Lessons from Country Experience

Based on the case studies, fiscal multiplier discussions between staff and officials were typically quite limited, with few examples in which staff offered more than ad hoc adjustments to standard multipliers. Staff reports supporting program requests rarely provided explicit analysis of multiplier assumptions.³¹ In some cases, staff sought to modify standard assumptions, for example lowering the assumed multiplier in Grenada (2014) as a small very open economy, in Honduras (2014) to reflect beneficial effects through improved confidence, in Latvia (2008) because the tightening followed a boom, and in Mongolia (2009) because large cutbacks in public investment would mainly affect imports. In contrast, it appears that a mis-specified multiplier assumption contributed to a growth shortfall in Jordan and Ukraine. In the case of Jordan, while multiplier assumptions were not discussed explicitly in program documents, the case study cautiously pointed to a possibility that too low a multiplier assumption may have contributed to the program’s growth shortfalls. In the case of Ukraine, staff suggested in ex post analysis that the deeper than expected contraction in output reflected excessive reliance on expenditure restraint rather than revenue measures to achieve fiscal objectives given their estimation of a higher expenditure multiplier (Mitra and Poghosyan, 2015).

In practice, in programs with a moderate degree of fiscal adjustment (typically PRGT programs and GRA programs out of a crisis), fiscal drag seems to have been quite modest, and countries were able to grow at or even better than their IEO-calculated benchmark, especially when benefiting from favorable supply conditions. By contrast, in a number of GRA crisis cases (e.g., Latvia 2008, Mongolia 2009, and Romania 2009), deep fiscal adjustment was associated with even deeper output contractions than projected. However, it is hard to assess the counterfactual given that weaker policy commitments could have undercut confidence gains and other factors such as credit constraints were also at play.

Revenue Mobilization

Revenue mobilization is growth friendly to the extent that it provides additional resources for priority spending, occurs in a non-distortionary way, and avoids imposing a burden on low-income groups. Evidence of improvements in tax mobilization is particularly marked in PRGT programs (Table 3). Relative to the pre-program period, the tax-to-GDP ratio rose on average by 1.1 percentage points in the post-program period, largely aided by increases in taxes on goods and services. Not surprisingly, tax mobilization was substantially stronger in completed programs (where the tax-to-GDP ratio increased on average by 3 percentage points between pre- and post-program periods) than in programs that went off track. Those countries with completed programs collected more revenues not only from taxes on goods and services but also taxes on income, thereby making the tax structure more progressive by bringing individuals and businesses with rising incomes into the tax net.

^{Table 3}

Tax and Expenditure Trends Associated with IMF-Supported Programs

(In percent of GDP)

Source: Gupta (2021). Note: “Pre-program” captures the three years prior to a program’s starting year; “Program” captures program years; “Post-program” captures the three years following a program’s ending year.

^{Table 3}

Tax and Expenditure Trends Associated with IMF-Supported Programs

(In percent of GDP)

		PRGT Programs			GRA Programs
		Pre-program	Program	Post-program	Pre-program	Program	Post-program
Taxes		13.6	14.4	14.7	20.1	20.5	20.4
	On income, profits, and capital gains	5.5	6.3	5.6	9.8	11.3	9.7
	On goods and services	3.7	4.5	4.3	7.6	6.9	8.2
	On international trade and transactions	2.7	2.6	2.7	2.3	2.1	1.8
Social contributions		0.9	0.9	0.8	6.9	6.6	6.3
General government total expenditure		25.1	24.8	26.2	35.7	33.7	33.0
General government expense		17.4	16.8	17.2	32.2	29.9	30.1
	Compensation of employees	6.8	6.4	7.3	9.5	9.3	9.2
	Purchases/use of goods and services	4.3	4.1	3.7	5.6	5.6	4.9
	Interest	1.2	1.5	1.6	2.9	2.9	2.8
	Social benefits	2.6	3.0	2.2	9.4	9.2	9.3
Capital expenditure		7.3	7.4	8.4	5.0	4.5	3.7
Net acquisition of financial assets		0.1	0.4	0.1	0.6	-0.1	-1.4

Source: Gupta (2021). Note: “Pre-program” captures the three years prior to a program’s starting year; “Program” captures program years; “Post-program” captures the three years following a program’s ending year.

View Table

^{Table 3}

Tax and Expenditure Trends Associated with IMF-Supported Programs

(In percent of GDP)

		PRGT Programs			GRA Programs
		Pre-program	Program	Post-program	Pre-program	Program	Post-program
Taxes		13.6	14.4	14.7	20.1	20.5	20.4
	On income, profits, and capital gains	5.5	6.3	5.6	9.8	11.3	9.7
	On goods and services	3.7	4.5	4.3	7.6	6.9	8.2
	On international trade and transactions	2.7	2.6	2.7	2.3	2.1	1.8
Social contributions		0.9	0.9	0.8	6.9	6.6	6.3
General government total expenditure		25.1	24.8	26.2	35.7	33.7	33.0
General government expense		17.4	16.8	17.2	32.2	29.9	30.1
	Compensation of employees	6.8	6.4	7.3	9.5	9.3	9.2
	Purchases/use of goods and services	4.3	4.1	3.7	5.6	5.6	4.9
	Interest	1.2	1.5	1.6	2.9	2.9	2.8
	Social benefits	2.6	3.0	2.2	9.4	9.2	9.3
Capital expenditure		7.3	7.4	8.4	5.0	4.5	3.7
Net acquisition of financial assets		0.1	0.4	0.1	0.6	-0.1	-1.4

Source: Gupta (2021). Note: “Pre-program” captures the three years prior to a program’s starting year; “Program” captures program years; “Post-program” captures the three years following a program’s ending year.

In GRA programs, the increase in the average tax-to-GDP ratio was much more modest, less than ½ percentage points. However, there is some evidence that the tax structure became more growth friendly in the post-program period. The dependence on distortionary trade taxes fell by ½ percent of GDP while reliance on more efficient taxes on goods and services increased by 0.6 percent of GDP. Crisis programs were no different in terms of tax mobilization outcomes from other completed GRA programs.

Expenditure Trends

Growth-friendly expenditure adjustment prioritizes resources for capital spending that can bring long-term benefits for the economy and social spending that supports low-income and vulnerable groups. Overall, total general government outlays increased in PRGT programs, buttressed by higher tax revenues (see Table 3, which is based on reported budgetary data). Moreover, there was a clear shift in the composition of spending in favor of capital projects. Capital spending rose on average by 1.1 percent of GDP between pre- and post-program periods, a desirable outcome for addressing the country’s infrastructure needs. However, little progress was made in raising social benefits—these rose during the program itself but subsided thereafter.³² Outlays for compensation of government employees increased during the post-program period, but this data set does not distinguish compensation to health and education workers, which could be associated with providing social support. As with revenues, performance was significantly stronger in on-track programs than in off-track programs, suggesting that program completion matters not just in the short run but also in the medium run.

In contrast, government spending declined in GRA programs by about 3 percent of GDP between pre- and post-program periods, reflecting in part greater focus on bringing down the fiscal deficit and providing room for the private sector to grow, as well as the more modest progress in raising revenues. Capital spending declined by 1.3 percent of GDP between pre- and post-program periods, while social benefits remained fat.

Data from the World Development Indicators allow for a more focused look at trends in education and health spending related to IMF-supported programs, and suggest that overall progress in raising such social spending in the program context was limited.³³ In GRA programs, health and education spending showed little movement relative to GDP, although health spending increased in the post-program period as a share of the total budget, suggesting some increased prioritization of the health sector in government budgets (Table 4). In PRGT programs where such spending was relatively low, health spending remained unchanged as a share of GDP while spending on education declined somewhat after the program ended. These results occurred even though PRGT program conditionality actively sought to protect or increase social spending, suggesting that program conditionality to protect or raise social spending in the short term was not enough to achieve sustained increases in such spending.³⁴

^{Table 4}

Trends in Health and Education Spending in IMF-Supported Programs

Source: Gupta (2021). Note: “Pre-program” captures the three years prior to a program’s starting year; “Program” captures program years; “Post-program” captures the three years following a program’s ending year.

^{Table 4}

Trends in Health and Education Spending in IMF-Supported Programs

GRA PROGRAMS
	In Percent of GDP			In Percent of Government Expenditure
Expenditure	Pre-program	Program	Post-program	Pre-program	Program	Post-program
Health	3.5	3.5	3.6	10.2	10.3	11.3
Education	4.7	4.2	4.6	13.8	12.9	13.8

PRGT PROGRAMS
	In Percent of GDP			In Percent of Government Expenditure
Expenditure	Pre-program	Program	Post-program	Pre-program	Program	Post-program
Health	1.9	2.0	1.9	7.1	7.7	6.3
Education	4.7	4.5	4.2	17.6	16.9	16.3

Source: Gupta (2021). Note: “Pre-program” captures the three years prior to a program’s starting year; “Program” captures program years; “Post-program” captures the three years following a program’s ending year.

View Table

^{Table 4}

Trends in Health and Education Spending in IMF-Supported Programs

GRA PROGRAMS
	In Percent of GDP			In Percent of Government Expenditure
Expenditure	Pre-program	Program	Post-program	Pre-program	Program	Post-program
Health	3.5	3.5	3.6	10.2	10.3	11.3
Education	4.7	4.2	4.6	13.8	12.9	13.8

PRGT PROGRAMS
	In Percent of GDP			In Percent of Government Expenditure
Expenditure	Pre-program	Program	Post-program	Pre-program	Program	Post-program
Health	1.9	2.0	1.9	7.1	7.7	6.3
Education	4.7	4.5	4.2	17.6	16.9	16.3

Source: Gupta (2021). Note: “Pre-program” captures the three years prior to a program’s starting year; “Program” captures program years; “Post-program” captures the three years following a program’s ending year.

Composition of Fiscal Adjustment

The case studies reveal a wide variety of country experience within the aggregate data presented earlier. While most PRGT programs in the case studies envisaged growth-friendly revenue-based fiscal consolidation, in some programs such as Ghana (2009, 2015), Honduras (2014), and Malawi (2010, 2018), significant upfront reductions in expenditures were incorporated, given that the reforms needed to raise revenue would take time to accomplish.

Initial commitment to growth-friendly fiscal adjustment often needed to be adapted in the face of implementation slippages. Programs had to pull back from growth-friendly fiscal strategies during implementation owing to weak revenue results, shortfalls in grants, and/or pressures on current spending, including strong resistance to public wage restraint. In a number of cases, these pressures led to cuts, as opposed to programmed increases in public investment (Benin, Jordan, Malawi, and Tunisia). In some cases, weak expenditure controls contributed to large accumulation of public domestic arrears with a negative impact on growth (e.g., Benin, Cameroon, Grenada, Malawi, and Senegal); subsequent clearance or reduction of arrears supported growth by freeing up resources for the private sector (Grenada, Romania, and Senegal).

In many case studies, revenue mobilization efforts had modest success during the program period. In a number of cases, tax reforms faced substantial internal opposition and complicated implementation challenges and were delayed or watered down. In some PRGT cases, an increase in tax revenue was achieved after the program or during a successor program (Benin), while in others it remained elusive even after reform was rolled out (Bangladesh). Revenue measures proved challenging in some GRA countries too, particularly in the face of political resistance (Jordan 2016).

Public Investment for Growth

In a number of the case studies, programs were able to accommodate significant increases in public investment. For example, in the case of Honduras, increased revenues and efforts to contain non-essential spending were instrumental to increase spending on infrastructure. In Senegal, increased access to external finance helped by use of IMF signaling instruments and flexibility in adapting the fiscal program to shortfalls in adjustment allowed a build-up in such spending. In Egypt, increases in capital spending by state-owned enterprises were not restrained by program conditionality. Targeted efforts at protecting public investment were generally successful. Bangladesh (2012), for example, protected 10 mega public investment projects (in areas such as transport, ports, and power generation) that were already underway at the start of the program, despite the associated strain on public finances and the current account. In other cases, public investment spending was ring-fenced from budget cuts and supported by external financing. In Latvia and Romania, pro-growth capital spending was supported by EU financing.

At the same time, case studies also raised questions about the medium-term growth impact of higher public investment, highlighting the need for strong management and governance, including greater attention to the assessment of infrastructure gaps, transparent selection of projects based on cost-benefit analysis, effective monitoring and execution, and improved debt management. While scaling up infrastructure spending to support growth was a primary fiscal objective of many PRGT programs, some countries accumulated debt rapidly and did not always see the expected productivity gains from increased public capital, raising concerns about the quality and growth impact of public investment as well as the risks created for debt sustainability (Benin and Senegal). In Senegal, for instance, debt tripled between 2008 and 2019. A Public Investment Management Assessment (PIMA) highlighted the low quality of public investment spending in Benin, Cameroon, and Senegal. Another factor leading to rapid increases in debt was of-balance sheet operations, including from state-owned enterprises, and hidden domestic arrears, pointing to the need for monitoring comprehensive fiscal targets and debt (Cameroon).

In response to governance concerns surrounding public investment, the IMF has provided TA to enhance the efficiency of public investment (Cameroon and Senegal) and to establish a legal framework for public private partnerships to encourage private sector participation in public investment (Benin). Partner institutions have also contributed to this effort. The World Bank provided extensive TA on public investment management for Cameroon, Jordan, Malawi, and Mongolia. The USAID also joined in this effort for Jordan, and the EU for Malawi. However, the results of such efforts have typically taken considerable time to materialize.

Subsidy Reform and Social Spending for Inclusive Growth

Case studies show widespread attention in program discussions to supporting social spending and inclusive growth. Approaches followed were adapted to account for institutional capacity and national preferences, and typically involved considerable support from the World Bank and regional development banks with deeper experience and expertise in this policy topic. Particular areas of focus included reforms of costly and distortionary energy and food subsidies and provision of social transfers to protect vulnerable groups.

Varying success was achieved in reforming energy subsidies. In Egypt (2016), energy subsidy reforms were extensively prepared, had strong political support and were accompanied by social assistance measures to protect the most vulnerable, such as additional food subsidies, cash transfers to the elderly and poor families, and other targeted social programs. Notwithstanding progress, however, concerns remained about the adequacy of targeting. In Jordan, energy subsidy reforms gained less traction because the cash transfer intended to protect the vulnerable from the impact of the automatic fuel price adjustment was poorly targeted, resulting in 70 percent of the population receiving this cash transfer. In some cases, political factors were behind delays or failures in energy sector reforms. In Ukraine, meaningful progress on curtailing gas losses was eventually achieved in the 2015 EFF after serious shortfalls in the 2008 and 2014 SBAs. Similarly, in Pakistan, progress was made in power sector reforms in the 2013 EFF after failure in this area in the 2008 SBA.

Two lessons emerge from this varied country experience. First, strong domestic political commitment and ownership are crucial to overcome resistance from vested interests. Second, particularly where progress is key to program success, the Fund may need to pay more attention to mobilizing technical support for reforms with development partners, including applying leverage related to the IMF’s access to senior decision makers, rather than taking a “hands of” attitude and effectively delegating responsibility to other agencies.

Turning to social safety nets, the case studies confirm that effective steps to safeguard vulnerable groups from the adverse impact of fiscal adjustment and energy subsidy reform can play an important part to maintain domestic support (Grenada, Honduras, and Malawi). However, in some cases (e.g., Romania 2009), reforms to the social safety net were too slow to provide protection to the most vulnerable. Nearly all programs aimed at improving the efficiency of social spending through better targeting. In some cases, the Fund adapted their recommended approach in view of national preferences (e.g., Latvia 2008) or faced a lack of political buy-in (e.g., Mongolia 2017). Given limited expertise of Fund staff in this area, social programs were often implemented with TA from the World Bank and other multilateral institutions (Bangladesh, Benin, Grenada, Honduras, Latvia, Malawi, Mongolia, Romania, Tunisia, and Ukraine).

More broadly, all PRGT programs have been required to include indicative targets establishing floors on social spending. Some GRA programs also established indicative floors on social spending (Jordan and Tunisia). Design of these floors has had to be adapted to data availability and are often broad rather than well targeted. As a result, even though the floors have typically been observed, in some cases, they have not been effective for safeguarding social expenditures for low-income and vulnerable groups (IEO, 2017). Malawi provides a case in point as its indicative floor on social spending was progressively lowered and the coverage was broadened over time.

The IMF’s increased attention to social spending was well-received by many, though not all, country authorities. Ghanaian authorities, for example, “appreciated that they had flexibility on the choice of fiscal measures” that allowed them to implement free high school education. Latvian authorities, however, expressed “concerns about program costs, possible adverse impacts on work incentives, and cultural preferences for family-based rather than government-based safety nets.” Malawi authorities, meanwhile, thought growth-promoting development programs should be prioritized ahead of social spending, which includes items that do not directly benefit the most vulnerable such as the wages and salaries of government employees. In some cases, authorities preferred to follow universal rights rather than targeted approaches usually advocated by the Fund, arguing that a targeting approach can be more expensive than universal programs and/or exclude large segments of vulnerable populations.⁴⁰ In Mongolia, for instance, the 2009 SBA included conditionality to shift from universal to targeted social transfers. But progress was difficult without a clear social and political consensus.

One general issue raised in many of the case studies relates to the lack of monitoring and reporting on the social impact of the overall program and specific policies to protect vulnerable groups. Very few staff reports provided much information in this area, which made it hard for this evaluation to reach conclusions on the extent to which the undoubted attention to protecting the vulnerable was actually successful in achieving its objectives. Even more important, the lack of a capacity to track effectiveness made it hard to identify emerging risks and assess the need for further reinforcing actions.

Fiscal Reforms and Growth

A clear lesson across the case studies is that meaningful fiscal reforms often take a long time to put in place, implying that the growth benefits are slow to materialize, often well beyond the program period. This experience implies a need for realism in setting program timelines for reform progress and in anticipating growth benefits. It also underlines the importance of building political support for reforms and sustaining reform efforts well after programs are completed.

A case in point is Bangladesh (2012) where the introduction of the VAT was the centerpiece of fiscal reform of the program. Notwithstanding the enactment of the VAT law in 2012 and a short extension of the program to give time for reform implementation, the tax reform was not implemented because of strong opposition from vested interests. Eventually, Bangladesh implemented a VAT four years after program completion, but a positive revenue impact has been slow to materialize.

The case of Mongolia also highlights the importance of continued efforts to preserve program achievements after the program; otherwise the hard-earned reform benefits may be lost. The major achievement under the 2009 SBA was the establishment of the fiscal stability law (FSL) which was set up to avoid boom-bust fiscal cycles related to world mineral prices and to ensure saving of a greater proportion of windfall revenues during the good times. However, the fiscal rules under the FSL were circumvented after the SBA was completed and, as a result, the boom intensified during 2011–14 as mineral production expanded, and commodity prices rose. When the boom turned to bust, authorities sought a new three-year program in 2017.

Country case studies on Egypt, Jordan, and Tunisia illustrate the importance of sequencing fiscal adjustments, garnering political support and avoiding reform fatigue. All three countries were affected by economic and social disruptions in the context of the Arab Spring uprisings and protracted regional conflicts. They all asked for Fund financial support to attain two key objectives—regaining macroeconomic stability and reinvigorating growth to address longstanding social problems at the root of the uprisings. Among the three, only Egypt was able to attain both the adjustment and growth goals although it took a number of years for Egypt to lay the groundwork before a program could be put in place. At the start of the program, the Egyptian authorities took decisive policy measures, centered on fiscal consolidation and the liberalization of the foreign exchange market, and used the savings from the elimination of fuel subsidies to strengthen social programs for the most vulnerable. Such early decisive actions helped the authorities to avoid reform and donor fatigue which emerged in the other two countries over time and fueled skepticism and opposition to reform.

The case studies illustrate the critical role of program ownership in implementing fiscal adjustment and reaping the benefit of fiscal reforms. In Latvia (2008), which involved bold, front-loaded fiscal adjustments, authorities retained strong political support despite a sharp initial growth downturn by clearly communicating to the public the national strategy for sustaining progress towards closer EU integration. The program achieved its goals and resolved external and fiscal imbalances relatively quickly. After three years of recession, growth resumed at a respectable pace and Latvia joined the Eurozone in January 2014. By contrast, Pakistan (2008 SBA) is a case in which limited political support undermined critical tax reform implementation. Despite massive IMF TA and program extensions to allow more to time to unlock the VAT reform, the tax reform failed to be implemented before the program expired in 2011 due to lack of political support. As the VAT reform remained politically infeasible, another program in 2013 sought revenue mobilization via a combination of several incremental reforms (e.g., scaling back tax exemptions, broadening the tax base, increasing goods and services taxes, and improving tax administration) and achieved partial success.

Domestic Arrears

A number of case studies highlight that large accumulation of public domestic arrears has a negative impact on growth and business climate (e.g., Benin, Cameroon, Grenada, Malawi, Romania, and Senegal). These strains typically received considerable attention in program design although implementation of needed reforms was often challenging. In Senegal (2015), a large increase in domestic arrears was not reported on a timely basis and required introduction of appropriate program conditionality. In Malawi (2012), program efforts to control government spending were partly circumvented by the accumulation of arrears to domestic suppliers, which led to increased nonperforming loans (NPLs) in the banking sector. Creditors were issued promissory notes amounting to a cumulative 9 percent of GDP between 2013 and 2018. The 2012 program also included conditionality on public financial management designed to better limit and track new arrears. Cameroon (2017) aimed at a gradual elimination of domestic arrears to contractors to reduce banks’ NPLs and unlock bank credit to the private sector, but efforts have not been effective, resulting in persistent domestic arrears. In Romania (2009, 2011), both authorities and staff viewed efforts to strengthen public sector arrears management as having particularly benefited growth and business climate as payment arrears by state-owned enterprises, local governments, and health sector bodies were an important burden on the business community. Pre-existing arrears were regularized, and disciplinary rules were established to deter new arrears. In Grenada (2014), the private sector benefited from improved fiscal management as the government had a history of accumulating significant arrears on domestic payables, which were eliminated under the program. In Benin (2017), business climate benefited from the completion of an audit of government arrears to domestic suppliers.