Journal Issue
Share
Article

2018 Review of Program Design and Conditionality—Case Studies

Author(s):
International Monetary Fund. Strategy, Policy, & Review Department
Published Date:
May 2019
Share
  • ShareShare
Show Summary Details

Growth Optimism

A. Lessons

  • Conservative growth forecasts can support program performance. Avoiding optimistic assumptions regarding reform payoffs and the impact of fiscal policy on growth can help meet program targets. Furthermore, inherently volatile growth in small island states warrants a conservative approach.
  • Discussing downside risks and developing contingency plans can help prepare for the potential materialization of downside scenarios. Continuously assessing the macroeconomic framework is important for sustaining program implementation in the face of shocks.
  • Taking advantage of early reform momentum and frontloading reforms can facilitate program success. However, some reforms take longer, and proper sequencing of reforms remains important.
Table 1.Case Studies: Growth Optimism
Program Completion 1/Were initial growth f o recasts conservative?Did the program request staff report discuss fiscal multipliers?Were downside risks sufficiently discussed, including with contingency measures?Did external shocks result in significant deviations from baseline growth projections?Was there a frontloading of reforms?Did program design reflect underlying vulnerabilities?Comments
Cyprus – EFF 2013C
Grenada – ECF 2014C
Mozambique -SCF 2015QOTPartlyPartlyPartly
Solomon Islands – ECF 2012C** Baseline was gradually adjusted to become more conservative.
Ukraine – SBA 2014RPartlyPartly
Source: IMF staff.

Program completion status as of end-September 2018: C = completed all reviews, IP = in progress, L = largely implemented, OT= off track, QOT= quickly off track, R = replaced.

Source: IMF staff.

Program completion status as of end-September 2018: C = completed all reviews, IP = in progress, L = largely implemented, OT= off track, QOT= quickly off track, R = replaced.

B. Cyprus: 2013 EFF

1. A three-year extended arrangement under the Extended Fund Facility (EFF) was requested on the back of increasing imbalances in the run-up to the Global Financial Crisis. The objectives of the authorities’ program were to restore financial sector stability and implement structural reforms to achieve sustainable public finances.

2. The program maintained conservative assumptions regarding growth payoffs. Ultimately, growth surprised on the upside. Despite a severe recession, the growth outturn— a Growth Domestic Product (GDP) decline of close to 6 percent in 2013—was nonetheless less severe than projected and continued to surprise on the upside (Figure 1).

Figure 1.Cyprus: Growth Forecast

Sources: WEO and IMF staff calculations.

3. The forecast internalized the impact of fiscal consolidation on growth. The document requesting the arrangement discussed that fiscal consolidation would adversely impact growth. In turn, citing recent literature and experience from other program countries, staff assumed a one-for-one impact of fiscal consolidation measures on growth (Blanchard and Leigh, 2013; and IMF, 2012a).

4. Staff highlighted significant downside risks and pointed to contingency measures. Downside risks included weaker-than-expected growth, unexpected pressure from the banking system, and weakening ownership. These risks were sufficiently covered as part of discussions on the macroeconomic outlook and program modalities. The staff report also noted that some buffers were built into the financing envelope and capitalization requirements for banks to mitigate downside risks. The government agreed with staff advice to stand ready to take additional measures to preserve the medium-term fiscal balance and debt objectives. Ultimately, significant risks did not materialize, and—as mentioned above—actual outturns largely exceeded growth projections.

5. The program benefitted from early reform momentum and frontloading reforms. The Cypriot authorities were already implementing reforms in the public sector prior to arrangement approval. Subsequently, the program aimed to advance key structural reforms in the financial sector and public sector, while maintaining conservative growth dividends.

6. Program design was appropriate in addressing the high level of public debt and vulnerabilities in the financial sector. Program conditionality focused on fiscal consolidation and financial sector restructuring. Quantitative performance criteria (QPCs) aimed at containing the fiscal deficit and arrears. Structural benchmarks (SBs) were geared toward a restructuring of the banking sector. Out of 34 SBs during the completed 9 reviews, 19 SBs were related to financial sector stability: 9 on financial sector legal reforms, regulation, and supervision; 7 on restructuring and privatization of financial institutions; and 3 on capital account restrictions.

7. The program was completed with significant achievements. Amid a significant fiscal overperformance—driven by strict budget controls, improved labor market conditions, and resumed GDP growth—public debt was put on a downward trajectory. Public debt peaked at around 108 percent of GDP, about 18 percentage points of GDP below projections.1 Cyprus regained access to international capital markets.

8. Yet, some challenges remained. In the banking sector, still-high non-performing loans (NPLs) at 46 percent of total loans at end-2016 remained a concern. Structural reforms, which had moved forward initially, experienced delays toward the end of the program in the politically sensitive areas of private debt restructuring, privatization, and public administration. Delays in implementing the Public Financial Management (PFM) and revenue administration (RA) reforms were less severe and largely related to capacity constraints. While program ownership remained high, the minority position of the governing party in parliament and elections in the middle of the program period complicated the implementation of structural reforms.

C. Grenada: 2014 ECF

9. As the Global Financial Crisis had derailed Grenada’s protracted recovery from a decade of natural disasters and economic shocks, a three-year extended arrangement under the Extended Credit Facility (ECF) was approved in 2014. This followed several Fund supported programs, which were evaluated in the 2014 Ex Post Assessment (IMF, 2014a).

10. Growth projections for the 2014 program were intentionally conservative. The Ex Post Assessment for the predecessor program had highlighted difficulties tackling shocks given institutional limitations, overoptimistic growth projections, overambitious structural reforms, and weak ownership. In turn, the 2014 program internalized these concerns, and growth was expected to remain constrained by large fiscal adjustment and weak credit growth, amid bank balance sheet repair.

11. The program paid close attention to the potential adverse impact of fiscal consolidation on growth. Both staff and the authorities noted that the drag of the large and front-loaded fiscal adjustment could be greater than projected. At the time, the envisaged consolidation of 7¾ percent of GDP was in the 97th percentile of adjustment cases over the preceding two decades for countries with debt over 60 percent of GDP. To minimize the impact of the adjustment, the program aimed at protecting social spending and growth-enhancing capital investments.

12. Downside risks were perceived as high and given significant prominence. Key risks included underestimation of the drag from fiscal adjustment; faltering confidence if the fiscal crisis remains unaddressed; continued weak global recovery; natural disasters; and fragile regional markets. Based on historical variance of growth outcomes, staff estimated the probability of negative growth during the program at about 40 percent. Taking into account these risks, the program identified contingency measures on both revenues and expenditures. Furthermore, at the time of the request for the arrangement, the government noted its plans to purchase additional natural disaster insurance for the duration of the program to ensure that natural disaster shocks would not derail its implementation.

13. External shocks did not materialize, and growth remained strong. Growth overperformed markedly relative to projections, reaching more than 7 percent in 2014 and averaging 5½ percent during 2014–17, reigniting confidence (Figure 2). The robust performance was underpinned by tourism, related construction, agriculture, and continued growth in private education, and was supported by the U.S. expansion. Fiscal adjustment (9½ percent of GDP), debt restructuring, and private balance sheet repair unfolded as planned. The process was underpinned by strict implementation of the new Fiscal Responsibility Law.

Figure 2.Grenada: Growth Forecast

Sources: WEO and IMF staff calculations.

14. Fiscal consolidation was aided by the frontloading of measures. About three quarters of the adjustment was planned for the first two years of the program. Most revenue measures that were needed for the entire adjustment were pre-approved, focusing on widening the tax base.

15. Program performance was strong. The program was fully completed, with 93 percent of QPCs met and 75 percent of SBs met or implemented with delay. The focus of structural conditionality on the underlying institutional causes of fiscal weaknesses and extensive technical assistance (TA) contributed to success. Moreover, conservative growth forecasts supported program performance. However, implementation of some SBs was delayed due to capacity constraints. Overall, the positive outcome exceeding expectations was a product of program design—including to address underlying vulnerabilities—and the fact that risks did not materialize.

D. Mozambique: 2015 SCF

16. In 2015, an 18-month Standby Credit Facility (SCF) was approved, supplementing the existing program under the Policy Support Instrument (PSI). This occurred in response to what was seen as a temporary commodity price shock and a decline in foreign exchange inflows. Hence, program objectives included addressing the BoP need and building reserves to strengthen external buffers.

17. The program assumed a robust growth recovery, supported by large investment in natural gas projects and higher coal production. Following a steady strengthening after arrangement approval, medium-term growth was projected to recover to 7½–8 percent, supported by large investments in natural gas projects and higher coal production (Figure 3). Reflecting the limited planned fiscal consolidation, discussions of the size of fiscal multipliers were not included in the document requesting the arrangement.

Figure 3.Mozambique: Growth Forecast

Sources: WEO and IMF staff calculations.

18. Several reforms were frontloaded, as were disbursements. A strong fiscal adjustment policy package (at the outset fiscal adjustment included a 1 percent revenue increase and about 1.5 percent of GDP spending reduction), continued exchange rate flexibility, tight liquidity management, and structural reforms were implemented under the SCF. To cushion the external BoP shocks and support the government’s goals on poverty reduction and inclusive growth, disbursements under the SCF were frontloaded at 75 percent of quota.

19. The staff report identified substantial downside risks to the outlook as well as to the program. Risks to the outlook included (i) persistently weak international commodity prices; (ii) significant slowdowns in China and other key economies, which would delay coal mining expansion; and (iii) further delay in the construction of liquefied natural gas (LNG) plants. Program-specific risks were also highlighted: (i) political instability eroding program ownership; (ii) further deterioration of the external environment; (iii) delays in the negotiation of large investment projects in the natural resource sector; (iv) adverse weather conditions which would significantly affect agricultural production; and (v) capacity constraints that can delay reform implementation (including on revenue administration and budget execution). However, the staff report included only a very brief mention of contingency plans in the event of increasing international oil prices.

20. Downside risks materialized, and most macroeconomic outcomes deviated markedly from projections. In 2016, undisclosed borrowing (equivalent to about 11 percent of GDP) by the Proindicus and Mozambique Asset Management public companies emerged and led to suspension of donor budget support. In addition, Mozambique was exposed to a series of shocks, including lower commodity prices and adverse weather conditions. In turn, growth declined, and public debt rose to an unsustainable level.

21. Program design did not fully factor in vulnerabilities, while weak ownership also played a role. The program did not appear to fully reflect shortfalls in program implementation (against assessment criteria under the 2013 PSI) and in debt management capacity. However, the emergence of undisclosed debts points to weak ownership as a key factor that inhibited program success.

22. The program went quickly off track. The disclosure of hidden public debt highlighted governance concerns. In April 2016, this resulted in misreporting under the PSI and a breach of obligations under Article VIII, Section V. The sixth review under the PSI and the first review under the SCF arrangement did not take place as scheduled and both the PSI and the SCF lapsed.

E. Solomon Islands: 2012 ECF

23. The ECF was intended to tackle deep institutional and structural issues and followed a successful one-year precautionary arrangement under the SCF. The ECF emphasized the need for larger fiscal and external buffers and increased resilience (via climate adaptation and mitigation). Institutional capacity was a constraint, particularly in the context of the withdrawal of RAMSI (regional assistance).

24. Growth projections were complicated by the strong rebound that preceded the program. After rebounding strongly from the 2009 recession, staff factored in a gradual moderation over the medium term (Figure 4). The documents requesting the arrangement did not include a discussion of the impact of fiscal consolidation on growth. Overall, gold mining and infrastructure spending were expected to be the main growth drivers, offsetting a continued decline in logging.

Figure 4.Solomon Islands: Growth Forecast

Sources: WEO and IMF staff calculations.

25. Staff noted that risks were tilted to the downside, albeit initially not all key risks were reflected. Risks emanated from a slowdown in China and other Asian emerging market trading partners, which could reduce commodity and forestry exports; domestic political uncertainty; and slippages in implementing structural reforms. Correspondingly, the report discussed contingent monetary and exchange rate policies, as well as a downside scenario, which could potentially justify augmentation of access under the low-access ECF. However, key areas such as vulnerability to natural disasters and dependency on donor aid were not mentioned at arrangement approval.

26. The materialization of risks led to disappointing growth outcomes. Growth fell short of projections owing to unfavorable weather, lower export prices, reduced gold production (which was impacted by massive floods), and a larger-than-expected decline in logging and donor support. Subsequently, potential output was revised down from 4 percent at the time of the request for the arrangement to 3.5 percent early on in the program, and further to 3 percent at the end of the program.

27. Gradually, the program adopted a more conservative baseline. Amid downside shocks, staff gradually recalibrated the baseline to better factor in vulnerabilities, internalizing growth disappointments and implementation experience. Risks became more balanced, reflecting the ongoing decline in logging and donor support, weaker prospects for mining, and recognition of the impact of frequent natural disasters. At the final review, estimates suggested that losses/damages from natural disasters reduced GDP by 0.3 ppts in the year of the disaster.

28. Program performance was solid. Continuously assessing the macro-framework became important for sustained program implementation in the face of external shocks, and the program was fully completed. About 90 percent of QPCs were met, in part reflecting a significant loosening of cash balance targets to accommodate the authorities’ flood response. While there was some progress made on structural reforms, implementation was significantly hampered by capacity constraints. Roughly half of SBs were met (or implemented with delay), with the other half still outstanding at program completion.

F. Ukraine: 2014 SBA

29. In a context of heightened geopolitical tensions and sizable losses in the energy sector, a two-year Stand-By Arrangement (SBA) was approved in April 2014. The objectives of the program were to restore macroeconomic stability, strengthen economic governance and transparency, restore sound public finances, improve the business environment, and lay the foundation for robust and balanced economic growth.

30. Growth projections were optimistic considering the prevalence and severity of downside risks. The program had factored in a growth decline in 2014, followed by a V-shaped recovery (Figure 5). Downside risks were perceived to be significant, however, stemming from (i) uncertainty about policy implementation given the presidential election cycle; (ii) resistance toward governance reforms from vested interests; and (iii) geopolitical tensions related to developments in the East. While acknowledging the impact of fiscal tightening on growth, the staff report did not discuss the size of fiscal multipliers.

Figure 5.Ukraine: Growth Forecast

Sources: WEO and IMF staff calculations.

31. Risks were well-recognized but their extent was underestimated. For example, the adverse scenario at the time of the first review showed the debt ratio remaining above the 70 percent risk threshold over the medium term (but well below their outturn, see below) if the conflict in the East were to intensify. In hindsight, the Ex Post Evaluation (IMF, 2016) notes that an earlier debt operation could have helped the financing envelope, creating policy space to meet policy objectives.

32. As most downside risks materialized, macroeconomic outcomes deviated fundamentally from projections and public debt dynamics worsened. Soon after arrangement approval, the conflict in the East led to a disruption of trade and industrial production, loss of confidence, and capital outflows. The sharp exchange rate depreciation put banks under increasing stress. By end-2014, Ukraine’s public debt was 15 percentage points higher than projected under the program; the latest estimates for the 2015 debt ratio were 30 percentage points higher than envisaged at program approval. This put the program at the highest debt surprise for crisis programs (among non-restructuring cases) following one year from arrangement approval. The main driver was the larger-than-expected hryvnia depreciation, as two-thirds of debt was denominated in foreign currency. The depreciation also added to contingent liabilities by raising bank recapitalization needs and financing requirements for the state-owned gas company Naftogaz. GDP declined markedly.

33. Conditionality had been comprehensive and front-loaded, supporting program objectives while seeking to mitigate implementation risks. The program significantly frontloaded structural conditionality, including with 12 prior actions (PAs) at approval. Structural conditions covered a broad range of policy areas, including energy, fiscal and financial sector policies, as well as governance reforms. As such, reform plans were more comprehensive than in Ukraine’s past programs or other contemporaneous exceptional access arrangements.

34. Amid the materialization of risks and weak program performance, the arrangement was canceled, and a new arrangement approved. The documents requesting the arrangement in 2014 did not contain clearly articulated adverse scenarios with contingency plans, nor did it internalize key underlying structural challenges, which required a longer adjustment period than what could be achieved under the SBA. Amid downside shocks, program performance was weak. Less than one third of QPCs were met. The Ex Post Evaluation (IMF, 2016) noted that the baseline appeared to be a best-case scenario subject to high risks. Ukraine’s balance of payment (BoP) and adjustment needs increased beyond what could be achieved under a two-year program. The SBA was cancelled, with only the first review completed, and a four-year EFF was approved in February 2015, reflecting the longer-term nature of the needed adjustment and structural reforms (see Section II).

Quality of Fiscal Adjustment

A. Lessons

  • Fiscal consolidation should be mindful not only of the size but also of the composition of the adjustment. Improvements in fiscal institutions and PFM reforms can be important complements. Granular conditionality on fiscal targets and reforms, when implemented, helped achieve the targeted composition of fiscal adjustment.
  • Comprehensive reform strategies are beneficial to domestic revenue mobilization. Multi-year reform strategies can be effective in mobilizing revenue. These include measures to strengthen basic institutions, broaden the tax base, and modernize tax administration.
  • Cushioning the adjustment impact on the poor is essential. To this end, defining clear objectives to alleviate social costs—combined with carefully designed conditionality in accordance with the goal—can help ensure its implementation. Careful attention to mitigating adverse impacts of reforms on the poor and vulnerable groups can also help build political support for structural reforms. Collaboration with development partners can strengthen program design and tailoring.
Table 2.Case Studies: Quality of Fiscal Adjustment
Program Completion 1/Was the composition of fiscal adjustment as envisaged?Did reforms effectively mobilize domestic revenue?Was capital spending protected?Did the program have conditionality to support the composition of adjustment?Did program design pay appropriate attention to social objectives?Did collaboration with development partners support program design to protect vulnerable groups?Comments
Bangladesh -ECF 2012C
Jamaica – EFF 2013C
Mauritania -ECF 2010C
Serbia – SBA 2015C2/*Mostly* Amid high ownership, flexibility allowed for changing the composition of fiscal adjustment.
Tunisia – SBA 2013L
Ukraine – SBA 2014/EFF 2015R/ongoing
Source: IMF staff.

Program completion status as of end-September 2018: C = completed all reviews, IP = in progress, L = largely implemented, OT= off track, QOT= quickly off track, R = replaced.

Source: IMF staff.

Program completion status as of end-September 2018: C = completed all reviews, IP = in progress, L = largely implemented, OT= off track, QOT= quickly off track, R = replaced.

B. Bangladesh: 2012 ECF

35. A three-year ECF was approved in April 2012 following intensifying macroeconomic pressures and loss of international reserves. Program objectives were to restore macroeconomic stability, strengthen the external position, and engender higher, more inclusive growth. The program aimed to end the status quo of low tax revenues and low capital spending. Energy subsidy reform was included as a key factor to strengthen the fiscal position.

36. The composition of fiscal adjustment differed markedly from plans, with revenues and capital spending underperforming (Figure 6). The program aimed at a moderate fiscal consolidation over the medium term, allowing space to ramp up Annual Development Program (ADP) spending. Domestic revenues were to be lifted by strengthening revenue administration and broadening the tax base. However, amid weak revenue outturns and delays in VAT implementation, revenue mobilization disappointed. Capital spending under the ADP by 2015 turned out lower than planned by more than 1½ percent of GDP, in part due to weak implementation capacity.

Figure 6.Bangladesh: 2012 ECF 1/

Sources: WEO and IMF staff calculations.

1/ T=0 is the year of program approval.

37. Conditionality was targeted to support the core objectives of the program. It introduced ITs, PAs, and SBs to increase tax revenue and to protect social spending. The program had an IT (floor) on tax revenue and an IT (floor) on social spending. PAs and SBs focused on the introduction of a new VAT law, removal of tax concessions and exemptions, tax RA reforms, and energy subsidy reform. Notably, the adoption of an automatic fuel price adjustment mechanism aimed to free up resources to compensate vulnerable households from rising fuel and food prices. No conditionality was introduced to specifically preserve capital expenditure.

38. The program focused on safeguarding social spending, building on advice from development partners, including the World Bank. In the documents requesting the arrangement, Fund and World Bank staff prepared a joint note detailing the type of social safety net measures in place and assessed the effectiveness in mitigating the impact of higher food and energy prices. Further, it took stock of existing social safety net programs, aiming at identifying weaknesses in the current system. The World Bank also provided support to construct a poverty database, helping to streamline social safety net programs and improve their targeting.

39. Despite some delays toward the end of the program, which prompted two arrangements extensions, the program was ultimately fully completed. Performance against quantitative criteria and ITs was generally strong. The IT on social spending was met in all (but one) reviews. Social spending was generally protected and maintained as a share of GDP. Poverty continued to come down, despite the cut in energy subsidies. Efforts were also made to strengthen the efficiency and transparency of social safety net programs, including by developing a comprehensive poverty database. However, the IT on tax revenue was consistently missed by small margins throughout the program, while most SBs were implemented, albeit often with delays. Significant delays and uncertainty in the implementation of the new VAT prompted two three-month extensions of the program. The delays in tax measures contributed to a suboptimal composition of fiscal adjustment.

C. Jamaica: 2013 EFF

40. Jamaica requested a four-year EFF in 2013. The 2013 EFF aimed to avert immediate crisis risks and create the conditions for sustained growth through a significant improvement in public debt and competitiveness. The program included structural reforms to boost growth and employment; actions to improve price and non-price competitiveness; upfront fiscal adjustment and extensive fiscal reforms; debt reduction; and improved social protection programs.

41. Upfront and sustained fiscal tightening was achieved under the program (Figure 7). The adjustment was supplemented with a debt exchange to place public debt on a sustainable path, while protecting financial system stability and improving social protection programs. The program targeted a central government primary surplus of 7½ percent, with an upfront adjustment of 2 percent of GDP and a balanced budget for the public entities throughout the program period. Fiscal adjustment included a combination of revenue enhancing and expenditure reduction measures. On revenues, tax measures aimed to broaden the tax base and equalize rates, as well as increase rates and fees. Expenditure measures included a multiyear wage agreement to limit the nominal wage increase and reduce transfers to local governments.

Figure 7.Jamaica: 2013 EFF

Sources: WEO and IMF staff calculations.

1/ T=0 is the fiscal year of program approval.

42. The program effectively mobilized domestic revenue but fell short of plans on wages and capital spending. Efforts to overhaul tax administration, eliminate tax incentives, and to generally rebalance the tax system from direct to indirect taxation (e.g., consumption tax, higher excises on fuel) helped Jamaica overperform revenue targets. Capital spending was initially protected and the multiyear wage agreement created some fiscal space but both fell short of projections over the course of the program.

43. The program included conditionality to support the composition of adjustment. Along with a performance criterion limiting the overall deficit of the wider public sector, the program included an IT on tax revenues to underline the importance of improving tax administration and tax reform, and to help avoid reliance on ad hoc spending cuts for meeting the fiscal targets. Social expenditures were protected through a floor on targeted social spending.

44. Program documents discussed in detail the authorities’ commitment to reduce the adverse impact of adjustment on vulnerable groups. The authorities planned to (i) improve training and certification for labor market participants to tackle the high unemployment rate; (ii) enhance benefits and improve effectiveness and targeting under the social assistance program PATH, a conditional cash transfer program; and (iii) implement welfare-to-work exit strategies for vulnerable households. Conditionality was designed in accordance with the goal, including not only quantitative targets on social spending floors with clear definitions but also adding SCs to strengthen education and health spending and social safety nets. Overall, social spending was increased under the 2013 EFF.

45. The program leveraged TA by the IMF and other development partners. IMF TA aimed at improving the efficiency and targeting of social spending. Coordination with other development partners, such as the World Bank and the Inter-American Development Bank, also supported the program. Policy advice on social spending included detailed suggestions to improve the efficiency and effectiveness of education and health spending, including by better balancing the student-teacher ratio within and between schools and improving the efficiency of health care.

46. The program was completed while successfully mitigating the effects of fiscal consolidation. Nearly all fiscal QPCs were met SBs on institutional fiscal reforms, tax reform, and tax administration were met with few exceptions. Conditionality was clearly defined. Internal and external expertise supported the design of social sector reforms, helping strengthen program design and conditionality for example through an SB on improving data monitoring. In addition, ownership was strong, ensuring the implementation of program measures, including to improve social spending.

47. However, some challenges remained. These relate to low growth, high poverty and unemployment, and crime and security issues. In turn, Jamaica requested a 36-month precautionary SBA in 2016 to provide a backstop against adverse external economic shocks. In addition to reform efforts to deliver better growth and job outcomes and reduce poverty, the program focused on transitioning toward inflation targeting and exchange rate flexibility as well as on strengthening Jamaica’s institutions.

D. Mauritania: 2010 ECF

48. Prior to the three-year 2010 ECF, Mauritania had several previous programs, with only a few reaching completion. Factors at play included (i) food and fuel price increases; (ii) the global economic slowdown; (iii) a domestic political crisis; (iv) a decline in donor financing; and (v) a significant deterioration in the fiscal position.

49. The 2010 program marked a turnaround. The program was initiated after a peaceful presidential election, which established a strong basis for a resumption of the reform agenda and financial support from the international community. A key program objective was to reduce public debt and create fiscal space for social and infrastructure spending. On the expenditure side, the program envisioned lowering the wage bill and reducing transfers to SOEs. While no increase in revenue was envisaged, the program included measures to broaden the tax base, curtail exemptions, and strengthen tax and customs administration.

50. A number of complementary reforms resulted in significant revenue overperformance. Efforts to mobilize revenues included the VAT regime in mining (IMF, 2015), the abolishment of corporate income tax exemptions, the replacement of the global income tax with a dual tax system, and the introduction of a withholding tax of 15 percent on payments to nonresidents to protect the tax base and reduce profit-shifting abroad. TA to improve taxpayer registration and build audit capacity in the large and medium-sized taxpayer offices aimed to minimize revenue leakages. As a result, tax revenue collection improved markedly.

51. The composition of fiscal adjustment became more growth-friendly than envisaged (Figure 8). Revenue overperformance allowed for a notable scaling-up of the public investment program and infrastructure projects related to energy, roads, water, and agriculture. This helped support growth (IMF, 2012b). While the wage bill and transfers to SOEs were reduced, a modest loosening of the fiscal stance became necessary to mitigate the impact of higher fuel and food prices on the poor through an increase in food subsidies (drought emergency program). The repayment of domestic arrears (accumulated in 2011) and additional electricity subsidies (following the delay of the planned review of electricity tariffs) added to expenditures.

Figure 8.Mauritania: 2010 ECF 1/

Sources: WEO and IMF staff calculations.

1/ T=0 is the year of program approval.

52. Program conditionality supported the composition of adjustment. Quantitative targets (a ceiling on net domestic assets of the central bank and a floor on the balance of government non-oil operations) helped support fiscal sustainability. In addition, SBs covered domestic revenue mobilization (tax base broadening, strengthening tax administration); improved budget execution; public service reform to contain the wage bill; energy subsidy reform; and reforms to reduce budgetary transfers to SOEs.

53. Strengthening social protection and safety nets was a stated program objective. With the support of UNICEF, a study on social protection was launched to review the social safety nets in place, assess priority needs, and issue recommendations to guide the government and its partners in preparing the Poverty Reduction and Strategy Paper (PRSP). Based on the study’s recommendations, the program promoted the exit from a surge in subsidy payments through the replacement of existing subsidy schemes by well-targeted and more cost-effective social safety nets. The program envisioned raising poverty-reducing expenditures and had an IT on poverty-related expenditures and an SB on social policy to better target protection of vulnerable households.

54. Overall program performance was strong. The ECF was fully completed, with all QPCs and most SBs (related to the fiscal composition and social policy) met. At the end of the program, the authorities had achieved a more prudent fiscal balance and improved policy space.

E. Serbia: 2015 SBA

55. When the three-year precautionary SBA was approved in 2015, Serbia’s economy faced large fiscal imbalances and protracted structural challenges. Concerns arose from (i) declining revenues, despite tax rate hikes; (ii) rising mandatory spending, especially public wage and pension bills; (iii) expanding state aid to ailing SOEs, usually in the form of direct subsidies and guarantees for borrowing; and (iv) the cost of resolving ailing public banks.

56. The program envisaged sizable fiscal consolidation (Figure 9). Consolidation plans of more than 4 percent of GDP focused on durable expenditure measures (curbing mandatory spending and reducing state transfers to SOEs), as tax rates had been raised during the two years prior to the program. Efforts to improve tax collection efficiency and broadening the tax base—while taking a cautious approach to the assumed fiscal gains in the macroeconomic framework—were also envisaged. Prior to arrangement approval, the authorities had initiated reforms of the SOE sector, which fed into program design.

Figure 9.Serbia: 2015 SBA 1/

Sources: WEO and IMF staff calculations.

1/ T=0 is the year of program approval.

57. Over time, the fiscal adjustment mix shifted, and outturns exceeded plans. At the outset, a 7 percent of GDP adjustment in primary current spending was envisaged, while revenues were to decline by more than 2 percent of GDP. Over time, the composition was adjusted with both revenues and primary current spending contributing equally to a fiscal adjustment of 6 percent of GDP. This was facilitated by positive growth surprises—partly on the back of strengthened confidence and financial sector intermediation—which supported revenue overperformance. Value Added Tax (VAT), corporate income tax (CIT), and excise taxes played an increasingly important role. Current expenditures (particularly wage and pension expenses and state transfers) were contained, and capital spending was broadly protected relative to programmed projections. However, capacity constraints prevented the full execution of additional public investment enabled by revenue overperformance.

58. Program conditionality focused on limiting fiscal risks and strengthening institutions to support the envisaged fiscal consolidation. A QPC (ceiling) was included to constrain current primary expenditure. In addition, SBs aimed to strengthen the public wage system, reduce budget subsidies and state guarantees, and improve PFM and tax administration. However, some SBs were implemented with a delay or had limited impact.

59. With expenditures on social programs roughly comparable with that of peers, the program supported improvements in the existing social safety net. While there was no explicit conditionality related to social protection, the program supported amendments to the Law on Social Protection with the aim to improve the effectiveness and targeting of cash welfare allowances. In light of concerns regarding the social impact of reforms—in particular related to electricity tariff increases—the program pointed to the World Bank’s assistance in lessening such impact by improving the efficiency of social spending and safety nets.

60. The program was successfully completed. All reviews were completed with all QPCs met. Serbia succeeded in addressing macroeconomic imbalances and restoring confidence and growth. Fiscal sustainability was restored, and the external position realigned with fundamentals. Success factors included conservative initial growth forecasts and—supported by strong ownership—the flexible implementation of durable fiscal adjustment, which largely protected capital spending. With no BoP need at the end of the program, Serbia requested Fund engagement through a Policy Coordination Instrument (PCI) to support the authorities’ continued reform agenda.

F. Tunisia: 2013 SBA

61. The 2013 two-year SBA was designed to assist Tunisia in maintaining its macroeconomic stability in the aftermath of the Arab Spring. It incorporated an ambitious reform agenda to support private sector development, tackle high unemployment, and reduce regional disparities.

62. The program targeted expenditure cuts combined with revenue-neutral reforms (Figure 10). Reductions in primary current spending were envisaged to carry the bulk of the planned adjustment. Amid favorable tax collection relative to GDP and relative to peers, the focus on the revenue side was largely on revenue-neutral reforms. These included (i) rationalizing tax benefits/incentives; (ii) addressing the regressivity of the income tax system; (iii) simplifying the indirect tax system (particularly excises); (iv) strengthening transparency; and (v) improving tax administration to address a fragmented revenue collection process and complex procedures.

Figure 10.Tunisia: 2013 SBA 1/

Sources: WEO and IMF staff calculations.

1/ T=0 is the year of program approval.

63. The composition of fiscal adjustment was ultimately markedly different than planned. Current expenditures increased beyond the programmed level, primarily owing to an increase in the wage bill. Instead, a significant under-execution of the capital budget—with public investment reaching historically low levels—contributed to containing overall expenditures. This exacerbated the growth slowdown, which had already disappointed relative to projections. To support the overall adjustment, a tax reform strategy was introduced toward the end of the program, including permanent base broadening measures to help raise revenue and promote greater efficiency, equity, and simplification. Nonetheless, lower oil production and weaker economic activity, in particular in 2015, became a drag on revenues, which ultimately underperformed.

64. Conditionality to support the adjustment included primarily fiscal structural measures. The Tunisian program included 16 fiscal SBs—mainly to strengthen revenue collections but also to help reduce expenditures. On the revenue side, reforms focused on (i) reforming the tax system to enhance equity and efficiency; (ii) introducing a new tax code; and (iii) strengthening tax administration. On the expenditure side, reforms aimed at addressing financial weaknesses in State-Owned Enterprises (SOEs) and reducing energy subsidies. An indicative target (IT) (ceiling) was initially set on current primary expenditure; later, it was converted to a QPC as the composition of the adjustment worsened.

65. The program put emphasis on reducing inequality and strengthening social protection. To reduce income disparities, the program aimed to assess the social impact of the envisaged reforms and strengthen social assistance mechanisms. Furthermore, the existing cash transfer system for the poor was expanded to cover families at twice the 2010 level—close to 60 percent of the estimated poor. The program also introduced a social electricity tariff to protect poorer households as part of a household compensation strategy to accompany energy subsidy reform. This was informed by a World Bank study, which had assessed the impact of an increase in energy prices on vulnerable households and different productive sectors.

66. The program was largely completed, but with many implementation shortfalls. All but one review was completed. Nonetheless, frequent changes in government, social tensions (including strikes and work stoppages), and security tensions (impacting tourism and growth more broadly) added challenges to reform implementation (see also Section IV). Amid opposition by vested interests and a resulting search for consensus, reforms were often less ambitious than planned, with a number of SBs not being met. This contributed to a suboptimal composition of fiscal adjustment.

G. Ukraine: 2014 SBA and 2015 EFF

67. The 2014 SBA and subsequent 2015 EFF aimed to restore macroeconomic stability, in particular public finances, and strengthen economic governance and the business environment. Following sizable losses in the energy sector, the 2014 SBA aimed at laying the foundation for robust and balanced growth. However, as longer-term adjustment and reform needs were necessary to reach program objectives, the two-year SBA program was canceled in 2015 and a four-year EFF was approved (see Section I).

68. The 2014 and 2015 programs envisaged sizable fiscal adjustment (Figure 11). The programs aimed to strengthen fiscal sustainability through expenditure-led adjustment and frontloaded price increases to reduce energy subsidies, thereby bringing gas and heating prices to cost recovery. Revenue mobilization was less prominent. In the 2015 EFF, fiscal adjustment plans included the elimination of Naftogaz’s (Ukraine’s state-owned oil and gas company) deficit by 2017. This adjustment was supported by two performance criteria (a ceiling on the cash deficit of the general government and a ceiling on the cash deficit of the general government and Naftogaz) and an IT (a ceiling on the current primary expenditure of the state budget). Significant pension savings were to be achieved by effectively freezing pensions during most of 2015 and tightening eligibility for early retirement.

Figure 11.Ukraine: 2015 EFF 1/

Sources: WEO and IMF staff calculations.

1/ T=0 is the year of program approval.

69. Fiscal consolidation was carried out broadly as planned in terms of size and composition. Supported by spending control and bold increases in energy tariffs in 2015 and 2016, the overall fiscal deficit (including the energy sector’s quasi-fiscal deficit), which had swelled to 10 percent of GDP in 2014, declined to just above 2 percent of GDP in 2017. Capital spending was broadly protected at the level envisaged at the beginning of the program. Notwithstanding a recession in 2015, strong policies and fiscal discipline helped restore confidence and reignite growth in 2016.

70. Program design included notable steps to improve the targeting of social spending. With gas and heating prices set to increase, both programs highlighted measures to mitigate the impact of higher energy prices on the poor and most vulnerable. The 2014 SBA included a PA for the government to introduce a new social assistance scheme to protect vulnerable households. The 2015 EFF considered areas where efficiency could be improved, including by improving the targeting of existing social assistance programs. The program also aimed at opening the healthcare sector to private financing and gradually moving to a medical insurance system, as well as improving the quality and efficiency of education spending. Progress was achieved on multiple fronts, including by improving social assistance programs and education.

71. Collaboration with the World Bank and other development partners supported Ukraine’s efforts by helping to tailor measures to specific needs. With the objective of reaching stronger, inclusive growth, program design included reforms across several sectors. In healthcare and education reforms, the authorities drew from advice by the World Bank and other specialized institutions. The World Bank also assisted in developing a comprehensive pension reform package, aiming at putting pension fund finances on a sustainable footing, although implementation was not fully successful.

72. After the approval of the EFF, performance strengthened. Amid the materialization of downside risks, only one review had been completed under the SBA before it was cancelled and the EFF was approved to help Ukraine transform its economy. Performance under the 2015 EFF improved, with generally strong implementation of fiscal adjustment. However, structural reform implementation proved challenging. Despite some initial success in advancing energy and banking sector reforms and setting up anticorruption institutions, reforms (e.g., anticorruption efforts, restructuring and privatization of SOEs, land reform) increasingly faced resistance, causing delays in the completion of reviews. Ultimately, the four-year EFF was cancelled in December 2018 and a new 14-month SBA approved to help cover Ukraine’s BoP needs by bolstering confidence, unlocking external financing, and anchoring economic policies during the 2019 election period.

Public Debt

73. Case studies are based on PRGT programs that saw large public debt projection errors, which did not involve debt restructuring and were approved before end-2015. Projection errors were calculated by comparing public debt projections at program approval T through T+3 with actual outcomes over the same period, drawing on data from the program approval Debt Sustainability Analysis (DSA) and the most recent DSA available. Drivers of projection errors were then decomposed into (i) primary balance; (ii) real interest rates; (iii) real GDP; (iv) exchange rate (valuation) effects; and (iv) any other factors, including residuals.2

A. Lessons

74. The studies shed further light on the factors driving the large debt projection errors and highlight broader lessons for program design. The RoC looked at a diverse sample of the programs with very large errors, ranging from those that went off track quickly (The Gambia and São Tomé and Príncipe) to those that continued to completion (Malawi, Niger, and Rwanda) (Figure 12). Key findings include:

  • Underlying drivers of residuals in projection errors were partially identified, but a large part remains unexplained. In The Gambia, for example, only around 40 percent of the residuals is attributable to the off-budget transfers to parastatals. In all cases, it is important to fully reconcile flow and stock changes ex-post.
  • Over-optimistic program baselines were a key factor behind projection errors in some cases. In the Rwanda and São Tomé and Príncipe programs, debt overshooting was driven by factors that were somewhat anticipated at approval (i.e., planned investment and risks to oil production, respectively) but only reflected in alternative scenarios rather than the baseline. This finding is consistent with the broader finding that macroeconomic assumptions were too optimistic and raises questions about whether these plans or risks should have been included in a more balanced baseline. In general, the baseline scenario should be chosen conservatively. In addition, in some cases it may be reasonable to partially integrate alternative scenarios into the baseline, for instance for a partial execution of an investment scenario or a failure of bringing oil production online.
  • The materialization of contingent liabilities and off-budget guarantees was a recurrent theme. In The Gambia and Malawi programs, contingent liabilities and off-budget guarantees were an important driver of the higher-than-projected debt. This highlights the need for more transparency in debt instruments, including appropriate coverage of debt, and consideration of contingent risks in program design.
  • Conditionality evolved in response to shocks, but some programs appear to have repeatedly accommodated fiscal slippages relative to the revised targets. In the completed programs, conditionality was updated to incorporate additional borrowing, along with further adjustment. All five programs at the outset also included zero non-concessional borrowing (NCB) limits, though in most cases the limit was loosened or breached.3 In Rwanda, debt overshooting appears justified as higher borrowing mainly financed a scaling-up in public investment against the background of good program performance. While debt overshooting in Malawi and Niger was in part explained by exogenous shocks, significant fiscal slippages and rising contingent liabilities in excess of the relaxed program targets (which were repeatedly accommodated) appear to have been important additional factors. This raises the question of whether policy slippages should be accommodated, including through higher debt limits.

Figure 12.Public Debt Projection Errors

Sources: IMF staff calculations.

Table 3.Case Studies: High Public Debt Overshooting in Other Developing Programs
Decomposition of debt projection error 1/Drivers
CountryProgram Completion 2/TotalPrimary deficitReal interest rateReal GDP growthExchange rate depreciationOtherExogenousDomesticComments
WeatherCommodity pricesSecurityOtherPolicy slippageContingent liabilities / guaranteesGovernance / PFMOther
Gambia, The -ECF 2012QOT3611113-414Severe droughts; policy slippages; materialization of contingent liabilities in the form of off-budget transfers to parastatals.
Malawi – ECF 2012C32414419Cashgate scandal caused large spillovers to the exchange rate and growth; deterred donor support; poor spending controls also played a role.
Niger – ECF 2012L88-535-2Drought; decline in commodity prices; deterioration of the security environment; failure of envisaged oil export pipeline to materialize.
Rwanda – PSI 2013IP21100614Currency depreciation due to commodity price shock; higher-than-anticipated investment in the form of off-budget guarantees.
São Tomé and Príncipe – ECF 2012OT35-11013330Oil production did not materialize, resulting in lower-than-anticipated growth and debt-financing (instead of drawdown of National Oil Account).
Sources: IMF staff calculations.

Percentage points of GDP; cumulative from T to T+3.

Program completion status as of end-September 2018: C = completed all reviews, IP = in progress, L = largely implemented, OT= off track, QOT= quickly off track.

Sources: IMF staff calculations.

Percentage points of GDP; cumulative from T to T+3.

Program completion status as of end-September 2018: C = completed all reviews, IP = in progress, L = largely implemented, OT= off track, QOT= quickly off track.

B. The Gambia: 2012 ECF

75. Public debt surged during and after The Gambia’s 2012 ECF, ending 36 percentage points of GDP higher than anticipated.4 The projection deviation between approval and the third year of the program can be decomposed as follows (Figure 12):

  • Primary deficit. Primary balance slippages account for 11 percentage points of the error.
  • Real interest rates. Another 11 percent of the projection error was caused by higher than expected real interest rates, due in part to a sharp increase in domestic borrowing.
  • Residuals. Finally, 14 percentage points of the error is accounted for by residuals, reflecting in part off budget transfers to parastatals (4 percent) and an unexplained component (10 percent).
  • Exchange rate and growth. Effects from a greater than anticipated rate of exchange rate appreciation and lower than projected growth were largely offsetting.

76. Despite the significant increase in the debt ratio, the risk of debt distress improved from high to moderate.5 The upgrade occurred in the second year of the program, reflecting an improvement in The Gambia’s Country Policy and Institutional Assessment (CPIA) score, an increase in the proportion of domestic debt, and the inclusion of re-exports in the external debt-to-exports ratio.6 Notably, the debt rating improvement occurred before most of the deterioration noted above, which happened after the program went off track. At T+3, the rating was still moderate but at T+5 had slipped back to high risk of distress due to additional debt accumulation and a downgrade in The Gambia’s CPIA rating.

77. Exogenous shocks account for roughly half of the debt projection error. The spillover from the Ebola epidemic and drought accounted for revenue loss and additional spending equivalent to about 2.5 percent of GDP over 2014–15. In addition, lost tourism receipts and crop damage led to a decline in real GDP of about ¼ percent in 2014. Fiscal slippages led to increased domestic borrowing, part of which was financed by the central bank, leading to a consequent spike in domestic interest rates.

78. The other half was caused by domestic factors, principally contingent liabilities and an unexplained residual. The government also incurred substantial outlays on assistance to various parastatals, most of which are accounted for as a residual. These comprise the unbudgeted injection of capital equivalent to 1¾ percent of GDP in two distressed banks and transfers to the National Water and Electric Company (NAWEC) to assist the entity in meeting its debt obligations (2½ percent of GDP). There is an unexplained residual of about 10 percent of GDP, which could be related to the realization of additional contingent liabilities not quantified in program documents.

79. Program performance was mixed, and the program went off track after the first review. The ECF aimed to reduce net domestic borrowing via increased domestic revenue mobilization and reduced fuel subsidies, but this became untenable due to the factors noted above. Not surprisingly, implementation of QPCs for the program as a whole was mixed, with nine met and three waived, all of which were key policy anchors of the program—the ceilings on net domestic borrowing and net domestic assets, and the floor on net international reserves. Ten out of twelve SBs were not met or met with delay. Staff’s proposed corrective actions, including revenue base broadening, strict cash budgeting, and containment of extrabudgetary spending, failed to gain traction.

C. Malawi: 2012 ECF

80. The accumulated debt forecast deviation was 32 percent of GDP. The forecast deviations occurred mostly in the first and second years of the program and were driven principally by (Figure 12):

  • Residual. A large residual of 19 percent of GDP resulting in part from spillovers that arose in connection with the embezzlement of public funds known as the Cashgate scandal; an unanticipated non-concessional loan for military equipment; the conversion of domestic debt to external debt as part of an operation to reduce financing costs; and a large unexplained component (about 10 percent of GDP);
  • Exchange rate. Greater than anticipated exchange rate depreciation, accounting for 4 percent of the slippage, due to a loss of market confidence as news of the scandal emerged and donors cut off support;
  • Growth. Lower than expected GDP growth, mainly caused by drought and the deterioration of market confidence (4 percent); and
  • Higher primary deficits (4 percent) caused by spillovers from lower growth.

81. Despite this large slippage, the external debt risk rating remained unchanged at moderate through the program period. As donors cut back assistance in the wake of the scandal, the government resorted to massive domestic borrowing to meet the consequent financing gap, causing an increase in domestic borrowing relative to external borrowing. The use of a higher discount rate for debt service payments starting in 2013, in line with Fund guidance, helped stabilize Malawi’s external debt indictors. By the third year of the program, all external debt indicators remained below their respective thresholds under the baseline scenario though the margin between baseline values and thresholds had shrunk, in part due to the slippages/shocks noted above and the downgrading of Malawi’s CPIA rating to weak at the fifth/sixth reviews.

82. The debt projection error was predominantly the result of domestic factors:

  • The single most important factor was the Cashgate Scandal—a large-scale theft of public funds, which revealed severe weaknesses in PFM. This led most immediately to the loss of donor confidence and withdrawal of donor budget support (equivalent to 4.5 percent of GDP and half of total budget support in 2013).
  • The shock had cascading effects:
    • ➢ The government recapitalized the central bank following losses that arose from the 2012 devaluation of the exchange rate (about 2.3 percent of 2014 GDP) following the revelation of the scandal, and injected capital to cover the bad loans of one public bank that was privatized.
    • ➢ Fallout from the scandal eroded confidence and lowered growth. Despite an improvement in domestic revenue mobilization, the primary balance was worse than expected due to a substantial shortfall in donor grant financing, primarily in reaction to the Cashgate scandal, which was not offset by commensurate spending cuts and contributed to arrears accumulation.
  • A non-concessional loan of $145 million for military equipment was contracted by the previous government, but it was not revealed to the Fund until mid-way through the program. The contract was subsequently renegotiated, and the loan was reduced to $33 million (1 percent of GDP).
  • In a bid to reduce the costs of servicing domestic debt, in December 2014 the authorities restructured some kwacha-denominated domestic debt (amounting to 6.7 percent of GDP) by selling it for $250 million (in foreign currency) to a regional development bank. The restructuring had the effect of increasing external debt.
  • There is also an unexplained residual of about 10 percent, representing the difference between the estimated impact of contingent liabilities related to the Cashgate scandal and non-concessional loan, and the DSA residual of 19 percent. The difference may represent additional, quasi-fiscal support to banks and/or SOEs that were adversely impacted by the spillover from arrears in payments to suppliers.

83. An exogenous shock—a severe drought in 2015—also undermined growth. Maize production fell by 42 percent, contributing to a 3½ percent underperformance of GDP growth in 2015 compared to initial program projections.

84. Policy slippage and shocks were generally accommodated. The primary objective of the ECF was to maintain macroeconomic stability, promote inclusive growth, and reduce poverty. However, emphasis shifted to macroeconomic stabilization and improving fiscal governance to regain domestic and external confidence, following the Cashgate scandal and other shocks. Some elements of program conditionality were adapted to address shocks and policy slippages. For example, while the program was initially based on zero net domestic financing, this became untenable in the face of continued shortfalls in external financing and uncertainties regarding the resumption of donor support, requiring an increase in domestic financing and the net domestic assets (NDA) ceiling. Primary fiscal balance targets were also relaxed to accommodate the shortfall in external financing. A waiver for a breach of the ceiling on non-concessional borrowing was granted for the military equipment loan. In 2016, the Executive Board approved an augmentation of access (25 percent of quota) to address the BoP needs caused by the drought.

85. Despite accommodations, overall program performance was mixed. For nearly half of QPCs, the Board granted waivers for non-observances or modified them, including for repeat misses on net domestic borrowing, NDA and non-concessional borrowing targets. Sixty percent of SBs were not met or implemented with delay. Fiscal performance fell notably short of program targets, despite an increase in domestic revenue mobilization through improved tax administration. Implementation of the structural reform agenda was repeatedly delayed, especially in PFM, leading to multiple program extensions. While the ECF expired in 2017 with conclusions of all reviews, several reviews were combined in response to shocks and slippages.

D. Niger: 2012 ECF

86. The actual change in gross public debt to GDP was 8 percent of GDP higher than expected at arrangement approval. This was driven by (Figure 12):

  • Primary deficit. Primary balance slippages, mostly resulting from trade and security-related shocks, were responsible for 8 percent of the deviation.
  • Exchange rate depreciation. The depreciation of the CFA franc accounted for 5 percent of the error.
  • Growth. Weaker growth contributed 3 percent of the deviation.
  • Interest rates and residual. The previous factors were partially offset by lower than anticipated real interest rates (-5 percent) and a small residual (-2 percent).

87. The debt risk rating remained unchanged at moderate. By the third year of the program, baseline values relative to thresholds had narrowed, in large measure because of an increase in the debt stock due to fiscal slippages, as well as external financing for natural resource projects, including the Soraz oil refinery.7

88. A variety of exogenous shocks account for most of the forecast deviation. The deterioration in the primary balance was mostly the result of a confluence of external shocks. Drought, a decline in commodity prices, a deterioration in the security environment and resultant refugee crisis, and the non-completion of the envisaged oil export pipeline (due to the oil price and security shocks) led to a decline in revenues and increased spending pressures. Between 2012 and 2015, for example, tax revenues to GDP fell 8 ½ percent below initial forecasts. The external shocks that impacted fiscal performance also affected growth. Actual growth vs. that projected at arrangement approval was weaker by a cumulative 6 percent of GDP. A depreciated CFA franc, reflecting the strong appreciation of the U.S. dollar against the euro, also contributed.

89. Shocks were largely accommodated via revisions to program conditionality. Under the program, rising receipts from oil production and domestic revenue mobilization were expected to yield a 3½ percent of GDP increase in total revenue, supporting an immediate reduction of the basic fiscal deficit from 4 percent to 1½ percent in 2012–14 and an expansion of development spending. However, the program went off track after the first review as drought and a deterioration in the security situation contributed to fiscal slippage through revenue shortfalls and increased security spending. Staff accommodated part of this slippage by loosening the basic fiscal balance target by 2½ percent of GDP in 2014 and by extending the arrangement and rephasing remaining purchases. The fiscal balance was relaxed again at the (combined) fourth and fifth, and sixth and seventh reviews to accommodate further underperformance of domestic revenue, a shortfall in external financing, and heightened security spending. The arrangement was extended and access was augmented at the sixth and seventh reviews to accommodate larger BoP needs. The ECF expired at the end of 2016 with the completion of eight reviews.

90. Implementation of program performance criteria was mixed, but performance on structural conditionality was reasonably strong. By the end of the program, the Board granted waivers for non-observance of nearly half of the QPCs or modified them, mostly as a result of missed targets for domestic financing and domestic arrears caused by the aforementioned shocks. The completion rate for SBs was higher, however, with three-quarters of SBs met. The basic fiscal deficit was 3.1 percent of GDP larger than anticipated at program inception with the extra financing filled via increased domestic financing and, secondarily, increased external borrowing.

E. Rwanda: 2013 PSI

91. Public debt overshot initial program projections by a total of around 20 percent of GDP by end-2016. Gross public debt was initially projected to fall slightly from around 29 percent of GDP to 27 percent of GDP through end-2016, but instead rose to around 45 percent of GDP at end-2016. Three-quarters of the deviation from projection was due to much higher residuals (16 ppts), over a third of which reflected the authorities’ investment push (Figure 12).8 The significant depreciation of the Rwandan Franc accounted for the bulk of the remaining deviation (6 ppts). Primary deficit slippages played a limited role (1 ppts), while the deviation in other debt-creating flows had a small offsetting effect (-2 ppts).

92. Despite the large increase in public debt, the risk of debt distress remained low. The program approval DSA deemed risks to be low, with debt remaining below policy thresholds in the baseline and in shock scenarios. Notably, the baseline did not include a number of large investment projects under consideration, because of the early stage of deliberations. Instead, alternative scenarios assessed the impact of additional non-concessional borrowing, up to US$1 billion. As no policy thresholds were breached, staff concluded that some fiscal space existed. The subsequent higher present value of debt-to-GDP did not imply a change of the low risk rating in Rwanda’s subsequent DSAs. Given rollover expectations, the scheduled repayment of the 2013 Eurobond breached the debt-service-to-revenue threshold only temporarily in 2023. Other indicators remained below thresholds. At end-2017, the present value of debt-to-exports was around 145 percent, in line with the original “US$1 billion additional borrowing” scenario.

93. Exogenous factors explain around half the deviation of public debt from projection:

  • The large exchange rate depreciation—15 percent between mid-2015 and end-2016—reflected the impact of the massive commodity price shock and the use of the exchange rate as the principal shock absorber and adjustment tool in line with Fund advice. Moreover, the shock also led to the authorities’ request for, and successful completion of, an SCF in 2016, which further increased debt relative to the original PSI program projections by around 2.5 percent of GDP.
  • The small deviation caused by the primary deficit slippage was also driven by external factors. Besides a weaker growth environment, implementation of East African Community (EAC) internal tariffs and harmonization of external EAC tariffs at lower levels hampered domestic revenue mobilization efforts, a key element of the program. Also, grants declined faster than anticipated. For 2016/17, total revenue and grants were over 2 percent of GDP lower than initially expected, while expenditures remained broadly in line with program projections.

94. The rest of the deviation from projections reflects the design and evolution of the program, particularly the incorporation of the public investment push. A key objective of the PSI was to maintain a sustainable fiscal position and create space for investment. The initial baseline debt profile was broadly flat, since it excluded some anticipated potential investment projects, and tight Quantitative Assessment Criteria (QACs) were set for domestic financing, alongside a zero ceiling for non-concessional borrowing. However, the PSI flagged the large investment needs, limited access to concessional resources, a pipeline of investment projects, and the existence of fiscal space. Conditionality subsequently evolved to accommodate large-scale investment projects, which staff assessed to be productive and worthwhile investments mostly via government guarantees provided to SOEs. These fell outside the general government primary deficit but within the program’s NCB limit, which was raised to $250 million at the first review and $500 million at the third review.9 Key projects included: RwandAir expansion (2.5 percent of GDP); completion of Kigali Convention Center (2 percent of GDP), bridge financing for a new international airport (1 percent of GDP) and guarantees for the insurance and hotel sectors (1 percent of GDP).

95. The PSI remained on-track and implementation remained strong. By end-2018, ten reviews were completed, with conditionality largely met (i.e., with only minor IT breaches), and the SCF was a success, notably overperforming on fiscal adjustment10

F. São Tomé and Príncipe: 2012 ECF

96. Public debt significantly overshot initial projections, with the cumulative deviation of around 35 percent of GDP by end-2015. At arrangement approval, public debt was projected to fall from 54 percent of GDP at end-2012 to 34 percent of GDP at end-2015. In fact, debt rose substantially from a revised end-2012 base of 49 percent of GDP to 62 percent of GDP by end-2015. Residuals accounted for around 30 ppts of deviation, as the anticipated drawdown of government assets was replaced by an increase in debt-creating flows (Figure 12). Automatic debt dynamics accounted for another 16 ppts of deviation, mostly through lower-than-expected growth and some real exchange rate depreciation. This was offset by overperformance of 11 ppts in the primary deficit.

97. The risk of debt distress remained high throughout the duration of the ECF. At arrangement approval, debt was assessed to be at high risk of distress, given the narrow export base. Debt dynamics were considered to be manageable under the “oil production” baseline, but under an alternative “non-oil” scenario, debt ratios were projected to remain higher for longer, particularly the present value of debt-to-exports. This “non-oil” scenario was adopted as the baseline at the second review. The DSA maintained a high risk of debt distress rating through the two completed reviews until the program was cancelled in 2014. Although a better recording of services (travel and tourism) and strong cocoa exports had significantly improved the debt-to-export ratio by 2015, by most metrics the position was worse than originally anticipated, even under the “non-oil” alternative scenario.

98. Almost all the deviation from projection is explained by the failure of the “oil production” scenario to materialize in 2015. The absence of oil production created the very large residuals, as the primary deficit was financed by debt, rather than by the anticipated large drawdown from the Nation Oil Account. Without oil production, growth came in at around 4 percent in 2015, significantly underperforming projected growth of 38 percent at program approval.11

99. The 2012 ECF balanced the positive prospects for oil production with the high risk of a debt distress rating. The ECF strategy focused on relying on non-debt financing, while mobilizing domestic revenues and containing non-priority spending. Oil revenues were expected to increase fiscal space for infrastructure and pro-poor spending from 2015 onwards. In this context, the consolidation plan initially targeted a small adjustment of around ¼ percent of GDP for 2013 and 2014, and an additional 1 percent of GDP over the medium term. However, the program anticipated the risk of either delayed or disappointing oil production, and, in this scenario, the authorities had committed to additional fiscal measures of 1 percent of GDP by 2015 to achieve manageable debt dynamics. Program conditionality reflected additional adjustment and some accommodation of shocks for the first and second reviews, but the zero ceiling on the contracting or guaranteeing of new non-concessional external debt by the central government or the central bank remained in place.

100. While performance was initially satisfactory, the program went off track due to the contracting of non-concessional borrowing and fiscal slippages. Conditionality was largely met through the first and second reviews, with notable fiscal overperformance in 2013 due to higher non-tax revenues from license fees and much lower-than-anticipated primary expenditures (despite higher personnel costs). However, the program subsequently went off track in 2014, when the authorities contracted a $40 million (11 percent of GDP) loan from Angola with a grant element well below the 50 percent floor specified under the program. The loan was renegotiated to near concessional terms (46 percent) but expenditure slippages in the run-up to elections (1½ ppts of GDP on public sector wages) delayed program resumption, by which time key assumptions had changed with a lower probability of oil production. Most of the Angolan loan was spent on investment projects, which staff ultimately assessed to be worthwhile, but some was spent on current outlays before elections.

101. STP’s successor 2015 ECF maintained a focus on grants but relaxed the concessionality threshold for borrowing. With oil production unlikely for many years, the new ECF recognized the need to explore alternative financing options for investment. The DSA made the case for concessional borrowing of 6.6 percent of GDP annually over the program period and emphasized that lowering the concessionality threshold to 35 percent was possible without jeopardizing sustainability, given improvements in the DSA and clearly identified growth-enhancing investments.

Structural Conditionality and Program Design

A. Lessons

  • Ownership and implementation capacity are critical. Risks are particularly high in the absence of a clear track record of program implementation performance and weak implementation capacity.
  • Conditions most critical to program success may be in shared or non-core areas of Fund responsibility. Besides building in-house expertise in critical shared areas, close collaboration with other institutions and technical assistance can help avoid a bias toward core areas of responsibility.
  • Implementation of structural reforms often requires significant time. Recognizing the complexity of structural reforms, careful sequencing of reform steps and realistic timetables could improve prospects for successful implementation.
Table 4.Case Studies: Structural Conditionality and Program Design
Program Completion 1/Did structural conditionality address key gaps identified in surveillance?Were aspects of the political economy considered in program design?Were conditions sufficiently critical and parsimonious?Did collaboration with other institutions help identify reform needs and address capacity constraints?Was technical assistance aligned with the structural reform agenda?Comments
Ireland – 2010 EFFCPartly *N/A* The scale of banking sector problems were not fully anticipated.
Latvia – SBA 2008C✓ *Partly* Structural reform areas were expanded over the course of the program.
Tunisia – SBA 2013LPartlyPartly
Source: IMF staff.

Program completion status as of end-September 2018: C = completed all reviews, IP = in progress, L = largely implemented, OT= off track, QOT= quickly off track, R = replaced.

Source: IMF staff.

Program completion status as of end-September 2018: C = completed all reviews, IP = in progress, L = largely implemented, OT= off track, QOT= quickly off track, R = replaced.

B. Ireland: 2010 EFF

102. The three-year EFF was approved as the global financial crisis exposed deep vulnerabilities in the Irish banking system, triggering a crisis. The Irish authorities initially took a number of actions to restore financial stability, including a blanket guarantee scheme, interventions in major banks, and establishment of a state-owned restructuring agency. However, as the initial crisis response proved insufficient, in December 2010, the Fund approved an EFF involving exceptional access for SDR 19.466 billion (2,322 percent of quota) in support of Ireland’s crisis response and as part of a broader European-supported financing package, in close coordination with the European Commission and the European Central Bank.

103. The program focused on the following objectives: (i) restoring the banking system to health, including by establishing a smaller banking sector with high capital buffers and more stable funding sources; and (ii) securing fiscal sustainability while limiting the near-term demand drag from fiscal consolidation. At the outset of the program, it was decided not to address some policy challenges upfront due to both political constraints and the weak economy. These constraints impeded, for example, a swift and durable workout of mortgage and Small- and Medium-Sized Enterprise (SME) arrears.

104. Structural conditions focused on addressing the Irish banking crisis. The banking sector was impacted by the burst of a property bubble following the Global Financial Crisis, but the scale of problems was underestimated initially. The need for fiscal consolidation—unavoidable given the economic correction—was identified in surveillance, and the programmed fiscal adjustment was implemented. Consistent with surveillance insights, no other major structural weakness was present.

105. Overall, the Irish program benefitted from a national reform plan taking into account political economy considerations. The authorities’ national reform plan served effectively as the basis for the Fund-supported program, underpinned by strong ownership and effective communication to the general public and market participants. Historical and political constraints made some options for NPL workout not viable in the Irish context, including repossessions of a considerable number of owner-occupied residences. Hence, NPL workout experienced a delayed start as insolvency reforms took time. Nonetheless, the incipient recovery and a more prescriptive stance of the Central Bank of Ireland, including by introducing quantitative SME and mortgage arrears restructuring targets and further strengthening of loan provisioning and classification standards, helped advance loan restructuring towards the end of the program.

106. The expertise offered by the IMF, European Commission, and European Central Bank was mutually complementary. The provision of additional, very large financial resources by the ECB and the EU was essential to successful gradual adjustment.

107. Conditionality was parsimonious, focusing on the financial sector and selected fiscal issues which were identified as critical. To avoid over-burdening implementation capacity and recognizing the structural strengths and competitiveness of the Irish economy, Fund conditionality did not contain other structural reforms.

108. Program implementation was strong, and the program was completed. The program succeeded in stabilizing the banking sector and reducing its size. All performance criteria under Ireland’s EFF were met. Almost all structural conditions were met, albeit a few with some delay or partial implementation. Fiscal deficit and debt outcomes were broadly as anticipated. While the first review was delayed and combined with the second review amid early elections, all other reviews were concluded as originally scheduled. While in hindsight Irish regulators could have acted earlier (e.g., by using their prudential powers to improve classification of NPLs and restructured loans; strengthen provisioning and write off practices; and setting targets), the accommodation of political constraints contributed to high ownership throughout the program. Allowing time for NPL resolution also enabled development of new approaches to mortgage NPL workout, for instance by establishing an Insolvency Service and developing durable loan modifications.

C. Latvia: 2008 SBA

109. Years of unsustainably high growth and large current account deficits caused a financial and BoP crisis, which led to a request for a 27-month SBA. Between August 2007 and December 2008, private sector deposits fell by 10 percent, led by a run on the Parex Bank, the second largest bank. Official reserves dropped by almost 20 percent, as the central bank sold foreign currency to defend the peg to the euro. Against this background, the program aimed to address an immediate liquidity crisis and ensure long-term external stability while maintaining Latvia’s exchange rate peg.

110. Structural reform areas were expanded during the program period. The program focused on fiscal and financial sector issues, including private-sector debt restructuring. While conditionality on labor market reforms were not initially included, three SBs were introduced over the course of subsequent reviews. These included promoting wage restraint by establishing a committee to help ensure fiscal sustainability and active labor market policies.

111. Structural conditionality was broadly in line with surveillance gaps. The 2006 Article IV consultation report identified fiscal, monetary and financial sector challenges. Furthermore, surveillance gaps in several shared areas were identified, including low labor force participation and low value-added exports. The 2008 SBA drew appropriate insights from surveillance, although SCs on labor market reforms were introduced only later in the program.

112. The authorities were strongly dedicated to the program in order to safeguard Latvia’s future membership in the euro area. Hence, ownership was exceptionally high.

113. Latvia’s 2008 SBA was part of a coordinated effort involving several institutions. The Fund cooperated closely with staff of the European Commission, ECB, World Bank, EBRD, Swedish Ministry of Finance and Riksbank, and Nordic country governments. The European Commission participated fully in the Fund mission, along with representatives from the ECB (given Latvia’s membership in the European Exchange Rate Mechanism II), as well as Sweden and other Nordic countries given their banks’ exposures.

114. The number of SCs was larger than in comparator programs. The 27 month SBA included 48 SCs. In part, this above-average number of conditions resulted from modifying benchmarks and specifying smaller, intermediate steps toward an end goal. While the conditions were critical to achieve program objectives, strong ownership would suggest that conditionality could have been streamlined.

115. The Fund and the World Bank provided TA in essential areas of the program. The program request document noted that there would be a need for substantial TA to relieve institutional constraints to program implementation. The authorities requested TA from the Fund to strengthen PFM and develop a comprehensive private debt restructuring strategy. Additionally, they received World Bank TA on the comprehensive reforms of the education, civil service, state administration, and healthcare systems, among others.

116. Implementation was strong, and the program was completed. The government undertook significant fiscal consolidation and labor market reforms to facilitate an internal devaluation. This was complemented via actions to promote financial stability and corporate sector restructuring. Nonetheless, unemployment remained high at the end of the program, underlining the impact of the protracted adjustment.

D. Tunisia: 2013 SBA and 2016 EFF

117. Amid deep structural challenges, the two-year SBA and the four-year EFF aimed at short-term macroeconomic stabilization as well as structural reforms for stronger, inclusive growth. Structural challenges to be addressed spanned almost all sectors. Hence, the programs included many structural conditions (2013 SBA: 51; 2016 EFF: 29), of which about one third covered gaps identified in previous Article IV consultation reports. Most SCs were related to the core areas of Fund responsibility, even though half of the gaps identified in surveillance fell in shared or non-core areas. While assessing the social impact of reforms, the programs targeted strengthening social assistance.

118. Despite having broad objectives of achieving stronger and inclusive growth, the programs only partly addressed structural gaps identified in surveillance. Overall, close to half of the gaps identified in Article IV consultation reports were not covered by SCs, in particular, the macro-structural area. More specifically:

  • 2013 SBA. Frontloading high-priority structural reforms, the program focused on reforms achievable by a transition government preparing for a new constitution. SCs covered PFM, revenue administration, central bank policies, the financial sector, and energy subsidies. Labor and product markets, pension and social issues, and capital market development were not included, notwithstanding significant gaps identified in surveillance.
  • 2016 EFF. Assuming an increased capacity of the Parliament to pass transformative legislation, the EFF included more ambitious reforms. A majority of SCs focused on monetary and banking issues, energy subsidies, SOE reforms, corruption, and the pension and civil service systems. Other issues highlighted in the preceding Article IV consultation report, such as labor and product market gaps or data transparency, were subject to review discussions but were not addressed through SCs.

119. Political economy risks were significant, particularly for the 2013 SBA. Given the political transition, staff reports highlighted the political situation as the main risk to the program, in particular given the limited implementation track record. Yet, politics turned out to be more complex than anticipated: between 2010 and 2016, seven prime ministers took office, elections were delayed, and numerous serious social and security tensions emerged.

120. Staff collaborated closely with other institutions. During the 2013 SBA and the 2016 EFF, the World Bank and other institutions provided extensive TA. To support the authorities’ reform agenda through donor activity, staff coordinated closely with the World Bank, the European Commission, the EBRD, the ADB, and Tunisia’s key bilateral partners.

121. Reducing the number of SCs did not prove sufficient to overcome weak implementation. Despite drastically reducing the number of SCs from the 2013 SBA to the 2016 EFF and stronger tailoring, program implementation remained weak.

122. To mitigate capacity constraints, extensive TA was provided in both programs. During the 2013 SBA, Fund TA focused on the financial sector, PFM and RA, as well as data and energy subsidies, in collaboration with the World Bank. World Bank TA covered several governance areas, including the public investment code. During the 2016 EFF, other institutions, including the ILO, the Banque de France, and the KfW also provided TA.

123. The objectives of the 2013 SBA were only partially met, and the 2016 EFF is still ongoing. While the 2013 SBA drew on an existing national program, political fragility and vested interests hindered progress in several reform areas.

Ownership

A. Lessons

  • Factors important for ownership include the integration of national reform and poverty reduction plans and a clear communication strategy. The latter includes outreach and engagement with Civil Society Organizations (CSOs), which can support broad consensus on the needed policies.
  • Factors that can significantly weigh on program performance, if not sufficiently accounted for, include capacity constraints, political transition and risks, and vested interests. Careful analysis of institutional and political constraints and setting out a realistic timetable can help ensure program objectives are achieved.
Table 5.Case Studies: Ownership
Program Completion 1/Was a national reform plan well integrated with program design?Did the authorities have a clear communication strategy?Did capacity constraints significantly hinder reform implementation?To what extent did the political cycle impact program design and implementation?To what extent did vested interests play a role in reform implementation?Comments
Jamaica – EFF 2013CLowLow
Kenya – SCF-SBA 2015/2016C/QOTHighHigh
Romania – SBAs 2009/2011/2013C/C/OT*PartlyHighHigh* Communications aspects were included in European Semester documents.
Rwanda – PSIs 2010/2013, SCF 2016C/C/CLowLow
Seychelles -EFFs 2009/2014C/CPartlyModerateLow
Source: IMF staff.

Program completion status as of end-September 2018: C = completed all reviews, IP = in progress, L = largely implemented, OT= off track, QOT= quickly off track, R = replaced.

Source: IMF staff.

Program completion status as of end-September 2018: C = completed all reviews, IP = in progress, L = largely implemented, OT= off track, QOT= quickly off track, R = replaced.

B. Jamaica: 2013 EFF

124. The 2013 four-year EFF was preceded by years of stagnant growth and large fiscal deficits. A succession of Fund-supported programs, including a 27-month SBA in 2010, had failed to gain traction. By 2013, low growth and weak public-sector finances had crowded out private credit and investment, leading to high and unsustainable debt ratios. The 2013 EFF therefore aimed to avert immediate crisis risks and create the conditions for sustained growth through significant fiscal adjustment and improvements in competitiveness. The program was supported by a domestic monitoring mechanism, the Economic Programme Oversight Committee (EPOC), which was established in the context of the 2013 EFF and comprised members from the private and public sectors, and civil society.

125. The relative success of the program was aided by a national economic program and outreach to civil society. The program was closely linked to the authorities’ own economic program. Though the authorities did not have an explicit communications strategy, significant outreach to CSOs was carried out. This included communicating the important objective of alleviating social costs. Further, EPOC was effective, supported by regular press releases and press conferences as well as broad participation.

126. The 2013 EFF was replaced in 2016 by an SBA treated to be precautionary. Under the EFF, all except the last two reviews were completed, and significant strides were made in restoring macroeconomic stability. Thanks to strong program ownership, the authorities adopted a new fiscal rule and substantially reduced public debt via fiscal consolidation and a debt exchange. They also made improvements in tax policy, financial sector resilience, and the investment environment. A large majority of QPCs and SBs were implemented on time. The election of a new government in 2016 did not negatively affect program performance, suggesting strong and broad-based ownership. The precautionary successor SBA agreed in 2016 was intended to provide insurance against unforeseen adverse external economic shocks, while focusing reform efforts to deliver better growth and job outcomes, as well as reduce poverty.

C. Kenya: 2015 and 2016 Blends: SCF and SBA

127. Following successful completion of its 2011–13 ECF, Kenya was supported by two successive blended SCF-SBAs. The first was approved in 2015 (for 12 months) and the second in 2016 (initially approved for two years and extended by six months). The authorities treated the SCF-SBAs as a precautionary buffer against potential exogenous shocks.

128. The 2015 program was not explicitly linked to a national reform plan. While the successful 2011 ECF built on a national development plan, the 2015 SCF-SBA did not explicitly link program objectives and measures to the authorities’ own strategy other than through brief references to the authorities’ “Vision 2030.” Also, communication issues were not discussed in the documents requesting the arrangements.

129. Capacity constraints were not a major obstacle to reform implementation. Significant TA support was provided, including TA on cross-border financing supervision from Afritac East. Capacity development support, however, was not emphasized in staff reports.

130. The political cycle was a factor in program design in the run-up to the 2017 election. The 2015 program recognized the risks of political instability but did not explicitly factor them into program design, emphasizing the strong track record under the previous ECF. The 2016 program, in contrast, explicitly recognized the constraints of the upcoming 2017 election.

131. Politics affected the implementation of key program objectives. Although the program supported a scaling up of infrastructure spending consistent with debt sustainability, there were political pressures for even more expansionary macroeconomic policies. The adoption of interest rate controls, against staffs advice, and subsequently, Parliament’s rejection of the bill to remove the rate ceiling also signaled limited ownership.

132. Performance under the 2015 program was satisfactory but deteriorated under the 2016 program. Both reviews under the 2015 program were completed. Quantitative targets were generally met, though SB implementation was mixed. While the first review under the 2016 program was completed, some SBs were not met. In addition, the fiscal deficit widened, reflecting shocks like drought, revenue shortfalls, and pre-election spending pressures. Subsequent to the adoption of interest rate controls, bank lending to the private sector fell sharply, and the monetary policy framework was weakened. Despite an extension of the SBA in 2018, the second review could not be completed before program expiration.

D. Romania: 2009, 2011, and 2013 SBAs

133. Romania has had a total of 10 Fund arrangements since 1991. Romania requested three SBAs involving exceptional access in the wake of the Global Financial Crisis, making purchases under the 2009 arrangement and treating successor arrangements (2011 and 2013) as precautionary. The programs benefitted from deep collaboration with the EU (integrating the National Reform Programme and the Convergence Programme) and the World Bank. While communication issues were not discussed in the program request staff reports, European Semester documents included some communications aspects.

134. Challenges under the programs included opposition from vested interests alongside a shift in program objectives toward more difficult structural reforms. Vested interests opposed key reforms, in particular those related to increased private-sector involvement in the energy and transportation sectors. This inhibited capital investments and delayed efficiency-enhancing reforms. Implementation of corporate governance reforms for SOEs also disappointed. Further, frontloading of difficult structural reforms in the later programs increasingly encountered implementation capacity constraints.

135. Political tensions added to challenges. There were delays in completing the second and third reviews under the 2009 SBA and the seventh and eighth reviews of the 2011 SBA due to political reasons. This had particular importance for SOE reforms, which faced strong political resistance and, amid political turmoil throughout 2012 (with three changes in government), encountered a number of postponements and reversals (IMF, 2014b). Further, on the back of reduced incentives for undertaking difficult structural reforms following the 2014 elections and improved macroeconomic and financing conditions, the Ex Post Evaluation (IMF, 2017a) noted that “an assessment of political economy constraints could figure more prominently in program design.”

136. Performance was originally strong, but it deteriorated as program objectives shifted from stabilization to deeper structural reforms through the series of SBAs.

  • 2009 SBA. The central objectives of the 2009 arrangement were achieved on the back of strong ownership, and flexibility in program design (IMF, 2012).
  • 2011 SBA. Program objectives under the 2011 SBA were also largely met, but progress on the structural reform agenda was uneven. Over time, the program increasingly relied on PAs, largely reflecting delays in implementing the structural reform agenda.
  • 2013 SBA. The 2013 SBA could not be completed. Politically difficult reforms from the 2011 SBA were carried over to the 2013 SBA and frontloaded. In turn, program performance deteriorated as vested interests opposed key reforms. Ultimately, the Ex Post Evaluation (IMF, 2017a) concluded that ownership for the program was insufficient and that implementation of the ambitious reform agenda was undermined by capacity constraints.

E. Rwanda: 2010 and 2013 PSI, and 2016 SCF

137. Rwanda has a long history of Fund engagement. PSIs covering 2010–13 and 2013–18 aimed to support the authorities’ poverty reduction strategy and preserve macroeconomic stability and strengthen inclusive growth. The PSI was accompanied by an SCF arrangement in 2016–2018 in response to a commodity price shock, external risks from a political crisis in neighboring Burundi, and a drought.

138. Ownership and transparency were supported by building on national plans and effective communication.

  • National plans. National reform plans included the Poverty Reduction Strategy (2002–06) and Economic Development and Poverty Reduction Strategies (EDPRS I: 2008–2013; EDPRS II: 2013–2018) which fed into program design.
  • Communication. While the document requesting the arrangements put little emphasis on externally communicating the program objectives, the programs did pay close attention to the need for strategies to communicate program policies. In 2015, staff called for strengthened communications between the Ministry of Finance, Planning and Economic Development (MoFPED), and the National Bank of Rwanda (BoR), and with market players. In turn, MoFPED agreed to overhaul their communication strategy to better influence economic expectations and prevent market confusion or misperceptions. In 2016, the BoR further enhanced its communication strategy by creating an interactive platform for exchanging information with stakeholders. Other communications included educating the public on credit reporting bureau activities, capital market development, and payment systems modernization. IMF staff met regularly with CSOs.
  • Political and capacity constraints. Program performance was aided by broad support for policies. Neither capacity constraints nor the political cycle hindered reform implementation.

139. Program objectives were met, and the programs were completed. All scheduled reviews were completed, with most conditions met among both quantitative assessment/performance criteria, ITs, and SBs, although most SBs were of low or medium depth. Poverty and social indicators, including life expectancy, improved. Rwanda’s per capita income, while still low at US$729 in 2016, nearly doubled during the past decade.

F. Seychelles: 2009 and 2014 EFF

140. Fund engagement supported Seychelles’ recovery from an economic and financial crisis. Supported initially by the 2008 SBA, the authorities pursued macroeconomic stabilization and structural reforms to restore fiscal and monetary policy credibility. A debt restructuring was agreed with Paris Club creditors in April 2009. During 2009–13 and 2014–17, two successive EFFs supported “second-generation” structural reforms to enhance PFM, reduce the role of the state in the economy, and strengthen the financial system. In late 2017, engagement under a PCI was approved focusing on building resilience to potential shocks.

141. The 2014 program aimed at adopting a development plan. Although the 2009 EFF did not feature an integrated national plan, the 2014 EFF included benchmarks on adopting medium-term development (MTNDS) as well as financial sector strategy plans (met with delay) and envisaged civil society participation in the development plan. However, discussions of program communication issues were absent in the documents requesting the arrangement.

142. Capacity constraints were the main factors slowing reform implementation. The very high share of measures that were implemented with delays is consistent with an underlying willingness to reform amid capacity constraints. The 2017 Ex Post Assessment (IMF, 2017b) attributed most delays to technical and administrative hurdles. In this respect, while Fund TA was substantial, program timelines could have taken into account better the timing and payoffs from this TA.

143. The political cycle was taken, to some extent, into consideration during the program. The 2014 EFF request noted presidential elections scheduled for 2016, but discussion of political factors was otherwise limited. Although the authorities’ track record was strong, future public support for reforms was not guaranteed. In the context of policy slippages and corrections, the staff report for the fourth and fifth reviews included more coverage of political factors.

144. Program objectives under both EFFs were met, and the programs were completed. All reviews were completed. Performance against quantitative targets was generally strong, though with some slippage in 2015–16 amid social concerns over economic inequality and election-related spending pressures. Post-election, the authorities took measures to contain the slippages, bringing the program back on track. The structural reform agenda—including the adoption of a medium-term development plan with participation of civil society, as well as financial sector strategy plans— experienced delays but was largely implemented. The Ex Post Assessment (IMF, 2017b) highlighted strong ownership by the authorities as a key success factor for the Fund-supported programs.

References
1

Part of the decline in the debt ratio is explained by a large revision in the national accounts.

2

For these case studies, public debt is assessed on a currency (not residency) basis.

3

In Rwanda, the limit was raised substantially to reflect planned investment. In Malawi, the ceiling was breached twice and significantly for an unanticipated loan for military equipment, which was subsequently renegotiated, resulting in the granting of a waiver. A waiver was also granted for a small breach of the limit in the Niger program. In The Gambia, an exception was planned for a small project before the program went off track. In São Tomé and Príncipe, there was a significant breach of the limit, and although the loan was successfully renegotiated to reduce costs, the program ultimately went off track.

4

The projection error in all case studies is the difference between the actual change in gross public debt from t to t+3 based on the latest DSA and the anticipated change at program approval measured over the same period.

5

In all case studies, the change in the risk of debt distress is that observed between t and t+3.

6

At the end of the program (first review), all external debt indicators were below their respective thresholds under the baseline scenario compared to three breaches in the baseline at program approval. However, under the most extreme shock, the end-of-program DSA continued to show breaches in all debt ratios except the debt service to export ratio.

7

On net, the size of breaches under the various shock scenarios were largely unchanged. In the public DSA, increased reliance on bond financing from the regional market resulted in higher domestic debt stock.

8

Identified factors generating the large deviation in residuals, include 6.5 ppts of investment guarantees (see below), and 2.5 ppts Fund SCF, which were not included in the arrangement approval baseline. Roughly 7 ppts of the deviation in residuals is unexplained.

9

The NCB ceiling was subsequently raised to US$800 million at the ninth review. Other conditionality changes occurred at the fourth and fifth review to reflect the updated Debt Limits Policy. The QACs on net domestic financing and NCB were replaced with ITs, and the IT on public domestic debt was eliminated.

10

The IT on public enterprise debt (formerly non-concessional borrowing ceiling) was later breached by a large margin in 2018 (US$134 million) due to earlier-than-anticipated leasing of aircraft by Rwandair to obtain more favorable terms.

11

At arrangement approval, it was assumed that a large asset drawdown would be used to finance the primary deficit (e.g., particularly investment spending) from 2013–2015, with an accumulation of oil revenues once oil production began. In the DSA at arrangement approval, this was reflected in very large negative change in assets from 2013– 2015 (-20 ppts), and positive change in assets from 2015 onwards. In fact, residuals proved to be positive and relatively large (10ppts) – these remain unexplained.

Other Resources Citing This Publication