A. Introduction

1. Uganda’s program relation with the Fund dates to the 1980s. After several Poverty Reduction and Growth Facility arrangements, Uganda was the first country to request the just created PSI in 2006.² Since then, Uganda has had four successive PSIs (with the fourth one expiring in July 2017). The PSIs have been instrumental in supporting Uganda’s growth agenda. Technical assistance has complemented the objectives of the PSIs, and focused on tax policy and administration, public finance management, monetary policy, financial sector supervision and macroeconomic statistics.

2. Uganda has been assessed to perform well under successive PSIs.³ Growth remained strong during the PSI periods, though since 2010 trend growth has slowed.⁴ Inflation was maintained within reasonable bands—except for the 2011 inflationary spike—and reserves within adequate levels, while the fiscal deficit has been generally low. Despite the good performance, there were some weaknesses related to public financial management, including persistent use of supplementary budgets, under-execution of development budgets, “procurement problems, inability to enforce contracts, and limited implementation capacity”.⁵ The structural agenda also suffered some “technical and political limitations as well as ‘reform fatigue.”⁶

B. The Current PSI

3. The current PSI was approved by the IMF’s Executive Board in June 2013 with an initial duration of three years. Following a one-year extension in 2016, it was set to expire in June 2017, before being extended again through end-July 2017. The overarching objective of the 2013 PSI was to support inclusive growth through macroeconomic stability and structural reforms. Specific priorities included (i) enhancing revenue through measures to broaden and deepen the tax base and improve tax administration; (ii) improving the effectiveness of PFM; (iii) preparing the economy for oil production and management of petroleum revenues; (iv) moving from inflation targeting ‘lite’ to full-fledged inflation targeting; and (v) improving the business environment, supporting the development of the financial sector, and continuing to maintain financial sector stability.⁷ In practice, a critical objective was to ensure that the scaling-up of infrastructure investment was properly implemented, while safeguarding the debt sustainability low risk of distress rating.

4. The external environment influenced Uganda significantly, as a small open economy with a freely-floating exchange rate. In FY2014/15, the Ugandan shilling declined by 27 percent vis-à-vis the US dollar year-on-year (with the depreciation reaching nearly 40 percent in August 2015), as the economy was affected by global liquidity concerns, negative shocks in neighboring countries and trading partners, and election-related nervousness.⁸ As a commodity importer, Ugandan imports benefited from the oil price shock, even if in the medium run it represents a risk which could have delayed investment decisions in the domestic oil sector.⁹ Uganda was however also negatively affected by the commodity price decline, including for coffee, Uganda’s largest export commodity. Security concerns in the region have also impacted sentiment, and more recently, the South Sudan crisis has led to a decline in Ugandan exports and remittances, and an exponential increase in refugee arrivals. Droughts have affected the economy and triggered a spike in food prices in 2013 and more recently in 2016-17. Reduced development partners’ aid budgets as the 2013 PSI started also spurred domestic borrowing requirements.¹⁰

5. Rebasing led to an upwards revision in GDP in November 2014. Improved methodologies and coverage showed a FY2009/10 GDP that was 17 percent larger than previously believed. Thus, some indicators experienced significant change, including the debt-to-GDP ratio, (almost 4 percentage points lower, and the tax-to GDP ratio (about I percentage point lower).¹¹

C. Performance Under the Current PSI

Quantitative conditionality

6. Performance under the quantitative assessment criteria was generally assessed to be sound, with most quantitative assessment criteria met. However, projections consistently overestimated infrastructure investment and underestimated current expenditure (and thus domestic financing), which contributed to an overestimation of growth.

7. Growth performance was consistently overestimated. The Ugandan economy was projected to growth on average by 6.3 percent in the initial three years covered by the 2013 PSI, based on projections in June 2013, June 2014, and June 2015. However, the actual growth was on average 5 percent, while the FY2016/17 growth is projected at about 3.9 percent. Such performance confirms the decline in trend growth recorded since 2010.¹² Over the same period, IMF’s global and Sub-Saharan Africa forecasts was also repeatedly revised down, suggesting that some of the growth optimism reflected expectations over an improved external environment.

View Full Size

Growth and Inflation: Outturn vs. Projections

Citation: IMF Staff Country Reports 2017, 207; 10.5089/9781484309360.002.A001

Sources: Ugandan authorities and IMF staff calculations

• Download Figure
• Download figure as PowerPoint slide

8. Inflation during the period was kept close to the 5 percent core inflation target. BoU transitioned to an inflation targeting regime in July 2011, initially referred to as “lite”. The 2013 PSI introduced an inflation consultation clause (ICC), which during the first year of the PSI co-existed with the previous base money target. ¹³ Performance under the PSI in this area was good, with only one breach of the inner (lower) band of the ICC in December 2014, when the 12-month average core inflation recorded 3.1 percent, well below the target of 5.7 percent. This was driven by subdued food crop prices and low import prices. Such a breach triggered consultations with staff in line with the ICC. During the consultations, BoU explained that the prospects for loosening earlier had been constrained by “(i) the start of a shilling depreciation trend; (ii) an expected increase of food prices; and (iii) unclear fiscal prospects.”¹⁴

9. Inflation targeting has served Uganda well. The system has continued to strengthen, with better forecasting, clear arrangements for monetary policy formulation, and revamped policy instruments. Despite the shallowness of the financial markets, the monetary transmission has worked reasonably well, though faster when rates increase.¹⁵

10. International reserves have been built up and properly managed. The authorities’ needs for foreign exchange during the period have been important as they prepared for the rise in government imports and made large payments in USD related to the repayment of loans for large infrastructure projects. This resulted in BoU having to carefully plan a reserve accumulation policy. BoU generally met PSI reserve targets and kept reserves within the adequacy range. Foreign exchange interventions have been limited to smoothing excessive volatility. Furthermore, the modalities of foreign exchange purchases in the market have been generally sound and evolved throughout the duration of the PSI, initially with pre-announced daily auctions, moving to a pilot whereby BoU agreed to buy a fixed amount from each of seven banks in 2015, moving to the current, more discretionary system. ¹⁶

View Full Size

Reserves

(in months of imports and stock in USD mil)

Citation: IMF Staff Country Reports 2017, 207; 10.5089/9781484309360.002.A001

• Download Figure
• Download figure as PowerPoint slide

11. The key objective of increasing the tax to GDP ratio by at least ½ percent of GDP per year was not consistently met annually, though it was met on cumulative terms over the PSI period. Shortfalls in the tax-to-GDP ratio target for FY2013/2014¹⁷ were compensated by significant improvements in FY2014/15, which contributed to an aggregated increase in the tax-to-GDP ratio of around 2.2 percentage points of GDP between FY2012/13 and FY2015/16. Both tax policy and administrative measures played a role, with initiatives being supported by technical assistance, including in preparing a VAT-gap analysis.

View Full Size

Main Fiscal Indicators: Outturns vs. Projections

Citation: IMF Staff Country Reports 2017, 207; 10.5089/9781484309360.002.A001

Sources: Uganda authorites and IMF staff.

• Download Figure
• Download figure as PowerPoint slide

12. The fiscal strategy was based on the decision to prioritize infrastructure investment to remove key bottlenecks to growth.

• Development spending and net lending¹⁸ increased by 2.1 percentage points, from 6.5 percent of GDP in 2012/13, to 8.6 percent in 2015/16. Despite the increase, such amount was significantly lower than originally projected, particularly in the externally-financed part. The forecast error on the foreign financed and net lending projections was, on average for the three years, about 1.1 percent of GDP.

• The main reasons behind the lower-than projected infrastructure expenditure include capacity issues—in the areas of procurement, contract management, compensation— and in some cases poor design of projects. As the low implementation and execution rate became more apparent, the projections under the PSI started to discount accordingly the infrastructure investment budget releases, resulting in more accurate projections. The delays in the absorption of funds and project implementation were among the key reasons behind the World Bank’s decision to temporarily withhold new lending effective August 22, 2016 (which lasted until May 2017, when it was decided to resume new lending). Given the relevance of this area, the PSI included benchmarks in this area since the third review, and technical assistance was provided to improve procedures and strengthen capacity and controls. Recent work has been undertaken to support the design and implementation of the Action Plan to improve Public Investment Management.

View Full Size

Current Spending and Domestic Issuances: Outturns vs. Projections

Citation: IMF Staff Country Reports 2017, 207; 10.5089/9781484309360.002.A001

Sources: Ugandan authorites and IMF staff.

• Download Figure
• Download figure as PowerPoint slide

• In parallel, current expenditures increased by 1.9 percentage points of GDP. In most cases (particularly towards the end of the period) such expenditures were higher than originally projected, and led to higher-than-projected domestic issuances. The average underestimation of domestic issuances was about 0.3 percent of GDP over the three-year period. Higher domestic issuances could generate crowding out effects on the private sector and raise lending rates.¹⁹

• The overall deficit was lower than projected, though the target was not always met. The under-execution of development spending offset occasional revenue shortfalls and current expenditures overruns, leading to a lower-than-projected deficit most of the PSI period. For purposes of the PSI²⁰, the deficit definition was narrower, as it excluded externally-financed development spending and other one-off flows like oil-related revenue. With this definition, the deficit target ceiling was respected during the first two years of the PSI, but missed for FY2015/16 in the context of softer-than-anticipated nominal growth and higher-than-anticipated election-related spending.

View Full Size

Fiscal Targets: Outturns vs. Projections

Citation: IMF Staff Country Reports 2017, 207; 10.5089/9781484309360.002.A001

Sources: Ugandan authorites and IMF staff.

• Download Figure
• Download figure as PowerPoint slide

13. The PSI addressed government advances from the BoU. In the context of higher than anticipated domestic financing needs and tight liquidity conditions (particularly in the first part of the year), the authorities used a temporary advance from the BoU to finance operations, which was not repaid within the year. The PFM Act requires that advances from BoU be repaid within the fiscal year. Furthermore, the PSI included a new indicator to monitor the repayments of these advances in the 5^th PSI review (November 2015), in the context of a higher-than-usual recourse to advances which were complicating monetary policy implementation within the inflation targeting framework. The PSI’s definition of advance is more precise and stringent than that of the PFM Act. To further address this issue, the PSI also introduced a structural benchmark in the 5^th review on the “introduction of specific rules and controls on the intra-year use and balances of the Uganda Consolidated Fund (UCF) and UCF/TSA and institutionalize these rules in a framework agreement with the BoU and in the charter of fiscal responsibility”, which was, however, not met.

View Full Size

Social Spending and Outcome Indicators

Citation: IMF Staff Country Reports 2017, 207; 10.5089/9781484309360.002.A001

Sources: Uganda authorities, World Development Indicators, and IMF staff calculations.

• Download Figure
• Download figure as PowerPoint slide

14. The PSI also included a social spending floor. The floor was defined as the sum of the domestic expenditures inclusive of salaries and wages for the Health, Education, Water and Environment, and Agriculture sectors. The program targets were exceeded. However, the overall social spending level has remained stable, at around 4 percent of GDP, since FY12/13. This level is significantly lower than the EAC average for related social categories (Health, Education, Pensions and Social Assistance). While it is difficult to measure improvements in social indicators in such a short period, some progress was achieved in terms of poverty reduction during the period of the PSI, and some social indicators also improved. However, other social indicators, such as malnourishment, deteriorated.

15. The PSI also put emphasis on the need to strengthen social protection. Though not part of the formal conditionality, staff and authorities engaged in frequent discussions about the merits of social protection in a country like Uganda, its affordability and likely positive impact reducing poverty, improving social indicators and enhancing growth and productivity. In the course of the PSI, a Social Protection Policy was approved in November 2015, and Government committed to deepening the existing social assistance programs as well as expanding new ones gradually.

16. Public debt increased over the period from 26.2 percent of GDP in FY12/13 to 34½ percent in FY15/16. Projections were biased towards higher debt in the first years of the PSI—in line with the originally expected higher development expenditure and higher deficits, but also as a result of a purely statistical effect from GDP-rebasing prior to the fourth review.²¹ On average in the three-year period, public debt was about 4.2 percentage points lower than the average PSI projections. While in principle the lower-than-projected infrastructure investment should have resulted in an even lower debt-to-GDP ratio, considering the rebasing, some factors played in the opposite direction. This include lower-than-projected growth a more depreciated exchange rate, and higher-than-projected domestic borrowing.²² Furthermore, delays in the execution rates of projects resulted in a divergence between the contracted and effective loans (“debt” was considered contracted or guaranteed following the approval by a resolution of Parliament under the PSI’s definition during most of the PSI period) and the executed capital expenditure recorded in the fiscal accounts.²³

17. Uganda has remained at low risk of distress through the duration of the PSI. Debt is expected to remain sustainable over the medium term, provided the key assumptions behind the macroeconomic framework –mostly in terms of revenue increases, fiscal consolidation as projects are completed, and projects yielding the expected growth dividend. However, vulnerabilities have recently increased.²⁴

18. The PSI included a criterion establishing that all oil-related resources had to be placed under the Oil Fund, which was consistently met. The Oil Fund was ring-fenced for the financing of infrastructure projects. Furthermore, the PSI followed closely all the use and movements of the oil fund, with the MEFP explaining in detail the stock of the Oil Fund and its currency composition over time. However, when the Oil Fund was closed to open a Petroleum Fund in line with the new PFM Act, the resources from the Oil Fund were not moved to the Petroleum Fund as originally agreed, but were pooled into the Consolidated Fund. Thus, the ringfencing of the Oil Fund resources for infrastructure was lost. Preparations for oil revenue management and taxation also started and benefited from technical assistance to support the design and improvement in the fiscal regime for extractive industries; the Income Tax and the Valued Added Tax were updated with IMF TA.

Structural reforms

19. The emphasis of the structural reform PSI under the program fell on the fiscal side. 74 percent of all structural benchmarks were of fiscal nature, and more specifically related to public financial management (about half of the benchmarks). Since the 2013 PSI started just after the end-2012 scandal of the theft of donor funds in the Office of the Prime Minister, many reform areas were focused on strengthening accountability and closing the loopholes that contributed to the problem.

View Full Size

Structural Benchmarks

Citation: IMF Staff Country Reports 2017, 207; 10.5089/9781484309360.002.A001

• Download Figure
• Download figure as PowerPoint slide

20. Structural benchmarks were partly met. The IMF’s database for Monitoring of Fund Arrangements (MONA) —which tracks the performance of countries in terms of reviews, quantitative and structural conditionally—shows that less than a third of the structural benchmarks were met, even when including the ones implemented with a delay. Nearly 1/5 of the SB had to be modified. However, performance is better when looking at the ultimate outcome: 56 percent of structural reforms were ultimately achieved. ²⁵ Potential reasons behind the declining performance in the structural side could include insufficient buy-in resulting from insufficient elaboration between the benchmark and its ultimate objective, or insufficient attention and commitment to the implementation of the agreed agenda, together with technical lapses or failures. The number of structural benchmarks remained relatively constant at around 12 structural benchmarks per PSI review on average, and do not seem to have become more complicated or ambitious.

View Full Size

Indicators on Structural Reform Performance^*

Citation: IMF Staff Country Reports 2017, 207; 10.5089/9781484309360.002.A001

Sources: IMF staff.^*All PSIs refer to all PSI programs since the creation of the instrument. Also, one structural benchmark can be counted multiple times if it is included in multiple reviews.

• Download Figure
• Download figure as PowerPoint slide

21. Performance of some of the key structural benchmarks are the following:

• Publication of the stock or arrears and unpaid bills. In the initial version of the PSI, the requirement was to submit to cabinet regular quarterly reports on unpaid bills on nine ministries based on data in the Commitment Control System for the previous quarter. Such benchmark was not met. In the second PSI review, it was agreed to expand the requirement to all ministries and require publication in the context of efforts to streamline “institutional arrangements to deal with reporting mechanisms” (second review, para. 23 MEFP). Capacity issues were mentioned as key reasons behind the inability to deliver. To try to simply reporting requirements, the reporting frequency under the benchmark was reduced from quarterly to semi-annually. Furthermore, a technical MoU was prepared clarifying the definitions, which had been blamed as a factor of confusion (second review). A new indicative target was included in the context of the first review of the PSI on the reduction of the stock of arrears; this scaling up of conditionality was useful to place prominence to the problem. In 2016, a reconciliation and verification exercise was undertaken that provided a clear basis to understand the magnitude and sources of the problem. The reconciliation was followed by TA support, and a strategy is expected to be prepared in line with its recommendations to prevent future accumulation of arrears, and deal with the existing stock.

• The presentation of the BoU Act amendments to Parliament was delayed. The amendments are aimed at further strengthening the independence of the BoU and facilitate operations. They include provisions for automatic recapitalization to ensure a level of BoU’s capital as a share of its monetary liabilities. Extensive consultations and discussions explain part of the delays.

• Benchmarks also addressed the need to recapitalize BoU with marketable securities, while the BoU Act amendmenst came into effect. These benchmarks were met, and addressed the most pressing need, which could have removed the sense of urgency about the need for reform for the BoU Act. However, the temporary recapitalizations were meant to go hand in hand with a streamlining of the operational and administrative costs of the BoU. Progress is still ongoing in this front.

• Consistent with the assessment that the financial sector was generaly sound, the PSI’s structural benchmarks in the financial sector field focused on assuring a prompt exit of the FATF grey list. Though progress took longer than anticipated, the required legislation has been just passed.

• Although not a structural benchmark in itself, the PSI called for improved fiscal-monetary coordination, with better communication to ensure better policy aligment. The central bank would benefit from better certainty about the extent and timing of issuances of securities for fiscal purposes; and of limits in the overdrafts in the main government account or use of government deposits. Progress is still ongoing in this area.

• PFM experienced significant progresss. Key areas include the adoption of the new PFM Act, the enactment of a Charter of Fiscal Responsibility; the upgrading of payment and payroll systems (with the electronic integrated financial management system (IFMS) and the integrated personnel and payroll system (IPPS), which helped identify ghost workers and pensioners); the introduction of a Treasury Single Account (preceeded by the closing of many dormant accounts, which were instrumental in the 2012 financial impropriety event); the removal of some tax exemptions; the publication of a VAT-gap analysis. More recently, benchmarks on enhancing public investment management were introduced, and generally complied with, though sometimes with delays.

D. Conclusions

22. Successive PSIs have contributed to growth by maintaining macroeconomic stability and implementing reforms in the fiscal, monetary, and financial sectors. Furthermore, focusing on the current PSI, most of the quantitative targets have been achieved, and some structural ones. However, the under-execution in infrastructure development and over-execution of current spending contributed to lower growth than originally projected, expenditure composition shifts, and generally lower deficits. While the public debt-to-GDP ratio declined, the reduction was partly due to GDP-rebasing.

23. Key achievements under the PSI include (i) domestic revenue mobilization improved significantly, and progress is ongoing to continue moving in this direction; (ii) PFM experienced significant progress, with the adoption of the PFM Act and the introduction of the TSA, that is allowing for more efficient and transparent cash management; (iii) debt remained at low risk of distress, even if vulnerabilities increased; (iv) the transition to inflation targeting was conducted smoothly, and the capabilities and procedures at the central bank continue to be strengthened and become more sophisticated, with inflation having been kept in check.

24. However, some areas continue to suffer from delays and require further strengthening. These including PIM (particularly to address weak implementation capacity); budgeting processes, arrears and fiscal rules; the business environment; deepening and expanding access to the financial sector, which remains shallow²⁶; and further improving fiscal-monetary policy coordination.

25. In terms of processes, the experience shows that it is essential to get sufficient commitment to implement the agreed structural agenda. Explaining more clearly what is the economic rationale behind each benchmark and their relevance could help. The support with TA when necessary (ensuring regular follow up of the implementation of the TA reports), could also contribute to enhancing implementation, and a more serious analysis and demand for accountability in case of delays. Ensuring buy-in from the key players at the highest level can also make a significant difference; when reforms have been supported by a champion, progress has been faster and more constant. A broader commitment to the PSI beyond MoFPED and BoU could also help ensure smoother implementation, so bringing on board more actively other key stakeholders could be useful, including Parliament, or the Prime Minister.

26. The fact that debt has increased without fully reflecting the expected increase in capital spending due to partly low execution rates confirms the importance of ensuring further progress in public investment management. While work is in progress, it is important to consider these limitations to ensure forecast errors are minimized and projections better reflective of realities.