central PPP unit; Department of Strategy and Budget provides technical support.
  - Advanced Europe: PPP units broadly similar; few carry gatekeeping functions (examples include Ireland and Portugal).

### Procurement, renegotiation, and dispute resolution
- Public procurement and direct negotiations:
  - Direct negotiations should be tightly restricted; allowed in Belarus, Moldova, Montenegro, North Macedonia, and Ukraine are prohibited.
  - In Western Balkans: direct negotiations allowed under certain conditions in some countries; vague conditions (e.g., “unforeseeable events”) in Albania, Bosnia and Herzegovina, Kosovo.
  - Unsolicited proposals (USPs): should be avoided or strictly scrutinized; Albania and Bosnia and Herzegovina subject USPs to competitive procedure but USPs represent sizable share of PPP contracts in Albania (notably in energy).
  - Belarus: USPs not subject to competitive bidding or proper feasibility study.
- Renegotiations and ADRs:
  - Long-term PPP contracts need standardized renegotiation and ADR provisions.
  - Western Balkans:
    - Albania and Serbia: laws detail circumstances (including Force Majeure) for renegotiation; Serbia allows renegotiations if they do not increase initial contract value by more than 3 percent.
    - Some laws require Ministry of Finance or cabinet approval for renegotiations (Albania, Serbia, Turkey).
    - Bosnia and Herzegovina and Moldova: laws do not include standard renegotiation clauses but require contracts to include them—potential inconsistency.
    - Most laws reference ADRs.
  - Advanced Europe: renegotiations limited to “substantial” changes; EU norms allow contract amendments below 10 percent (service/supply) or 15 percent (works) of initial contract.

### Management and reporting of fiscal risks
- Comprehensive fiscal risk management includes:
  1. Identification of fiscal risks.
  2. Assessment of magnitude and likelihood.
  3. Reporting of fiscal risks.
  4. Actions to mitigate and manage exposures.
- PFRAM (IMF-World Bank PPP Fiscal Risk Assessment Model):
  - Gathers portfolio information via five questions on initiation, control, payment, government support, and contract risk allocation.
  - Generates portfolio outcomes using project-specific and macro data.
  - Several countries in the region have received PFRAM technical assistance.
- Regional status:
  - Management and reporting of fiscal risks: among weakest areas in Western Balkans and Eastern Europe.
  - Systems to account for explicit/implicit PPP liabilities largely absent or in early stages.
  - Fiscal Risk Units exist in some Ministries of Finance (e.g., Albania, Serbia, Moldova) but are yet to operationalize PPP assessments.

### PPP ceilings, accounting, and disclosure
- PPP ceilings:
  - Ceilings can cover stock and annual flow of PPP payments to ensure fiscal affordability.
  - In Western Balkans, only Albania has an enacted ceiling: total government payouts to PPPs capped at 5 percent of the previous year’s tax revenue.
- Accounting and statistical standards:
  - IPSAS 32 and GFSM 2014 guide on on-/off-balance-sheet treatment based on asset control and economic ownership.
  - Recommendations for cash-accounting countries: treat PPPs as part of government accounting by including PPP partner spending/debt as government spending/debt.
  - Progress: most countries still use cash accounting, lack tests for on-balance-sheet treatment, and have not fully adopted GFSM 2014 or IPSAS 32.
  - Turkey: adopted accrual accounting, GFSM 2014, and IPSAS 32; developed a test to classify PPP-related assets but has not fully applied test results.
  - Albania: determined PPPs need to be on balance sheet if public partner bears construction, availability, or demand risk, but has not registered PPP transactions systematically.
  - Most advanced European countries follow IPSAS 32 and ESA 10.

### Transparency and disclosure
- Recommended disclosures in budget documents:
  - Baseline forecasts for PPP payments.
  - Discussion of risks to the baseline.
- Current practices:
  - None of the Western Balkan countries disclose PPP-related costs in their budget documents.
  - Several countries (Albania, Kosovo, North Macedonia, Serbia, Moldova, Belarus, Ukraine) publish a full list of active PPP projects with information on each.
  - All advanced European countries in the sample publish both lists of PPP projects and their costs in budget documents.

### Scorecard assessment and main governance gaps
- Scorecard: 19 questions across institutional control, procurement and renegotiation, and management/reporting of fiscal risks; benchmarks against IMF and World Bank best practices; color-coded assessments.
- Key gaps identified for Western Balkan countries:
  - No common PPP ceilings (question 13).
  - Frameworks to assess contingent fiscal risks often absent or under development (question 14).
  - Accounting standards and adoption of GFSM 2014/IPSAS 32 vary (questions 15 and 16).
  - Tests for on-balance-sheet treatment often lacking/not fully applied (question 17).
  - PPP costs frequently not included in budget documents (question 18).
  - Publication of active PPP project lists is uneven (question 19).

### Concluding policy priorities and recommendations
- Governance priorities associated with successful PPP programs:
  - Institutional control: establish clear legal framework; integrate PPPs with national investment strategy and budget; provide strong Ministry of Finance oversight and an active gateway approval process.
  - Procurement and contract management: ensure transparent, fair, competitive bidding; restrict direct negotiations; prohibit or tightly regulate USPs; specify renegotiation circumstances in legislation; promote ADRs.
  - Management and reporting: lifecycle reporting of fiscal costs and risks; adopt improved accounting standards (GFSM 2014/IPSAS 32); disclose PPP costs in national budgets.
- Recommended actions for Western Balkan countries:
  - Establish integrated public investment planning and pipelines aligned with national investment strategies.
  - Grant the Minister of Finance an active role in gateway approval and fiscal assessment of PPPs.
  - Consolidate PPP planning, promotion, and negotiation within a centralized PPP unit staffed with qualified personnel; separate gatekeeping from procurement/contract management.
  - Tighten regulations on direct negotiations; prohibit unsolicited proposals as first best or require them to follow the same competitive procedure and verification of feasibility.
  - Require independent verification or re-assessment of feasibility studies for unsolicited proposals.
  - Specify circumstances for contract renegotiation in PPP legislation and promote ADR before formal mechanisms.
  - Establish frameworks and comprehensive databases to assess and account for explicit and implicit PPP liabilities; operationalize assessments to manage fiscal risks.
  - Improve accounting standards and disclosure in national budgets to better reflect PPP risks and costs.

### Empirical macro-institutional analysis (Annex IV) — data and main results
- Sample period: 1996–2019; around 100 countries.
- Data source: IMF Investment and Capital Stock Dataset (ICSD) merging World Bank PPI and EIB EPEC databases.
- General sample PPP investment, % of GDP (medians [means]):
  - 1996–2001 = 0.21 [0.46]
  - 2002–2007 = 0.18 [0.35]
  - 2008–2013 = 0.10 [0.54]
  - 2014–2019 = 0.25 [0.40]
- Per capita PPP capital stock (full sample medians [means]):
  - 1996–2001 = 30.11 [103.64]
  - 2002–2007 = 97.25 [203.35]
  - 2008–2013 = 112.05 [299.77]
  - 2014–2019 = 121.91 [428.71]
- Subgroup regularities:
  - PPP investments higher in countries with higher growth and lower income levels (preliminary and not consistent across periods).
  - High-inflation countries tend to have lower PPP capital stock per capita after 2002.
  - High-debt countries may have higher PPP capital stock per capita (statistically significant after 2007).
  - Countries with stronger governance tend to have higher PPP capital stock per capita.
- Regression evidence (cross-country and panel):
  - Results are generally weak and not robust across specifications and samples.
  - Most robust positive association: better governance associated with greater use of PPPs.
  - Supporting but less robust results: high inflation adversely affects PPP investments; high debt may be associated with increased PPP use.
  - Panel estimation diagnostics: lagged PPP investment (% of GDP, lag) is consistently positive and statistically significant across many specifications (examples: 0.404*** (0.032); 0.592*** (0.056); 0.740*** (0.041)).
  - Governance indicators (Regulatory Quality, Control of Corruption, Rule of Law) show mixed signs and significance across specifications.
  - Fiscal and macro controls (public capital stock per capita, GG gross debt, inflation, growth) yield varied signs and occasional significant coefficients across specifications.
- Limitations:
  - Macro variables may not capture sectoral/project-specific drivers of PPPs.
  - Political economy and governance quality, project design, and private financing factors are important but incompletely captured.
  - Data limitations and heterogeneity across countries reduce robustness.

*Source: IMF Working Paper — The Future of PPPs in the Western Balkans (wpiea2023031-print-pdf).*

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

### wpiea2023031-print-pdf - References (Extracted content)

### Key findings and context
- Public infrastructure in the Western Balkans remains underdeveloped relative to the EU and other regional peers, with the 2019 Global Competitiveness Index ranking the Western Balkans at 78 (out of 141 countries) for infrastructure quality versus EU-15 at 17 and CESEE at 51.
- Variation across the Western Balkans is large: Serbia ranked 51 and Albania ranked 98 (out of 141) in 2019 infrastructure quality rankings.
- Public debt levels as of 2021 range between 36 percent of GDP in Bosnia and Herzegovina and 87 percent of GDP in Montenegro.
- The regional average public debt increased from 54 percent in 2019 to a projected 62 percent in 2022.
- Since 2012, PPP investments in the Western Balkans averaged around 0.5 percent of GDP—higher than the EU and global averages for the same period.

### PPP rationale, benefits, and risks
- PPPs are long-term, risk-sharing contracts between a government and a private company, where the private party typically invests in an asset and provides services subject to performance criteria.
- Two broad types of PPPs by funding source (Box 1):
  - Government-funded or availability-based PPPs: government makes predetermined payments for services over the contract duration or payments per volume of service.
  - User-funded PPPs (often “concessions”): users pay fees for services; sizeable government spending may still be needed to subsidize investment or guarantee private debt/revenue.
  - Mixed PPPs combine elements of both government- and user-funded types.
- Potential benefits: can increase short-term output via aggregate demand and longer-term output by expanding productive capacity; contribute to job creation; help meet energy transition goals by pooling capital and private-sector technological expertise.
- Fiscal and other risks:
  - PPPs can introduce large fiscal risks and increased costs if poorly designed or managed, due to complex risk allocation and long-term commitments.
  - PPPs may create “fiscal illusion” because their long-term cashflow profiles can mask fiscal commitments.
  - Risks can materialize during shocks (e.g., COVID-19 pandemic, war in Ukraine), affecting the resilience of PPP-financed projects.
  - PPPs have been disrupted historically during financial crises (AFC 1997–98 and GFC 2008–09), with EU PPP transaction values declining by 50 percent from 2007 to 2009 and PPP project counts dropping around 40–50 percent during the GFC and Euro Area crisis.

### Stylized facts and global/regional trends
- Global median PPP capital stock:
  - 0.6 percent of GDP in 2001
  - 1.8 percent of GDP in 2011
  - 2.5 percent of GDP in 2019
- At end-2019:
  - Emerging markets had the largest PPP capital stock as a share of GDP at about 3.7 percent, driven by Asian emerging markets.
  - Advanced economies had the largest PPP capital stock in per capita terms.
  - Ten European economies were among the top 20 countries with the largest per capita PPP capital stocks.
- Median annual PPP investments:
  - Peaked at 0.2 percent of GDP in 2010
  - Declined to about 0.1 percent of GDP thereafter
  - In low-income countries, median PPP investments remained around 0.2 percent of GDP over the past decade
- Geographic and sectoral patterns:
  - Europe: PPPs concentrated in transportation by project value; education and healthcare dominated by number of projects in the mid-2000s.
  - Emerging market and developing economies: PPP infrastructure investment concentrated in energy, followed by transportation.
  - Since 2020, the number and value of user-funded PPP projects in Europe increased.

### Governance and policy implications (summary of main conclusions)
- PPPs can be a useful tool but require strong governance to avoid fiscal costs and risks. Policymakers should carefully design, implement, monitor, and manage PPPs.
- Key areas for improvements in Western Balkans PPP frameworks:
  - Strengthen institutional control over PPPs, including:
    - Establishing common public investment planning processes that apply to both PPPs and traditional public procurement projects.
    - Providing the Minister of Finance with effective powers to ringfence PPP risks.
  - Enhance competitive procurement processes, including restricting scope for direct negotiations and unsolicited proposals.
  - Ensure more resilient contracts by specifying clear criteria for contract renegotiation.
  - Improve management and reporting of fiscal costs and risks, particularly accounting for and disclosure of PPPs.
- PPPs may support energy transition objectives by mobilizing private capital and technology, but certain segments (e.g., hydropower, nuclear) involve relatively high participation by SOEs.

### Paper structure and scope
- The paper provides:
  - Historical overview of PPPs in the Western Balkans and their place in global and European context.
  - Review of literature on PPP benefits and drawbacks and survey of empirical evidence.
  - Framework for assessing PPP governance focusing on institutional elements to manage fiscal costs and risks.
  - Lessons and policy recommendations tailored to Western Balkans policymakers.

*Source: IMF Working Paper extract (Introduction, Sections I–II, Figures and Box 1 as provided in the source content).*

### 16. In  the Western Balkans, while the number of PPP projects has remained low relative to the EU-

### 16. In the Western Balkans, while the number of PPP projects has remained low relative to the EU-15 countries and CESEE peers, they account for a relatively large share of GDP and total capital stock.

### Key findings on PPP prevalence and scale in the Western Balkans
- PPP capital stock as a share of GDP and of total capital stock in the Western Balkan countries have been rising since 2005, with a sharp acceleration in 2012/13.
- As of 2019, this share is higher in the Western Balkans than in the EU-15 or CESEE countries.
- Annual PPP investment as a share of GDP is more volatile but has also grown since 2005, particularly in Albania, Bosnia, and Serbia.
- These data broadly suggest that PPP projects in the Western Balkans tend to be larger in percent of GDP compared to PPP projects in the EU.
- Implication: the failure of an individual PPP project could create a larger burden for these countries.
- Note on data: The coverage of PPP projects drawn from different sources might vary across countries.

### Conceptual arguments — Efficiency
- Pro-PPP argument:
  - PPPs can complement public provision of capital and social infrastructure to generate a more efficient allocation of public resources.
  - Typical division of labor: public authorities focus on planning, policy making, and regulation; private companies provide specialized services.
  - Well-designed PPPs can generate efficiency gains that can more than compensate the higher cost of private financing.
- Anti-PPP argument:
  - Complexity of PPP structures increases cost as authorities implement processes outside normal competence.
  - Long-term commitments can be insufficiently flexible to handle unforeseen changes.
  - PPP procurement costs can reach 5-10 percent of the capital cost for a large project, and do not decline proportionately for smaller projects.
  - Cost-efficiency assessments should incorporate whole economic picture, including social returns.

### Conceptual arguments — Fiscal space
- Claimed benefit:
  - Private financing of PPPs is sometimes presented as creating fiscal space because governments need not issue debt or use public resources upfront.
- Counterarguments:
  - The fiscal space argument can be a “fiscal illusion”: resources “saved” upfront are offset by future revenue flows to the private company.
  - PPPs may only affect intertemporal allocation of revenues/expenditures rather than provide extra fiscal space in present value terms.
  - Fiscal illusion and postponement of costly disclosures can lead to low-quality, fiscally costly projects that would otherwise be rejected.
  - Poor contract design, optimistic revenue assumptions, and minimum income guarantees often result in large fiscal risks or eventual fiscal costs.
  - Political pressure to lower prices can weaken PPP finances, cause underinvestment and quality problems, and can eventually result in nationalization.

### Conceptual arguments — Fiscal risks
- Long-term contracting through PPPs does not remove infrastructure project risks; PPPs often involve repeated renegotiations as risks materialize.
- Government remains ultimately responsible for provision of public services; bargaining advantage often shifts to private company, increasing fiscal risks.
- Long contractual horizons add management risks and implicit risks where infrastructure governance is weak.
- Institutional PPPs (joint public-private ownership) can lead to inefficient allocation of risk favoring the private partner.
- PPP projects create direct and contingent liabilities for the public sector:
  - Direct liabilities: timing and amounts known.
  - Contingent liabilities (including guarantees): depend on uncertain future events and are a natural source of large contingent liabilities.
- Monitoring and reporting frameworks for contingent liabilities are limited in most countries; consolidation and probability assessment techniques are important.
- Recommendation: include provision for potential defaults in both the medium-term budget and annual budget.

### Empirical evidence on contingent liabilities and payout realizations
- Official data on contingent government liabilities related to off-balance-sheet PPPs are available only for a limited number of European countries, with ratios lower than 2.5 percent of GDP.
- Example data and dataset constraints:
  - Open Data Albania reports that PPP-related contingent liabilities as of September 2020 accounted about 1.5 billion euro (about [10] percent of GDP).
  - Bova, Ruiz-Arranz, Toscani, and Ture (2019) dataset (1990-2014, 80 countries) excludes realizations with a gross payout lower than 0.2 percent of GDP; covers 192 contingent liability realizations with only 8 related to PPPs, and 4 of them are from Europe.
  - Examples of PPP-related realized payouts:
    - Hungary: payout around USD 1.7 billion ( 1.5 percent of GDP) — resulted in nationalization of the PPP on motorways in 2005.
    - United Kingdom: payout USD 6.1 billion USD (0.2 percent of GDP) — resulted in nationalization of the PPP on the London Underground in 2010.
    - Portugal: payout around USD 1.4 billion (0.6 percent of GDP) — led to reclassification of a large stock of PPPs into central government debt in 2014.

### Empirical evidence on PPP performance and failures
- Empirical literature is constrained by scope, coverage, point-in-time analyses, anecdotal evidence, and case studies; attribution and counterfactuals are often debatable.
- Some evidence of efficiency gains exists, but results are context-dependent:
  - Efficiency gains depend on type and size of projects, regulatory environment, governance, and historical/political landscape.
  - Improvements in sector performance and quality have been observed in PPIs, though findings are sensitive to control variables.
  - Cost-savings do not always translate into service price reductions for final consumers.
  - Value for money (VfM) studies offer mixed results.
- Many anecdotal and empirical examples document mixed performance and numerous failed PPPs with poor design and excessive costs across Europe.
- Contract cancellations or renegotiations are perceived as PPP failures due to high costs and service disruptions; renegotiations are commonly linked to increased costs and delayed completion.
- Safeguards to mitigate renegotiation perverse incentives:
  - Example safeguard: limit the present value of private sector compensation to the amount in the original bid and require auctioning additional work for the lowest bid.
  - India example: government reserves renegotiations for very exceptional circumstances to minimize moral hazard in bidding and project specification.

### Policy and governance implications emphasized
- Strengthen monitoring and reporting frameworks for contingent and direct PPP-related liabilities.
- Develop techniques to consolidate PPP-related contingent and direct liabilities and calculate probability of realization.
- Ensure medium-term and annual budgets include provisions for potential defaults and PPP-related shocks.
- Apply contractual safeguards to limit renegotiation incentives and protect public finances.
- Recognize that PPP suitability depends on project size, institutional context, governance quality, and regulatory environment; treat PPPs within the broader investment strategy rather than in isolation.

*IMF Working Paper: The Future of PPPs in the Western Balkans*

### 29. Most countries with successful PPP programs benefit from a clear and well-designed

### 29. Most countries with successful PPP programs benefit from a clear and well-designed governance framework

### Importance of a strong PPP governance framework
- Poor infrastructure governance is a key factor contributing to the failure and increasing fiscal risks from PPPs.
- A strong set of PPP governance policies and practices is critical to:
  - Maximize possible benefits of PPPs.
  - Increase authorities’ ability to monitor and mitigate accompanying risks.
  - Ensure rights and responsibilities of public and private parties are clearly delineated.

### PPP lifecycle coverage and safeguards
- An adequate framework should cover all stages of the PPP lifecycle, including preparation, procurement, and contract management (World Bank, 2014).
- Preparation phase:
  - Priority investment projects should be identified regardless of financing modality.
  - Possible PPP projects should be identified based on an appraisal showing PPP is expected to maximize value for money compared to conventional financing tools.
- Procurement phase:
  - PPP contracts are designed to prepare the project for procurement.
- Contract management phase:
  - Management of the PPP contract over its lifetime after contract signature.
- Safeguards:
  - Appropriate safeguards should ensure PFM controls are not bypassed.
  - Government exposure to PPPs-related fiscal risks should be controlled (IMF, 2021).

### Assessment scope and methodology
- This section assesses PPP governance frameworks in the Western Balkans with focus on elements needed to support effective monitoring and management of fiscal costs and risks.
- Comparators include Turkey, three Eastern European countries (Belarus, Moldova, and Ukraine), and a set of comparators in advanced Europe (Belgium, Denmark, France, Germany, Ireland, Portugal, Spain, and the UK).
- The assessment reviews 19 specific aspects which play a role in managing fiscal risks and costs in PPPs, with standards consistent with Fund advice.
- The evaluation for each subcomponent supports construction of a country score matrix (see Annex II).

### Three main categories of assessment
- Institutional control over PPPs:
  - Weaknesses in early stages may lead governments to select poor quality projects and increase future fiscal risks.
- Practices for public procurement and contract management:
  - PPP contracts create present and contingent public sector liabilities that can be a source of fiscal risks if poorly designed or executed.
- Management and reporting of fiscal risks:
  - Fiscal costs and risks of PPPs must be monitored and controlled along different stages of the cycle (IMF, 2021).

### Institutional control of PPPs — key findings and standards
- Strong institutional framework should cover preparation and planning stages, including:
  - Legal framework governing PPPs.
  - Integration of PPPs with investment strategy, infrastructure priorities and budget.
  - Strong powers for the Ministry of Finance to control fiscal consequences (OECD, 2012; Schwartz and others, 2020).
- Common pitfalls:
  - PPPs placed on separate tracks from other public investment projects.
  - Fiscal illusion that PPPs are “free” infrastructure, despite fiscal implications materializing later.

A. Scope and Coverage of the PPP Legal Framework
- A sound legal and regulatory framework should:
  - Define PPPs clearly and apply uniformly to all PPPs.
  - Include policies, procedures, institutions and rules for project selection, procurement, assessment and monitoring of fiscal risks, service delivery, and project sustainability.
- Forms of legal frameworks:
  - PPP-specific legislation and regulations.
  - Official PPP-specific government guidelines, manuals, or other legally-binding official documents.
  - Other non-PPP-specific legislation relevant to PPPs (sector-specific law, general procurement, contract, company, arbitration, or environmental law).
- Observations:
  - Most Western Balkans countries have established legal frameworks, though not all have PPP-specific frameworks.
  - Albania, Kosovo, Montenegro, and North Macedonia as well as Belarus and Moldova have adopted a single PPP-specific primary legislation.
  - In other countries (e.g., Serbia, Bosnia and Herzegovina, Turkey) PPP governance is spread across multiple laws, implying non-uniform treatment of PPP projects.
  - Some PPP frameworks exclude sectors (examples given: Albania excludes air transport, roads, ICT, some renewable energy; Serbia excludes some ICT, water and irrigation, some ground transportation; Montenegro excludes certain natural resource and energy projects; large energy PPPs in Ukraine not governed by PPP law).
  - Advanced European countries also vary; France, Spain, Germany, UK, Denmark have multiple PPP or concession laws; most EU member PPP contracts regulated by EU directives.

B. Integration of PPPs with the National Investment Strategy and Budget Process
- Best practice:
  - PPP projects should originate from a national investment strategy and be aligned with public priorities.
  - All public investment projects need to be assessed through the same public investment planning process and integrated into medium-term national investment strategies (Schwartz and others, 2020 and OECD, 2012).
- Budget integration:
  - All PPP fiscal costs and risks (explicit or contingent) should be approved in national budgets, preferably through the Medium-Term Budget Framework (MTBF) to prevent circumvention of budgetary restrictions (IMF, 2021; Queyranne, Daal, and Funke 2019).
- Observations in Western Balkans:
  - PPPs tend to be planned through parallel investment planning processes rather than integrated with public investment processes.
  - Legislation often requires alignment with national or sectoral strategies, but projects are typically appraised and prioritized through a separate PPP pipeline.
  - Albania’s law allows contracting authorities to select projects to be financed via PPPs; in practice PPPs remain managed outside PIM and budget processes (Funke, 2019; Darcy and Steger, 2022).
  - Among advanced European countries, Germany, Ireland, Spain, and UK require all investment projects be assessed through the same investment planning process.
  - In all Western Balkan countries, PPPs are not approved through the budget (Figure 6). Albania is the only country where law requires contracting authorities to ensure future PPP obligations are covered in the approved annual budget and MTBF, yet practice still manages PPPs outside PIM and budget.

C. Ministry of Finance Veto Power
- Best practice:
  - The Minister of Finance should have authority to veto projects throughout their lifecycle if they lack value for money, are unaffordable, or entail significant fiscal risks (Irwin, Mazraani, and Saxena, 2018; IMF, 2021).
  - Approval by the minister should be required at early stages (pre-feasibility and feasibility) and be needed through contract signature.
- Observations:
  - Western Balkans PPP laws require Finance Minister involvement to varying degrees (Figure 6):
    - Albania, Bosnia and Herzegovina, Kosovo, and Montenegro: Finance Minister approval required both before procurement launch and before contract signature.
    - Albania: decisions remain politicized in practice (Darcy and Steger, 2022).
    - North Macedonia and Ukraine: Ministry of Finance can approve before procurement launch but not before contract signature.
    - Serbia (projects exceeding EUR 50 million), Turkey and Ukraine: Finance Ministers may only issue non-binding opinions.
    - North Macedonia and Moldova: ministerial assessment required only if a project requires direct budget participation.
  - Advanced Europe:
    - Gateway process originated in the UK and is used in France, Germany, and Portugal.
    - In Denmark, France and Portugal, a financial analysis comparing PPP cost to traditional procurement is required.

D. PPP Units
- Roles and benefits:
  - Provide technical support: legal, technical, financial expertise for project identification, assessment, standardization of procurement, capacity building, and demonstration of political commitment (OECD, 2010; World Bank, 2007).
  - Improve coordination and reduce transaction costs by filling information gaps.
  - Staff should have negotiation, finance, and regulation skills (OECD, 2010).
- Conflict of interest concerns:
  - Dual promotion and gatekeeping functions can create conflicts of interest.
  - Preferable to separate technical support/promotion from gatekeeping (e.g., Chinese walls, separate units, or independent fiscal management agency) (IMF, 2021).
- Observations in Western Balkans:
  - Albania, Bosnia and Herzegovina, Kosovo and Serbia: PPP units assist in preparing and evaluating projects and promote PPPs; Albania’s unit is dependent on Ministry of Finance but not mandated with fiscal management functions.
  - North Macedonia: PPP units have more limited or advisory roles (World Bank, 2020a).
  - Montenegro: new agency (Montenegrin Investment Agency) tasked with PPP promotion and support in preparation; role in procurement and implementation unclear. Concession Commission supports implementation but not full cycle responsibilities.
  - Turkey: no central PPP unit; Department of the Strategy and Budget provides technical support to the public investment program which includes PPPs.
  - Kosovo and Serbia: PPP units carry both promotion and fiscal management functions, raising conflict of interest concerns.
  - Advanced Europe: PPP units broadly similar; few carry gatekeeping functions (examples include Ireland and Portugal).

### Overall conclusions on institutional control
- PPP governance frameworks in Western Balkans have improved over time, but institutional shortcomings remain:
  - Fragmentation of PPP legal framework in some countries, sometimes involving sector-specific legislation.
  - Need to establish integrated public investment planning processes to create a pipeline of PPP projects.
  - Need to grant the Minister of Finance a key and active role in a gateway process to approve PPP projects and assess PPP fiscal costs and risks.
  - PPP units, particularly recent ones, require effective and clear-cut roles.
- Advanced Europe frameworks are broadly stronger, but shortcomings persist where PPP selection is not integrated within broader planning processes.

*Source: IMF staff*

### 50. Strong  legal and operational frameworks are needed to govern PPP procurement and

### 50. Strong legal and operational frameworks are needed to govern PPP procurement and renegotiation

### A. Public procurement and direct negotiations
- PPP transactions are expensive, lengthy, and risky; procurement laws should ensure bidding is transparent, fair, and competitive to reduce perceptions of corruption and attract high-quality bidders.
- Direct negotiations should be strictly regulated and only permitted under limited circumstances such as for very small or very urgent projects, or if there is only one possible counterpart.
- In the Western Balkans:
  - Direct negotiations are prohibited in Belarus, Moldova, Montenegro, North Macedonia, and Ukraine.
  - Direct negotiations are allowed under certain circumstances in the remaining Western Balkan countries; some circumstances are well-defined, others refer to vague conditions such as “unforeseeable events” (Albania, Bosnia and Herzegovina, and Kosovo).
  - In Turkey, direct negotiations are allowed when procurement does not exceed a minimum amount.
- In advanced European countries, direct negotiation is allowed under tight restrictions, except for Belgium and Portugal where it is prohibited.
- Unsolicited proposals (USPs):
  - USPs should normally be avoided or, at minimum, subject to strict scrutiny because they can divert investment from strategic priorities and risk poor value for money.
  - If USPs are allowed, procurement laws should (i) require the government undertakes independent feasibility analysis; and (ii) ensure a competitive bidding process that gives other bidders adequate time to prepare bids.
  - Evidence indicates the vast majority of tenders for unsolicited PPPs present a single bidder, suggesting USPs may not lead to competition even under competitive procurement.
- Regional treatment of USPs in the Western Balkans:
  - Albania and Bosnia and Herzegovina: USPs subject to the same competitive procedure as government-initiated projects and require feasibility studies; nevertheless, in Albania USPs represent a sizable share of PPP contracts, especially in the energy sector.
  - Montenegro and Serbia: USPs are awarded through the same competitive process as other PPPs, but there is no requirement for independent feasibility studies; except for Serbia, feasibility studies for USPs are not explicitly required in most Western Balkan countries.
  - Ukraine: USPs are awarded through competitive bidding (provided there are multiple bidders).
  - Belarus: USPs are not subject to a competitive bidding process or a proper feasibility study.
  - Most advanced European countries do not reference USPs explicitly, with exceptions noted elsewhere.

### B. PPP renegotiations and dispute resolution
- PPP contracts often span beyond 20 years and should provide flexibility and certainty through mechanisms for renegotiation and dispute resolution.
- The legal framework, rather than individual contracts, should standardize required provisions to ensure consistency.
- Renegotiations create corruption and collusion risks and therefore should be restricted; examples include prohibiting contract amendments during the first few years or limiting the number of amendments.
- The PPP legal framework should describe institutional mechanisms for alternative dispute resolution mechanisms (ADRs).
- Western Balkans practices:
  - Only the PPP laws in Albania and Serbia describe in detail specific circumstances (including Force Majeure) that would justify a contract renegotiation.
  - Serbian law allows renegotiations if these do not increase the initial contract value by more than 3 percent.
  - Some laws require renegotiations to be approved by the Ministry of Finance (Albania, Serbia) or the cabinet (Turkey).
  - Bosnia and Herzegovina and Moldova: PPP laws do not include standard renegotiation clauses but require specific PPP contracts to include them, creating potential inconsistency.
  - Most laws in the Western Balkans reference ADRs.
- Advanced European practices:
  - In the UK, Ireland, and France, renegotiations are limited to changes affecting the risk allocation or other amendments considered “substantial.”
  - EU member countries adhere to norms allowing contract amendments below 10 percent of the amount of the initial contract for service and supply contracts or 15 percent of the amount of the initial contract for works contracts.
  - Following the COVID-19 pandemic, definitions of force majeure and renegotiation processes have been revised in several European countries (e.g., the UK, Spain, Portugal) to add clarity, flexibility, and accountability.
- Policy implications:
  - PPP legislations should tighten regulations for recourse to direct negotiations.
  - Where USPs are not prohibited, policy options include prohibiting them as a first best solution or requiring the same competitive procedures as for other proposals.
  - PPP legislations (rather than individual contracts) should specify circumstances for contract renegotiation and require space for ADRs before resorting to formal mechanisms.

### C. Management and reporting of fiscal risks
- PPPs distribute risks to the public sector, including construction risk (design problems, cost and schedule overruns), financial risk (cash flow shortfalls), and demand risk.
- Adequate risk management frameworks and fiscal institutions (fiscal rules/targets, budgeting procedures, accounting and auditing standards, assignment of responsibilities) influence the magnitude of fiscal risks.
- A comprehensive fiscal risk management framework includes:
  1. Identification of fiscal risks.
  2. Assessment of their magnitude and likelihood of realization.
  3. Reporting of fiscal risks.
  4. Actions to mitigate and manage exposures.
- The IMF-World Bank PPP Fiscal Risk Assessment Model (PFRAM) can be used to quantify fiscal costs and risks of PPPs.
  - PFRAM is an analytical tool that helps quantify macro-fiscal implications and identify mitigation measures.
  - PFRAM gathers portfolio information via five main questions:
    1. Who initiates the project?
    2. Who controls the asset?
    3. Who ultimately pays for the asset?
    4. Does the government provide additional support to the private sector?
    5. What does the PPP contract risk allocation tell us about macro-fiscal risks?
  - Using project-specific and macroeconomic data, the module generates outcomes for a portfolio of PPP projects.
  - Several countries in the region have received technical assistance on PFRAM from the World Bank or the IMF’s Fiscal Affairs Department.
- Current regional status:
  - Management and reporting of fiscal risks is one of the weakest areas in the Western Balkans and Eastern Europe; none of the developing European countries in the sample have well-established frameworks to assess fiscal risks from PPPs.
  - Systems to account for explicit or implicit liabilities from PPPs are largely absent or in early stages; comprehensive databases are still being established in some countries.
  - A few countries have Fiscal Risk Units (e.g., Albania, Serbia, and Moldova) in their Ministries of Finance, though these units are yet to operationalize assessments of fiscal risks from PPPs.

### D. PPP ceilings
- Ceilings or limits on PPPs (covering stock and annual flow of PPP payments) can help ensure PPPs remain fiscally affordable and incentivize investment prioritization.
- Except for Albania, no country in the Western Balkan region has enacted ceilings on PPPs.
  - Albania has a ceiling on total government payouts to PPPs set at 5 percent of the previous year’s tax revenue.
- In the sample of advanced European countries, only a few ceilings are established, though several countries have checks and balances through assessment of contingent liabilities and transparent accounting.

### E. Accounting and statistical standards
- Modern public sector accounting (IPSAS 32) and statistical standards (GFSM 2014) provide guidance on determining whether PPP-related assets are on- or off-balance sheet based on asset control and economic ownership.
- Where full adoption of IPSAS 32 and GFSM 2014 is not immediate, countries using cash accounting are advised to treat PPPs as part of the government for accounting purposes by including PPP partner spending/debt as government spending/debt.
- Progress in developing countries in the sample has been slow:
  - Most countries still use cash accounting, lack tests to determine on-balance-sheet treatment, and have not fully adopted GFSM 2014 or IPSAS 32.
  - Turkey is an outlier: it has adopted accrual accounting, GFSM 2014, and IPSAS 32, and developed a test to classify PPP-related assets as public or private, though it has not fully applied the tests’ results and currently records only PPP liabilities that are clearly on balance sheet and above the line in the budget.
  - Albania has determined PPPs need to be on balance sheet if the public partner bears construction, availability, or demand risk, but in practice has not registered PPP transactions systematically.
  - Most advanced European countries follow some form of IPSAS 32 and ESA 10 (a close substitute of GFSM 2014).

*Italicized source: IMF Working Papers — The Future of PPPs in the Western Balkans*

### 66. A  transparent disclosure of  information on  PPP projects is essential to  informing  all

### A transparent disclosure of information on PPP projects is essential to informing all

### Transparency and disclosure: standards and importance
- Budget documents should include details on government obligations under PPPs, covering both direct and contingent liabilities, as recommended in the IMF’s Fiscal Transparency Handbook (Irvin and others, 2018).
- Disclosures should include:
  - Baseline forecasts for payments under PPPs.
  - A discussion of risks to the baseline.
- Systematic disclosure of information related to PPP projects — including contracts, where feasible — is important for accountability and good governance (World Bank, 2017).

### Current practice in the Western Balkans versus advanced European countries
- None of the Western Balkan countries discloses PPP-related costs in their budget documents.
- Several countries in the paper’s sample (including Albania, Kosovo, North Macedonia, Serbia Moldova, Belarus, and Ukraine) publish a full list of active PPP projects, with information on each project.
- All the advanced European countries in the paper’s sample publish both:
  - A list of PPP projects, and
  - Their costs in budget documents.

### Managing and reporting fiscal risks of PPPs (assessment framework and findings)
- The paper evaluates PPP governance frameworks using a PPP governance “scorecard” based on answers to 19 specific questions covering three areas: institutional control of PPPs, PPP procurement and renegotiation, and management and reporting of fiscal risks.
- The assessment translates benchmarked answers into colors in a chart (green = highest compliance; red = lowest). For questions with 2, 3, or 4 answers the assigned colors are red/green; red/orange/green; and red/orange/yellow/green respectively. Absence of sufficiently clear information is grey.
- Key management and reporting gaps identified for Western Balkan countries:
  - No ceilings on PPPs commonly in place (question 13 status varies by country).
  - Frameworks to assess contingent fiscal risks are often absent or under development (question 14).
  - Accounting standards and adoption of GFSM 2014/IPSAS 32 vary (questions 15 and 16).
  - Tests for on-balance-sheet treatment of PPPs are often lacking or not fully applied (question 17).
  - PPP costs are frequently not included in budget documents (question 18).
  - Publication of lists of active PPP projects is uneven (question 19).

### Concluding remarks: benefits, risks, and governance priorities (paragraphs 68–71)
- Role of PPPs:
  - PPPs are increasingly important for Western Balkan countries to increase investment and reduce infrastructure gaps.
  - The infrastructure gap with the EU-15 has narrowed somewhat over the past 15 years but remains large; infrastructure quality compares poorly with other CESEE countries.
  - Next-Gen EU funding and broader regional infrastructure pushes could increase pressure on Western Balkan countries to scale up investment.
- Benefits and drawbacks:
  - Well-designed and executed PPPs can support infrastructure development and enhance public service delivery through efficiency gains.
  - PPPs carry potential for large fiscal risks and increased costs if not managed well; complexity and long-term deviations can reduce cost efficiency in practice.
  - PPPs can create a “fiscal illusion” leading to uptake of low-quality and fiscally costly projects.
  - Direct and contingent liabilities from PPPs are often not adequately recognized, recorded, and/or managed, posing significant fiscal risks.
- Governance framework elements associated with successful PPP programs:
  - Institutional control: legal framework, integration with investment strategy and budget, strong Ministry of Finance oversight over fiscal consequences, an active gateway process for approvals.
  - Public procurement and contract management: transparent, fair, competitive bidding; well-designed and flexible contracts with provisions on renegotiation and efficient dispute resolution.
  - Management and reporting: lifecycle reporting of fiscal costs and risks; improved accounting standards and disclosures in national budgets.
- Recommended priorities for Western Balkan countries:
  - Establish integrated public investment planning and pipelines aligned with national investment strategies.
  - Provide the Minister of Finance an active role in gateway approval and fiscal assessment of PPPs.
  - Consolidate PPP planning, promotion, and negotiation within a centralized PPP unit staffed with qualified personnel; separate gatekeeping from procurement/contract management to reduce conflicts of interest.
  - Tighten regulations on recourse to direct negotiations; prohibit unsolicited proposals as first best or require them to follow the same competitive procedure.
  - Ensure feasibility studies for unsolicited proposals are verified or re-assessed.
  - Specify in PPP legislation circumstances for contract renegotiation and promote alternative dispute resolution before formal mechanisms.
  - Establish frameworks and comprehensive databases to assess and account for explicit and implicit liabilities from PPPs; operationalize assessments to manage fiscal risks.
  - Improve accounting standards and disclosure in national budgets to better reflect PPP risks and costs.

### Annex highlights
- Country coverage definitions:
  - Western Balkans: Albania; Bosnia and Herzegovina; Kosovo; Montenegro; North Macedonia; Serbia.
  - EU-15: Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain, Sweden, United Kingdom.
  - CESEE: Albania, Belarus, Bosnia and Herzegovina, Bulgaria, Croatia, Czech Republic, Estonia, Hungary, Kosovo, Latvia, Lithuania, Moldova, Montenegro, North Macedonia, Poland, Romania, Russia, Serbia, Slovak Republic, Slovenia, Turkey, and Ukraine.
- Evaluation methodology:
  - The scorecard uses answers to 19 questions, drawn from IMF staff desk reviews, PIMA reports, a World Bank survey, and country legal documents and databases. Responses are benchmarked against IMF and World Bank best practices and translated into color-coded assessments.
- Historical evidence on cancellations and distressed PPPs (Annex III):
  - Less than 4 percent of PPP investments have been cancelled between 1990 and 2021, with a total value of approximately USD 80 billion.
  - Between 1990 and 2021, the average investment value of cancelled projects was about USD 410 million versus USD 222 million for active projects.
  - The average declines to about USD 280 million without a large cancelled PPP project in Brazil (footnote context noted in source).
  - Cancellations tend to occur in larger projects, in the water sector, and during years of financial crises.
  - The water sector had the highest cancellation rate with more than 15 percent of projects being cancelled compared to 6 percent or less in the energy and transport sectors.
  - Almost half of all canceled PPP investments in nominal terms occurred in the transport sector.

*Source: IMF staff.*

### Annex IV. Macro-Institutional Determinants of

### Annex IV. Macro-Institutional Determinants of PPPs

### Data
- Sample period: 1996–2019 for a sample of around 100 countries.
- Source dataset: IMF Investment and Capital Stock Dataset (ICSD) which merges and harmonizes the World Bank PPI database and the EIB EPEC database.
  - PPI database covers 137 emerging economies and developing countries.
  - EPEC database covers 27 advanced and emerging European economies and includes social infrastructure sectors in addition to economic infrastructure sectors.
- Key harmonization notes:
  - EPEC reports project value (a stock variable); PPI reports investment amounts over years (flows).
  - The IMF database adjusts and harmonizes series to be consistent over time and across countries.

### PPPs and macro-fiscal and institutional characteristics — empirical comparisons
- Variables explored:
  - Size of the economy: growth, inflation, income levels, capital stock per capita.
  - Fiscal indicators: debt, fiscal balances, fiscal vulnerability index.
  - Institutional characteristics: corruption (control of corruption), regulatory quality, rule of law (Worldwide Governance Indicators).
- Method: sample split into high vs low groups using median thresholds; statistics reported as medians with means in brackets; analysis also divided 1996–2019 into four sub-periods (1996–2001, 2002–07, 2008–13, 2014–19).

Findings from subgroup comparisons (selected empirical regularities and exact reported sample statistics):
- General sample PPP investment, % of GDP (medians [means]):
  - Full sample: 1996–2001 = 0.21 [0.46]; 2002–2007 = 0.18 [0.35]; 2008–2013 = 0.10 [0.54]; 2014–2019 = 0.25 [0.40].
- Growth and income:
  - Preliminary evidence: PPP investments higher in countries with higher growth and lower income levels (Annex Table 1A), though not consistent across all time periods.
  - After 2007, PPP investments (share of GDP) are larger and statistically significant in high-growth countries and those with per-capita income lower than the median.
- Inflation:
  - Using PPP capital stock per capita (Annex Table 1B): after 2002, high-inflation countries tend to have significantly lower PPP capital stock per capita than low-inflation countries.
- Debt and fiscal vulnerability:
  - Some evidence that high-debt countries have higher PPP capital stock per capita than low-debt countries (statistically significant after 2007).
  - Countries with high fiscal vulnerabilities tend to have a lower PPP capital stock per capita versus countries with lower vulnerabilities.
- Governance:
  - Countries with stronger governance (better control of corruption, higher regulatory quality, better rule of law) tend to have higher PPP capital stock per capita (Annex Table 1C).
  - Caveat: this may reflect correlation between good governance and higher per-capita income/capital rather than a causal inclination to favor PPPs.

Selected exact medians from Annex Table 1.B (Per capita PPP capital stock, medians):
- Full sample medians:
  - 1996–2001 = 30.11 [103.64]
  - 2002–2007 = 97.25 [203.35]
  - 2008–2013 = 112.05 [299.77]
  - 2014–2019 = 121.91 [428.71]
- High inflation countries medians (selected):
  - 1996–2001 = 20.48 [63.45]
  - 2002–2007 = 35.64 [145.25]
  - 2008–2013 = 56.55 [191.72**]
  - 2014–2019 = 69.68 [258.84***]
- High income per capita countries medians (selected):
  - 1996–2001 = 63.36*** [143.58**]
  - 2002–2007 = 107.49*** [311.50***]
  - 2008–2013 = 198.67*** [430.42***]
  - 2014–2019 = 390.64*** [566.75**]

(Note: ***, **, * indicate significance at 1%, 5%, and 10% levels respectively as reported in tables.)

### Regression results — cross-country and panel specifications
- Cross-country regressions (Annex Table 2; averages used per episode):
  - General pattern: higher income levels lead to a decline in PPP investments in advanced and emerging market economies (statistically significant mainly in periods following the GFC).
  - During 1996–2001, high inflation led to lower PPP investments in the full sample and in the emerging markets sample.
  - Many results are not robust to changes in country sample or time period.
- Panel regressions (Annex Tables 3A and 3B): three techniques used — panel fixed effects (FE Panel), generalized least squares (GLS), generalized method of moments (GMM).
  - Overall assessment: results are weak and not sufficiently robust.
  - Reported empirical summaries (from panel results):
    - Higher growth leads to a small increase in PPP investments in advanced economies (GMM specification); the opposite found for emerging markets.
    - Inflation leads to fewer PPPs (GLS for overall sample).
    - Higher income level leads to fewer PPPs, suggesting LICs and other countries with infrastructure needs tend to invest more through PPPs (some EM specifications show the opposite).
    - Higher debt levels lead to lower PPPs in advanced economies; opposite signs can be found in EM and LIC estimations.
    - Controls for corruption are associated with more PPPs in all countries (GLS for overall sample and EMs; GMM for AEs).
    - Regulatory quality leads to more PPPs in EMs (GLS) and to less PPPs in AEs (GLS and GMM).
    - Rule of law results are inconsistent across country groups (panel FE).
- Exact model diagnostics and coefficients: Annex Tables 2 and 3 report coefficients, standard errors, observations, R-squared and numbers of countries for each specification and episode (e.g., cross-country Observations range: 69, 113, 62, 29, 51, 44, 93, 2 in the reporting layout; panel regressions report Number of observations such as 1,573; 1,686; 1,492; 1,592 and Number of countries reported as 99, 100, 99, 100, 111 across panels).

### Main interpretative conclusions and limitations
- Most robust positive result: better governance is associated with greater use of PPPs.
- Supporting, but less robust, results:
  - High inflation (macroeconomic instability) adversely affects PPP investments.
  - High debt may be associated with increased PPP use, reflecting fiscal constraints and resorting to PPPs for financing infrastructure.
- Reasons for lack of robustness and empirical challenges:
  - PPP investments are treated similarly to FDI/private investment but PPPs respond to sector-specific government failures, project design, and private financing needs that are not fully captured by macro variables.
  - Political economy and political preferences may drive PPP engagement in specific countries.
  - Successful PPP implementation depends on well-designed governance frameworks across the PPP lifecycle, which are not fully quantified in cross-country macro analysis.
  - Data limitations: PPP investments are recent or limited in many sample countries.
  - Macro analysis did not take into account exact PPP project form or sector risk, which can affect PPP attractiveness.

*IMF Working Paper — Annex IV. Macro-Institutional Determinants of PPPs*

### Annex Table 3.B. Regression Results including Governance  Indicators: Panel Data

### Annex Table 3.B. Regression Results including Governance Indicators: Panel Data

### Regression setup and estimation methods
- Dependent variable: PPP investment, % of GDP
- Estimation types referenced: Panel FE (panel regressions with country fixed effects), GLS (generalized least squares), and GMM (generalized method of moments).
- Significance notation: ***, **, * indicate significance at 1%, 5%, and 10%, respectively.
- Standard errors are in parenthesis.

### Key estimated coefficients (selected regressors across specifications)
- PPP investment, % of GDP, lag:
  - 0.404*** (0.032)
  - 0.405*** (0.031)
  - 0.592*** (0.056)
  - 0.652*** (0.059)
  - 0.740*** (0.041)
  - 0.725*** (0.040)
  - 0.415*** (0.053)
  - 0.420*** (0.052)
- Regulatory Quality, Estimate (multiple specifications):
  - -0.242 (0.215)
  - -0.201 (0.203)
  - -0.003 (0.062)
  - -0.015 (0.060)
  - -0.126 (0.255)
  - -0.169 (0.254)
  - -0.206 (0.156)
  - -0.210 (0.157)
  - -0.220*** (0.070)
  - -0.011 (0.084)
  - -0.141** (0.065)
  - -0.112 (0.069)
  - 0.094 (0.130)
  - 0.124 (0.148)
  - 0.152*** (0.045)
  - 0.165*** (0.043)
  - -0.085 (0.157)
  - -0.127 (0.147)
  - -0.701 (0.549)
  - -0.758 (0.572)
  - 0.016 (0.194)
  - -0.245 (0.177)
  - -0.365 (0.471)
  - -0.266 (0.449)
- Control of Corruption, Estimate (multiple specifications):
  - 0.280 (0.251)
  - 0.290 (0.253)
  - 0.276*** (0.077)
  - 0.276*** (0.076)
  - 0.380 (0.302)
  - 0.483* (0.283)
  - 0.055 (0.135)
  - -0.014 (0.177)
  - -0.014 (0.062)
  - 0.029 (0.070)
  - 0.160** (0.066)
  - 0.168** (0.072)
  - -0.197 (0.149)
  - -0.225 (0.162)
  - 0.124* (0.069)
  - 0.113* (0.067)
  - 0.096 (0.155)
  - 0.035 (0.146)
  - 0.803 (0.559)
  - 0.919 (0.589)
  - 0.290 (0.207)
  - 0.457** (0.200)
  - 0.143 (0.530)
  - 0.199 (0.518)
- Rule of Law, Estimate (multiple specifications):
  - 0.481 (0.315)
  - 0.411 (0.269)
  - -0.241*** (0.093)
  - -0.172* (0.089)
  - 0.431 (0.355)
  - 0.214 (0.343)
  - 0.213 (0.160)
  - 0.477*** (0.137)
  - 0.044 (0.074)
  - 0.160* (0.085)
  - 0.098 (0.088)
  - 0.205** (0.090)
  - 0.131 (0.158)
  - 0.075 (0.228)
  - -0.132* (0.068)
  - -0.071 (0.063)
  - 0.215 (0.192)
  - 0.341* (0.175)
  - 0.972 (0.688)
  - 0.914 (0.656)
  - -0.068 (0.264)
  - -0.107 (0.263)
  - 0.427 (0.530)
  - 0.167 (0.512)
- Real GDP growth, %, lag (multiple specifications):
  - -0.016* (0.009)
  - -0.013* (0.008)
  - 0.005 (0.006)
  - 0.001 (0.006)
  - 0.007 (0.014)
  - 0.007 (0.014)
  - 0.002 (0.004)
  - -0.002 (0.004)
  - 0.009** (0.004)
  - 0.010* (0.005)
  - 0.008*** (0.003)
  - 0.006** (0.003)
  - -0.015** (0.006)
  - -0.010** (0.005)
  - -0.005 (0.005)
  - -0.006 (0.005)
  - -0.009 (0.007)
  - -0.011* (0.007)
  - -0.022 (0.020)
  - -0.021 (0.021)
  - 0.025 (0.015)
  - 0.011 (0.015)
  - -0.002 (0.020)
  - -0.008 (0.019)
- Inflation, %, lag (multiple specifications):
  - 0.005* (0.003)
  - 0.003 (0.002)
  - -0.009*** (0.003)
  - -0.009*** (0.003)
  - -0.002 (0.008)
  - -0.002 (0.008)
  - 0.001 (0.026)
  - -0.002 (0.029)
  - -0.001 (0.013)
  - 0.025 (0.016)
  - -0.000 (0.009)
  - -0.003 (0.010)
  - 0.005** (0.002)
  - 0.002 (0.002)
  - -0.004* (0.002)
  - -0.004* (0.002)
  - -0.005 (0.005)
  - -0.006 (0.005)
  - -0.003 (0.007)
  - -0.001 (0.008)
  - -0.012 (0.010)
  - -0.019** (0.010)
  - -0.005 (0.011)
  - -0.005 (0.011)
- log(Real GDP per capita), lag (multiple specifications):
  - 0.149 (0.264)
  - 0.212 (0.255)
  - -0.347*** (0.056)
  - -0.321*** (0.052)
  - -0.435 (0.764)
  - -0.112 (0.697)
  - -0.331 (0.234)
  - 0.358* (0.177)
  - -0.746*** (0.088)
  - -0.600*** (0.110)
  - -0.345** (0.148)
  - 0.157 (0.119)
  - 0.236 (0.306)
  - 0.277 (0.352)
  - -0.229*** (0.051)
  - -0.198*** (0.048)
  - 0.671* (0.367)
  - 0.510* (0.303)
  - -0.399 (0.558)
  - -0.326 (0.534)
  - -0.128 (0.158)
  - -0.357** (0.145)
  - -0.083 (1.137)
  - 0.291 (1.050)

### Public capital, fiscal, and fiscal-risk controls (selected estimates)
- log(Public capital stock per capita), lag:
  - -0.406** (0.194)
  - -0.463** (0.201)
  - 0.175*** (0.043)
  - 0.179*** (0.041)
  - -1.446*** (0.498)
  - -1.378*** (0.488)
  - 0.705 (0.491)
  - -0.155 (0.446)
  - -0.303*** (0.050)
  - -0.223*** (0.066)
  - 0.153 (0.257)
  - -0.596*** (0.202)
  - -0.413* (0.232)
  - -0.367 (0.227)
  - 0.050 (0.035)
  - 0.031 (0.034)
  - -0.645*** (0.202)
  - -0.650*** (0.195)
  - -0.380 (0.283)
  - -0.374 (0.297)
  - 0.349*** (0.095)
  - 0.344*** (0.092)
  - -0.383 (0.564)
  - -0.482 (0.577)
- GG gross debt, % GDP, lag:
  - -0.002 (0.001)
  - 0.002** (0.001)
  - 0.000 (0.003)
  - -0.008*** (0.002)
  - -0.007*** (0.001)
  - -0.005*** (0.001)
  - -0.003 (0.002)
  - 0.002*** (0.001)
  - 0.003 (0.002)
  - 0.001 (0.003)
  - 0.006** (0.003)
  - 0.000 (0.006)
- GG primary budget balance, % GDP, lag:
  - 0.016 (0.014)
  - -0.003 (0.006)
  - 0.013 (0.012)
  - -0.008 (0.005)
  - -0.021*** (0.004)
  - -0.004 (0.003)
  - 0.008 (0.006)
  - 0.008* (0.005)
  - -0.010 (0.008)
  - 0.027 (0.027)
  - -0.009 (0.014)
  - -0.013 (0.021)
- GG effective interest rate, %:
  - -0.018 (0.034)
  - 0.006 (0.009)
  - 0.001 (0.040)
  - 0.075** (0.030)
  - 0.116*** (0.022)
  - 0.046** (0.023)
  - -0.043** (0.020)
  - 0.022*** (0.007)
  - 0.018 (0.022)
  - 0.021 (0.063)
  - -0.092*** (0.032)
  - 0.045 (0.066)
- log(Fiscal risk index), lag:
  - 0.001 (0.031)
  - 0.056*** (0.018)
  - 0.103 (0.074)
  - -0.012 (0.023)
  - 0.048*** (0.010)
  - -0.000 (0.010)
  - 0.072 (0.056)
  - 0.076*** (0.014)
  - 0.016 (0.031)
  - -0.122 (0.113)
  - -0.131 (0.081)
  - 0.146 (0.155)

### Sample sizes and country coverage
- Number of observations (selected blocks/specifications):
  - 1,284
  - 1,293
  - 1,284
  - 1,293
  - 1,102
  - 1,107
  - 75
  - 169
  - 175
  - 169
  - 158
  - 150
  - 692
  - 695
  - 692
  - 695
  - 606
  - 609
  - 417
  - 429
  - 417
  - 429
  - 338
  - 348
- Number of countries (selected blocks/specifications):
  - 95
  - 96
  - 95
  - 96
  - 111
  - 111
  - 11
  - 11
  - 11
  - 11
  - 11
  - 11
  - 49
  - 49
  - 49
  - 49
  - 35
  - 36
  - 35
  - 36
  - 35
  - 36
  - 35
  - 36

### Notable patterns and statistically significant relationships reported
- The lagged dependent variable (PPP investment, % of GDP, lag) is consistently positive and statistically significant across many specifications (examples: 0.404***, 0.592***, 0.740***, 0.725***, 0.415***).
- Regulatory Quality shows mixed signs and significance across specifications, with at least one strongly significant negative estimate: -0.220*** (0.070), and at least one significant negative: -0.141** (0.065). Other specifications show positive but insignificant estimates (e.g., 0.152*** (0.045) in a subset).
- Control of Corruption estimates vary in sign; examples of significant positive estimates include 0.276*** (0.077), 0.160** (0.066), and 0.168** (0.072); other specifications show insignificant or negative estimates.
- Rule of Law displays both positive and negative significant coefficients depending on specification, including -0.241*** (0.093) and 0.477*** (0.137).
- Real GDP growth, lag, shows small coefficients often with mixed significance (e.g., -0.016* (0.009), 0.008*** (0.003)).
- Inflation, lag, exhibits negative significant coefficients in some specifications (e.g., -0.009*** (0.003)) and small positive coefficients in others (e.g., 0.005* (0.003)).
- log(Public capital stock per capita), lag, includes large negative and significant estimates in several specifications (e.g., -1.446*** (0.498), -1.378*** (0.488), -0.303*** (0.050), -0.223*** (0.066)), and some positive significant estimates in other subsets (e.g., 0.175*** (0.043), 0.179*** (0.041), 0.349*** (0.095), 0.344*** (0.092)).
- Fiscal variables (GG gross debt, GG primary balance, GG effective interest rate, log(Fiscal risk index)) show varied signs and occasional statistical significance across specifications (examples include GG gross debt 0.002** (0.001); GG effective interest rate 0.116*** (0.022); log(Fiscal risk index) 0.056*** (0.018); GG primary budget balance -0.021*** (0.004)).

*Source: Annex Table 3.B. Regression Results including Governance Indicators: Panel Data, The Future of PPPs in the Western Balkans, Working Paper No. WP/23/31*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023031-print-pdf.pdf_