Abstract
This paper analyzes the efforts taken to create fiscal space for the implementation of the fifth national development plan and the risk associated with it, examines the role of monetary policy in determining inflation, and discusses policy options to achieve low inflation. It also identifies areas where reform strategy needs more attention and suggests that reforms of financial system regulation need to be accelerated to ensure stability of the system. It analyzes traditional reserve adequacy measures, and finds looming power crisis as an obstacle to growth.
I. Creating Fiscal Space to Implement the Fifth National Development Plan1
1. If the government is to implement its development strategy, more fiscal space is needed. The Fifth National Development Plan (FNDP)—Zambia’s PRSP—sets out a policy framework for boosting growth and enhancing employment and income opportunities for the poor. The five-year plan (2006–10) has a strategic focus on investments in infrastructure and human resources, and provides detailed expenditure plans in these areas. However, the estimated financial requirements exceed projected resources by about 1.4 percent of GDP a year through 2010. This paper assesses how fiscal space can be created to fully implement the FNDP.
2. Full implementation of FNDP projects will not by itself ensure rapid progress toward the Millennium Development Goals (MDGs). The cost of meeting the MDGs is uncertain, in part because the relationship between government spending and outcomes is not clear.2 Some studies indicate that spending on education and health has a positive and significant impact on growth but stress that other policy interventions are also important, such as improving governance and containing inflation.3 Capacity to efficiently implement a larger number of projects also needs to be evaluated. Finally, the development of a vibrant private sector is critical, which requires a better business environment.
3. In this paper, Section A defines fiscal space and outlines channels through which it can be created. Section B briefly summarizes recent developments and the broad outlook for fiscal space. Section C then discusses each channel and assesses which are the most feasible for Zambia at the moment. Section D presents the conclusions.
A. Defining Fiscal Space
4. While the term “fiscal space” is relatively new, the concept is not. Fiscal space is generally understood as room in the government’s budget that allows it to provide resources for a desired purpose without jeopardizing the sustainability of its financial position or the stability of the economy. Though this challenge has always confronted governments, it has recently been receiving considerably more attention as low-income countries look for ways to finance the costs of meeting the MDGs.
5. Fiscal space can be created in many different ways. One set of channels involves increasing the resource envelope. This can be done through revenue mobilization, securing additional aid inflows, or borrowing resources from domestic or external sources. The other set of channels involves better utilizing existing resources. Specifically, lower priority spending can be cut to make room for higher priority spending, or existing expenditure could be made more effective. In each situation, the fiscal space must be created in such a way that fiscal sustainability and macroeconomic stability are protected.
6. The fiscal sustainability constraint suggests utilizing a medium-term expenditure framework. Higher expenditures in the short term, and any associated increase in future expenditures (e.g., to maintain capital investments or continue programs that require a long-term spending commitment, such as antiretroviral treatment) must be matched by an increase in current and future resources. This requires the difficult assessment of the impact of expenditures on growth and the future revenue base, and the sustainability of an increase in resources, which is particularly difficult for aid inflows.
7. The macroeconomic stability constraint also forces attention on the medium term. This constraint is most obvious in considering the impact higher aid flows may have on the exchange rate and the competitiveness of export industries (i.e., Dutch disease). However, it is also important for other channels. For example, eliminating accelerated depreciation allowances may increase company income tax revenue in the short term, but it may also deter capital investment, jeopardizing future growth. This constraint also rules out printing money to finance additional government spending.
B. Recent Developments and Outlook for Fiscal Space
8. Fiscal space in Zambia seems to have shrunk in recent years. In 2003–06, declining donor inflows and poor revenue performance, combined with the need to reduce domestic borrowing to stabilize the economy, caused a sharp compression on expenditures (Table 1). However, the 2007 Article IV review suggests that this should now reverse. Specifically, higher revenues and falling domestic interest payments should more than offset a reduction in domestic borrowing. The external contribution is expected to hold steady. In regard to the use of fiscal space, capital expenditure is projected to increase quite sharply in line with the priorities of the FNDP. These issues are discussed in detail below.
Fiscal Space Indicators
Based on IMF staff projections.
A positive number indicates domestic interest savings.
Net of amortization and interest payments.
Reflects cash float. Nature of expenditure cannot be determined.
Fiscal Space Indicators
| 2003-2006 | 2006-20101 | |||
|---|---|---|---|---|
| Source of fiscal space | −5.2 | 1.1 | ||
| Domestic contribution | −2.7 | 1.3 | ||
| Revenues | −1.1 | 2.1 | ||
| Domestic interest payments2 | 1.0 | 0.6 | ||
| Domestic financing (net) | −2.6 | −1.4 | ||
| External contribution | −2.5 | −0.2 | ||
| Grants | −2.4 | −0.6 | ||
| External financing (net)3 | −0.1 | 0.4 | ||
| Use of fiscal space | −5.2 | 1.1 | ||
| Wages | −1.2 | 1.4 | ||
| Other current spending | 2.6 | −2.5 | ||
| Capital expenditure | −7.3 | 3.5 | ||
| Discrepancy4 | 0.7 | −1.2 | ||
Based on IMF staff projections.
A positive number indicates domestic interest savings.
Net of amortization and interest payments.
Reflects cash float. Nature of expenditure cannot be determined.
Fiscal Space Indicators
| 2003-2006 | 2006-20101 | |||
|---|---|---|---|---|
| Source of fiscal space | −5.2 | 1.1 | ||
| Domestic contribution | −2.7 | 1.3 | ||
| Revenues | −1.1 | 2.1 | ||
| Domestic interest payments2 | 1.0 | 0.6 | ||
| Domestic financing (net) | −2.6 | −1.4 | ||
| External contribution | −2.5 | −0.2 | ||
| Grants | −2.4 | −0.6 | ||
| External financing (net)3 | −0.1 | 0.4 | ||
| Use of fiscal space | −5.2 | 1.1 | ||
| Wages | −1.2 | 1.4 | ||
| Other current spending | 2.6 | −2.5 | ||
| Capital expenditure | −7.3 | 3.5 | ||
| Discrepancy4 | 0.7 | −1.2 | ||
Based on IMF staff projections.
A positive number indicates domestic interest savings.
Net of amortization and interest payments.
Reflects cash float. Nature of expenditure cannot be determined.
C. Channels for Creating Fiscal Space
Revenue mobilization
9. Revenue mobilization is a favored source of fiscal space because it avoids many of the difficulties with aid and borrowing (see below). However, the extent to which revenues can be increased is limited.4 One consideration is that revenues involve a transfer of resources from the private to the public sector. The benefit of greater public spending therefore needs to be balanced against the costs of reducing the resources of the private sector. An additional consideration is that taxes tend to have a distortionary impact on the decisions of firms and employees. Such distortions can be kept to a minimum by broadening the tax base so that rates can be kept as low as possible.
Central Government Revenue
(Percent of GDP)
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Source: WETA.Central Government Revenue
(Percent of GDP)
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Source: WETA.Central Government Revenue
(Percent of GDP)
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Source: WETA.10. There seems to be scope to increase revenue to at least 19 percent of GDP. While Zambia’s revenue effort is reasonable by regional standards, the decline in 2001–2006 contrasts starkly with an increase on average elsewhere. Explanations for the decline include an increase in tax evasion, tariff reductions, frequent suspension of excises and customs duties on petroleum products, and, in 2006, a sharp appreciation of the kwacha. The FNDP projects that revenues will increase from 16.8 percent of GDP in 2006 to 18.2 percent by 2010 as administration improves and the contribution of the mining sector increases (see below). The projections have since been revised upward, but a financing gap remains.5
11. Strengthening administration will be a key to increasing revenue. A significant decline in VAT productivity during a period of largely unchanged exemptions suggests an increase in tax evasion. After peaking at almost 36 percent in 2001, productivity fell to 26 percent in 2006, well below the average in selected African countries (Table 2).6 Current efforts to improve administration focus on two areas. First, funding to the Zambia Revenue Authority (ZRA) has been increased to correct the gradual erosion in staffing and enforcement. Second, the ZRA is moving toward an integrated and taxpayer-segmented structure. These reforms, similar to those undertaken in Kenya, Uganda, and elsewhere should in time boost revenues and efficiency. An increase in productivity to the regional average would generate almost 1 percentage point of GDP in additional revenue.
Cross-Country Comparison on VAT Productivity
Cross-Country Comparison on VAT Productivity
| Country | Applicable Year | Standard VAT Rate (percent) | VAT Revenue (percent of GDP) | VAT Productivity (percent) |
|---|---|---|---|---|
| South Africa | 2004 | 14.0 | 7.0 | 50.0 |
| Ethiopia | 2005 | 15.0 | 5.7 | 37.7 |
| Botswana | 2006 | 10.0 | 3.6 | 36.2 |
| Malawi | 2004 | 17.5 | 6.3 | 36.0 |
| Kenya | 2004 | 16.0 | 5.4 | 33.8 |
| Ghana | 2005 | 12.5 | 4.2 | 33.7 |
| Namibia | 2004 | 15.0 | 5.0 | 33.3 |
| Mozambique | 2004 | 17.0 | 4.7 | 27.5 |
| Tanzania | 2005 | 20.0 | 5.3 | 26.5 |
| Zambia | 2006 | 17.5 | 4.5 | 25.7 |
| Madagascar | 2005 | 18.0 | 4.4 | 24.5 |
| Uganda | 2005 | 18.0 | 4.2 | 23.3 |
| Average (excluding South Africa) | … | 16.0 | 4.8 | 30.7 |
Cross-Country Comparison on VAT Productivity
| Country | Applicable Year | Standard VAT Rate (percent) | VAT Revenue (percent of GDP) | VAT Productivity (percent) |
|---|---|---|---|---|
| South Africa | 2004 | 14.0 | 7.0 | 50.0 |
| Ethiopia | 2005 | 15.0 | 5.7 | 37.7 |
| Botswana | 2006 | 10.0 | 3.6 | 36.2 |
| Malawi | 2004 | 17.5 | 6.3 | 36.0 |
| Kenya | 2004 | 16.0 | 5.4 | 33.8 |
| Ghana | 2005 | 12.5 | 4.2 | 33.7 |
| Namibia | 2004 | 15.0 | 5.0 | 33.3 |
| Mozambique | 2004 | 17.0 | 4.7 | 27.5 |
| Tanzania | 2005 | 20.0 | 5.3 | 26.5 |
| Zambia | 2006 | 17.5 | 4.5 | 25.7 |
| Madagascar | 2005 | 18.0 | 4.4 | 24.5 |
| Uganda | 2005 | 18.0 | 4.2 | 23.3 |
| Average (excluding South Africa) | … | 16.0 | 4.8 | 30.7 |
12. Broadening the tax base would further increase revenues. Most VAT systems contain some exemptions due to equity considerations. However, this needs to be weighed against the cost of complicating administration and distorting decision making. Moreover, well-off families typically benefit more in absolute terms because they tend to consume more of the exempted items than poor families. Studies in Ethiopia (Munoz and Sang-Wook Cho, 2003) and South Africa (Alderman and del Ninno, 1999) suggest that exempting basic food items can help make VAT more progressive, but other exemptions tend to be regressive. Eliminating selected exempt and zero-rated items would boost revenue slightly.7 Amending Section 89 of the Customs and Excise Act to limit its use to reduce rates—it has been used to suspend or reduce taxes on petroleum products when the refinery is shut down to protect the consumer from higher prices—would also help to protect revenues.
13. The use of tax incentives should be strictly limited. Investment incentives based on tax holidays were reintroduced in early 2007 for companies operating in multi-economic facility zones and a long list of priority sectors. While additional investment is needed, international experience demonstrates that investors give lower priority to tax incentives than to a competitive investment climate characterized by sound economic policies and institutions; political and economic stability; quality infrastructure; a productive workforce; and transparent tax rules and administration.8 In any case, a recent study by the Foreign Investment Advisory Service showed that the effective tax burden in Zambia is already regionally competitive. The government should take stock of all tax incentives and begin reporting them in the budget documents, with a view to quantifying the cost and streamlining them to broaden the revenue base.
14. There is limited scope to raise tax rates. The VAT rate (17.5 percent) and the standard corporate income tax rate (35 percent) are already quite high, there is a desire to reduce rather than increase the burden on personal income taxpayers, and import duties are declining as trade is being liberalized. Rate increases should only be contemplated for excises and in terms of unifying corporate rates to reduce tax-induced distortions.9 However, unifying corporate rates at 30 percent would involve a revenue loss because the increase from sectors currently paying 15 percent would be less than the reduction in revenue from those now paying the higher rates.
Increased from 25 to 30 percent for new investments in the 2007 Budget.
| Corporate tax rates | ||
| Agriculture, nontraditional exports, fertilizer | 15 | |
| Mining1 | 30 | |
| Banking | ||
| First K250 million | 35 | |
| Remainder | 40 | |
| Manufacturing and other | 35 | |
Increased from 25 to 30 percent for new investments in the 2007 Budget.
| Corporate tax rates | ||
| Agriculture, nontraditional exports, fertilizer | 15 | |
| Mining1 | 30 | |
| Banking | ||
| First K250 million | 35 | |
| Remainder | 40 | |
| Manufacturing and other | 35 | |
Increased from 25 to 30 percent for new investments in the 2007 Budget.
15. Generous fiscal terms and the write-down of large investments have limited fiscal revenues from the mining sector. As part of the privatization process that began in late 1994, the government entered into long-term development agreements with the companies taking over public mining assets. The agreements provided generous fiscal terms and helped to promote investment in the sector when copper prices were low. The fiscal terms included a royalty of 0.6 percent of gross value; corporate income tax at 25 percent; depreciation for tax purposes at 100 percent; withholding taxes at zero percent, except on construction and technical services supplied by nonresidents; customs duty exemptions for capital equipment imports; and limits on duties payables for consumables. These terms and the write-down of large investments have limited the government’s share from the copper boom, although revenues have started to pick up (Figure 1) and the dates at which companies begin to declare taxable profits has been brought forward.
Projected Government Revenue from the Mining Sector
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Projected Government Revenue from the Mining Sector
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Projected Government Revenue from the Mining Sector
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
16. The fiscal regime for new mining operations has been revised and the government is renegotiating development agreements. In recognition of high copper prices and the generally more favorable investment climate, the government has revised the mining sector fiscal regime to be more consistent with international standards. Changes announced in the 2007 budget included raising the royalty rate to 3 percent, the corporate income tax rate to 30 percent, and the withholding tax rate to 15 percent. These terms currently apply only to new investments, but government is renegotiating with mining companies the fiscal terms of current development agreements, arguing that copper prices far exceed the levels anticipated by either party when most of the agreements were signed.10
17. Government revenues from the mining sector could increase significantly. To estimate possible future revenues, a simulation model was used to project revenues from the three largest mines. The model excludes large tax payments from a fourth mine that imports ores from outside Zambia. Estimates were generated using three alternative copper price projections (IMF World Economic Outlook projections, a more gradual decline, and a sharper decline) and using the fiscal terms in existing development agreements and the revised regime from 2008 (Figure 1). The results, which should be viewed as illustrations of possible outcomes rather than precise forecasts, show that mining revenues could rise significantly. However, the size and sustainability of the increase will be sensitive to changes in copper prices and the outcome of renegotiations, and the sharp pick up projected for 2008 could be delayed.11
Changing the composition of expenditure
18. Changing the composition of spending in favor of priority sectors is another favored source of fiscal space. A sharp decline in total expenditure—brought on by the fall in revenue, a decline in donor financed spending, and a reduction in borrowing to more prudent levels (see below)—highlights the need to prioritize spending. The data on health and education spending in sub-Saharan Africa is poor because donor-financed off-budget activities are not completely covered, but the data available suggests that the share Zambia allocates to health is reasonable by regional standards but still below indicative targets12; funding for education is considerably below average (Figure 2).
Government Spending on Health and Education
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Source: World Health Organization. Figures are compiled to ensure comparability across countries; official statistics of the countries may be based on alternative methods.Government Spending on Health and Education
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Source: World Health Organization. Figures are compiled to ensure comparability across countries; official statistics of the countries may be based on alternative methods.Government Spending on Health and Education
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Source: World Health Organization. Figures are compiled to ensure comparability across countries; official statistics of the countries may be based on alternative methods.
Central Government Expenditure
(Percent of GDP)
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Source: WETA.Central Government Expenditure
(Percent of GDP)
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Source: WETA.Central Government Expenditure
(Percent of GDP)
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Source: WETA.
Education Expenditure in Percent of Total Government Expenditure
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Source: IMF country documents and World Development Indicators. Data for Kenya are for 2002 and 2004.Education Expenditure in Percent of Total Government Expenditure
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Source: IMF country documents and World Development Indicators. Data for Kenya are for 2002 and 2004.Education Expenditure in Percent of Total Government Expenditure
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Source: IMF country documents and World Development Indicators. Data for Kenya are for 2002 and 2004.19. Changing the composition of expenditure requires sound budget planning. With limited resources and many competing priorities, governments must assess the marginal benefit of increased spending to some sectors against the marginal cost of reducing spending elsewhere. Such an assessment is complicated by the fact that many line ministries lack the capacity to present a compelling case for how additional funds would be used to produce specific outcomes. Reprioritization in the short term is made even more difficult because a considerable share of spending is not discretionary—interest payments and wages comprised 40 percent of Zambia’s total expenditure in 2006.
20. The FNDP calls for a significant increase in the share of the budget allocated to priority sectors. However, slippages have already occurred with the 2007 budget allocation for education and health below FNDP targets (Table 3). It is vital that future targets are met, or priority sectors will be underfunded and the true financing gap—what is needed to meet objectives—may be greater than currently estimated. To achieve this, the government will need to constrain non-priority expenditures, including by stepping up implementation of civil service and pay reforms, and, more broadly, to better integrate the FNDP, the medium-term expenditure framework and annual budgets.
FNDP Targets for Budget Shares1
(Percent of GDP)
Includes government and donor funds. The figures for 2006 are actual spending.
Budget allocation.
Excludes some major infrastructure needs (for example, in the energy sector).
FNDP Targets for Budget Shares1
(Percent of GDP)
| 2006 | 2007 | 20072 | 2008 | 2009 | 2010 | |
|---|---|---|---|---|---|---|
| Health | 10.7 | 10.7 | 10.7 | 12.5 | 13.2 | 14.0 |
| Education | 16.1 | 17.1 | 15.0 | 17.8 | 19.5 | 20.3 |
| Agriculture | 5.7 | 6.9 | 8.8 | 8.5 | 7.8 | 9.2 |
| Infrastructure3 | 12.5 | 12.0 | … | 10.7 | 11.3 | 9.2 |
Includes government and donor funds. The figures for 2006 are actual spending.
Budget allocation.
Excludes some major infrastructure needs (for example, in the energy sector).
FNDP Targets for Budget Shares1
(Percent of GDP)
| 2006 | 2007 | 20072 | 2008 | 2009 | 2010 | |
|---|---|---|---|---|---|---|
| Health | 10.7 | 10.7 | 10.7 | 12.5 | 13.2 | 14.0 |
| Education | 16.1 | 17.1 | 15.0 | 17.8 | 19.5 | 20.3 |
| Agriculture | 5.7 | 6.9 | 8.8 | 8.5 | 7.8 | 9.2 |
| Infrastructure3 | 12.5 | 12.0 | … | 10.7 | 11.3 | 9.2 |
Includes government and donor funds. The figures for 2006 are actual spending.
Budget allocation.
Excludes some major infrastructure needs (for example, in the energy sector).
Increasing the efficiency of expenditure
21. Increasing the efficiency of expenditure is desirable whether or not there is a need for fiscal space. While higher spending on priority sectors is important, greater efforts are needed to ensure that current and additional funding translates into improved outcomes. If this can be done, it may be possible to reach FNDP objectives and the MDGs with fewer resources than are currently envisaged. Moreover, if the health and education ministries can demonstrate a capacity to effectively use increased funds, government and donors may be more willing to commit the resources that are needed.
22. The efficiency of spending on health and education can be assessed by measuring how effective Zambia is in producing outcomes relative to other countries. Data Envelopment Analysis (DEA) is a nonparametric technique that estimates efficiency by relating inputs (e.g., health expenditure) to outputs (e.g., immunization rates).13 The countries that produce the maximum output for a given level of inputs (or the minimum input for a given level of output) define the best-practice frontier; the distance of a country to the frontier is a measure of relative efficiency. The results, which should be interpreted with caution, indicate that in Zambia education spending is relatively more efficient than health spending (Appendix). Public expenditure tracking surveys can also be used to identify inefficiency and inform policy design (Box 1).
23. Reforming agricultural policy should be a priority. Donors have heavily criticized the food strategic reserve and fertilizer support programs, which should end as planned starting with the 2009 budget. In the former program the Food Reserve Agency buys surpluses at substantial cost to the budget (2.2 percent of the 2007 domestic budget) in favor of allowing maize exports. The fertilizer subsidy program also absorbs significant resources (1.6 percent of the 2007 domestic budget). However, a recent poverty and social impact analysis found that the most vulnerable still had limited access to fertilizer, and those who do acquire it do not always use it effectively.14
Public Expenditure Tracking Survey (PETS) on Education
PETS is a method used to track the flow of government resources through different levels of government and administration. It provides information on the leakage of funds, the degree of targeting, and household responses.
A 2004 World Bank study on education in Zambia distinguished between funds that were disbursed to provincial and district offices and then allocated to schools at the discretion of education officers and those that were allocated directly to schools through a transparent rule. It found that schools got less than 20 percent of discretionary funds but more than 90 percent of rule-based funds. Discretionary funds were also found to be allocated less progressively.
However, the study found that even rules-based funds did not improve test scores. One reason for this is that households responded to an increase in public funds by cutting back the same amount in private contributions. The other is that schools found it difficult to spend the money on much-needed teachers because there were few trained personnel (this would have reduced the first problem in that teachers are an input that households are unable to purchase).
These findings support the free basic education policy announced in 2002. The policy has been very successful in its stated aim of increasing attendance (primary school enrollment went from 65 percent in 2002 to 89 percent in 2005). Moreover, since households no longer have to contribute, it has also been beneficial in that an increase in government funds can no longer by offset by a reduction in private expenditure.
A mini-PETS update was carried out in 2007. This study documented that schools in the sample reported very high pupil-teacher ratios, insufficient infrastructure, and a lack of learning material. Despite an increase in overall funding, improvements in these areas have been slow because enrollment rose so sharply and teachers salaries were upgraded (to regionally comparable levels). The report also noted that funding is still substantially below the regional average, and the bulk of it is not received until the fourth quarter, which complicates budget execution.
24. Current efforts to strengthen public financial management (PFM) should make spending more efficient. Introduction of the automated payroll management and establishment control system has helped to eliminate a large number of ghost workers from public sector payrolls. However, a recent PFM assessment15 identified other areas for improvement. The areas cover both budget planning and execution. In particular, the variation between budget allocations and outturns needs to be minimized, and funds need to be released earlier to enable ministries to implement programs as intended. Improved fiscal reporting and more active parliamentary, press, and civil society oversight would also help enforce accountability. The public expenditure management and financial accountability program (2005-09) aims to address many of these issues; progress is being made. Careful implementation of the fiscal decentralization plan may also help to improve the quality, efficiency, and targeting of public services.
25. The donor community can help to improve priority-setting and monitoring of poverty-reducing expenditures. Large off-budget activities financed by donors hampers budget planning and monitoring. For example, the 2005 Financial Report suggests that the Ministry of Health received only 49 percent of budgeted funds. However, this figure is highly misleading. The budget incorporates only very approximate estimates of donor funding and related expenditures; actual receipts and expenditures are not monitored at all because they take place outside the government’s financial management system. To improve priority-setting and monitoring, donors should give government timely and detailed information on off-budget project support, and increase the share of assistance delivered as budget support based on multiyear commitments (see below).
Increased donor support
26. A number of challenges need to be dealt with to maximize the benefits of any scaling up of aid. In particular, the timing, magnitude, and duration of aid disbursements tends to be more uncertain and volatile than revenues.16 This complicates medium-term fiscal planning and can translate into spending volatility. Donors can help by moving to predictable multiyear commitments.17 Zambia can itself take some steps, such as subjecting medium-term expenditure plans to stress tests; smoothing expenditures by saving resources during aid surges; identifying ahead of time priority spending that would be protected from cuts if there is an aid shortfall; and incorporating flexibility into nonpriority spending programs. Another challenge is to build capacity to effectively absorb more aid, which could be demonstrated through PFM reforms and execution of annual budgets that are in line with the FNDP. The possibility that scaled-up aid would reduce incentives to mobilize revenues also needs to be guarded against.
27. The macroeconomic consequences of scaling up need to be managed. This would require that fiscal and monetary policy be better coordinated. Decisions need to be made on whether to spend (increase the fiscal deficit excluding aid) and absorb (widen the current account deficit excluding aid). The economic effects will depend on the policy combination chosen; the appropriate choice will depend on initial conditions (see the matrix below). Circumstances in Zambia support a spend and absorb approach, although for short periods choosing not to spend nor absorb part of the aid could help build foreign exchange reserves and smooth the impact of aid volatility. Any exchange rate pressure should be managed by adapting the timing of foreign exchange sales rather than spending but not absorbing the aid, as some countries have done.18
Matrix: Spending and Absorbing Aid
Actions, Consequences, and Appropriateness for Zambia
Matrix: Spending and Absorbing Aid
Actions, Consequences, and Appropriateness for Zambia
| Absorbed | Not Absorbed | |
|---|---|---|
| Spent | Bank of Zambia sells aid dollars and fiscal deficit rises as aid is spent. No change in foreign exchange reserves or the money supply. Risk of Dutch disease. Appropriate for Zambia. Appreciation pressures can be mitigated by initially using aid to build productive capacity and import goods, and back-loading expenditures on health and education. However, this may delay attainment of some MDG objectives. | Bank of Zambia holds on to aid and fiscal deficit rises as aid is spent. Foreign exchange reserves increase. if unsterilized, the money supply will rise, with a risk of inflation. If sterilized, the increase in the money supply will be contained but the private sector will be crowded out as domestic debt and interest rates rise. Not appropriate. Equivalent to fiscal stimulus in absence of aid. |
| Not Spent | Bank of Zambia sells aid dollars but fiscal deficit remains unchanged. Substitutes aid for domestic financing of government deficit. Can be useful in the short term to stabilize the economy if initial domestic financing is too high. However, this is not the case in Zambia. | Bank of Zambia holds on to aid and fiscal deficit remains unchanged. Foreign exchange reserves will increase. Can be useful in the short term to smooth volatile aid flows. However, it is not appropriate in the long run because it prevents a real resource transfer. |
Matrix: Spending and Absorbing Aid
Actions, Consequences, and Appropriateness for Zambia
| Absorbed | Not Absorbed | |
|---|---|---|
| Spent | Bank of Zambia sells aid dollars and fiscal deficit rises as aid is spent. No change in foreign exchange reserves or the money supply. Risk of Dutch disease. Appropriate for Zambia. Appreciation pressures can be mitigated by initially using aid to build productive capacity and import goods, and back-loading expenditures on health and education. However, this may delay attainment of some MDG objectives. | Bank of Zambia holds on to aid and fiscal deficit rises as aid is spent. Foreign exchange reserves increase. if unsterilized, the money supply will rise, with a risk of inflation. If sterilized, the increase in the money supply will be contained but the private sector will be crowded out as domestic debt and interest rates rise. Not appropriate. Equivalent to fiscal stimulus in absence of aid. |
| Not Spent | Bank of Zambia sells aid dollars but fiscal deficit remains unchanged. Substitutes aid for domestic financing of government deficit. Can be useful in the short term to stabilize the economy if initial domestic financing is too high. However, this is not the case in Zambia. | Bank of Zambia holds on to aid and fiscal deficit remains unchanged. Foreign exchange reserves will increase. Can be useful in the short term to smooth volatile aid flows. However, it is not appropriate in the long run because it prevents a real resource transfer. |
28. The prospect of greater donor support has increased with renewed international focus on accelerating progress toward the MDGs, but it is yet to materialize. At the 2005 Gleneagles meeting, the G8 leaders agreed to double aid to Africa by 2010. Leaders reaffirmed this commitment at the 2007 G8 meeting but details, including a schedule for the increased funding, have yet to be decided. From 1996 through 2005, aid flows to Zambia averaged 10 percent of GDP; however, in 2006 support fell to 5.2 percent of GDP. After discussions with donors, IMF baseline projections do not include any significant scaling up in the medium term.
29. The impact of debt relief on fiscal space is obscured by the recent decline in aid and the difficulties Zambia would have had in servicing the debt. HIPC, MDRI, and bilateral debt relief has sharply reduced debt service; forward projections illustrate the significance of future debt-service savings from MDRI alone (Figure 3). However, just as the government would have had difficulty in servicing the debt, it will find it difficult to increase expenditure by the full amount of projected savings.
External Debt Service and Official Development Assistance
(Percent of GDP (LHS) and US$ millions (RHS))
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
External Debt Service and Official Development Assistance
(Percent of GDP (LHS) and US$ millions (RHS))
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
External Debt Service and Official Development Assistance
(Percent of GDP (LHS) and US$ millions (RHS))
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Official Development Assistance
(Percent of GDP)
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Official Development Assistance
(Percent of GDP)
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Official Development Assistance
(Percent of GDP)
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Borrowing
30. Debt relief has provided scope for external borrowing to finance a limited number of projects. Domestic borrowing is expensive, with domestic interest rates far exceeding international rates. Moreover, a heavy reliance on domestic borrowing may place further upward pressure on interest rates because the pool of domestic savings is limited, crowding out private sector investment and increasing the fiscal burden—unwinding the benefits brought by recent fiscal consolidation.19 Concessional external loans are less expensive than domestic borrowing, while debt relief has raised the prospect of nonconcessional external borrowing. As past difficulties attest, however, nonconcessional borrowing should be used with caution and only be for amounts that ensure debt sustainability. A public debt management strategy and guidelines for effective project selection and monitoring should be in place before nonconcessional debt is contracted.
Central Government Balance Including Grants
(Percent of GDP)
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Source: WETA.Central Government Balance Including Grants
(Percent of GDP)
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Source: WETA.Central Government Balance Including Grants
(Percent of GDP)
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Source: WETA.D. Conclusion
31. Efforts to create fiscal space need to be spread across a number of areas. Strengthened administration should help to increase revenues over time. An increase in the contribution from the mining sector could provide a quick and substantial boost, particularly if the government can renegotiate with mining companies the fiscal terms of current development agreements. Options to broaden the tax base, including by eliminating unwarranted VAT and import duty exemptions and restricting the use of tax incentives, should be explored, although the gains that can be expected are limited. More effective use of resources by constraining nonpriority spending and reforming PFM to increase spending efficiency is a priority. Better PFM should also help the authorities in their efforts to encourage donors to scale up aid. If donor assistance does not increase, nonconcessional external borrowing could be considered for a limited number of clearly viable projects that do not jeopardize debt sustainability.
32. Risks associated with full implementation of the FNDP will need to be carefully managed. While there is an urgent need to increase spending on infrastructure and human resources, the authorities will need to closely monitor the effects of accelerated public spending on inflation and the real exchange rate. Front-loading infrastructure investments and other projects that are import-intensive and increase the productive capacity of the economy would help to limit crowding-out of the private sector. However, delaying social spending could slow the initial advance of the non-income-related human development MDGs. If increased spending is financed by a scaling up of aid, close coordination of fiscal and monetary policy will be important. The authorities’ capacity to efficiently implement a larger number of projects also needs to strengthen. It would be helpful for donors to increase the predictability of their aid and harmonize it with existing government programs.
Appendix
33. Based on a sample of 40 African countries, efficiency frontiers were estimated for six education indicators and six health indicators (data were not available for all countries for all indicators).20 The education indicators were literacy rate, primary school enrollment, persistence until grade 5, ratio of girls to boys, ratio of young literate females to males, and trained teachers in primary education. The health indicators were the measles immunization rate, infant survival rate; under-5 survival rate, maternal survival rate, health worker density ratio, and births attended by skilled health staff. All indicators reflect the latest data from the World Bank World Development Indicators (WDI) database. On the input side, data were sourced from IMF country documents, the WDI database, and the World Health Organization, and are measured in purchasing power parity (PPP) dollars to correct for differences due to relative prices.
34. DEA methods have a number of limitations:
a. First, the results are sensitive to the quality of data, which raises a number of concerns. On the inputs, donor funding is very important for the health sector but is not always fully captured, and the data on education only cover public spending. On the outputs, indicators may not accurately reflect outcomes; for example, enrollment rates are not a direct measure of learning. The use of aggregate input data and disaggregated output data is also an issue. It would have been preferable to split education expenditure into primary and secondary education.
b. Second, outcomes can be influenced by factors other than spending, such as culture, lifestyle, and the environment.
c. Finally, the results can be very sensitive to the sample and the presence of outliers.
35. Figure A1 illustrate the efficiency frontier for representative input-output combinations. Based on efficiency scores (distance to the frontier), Zambia had an average country ranking of 3.5 across the six education indicators (average sample size: 27 countries) and 16.5 across the six health indicators (average sample size: 39 countries). Gupta and Verhoeven (2001) found that the efficiency of education spending is lower in African countries than in Asian and Western Hemisphere countries, which suggests that if the sample were broadened, Zambia’s relative standing—along with the other countries in the sample—might decline.
Relative Efficiency of Public Expenditure on Education and Health for 39 Countries from Sub-Saharan Africa
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Relative Efficiency of Public Expenditure on Education and Health for 39 Countries from Sub-Saharan Africa
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
Relative Efficiency of Public Expenditure on Education and Health for 39 Countries from Sub-Saharan Africa
Citation: IMF Staff Country Reports 2008, 029; 10.5089/9781451841329.002.A001
References
Alderman, H. and C. del Ninno, 1999, “Poverty Issues for Zero Rating VAT in South Africa,” Journal of African Economies, Vol. 8, No. 2, pp. 182 –208.
Baldacci, E., B. Clements, S. Gupta, and Q. Cui, 2004, “Social Spending, Human Capital and Growth in Developing Countries: Implications for Achieving the MDGs,” IMF Working Paper 04/217 (Washington: International Monetary Fund).
Berg, A., S. Aiyar, M. Hussain, S. Roache, T. Mirzoev, and A. Mahone, 2007, The Macroeconomics of Scaling Up Aid: Lessons from Recent Experience, Occasional Paper No. 253 (Washington: International Monetary Fund).
Bulir, A., and A. Hamann, 2006, “Volatility of Development Aid: From the Frying Pan into the Fire?” IMF Working Paper 06/65 (Washington: International Monetary Fund).
Gupta, S., and M. Verhoeven, 2001, “The Efficiency of Government Expenditure: Experiences from Africa,” Journal of Policy Modeling, Vol. 23, pp. 433 –67.
Heller, P. 2005, “Understanding Fiscal Space,” IMF Policy Discussion Paper 05/4 (Washington: International Monetary Fund).
Herrera, S., and G. Pang, 2005, “Efficiency of Public Spending in Developing Countries: An Efficiency Frontier Approach,” World Bank Research Working Paper No. 3645.
Jorgensen, S., and Z. Loudjeva, 2005, “A Poverty and Social Impact Analysis of Three Reforms in Zambia: Land, Fertilizer and Infrastructure,” Social Development Papers No. 49 (Washington: World Bank).
Mphuka, C., 2005, The Cost of Meeting the MDGs in Zambia, A Research Report Commissioned by the Civil Society for Poverty Reduction, Jesuit Centre for Theological Reflection, and The Catholic Centre for Justice Development and Peace.
Munoz, S., and S. Sang-Wook Cho, 2003, “Social Impact of Tax Reform: The Case of Ethiopia,” IMF Working Paper 03/232 (Washington: International Monetary Fund).
Zee, H., J. Stotsky, and E. Ley, 2002, “Tax Incentives for Business Investment: A Primer for Policy Makers in Developing Countries,” World Development, Vol. 30, pp. 1497 -1516.
Prepared by Brenton Goldsworthy.
Mphuka (2005) estimates that to reach the MDGs both the government and cooperating partners would need to double their financing to this area between 2006 and 2015.
See, for example, Baldacci, Clements, Gupta and Cui (2004).
According to Heller (2005), a minimum objective for low-income countries should be to raise the tax share to at least 15 percent.
The draft MTEF 2008–2010 projects that revenues will reach 18.7 percent of GDP by 2010 (updated staff estimates suggest revenues could reach 19 percent), reflecting a higher expected base in 2007.
VAT productivity here is VAT revenue expressed as a percentage of GDP divided by the VAT rate. A fall in the share of consumption in GDP could explain part of the decline in reported productivity. Similarly, variances across countries may in part be explained by different consumption shares.
A 2006 IMF technical assistance mission estimated that eliminating certain exempt and zero-rated items (e.g., package tours and nonbasic food items) would boost VAT revenue by 0.16 percentage points of GDP.
See, for example, Zee, Stotsky, and Ley (2002).
In the 2007 budget, excise rates were increased on cigarettes, clear beer, and motor vehicles, and the personal income tax burden was reduced by adjusting the thresholds and reducing the top marginal rate.
Some development agreements were signed as late as 2006 when prices were already very high. IMF technical assistance missions have emphasized that fiscal regimes should not depend on output price forecasts. A resource rent tax or variable income tax are useful devices for securing a share of additional mineral rents for the government if mineral prices rise unexpectedly.
The large increase in 2008 reflects the model projection that the two mines not currently paying company income tax will begin to do so. If those mines are still writing off capital investments not captured in the model, the date at which they start recording taxable profits could be delayed beyond 2008.
At a 2001 health summit in Abuja, Nigeria, member governments of the Organization of African Unity set a target of allocating at least 15 percent of their annual budgets to improvement of the health sector. The WHO Commission on Macroeconomics and Health suggests that US$40 per person is the minimum amount necessary for effective delivery of a basic public health intervention.
There is a growing literature that uses the DEA method to measure the relative efficiency of public spending. See Herrera and Pang (2005) for a detailed description of the methodology. It is nonparametric in that there is no need to specify a functional form for the relationship between inputs and outputs. This is important because little is known about this relationship for education and health.
“Public Financial Management Performance Report and Performance Indicators” December 2005.
See Bulir and Hamann (2006) for a multicountry analysis of the volatility and predictability of aid flows. In Zambia, the standard deviation of revenue was 1.1, compared with 2.8 for gross aid for 1996–2006.
Donors pledged to do this in the Paris Declaration on Aid Effectiveness that was issued at The Paris High Level Forum in 2005.
Berg et al. (2007) and others have found that conflicting priorities of central banks and fiscal authorities have led to incomplete absorption.
Domestic interest payments fell by over 1 percentage point of GDP from 2004 to 2006 as the debt burden and interest rates declined.
To account for the influence that GDP per capita has been estimated to have on efficiency, the country sample was restricted to countries in Africa where data were available.
