Chapter

CHAPTER 4 The Policy Record

Author(s):
Susan Schadler, and Hugh Bredenkamp
Published Date:
June 1999
Share
  • ShareShare
Show Summary Details
Author(s)
Louis Dicks-Mireaux, Jean Le Dem, Steven Phillips and Kalpana Kochhar 

This chapter examines developments in the macroeconomic and structural policies of the ESAF countries during the period under review and takes stock of the progress achieved toward greater macroeconomic stability and more open, flexible, and market-driven economies. Three critical areas are covered: fiscal policy, exchange rate and monetary policies, and structural reform. The analysis of structural reform focuses on five key areas: domestic markets and prices, the exchange and trade system, financial sector reform, public enterprise reform, and fiscal reform. Significant advances, albeit uneven, were made in all three policy areas over the past ten years. Looking ahead, there remains considerable scope for further improvement in some areas.

The approach in this chapter is to review policies over a long period falling under the influence of SAF/ESAF-supported programs, rather than just under the limited (and often intermittent) periods when an ESAF-supported program was actually in effect. In addition, the section on fiscal policy examines the objectives, quantitative targets, and outcomes of the fiscal programs under SAF/ESAF-supported programs. The first SAF/ESAF arrangements for most countries began during 1986–88, and policies followed during the ten-year period from 1986–95 are reviewed in relation to policies in the pre-SAF/ESAF period of the early 1980s. In light of the severe data limitations for transition economies and the fact that SAF/ESAF arrangements for all but one of these countries began in 1993–94, much of the discussion is confined to nontransition economies.

Fiscal Policy

Fiscal policy in ESAF countries aimed both to strengthen the government’s financial position in the medium term and to reform the structure and administration of the revenue system, as well as public expenditure management and spending priorities. The latter reforms were often as important as improving the fiscal position itself, in as much as they were needed to sustain fiscal adjustment. The main focus in this section is on the policy goals and outcomes for the fiscal balance and its major components; structural reform in the government sector and developments in the structure of revenue and expenditure are discussed in depth in Abed and others (1998). The record of fiscal adjustment is assessed in terms of the primary balance—that is, the overall balance excluding grants and net interest payments. This measure captures the fiscal adjustment effort under the authorities’ control and provides an indication of the sustainability of the fiscal position. Other perspectives on fiscal adjustment are relevant for other important objectives of ESAF-supported programs. The overall balance provides an indicator of pressures on the external current account and of the crowding out of private sector activity; hence, data on the overall balance are also reported. The record of adjustment under programs and longer-term developments in the overall balance were broadly parallel with those in the primary balance.1

For two-thirds of the ESAF countries, the fiscal accounts cover only the central government. In the four Western Hemisphere countries, consolidated public sector accounts (including central bank profits and losses) are available.2 In the remaining countries, the operations of the consolidated central government, which integrates some extrabudgetary accounts, are covered. In practice, this implies for some countries the omission from the fiscal accounts of several activities that are known to have been an important part of the public sector: public-enterprises and marketing boards, extrabudgetary funds, and the quasi-fiscal losses of the central bank and state-owned commercial banks. Hence, the focus on a narrow coverage of government accounts may be misleading insofar as activities are shifted between different levels of the public sector. However, in some cases the performance of noncentral government public sector entities were addressed directly in SAF/ESAF-supported programs.

Fiscal Trends, 1981–95

During the first half of the 1980s, before the creation of the SAF in 1986, ESAF countries consolidated their fiscal positions in response to the collapse in the commodity price boom of the late 1970s that had fueled a massive expansion of government spending (Figure 4.1 and Table 4.1).3 These efforts were most pronounced in those countries—typically within non-CFA Africa and the Western Hemisphere—with the largest imbalances.

Figure 4.1Fiscal Trends in ESAF Countries

(Average; in percent of OOP)

Source: IMF staff estimates.

1Excluding grants.

2Excluding Equatorial Guinea.

3Excluding Guyana.

Table 4.1Fiscal Developments in Nontransition Economies(Averages; in percent of GDP)
1981–8511986–901991–931994–95
Overall balance, including grants-8.3-6.0-4.5-4.2
Overall balance, excluding grants-10.6-9.3-8.4-8.0
Primary balance, excluding grants-7.8-6.1-4.9-4.2
Revenue17.517.417.317.6
Noninterest expenditures225.323.422.221.9
Current16.315.514.514.2
Capital8.88.27.97.9
Source: IMF staff estimates.

Excludes Bangladesh, Ghana, and Guinea, for which data on a comparable basis to later years are not available.

Components do not sum to totals because of missing observations for noninterest current and capital spending.

Source: IMF staff estimates.

Excludes Bangladesh, Ghana, and Guinea, for which data on a comparable basis to later years are not available.

Components do not sum to totals because of missing observations for noninterest current and capital spending.

Subsequently, consolidation stalled in the African countries during the late 1980s and early 1990s, their difficulties compounded by adverse terms of trade shifts and faltering growth. In the non-CFA African countries, budget deficits held steady on average, but the record of consolidation was diverse: in Ghana, Malawi, Sierra Leone, and Tanzania, early consolidation was reversed in the early to mid–1990s; in Kenya, Lesotho, and Mauritania, improvements were sustained and, by 1995 if not sooner, primary surpluses were registered; in the remaining countries, there was a steady deterioration in the primary deficit. In most CFA African countries, however, the fiscal position steadily worsened—Benin being the exception. By contrast, Asian and Western Hemisphere countries continued to trim budget deficits during this period. During 1994–95, progress resumed in Africa and continued in most other regions as countries took advantage of an upturn in export demand, growth, and the terms of trade to narrow budget deficits, mainly by containing expenditure in relation to GDP. The CFA countries achieved substantial reductions in expenditure (relative to GDP), especially in the wage bill, by containing nominal wage growth following the devaluation of the CFA franc. Overall, comparing 1995 with 1985, 16 of the 28 nontransition economies had improved their primary balance (excluding grants); for the remainder, the deficit had held steady or worsened.

The improvement in the primary balance during 1985–95 was based, on average, on curtailing non-interest expenditure in relation to GDP. Revenue barely increased. The combination of changes in revenue and expenditure varied across countries. Among the transition economies, in Albania, the Kyrgyz Republic, and Mongolia, the scope of government was scaled back during the first half of the 1990s, with revenue and spending cut sharply from relatively high initial levels. By contrast, in Cambodia, the Lao People’s Democratic Republic, and Vietnam, both revenue and expenditure rose from relatively low levels during 1991–95. In general, the magnitude of changes in the revenue effort or spending was inversely related to initial levels in the mid–1980s; in fact, countries with the highest initial levels of revenue (relative to GDP) registered declines. Overall, there was a convergence over time in both revenue and expenditure in relation to GDP across countries.

Medium-Term Program Objectives

Policies supported by SAF/ESAF resources were set within a three-year program horizon, and hence the objectives (and outcomes) of fiscal policy are examined here in terms of the three-year targets set at the outset of the multiyear arrangement.4 For a variety of data-related reasons, medium-term projections for only about two-thirds of the 68 SAF/ESAF arrangements are considered.5 In the background paper on fiscal issues (Abed and others, 1998), the analysis of program objectives for the level and composition of revenue and expenditure is based on (revised) one-year targets.6 This short-term approach stems from the absence of medium-term projections for the detailed structure of the fiscal accounts in most programs.

The different methodologies used in this study and in Abed and others (1998) offer alternative perspectives on the goals and outcomes of SAF/ESAF-supported programs. In broad terms, the analysis of three-year objectives points to a strategy of significant revenue mobilization and spending restraint that only partially succeeded. From a one-year perspective, annual programs on average sought to increase revenues relative to GDP only slightly and to maintain expenditure ratios broadly unchanged; not surprisingly, countries came closer to meeting these objectives than the original three-year goals. Taken together, the two analyses suggest that annual programs lowered their goals as SAF/ESAF arrangements evolved, incorporating the effects of and responses to past policy slippages or overperformance, unforeseen negative exogenous shocks, and perhaps reassessments of the time needed to accomplish the required underlying fiscal reforms. Over the SAF/ESAF period as a whole, progress fell substantially short of strategic medium-term goals but was more in line with annual program targets.

Primary Balance

SAF/ESAF-supported programs, on average, aimed to cut primary budget deficits (in relation to GDP, excluding grants) by just under half—or 3 percentage points of GDP (Table 4.2). African countries sought a greater-than-average measure of adjustment, including in programs supporting the devaluation of the CFA franc in 1994. On average, the greatest adjustment was targeted in cases where initial deficits were largest (Box 4.1 and Figure 4.2). Although most programs targeted fiscal consolidation over the medium term, 8 (out of 47) SAF/ ESAF-supported programs allowed for a deterioration in the fiscal position. In three of these (Ghana 1987 and 1988 and Honduras 1982), the initial primary balance was in surplus, and the envisaged worsening was slight. In another three cases (Lao P.D.R. 1989, Mozambique 1990, and Niger 1989), the sizable deterioration in the deficit stemmed from a recovery in public investment financed by foreign grants and concessional loans. In Bolivia (1987), the widening of the deficit largely reflected the unwinding of a tighter-than-expected squeeze in public enterprise investment under a prior Stand-By Arrangement. In Togo (1989), a slight increase in the primary deficit was expected despite a targeted cut in spending, but in terms of the overall deficit this was more than offset by a drop in interest payments.

Figure 4.2Fiscal Adjustment in SAF/ESAF-Supported Programs, by Initial Conditions1

(Average; in percent of GDP)

Source: IMF staff estimates.

1Ranked by preprogram levels as estimated at the start of the program.

Table 4.2Fiscal Adjustment in SAF/ESAF-Supported Programs: Targets and Outcomes(Average; in percentage points of GDP)
Overall Balance Including GrantsOverall Balance Excluding GrantsPrimary BalanceRevenuesNoninterest Expenditures
Preprogram year1
All programs-6.1-9.6-6.317.623.9
Nontransition economies-5.4-9.0-5.517.623.1
CFA Africa-5.5-8.4-5.115.420.5
Non-CFA Africa-4.6-10.0-6.720.126.8
Asia-7.8-8.9-5.614.420.0
Western Hemisphere-2.7-6.2-0.521.421.9
Transition economies-16.3-19.2-17.718.436.1
Three-year targets (change from preprogram year)
All programs2.83.33.01.1-1.9
Nontransition economies2.73.03.00.8-2.1
CFA Africa3.53.53.01.0-2.0
Non-CFA Africa3.03.83.70.5-3.2
Asia1.71.72.21.2-1.1
Western Hemisphere-0.40.30.21.00.8
Transition economies5.98.22.94.41.5
Three-year outcomes (change from preprogram year)
All programs1.71.71.60.2-1.4
Nontransition economies1.31.21.10.1-0.9
CFA Africa2.21.61.2-0.1-1.3
Non-CFA Africa1.51.51.60.2-1.4
Asia0.80.80.8--0.9
Western Hemisphere-3.2-2.1-2.20.73.4
Transition economies8.110.010.01.7-8.4
Deviations from targeted changes2
All programs-1.2-1.7-1.3-0.80.5
Nontransition economies-1.3-1.8-1.9-0.71.2
CFA Africa-1.3-1.9-1.8-1.10.7
Non-CFA Africa-1.3-2.0-2.1-0.21.9
Asia-1.1-1.1-1.4-1.10.2
Western Hemisphere-2.8-2.4-2.3-0.32.5
Transition economies1.92.07.1-2.7-9.9
Source: IMF staff estimates.

As estimated at the start of the program.

Actual change minus targeted change.

Source: IMF staff estimates.

As estimated at the start of the program.

Actual change minus targeted change.

The Mix of Revenue and Expenditure Adjustment

The envisaged reliance on expenditure cuts relative to revenue increases to bring about fiscal consolidation varied widely, depending to some degree on initial conditions. Targets for changes in expenditures were far more dispersed than those for changes in revenues, and planned cuts in spending were typically larger than projected increases in revenues (Figure 4.3).7 Although programs, on average, envisaged cuts in noninterest expenditure relative to GDP of 2 percentage points over three years, almost one-third of programs targeted increased spending—primarily in capital outlays—from preprogram levels. The deepest (and most frequent) cuts—averaging more than 2½ percentage points of GDP8—were sought in Africa, where initial spending levels were highest (with the exception of the transition economies). In the other country groups, the targeted fiscal adjustment was to come primarily from higher revenue, with programs in Asia and the Western Hemisphere aiming to raise revenue relative to GDP by about 1 percentage point of GDP over three years. In almost one-half of programs, expenditure cuts combined with revenue increases were expected; in about one-fourth of programs, targets sought to raise both revenue and spending from relatively low levels.

Figure 4.3Fiscal Primary Balance Targets: Changes from Preprogram Year1

(In percentage points of GDP)

Source: IMF staff estimates.

1Restricted sample of 28 countries and 47 SAF/ESAF-supported programs.

In general, the targeted adjustment in revenue and expenditure was inversely related to the preprogram level (see Box 4.1 and Figure 4.2). The range of targets also implies that when revenue was relatively high and expenditure relatively low (as a share of GDP), additional increases in revenue and cuts in expenditure were considered undesirable or infeasible. When initial expenditure in relation to GDP was less than 16 percent (about one-fourth of programs) programs tended to provide for an increase in spending. When the starting level of revenue was at the high end of the sample range—24 percent of GDP or more—programs were more likely to target cuts in revenue. Thus, the distribution of targets across programs implied some convergence of revenue and spending ratios to GDP over rime. Targeted changes in revenue and expenditure, however, do not appear to have been related to income per capita or the initial size of the primary balances. The latter finding implies that the particular mix of revenue and expenditure changes targeted to achieve consolidation was not influenced by the required scale of deficit reduction, but rather by factors specific to the revenue effort and expenditure needs.

Composition of Spending and Revenue

On average, programs aimed, over a three-year horizon, to reorient government noninterest expenditure from current to capital spending through cuts in the former (in relation to GDP) of some 2½ percentage points of GDP and increases in the latter of almost ¾ of 1 percentage point of GDP (Table 4.3).9 About three-fourths of programs envisaged increases in capital expenditure (in relation to GDP), and roughly the same proportion planned cuts in current spending; about one-half planned both to cut current spending and to raise capital expenditures. The planned adjustment in the ratio of current noninterest spending to GDP tended to be inversely related to the preprogram level—programs sought the largest cuts from the highest starting positions.10 Such a relation, however, was not found for capital spending targets, reflecting in part that such targets are influenced considerably by the availability of foreign financing and assessments of the productivity of proposed investments.

Table 4.3Composition of Noninterest Expenditures in SAF/ESAF-Supported Programs: Targets and Outcomes1(Averages; in percentage points of GDP)
Current Noninterest ExpendituresCapital Expenditures
Change from preprogram yearChange from preprogram year
Preprogram levelThree-year targetThree-year outcomeDeviation from target2Preprogram levelThree-year targetThree-year outcomeDeviation from target2
All programs15.2-2.6-1.01.68.30.70.1-0.6
CFA Africa14.1-2.6-1.11.56.40.7-0.2-0.8
Non-CFA Africa15.2-2.7-0.52.210.20.1-0.1-0.2
Asia10.7-1.30.41.79.60.2-1.2-1.4
Western Hemisphere16.6-1.11.42.58.92.03.61.6
Transition economies30.2-6.0-8.7-2.75.74.52.2-2.3
Source: IMF staff estimates.

Based on a sample of 37 programs for which three-year targets and outcomes are available.

Actual changes minus targeted changes.

Source: IMF staff estimates.

Based on a sample of 37 programs for which three-year targets and outcomes are available.

Actual changes minus targeted changes.

Projections for the structure of revenues and spending over a three-year horizon are not available. Abed and others (1998), however, examine one-year-ahead targets and find some common elements in program objectives. On the revenue side, programs aimed to shift the effort from nontax to tax revenues, from direct to indirect taxation (often through the introduction of a general consumption-based tax), and to boost tax revenues from international trade (especially through strengthened customs administration). Within expenditures, about half of the saving from current spending was to come from the government wage hill and half from subsidies and transfers. In broad terms, most programs sought to strengthen health and education spending.

Fiscal Outcomes

On average, half of the targeted reduction in the primary budget deficit was achieved (see Table 4.2). Performance varied widely, however, with almost three-fourths of programs achieving no improvement in the primary balance over three years. While the frequency of shortfalls was broadly similar in the different country groups, undershooting of targets was greatest in Africa and the Western Hemisphere countries. Program targets for revenue mobilization were also frequently missed: revenues fell shore of target in two-thirds of programs, and on average revenues were barely above preprogram levels in relation to GDP (Figure 4.4). With respect to noninterest expenditures, spending exceeded targets in about 40 percent of programs: in those programs that aimed to cut spending (24 out of 47), three-fourths fell short; in those that sought to raise-spending (15 out of 47 programs), targeted levels were surpassed in one-fourth of programs. The propensity for revenue shortfalls and spending overruns was somewhat greater in Africa—affecting about three-fourths of programs—than in other regions. With the exception or transition economies, however, African countries also achieved the largest spending cuts in relation to GDP—just under 1½ percentage points of GDP, or slightly more than half of that targeted.

Figure 4.4Revenues, Expenditures, and the Primary Balance: Deviations from Targeted Changes1

(In percentage points of GDP)

Source: IMF staff estimates.

1Three-year actual change minus three-year target change. Lao P.D.R. (expenditure and revenue shortfalls of 19.7 and 1.8 percent of GDP, respectively) is not shown because of scale considerations.

Box 4.1Program Objectives and Initial Conditions

Fiscal targets are influenced by several considerations such as initial conditions, national preferences regarding the size of government, and the sustainability of the fiscal position. An assessment of the last would depend, among other things, on the sustainable level of concessional financing or grants, the initial ratio of public debt to GDP, and the consistency of the fiscal position with acceptable levels of inflation.

The link between initial positions and targeted changes in the primary balance, revenues, and noninterest expenditures (all expressed relative to GDP) was examined using a simple regression. On the assumption that the targeted level of the fiscal variable would be related to some long-run objective with a partial adjustment toward this objective, targeted changes in the variable were related to preprogram levels and a constant. The model tested is

ΔT = αT(-1) + β TRANS + γ,

where ΔT is the three-year target change in the fiscal variable, T(-1) is the preprogram level of the variable, and TRANS is a dummy for transition economies (dummy terms for the other country-groups were found to be statistically insignificant).

The regression results (see table) indicate the following:

  • The model has low explanatory power for the targeted adjustment in the primary balance (excluding grants) but explains 30–40 percent of the variation in revenue and expenditure targets.
  • The targeted adjustment for revenue, spending, and the primary balance is inversely related to the starting position; this relationship is statistically significant at the 99 percent confidence level for revenue and spending, and at the 95 percent level for the primary balance.
  • Neither income per capita (measured in terms of U.S. dollar GDP) nor the initial level of the primary balance bad a statistically significant influence on the targeted adjustment in revenue and spending, when included in the estimated regression.
Regression Results
Targeted ChangesPreprogram Level

(α)
Transition Dummy

(β)
Constant

(γ)
R2
Primary balance-0.2-2.91.70.1
(2.1)*(1.1)(1.9)
Revenue-0.12.83.10.3
(3.1)**(3.2)**(3.9)**
Noninterest expenditure-0.37.54.80.4
(5.0)**(3.6)**(3.3)**
Source: IMF staff estimates.Note: t-Statistics are indicated in parentheses under each estimated coefficient; ** indicates statistical significance at the 1 percent level and * at the 5 percent level, and R2 is the coefficient of determination.
Source: IMF staff estimates.Note: t-Statistics are indicated in parentheses under each estimated coefficient; ** indicates statistical significance at the 1 percent level and * at the 5 percent level, and R2 is the coefficient of determination.

What distinguished programs in which the primary balance improved from those in which it worsened? The likelihood of deficit overruns depended primarily on whether expenditure exceeded its targeted level;11 also the size of deficit overruns was largely influenced by the performance of expenditure (see Figure 4.4).12 In practice, in the face of expenditure overruns, the scope for raising revenues to contain the deficit was limited. By contrast, there was in general no significant correlation between revenue shortfalls and deficit overruns. This reflected that, in half of the programs in which revenue was lower than expected, expenditure was cut hack from targeted levels. The brunt of this offsetting adjustment in expenditure was borne by capital spending, which was below target in three-fourths of programs that experienced both revenue and expenditure shortfalls.

The experience with fiscal adjustment under programs suggests that the degree of success in meeting program targets for the primary balance and noninterest expenditure was inversely related to the size of the adjustment sought; for revenue, only a weak correlation was found. However, differences across programs in the size of the targeted adjustment “explain” only a small fraction of the deviations from targets (Figure 4.5).13 The analysis of one-year-ahead objectives and outturns for the components of revenue and expenditure (see Abed and others, 1998) points to a number of other contributory factors, including the sustainability or permanence of adjustment measures (for example, in efforts to scale back the civil service wage bill).

Figure 4.5Fiscal Targets and Outcomes1

(In percentage points of GDP)

Source: IMF staff estimates.

1Dashed lines indicate the relationship between targeted changes and outcomes estimated by regressing the targeted change on the deviation of outcomes from targets. This is not shown for revenues because a statistically significant relationship was not found.

Was the degree to which medium-term targets mere met related to the phasing of adjustment within the three-year program horizon? On average, targets for deficit reduction and expenditure cuts were not front-loaded.14 About one-third of the targeted adjustment in these variables was planned for the first program year. For revenue, about one-half of the targeted effort was programmed for the first year. After the event, the degree by which targets were missed was not related, of itself, to the amount by which the programmed adjustment was front-loaded.15 This would support the view that success in meeting fiscal targets was influenced more by the quality and strength of policy measures than by the size of the adjustment sought.

On average, the targeted reorientation of non-interest expenditure from current to capital outlays did occur, although to a lesser extent than envisaged under programs. Overruns in current expenditures were widespread—occurring in almost three-fourths of programs—while capital expenditure plans were undershot in almost two-thirds of programs. Although, on average, there was a shift in the mix of spending in favor of capital spending, this was based primarily on cuts in current spending (relative to GDP) rather than on higher capital outlays, which were virtually unchanged in relation to GDP (see Tables 4.1 and 4.3). Progress toward the desired rcstrucruring of expenditure was particularly limited in non-CFA Africa and Asia.

The shortfall in capital expenditure from programmed levels reflected a number of factors, including cutbacks in the face of shortfalls in revenues or overruns in current outlays, as well as lower-than-expected levels of concessional financing and grants destined for government investment. In most programs in which capital outlays were undershot, a combination of these factors was at play. Of the 24 programs in which capital outlays fell short of planned levels, over two-thirds experienced revenue shortfalls—in just over half of these cases current expenditures exceeded program targets—and in 40 percent of the programs official transfers were smaller than projected.

As regards the composition of expenditures and revenues, the following observations can be made, based on the analysis of one-year targets in Abed and others (1998). For expenditure, governments often failed to scale back the wage bill as intended, while subsidy costs also tended to exceed targets; modest increases in the share of health and education spending in GDP were achieved, on average; and declines in military spending (in relation to GDP) were recorded in many countries. For revenue, targets for direct and indirect taxes were achieved on average, but the intention to raise tax revenues from foreign trade was only partially realized.

Was Fiscal Consolidation Adequate?

Although fiscal targets under SAF/ESAF-sup-ported programs in nontransition countries were on average undershot, progress was made during programs—and, more broadly, over the period 1985–95—in moving budget deficits toward a more sustainable position (Table 4.4).16 But progress was not universal under programs—deficits widened in 12 of the 27 countries during the course of SAF/ESAF-supported programs. Outside the CFA, ESAF countries on average cut their budget deficits both during program years and outside of program periods, implying a sustained policy effort aimed at fiscal consolidation. However, in the CFA countries, the reduction in the primary deficit accomplished during program years was undone by slippages in other years in about half of the countries; on average, during 1985–95 the deficit widened in these countries.

Table 4.4Budget Consolidation(Averages; in percent of GDP)
Primary Balance

(1985)
Change in Balance

(1985–95)
Cumulative Change in Balance Under SAF/ESAF ProgramsNumber of Program YearsPrimary Balance

(1995)
Nontransition economies1,2-5.31.92.65.8-4.2
CFA Africa 1,3-2.8-1.42.85.6-4.2
Non-CFA Africa-6.32.30.95.8-4.0
Asia-6.63.22.16.0-3.4
Western Hemisphere2-3.53.40.27.0-0.1
Source: IMF staff estimates.

Excluding Equatorial Guinea, Guyana, and Nicaragua.

Excluding Nicaragua and Guyana. Initial year for Bolivia is 1986. Figures for Western Hemisphere would be distorted by large changes for Nicaragua; including this country gives a change in the balance of 6.8 percent of GDP (under programs, –0.9 percent of GDP).

If Côte d’Ivoire, which had exceptionally high revenues in 1985, were excluded, the average balance in 1985 would be –5.2 percent of GDP.

Source: IMF staff estimates.

Excluding Equatorial Guinea, Guyana, and Nicaragua.

Excluding Nicaragua and Guyana. Initial year for Bolivia is 1986. Figures for Western Hemisphere would be distorted by large changes for Nicaragua; including this country gives a change in the balance of 6.8 percent of GDP (under programs, –0.9 percent of GDP).

If Côte d’Ivoire, which had exceptionally high revenues in 1985, were excluded, the average balance in 1985 would be –5.2 percent of GDP.

Assessments of what constitutes a sustainable fiscal position are fraught with difficulty: there is no unique standard for fiscal sustainability; data are inadequate; and any specific assessment is subject to change as economic conditions evolve. However, without such assessments it is difficult to ascertain how much fiscal adjustment is needed, and where along the path toward sustainability a country stands. In what follows, one relatively simple but partial measure of sustainability that is based on debt dynamics (Box 4.2) and can be calculated using readily available data is used to shed some light on this issue.

Specifically, an attempt is made to assess whether, on the basis of policies during 1993–95, the debt-to-GDP ratio would be rising or falling to a stable long-run level.17 A comparison of the actual debt ratio to the long-run debt ratio would indicate whether a country was on a path toward lower indebtedness and increased sustainability, on the assumption that lenders would continue financing the deficit during the transition.

The debt-dynamics calculations indicate that, on the basis of economic conditions, policies, and the level of external aid during 1993–95, countries were, on average, on a path of declining debt-to-GDP ratios (Table 4.5).18 The debt dynamics indicate that after three years the average debt-to-GDP ratio would have fallen by 17 percentage points of GDP, to 118 percent of GDP (with a saving in interest payments of about ½ of 1 percentage point of GDP). However, there was considerable variation across countries, with 20 of the 27 countries on a path toward lower levels of indebtedness relative to GDP.19 Of the remaining seven countries, Burundi, The Gambia, and Sierra Leone recorded negative real GDP growth during 1993–95 and hence faced an explosive rising path of the debt ratio. A return to positive growth rates would, however, stabilize the debt dynamics and move these countries to lower debt ratios in the long run. The debt dynamics also point to very high long-run debt ratios for Kenya, Sri Lanka, and Zimbabwe because the effective interest rate on debt is significantly higher than the real GDP growth rate in these countries. For Bangladesh, a slight rise in the debt ratio was projected.

Table 4.5Fiscal Debt Dynamics1(Averages; in percent of GDP unless otherwise noted)
1993–952Projections
Primary balance

(excluding grants)
Real GDP growth

(in percent)
Inflation3

(in percent)
Base moneySeigniorage revenue4Official grantsPrimary balance less seigniorage less grantsPublic debt ratio

(1995)
Average interest rate

(in percent)
Debt ratio after 3 yearsDebt ratio after 10 yearsLong-run debt ratio
Nontransition economies-4.04.215.310.31.23.60.81353.811482355
CFA Africa-3.75.511.512.61.33.71.31023.28044-24
Non-CFA Africa-4.73.618.79.71.14.71.11343.611486-65
Asia-3.25.28.812.11.30.7-1.2575.755522466
Western Hemisphere-4.03.114.56.71.33.20.53152.826417810
Sensitivity analysis7
Baseline-4.04.215.310.31.43.60.91353.811886-27
Lower primary deficit (1 percentage point of GDP)-3.04.215.310.31.43.61.91353.811577-55
Higher primary deficit (1 percentage point of GDP)-5.04.215.310.31.43.6-0.11353.8121942
1 percent higher real GDP growth-4.05.215.310.31.43.60.91353.811476-20
1 percent lower real GDP growth-4.03.215.310.31.43.60.91353.812297-40
5 percent higher inflation-4.04.220.310.31.33.60.91353.811782-23
5 percent lower inflation-4.04.210.310.31.33.60.91353.812090-29
1 percent a year real exchange rate depreciation (for 10 years)8-4.04.215.310.31.33.60.91353.812297-27
Source: IMF staff estimates.

The foreign inflation rate used in the debt-dynamics calculation is the weighted-average inflation rate for 1993–95 (2 percent) in the countries included in the SDR basket, with weights corresponding to their currency shares in the basket.

1995 for CFA Africa.

Measured in terms of the GDP deflator.

Calculated using a quadratic function linking seigniorage to inflation that was estimated by Easterly, Rodriguez, and Schmidt-Hebbel (1994). Baseline values differ from the average of nontransition economies because baseline seigniorage is calculated from a quadratic function using the average value of inflation.

Excludes Burundi, The Gambia, Kenya, Sierra Leone, and Zimbabwe, for which the long-run debt ratio would rise indefinitely under 1993–95 fiscal policies and economic conditions.

This average long-run debt ratio is distorted by Sri Lanka; when excluded, the debt ratio would be 20 percent of GDP.

The analysis considers only the partial effects of the change in policies or economic conditions on the debt dynamics; for example, growth, inflation, and grants are held constant in the case of a real exchange rate depreciation.

Assumes the real exchange rate depreciates at a rate of 1 percent a year for the first ten years and thereafter is constant.

Source: IMF staff estimates.

The foreign inflation rate used in the debt-dynamics calculation is the weighted-average inflation rate for 1993–95 (2 percent) in the countries included in the SDR basket, with weights corresponding to their currency shares in the basket.

1995 for CFA Africa.

Measured in terms of the GDP deflator.

Calculated using a quadratic function linking seigniorage to inflation that was estimated by Easterly, Rodriguez, and Schmidt-Hebbel (1994). Baseline values differ from the average of nontransition economies because baseline seigniorage is calculated from a quadratic function using the average value of inflation.

Excludes Burundi, The Gambia, Kenya, Sierra Leone, and Zimbabwe, for which the long-run debt ratio would rise indefinitely under 1993–95 fiscal policies and economic conditions.

This average long-run debt ratio is distorted by Sri Lanka; when excluded, the debt ratio would be 20 percent of GDP.

The analysis considers only the partial effects of the change in policies or economic conditions on the debt dynamics; for example, growth, inflation, and grants are held constant in the case of a real exchange rate depreciation.

Assumes the real exchange rate depreciates at a rate of 1 percent a year for the first ten years and thereafter is constant.

The debt-dynamics indicator of sustainability provides only one of several possible perspectives on the soundness of public finances. In particular, a full assessment of the sustainability of the fiscal position would need to take account of the willingness of lenders to continue to finance the deficit. Also, the measure of sustainability considered here does not address whether the fiscal position is consistent with other macro-economic objectives, including those for national saving, inflation, and the external current account; the permanence and quality of fiscal measures used to achieve a balance deemed sustainable is also important. When these other considerations are taken into account, fiscal adjustments cannot be seen as complete in most of the ESAF countries, even in those that had primary surpluses in the mid–1990s and prospects for improving public debt positions. In these countries, saving rates are very low, inflation is frequently in the double digits (in 12 of the 21 non-CFA countries in the debt-dynamics exercise), there was substantial reliance on exceptional financing, and official budgetary grants—for which prospects are inevitably uncertain—were sizable.

Box 4.2Fiscal Debt Dynamics

Assessing the sustainability of the fiscal position is a complex and judgmental exercise that depends on many variables; see Easterly, Rodriguez, and Schmidt-Hehbel (1994) for a detailed discussion. There is no unique standard of fiscal sustainability. Here, a simple indicator based on the debt dynamics of the budget is described.

Fiscal deficits are financed from four sources: seigniorage revenues comprising the inflation tax on the monetary base; seigniorage that accrues from an increase in real money balances; increases in the stock of domestic and external debt; and external grants. The intertemporal budget constraint is

where pd is the primary deficit excluding grants; m is base money; d is public domestic debt; d* is public external debt; u is grants (all expressed as shares of GDP); π and π* are rates of domestic and foreign inflation, respectively (in terms of the GDP deflator); g is the real GDP growth rate; r and r* are the average real rate of interest on domestic and external debt, respectively; and ε is the rate of depreciation of the real exchange rate. Overdots indicate changes.

Sustainability can be assessed in two ways: in terms of the primary balance needed to keep the public debt-to-GDP ratio constant, and in terms of the long-run public debt-to-GDP ratio that would he reached for a particular level of the primary deficit relative to GDP. The sustainable primary balance that will hold the public debt-co-GDP ratio constant is

where s=m˙+(π+g)m is the seigniorage revenue. The long-run debt-to-GDP ratio (td = d + d*) implied by a given primary balance is

where δ* is the share of external debt in total public debt, ρ = (r + π) (1 – δ*) + (r* + π*)δ* is the average nominal interest rate on total public debt, and e = ε + π – π* is the rate of nominal exchange rate depreciation.

The numerator is the primary balance including seigniorage revenue and grants and indicates whether noninterest operations are adding to the debt ratio. The denominator gives the growth of the debt ratio needed to finance debt service even if the primary balance (after seigniorage and grants) is zero. It the rate of interest and the exchange-rate-induced increase in the domestic currency value of external debt is higher than the nominal rate of growth, the debt ratio will be rising.

Other things being equal, the long-run debt-to-GDP ratio is higher the slower is the rate of real GDP growth (the rate at which resources are generated to repay the debt principal and interest), the higher is the interest rate, and the larger is the depreciation of the nominal exchange rate (which increases the domestic currency value of external debt). The impact of domestic inflation is more complex. In a situation where inflation rates remain moderate, a rise in inflation will reduce the long-run debt ratio because it will increase the inflation tax on money balances and erode the value of domestic debt. When inflation rises above a certain level, however, the inflation tax will decline because the base money ratio is endogenous to the rate of inflation, and the tax base will shrink; at high rates of inflation, seigniorage revenue will become negative.

The calculation of the long-run debt-to-GDP ratio reported in Table 4.5 assumes that the share of external debt, the average interest rate on (and implicitly the degree of concessionality of) debt, the share of grants in GDP, and the real exchange rate are expected to be constant. To obtain an estimate of seigniorage revenues in which money demand has fully adjusted to the rate of inflation, seigniorage is calculated using a quadratic relationship between inflation and seigniorage estimated from a cross-country regression by Easterly, Rodriguez, and Schmidt-Hebbel (1994).

A comparison of the maximum and actual debt ratio shows whether an unchanged primary balance would reduce or increase indebtedness relative to GDP. The fiscal position is sustainable only if creditors are willing to finance it.

Two other more technical caveats also apply. First, the calculations assume a constant real exchange rate, whereas in most ESAF countries a depreciation of the real exchange rate is more typical; indeed, during 1985–95 the ESAF countries experienced, on average, a real effective exchange rate depreciation of just over 4 percent a year. In the debt-dynamics calculations presented here, a real exchange rate depreciation of 1 percent a year for ten years would, other things being equal, raise the long-run debt ratio by about 10 percentage points of GDP, evaluated relative to a baseline of the “average” ESAF country.20 Second, in the past there have been sizable unanticipated additions to the government’s debt stock—for example, because of the assumption of (government-guaranteed) debts of insolvent public enterprises, balance sheet restructuring of distressed banks, and compensation to central banks for quasi-fiscal losses.21 These may occur in the future in some ESAF countries, and they are not factored into the debt-dynamics calculations.

To illustrate this point, a similar debt-dynamics exercise based on policies and economic conditions in 1983–85 would have projected a decline, on average, in debt ratios between 1985 to 1995. By contrast, the debt ratio actually rose, from 80 percent to 135 percent of GDP A substantial part of this increase would appear to be attributable to real exchange rate effects and quasi-fiscal additions to government debt. More generally, economic conditions and policies are likely to change, and Table 4.5 provides some analysis of the sensitivity of the projected debt ratios to deviations in primary balances, real GDP growth, inflation, and the real exchange rate from baseline projections.

Exchange Rate and Monetary Policies

Exchange Rate Policies

Exchange rate arrangements varied across the countries under review and evolved over time. All of the CFA countries had an exchange rate peg that provided a nominal anchor throughout the period. Elsewhere, use of exchange rate anchors was both less prevalent and declining over time (Figure 4.6). Thus, until 1990 about three-fourths of the nontransition countries employed an exchange rate peg, but by 1995 this share had fallen to 40 percent, with the majority being CFA countries.22

Figure 4.6Formal Classification of Exchange Rate Arrangements1

Source: IMF, International Financial Statistics (IFS); classification at year end, as reported to the IMF by member countries.

1Excluding transition economies.

According to the formal classification system used in countries’ reports to the IMF, most SAF/ESAF users giving up exchange rate pegs during the period under review moved to an “independently floating” system; cases in the “managed floating” category remained a small minority. Experience shows, however, that exchange rates are often extensively managed even in the former group.23 Indeed, with the move away from explicit pegging, actual exchange rate practices have become more difficult to characterize precisely, since more flexible regimes can encompass a wide variety of policy rules and discretion.24

Exchange rates were explicitly used as nominal anchors in about one-sixth of the 35 SAF/ESAF-supported programs that aimed to bring inflation down from double-digit levels.25 By contrast, in one-fifth of these programs it was noted that nominal exchange rate policy would pursue a real exchange rate target. Such a policy of adjusting the exchange rate in response to inflation developments may reflect a strong priority on objectives for the current account, but it also may make disinflation particularly difficult (Chapter 6 examines evidence on this point).

Exchange Rate Developments

Many three-year SAF/ESAF-supported programs began with a sizable depreciation of the nominal effective exchange rate just before or during their first year: this is true of the three programs in CFA countries begun in 1994, as well as programs in Bangladesh (1986), Burundi (1986), The Gambia (1986), Ghana (1987), Lao P.D.R. (1989), Madagascar (1987), Mozambique (1987) and (1990), Zimbabwe (1992), and most of the high-inflation cases. As regards the real effective exchange rate (REER), there was a widespread tendency (apparent in about 85 percent of all programs) for a cumulative real depreciation over the three-year program period.

Indeed, throughout the period under review, measured REERs followed a depreciating trend (Figure 4.7).26 For many countries, the steep declines apparent in such measures in the mid- to late 1980s largely reflected the abandonment of much overvalued official exchange rates, presumably bringing the measured REER closer to its equilibrium value. Subsequently, REERs have been more stable, showing a more gradual depreciation. In part this trend may be related to the terms of trade deterioration affecting many countries during the second half of the 1980s and early 1990s (indicated in Figure 4.7) and the implied depreciation of equilibrium REERs. After 1993, there appears to have been a strengthening of incentives for the export sectors in Africa and the Western Flemisphere, with the devaluation of the CFA franc, stable or depreciating real exchange rates elsewhere, and more generally a sharp upturn in the terms of trade.

Figure 4.7Real Effective Exchange Rate and Terms of Trade

(Average; 1985 = 100)

Sources: IMF, Information Notice System; World Economic Outlook; and IMF staff estimates.

Money, Credit, and Interest Rates

With the exception of the low-inflation CFA cases, most SAF/ESAF-supported programs projected a deceleration of money growth aimed at bringing inflation down to single-digit levels over a three-year period. The operational design of these programs focused on limiting the contribution of domestic credit or net domestic assets to money growth. During many programs (more than half of the non-CFA programs), countries did indeed slow the pace of credit expansion, relative to preprogram experience, although often by less than envisaged. Frequenrly, however, monetary growth was sustained by an accelerated growth of net foreign assets (Figure 4.8). Those countries that experienced high inflation—mainly in the Western Hemisphere and transition economy groups—eventually reduced both money growth and inflation sharply. Elsewhere, growth of broad money and inflation remained broadly unchanged on average during the late 1980s and the first half of the 1990s. These developments, and the resulting trends in inflation during programs and over time, are discussed in depth in Chapter 6.

Figure 4.8Broad Money Growth

(Median; in percent a year)

Source: IMF staff estimates.

SAF/ESAF-supported programs generally aimed to bring about positive real interest rates, partly for monetary policy reasons (to help curtail demand for credit) and partly for structural reasons (to promote financial intermediation and improve resource allocation—see the next section). The record in this respect was mixed, particularly with respect to rates on deposits (Figure 4.9), Some progress was made, notably in reducing the incidence of extremely negative (less than—10 percent) real rates. However, on average, during the most recent five-year period, many countries still had negative rates.27 Performance varied considerably by region. Throughout the 1986–95 period, real deposit interest rates in the CFA countries tended to be strongly positive.28 In general, the non-CFA African and Western Hemisphere countries started out with the most distorted (highly negative) real interest rates, but countries in both these groups had broadly succeeded in establishing positive real rates by the mid–1990s (Figure 4.10). In contrast, the Asian economies—with their lower inflation—have avoided extremely negative real deposit rates but have shown little trend change over time: performance has remained split between Bangladesh and Sri Lanka (usually positive rates) and Nepal and Pakistan (usually negative rates).

Figure 4.9Distribution of Real Deposit Interest Rates1

Source: IMF staff estimates.

1Ex post real interest rates calculated using contemporaneous end-period consumer price inflation, where available. Transition economies are excluded because earlier data tend not to be available.

2Represents number of cases below –20 percent.

3Represents number of cases above 10 percent.

4Included in the negative observations for the 1991–95 period are five CFA African countries that typically had positive real rates. Data shown here reflect impact of 1994 CFAF devaluation.

Figure 4.10Real Deposit Interest Rates, Before First SAF/ESAF-Supported Program and in 1993–95

(Median; in percent a year; t denotes first year of program)

Sources: IMF, IFS; and IMF staff estimates.

One of several considerations motivating fiscal consolidation in ESAF countries has been to reduce the government’s claim on domestic credit, making a larger share available to the private sector. The record indicates progress on this front for many countries.29 The share of the public sector in total domestic credit extended by the banking system was on average on a persistent downward trend in both the non-CFA African and Asian countries, from around onehalf in the early 1980s to less than 30 percent in the mid–1990s.30 The experience in the Western Hemisphere countries was more volatile, but the public sector’s share did tend to decline through the early 1990s—although more recently it has increased in most of these countries. In the CFA countries, the share of credit to the government was relatively low, averaging only about one-fifth through 1991, reflecting statutory limits. In several of these countries, however, this share rose considerably in the 1990s, reflecting the assumption by the government of loans previously classified as private, in the context of bank restructuring.

Structural Reforms

The extensive and severe structural weaknesses of the ESAF countries necessitated broadly based reforms. Programs were intended to reduce the government’s intervention in the allocation of resources, make the economy more outward-oriented, and strengthen the finances of the public sector. The World Bank contributed in important ways to the policy advice, technical assistance, and financial support and played a leading role in those areas of reform under its purview. In this section, the goals, strategies, and extent of progress in five principal areas of structural reform—domestic markets and prices, exchange system, external trade, financial sector, and public enterprises—are examined for the 30 nontransition countries covered by the review over the period 1981–95.31 There is also some discussion of progress with regard to civil service reform and property rights. The analysis updates and expands the survey of structural reforms in the last ESAF review (Schadler and others, 1993), which covered developments in 19 countries up to 1991.

This section also tries to put the assessment of progress across countries on a more systematic and comparable basis through the use of structural reform indices, specially constructed for this study.32 The indices combine various indicators of structural policy and provide a snapshot of the scope of structural distortions during three five-year periods (1981–85, 1986–90, and 1991–95); see Box 4.3. Broadly speaking, the 1981–85 period corresponds to the pre-SAF/ESAF period for most nontransition countries, and the two subsequent periods can be considered the adjustment phase.33

Box 4.3Indices of Structural Reform

The five indices of structural reform discussed in this section of the chapter are each a composite of several specific criteria. The indices use a six-point scale along which the extent of reform can be classified as low (a score of 1–2), moderate (3–4), or high (5–6) relative to a specified notion of “best practices,” or to an average for all developing countries. The indices are based on the following:

  • Pricing and marketing: the extent of overall price controls; the extent of price controls and state intervention in the import and marketing of petroleum products; and the extent of price controls and state intervention in agricultural products—fertilizers, major domestic foodstuffs, and exported agricultural commodities.
  • Exchange system: the level of the premium in the parallel exchange market, and the extent of surrender requirements and nonmarket foreign exchange allocation.
  • Trade system: the coverage of quantitative restrictions, the level and dispersion of tariffs, and the extent of import exemptions.
  • Financial sector: the presence or otherwise of controls on credit or interest rates, real interest rates, and interest rate spreads that are above or below certain thresholds; quantitative measures of financial intermediation (ratio of currency to broad money), financial deepening (ratio of broad money to GDP), and private sector access to credit (ratio of private sector credit to GDP); and the existence or otherwise of a diversified banking system and markets for interbank funds, government, or central bank securities and stocks.
  • Public enterprise sector: measures of the public enterprises’ role in economic activity—share in total output, total employment, and total bank credit; and their claim on financial resources—their overall financial balance and net financial flows from government.

The detailed methodology for constructing these indices is explained in Appendix 4.1.

The last ESAF review concluded that, over the first six years of SAF/ESAF-supported programs, progress in structural reform had been significant but uneven. Important gains had been achieved in exchange and trade liberalization, price liberalization and market reform, and the freeing of interest rates; less progress was made in other areas, notably in restructuring and reforming the public enterprise sector and banking system. Four years on, the overall assessment of structural reforms holds true. On a positive note, reforms continued to move ahead across a broad front in all countries to varying degree. Thus, although the process has been protracted and faltering at times, all the countries now have economies that are significantly more flexible, and guided more by market forces, than they were ten years ago and prior to their SAF/ESAF arrangements. Also, there are signs that in broad terms the pace of reform has picked up in the 1990s, notably in Africa. But reforms affecting public enterprises and bank soundness have continued to lag behind other areas. Chapter 8 takes a closer look at why this was, and how the IMF’s contribution to the public enterprise and banking reform process could be strengthened. There also remains, in other areas such as trade liberalization and strengthening of property rights, considerable need for further reform.

Structural reforms were implemented with varying degrees of aggressiveness and success. In general, the push for reforms co price controls and marketing and to the exchange and trade system came early in the SAF/ESAF period and spawned major successes early on; since then progress has continued. Thus, it is in these areas that the most progress has been achieved (Figure 4.11). In the early 1980s, controls on prices and markets were extensive in most countries, but by 1995 widespread price controls had been eliminated, and market intervention had been substantially reduced. Similar rapid progress was achieved in eliminating distortions in the exchange system. In the trade system, quantitative restrictions were largely removed by 1995, and a substantial measure of success in rationalizing and reducing tariff rates was achieved. In general, the administrative capacity needed to make these changes was less than in other areas of reform. Also, for prices and the exchange and trade system, reforms tended to be met with less resistance, especially when a large share of transactions were already made in parallel markets. However, reform in these areas is not complete. Several elements still need to be addressed, including the elimination (or replacement with targeted subsidies) of remaining selected price controls (notably on staple foods); a further improvement in the structure of trade tariffs, especially reduction and consolidation into tariffs of various duties and charges targeted at imports; and further reductions in the level of external tariff rates.

Figure 4.11Status of Structural Reform in ESAF Countries

(Indices of structural reform; six-point scale)1

Source: Appendix 4.1.

1Higher values indicate better structural policies or fewer distortions.

Financial sector reforms played an important role in virtually all SAF/ESAF-supported programs and were aimed at the complementary goals of enhancing the role of market forces in the allocation of credit, shifting to indirect instruments of monetary control, and strengthening banking system soundness. A good deal of progress was achieved in strengthening market mechanisms in the determination of interest rates and the allocation of credit and in introducing indirect instruments of monetary control. Here, the process of reform began gradually but has picked up pace since the late 1980s (Figure 4.12). Although progress in these areas has improved the efficiency of financial intermediation, it has not, in general, yielded any substantial increase in financial deepening and monetization of the economy. These financial processes are influenced by a wide range of factors—such as the inadequacy of property rights for loan collateral—in which complementary improvements were not always forthcoming. The slow pace of progress in restoring, on a sustainable basis, a financially healthy banking system in many ESAF countries also likely contributed to the limited increase in the depth of financial intermediation.

Figure 4.12Pace of Structural Reform

(Changes in indices of structural reform)1

Source: Appendix 4.1.

1Increase in indices denotes improvement in structural policies or decline in distortions.

The widespread need for restructuring banking systems became increasingly apparent during the second half of the 1980s, and multiyear reform programs were implemented in most countries. However, progress was uneven and slow. In many countries, the initial steps needed for reform—the passage of legislation and restructuring of balance sheets—were taken, but there was only limited progress in ensuring that banking activities are conducted on a commercially sound and sustainable basis. In some instances, initial improvements in balance sheets were reversed. A key factor contributing to the protracted pace of bank restructuring was the equally slow speed of public enterprise reform. The health of the banking sector ultimately mirrors the vigor of the real sector, and in the ESAF countries the public enterprise sector was a major client of banks.

Public enterprise reform commenced early in the SAF/ESAF period, but the reform process was protracted and subject to many slippages. Overall, the improvement was modest, with a reduction in the size and financial burden of public enterprises (based on the limited data available) in about half of the ESAF countries. Progress was notably limited with respect to the enterprises of greatest economic significance. This reflected a lack of success with early efforts to reform large strategic enterprises, where the emphasis was on restructuring rather than privatization. Performance contracts did little to change the behavior of managers, and the enforcement of hard budget constraints proved difficult, thereby reducing the pressure on public enterprises to reform. However, since the early 1990s, reform efforts for larger strategic public enterprises have been refocused toward privatization.

Many factors contributed to the differing success and pace of reforms, but a few key recurring elements stand out from this review. First, the strength of political commitment to reform was critical. In particular in public enterprise reform and bank restructuring, governments found it difficult in practice to break from past views about the need to steer the development process, and they were subject to strong pressures from interest groups benefiting from the status quo. Second, the administrative capacity to implement and to sustain the reform process was extremely weak, notably for financial sector and public enterprise reform. Furthermore, in some specialized sectors such as banking, there was a shortage of private sector skills. In an effort to build countries’ administrative capacity, the IMF, World Bank, and other bilateral and multilateral agencies provided extensive technical assistance.

The process of reform and the economic activities it put in motion also revealed other areas deserving greater attention: improved public administration and civil service reform, and the need for well-defined property rights and a well-functioning judicial system. Also, structural reforms had an important social impact. On the positive side, reforms contributed to reducing poverty through increasing production and employment and redistributing income to the poorer rural sector. However, reforms also involved short-term costs to the poor and other segments of the population. In some areas, such as civil service reform and the restructuring or closure of public enterprises, concerns over their social impact led to delays in their implementation. To mitigate the social costs of adjustment, social safety nets were increasingly integrated in SAF/ESAF-supported programs.34

Domestic Markets and Prices

In the first half of the 1980s, official controls on prices and markets were widespread in most countries. In all but seven countries there were extensive controls on prices; in the remainder only prices of selected items considered to be of strategic importance were administered. The marketing of many strategic items was typically controlled by state trading monopolies (imports of cement, fertilizer, petroleum products, and foodstuffs), agricultural marketing boards, and public enterprises (transportation, telecommunications, and utilities).

The distortionary impact of price controls was influenced by whether they were designed to subsidize production or consumption, shield domestic producers and consumers from fluctuations in international prices, or raise revenues. Their adverse impact was perhaps most evident in the agricultural sector and markets for foodstuffs. In several countries, however, parallel markets and widespread evasion (for example in Bolivia and Guinea) kept prices close to market levels despite official controls.

In some countries—Lesotho, and to a lesser extent Bolivia and Guinea—there were relatively few controls at the start of their first SAF/ESAF arrangements, reflecting previous reform efforts in the early and mid–1980s. In most, however, deregulation was begun at an early stage of the SAF/ESAF period and was continued thereafter. During 1986–95, virtually all countries dismantled the extensive network of controls on marketing and prices. By 1995, widespread controls were a relic of the past: they applied to a relatively small range of products in almost half of the countries, and to a somewhat broader group of goods in the remainder.

Extensive price controls on goods, except utilities and petroleum products (see below), were in place in over two-thirds of the countries in the early 1980s but were lifted early in the SAF/ESAF period (Figure 4.13). By 1990, pervasive controls existed in only nine countries,35 and by 1995 they were left only in Equatorial Guinea. By 1995, controls had been abolished in seven countries, and in the remainder general controls on prices and profit margins had been replaced by selective controls. Typically, these covered a few (but important) staple foodstuffs, transportation, and cement or steel products.

Figure 4.13Structural Reform in Pricing and Marketing

(Indices of structural reform; six-point scale)

Source: Appendix 4.1.

1Higher values indicate better structural adjustment policies or fewer distortions.

2Increase in indices denotes improvement in structural policies or decline in distortions.

The withdrawal of government intervention in petroleum products and utilities was less rapid, particularly for utilities. Only in Bolivia is the marketing of utilities fully privatized, with prices subject to regulatory approval; most other countries have worked with the World Bank to establish prices that are set on a cost-recovery or long-run marginal cost basis. Somewhat more progress was made in petroleum products. Rigid administered prices were replaced during 1986–95 in most countries (except Equatorial Guinea, Nicaragua, and Tanzania) by mechanisms linking domestic prices to changes in international prices and exchange rates, or (in six African countries) by freely market-determined prices. Extensive state-controlled marketing of petroleum products at the wholesale or retail level had been less widespread in the early 1980s—applied in only half of the countries—and was dismantled in all but seven countries by 1995. In all, six countries had fully decontrolled both prices and markets for petroleum products by 1995; in a further six, full price decontrol was planned in the near future.

For many countries the most important reforms of marketing systems and prices affected agricultural products: fertilizer inputs, export commodities, and domestic foodstuffs. Intervention in these areas was present in all countries and was often highly restrictive. The dismantling of restrictions was tackled early in the SAF/ESAF period, and a major degree of success was achieved. Inefficient distribution systems and subsidies for fertilizers were prevalent in the early 1980s, except in Bolivia, Guinea, Guyana, and Lesotho. The removal of subsidies proceeded only gradually and was accompanied by measures to give the private sector a greater role in the import and distribution of fertilizers. By 1995, 17 countries had lifted subsidies and marketing controls on fertilizers.

The reform of the marketing of export commodities and domestic foodstuffs (imported and domestically produced) sought to eliminate the monopoly of state marketing boards and trading monopolies in the purchase, transportation, processing, and marketing of these items; to allow market forces to determine producer prices; and either to improve the financial position of marketing boards and monopolies or to abolish them. In the early 1980s, state-controlled producer prices and heavy state involvement in exporting were present in almost half of the countries—only in Niger was a liberal system in place. By 1995, however, following a pickup in reforms after 1990, such heavy intervention no longer existed: in half of the 30 countries producer prices were no longer controlled or were linked to world prices, and state involvement in export-related activities was limited. A similar turnaround also took place with respect to the trading and marketing of domestic foodstuffs. Overall, some 14 countries had a largely deregulated agricultural sector by 1995.

Exchange and Trade System

In the early 1980s, with the exception of the CFA franc zone countries, distortions in the exchange and trade system were common in the ESAF countries. Outside the CFA franc zone, only Lesotho had an exchange system without surrender requirements or controls on foreign exchange allocation. In all other countries, parallel market premia exceeded 10 percent, and in nine countries surpassed 50 percent.

Reform of the exchange and trade system was an area in which most countries were able to make the most forceful changes to eliminate distortions—particularly in the exchange system, where the reform process began early. Trade reform got off to a slower start but picked up in the 1990s (see Figures 4.11 and 4.12). The intensity of reforms reflected the severity of initial distortions, with the non-CFA African countries making the greatest strides from the weakest starting position. Because exchange and trade restrictions had usually been aimed at officially controlling the allocation of foreign exchange, reforms needed to be supported by changes in exchange rate policy. In fact, countries did devalue or allow the value of their currencies to float down as they undertook reforms.

The elimination of distortions in the exchange system was tackled early (1986–90) in the SAF/ESAF period: already by 1990, five of the nine countries that had comprehensive surrender requirements or controls on foreign exchange allocation in the early 1980s had eliminated them; they remained in Bangladesh, Mauritania, Tanzania, and Zimbabwe.36 By 1995, all countries except Madagascar had eliminated controls on foreign exchange allocation and had either no, or limited, surrender requirements.37 All of the nontransition ESAF countries in this review have accepted the obligations of Article VIII of the IMF’s Articles of Agreement except for Burundi, Lesotho, Mauritania, and Mozambique.38 The removal of controls was accompanied by a reduction in the size of parallel market premia, although their continued presence (in excess of 10 percent in 12 Asian and non-CFA African countries) reflected the effect of remaining restrictions on capital transactions.

Trade reform featured prominently in programs, except in Bolivia, Guyana, Equatorial Guinea, Lesotho, and Niger, which had relatively open systems from the outset. The reform of the trade system began in the second half of the 1980s, most aggressively in the non-CFA African countries. In the other regions, the pace of reform picked up sharply in the 1990s, especially in CFA Africa, after a relatively slim sun. In CFA Africa, the devaluation of the CFA franc and the launching of regional initiatives toward a common economic market and currency union in 1994 gave a modest additional impetus to trade reform.

Quantitative restrictions and export and import licensing requirements were used widely in the early 1980s. An early start was made to reduce these restrictions: by 1990, the number of countries with extensive controls had fallen by more than half to seven, and controls were limited largely to security and health-related restrictions in one-third of the countries. By 1995, no country had extensive controls, and over two-thirds reportedly maintained only security and health-related restrictions; thus, trade systems had become primarily tariff-based.

Tariff reform progressed more slowly until 1991–95. During this period, the number of countries with high and dispersed tariffs, as well as widespread exemptions, fell sharply from 17 to a handful (Burundi, Mozambique, Pakistan, and Zimbabwe). However, important advances in liberalization—to a narrow band or a unitary tariff, low (10–20 percent) average tariffs, and minimal exemptions—took place in only six countries (Benin, Ghana, Guinea, Honduras, Kenya, and Togo), and in over half of the ESAF countries the trade regime was still “moderately” restrictive in 1995.’39 Moreover, in some cases, notably in Africa, tariffs were supplemented by other duties and charges targeted at imports, which are not reflected in the structural indices: examples include a 10 percent sales tax on all merchandise imports in The Gambia; in Burkina Faso, a protective tax of 10–30 percent on imports of goods that are also locally produced; and relatively high customs service fees and duties in several other countries.

Financial Sector Reform

In the early 1980s, governments in most of the countries exercised strict control over interest rates and the allocation of credit—interbank markets were largely nonexistent; only in Kenva, Sri Lanka, and Zimbabwe was there anything much more than a rudimentary or largely uncompetitive financial system in place. Interest rates were either administered directly or were controlled through floors and ceilings or spreads (except in Sri Lanka), and during 1981–85 real deposit interest rates were negative in over two-thirds of the countries. The allocation and volume of credit was controlled through sectoral or bank-specific quantitative ceilings in all but five countries. These credit policies often accommodated the deficits of the parastatal sector on preferential terms, and, because this credit frequently turned sour, nonperforming loans to public enterprises became a significant share of banks’ assets. Overall, the high level of financial repression, weak loan portfolios, and high operating costs contributed to a relatively low level of financial intermediation and widespread bank distress.

Under SAF/ESAF-supported arrangements, financial sector policies were intended to enhance the role of market mechanisms in allocating credit, shifting the means of monetary control to indirect instruments, eliminating the quasi-fiscal activities of the banking system, and strengthening the soundness of banks. In the first two areas of reform, considerable progress was achieved during 1985–95, with a marked acceleration in the pace of reform in the 1990s (Figure 4.14). Quasi-fiscal restrictions on the portfolios of commercial banks were reduced considerably, but they were often replaced by informal pressures on lending activities; in addition, central banks were often called upon to provide soft loans to strategic enterprises and distressed banks. Furthermore, a sustained improvement in the financial health and operational efficiency of banks and in the effectiveness of the prudential and supervisory framework has proved to be elusive.

Figure 4.14Structural Reform in the Financial Sector

(Indices of structural reform; six-point scale)

Source: Appendix 4.1.

1Higher values indicate better structural adjustment policies or fewer distortions.

2Increase in indices denotes improvement in structural policies or decline in distortions.

Interest Rates, Credit, and Instruments of Monetary Control

In non-CFA Africa, Asia, and Western Hemisphere countries there was some modest progress in moving real interest rates on time deposits toward positive levels during 1981–95, but negative real deposit rates prevailed in most years in some three-fourths (or 15–17) of the countries (Figure 4.15). In all of the CFA countries, real deposit rates were positive and generally rising during the late 1980s and early 1990s, in support of the CFA franc. Real rates turned sharply negative as inflation surged temporarily following the CFA franc devaluation in 1994.40 Real lending rates were positive in most ESAF countries throughout 1981–95. In a handful of non-CFA African countries, negative real lending rates had been brought to positive levels by the early 1990s; in Western Hemisphere countries, rates fluctuated sharply, reflecting the volatility of inflation.

Figure 4.15Real Interest Rates1

(Median; in percent a year)

Sources: IMF, IFS; and IMF staff estimates.

1Computed using the contemporaneous rate of inflation of the consumer price index; end-year inflation rates are used where available, otherwise period averages. Nominal deposit interest rates are those on three-month deposits (or closest maturity) and for lending rates, those on one- to three-year loans (or closest proxy).

In general, except in Bolivia and Sri Lanka where interest rates were free of controls prior to their first SAF/ESAF arrangements, interest rates were, initially at least, set administratively. By 1990, three more countries (The Gambia, Ghana, and Malawi) had progressed to full liberalization. There was a strong push toward liberalization during 1991–95, and by 1995 almost two-thirds of the countries had liberalized interest rates; most of those retaining controls were in Africa. However, because banking systems in most countries comprised only a relatively small number of banks, often with significant public ownership, the scope for competitively determined interest rates remained limited. In general, real lending rates rose by more than real deposit rates as banks sought to raise their profitability, and interest rate spreads widened; the number of countries with moderate spreads of 1–5 percent fell from seven to four (Burundi, Guinea, Pakistan, and Sri Lanka) during 1981–95.

Along with liberalizing interest rates, reforms were aimed at deregulating the allocation of credit and introducing indirect means of monetary control. The number of countries without quantitative credit controls rose from five to almost half of the countries by 1990, largely reflecting reforms in the CFA countries that were introduced on a regional basis by the West African Monetary Union (WAMU) beginning in 1989. By 1995, controls on the allocation or volume of credit remained in only seven countries (five non-CFA African countries, Nepal, and Pakistan). The shift toward indirect instruments of monetary control involved a wide range of measures, including changes in reserve requirements, measures to make central bank refinancing more market-oriented, the introduction of government and central bank securities (sometimes with regular auctions), and the development of active secondary markets for government securities and interbank markets.

An indicator of the pace and extent of the switch to indirect monetary instruments is given by the introduction of government and central bank bills and the creation of interbank markets. Markets for government securities existed in eight countries in the early 1980s and were steadily introduced in another 15 countries during 1986–95; by 1995 they were absent only in Mozambique and CFA Africa.41 Interbank markets were established at a slower pace. By 1990, nine countries had such markets, but during 1991–95 the number of countries more than doubled, in large part reflecting reforms in the WAMU.42 Initially, reflecting countries’ inexperience in the use of indirect monetary controls, these new systems tended to supplement rather than replace traditional means of control; their effective use is also likely to have been impaired by the degree of financial distress in the banking system. However, as experience grew, increasing reliance was placed on indirect instruments, and IMF-supported programs relied more heavily on targeting the net domestic assets of the central bank rather than of the banking system.

Although much was accomplished in financial sector reforms, no significant increase in financial deepening or intermediation has occurred since the early 1980s: broad money and private sector credit in relation to GDP, and the share of deposits in broad money, have increased only marginally, if at all. The low rate of financial saving reflects in part the low level of per capita GDP, the persistence of low or negative real deposit interest rates, and a lack of confidence in the banking system. The last of these owes much to two interrelated factors: continued government intervention and the poor health of banks.

Bank Restructuring

Virtually all the ESAF countries have experienced episodes of banking system distress since the mid–1980s. The proximate causes of banks’ problems were typically macroeconomic. In part, these stemmed from policies taken as part of adjustment programs that exposed underlying structural flaws; for example, lacking the flexibility needed to adjust to major exchange rate realignments and the dismantling of protectionist barriers, inefficient private and public enterprises became uncompetitive and unable to service their debts. Because of the concentrated production base and loan portfolios of banks, exogenous shocks, especially the deterioration in the terms of trade during the first half of the 1980s, had a significant impact on bank performance.

The underlying causes of banking problems were, however, long-standing structural weaknesses. As noted above, in most countries the state exercised control over interest rates and the allocation of credit and accorded public enterprises preferential treatment. Also, accounting regulations were lax and supervision ineffective, and insider lending and political interference in loan decisions were commonplace. Moreover, bank managers were often politically motivated appointees, while commercial banking skills and internal control systems were weak. In most countries, the banking sector was dominated by state-owned (or majority-controlled) banks. In these circumstances, credit policies leaned toward accommodating the deficits of public enterprises and agricultural marketing boards, which resulted in the accumulation of large stocks of nonperforming loans. The deterioration of asset portfolios, together with poor incentives for depositors, eventually brought a number of significant banks, typically state-owned, to the point of insolvency.

In most of the countries, banking difficulties reflected a situation of distress (persistent solvency problems) rather than an outright liquidity crisis. The absence of a hill-blown crisis reflected in part the expectation of government protection (especially for state-owned banks) by depositors. The scope of distress was often large, and in about one-third of the ESAF countries, more than half of the banking system’s loan portfolio was nonperforming (Table 4.6). To address these problems, most countries implemented systemic bank restructuring reforms under SAF/ESAF-supported programs.

Table 4.6Banking System Distress in Nontransition ESAF Countries
CountryYears1Share of Nonperforming Loans in Banks’ Portfolios (in percent)
Africa
CFA
Benin1988–9078 (1988)
Burkina Faso1988–9434 (…)
Côte d’Ivoire1988–9012 (…)
Equatorial Guinea1983–85, 1995Two large banks liquidated (1986); principal bank’s main shareholder was placed in receivership (1995)
Mali1987–8975 percent of the largest bank
Niger1983-present50 (mid–1980s)
Senegal1983–8850 (1988)
Togo
Non-CFA
Burundi1994-present25 (1995)
Gambia, The1985–9210 (1992)
Ghana1983–8960 (1991)
Guinea1988-present26 (1995)
Kenya199366 percent of one-third of commercial banks
Lesotho1996Affects the dominant commercial bank
Madagascar1991–9525 (…)
Malawi
Mauritania1991–93
Mozambique1988-presentMost loans at end–1986
Sierra Leone1990-present40–50 (1995)
Tanzania1988-present60–80 (…)
Uganda1990-present40 (1995)
Zimbabwe1995-presentAffects two large banks
Asia
Bangladesh1971-presentA high percentage
Nepal198829 (1988)
Pakistan199635–45 (1996)
Sri LankaEarly 1990s20–25 (early 1990s)
Western Hemisphere
Bolivia1986–1987, 1994–9630 (1987); 6 (1995)
Guyana1993–95
HondurasNo crisis(1996)
NicaraguaLate 1980s-present50 (…)
Sources: IMF staff estimates; Lindgren, Garcia, and Saal (1996); and Caprio and Klingebiel (1996).

Based on information available through mid–1997.

Sources: IMF staff estimates; Lindgren, Garcia, and Saal (1996); and Caprio and Klingebiel (1996).

Based on information available through mid–1997.

Restructuring programs were cast in a medium-term framework—with many still ongoing—and sought to rehabilitate banks’ balance sheets and install sound commercial operating procedures in banks; strengthen the accounting and regulatory framework; improve supervision and compliance with prudential regulations; and promote competition. Some banks were reorganized or closed,43 but in general problem banks were recapitalized by the government, which took over the nonperforming loans. In only a few instances were existing private equity holders or new investors called upon to inject new capital (for example, Mauritania, 1993, and Tanzania, 1995), and depositors were expected to absorb losses in even fewer cases (such as Côte d’Ivoire, 1991). In many cases, non-performing assets were transferred to a centralized loan recovery unit. Because of the preponderance of state-owned banks in distress, the financial involvement of the government was substantial. However, governments also extended sizable support to largely private banks, in tacit recognition of their responsibility through directed credit policies, and because of concerns about confidence in the banking system. For those countries for which data are available, the fiscal cost was clearly large—15–25 percent of GDP in Benin, Côte d’Ivoire, Mauritania, and Senegal.

To strengthen banks’ income and capacity for sustainable earnings, and to prevent the recurrence of banking problems, reforms included changes in management, staff reductions and branch closures and other cost-cutting measures, as well as improvements in operating practices such as credit assessment and loss provisioning. These measures were implemented within the framework of performance or management contracts, and through the participation of foreign investors as well as privatization (involving a “core” investor) in order to introduce strong commercial banking skills.44 All of the countries introduced, to varying degrees, institutional measures to strengthen supervision and prudential regulation.

Comprehensive data on the implementation and effectiveness of the restructuring measures are not available. Nevertheless, there is considerable evidence of delays in implementation and weak enforcement of sound banking practices, and the reform process has proven to be protracted. In some countries, banking problems recurred (for example, in Albania and Lao P.D.R.) or worsened markedly (for example, in Tanzania), and in most countries the period of distress continued for several years after the onset of reform. While the dearth of domestic human capital in banking skills and the capacity to implement the reforms, especially with regard to monitoring the performance of banks, impeded progress, the level of political commitment was also a key factor, especially in view of the large role of state banks. An important aspect of political involvement was the pressure (often in the form of moral suasion) on banks to accommodate the continuing demand for credit by weak public enterprises and in some instances to grant politically motivated loans to the private sector. To the extent that the banking system mirrors the health of the real economy, delays in public enterprise reform held back the process of bank restructuring. Also, in most countries the banking sector continued to be dominated by a handful of banks and there was relatively little progress in enhancing market discipline.

Has bank restructuring led to improved banking system performance? Lack of adequate data make a comprehensive assessment impossible. Balance sheet data and quantitative measures of operational efficiency are especially fragmentary (and at times of questionable quality). Broad quantative assessments, however, indicate that in only a few countries had the banking systems been decisively restored to sound health on a sustainable basis. Comparison of four indicators of intermediation performance for the three five-year periods (1981–85, 1986–90, and 1991–95) suggests that in most countries there was little improvement between the pre- and postreform period in the extent and efficiency of financial intermediation. Of course, developments in these indicators were also influenced by factors other than banking system health.45

Public Enterprise Reform

Until the early 1980s, governments promoted the public enterprise sector in the belief that an interventionist role by the state was needed to promote development and industrialization. By the mid–1980s, public enterprises played an important role in production, investment, and formal sector employment and were involved in a wide range of activities, typically with a dominant presence in mining, manufacturing, utilities, and trading and marketing of imports and agricultural products.46 Furthermore, they were given preferential treatment, including monopoly and monopsony rights, government guarantees on external borrowing, and cheap access to bank credit. A composite measure of the role of public enterprises in the economy and their financial performance based on available data indicates that almost half of the nontransition countries under review had public enterprise sectors that were large relative to those of other developing countries, recorded substantial financial losses, were a drain on the budget and banking sectors, and imposed a relatively severe distortionary burden on the economy (Figure 4.16). In the remaining countries, public enterprises still had a significant, albeit less severe, adverse influence on the allocation of resources.47

Figure 4.16Structural Reform in the Public Enterprise Sector

(Indices of structural reform; six-point scale)

Source: Appendix 4.1.

1Higher values indicate better structural adjustment policies or fewer distortions.

2Increase in indices denotes improvement in structural policies or decline in distortions.

The reform of public enterprises was an important part of SAF/ESAF-supported programs in all of the countries, typically with the World Bank in a lead role. Reforms began in the mid- to late 1980s and in several countries were started prior to the first SAF/ESAF-supported programs. The primary objectives of reform were to improve the efficiency and financial performance of the public enterprise sector. From a macroeconomic perspective, this was essential to raise public saving, curb inflationary pressures, and strengthen banking systems. The first step in the reform process was typically a preparatory study to identify problems and propose solutions. The reforms themselves comprised the restructuring, liquidation, and privatization of public enterprises and encompassed both small and large strategic enterprises. In two-thirds of the countries, reform programs were broadly based, covering the sector as a whole, while in the others the approach was more piecemeal, often initiated in response to the emergence of significant financial difficulties in selected enterprises.48 In all of the countries, the reform of larger enterprises lagged that of others.

The emphasis and focus of reforms evolved over the SAF/ESAF period. Initially, most reform programs sought to restructure enterprises considered to be strategic—utilities, telecommunications, transport, heavy industries, and agricultural marketing boards and processing industries. Loss-making nonstrategic enterprises were to be liquidated, and commercially viable nonstrategic enterprises were to be privatized. In practice, this strategy targeted the divestment of a substantial number of small but economically less important public enterprises and the restructuring of only a small number of large public enterprises—often the major source of financial losses and economic distortions. Countries also enacted changes to the regulatory and legal framework needed to liquidate or privatize enterprises, to increase the autonomy of public enterprises (for example, in pricing decisions), and to strengthen market discipline by creating a more competitive environment.

Since the early 1990s, there has been a growing recognition that restructuring was not producing the desired results, in large part because performance contracts did little to change the outlook and behavior of existing managers and because the lack of comprehensive financial information frustrated attempts to enforce hard budget constraints on enterprises. Direct budgetary support was reduced or eliminated, but it was often replaced by indirect quasi-fiscal subsidies, such as continued bank borrowing with government guarantees and the nonpayment of taxes. As a result, a growing number of countries subsequently initiated more ambitious programs of privatization for strategic enterprises, including through the use of leases and concessions. To date, however, except for Bolivia and Ghana, actual progress in privatizing larger enterprises is still modest, reflecting a lack of government commitment and strong resistance from vested interests.

Information about the public enterprise sector is generally inadequate. The available evidence, however, points to a protracted process of reform over the past ten years, with only limited and spotty progress in improving the performance of public enterprises and in reducing the size of the sector. The pace of reform typically fell short of plans: in two-thirds of the ESAF countries, the broad objectives of their reform programs were achieved only after significant slippages of one or more years; in one-fourth of the countries (mainly African) the stated objectives of reform were never realized.

The size and weak financial performance of the public enterprise sector remained largely unchanged in about half of the nontransition ESAF countries. In the remaining half the record shows improvement between the 1980s and the early 1990s, including a fall in the number of countries (from 12 to 3) in which the public enterprise sector might be labeled as a “severe” burden on the economy.49 Moreover, although progress was generally slow, the pace of reform picked up in the 1990s.

Civil Service Reform

Since the mid–1980s, civil service reform has featured in all the countries under review, and in most the process was begun within a few years of the start of the SAF/ESAF period. Before the 1990s, reforms focused primarily on alleviating the heavy fiscal burden of large wage bills and the consequent crowding out of critical expenditures. During the 1990s, however, there has been increasing emphasis on improving the efficiency and quality of public administration. In support of these goals, countries undertook short-term measures to contain costs through wage restraint and downsizing, as well as medium-term programs to strengthen personnel management and training so as to improve skills and incentives. These programs included measures such as the redeployment of personnel, restructuring of ministries, introduction of formal hiring and promotion procedures, decompression of the wage structure to improve incentives, and establishing competitive wages for highly skilled civil servants.

In over two-thirds of the countries, a civil service census was an early step in identifying the magnitude of the problem. Efforts to downsize the civil service in most countries included the elimination of “ghost” workers and unfilled posts, as well as voluntary departure schemes. Only a handful of countries undertook the more sensitive actions of freezing recruitment and retrenchment; in Malawi, Mauritania, Senegal, and Tanzania health, education, and police workers were excluded from the recruitment ceilings. These measures were partially offset by new hiring, and a significant net reduction in the civil service was achieved in only about one-fourth of the countries since the beginning of the SAF/ESAF period (Table 4.7). In virtually all countries, however, the size of the civil service in relation to population fell slightly.

Table 4.7Change in the Size of the Civil Service During SAF/ESAF Period1
Increase of 5 Percent or MoreNo Significant ChangeDecrease of 5 Percent or More
Africa
CFABurkina Faso(5)Côte d’Ivoire(2)Benin(7)
Niger(6)2Senegal(10)Mali(7)
Togo(7)
Non-CFAGuinea(4)3Burundi(4)Madagascar(9)
Kenya(5)The Gambia(9)Uganda(9)4
Lesotho(5)Ghana(9)
Mauritania(10)Zimbabwe(3)
AsiaSri Lanka(7)
Western HemisphereBolivia(9)
Guyana(6)
Source: IMF staff estimates.

Percent change between the year preceding the first SAF/ESAF arrangement (or closest year) and 1995 (or latest year available). Number of years over this period are reported in parentheses. Information on civil service employment was available for only 20 nontransition countries. Major departures from the pre-SAF/ESAF year are indicated in footnotes.

1988/89–94.

1991–95.

1984/85–93/94.

Source: IMF staff estimates.

Percent change between the year preceding the first SAF/ESAF arrangement (or closest year) and 1995 (or latest year available). Number of years over this period are reported in parentheses. Information on civil service employment was available for only 20 nontransition countries. Major departures from the pre-SAF/ESAF year are indicated in footnotes.

1988/89–94.

1991–95.

1984/85–93/94.

Explicit limits on increases in the wage bill (underpinned by understandings on wage increases and employment levels) were incorporated in SAF/ESAE-supported programs in two-thirds of the countries. In many cases, including most of the non-CFA African countries, these limits provided for improvements in real wage levels relative to the nongovernment sector, but in some cases these were subsequently eroded by inflation. In several countries, selective wage increases and adjustments to the salary grade were implemented as part of efforts to reform the pay structure.50 In some countries (for example, Ghana and Mozambique), the range of civil service salaries was decompressed, while in others (The Gambia and Guinea) wages became more compressed (Nunberg and Nellis, 1995).

The impact of employment and pay reforms on the government’s wage bill (measured as a share of GDP) was mixed, with about one-half of the countries cutting their wage bill.51 On average, it was those countries with relatively larger wage bills prior to the SAF/ESAF period that reduced the wage bill; the greatest cuts, on average, were achieved in the CFA African countries (Table 4.8). However, in many cases these reductions tended to rely more on reduced real wages, and less on cutbacks in employment, than planned.

Table 4.8Civil Service Wage Bill During SAF/ESAF Period(Average; in percentage points of GDP)
Pre-SAF/ESAF2Change Between Pre-SAF/ESAF and 1995
Region11995TotalAverage annual change3
Region
Africa
CFA (8)7.85.8-2.0-0.4(7.8)
Non-CFA (14)6.86.1-0.7-0.1(8.6)
Asia (3)3.13.50.50.1(9.0)
Western Hemisphere (4)9.98.0-2.0-0.4(6.3)
Ranked by initial wage bill
Pre-SAF/ESAF median wage bill7.2(…)
Countries with initial wage bill greater than median in pre-SAF/ESAF period9.97.5-2.4-0.4(7.4)
Countries with initial wage bill smaller than median in pre-SAF/ESAF period4.24.50.30.01(8.9)
Sources: IMF staff estimates.

Number of countries in parentheses.

Average for three years preceding the first SAF/ESAF arrangement, or average of years for which data are available.

Number of years between pre-SAF/ESAF period and 1995 or most recent year for which data are available is shown in parentheses.

Sources: IMF staff estimates.

Number of countries in parentheses.

Average for three years preceding the first SAF/ESAF arrangement, or average of years for which data are available.

Number of years between pre-SAF/ESAF period and 1995 or most recent year for which data are available is shown in parentheses.

Property Rights

In the push toward more open, market-based systems, the need for well-defined property rights and legal systems and practices that support their respect and enforcement—essential elements of good governance—has become increasingly apparent and attracted more attention in ESAF-supported programs. For example, the lack of established property rights often delayed the start of loan recovery and workout exercises in bank restructuring and public enterprise privatization; an underdeveloped judicial system and lack of a property registry that would allow loans to be adequately collateralized were impediments to increasing financial intermediation and may have contributed to wider interest rate spreads; and new property rights were required for the privatization of land.

The strength or quality of property rights is inherently difficult to assess directly, particularly on a cross-country basis. However, attempts have been made by private agencies to rate countries on various criteria of “economic security” that are of particular relevance to actual or potential investors.52Figure 4.17 shows an index based on three criteria—the quality and acceptance of the tools of law and order, the quality of public administration and extent of government corruption, and the contract and expropriation risks to foreign investors—using data from the ICRGD. This evidence suggests that, in broad terms, the status of economic security improved between 1986–90 and 1990–91 in two-thirds of countries but deteriorated in one-third. Improvements appear to be concentrated mainly in Asian and Western Hemisphere countries, although some countries in Africa (Ghana, Tanzania, and Uganda) also made significant progress in this area.

Figure 4.17Index of Economic Security1

(Six-point scale)

Source: Center for Institutional Reform and the Informal Sector, International Country Risk Guide (College Park, Maryland: IRIS, University of Maryland).

1Higher values indicate better governance.

2Sufficient data are not available to construct an index for CFA Africa and non-CFA Africa.

3Excluding Benin and Equatorial Guinea, for which data are not available.

4Excluding Burundi, Lesotho, and Mauritania, for which data are not available.

5Excluding Nepal, for which data are not available.

It is difficult to assess the extent to which ESAF-supported programs helped to improve the status of property rights. Indirectly, structural reforms under programs are likely to have contributed in many ways—for instance by liberalizing and making more transparent foreign direct investment codes and assisting in the reform of public administration. More directly, public enterprise reforms and the restructuring of banking systems in many cases necessitated a strengthening of the legal framework, and land tenure systems were revised.

Appendix 4.1. Indices of Structural Distortions

Empirical investigations of the links between the removal of structural distortions and economic growth are hampered by difficulties in obtaining good measures or proxies of structural policies. Ideally, measures of structural distortions or reforms should be objective, comparable across countries, easy to compute and interpret, and hence yield unambiguous policy implications. In practice, few such measures exist. Empirical studies, therefore, proxy distortions and policies using indicators that are essentially observable outcomes, which, in turn, can also be affected by non-policy-related factors. For example, a widely used measure of openness—the ratio of total trade to GDP—fails to control for the fact that larger economies can generally support a more diversified productive base and thus tend to have lower trade ratios even if they are relatively open. Likewise, the ratio of trade taxes to total trade as a measure of autarky can be misleading if, in the course of trade liberalization, quantitative restrictions and bans are replaced with tariffs. The resultant increase in the ratio of trade taxes to trade may be seen as signaling an increase in barriers to trade, when in fact such barriers have been lowered. In general, such indicators have not yielded robust links between growth and structural reforms in econometric analyses.

Several studies have attempted a more direct approach to the measurement of structural distortions and reforms, by combining various indicators of policies to construct composite “scores” of distortions and policy performance. An early example of this approach is Agarwala (1983). More recent examples include Levine and Zervos (1993); Sachs and Warner (1995); Denizer, De Melo, and Gelb (1996); European Bank for Reconstruction and Development (1995); Fischer, Sahay, and Vegh (1996); and World Bank (1994 and 1996). Considerable caution, however, is needed in using and interpreting the results of such an approach. First, in many instances the indicators suffer the shortcomings discussed above and therefore may not accurately capture the strength of structural reform efforts. Second, there are structural elements of economies that are inevitably omitted, because they are very difficult to measure, but that can have a direct bearing on growth, such as the nature and quality of institutions and administrative systems. Third, a large element of judgment is involved in classifying and assigning scores to structural policies.

While recognizing these caveats, a scoring approach is used here to calibrate key structural aspects of the economy in the 30 nontransition ESAF countries under review. The aim of this exercise is twofold: first, to facilitate a more orderly discussion of the status of structural distortions and progress with structural reforms in each country (see the section on structural reforms, above); and, second, to permit a more systematic examination of the links between such policies and growth than would otherwise be possible (see Chapter 5). The focus is on structural distortions in five major areas typically covered in SAF/ESAF-supported programs—the exchange and trade systems, price controls and state intervention in marketing, the financial sector, and the public enterprise sector.53 This appendix is organized in two parts. The first contains a discussion of the scoring methodology; the second presents a list of the indicators that are used to assess the structural distortions and, through them, the strength of structural reforms.

General Aspects of the Scoring Methodology

The general approach is to use a variety of indicators to obtain a snapshot of structural distortions for each country at different times. As discussed above, one of the reasons that the links between structural policies and growth are difficult to establish empirically is that measurable proxies for such policies can, very often, give a misleading picture of the status of distortions in an economy. Moreover, there is no consensus in the literature on the clear superiority of any one measure or indicator of structural policies. To partially mitigate some of the problems associated with using any single measure of policies or distortions, the approach here uses several different indicators, within the constraints of data availability, to get a fuller picture of the status of structural distortions. Since different indicators can give a different picture of the relative severity of a country’s structural distortions, the issue arises of how to combine them so that the idiosyncracies of the weighting scheme do not themselves produce misleading results. In the absence of any clearly satisfactory method for determining weights, equal weight has been given to the different indicators in the structural policy scores.

Given the difficulty of obtaining reliable information on most of the chosen indicators on an annual basis, the period between 1991 and 1995 is divided into three non-overlapping five-year periods and at least one observation is made in each period for each indicator.54 These observations are not necessarily synchronous across countries and may not, as a rule, coincide with the implementation of SAF/ESAF-supported adjustment programs. It is, therefore, not possible to discuss progress with structural reforms strictly in terms of “before and after” SAF/ESAF-supported programs. In general, the period 1981–85 (the first five-year period) is, for many countries, a reasonable proxy for “before SAF/ESAF,” and the two subsequent five-year periods can be considered part of a broadly defined “adjustment” phase; however, six countries—Burkina Faso, Côte d’Ivoire, Guyana, Honduras, Nicaragua, and Zimbabwe—that had their first SAF/ESAF arrangements in the 1990s are exceptions.

In most cases, a six-point scale was used to classify distortions as low, moderate, or high according to some objective notion of “best practices.” In cases where there is no consensus on “best practices” (such as for indicators of public enterprise reforms), the data for each country were compared to developing country averages, and the scoring scale was calibrated accordingly.55 In each case higher scores indicate better policies or fewer distortions. For three of the financial sector indicators, relating to the depth of financial intermediation, it was possible to measure the extent of distortions along a continuous scale centered on the average value for developing countries and bounded by end-values of 2 and 6 (that is, on a comparable scale to that used for the other indices).

Barometers of Structural Policies Used to Construct Scores

The indicators used to measure structural policies and the scoring methodology are described in Table 4.9; individual country scores in each area of structural policy are contained in Tables 4.104.16. The individual country scores and group averages are reported using a six-point scale: scores can be classified as low (a score of 1–2), moderate (3–4), or high (5–6) relative to a notion of best practices (a score of 5–6) or to an average for all developing countries (a score of 3–4). The choice of indicators was guided by the objective of capturing the extent to which policies and institutions permit the free interplay of market forces to occur.

Table 4.9Indicators and Scoring System for Structural Policies(Higher scores indicate better policies or fewer distortions)
Structural Policy ElementScoreDescription of Policies
Exchange System
Black market exchange rate premium6Premium below 10 percent
4Premium between 10 percent and 50 percent
2Premium greater than 50 percent
Surrender requirements and foreign exchange allocation6Full retention/market-based foreign exchange allocation or full surrender but no restrictions on foreign exchange allocation
4Partial retention plus limited non-market foreign exchange allocation or full surrender plus some non-market foreign exchange allocation
2Full surrender or comprehensive non-market foreign exchange allocation
Trade Policies
Quantitative restrictions (QRs)6QRs on imports and exports only for security and health reasons
4Some QRs on imports and exports
2Extensive use of QRs, and import and export licensing
Tariff system6Low average tariffs of about 10–20 percent, maximum tariffs in the range of 20–30 percent and minimal exemptions
4Moderate average tariffs of about 20–25 percent, maximum of 30–40 percent with few exceptions to the maximum rate and few exemptions overall
2Very high and dispersed tariffs with proliferation of exemptions
Financial Sector
Interest rates and credit controls6Presence of three or more of the following:

  • - Ex post real deposit interest rates between 0 and 5 percent
  • - Spread between deposit and lending rates between 1 and 5 percent
  • - No credit ceilings or direct credit controls
  • - No interest rate controls
4Presence of two of the above
2Presence of one or fewer of the above
Financial depth
Financial deepening, measured by the ratio of broad money to GDP

Private sector’s access to credit, measured by the ratio of private sector credit to GDP

Degree of financial intermediation, measured by the ratio of currency to broad money
The scoring in this subsection is done by mapping values of these variables, averaged over the five-year period, to a continuous scale between 2 and 6; for the first two variables, higher values are assigned higher scores, and, for the third, lower values are assigned higher scores
Structure of the banking system and financial markets6Presence of three or more of the following:

  • - Banking system reasonably diversified and banks mainly private
  • - Interbank market exists
  • - Market for government or central bank securities exists
  • - Stock market exists
4Presence of two of the above
2Presence of one or fewer of the above
Public Enterprise Reform
(Scores are centered around developing country averages)
Role of public enterprises (PEs) in economic activity
Output6Share of PE output in economic activity of less than 8 percent
4Share of PE output in economic activity between 8 and 12 percent
2Share of PE output in economic activity over 12 percent
Employment6Share of PE employment in total employment of less than 5 percent
4Share of PE employment in total employment of between 5 and 8 percent
2Share of PE employment in total employment over 8 percent
Domestic bank credit6Share of credit to PEs in total credit less than 8 percent
4Share of credit to PEs in total credit between 8 and 12 percent
2Share of credit to PEs in total credit greater than 12 percent
Public enterprise sector’s financial performance
Overall financial balance6PEs’ overall balance before transfers (relative to GDP) of more than –2 percent
4PEs’ overall balance before transfers (relative to GDP) between –2 and –4 percent
2PEs’ overall balance before transfers (relative to GDP) of less than –4 percent
Net financial flows from government6Net financial flows from government to PEs (relative to GDP) of less than –2 percent
4Net financial flows from government to PEs (relative to GDP) between 0 and –2 percent
2Positive net financial flows from government to PEs
Price Controls and State Intervention in Marketing
Fertilizers6No controls on pricing and no intervention in marketing
4Controls on pricing or intervention in marketing
2Controls on pricing and intervention in marketing
Petroleum products
Pricing6Prices freely determined or automatic adjustment mechanism
4Prices linked to world prices and periodically adjusted
2Prices set at government discretion
Marketing6No controls on import, wholesale or retail marketing
4Controls on imports or wholesale or on retail marketing
2Controls on imports or wholesale and on retail marketing
Major foodstuffs
Pricing6No price controls
4Controls on few commodities (less than 10)
2Controls on many commodities (more than 10)
Marketing6No controls on marketing
4Imports of wheat or rice or marketing of domestic food crops state-controlled
2Imports of wheat or rice and marketing of domestic food crops state-controlled
State intervention in export activities
Pricing6No controls on producer prices or automatically linked to world prices
4State-controlled producer prices linked to world prices, periodically adjusted
2State-controlled producer prices, usually set below world prices or costs
Marketing6Insignificant (less than 10 percent of total exports) direct involvement in export-related activities by the state
4Moderate involvement (between 10 and 50 percent of total exports) by the state in export trading
2Heavy involvement (more than 50 percent of total exports) of state entities in trade of major export commodities
Overall price controls (excluding petroleum products and utilities)
6No price controls
4Few price controls (fewer than 25 goods)
2Many price controls (more than 25 goods)
Table 4.10Overall Structural Reform Index1

(Six-point scale)2

ScoreCumulative ScoreChange in Score
1981–85

(I)
1986–90

(II)
1991–95

(III)
1981–85

(I)
1981–90

(I + II)
1981–95

(I + II + III)
II

over I
III

over II
III

over I
(Averages of country scores)
Africa3.03.64.43.06.611.00.60.81.4
CFA3.13.64.43.16.811.20.50.81.3
Non-CFA2.93.64.42.96.410.90.70.91.5
Asia3.03.84.43.06.811.20.70.61.4
Western Hemisphere3.23.94.83.27.111.90.70.91.6
Africa(Individual country scores)
CFA
Benin2.93.54.52.96.411.00.61.01.6
Burkina Faso3.13.14.53.16.110.6-1.41.4
Côte d’Ivoire3.03.54.03.06.510.50.50.51.0
Equatorial Guinea3.23.53.73.26.710.40.30.30.5
Mali3.03.64.63.06.611.10.51.01.5
Niger3.74.35.03.78.013.00.60.71.3
Senegal3.13.84.43.16.911.30.70.51.4
Togo3.13.84.53.17.011.50.70.71.4
Non-CFA
Burundi3.03.54.33.06.510.80.50.81.3
The Gambia3.14.25.13.17.312.41.20.92.1
Ghana2.43.74.32.46.010.41.30.72.0
Guinea2.84.04.92.86.811.71.30.92.2
Kenya3.43.95.03.47.312.30.41.11.5
Lesotho4.04.04.14.08.112.2-0.10.1
Madagascar2.93.43.92.96.310.20.50.51.0
Malawi3.23.74.93.26.911.80.51.21.7
Mauritania3.03.74.23.06.610.90.70.61.3
Mozambique2.52.93.72.55.49.20.50.81.3
Sierra Leone2.83.74.72.86.511.20.91.11.9
Tanzania2.22.83.62.25.08.60.60.91.4
Uganda2.83.34.42.86.110.60.51.11.7
Zimbabwe2.73.14.42.75.810.20.41.31.7
Asia
Bangladesh2.63.54.12.66.110.20.90.71.5
Nepal3.23.74.43.26.811.20.50.71.2
Pakistan3.24.24.63.27.412.01.00.41.4
Sri Lanka3.13.84.43.16.911.30.70.61.3
Western Hemisphere
Bolivia3.74.75.43.78.413.81.10.61.7
Guyana3.33.94.53.37.211.70.70.51.2
Honduras3.33.85.13.37.112.20.51.21.8
Nicaragua2.53.04.42.55.59.90.51.31.9
Mean3.03.74.53.06.711.20.60.81.5
Minimum2.22.83.62.25.08.6-0.10.1
Maximum4.04.75.44.08.413.81.31.42.2
Standard deviation0.40.40.40.40.81.10.30.30.4
Source: Tables 4.114.16.

Unweighted average of scores for pricing and marketing, external sector, financial sector, and public enterprise sector indices.

The extent of reform is classified as low (score of 1–2), moderate (3–4), or high (5–6) relative to a specified notion of “best practices” (score of 5–6) or to an average for all developing countries (3–4).

Source: Tables 4.114.16.

Unweighted average of scores for pricing and marketing, external sector, financial sector, and public enterprise sector indices.

The extent of reform is classified as low (score of 1–2), moderate (3–4), or high (5–6) relative to a specified notion of “best practices” (score of 5–6) or to an average for all developing countries (3–4).

Table 4.11Indices of Structural Reform: Pricing and Marketing

(Six-point scale)1

ScoreCumulative ScoreChange in Score
1981–85

(I)
1986–90

(II)
1991–95

(III)
1981–85

(I)
1981–90

(I + II)
1981–95

(I + II + III)
II

over I
III

over II
III

over I
(Averages of country scores)
Africa2.73.84.92.76.511.41.11.12.2
CFA2.53.64.82.56.110.91.11.32.3
Non-CFA2.83.94.92.86.811.71.11.02.1
Asia2.93.84.22.96.710.90.90.51.3
Western Hemisphere2.93.44.82.96.311.10.51.41.9
Africa(Individual country scores)
CFA
Benin2.63.64.82.66.211.01.01.22.2
Burkina Faso2.23.04.82.25.210.00.81.82.6
Côte d’Ivoire2.42.84.42.45.29.60.41.62.0
Equatorial Guinea2.83.23.82.86.09.80.40.61.0
Mali2.44.45.62.46.812.42.01.23.2
Niger3.04.05.23.07.012.21.01.22.2
Senegal2.43.84.42.46.210.61.40.62.0
Togo2.23.65.62.25.811.41.42.03.4
Non-CFA
Burundi2.63.85.22.66.411.61.21.42.6
The Gambia3.05.05.43.08.013.42.00.42.4
Ghana3.04.04.43.07.011.41.00.41.4
Guinea3.84.65.03.88.413.40.80.41.2
Kenya2.43.25.42.45.611.00.82.23.0
Lesotho4.64.65.24.69.214.4-0.60.6
Madagascar2.03.43.62.05.49.01.40.21.6
Malawi3.23.25.63.26.412.0-2.42.4
Mauritania2.64.44.82.67.011.81.80.42.2
Mozambique2.24.45.02.26.611.62.20.62.8
Sierra Leone2.44.45.42.46.812.22.01.03.0
Tanzania2.03.44.02.05.49.41.40.62.0
Uganda3.04.25.63.07.212.81.21.42.6
Zimbabwe2.62.64.22.65.29.4-1.61.6
Asia
Bangladesh3.03.84.43.06.811.20.80.61.4
Nepal3.23.84.23.27.011.20.60.41.0
Pakistan3.23.84.23.27.011.20.60.41.0
Sri Lanka2.23.64.02.25.89.81.40.41.8
Western Hemisphere
Bolivia3.64.05.03.67.612.60.41.01.4
Guyana3.64.44.43.68.012.40.8-0.8
Honduras2.43.05.02.45.410.40.62.02.6
Nicaragua2.02.24.62.04.28.80.22.42.6
Mean2.83.74.82.86.511.31.01.02.0
Minimum2.02.23.62.04.28.8--0.6
Maximum4.65.05.64.69.214.42.22.43.4
Standard deviation0.60.60.60.61.11.40.60.70.7
Source: IMF staff estimates.

The extent of reform is classified as low (score of 1–2), moderate (3–4), or high (5–6) relative to a specified notion of “best practices” (score of 5–6).

Source: IMF staff estimates.

The extent of reform is classified as low (score of 1–2), moderate (3–4), or high (5–6) relative to a specified notion of “best practices” (score of 5–6).

Table 4.12Indices of Structural Reform: External Sector (Exchange and Trade Systems)1

(Six-point scale)2

ScoreCumulative ScoreChange in Score
1981–85

(I)
1986–90

(II)
1991–95

(III)
1981–85

(I)
1981–90

(I + II)
1981–95

(I + II + III)
II

over I
III

over II
III

over I
(Averages of country scores)
Africa3.54.35.23.57.813.10.80.91.7
CFA4.74.95.64.79.615.20.30.60.9
Non-CFA2.93.95.02.96.811.81.11.12.2
Asia3.13.44.63.16.511.10.31.31.5
Western Hemisphere3.44.05.63.47.413.00.61.62.3
Africa(Individual country scores)
CFA
Benin4.55.06.04.59.515.50.51.01.5
Burkina Faso4.54.55.54.59.014.5-1.01.0
Côte d’Ivoire4.55.05.54.59.515.00.50.51.0
Equatorial Guinea5.05.05.55.010.015.5-0.50.5
Mali4.54.55.54.59.014.5-1.01.0
Niger5.06.06.05.011.017.01.0-1.0
Senegal4.54.55.04.59.014.0-0.50.5
Togo5.05.05.55.010.015.5-0.50.5
Non-CFA
Burundi3.54.54.53.58.012.51.0-1.0
The Gambia4.04.55.54.08.514.00.51.01.5
Ghana2.04.06.02.06.012.02.02.04.0
Guinea2.55.56.02.58.014.03.00.53.5
Kenya3.03.55.53.06.512.00.52.02.5
Lesotho5.55.55.55.511.016.5---
Madagascar2.54.04.52.56.511.01.50.52.0
Malawi2.54.55.02.57.012.02.00.52.5
Mauritania3.04.04.53.07.011.51.00.51.5
Mozambique2.02.54.52.04.59.00.52.02.5
Sierra Leone3.03.55.53.06.512.00.52.02.5
Tanzania2.03.05.02.05.010.01.02.03.0
Uganda2.03.04.52.05.09.51.01.52.5
Zimbabwe2.53.04.02.55.59.50.51.01.5
Asia
Bangladesh2.02.04.02.04.08.0-2.02.0
Nepal3.03.55.03.06.511.50.51.52.0
Pakistan3.54.04.53.57.512.00.50.51.0
Sri Lanka4.04.05.04.08.013.0-1.01.0
Western Hemisphere
Bolivia4.56.06.04.510.516.51.5-1.5
Guyana3.54.05.53.57.513.00.51.52.0
Honduras3.52.56.03.05.511.5-0.53.53.0
Nicaragua2.53.55.02.56.011.01.01.52.5
Mean3.54.15.23.57.612.80.71.11.8
Minimum2.02.04.02.04.08.0-0.5--
Maximum5.56.06.05.511.017.03.03.54.0
Standard deviation1.11.00.61.11.92.30.70.80.9
Source: IMF staff estimates.

Unweighted average of the scores for the structural indices for the exchange and trade systems (Tables 4.134.14).

The extent of reform is classified as low (score of 1–2), moderate (3–4), or high (5–6) relative to a specified notion of “best practices” (score of 5–6).

Source: IMF staff estimates.

Unweighted average of the scores for the structural indices for the exchange and trade systems (Tables 4.134.14).

The extent of reform is classified as low (score of 1–2), moderate (3–4), or high (5–6) relative to a specified notion of “best practices” (score of 5–6).

Table 4.13Indices of Structural Reform: Exchange System

(Six-point scale)1

ScoreCumulative ScoreChange in Score
1981–85

(I)
1986–90

(II)
1991–95

(III)
1981–85

(I)
1981–90

(I + II)
1981–95

(I + II + III)
II

over I
III

over II
III

over I
(Averages of country scores)
Africa4.24.75.64.29.014.50.50.81.3
CFA5.96.06.05.911.917.90.1-0.1
Non-CFA3.34.05.33.37.312.60.71.32.0
Asia3.84.35.33.88.013.30.51.01.5
Western Hemisphere2.84.05.52.86.812.31.31.52.8
Africa(Individual country scores)
CFA
Benin6.06.06.06.012.018.0---
Burkina Faso6.06.06.06.012.018.0---
Côte d’Ivoire6.06.06.06.012.018.0---
Equatorial Guinea6.06.06.06.012.018.0---
Mali6.06.06.06.012.018.0---
Niger5.06.06.05.011.017.01.0-1.0
Senegal6.06.06.06.012.018.0---
Togo6.06.06.06.012.018.0---
Non-CFA
Burundi5.05.05.05.010.015.0---
The Gambia5.05.06.05.010.016.0-1.01.0
Ghana2.04.06.02.06.012.02.02.04.0
Guinea3.05.06.03.08.014.02.01.03.0
Kenya4.05.05.04.09.014.01.0-1.0
Lesotho6.06.06.06.012.018.0---
Madagascar3.04.05.03.07.012.01.01.02.0
Malawi3.05.05.03.08.013.02.0-2.0
Mauritania3.04.04.03.07.011.01.0-1.0
Mozambique2.03.05.02.05.010.01.02.03.0
Sierra Leone4.03.06.04.07.013.0-1.03.02.0
Tanzania2.02.05.02.04.09.0-3.03.0
Uganda2.03.05.02.05.010.01.02.03.0
Zimbabwe2.02.05.02.04.09.0-3.03.0
Asia
Bangladesh2.02.04.02.04.08.0-2.02.0
Nepal3.04.05.03.07.012.01.01.02.0
Pakistan5.06.06.05.011.017.01.0-1.0
Sri Lanka5.05.06.05.010.016.0-1.01.0
Western Hemisphere
Bolivia2.06.06.02.08.014.04.0-4.0
Guyana3.03.05.03.06.011.0-2.02.0
Honduras4.03.06.04.07.013.0-1.03.02.0
Nicaragua2.04.05.02.06.011.02.01.03.0
Mean4.04.65.54.08.514.00.60.91.5
Minimum2.02.04.02.04.08.0-1.0--
Maximum6.06.06.06.012.018.04.03.04.0
Standard deviation1.61.40.61.62.83.21.01.11.3
Source: IMF staff estimates.

The extent of reform is classified as low (score of 1–2), moderate (3–4), or high (5–6) relative to a specified notion of “best practices” (score of 5–6).

Source: IMF staff estimates.

The extent of reform is classified as low (score of 1–2), moderate (3–4), or high (5–6) relative to a specified notion of “best practices” (score of 5–6).

Table 4.14Indices of Structural Reform: Trade System

(Six-point scale)1

ScoreCumulative ScoreChange in Score
1981–85

(I)
1986–90

(II)
1991–95

(III)
1981–85

(I)
1981–90

(I + II)
1981–95

(I + II + III)
II

over I
III

over II
III

over I
(Averages of country scores)
Africa2.83.94.92.86.711.61.11.12.1
CFA3.53.95.13.57.412.50.41.31.6
Non-CFA2.43.94.82.46.311.11.40.92.4
Asia2.52.54.02.55.09.0-1.51.5
Western Hemisphere3.54.05.83.57.513.30.51.82.3
Africa(Individual country scores)
CFA
Benin3.04.06.03.07.013.01.02.03.0
Burkina Faso3.03.05.03.06.011.0-2.02.0
Côte d’Ivoire3.04.05.03.07.012.01.01.02.0
Equatorial Guinea4.04.05.04.08.013.0-1.01.0
Mali3.03.05.03.06.011.0-2.02.0
Niger5.06.06.05.011.017.01.0-1.0
Senegal3.03.04.03.06.010.0-1.01.0
Togo4.04.05.04.08.013.0-1.01.0
Non-CFA
Burundi2.04.04.02.06.010.02.0-2.0
The Gambia3.04.05.03.07.012.01.01.02.0
Ghana2.04.06.02.06.012.02.02.04.0
Guinea2.06.06.02.08.014.04.0-4.0
Kenya2.02.06.02.04.010.0-4.04.0
Lesotho5.05.05.05.010.015.0---
Madagascar2.04.04.02.06.010.02.0-2.0
Malawi2.04.05.02.06.011.02.01.03.0
Mauritania3.04.05.03.07.012.01.01.02.0
Mozambique2.02.04.02.04.08.0-2.02.0
Sierra Leone2.04.05.02.06.011.02.01.03.0
Tanzania2.04.05.02.06.011.02.01.03.0
Uganda2.03.04.02.05.09.01.01.02.0
Zimbabwe3.04.03.03.07.010.01.0-1.0-
Asia
Bangladesh2.02.04.02.04.08.0-2.02.0
Nepal3.03.05.03.06.011.0-2.02.0
Pakistan2.02.03.02.04.07.0-1.01.0
Sri Lanka3.03.04.03.06.010.0-1.01.0
Western Hemisphere
Bolivia5.06.06.05.011.017.01.0-1.0
Guyana4.05.06.04.09.015.01.01.02.0
Honduras2.02.06.02.04.010.0-4.04.0
Nicaragua3.03.05.03.06.011.0-2.02.0
Mean2.93.74.92.96.611.50.81.22.0
Minimum2.02.03.02.04.07.0--1.0-
Maximum5.06.06.05.011.017.04.04.04.0
Standard deviation1.01.10.91.01.92.41.01.11.1
Source: IMF staff estimates.

The extent of reform is classified as low (score of 1–2), moderate (3–4), or high (5–6) relative to a specified notion of “best practices” (score of 5–6).

Source: IMF staff estimates.

The extent of reform is classified as low (score of 1–2), moderate (3–4), or high (5–6) relative to a specified notion of “best practices” (score of 5–6).

Table 4.15Indices of Structural Reform: Financial Sector

(Six-point scale)1

ScoreCumulative ScoreChange in Score
1981–85

(I)
1986–90

(II)
1991–95

(III)
1981–85

(I)
1981–90

(I + II)
1981–95

(I + II + III)
II

over I
III

over II
III

over I
(Averages of country scores)
Africa2.62.83.62.65.59.10.20.70.9
CFA2.42.73.32.45.18.40.30.60.9
Non-CFA2.82.93.72.85.79.40.20.81.0
Asia3.14.34.33.17.311.61.20.11.3
Western Hemisphere2.83.54.22.86.410.50.70.61.3
Africa(Individual country scores)
CFA
Benin2.32.22.82.34.57.3-0.10.60.6
Burkina Faso2.82.23.52.85.08.6-0.61.30.7
Côte d’Ivoire3.03.73.73.06.810.40.7-0.10.6
Equatorial Guinea2.22.12.02.24.26.3-0.1--0.1
Mali2.22.23.52.24.37.8-1.31.3
Niger2.12.24.12.14.48.50.11.92.0
Senegal2.43.03.62.45.48.90.60.61.2
Togo2.33.83.12.36.19.21.4-0.70.7
Non-CFA
Burundi2.12.13.52.14.37.8-1.41.4
The Gambia2.34.24.32.36.510.82.0-2.0
Ghana2.13.43.52.15.59.01.3-1.4
Guinea2.73.44.72.76.110.80.71.32.0
Kenya4.44.44.54.48.713.3-0.20.2
Lesotho2.52.42.42.54.97.3--0.1-0.1
Madagascar2.22.23.52.24.37.9-1.41.3
Malawi2.32.34.22.34.68.8-0.12.01.9
Mauritania2.32.33.62.34.68.2-1.31.4
Mozambique3.72.92.73.76.69.3-0.9-0.2-1.0
Sierra Leone2.82.84.12.85.69.7-0.11.31.2
Tanzania2.32.22.92.34.57.4-0.10.70.6
Uganda3.42.74.13.46.110.2-0.71.40.7
Zimbabwe3.73.74.33.77.411.7-0.60.6
Asia
Bangladesh2.34.64.62.36.911.52.20.12.3
Nepal3.03.73.83.06.710.50.70.10.8
Pakistan3.13.73.73.16.810.50.7-0.7
Sri Lanka3.85.15.23.88.914.11.30.11.4
Western Hemisphere
Bolivia2.23.84.82.26.010.81.61.12.6
Guyana3.64.24.03.67.811.80.6-0.20.4
Honduras3.13.84.03.16.910.80.70.20.9
Nicaragua2.52.43.82.54.88.7-0.11.51.4
Mean2.73.13.82.75.89.60.40.61.0
Minimum2.12.12.02.14.26.3-0.9-0.7-1.0
Maximum4.45.15.24.48.914.12.22.02.6
Standard deviation0.60.90.70.61.31.80.70.70.8
Source: IMF staff estimates.

The extent of reform is classified as low (score of 1–2), moderate (3–4), or high (5–6) relative to a specified notion of “best practices” (score of 5–6) or, for indicators of financial deepening, to an average for all developing countries (3–4).

Source: IMF staff estimates.

The extent of reform is classified as low (score of 1–2), moderate (3–4), or high (5–6) relative to a specified notion of “best practices” (score of 5–6) or, for indicators of financial deepening, to an average for all developing countries (3–4).

Table 4.16Indices of Structural Reform: Public Enterprise Sector

(Six-point scale)1

ScoreCumulative ScoreChange in Score
1981–85

(I)
1986–90

(II)
1991–95

(III)
1981–85

(I)
1981–90

(I + II)
1981–95

(I + II + III)
II

over I
III

over II
III

over I
(Averages of country scores)
Africa3.13.44.03.16.510.50.30.50.9
CFA3.03.43.93.06.310.20.40.61.0
Non-CFA3.23.54.03.26.610.60.30.50.8
Asia3.03.74.43.06.711.10.70.71.4
Western Hemisphere3.64.64.73.68.212.91.00.11.1
Africa(Individual country scores)
CFA
Benin2.43.24.52.45.610.10.81.32.1
Burkina Faso2.72.74.02.75.39.3-1.31.3
Côte d’Ivoire2.02.52.52.04.57.00.5-0.5
Equatorial Guinea2.83.63.62.86.410.00.8-0.8
Mali3.03.23.63.06.29.80.20.40.6
Niger4.74.74.74.79.314.0---
Senegal3.24.04.53.27.211.70.80.51.3
Togo3.03.04.03.06.010.0-1.01.0
Non-CFA
Burundi3.63.64.03.67.211.2-0.40.4
The Gambia3.03.25.33.06.211.50.22.12.3
Ghana2.43.23.52.45.69.10.80.31.1
Guinea2.02.74.02.04.78.70.71.32.0
Kenya4.04.44.54.08.412.90.40.10.5
Lesotho3.63.63.53.67.210.7--0.1-0.1
Madagascar4.84.04.04.88.812.8-0.8--0.8
Malawi4.84.84.84.89.614.4---
Mauritania4.04.04.04.08.012.0---
Mozambique2.02.02.82.04.06.8-0.80.8
Sierra Leone3.04.04.03.07.011.01.0-1.0
Tanzania2.52.52.72.55.07.7-0.20.2
Uganda2.83.33.62.86.19.70.50.30.8
Zimbabwe2.03.05.02.05.010.01.02.03.0
Asia
Bangladesh3.03.53.53.06.510.00.5-0.5
Nepal3.63.64.53.67.211.7-0.90.9
Pakistan3.05.36.03.08.314.32.30.73.0
Sri Lanka2.42.43.52.44.88.3-1.11.1
Western Hemisphere
Bolivia4.45.25.64.49.615.20.80.41.2
Guyana2.43.24.02.45.69.60.80.81.6
Honduras4.76.05.34.710.716.01.3-0.70.7
Nicaragua3.04.04.03.07.011.01.0-1.0
Mean3.23.74.23.26.811.00.50.51.0
Minimum2.02.02.52.04.06.8-0.8-0.7-0.8
Maximum4.86.06.04.810.716.72.32.13.0
Standard deviation0.91.00.90.91.82.50.60.60.8
Source: IMF staff estimates.

The extent of reform is classified as low (score of 1–2), moderate (3–4), or high (5–6) relative to an average for all developing countries (score of 3–4).

Source: IMF staff estimates.

The extent of reform is classified as low (score of 1–2), moderate (3–4), or high (5–6) relative to an average for all developing countries (score of 3–4).

Exchange and Trade System

The indicators in this area were chosen to capture the extent of protective barriers to external trade, through restrictions on the volume and distortions to the price of trade, as well as restrictions on the system of payments for these transactions. There are, however, several problems with the chosen indicators and some important omissions, the latter forced by data constraints. For example, with respect to tariffs, an ideal measure would be an average of tariffs weighted by the amount of domestic production protected by the tariff or effective rates of protection, which measure the protection to the net value added in the production of a certain commodity—both measures are difficult to compute in practice. For this analysis, data are typically available on unweighted average tariffs and, in some cases, tariffs weighted by imports. Generally, the former tend to be biased upward, while the latter tend to be biased downward, relative to domestic-output-weighted average tariffs. Average tariffs are supplemented with information on the number of tariff bands and the spread between maximum and minimum rates to obtain an overall picture of the tariff regime.

The restrictiveness of quantitative restrictions or nontariff harriers is also very difficult to measure. A relatively good indicator is the world-trade-weighted average of tariff lines covered by quantative restrictions, but data on such averages are extremely scarce. Here, the indicator used broadly measures the value of domestic production of goods for which imports or exports are covered by quantative restrictions.

Pricing and Marketing Reforms

Price controls and marketing arrangements for domestic products, imports, and exports were considered for four major commodities or groups of commodities—fertilizers, petroleum products, major foodstuffs, and agricultural commodities—on the grounds that together they probably give a reasonably accurate picture of the pervasiveness of state intervention in the economy’s trading activities.56 An indicator of the overall extent of price controls is also used to supplement this information.57 The degree of detail used for each commodity group was dictated by the availability of data. For example, for major foodstuffs and petroleum it was possible to assign separate scores to price controls and state intervention in marketing. For fertilizers, the available information permitted only a binary classification based on the existence (or not) of price controls and controls on marketing.

Financial Sector Development

The indicators of structural policies in the financial sector are chosen to measure the degree of financial sector development and financial repression, as well as certain institutional characteristics of financial markets. The degree of financial development is measured by a combination of the ratio of broad money to GDP to measure the extent of monetization; the ratio of private sector credit to GDP to proxy the private sector’s access to credit;58 and the ratio of currency to broad money to proxy the degree of financial intermediation.59 The interpretation of these aggregates is not always straightforward. For example, in countries with highly developed financial markets, the ratio of broad money to GDP could actually be lower than in less developed financial markets. Likewise, a high ratio of private sector credit to GDP could be symptomatic of weak monetary control and credit risk assessment.

Financial repression arises from a combination of controls on interest rates and the allocation of credit. In addition to the presence or absence of controls on interest rates and credit allocation, indirect indicators of financial repression include ex post real interest rates, and the spread between lending and deposit rates. As pointed out by De Gregorio and Guidotti (1992), there may be an inverted U-shaped relationship between real interest rates and economic growth. Depending on the form of financial repression and inefficiency, they can result in either negative or very high real interest rates, the latter possibly reflecting a high-risk premium. Similarly, both very low and very high spreads between lending and deposit rates can be symptomatic of distortions and inefficiencies, including a high proportion of nonperforming loans in banks’ balance sheets. The thresholds for scoring these indicators attempt to take such features into account.

Institutional aspects of the financial sector are measured by the degree of concentration (in terms of state-owned banks) in the banking system, the existence of an interbank market and markets for government or central bank securities, and a stock market. These are very crude measures of the general status of money and capital markets, chosen only because of severe data constraints; for example, the existence of an interbank market or a government securities market does not, by itself, ensure more efficient financial intermediation. Measures such as the total value of stock trades to GDP or details about the functioning of interbank markets are not uniformly available and therefore could not be considered. Data limitations also precluded the construction of a measure of the health of the banking system that could incorporate data on the quality of bank portfolios, and information on the strength of supervision and prudential regulation.

Public Enterprise Reforms

While many benefits stem from the activities undertaken by certain public enterprises, the available evidence suggests that most public enterprises do not successfully mobilize or efficiently utilize productive resources and that they tend to be less efficient than their private sector counterparts. Information on public enterprises and their role in economic activity is extremely scanty and unsystematic and is not comprehensive enough to trace clearly the changing role of public enterprises in the countries under review. Only a few multicountry studies of the role of this sector and progress with reforms have been conducted—notably, Swanson and Wolde-Semait (1989); Shirley and Nellis (1991); and World Bank (1994 and 1995).60

In an attempt to get as consistent a picture of the public enterprise sector as possible, the scoring of public enterprise reforms has been based on five indicators. Three indicators pertain to the public enterprise sector’s role in economic activity—its share in value added, in employment, and in total domestic credit; and two concern financial performance—the overall balance of public enterprises and the size of net financial transfers between the government and the public enterprise sector, both expressed as a share of GDP.61 Not all of the indicators are available for all countries, forcing the use of an average of the scores based on available indicators for each country and each period; also, the coverage of public enterprises varies considerably across countries and is typically partial.62

Is the share of public enterprises in value added a reliable measure of the importance of this sector? In principle, price increases for the output of public enterprises could offset the effects of divestiture, so that a public enterprise sector that is smaller by other measures could remain unchanged in terms of value added. In most of the countries under review, however, the evidence suggests that price increases for public enterprise output did not, in general, keep pace with inflation and that divestment large enough to affect the overall size of the sector has been a relatively recent phenomenon in most countries where it has begun. The same justification is germane to the use of the public enterprise sector’s share of formal sector employment.

Nevertheless, one might be inclined to attach more weight to the two variables measuring the sector’s financial health, on the grounds that a large but profitable public enterprise sector gives rise to fewer distortions than one that is large but inefficient and unprofitable. However, aside from the fact that data on these measures are limited in their availability, there are also severe problems with the data on overall balances and net financial flows. Accounting in the public enterprise sector in many of the countries is extremely poor, and in many cases no reliable estimates can be made of the profitability of the sector.

References

    AbedGeorge T. and others1998Fiscal Reforms in Low-Income Countries: Experience Under IMF-Supported Programs Occasional Paper 160 (Washington: IMF).

    • Search Google Scholar
    • Export Citation

    AgarwalaRamgopal1983“Price Distortions and Growth in Developing Countries,”World Bank Staff Working Paper 575 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    CaprioGerard and DanielaKlingebiel1996“Bank Insolvencies: Cross-Country Experience,”Policy Research Working Paper 1620 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    De GregorioJosé and PabloGuidotti1992“Financial Development and Economic Growth,”IMF Working Paper 92/101 (Washington: IMF).

    DenizerCevdetMartha DeMelo and AlanGelb1996“From Plan to Market: Patterns of Transition,”Policy Research Working Paper 1564 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    EasterlyWilliamCarlos A.Rodríguez and KlausSchmidt-Hebbeleds.1994Public Sector Deficits and Macroeconomic Performance (New York: Oxford University Press).

    • Search Google Scholar
    • Export Citation

    European Bank for Reconstruction and Development1995Transition Report (London).

    FischerStanleyRatnaSahay and Carlos A.Végh1996“Stabilization and Growth in Transition Economies: The Early Experience,”Working Paper 96/31 (Washington: IMF).

    • Search Google Scholar
    • Export Citation

    GarciaGillian1997“A Framework for Analysis and Assessment,” in William E.Alexander and otherseds.Systemic Bank Restructuring and Macroeconomic Policy (Washington: IMF).

    • Search Google Scholar
    • Export Citation

    International Monetary FundExternal Evaluation of the ESAF: Report by a Group of Independent Experts (Washington: 1998).

    KnackStephen and PhilipKeefer1995“Institutions and Economic Performance: Empirical Tests Using Alternative Measures of Institutions,”IRIS Working Paper 109 (College Park, Maryland: Center for Institutional Reform and the Informal Sector, University of Maryland).

    • Search Google Scholar
    • Export Citation

    LevineRoss and SaraZervos1993“Looking at the Facts: What We Know About Policy and Growth from Cross-Country Analysis,”Policy Research Working Paper 1115 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    LindgrenCarl-JohanGillianGarcia and Matthew I.Saal1996Bank Soundness and Macroeconomic Policy (Washington: IMF).

    NunbergBarbara and John R.Nellis1995“Civil Service Reform and the World Bank,”Discussion Paper 161 (Washington: World Bank).

    SachsJeffrey D. and Andrew M.Warner1995“Economic Reform and the Process of Global Integration,”Brookings Papers on Economic Activity: I (Washington: Brookings).

    • Search Google Scholar
    • Export Citation

    SchadlerSusanFranekRozwadowskiSiddharthTiwari and DavidRobinson1993Economic Adjustment in Low—Income Countries: Experience Under the Enhanced Structural Adjustment Facility Occasional Paper 106 (Washington: IMF).

    • Search Google Scholar
    • Export Citation

    ShirleyMary M. and John R.Nellis1991“Public Enterprise Reform: The Lessons of Experience,”EDI Development Studies (Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    SwansonDaniel and TeferraWolde-Semait1989“Africa’s Public Enterprise Sector and Evidence of Reforms,”World Bank Technical Paper 95 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    World Bank1994Adjustment in Africa: Reforms Results and the Road Ahead (New York: Oxford University Press).

    World Bank1995Bureaucrats in Business: The Economics and Politics of Government Ownership (New York: Oxford University Press).

    World Bank1996World Development Report 1996: From Plan to Market (New York: Oxford University Press).

1Another measure of the fiscal adjustment effort is the primary current balance (excluding grants); this excludes net interest payments as well as capital expenditure, which tends to be driven by the availability of foreign grants. As with the overall balance, developments in this balance were similar to those in the primary balance except for a number of CFA African countries (Côte d’Ivoire, Niger, and Togo) in which the deterioration in the primary current balance was far greater than in the primary balance before the 1994 devaluation of the CFA franc. The implied compression of capital outlays relative to current spending reflected an inability to cut the latter in the face of a marked weakening in revenue performance, and hence also an inability to generate sufficient domestic counterpart funds for investment.
2For Honduras, medium-term fiscal targets (as established at the outset of the three-year ESAF arrangements) were available only for the central government. Hence, for consistency between the analysis of program performance and longer-term fiscal developments, the discussion in this section is based on the central government’s accounts.
3The discussion in this section includes 34 of the 36 countries covered by the review. Guyana is excluded because of likely inaccuracies and frequent revisions in the GDP data. Equatorial Guinea is excluded because, unlike all the other countries, the fiscal accounts do not include foreign-financed capital expenditures.
4In view of the frequent revisions to past data, particularly for GDP, made subsequent to the formulation of programs, the objectives are considered in terms of the targeted changes relative to the situation believed to prevail at the inception of the program. The outcomes or actual changes are measured using the most recent estimates for the preprogram and outturn years. Comparisons of targets with outcomes are made in terms of changes, not levels.
5Unless otherwise noted, the discussion of program targets and comparison with outcomes is based on a sample of 28 (out of 36) countries, and 47 (out of 68) SAF/ESAF arrangements. Programs were excluded from the sample because three-year fiscal targets were not provided or because of data inadequacies (severe discontinuities in time-series or unreliable GDP figures). The excluded countries are Equatorial Guinea (CFA Africa); Madagascar, Sierra Leone, and Tanzania (non-CFA Africa); Guyana (Western Hemisphere); and Cambodia, Kyrgyz Republic, and Mongolia (transition economies). In terms of programs, the coverage of the restricted sample is 15 out of 17 programs in CFA Africa, 18 of 28 in non-CFA Africa, 8 of 8 in Asia, 3 of 8 in the Western Hemisphere, and 3 of 7 in the transition economies.
6The targets used in Abed and others (1998) are those set either at the beginning of each annual SAF/ESAF arrangement or, if subsequently revised, those set at the time of the midterm review.
7Targeted revenue adjustments fell within the range of 2 to –2 percentage points of GDP in over 80 percent of programs, compared with about 40 percent of programs for changes in expenditure.
8When averaged only over programs that sought spending cuts, the average reduction is almost 4 percentage points of GDP.
9The discussion of current and capital spending targets is based on a subsample of 37 programs covering 24 countries for which three-year targets are available.
10A simple regression of the form described in Box 4.1 was estimated. The resulting regression had a coefficient of determination (R2) of 0.6, and the targeted adjustment in current noninterest spending was significantly inversely related to the preprogram level (with a confidence interval of 99 percent).
11The correlation between deviations from targeted changes of noninterest expenditures and the primary balance was positive and statistically significant at the 5 percent level. By contrast, the correlation between deviations from targeted changes of revenues and the primary balance was statistically insignificant even at the 10 percent level.
12On average, the size of expenditure overruns was small relative to revenue shortfalls (see Table 4.2). However, in those programs in which the deficit target was undershot, expenditure overruns were most common (see Figure 4.4).
13In a simple bivariate regression, the size of the deviation from the targeted change in each of the fiscal variables was found to be inversely related to the targeted change at the 5 percent significance level for the primary balance and expenditures, and at the 10 percent significance level for revenues. The R2 of each regression estimate, was modest: 0.2 for the primary balance, 0.3 for spending, and 0.05 for revenue.
14Adjustment was considered to be front-loaded if more than one-third of the three-year adjustment was targeted in the first program year.
15Correlations between the degree of front-loading and both the targeted adjustment and deviation of outcomes from targeted changes for revenue and expenditure were statistically insignificant. With regard to the primary balance, the extent of front-loading was negatively correlated with the targeted adjustment, and positively correlated with the deviation of outcomes from targeted changes, at a 10 percent level of statistical significance.
16Because of data availability, this section considers only the nontransition countries.
17This approach was taken in view of the relatively high public debt-to-GDP ratios of the ESAF countries—on average 135 percent in 1995, with an annual interest bill equivalent to 15–20 percent of government spending; an alternative computation would be to estimate the fiscal balance compatible with stability of the debt-to-GDP ratio at some point in time.
18By implication, the primary balance was greater than the “sustainable” balance required to keep the debt-to-GDP ratio constant.
19In 11 of these countries, the primary balance less grants and seigniorage revenue was negative, thereby contributing to a rise in the debt ratio. However, the debt ratio was on a declining path because the nominal growth rate of the economy exceeded the sum of the interest rate and the exchange-rate-induced increase in the domestic currency value of external debt, and the primary deficit was small. In this situation, a country can grow out of its debt.
20The baseline comprises the values for the relevant variables for the “average” ESAF country. At the same time, however, one might expect the depreciation to have a beneficial effect on the rate of output growth.
21Also, for lack of data the analysis does not take account of the stock of the government’s domestic nonbank debt, which in some of the countries is substantial.
22These data refer to exchange rate arrangements officially reported to the IMF and published in the IMF’s International Financial Statistics (IFS) as involving either a pegged rate or (in one case only) a rate with “flexibility limited in terms of a single currency.”
23Classification of an exchange rate arrangement as floating need not imply that the official exchange rate is a market-clearing one.
24For example, a country with a relatively clean float would be classified as “independently floating,” yet Bolivia was formally included in this category while it followed a real exchange rate rule.
25The presence of an exchange rate anchor is based on program documents at the outset of three-year programs rather than the classification published in IFS.
26The REER measures reflected in Figure 4.7 are based on domestic and foreign consumer price indices and official exchange rates; they should be interpreted as being merely suggestive of developments in the relative prices of traded and nontraded goods.
27Figure 4.9 (bottom panel) indicates a share for negative rates of more than one-half. However, if not for the 1994 CFA franc devaluation, it is likely that five more countries would have averaged positive rates. Also, if transition economies were included, a somewhat more favorable impression of the recent period would result; these economies had mostly positive real deposit rates by the early to mid–1990s. Transition economies are excluded from Figure 4.9 because earlier data were not generally available.
28The main exceptions to this were the rates of return influenced by the price-level jump associated with the 1994 devaluation.
29A number of data problems complicate an assessment of this question, especially because for most countries the private sector includes public enterprises, and different definitions of the government, public, and private sectors are used across countries. The discussion here emphasizes trends over time; comparisons across countries may he less meaningful.
30In the non-CFA African countries, the sharpest drop occurred in the late 1980s; the decline was more gradual in the Asian countries.
31The choice of a long-term perspective for this analysis effectively precluded the transition economies, most of which began reforms only late in the period under review.
32An index of the structural aspects of the composition and administration of taxation and expenditure was not compiled because of the difficulty of defining clear objective benchmarks or widely accepted “best practices” by which to calibrate progress. Fiscal structural issues are, however, discussed in Abed and others (1998).
33The three periods, however, do not as a rule coincide with the presence or absence of SAF/ESAF-supported adjustment programs.
34The design of social policies in programs has been examined in the external evaluation of the ESAF (IMF, 1998).
35Of these, four had not yet begun structural adjustment under SAF/ESAF-supported arrangements.
36Of these, Zimbabwe did not have an IMF-supported program until the early 1990s.
37An exchange system with surrender requirements together with an open market-based mechanism for allocating foreign exchange was considered to be free of distortions. In total, nine countries had no surrender requirements; in the remainder (except Madagascar), surrender requirements were combined with no restriction on the allocation of foreign exchange.
38Among the ESAF transition economies in this review, the Kyrgyz Republic and Mongolia have also accepted these obligations.
39A “moderately restrictive regime” is defined as comprising average tariffs of about 20–25 percent and maximum tariffs of 30–40 percent, with few exemptions to the maximum rate and few exemptions overall.
40During this episode, however, measured ex post real interest rates were probably a poor proxy for the unobservable (but more economically relevant) ex ante real interest rates.
41Government securities are not permitted in the CFA franc zone.
42The extent to which these markets played an important role in promoting efficient financial intermediation would depend in part on the volume of market transactions, for which comprehensive data are not available.
43Banks were liquidated in a quarter of the countries, mainly in the CFA countries.
44Under a management contract, management is contracted out for a performance-based fee, while under a performance contract managers are still appointed by the government.
45The four indicators comprise the real interest rate, the interest rate spread, the share of broad money in GDP, and the share of currency in broad money. The three periods correspond to prereform, first reform phase, and second reform phase. In some countries, the onset of reform is sufficiently recent to make meaningless the notion of a second phase. See also Caprio and Klingebiel (1996) and Garcia (1997).
46This picture of the public enterprise sector is based on fragmentary data. Because of the lack of comprehensive data on public sector activities for most countries, the structural index for the public enterprise sector should be interpreted cautiously.
47“Less severe” indicates values of the constructed index close to or below the average for developing countries.
48The strategy adopted in the transition economies differed significantly from the other countries: reforms were undertaken early in the transformation process and involved a sweeping privatization of small and medium-sized public enterprises.
49Again, “severe” means values for the composite public enterprise reform index that are worse than the average for all developing countries.
50Whereas public enterprise salary levels have often been excessive, the remuneration of upper- and middle-level civil service managers has typically been low compared with the private and parastatal sectors. In addition, salaries of upper-level managers were often severely compressed relative to lower-paid civil servants.
51In over half of the countries, explicit limits were placed on the size of the wage bill, but in many instances these were overrun.
52Ratings from several agencies have been compiled in an International Country Risk Guide Database (ICRGD) at the Center for Institutional Reform and the Informal Sector at the University of Maryland; see Knack and Keefer (1995).
53An index of the structural aspects of the tax and expenditure systems and the composition of taxation and spending was not compiled because of the difficulty of defining clear objective benchmarks or widely accepted “best practices.” Structural fiscal issues are, however, discussed in Abed and others (1998).
54For those indicators for which time-series data are available (such as interest rates, monetary and credit aggregates, public enterprise sector balances, financial flows between enterprise and governments), five-year averages were used. In most other cases, the observations generally were made at or close to the end of each five-year period. In a few cases, mostly for public enterprise policies and pricing policies, where data availability is scant, any available data relevant to the five-year period were used.
55For some of the subcomponents that were combined to construct the indices, the available information permitted only a binary score for the presence or the absence of the distortion.
56Prices and marketing of utilities are not considered here; in all of the countries in the sample there were public sector monopolies with regulated prices—as was also the case in most developing countries.
57The inclusion of the indicator of overall price controls introduces some double counting, but the justification for its inclusion is that, because of data availability, the major foodstuffs category typically captures only those goods that involve budgetary subsidies, whereas the overall category aims to get at a broader picture of the pervasiveness of price controls.
58In 17 of the 30 countries, private sector credit includes credit to public enterprises.
59The use of five-year averages would tend to smooth cyclical variations in these ratios.
60Much of the data used in the present study are taken from these sources. Information for the most recent years are taken directly from IMF and World Bank country reports.
61Net transfers to public enterprises are defined as the difference between all financial flows from the government to the enterprises, including government loans, equity contributions and subsidies, and all flows from enterprises to the government, including dividends and taxes. The main source for these data is World Bank (1995).
62Data were available on public enterprises’ investment and external debt but were not used here because of difficulties in interpreting these data. External debt statistics can grossly underestimate the amount of public enterprise borrowing, because governments often incur the debt (on paper) and on-lend the funds to the public enterprises. Likewise, public enterprise investment is typically not systematically monitored, and its share in total investment is difficult to compare across countries.

    Other Resources Citing This Publication