This Selected Issues paper analyzes sources of growth in Zambia. It assesses domestic debt dynamics and expenditure composition under two scenarios: first, the execution of the fiscal adjustment envisaged in the 2004 budget and in the medium-term framework; and second, the continuation of recent trends. The paper assesses movements in the external value of the kwacha using a number of different approaches. It also looks at developments in the real exchange rate and in differentials in rates of return on financial assets in Zambia and key financial centers.

Abstract

This Selected Issues paper analyzes sources of growth in Zambia. It assesses domestic debt dynamics and expenditure composition under two scenarios: first, the execution of the fiscal adjustment envisaged in the 2004 budget and in the medium-term framework; and second, the continuation of recent trends. The paper assesses movements in the external value of the kwacha using a number of different approaches. It also looks at developments in the real exchange rate and in differentials in rates of return on financial assets in Zambia and key financial centers.

II. Government Debt And Expenditure Composition 5

A. Summary

21. Recent fiscal developments in Zambia have been characterized by spending on nonpriority items in excess of budget allocations, which has led to increased domestic debt, higher real interest rates, and higher interest payments. Poverty-reducing spending has remained low, despite debt relief received at the decision point under the Heavily Indebted Poor Countries (HIPC) Initiative.

22. Against this background, this section assesses domestic debt dynamics and expenditure composition under two scenarios: (1) the execution of the fiscal adjustment envisaged in the 2004 budget and in the medium-term framework (“baseline scenario”), and (2) the continuation of recent trends (“alternative scenario”). Under the baseline scenario, structural reforms allow the authorities to phase out domestic financing, reduce interest payments, and significantly increase poverty-reducing spending. Under the alternative scenario, real interest rates remain constant at the 2003 average level, and all components of primary spending are crowded out by higher interest payments.

23. The analysis is conducted focusing not only on the usual indicators of debt sustainability (debt stock, debt service, and primary surplus as a share of GDP), but also on the main components of the primary balance: domestic revenues, primary non-poverty-reducing spending (mandatory and discretionary shares), and poverty-reducing spending. These additional indicators allow a consideration of the political economy of fiscal adjustment and a better assessment of the implications of a targeted primary surplus.

B. Recent Fiscal Developments

24. Before reaching the HIPC Initiative decision point in 2000, Zambia’s central government budget spent about 4 percent of GDP on external debt-service payments.6 Debt relief provided since 2001 has helped to reduce the burden on the budget. Projections factoring in the relief expected at the HIPC Initiative completion point indicate external payments declining from 2.3 percent of GDP in 2004 to 1.5 percent of GDP in 2008.

25. In contrast to the decline in external debt service, central government domestic interest payments rose from 1.4 percent to 2.8 percent of GDP over 2000–03 (Table II.1). Domestic debt increased from 7.7 percent to 21 percent of GDP over the same period.7 Recourse to domestic financing typically exceeded budget targets, as a result of expenditure overruns and shortfalls in donor assistance.

Table II.1.

Selected Fiscal and Macroeconomic Indicators, 2000-03

(As share of GDP, unless otherwise indicated)

article image
Sources: Zambian authorities; and staff estimates.

Debt service actually paid. Excludes debt service to IMF.

At cost value. Includes bonds issued in 2001 in favor of the Bank of Zambia (K 1,646 billion) and in 2002 in favor of the Zambia National Commercial Bank (K 248 billion).

Balance of central government operations excluding external resources and external payments.

Includes all grants and loans flowing in to the central government budget.

Program grants and loans minus external debt-service payments.

Simple average of real interest rates (annual average) on treasury bills and bonds.

26. The wage bill repeatedly exceeded budget allocations: From 2000 to 2003, it increased from 5.3 percent to 8.5 percent of GDP, as a result of both an increased head count (from 109,000 to 120,000, excluding military and security personnel) and large average wage awards (the average wage increased from 3.8 to 6.2 times per capita GDP). In response to these and other expenditure overruns, net program external aid (i.e., program grants and loans minus external debt service) declined.8 Lower-than-budgeted external aid could have been offset by some mix of increased domestic financing, higher revenues, or expenditure restraint. However, the authorities’ reliance on additional domestic credit exacerbated the debt dynamics and put pressure on interest rates.

27. The reliance on increased domestic financing, however, was not targeted to protect poverty-reducing programs. Execution of spending on those programs has been no more than 25 percent of the budgeted allocations since 2000. Moreover, spending on a subset of priority poverty-reducing programs remained, on average, below 1 percent of GDP, or 40 percent of the average annual HIPC Initiative assistance received since the HIPC Initiative decision point.9

C. Debt Dynamics: An Extended Equation

28. The starting point is the standard equation used to analyze debt dynamics.

dt=dt11+gt(1+it)pt(1)

where, in the year t, dt and pt are the domestic debt and the primary balance both expressed as a share of GDP, gt is the real GDP growth rate, and it is the average real interest rate applied to domestic debt.

This equation is expanded to identify a range of possible paths consistent with debt sustainability. The primary balances are looked at from “below the line” with domestic financing described as the difference between all the other budgetary inflows (domestic revenues and external grants and loans) and outflows (primary expenditure and external debt service):

dt=dt11+gt(1+it)[adrt+Δt+nestdpret(1αt)dnpretαtdnpret)](2)

where, as a share of GDP, adrt is the average of domestic revenue in the region (or in the countries at a similar level of development); Δt is the difference between the country’s domestic revenue and the regional average; nest is net external budgetary support (program grants and loans minus external debt service),10 dpret is domestic poverty-reducing spending; αt is the mandatory share of domestic primary non-poverty-reducing spending (i.e., wages, pensions, payments of arrears, and some counterpart fund for foreign-financed capital spending); and dnpret domestic primary non-poverty-reducing spending.11

29. The extended formula adds information to that provided by the usual macroeconomic indicators of sustainability (average primary surplus, debt stock and debt service as a share of GDP, revenues, or exports, in the case of external liabilities).12 In particular, the difference between the revenues collected and the average level collected in the region, Δt, as well as the scale or mandatory primary non-poverty-reducing spending, α · dnpre, allows us to assess the rigidity of budgetary revenues and expenditure. In this way, these indicators shed light on the feasibility of adjustment.

30. In Zambia, central government domestic revenues (19 percent of GDP in 2004) are in line with the average of the Common Market for Eastern and Southern Africa (COMESA) (excluding Seychelles), i.e., Δt = 0. This suggests relatively limited scope for adjustment on the revenue side. Primary non-poverty-reducing spending is estimated at 18 percent of GDP in 2003 (US$70 in per capita terms). The mandatory share of the primary non-poverty-reducing spending, α, is very high, at about 70 percent (12.5 percent of GDP). The discretionary spending is the residual 30 percent (5.5 percent of GDP). This suggests that expenditure policy reforms are key to moving the fiscal stance to a sustainable track while freeing up resources for rising poverty-reducing spending. Only modest savings can be generated by improving efficiency or tightening control on commitments.

D. Outlook Under Different Scenarios

The baseline scenario

31. The “baseline scenario” assumes, in line with the 2004 budget, that domestic financing is contained at 2 percent of GDP in 2004 and eliminated by 2008. The decline in domestic financing and domestic debt is accompanied by a fall in inflation and real interest rates (from 21.5 percent to 5 percent, and from 15.3 percent to 6 percent over 2003-08, respectively). Economic growth rates average almost 5 percent over the period.

32. The baseline scenario also envisages domestic revenues of 19 percent of GDP from 2004 onward, as well as savings on primary non-poverty-reducing spending due to policy and management reforms. Specifically, spending on primary non-poverty-reducing items (excluding items financed by project support from donors) is assumed to decline gradually—but significantly considering the large initial mandatory component—from 18 percent to 13 percent of GDP (or, in real per capita terms, from US$70 to US$52 per year).13 As a consequence of improved macroeconomic and structural performance, donors’ project and budgetary financial support—net of external debt service—remains above 7 percent of GDP over the period. Net external budgetary support increases by about 1 percentage point of GDP but remains slightly negative.

33. As a result of the fiscal adjustment, domestic debt would decline from 21 percent to 14 percent of GDP over 2003-08 (Table II.2 and Table II.3). Domestic interest payments remain above 3.5 percent of GDP in 2004 and 2005 but decrease significantly in the following years to 1.8 percent of GDP in 2008 (Figure II.1). Declining domestic borrowing underpins the credibility of the fiscal adjustment and accelerates the drop in real interest rates. A virtuous cycle operates, leading to rapid debt stabilization. Resources freed by structural reforms and declining domestic interest payments feed poverty-reducing spending, which rises by 3 percent of GDP to 5.4 percent of GDP, or from US$10 to US$22 per person each year.

Table II.2.

Main Results, 2003-08 (As share of GDP)

article image
Source: Staff calculations.
Table II.3.

Baseline Scenario-Main Assumptions and Results, 2003-08

article image
Source: Staff calculation.

Excludes expenditure financed by project support from donors.

Figure II.1.
Figure II.1.

Baseline Scenario: Debt Service

(As share of GDP)

Citation: IMF Staff Country Reports 2004, 160; 10.5089/9781451841206.002.A002

34. However, the increased spending on poverty-reducing programs in the baseline scenario may not be adequate to provide the additional resources required for progress toward the Millennium Development Goals (MDGs). Indeed, limiting the focus on health, the World Health Organization’s Commission on Macroeconomics and Health found that, on average, in the less-developed countries, the set of essential health interventions costs around US$34 per person per year. The commission also estimated that, by 2007, the less-developed countries should mobilize additional domestic resources of US$15 per person per year to supplement donors’ financial aid and adequately fund the programs in order to reach the MDGs on health.14 As indicated, total poverty-reducing spending in Zambia is projected to increase by US$12 over the same period. The projections under the baseline scenario therefore suggest that additional donor financial support may be necessary for Zambia to reach the MDGs.

The alternative scenario

35. Under the alternative scenario, no significant fiscal adjustment is undertaken. Consequently, in 2004, domestic financing remains close to 5 percent of GDP. Inflation accelerates to more than 40 percent in 2008, and real interest rates are assumed to remain constant at the 2003 levels (15.3 percentage points). Real GDP growth would fall back below population growth by the end of the period (from 3.5 percent in 2003 to 1.5 percent in 2008). Donors’ support of the budget would also decline. Despite declining primary expenditure, domestic financing reaches 10 percent of GDP, mainly to provide resources for rising domestic interest payments (Table II.4).

Table II.4.

Alternative Scenario-Main Assumptions and Results 2003-08

article image
Source: Staff calculation.

Excludes expenditure financed by project support from donors.

36. The alternative scenario results in mounting domestic financing and shift in expenditure composition required to meet the escalating domestic interest payments. Those payments rise to more than 12 percent of GDP in 2008 (Figure II.3 and II.4). The stock of domestic debt grows rapidly and climbs to 28 percent of GDP in 2008 (Table II.2). Increasing domestic borrowing and declining primary non-poverty-reducing spending do not protect primary poverty-reducing expenditure. The latter falls to 1.6 percent of GDP in 2008 or to US$6 per capita per annum.

Figure II.2.
Figure II.2.

Baseline Scenario: Expenditure Composition

(As share of GDP)

Citation: IMF Staff Country Reports 2004, 160; 10.5089/9781451841206.002.A002

Figure II.3.
Figure II.3.

Alternative Scenario: Debt Service

(As share of GDP)

Citation: IMF Staff Country Reports 2004, 160; 10.5089/9781451841206.002.A002

Figure II.4.
Figure II.4.

Alternative Scenario: Expenditure Composition

(As share of GDP)

Citation: IMF Staff Country Reports 2004, 160; 10.5089/9781451841206.002.A002

5

Prepared by Stefano Fassina.

6

External debt-service payments covered by budgetary resources exclude payments to the IMF, which are a liability of the Bank of Zambia.

7

The figures include the special bonds issued in 2001 in favor of the Bank of Zambia (K 1,647 billion or 12.6 percent of 2001 GDP), the bonds issued in 2002 in favor of the Zambia National Commercial Bank (K 248 billion or 1.5 percent of GDP), and the net balances of the Bridge Loan accounts in the Bank of Zambia.

8

Program grants and program loans are untied resources funding budgetary expenditure. Consequently, given a level of total expenditure, their volatility fully affects domestic financing. At the opposite end, project grants and project loans are resources tied to fund specific investments, and their amounts are fully reflected in expenditure. Consequently, their volatility does not affect domestic financing.

9

The set of poverty-reducing programs considered for the projections fully reflects the programs identified in the PRSP of Zambia. Consistently with this classification, expenditure on poverty-reducing programs excludes wages for education and health sector employees.

10

Project grants and loans net of foreign-financed capital expenditure.

11

The analysis also measures primary non-poverty-reducing spending in real U.S. dollars per capita.

12

The need to have additional indicators to assess debt sustainability in low-income countries (LICs) has been underlined in recent research and in an IMF board paper, where attention is focused on the primary surplus needed to reach a given target of debt-to-GDP ratio or debt-to-revenues or debt-to-export ratios. See A. Fedelino and A. Kudina, “Fiscal Sustainability in African HIPC Countries: A Policy Dilemma?” IMF Working Paper 03/187 (Washington: International Monetary Fund, 2003) and “Debt Sustainability in Low-Income Countries—Toward a Forward-Looking Strategy”.

13

Non-poverty-reducing spending (net of spending financed by project support from donors) is classified in the following economic categories (data refer to the 2003 estimated outturn): (i) wages (8.5 percent of GDP); (ii) retrenchment of civil service employees (0.1 percent of GDP); (iii) recurrent departmental charges (2.8 percent of GDP, excluding payments of domestic arrears); (iv) payments of domestic arrears (0.3 percent of GDP); (v) transfers to public sector institutions (constitutional bodies, universities, pension funds, the Board of Health, the Zambia Revenue Authority, etc.) and state-owned enterprises (4 percent of GDP in total); and (vi) domestically financed investments (1.4 percent of GDP, excluding counterpart funds for poverty-reducing investments).

14

World Health Organization—Commission on Macroeconomics and Health, Macroeconomic and Health: Investing in Health for Economic Development (2001). Currently, central government spending on health is about US$5 per person per year, including wages for health sector’s workers.

Zambia: Selected Issues and Statistical Appendix
Author: International Monetary Fund