Chapter

II Record of Fiscal Adjustment

Author(s):
Adam Bennett, Louis Dicks-Mireaux, Miguel Savastano, María Carkovic S., Mauro Mecagni, Susan Schadler, and James John
Published Date:
September 1995
Share
  • ShareShare
Show Summary Details
Author(s)
Adam Bennett,, Maria Carkovic, and Louis Dicks-Mireaux 

Unsustainable financial imbalances in the public sector were the source of developments that precipitated the need for IMF support in virtually all the 36 countries with Fund arrangements approved between mid-1988 and mid-1991 (see Chart 1-1). Fiscal adjustment was, therefore, at the heart of every program. Typically, large public sector borrowing requirements had led to combinations of unmanageable external current account deficits, heavy domestic and foreign indebtedness, reliance on arrears, crowding out of private activity, and high inflation. The mix of these consequences in any particular country depended most importantly on the size and duration of the fiscal imbalances and how they were financed.1

In broad terms, the countries under review faced one of three types of difficulties. First, many governments had accumulated sizable external debt in the past and bore a heavy interest burden. Many of these were countries that had had one or several previous arrangements with the IMF in the past decade, during which considerable progress had been made in adjusting the primary fiscal account, but the continuing large interest burden meant that the overall deficit could not be funded through normal external borrowing or noninflationary domestic finance. Second, many other countries had seen an increase in their fiscal deficits to unsustainable levels relatively recently; although they were often not heavily indebted, their access to international capital markets had never been on a large scale so the need to adjust quickly was urgent. Third, in Central Europe, problems often related less to the size of initial fiscal deficits than to the excessive role of the state in the economy.

This paper addresses five basic questions about fiscal adjustment during the IMF-supported programs of the 36 countries under review. First, to what degree were the fiscal strategies and targets in the programs tailored to the specific circumstances of each country? Second, what was the record of success in meeting the program targets? Third, what were the financing implications of the resulting borrowing requirements? Fourth, how were fiscal developments monitored during the arrangements and were monitoring methods appropriate? Fifth, what was the role of fiscal developments in programs that went off track?

To an important extent, answers to these questions were limited by the infrequency of analyses of the sustainability of targeted fiscal positions in program documents. Such analyses would have provided a standard for assessing the adequacy of fiscal adjustments programmed for each country over the medium term. Without such an analysis, it is possible to measure how closely targets were observed, but not how much of the movement toward a sustainable fiscal position programs aimed to accomplish.2 Depending, among other things, on the past and prospective sources of financing, such an assessment could take several forms: for countries that financed their deficits through market borrowing, the calculation of a fiscal position consistent with a stable and manageable ratio of government debt to GDP; for countries that received significant concessional financing or grants, the calculation of the fiscal position consistent with sustainable inflows of foreign financing; and for countries that relied heavily on monetary financing, the analysis of fiscal positions consistent with acceptable levels of inflation.

Many reasons can be adduced for why such analyses have not been undertaken: lack of data (particularly for domestic debt); sensitivity of both creditor and debtor governments to the inevitable, continuous revisions of such calculations; and the crisis environment in which IMF-supported programs are frequently agreed upon. Nevertheless, sharpening and making explicit the standard for medium-term fiscal adjustment should be a priority in future program design.

Coverage of the Accounts

In any study of this kind, differences in the coverage of the fiscal accounts and in the definitions of fiscal aggregates make it difficult to compare and draw inferences about fiscal performance.3 In principle, the fiscal accounts targeted and monitored should have the broadest coverage possible. In practice, the coverage has been far narrower than would have been desired—primarily because of a lack of reliable data in many countries on public entities outside the central government. For the Central European countries, assessing the fiscal position was especially difficult because at the outset of reform virtually the whole enterprise sector was effectively part of the public sector, and enterprises performed many functions typically carried out by governments in market economies.4 Strictly speaking, this difference in coverage prohibits intercountry comparisons of levels of fiscal balances. In fact, when coverage of the accounts is narrow, even within-country comparisons over time can be misleading because items can be shifted among levels of government. Owing to the want of fully consistent data, the comparisons made in this paper need to be recognized as flawed.

Developments in above-the-line items in the fiscal accounts are viewed in terms of the widest definition of the nonfinancial public sector for which comprehensive revenue and expenditure data exist—in 30 of the 36 countries under review, the central government only (see Appendix Table 2-A1 for definitions of the nonfinancial public sector). Omitted from most accounts, therefore, are public enterprises and extrabudgetary funds, which are known to have represented an important part of the overall public sector in all the countries. Despite their importance, separate data for public enterprises exist for only 15 countries and on a comparable consolidated basis (with a classification of revenues and expenditures) for only 6 countries. In these 6, data on the nonfinancial public sector are employed in this study.

There were several other differences in accounting practices among the countries. The accounts above the line are generally on a commitments basis, but for eight countries they are on a cash basis. Four of the countries undertook privatizations of significant value during the period reviewed: in Venezuela (1991) and Jamaica (1986–92) the receipts were recorded above the line as revenues, in Czechoslovakia (1991–92) they were recorded off budget, and in Mexico (1991–92) they were recorded below the line. Grants are considered a financing item in this study, although program documents tended to place grants above the line. In analyzing the financing of the fiscal balance, the widest definition for which comprehensive financing data exist has been employed for each country (Appendix Table 2-A1).5 In general, this entails a broader coverage of the non-financial public sector than for the above-the-line analysis.

Another important issue is the extent to which the quasi-fiscal losses of the central bank or even other publicly owned financial institutions and enterprises are incorporated in the fiscal accounts. Quantification often proved difficult. Some forms of quasi-fiscal losses, such as direct interest rate subsidies, are comparatively easy to measure, but others are almost impossible to quantify.6 Central bank losses were known to have been important in 11 countries and were quantified in 8 (Table 2-1). In 6 countries they were also subject to performance criteria by virtue of their inclusion in the relevant definition of the public sector deficit.

Overall and Primary Balances

In most programs, the focus for fiscal policy is on the overall balance of the central government or broader public sector. As a concise statement of the imbalance between government saving and investment, the overall position is an important indicator of fiscal pressures on the external current account and of the crowding out of private sector activity. Reflecting these linkages, which are critical to the aims of most programs, this study focuses on the overall position as the main indicator of fiscal adjustment.

It would be wrong, however, to suggest that the current account and private sector activity are the only final objectives of fiscal adjustment, that there is a one-to-one relationship between the overall deficit (and its financing) and even these final objectives, or that the causality is uni-directional. Other important final objectives require different perspectives on fiscal adjustment. Specifically, because interest payments are often beyond the control of current policies, the primary deficit—that is, the overall deficit excluding net interest payments—can be a better measure of the fiscal adjustment effort and the that are not explicitly considered are government saving (which indicates the government’s contribution to total resources for investment), the domestic deficit (which shows the government’s claim on domestic resources), and the operational deficit (which highlights the possibly distortive effects of high inflation on the measurement of the overall deficit). Some improvements in the structure of taxation or spending may not even show up in any measured fiscal aggregate but still entail significant gains in economic efficiency.

Initial Conditions

Most of the countries entered into the arrangements under review with large fiscal deficits. The four countries that had surpluses—Algeria, Czechoslovakia, Uruguay, and Yugoslavia—did not include in the accounts important public or quasi-public sector operations that were known to be a major financial drain on the public sector broadly defined. The average overall deficit for countries outside Central Europe was 8 percent of GDP: balances ranged from the small surpluses noted above to deficits as high as 31 percent in Brazil and over 20 percent in Zaїre and Egypt (Table 2-2).7 In Central Europe initial positions varied from near-balance (Czechoslovakia, Hungary, and Romania) to large deficits (Bulgaria and Poland).

Table 2-1.Central Bank Losses and Performance Criteria
1.Countries in which central bank losses are presently recorded in IMF documents as having existed at the time of the arrangement
Argentina

Costa Rica 1

Ecuador 1

El Salvador 1

Guatemala

Guyana

Honduras 2

Jamaica

Philippines

Uruguay

Venezuela

2.Countries in which the central bank was included in the definition of the public sector monitored under the arrangements
Argentina

Guatemala

Guyana

Honduras 2

Jamaica

Philippines

Uruguay

Venezuela
3.Countries in which the central bank was included in the definition of the public sector subject to performance criteria
Argentina

Honduras 2

Jamaica

Nigeria 3

Uruguay

Venezuela
Source: IMF reports.

Countries with central bank losses not identified at time of arrangement approval.

1990 only.

1989, interest payments only.

Source: IMF reports.

Countries with central bank losses not identified at time of arrangement approval.

1990 only.

1989, interest payments only.

Table 2-2.Fiscal Adjustment: Pre-Program Year1(In percent of GDP)
Interest PaymentsPrimary BalanceRevenuesNoninterest Expenditures
Overall Balance2CurrentCapital
All countries excluding Central Europe–8.2 (–8.4)6.0–2.221.116.37.0
Countries with several previous arrangements3–7.7 (–8.1)7.2–0.519.214.85.0
Countries with one previous arrangement–8.8 (–8.8)7.1–1.626.117.710.1
New users4–8.7 (–8.5)3.0–5.720.017.87.8
Central European countries5–3.5 (–5.4)3.3–0.250.543.77.0
Source: IMF staff estimates.

The pre-program year is defined as the year preceding the first arrangement. Data are as estimated at the time of the first arrangement and are the average of the countries in each group. Data exclude Guyana, where GDP is subject to serious measurement problems, and Brazil, Costa Rica, Pakistan, Papua New Guinea, Poland, and Yugoslavia, for which interest payments are not separately identified.

Full sample estimates for the overall balance including Brazil, Costa Rica, Pakistan, Papua New Guinea, Poland, and Yugoslavia are shown in parentheses.

Excluding Brazil, Costa Rica, and Yugoslavia, for which interest payments are not separately identified.

Excluding Pakistan and Papua New Guinea, for which interest payments are not separately identified.

Source: IMF staff estimates.

The pre-program year is defined as the year preceding the first arrangement. Data are as estimated at the time of the first arrangement and are the average of the countries in each group. Data exclude Guyana, where GDP is subject to serious measurement problems, and Brazil, Costa Rica, Pakistan, Papua New Guinea, Poland, and Yugoslavia, for which interest payments are not separately identified.

Full sample estimates for the overall balance including Brazil, Costa Rica, Pakistan, Papua New Guinea, Poland, and Yugoslavia are shown in parentheses.

Excluding Brazil, Costa Rica, and Yugoslavia, for which interest payments are not separately identified.

Excluding Pakistan and Papua New Guinea, for which interest payments are not separately identified.

Chart 2-1.Initial Fiscal Positions1

(In percent of GDP)

Source: IMF staff estimates.

1 Period t refers to the first year of the first arrangement under review for each country, excluding Central European countries, for which data in the periods prior to their first arrangement are not comparable.

2 Cameroon, El Salvador, Guatemala, Honduras, Jordan, Trinidad and Tobago, and Venezuela.

3 Argentina, Ecuador, Haiti, Jamaica, Madagascar, Mali, Mexico, Morocco, the Philippines, Uruguay, and Zaїre.

4 Algeria, the Congo, Egypt, Gabon, Nigeria, and Tunisia.

To a large degree, a country’s initial fiscal position and the adjustment problem that it faced reflected the length and duration of recent adjustment efforts. Thus, countries that had had at least two arrangements with the IMF in the previous five years had typically reduced their primary deficits in earlier years and started the programs under review from relatively strong primary positions (Chart 2-1). Apart from Mali and Zaїre, the initial primary positions of these countries were close to balance or even in surplus. Most countries that had had one arrangement in the past five years had achieved sizable reductions in primary deficits to relatively low levels. In both of these groups, however, heavy interest burdens meant that overall deficits averaged 8–9 percent of GDP. Countries outside Central Europe that had not recently used IMF resources tended to have relatively high primary deficits that had, on average, risen over the three years prior to their arrangements. Their interest burdens, however, were generally lighter than in countries that had taken advantage of large-scale market access in the late 1970s and early 1980s. In Central Europe, there was considerable diversity in initial positions: except in Poland and Bulgaria, primary positions were close to balanced; interest payments were also small except in Bulgaria.8

Several aspects of the fiscal imbalances differed among the various country groupings. First, most of the countries that had had one or several previous arrangements with the IMF had unsustainably high interest (domestic or external) burdens. For many of these countries, large primary deficits in the late 1970s and early 1980s had been funded by external borrowing. With the curtailment of access to credit markets and, for many, a sizable deterioration in the terms of trade in the mid-1980s, growth had slowed sharply, leaving high ratios of interest to GDP. In general, there had been relatively little change on average in interest payments relative to GDP since the mid-1980s.

Second, many countries that had had one previous arrangement with the IMF had relatively high non-interest expenditures. Most of these were oil producers that had increased spending, particularly for investment, on the strength of oil prices and the external borrowing they allowed, but then found it difficult to scale back as oil prices fell. (The ratio of capital spending to GDP in the pre-program year for these countries, however, is boosted by a large once-off transfer to financial institutions in Egypt.)9

Third, in many of the low-income countries, ratios of revenue to GDP were quite low owing to the structural weaknesses of the tax system.10 Very high inflation rates in a few countries had also weakened revenue ratios through the erosion of real tax revenues between accrual and collection—the Oliviera-Tanzi effect. Such countries were concentrated among those with several previous arrangements and the new users.

Fourth, in Central Europe, noninterest current expenditures and accordingly revenues relative to GDP were about twice the average level in the other countries. This reflected, in general, heavy spending on subsidies and social services.

This diversity of initial conditions suggests several groupings that could be followed in reviewing fiscal policy; beyond those based on the stage in the adjustment process, groupings of heavily indebted countries, low-income countries, high-inflation countries, and oil exporting countries suggest themselves. A critical difficulty with each of these groupings, except that of the stage in the adjustment process, is the sizable overlap in their coverage: for example, the oil exporters include several low-income and heavily indebted countries. Such overlaps prevent self-contained groupings and weaken the distinctions between groupings. In fact, after some experimentation with different groupings, it was clear that those based on stages in the adjustment process gave the most informative results.

Program Objectives

Ideally the programs under review would all have a medium-term time horizon. Medium-term projections, however, were provided for only about half of the countries reviewed; the other half had only one-year-ahead targets. This discrepancy raises the question of how to consider targets when many countries had several contiguous or multiyear programs during the period under review. There are two possible approaches. The first, employed in this study, is to compare the average of the targets set (over the course of the programs) for each program year with the original estimates for the pre-program year.11 This approach measures the average fiscal effort vis-à-vis the initial position but does not capture responses to slippages or overperformance during multiyear arrangements. The second approach is to compare targets for each program year with the position estimated for the preceding year. This approach permits consideration of the average annual change targeted, but the averaging process disguises the strength of adjustment efforts in multiyear programs when adjustment is frontloaded. An additional issue is whether the emphasis should be on programs or countries. Examining differences between average program targets and pre-program positions gives each country an equal weight in the analysis, regardless of the length of IMF involvement. Averaging individual program year targets shows the intentions of the typical program but gives a larger weight to countries that had several arrangements (or multiyear arrangements) than to those that had only one program year.

On average, outside Central Europe, countries targeted reductions in fiscal deficits relative to GDP of some 3 percentage points over an average 2–2½ years of IMF involvement (Table 2-3). The range of targets around this average tended to be related to the size of the initial deficit—the larger the deficit the larger the targeted cut (Chart 2-2).12 For countries with more than one program year, targeted changes in the deficit tended to be largest in the first program year.

While most countries targeted reductions in the fiscal deficit, six allowed for a deterioration in the fiscal accounts. Two of these (Uruguay and Yugoslavia) started from overall surpluses, and the envisaged worsening of the accounts was modest. In Romania, the fiscal position was in balance, but the reform process (in particular the loss of enterprise “profit” remittances to the state) was expected to push it into a small deficit. The significant widening of the deficit programmed for Brazil in 1988 largely reflected “water under the bridge”—there had been a sharp widening in the deficit in the first half of the year that planned corrective measures were not expected to offset for the year as a whole. For the Congo, the projected increase in borrowing by the central government was related to the restructuring of public enterprises involving intra-public-sector transactions. In Tunisia, the projected increase was due to spending associated with locust invasions and drought.13

On average, interest payments relative to GDP were expected to be stable. Several countries with high external interest burdens—concentrated among those with one or more previous arrangements—were engaged in negotiations for debt relief during their programs. Lower interest burdens that would result, however, were typically not reflected in targets except in multiyear arrangements that surpassed the completion of such operations (Mexico).14 Some countries (Argentina and Mexico) looked for significant reductions in the domestic interest burden too: in Argentina by redenominating domestic debt into U.S. dollar debt and in Mexico by lowering inflation and thereby domestic interest rates. Reschedulings, in contrast, tended to stretch out amortization but to increase the scheduled interest burden as late interest was either paid or added to debt. Back interest payments, however, were also not generally reflected in targets. In Cameroon and Côte d’Ivoire domestic debt restructurings were planned and involved a rise in the interest burden. Five other countries (Egypt, Honduras, Jordan, Trinidad and Tobago, and Venezuela) also expected interest payments to rise significantly, reflecting in part exchange rate devaluations prior to the program. The recent or planned introduction of treasury bill auctions in most of these countries was also expected to raise interest costs.

Table 2-3.Fiscal Targets: Change from Pre-Program Year1(In percentage points of GDP)
Interest PaymentsPrimary BalanceRevenuesNoninterest Expenditures
Overall Balance2CurrentCapital
All countries excluding Central Europe (2.3)3.4 (2.7)–0.13.31.7–0.8–0.8
Countries with several previous arrangements (2.5)33.6 (2.5)–1.62.01.7–0.50.2
Countries with one previous arrangement (2.2)2.9 (2.9)1.14.11.8–1.0–1.3
New users (2.0)43.5 (3.0)1.44.91.6–1.3–2.0
Central European countries (1.4)52.2 (4.4)–0.12.0–10.1–8.6–3.5
Source: IMF staff estimates.

Average of each country’s program target (or target average if multiyear program). Data exclude Guyana, where GDP is subject to serious measurement problems, and Brazil, Costa Rica, Pakistan, Papua New Guinea, Poland, and Yugoslavia, for which interest payments are not separately identified. Average number of program years per country are shown in parentheses next to group category.

Full sample estimates for the overall balance including Brazil, Costa Rica, Pakistan, Papua New Guinea, Poland, and Yugoslavia are shown in parentheses.

Excluding Brazil, Costa Rica, and Yugoslavia, for which interest payments are not separately identified.

Excluding Pakistan and Papua New Guinea, for which interest payments are not separately identified.

Source: IMF staff estimates.

Average of each country’s program target (or target average if multiyear program). Data exclude Guyana, where GDP is subject to serious measurement problems, and Brazil, Costa Rica, Pakistan, Papua New Guinea, Poland, and Yugoslavia, for which interest payments are not separately identified. Average number of program years per country are shown in parentheses next to group category.

Full sample estimates for the overall balance including Brazil, Costa Rica, Pakistan, Papua New Guinea, Poland, and Yugoslavia are shown in parentheses.

Excluding Brazil, Costa Rica, and Yugoslavia, for which interest payments are not separately identified.

Excluding Pakistan and Papua New Guinea, for which interest payments are not separately identified.

Reflecting the broad stability projected for average interest payments relative to GDP, targeted improvements in primary positions relative to GDP were about the same size as those in overall positions—just over 3 percentage points. As with the overall position, there was a significant link between the targeted improvements in the primary position and the size of the initial primary imbalance. Thus, countries that had had several previous arrangements and started from relatively strong primary positions tended to target smaller improvements. Moreover, countries expecting the largest increases in interest payments (concentrated among the new users) tended to seek the largest improvements in the primary account. Without explicit medium-term scenarios it is not possible to judge the adequacy of these targeted adjustments, but it is noteworthy that on average primary positions were targeted to be in surplus by over 1 percent of GDP.

On average, targeted improvements in primary accounts were evenly balanced between contributions from revenues and spending. Nevertheless, there were some distinct patterns in the reliance placed on expenditure restraint and revenue increases. The magnitude of targeted revenue increases and expenditure restraint was significantly negatively related to the initial levels of revenue and expenditure.15 Thus, most of the countries that had had several previous arrangements and had already reduced spending (especially on investment) to low levels planned to restrain noninterest spending modestly while relying heavily on revenue increases to meet their targeted primary adjustments. In contrast, more than half the countries that had had only one previous arrangement or were new users relied more heavily on expenditure restraint than on revenue increases for their adjustment.

Chart 2-2.Initial Fiscal Deficit and Targeted Change in Overall Balance1

Source: IMF staff estimates.

1 Deficit before grants; pre-program year.

2 Targeted change in first program year only. Positive number indicates targeted reductions; negative number indicates a deterioration in the fiscal accounts.

On average, outside Central Europe, reliance on noninterest current and capital expenditure restraint was about equal. However, because current non-interest expenditures were on average about twice the level of capital expenditures, the proportionate reduction in capital spending exceeded that of noninterest current spending. Once again, there were distinct differences between the newer adjusters, where restraint of capital spending relative to GDP was on average greater than that of current spending, and the countries with several previous programs, where capital spending relative to GDP was targeted to rise slightly and expenditure restraint was borne mainly by noninterest current spending.

In Central Europe, the objective was often not so much to reduce the fiscal imbalances, which, except in Bulgaria and Poland, were small, as to scale down the role of the public sector. Excluding Poland, where revenues relative to GDP were within the range of other countries, revenues were equivalent to at least 50 percent of GDP. Each of the countries except Poland, therefore, set as a goal a decline in the share of revenues relative to GDP—by over 10 percentage points in the countries that were newly transforming (Bulgaria, Czechoslovakia, and Romania). The projected decline in revenue ratios did not, however, reflect falling nominal revenues (these were expected to continue to rise) but rather projected rises in nominal GDP. Each country also aimed for a reduction in expenditure relative to GDP of equivalent or greater magnitude.

Table 2-4.Fiscal Adjustment: Deviation from Targeted Changes1(Average differences of changes in ratios to GDP)
Interest PaymentsPrimary BalanceRevenuesNoninterest Expenditures
Overall Balance2CurrentCapital
All countries excluding Central Europe0.4 (0.1)–0.30.1–0.3–0.3–0.2
Countries with several previous arrangements30.6 (0.4)–0.30.2–0.3–0.2–0.4
Countries with one previous arrangement41.0 (1.0)1.01.20.5–0.3
New users5–0.6 (–1.0)–0.5–1.0–2.0–1.30.3
Central European countries6–4.1 (–5.4)2.8–1.36.14.92.5
Source: IMF staff estimates.

Differences between targeted and actual changes. Data exclude Guyana, where GDP is subject to serious measurement problems; Argentina and Jordan, where GDP was substantially revised; Costa Rica, Pakistan, Papua New Guinea, and Poland, which did not target interest payments explicitly; and six countries where programs went off track immediately after approval (Brazil, the Congo, Guatemala, Madagascar, Uruguay, and Yugoslavia).

In parentheses are full sample estimates for overall balance including Costa Rica, Pakistan, Papua New Guinea, and Poland, for which disaggregated data are not available.

Excluding Brazil, Madagascar, Uruguay, and Yugoslavia (first purchase only); Argentina (GDP revisions); and Costa Rica (interest payments not targeted explicitly).

Excluding the Congo (first purchase only).

Excluding Guatemala (first purchase only), Jordan (GDP revisions), and Papua New Guinea and Pakistan (interest payments not targeted explicitly).

Excluding Poland (interest payments not targeted explicitly).

Source: IMF staff estimates.

Differences between targeted and actual changes. Data exclude Guyana, where GDP is subject to serious measurement problems; Argentina and Jordan, where GDP was substantially revised; Costa Rica, Pakistan, Papua New Guinea, and Poland, which did not target interest payments explicitly; and six countries where programs went off track immediately after approval (Brazil, the Congo, Guatemala, Madagascar, Uruguay, and Yugoslavia).

In parentheses are full sample estimates for overall balance including Costa Rica, Pakistan, Papua New Guinea, and Poland, for which disaggregated data are not available.

Excluding Brazil, Madagascar, Uruguay, and Yugoslavia (first purchase only); Argentina (GDP revisions); and Costa Rica (interest payments not targeted explicitly).

Excluding the Congo (first purchase only).

Excluding Guatemala (first purchase only), Jordan (GDP revisions), and Papua New Guinea and Pakistan (interest payments not targeted explicitly).

Excluding Poland (interest payments not targeted explicitly).

Outcomes Relative to Target

On average, outside Central Europe, except for the six countries where programs went off track (that is, did not meet performance criteria and were ineligible to make purchases under the arrangement) immediately after approval and three countries with severe data problems, countries were successful in meeting targeted changes in fiscal positions.16 The average improvement in the overall balance relative to GDP (that is, the difference between the average fiscal deficit during the program years and the deficit in the pre-program year) was about equal to that implied in the targets (Table 2-4). Outcomes were roughly evenly distributed between overperformers and underperformers, but there was quite a wide range. About half of the countries that had had several previous arrangements and nearly all with only one previous arrangement achieved larger-than-expected improvements in their fiscal balance (Chart 2-3). By contrast, most of the new users did worse than expected on the overall balance (Chart 2-4). In each of the Central European countries except Romania, deficits widened rather than falling as targeted.17

On average, outside the excluded countries, small shortfalls in revenue relative to GDP were slightly offset by greater-than-targeted restraint of spending relative to GDP. The shortfalls relative to targeted ratios of expenditure to GDP were, on average, evenly distributed among interest, noninterest current, and capital spending; however, with higher levels of noninterest current spending relative to capital spending and interest, shortfalls relative to targets were proportionately greater for capital and interest expenditures than for current spending. In Central Europe, the ratios of both revenues and expenditures to GDP were generally sharply above what had been programmed, but this situation owed much more to the unexpected declines in GDP that occurred during programs than to an overshooting of nominal targeted changes.

Chart 2-3.Fiscal Outcomes: Overperformers1

(Outturn less target, in percent of GDP)

Source: IMF staff estimates.

1 Excluding Central European countries and the six countries where programs went off track immediately after approval.

In general, overperformance and underperformance in the overall position relative to GDP were not large—averaging 1.4 percentage points for over-performers and 1.2 percentage points for underperformers. Although differences should not be exaggerated, a variety of influences appear to have distinguished the two groups. Perhaps the most striking was the terms of trade.18 Almost all of the over-performers benefited from stronger-than-expected terms of trade. Except in Nigeria and Zaїre, however, the improvement in the terms of trade was not reflected in significantly higher-than-targeted increases in revenue. In fact, excluding Nigeria and Zaїre, shortfalls in the ratio of revenue to GDP relative to targets were only slightly smaller in the countries that benefited from better-than-expected terms of trade than in those that did not.

The most significant difference between the over-performers and the underperformers was in non-interest expenditures. To some degree differences between projected and actual terms of trade may have influenced this outcome too. In general, stronger-than-expected terms of trade reflected higher-than-expected export prices rather than lower-than-expected import prices, so the effects on nominal expenditures should not have been great. Where the terms of trade appear to have played some role is in raising nominal GDP relative to program projections: nominal GDP growth exceeded targets by more in the overperformers than in the underperformers. This suggests that at least to some degree, countries planned spending in nominal terms and experienced reductions in ratios to GDP when GDP rose by more than expected.

Chart 2-4.Fiscal Outcomes: Underperformers1

(Outturn less target, in percent of GDP)

Source: IMF staff estimates.

1 Excluding Central European countries and the six countries where the programs went off track immediately after approval.

Nevertheless, there was also a distinction between the underperformers and overperformers in terms of shortfalls or excesses relative to targets for nominal spending, where the dominating factor appears to have been shortfalls in capital spending. While the increase in nominal noninterest current spending tended to overshoot in the overperformers, the increase in nominal capital spending was lower than targeted in all but two (Mexico and Nigeria)—in most by over 10 percentage points. In the underperformers, by contrast, noninterest current spending undershot in more than half, but nominal capital spending generally exceeded targets. It seems unlikely that this difference in outcomes for capital spending reflected the influence of the terms of trade on prices of capital goods. It is notable, however, that the overperformers on average targeted less restraint than the underperformers.19 Thus, targets for the underperformers reflected relatively ambitious efforts to restrain capital spending that left little room for policy slippage. By contrast, the less ambitious targets of the overperformers made them more tolerant of normal disruptions, such as delays in implementation, that resulted in shortfalls relative to target.

In Central Europe, much greater-than-expected declines in GDP resulted in nearly all ratios to GDP—revenue and spending—exceeding target. Nominal revenues were, however, below target in each country except Romania. Large nominal spending overruns occurred only in Bulgaria and Romania. In Bulgaria, this stemmed from higher-than-projected interest payments owing to the depreciation of the lev and the rise in domestic interest rates. In Romania, current noninterest expenditures overshot targets mainly because of a failure to rein in subsidies to enterprises. All the countries of Central Europe faced growing expenditure pressure from social security—particularly pensions and unemployment benefits. These were responsible for the smaller nominal spending overshoots of Hungary and Poland, and have since caused problems for the successor arrangements of Bulgaria, Czechoslovakia, and Romania.

Were Fiscal Adjustments Sustained?

Excluding the Central European countries and the six countries whose only programs went off track immediately after approval, average deficits fell from some 8 percent of GDP in the pre-program year to 5 percent during the programs under review (Table 2-5). The improvement was most pronounced on average for the countries with one previous program and smallest for the new users. (The apparent improvement for the countries with one previous arrangement overstates the underlying improvement, however, because a sizable portion of the large improvement in Egypt’s overall position reflected a rebound from a large capital expenditure to recapitalize financial institutions in the pre-program year. Excluding this effect, the improvement in the overall position would have been only slightly larger than for the countries with several previous arrangements.) In Central Europe deficits relative to GDP widened sharply.

Ideally, this study would sort out the degree to which these outcomes reflected discretionary policy changes, exogenous developments (external or domestic), or cyclical influences. In fact, the influences were so varied and, especially for the many countries with multiyear programs, interactive, that it is impossible to identify them separately with enough precision to make such an exercise meaningful. It is apparent from the influence of the terms of trade on the deviations of actual from targeted fiscal outcomes that the terms of trade—one of the most important exogenous variables for most countries—did affect fiscal outcomes. Moreover, the countries that experienced improvements in the terms of trade over the program period had larger improvements in their fiscal positions on average than the countries that experienced stable or declining terms of trade. Similarly, most countries that had above-average improvements in their fiscal positions had higher-than-average real GDP growth. A related issue is the degree to which improvements in the fiscal position reflected stopgap measures, such as the pushing forward of capital spending projects or unsustainable restraint in civil service wages. One broad indication of the influence of fundamental policy changes as opposed to that of exogenous, cyclical, or stopgap measures is the degree to which improvements in the fiscal position were sustained after the program period. For many countries, the program period extends through 1992, or subsequent programs outside the scope of this review started in 1992. Still, comparing the fiscal position in 1992 with that before the program period provides the best available indication of the degree to which fiscal changes were sustained.

More than half of the countries continued to strengthen their overall positions through 1992 (Chart 2-5).20 Most of these countries kept up revenue ratios and cut or contained spending ratios. A few (Côte d’Ivoire, Mexico, Tunisia, and Papua New Guinea) lost revenues but cut expenditure ratios (interest and noninterest) by more.

One third of the countries outside Central Europe experienced some weakening in their overall balances from the average position attained during the arrangements under review: five of these (Algeria, Cameroon, Haiti, Mali, and Pakistan) had larger overall deficits relative to GDP in 1992 than before their arrangements (Chart 2-6). In a few, exogenous setbacks—especially from the terms of trade or civil unrest—contributed importantly, but in most, policy slippages were the chief problem. (Adverse developments in the terms of trade were prevalent in the countries where fiscal positions weakened, although almost all countries that improved their fiscal positions after the programs reviewed also suffered terms of trade deteriorations.) A weakening in revenues occurred in five of the nine countries and explains most of the deterioration of the fiscal balance in Haiti, Mali, and Venezuela. In Haiti (in the context of civil unrest) and Mali these reflected reversals of previous gains in domestic tax revenues—indicative of the difficulties of raising revenue in low-income countries. Venezuela suffered sharply lower oil revenues in 1992. A rise in noninterest expenditures contributed significantly to the fiscal deteriorations in Algeria, El Salvador, and Nigeria—in Algeria owing to a relaxation of wage restraint, in El Salvador owing to a large increase in capital spending as civil strife ended, and in Nigeria reflecting a resumption of capital spending, which had been curtailed during the 1989 program. Several countries also faced higher interest payments. In El Salvador, Nigeria, and Pakistan the domestic interest burden was raised by the introduction of treasury bill auctions—with the effect growing as an increasingly larger share of domestic debt paid market-related interest rates. (Interest payments for nearly all countries liberalizing their financial markets from positions of initially negative real interest rates rose steadily over the period under review.)

Table 2-5.Fiscal Adjustment: Pre-Program Year Through 1992(In percent of GDP)
Overall BalanceInterest PaymentsPrimary Balance
Pre-program yearProgram average1992Pre-program yearProgram average1992Pre-program yearProgram average1992
All countries excluding Central Europe1–8.1–4.8–4.45.85.45.5–2.30.61.2
Countries with several previous arrangements2–6.7–3.4–2.77.45.64.80.72.22.2
Countries with one previous arrangement3–8.7–3.5–4.35.96.16.8–2.82.62.5
New users4–9.5–7.36.63.74.65.6–5.9–2.7–1.0
Central European countries–3.7–4.6–6.62.44.74.8–1.20.1–1.8
Noninterest Expenditures
RevenuesCurrentCapital
Pre-program yearProgram average1992Pre-program yearProgram average1992Pre-program yearProgram average1992
All countries excluding Central Europe121.222.322.616.816.015.86.75.75.6
Countries with several previous arrangements220.020.620.714.413.814.24.84.74.3
Countries with one previous arrangement324.327.227.717.116.917.010.07.88.2
New users420.621.321.719.618.317.06.95.75.8
Central European countries45.942.540.242.538.738.84.73.73.2
Source: IMF staff estimates.

Excluding Guyana, where GDP is subject to serious measurement problems; Zaїre, for lack of data through 1992; and six countries that did not proceed beyond the first purchase (Brazil, the Congo, Guatemala, Madagascar, Uruguay, and Yugoslavia).

Excluding Brazil, Madagascar, Uruguay, Yugoslavia (which made the first purchase only), and Zaїre.

Excluding the Congo (which made the first purchase only).

Excluding Guatemala (which made the first purchase only).

Source: IMF staff estimates.

Excluding Guyana, where GDP is subject to serious measurement problems; Zaїre, for lack of data through 1992; and six countries that did not proceed beyond the first purchase (Brazil, the Congo, Guatemala, Madagascar, Uruguay, and Yugoslavia).

Excluding Brazil, Madagascar, Uruguay, Yugoslavia (which made the first purchase only), and Zaїre.

Excluding the Congo (which made the first purchase only).

Excluding Guatemala (which made the first purchase only).

All the countries of Central Europe, except Bulgaria, continued to experience difficulties in containing the ratio of expenditures to GDP following the arrangements under review (Chart 2-7). Noninterest expenditure ratios rose in Czechoslovakia and Hungary and remained at relatively high levels in Poland and Romania. Interest payments also represented a growing burden in Hungary and Poland, following the increased use of treasury bill auctions for domestic financing. Against these developments, the generalized weakening or stagnation of revenues (as a share of GDP) that occurred in all countries resulted in budget deficits in 1992 being wider than during the arrangements under review, except in Bulgaria (which nevertheless had the largest deficit among the group).

Overall, outside Central Europe where expenditure control remains the principal problem, it is difficult to identify any single economic cause or policy change that distinguished countries that sustained fiscal adjustment from those that did not. Perhaps the only clear distinction is that, of the 12 countries that slipped, only 6 had successor arrangements in place in 1992, and 3 of those were in Central Europe, where continued fiscal problems might have been expected notwithstanding the involvement of the IMF.21 By contrast, all but one of the 15 countries that continued to improve their fiscal accounts had arrangements in place (6 being the tail end of multiyear arrangements reviewed in this study).22

Chart 2-5.Adjustment Sustained1

(In percent of GDP)

Source: IMF staff estimates.

1 For Costa Rica, Jamaica, Mexico, the Philippines, and Tunisia the (multiyear) arrangements under review encompassed 1992. All figures are latest estimates.

Chart 2-6.Adjustment Reversed1

(In percent of GDP)

Source: IMF staff estimates.

1 Venezuela’s EFF encompassed 1992, although no target was set for that year. All figures are latest estimates.

Financing the Fiscal Imbalance

The way in which fiscal deficits were financed had important implications for the level of external indebtedness, the future burden of interest payments, and bank credit creation. Most countries had four broad options for financing deficits. One was to run up domestic and external arrears, although this option had been pursued so often in the past it was disallowed during programs, which sought to redress past arrears. A second was external financing, in which case limits set in programs had to take into account already high external debt ratios in many countries or constrained access to capital markets in others. A third was domestic bank credit, in which case limits reflected goals for restraining money growth and reserve accumulation. A fourth source of financing was the domestic nonbank private sector (generally pension funds, insurance companies, large corporations, and post office savings banks as well as credit institutions outside the banking system), often a relatively untapped source in large part because countries had lacked the markets and instruments (bills and bonds) needed to tap it or were reluctant to pay typical interest rates.

Chart 2-7.Fiscal Performance: Central Europe1

(In percent of GDP)

Source: IMF staff estimates.

1 Latest estimates. Breakdown of expenditures differs from format for Charts 2-2 through 2-6 because of the importance of interest payment developments for Bulgaria and ambiguities in the classification of current versus capital expenditures.

An important question addressed in this section is the extent to which countries sought and achieved a shift in the structure of their fiscal financing (usually in the context of falling domestic financing requirements) from bank credit creation and arrears to non-bank financing. For most countries this was part of a process of financial sector development—in decontrolling interest rates and introducing auctions—geared toward both creating the channels for nonbank financing and improving the efficiency of financial markets. This part of the study therefore compares the financing arrangements of two categories: the “advanced reformers” (countries that had both liberalized bank interest rates and introduced auctions for treasury bills or equivalent instruments by 1992) and “slow reformers” (countries that had taken only one or neither of these steps by 1992). The advanced reformers are further subdivided into those where real interest rates were initially positive and those where they were initially negative. The former aimed to maintain, or even reduce, real interest rates while the latter typically sought to raise them. The experience reveals the trade-off, when domestic financing needs remain significant, between money growth, inflation, and negative real interest rates when reliance on bank credit is large and high real interest rates when nonbank financing is large.

Table 2-6.Financing of the Fiscal Balance Pre-Program Year1(Average in percent of GDP)
OverallExternalDomestic BorrowingArrears
TotalNonbankBank
All countries excluding Central Europe7.23.72.91.41.60.7
Of which:
Advanced reformers with:2
Negative real interest rates38.33.04.31.13.21.0
Positive real interest rates44.61.73.03.0–0.1
Slow reformers57.94.04.00.83.1
CFA countries69.67.50.30.10.31.8
Countries with several previous arrangements75.53.51.82.0–0.20.3
Countries with one previous arrangement88.84.72.41.01.51.6
New users99.33.45.40.94.50.4
Central European countries102.0–0.21.40.21.30.8
Source: IMF staff estimates.

The pre-program year is defined as the year preceding the first arrangement reviewed for each country. Data are as estimated at the time of the first arrangement. Data exclude countries for which satisfactory program financing figures are unavailable (Argentina, Brazil, Guatemala, Honduras, Yugoslavia, and Zaїre) and exclude Guyana. Overall balances may differ from those reported in Table 2-2 owing to the inclusion of different countries and to the definition of the overall balance employed.

Countries that had auctions for government or central bank paper and had freed bank interest rates during programs or by 1992. Slow reformers comprise the residual group of countries. Real interest rates, defined in Section IV of this volume, Behavior of Nominal and Real Interest Rates, are measured as of the year preceding the first arrangement.

Costa Rica, Ecuador, Egypt, El Salvador, Nigeria, Trinidad and Tobago, and Venezuela.

Jamaica, Mexico, Morocco, Papua New Guinea, the Philippines, Tunisia, and Uruguay.

Algeria, Haiti, Jordan, Madagascar, and Pakistan.

Cameroon, the Congo, Côte d’Ivoire, Gabon, and Mali.

Costa Rica, Côte d’Ivoire, Ecuador, Haiti, Jamaica, Madagascar, Mali, Mexico, Morocco, the Philippines, and Uruguay.

Algeria, the Congo, Egypt, Gabon, Nigeria, and Tunisia.

Cameroon, El Salvador, Jordan, Pakistan, Papua New Guinea, and Trinidad and Tobago.

Bulgaria, Czechoslovakia, Hungary, Poland (1991 only), and Romania.

Source: IMF staff estimates.

The pre-program year is defined as the year preceding the first arrangement reviewed for each country. Data are as estimated at the time of the first arrangement. Data exclude countries for which satisfactory program financing figures are unavailable (Argentina, Brazil, Guatemala, Honduras, Yugoslavia, and Zaїre) and exclude Guyana. Overall balances may differ from those reported in Table 2-2 owing to the inclusion of different countries and to the definition of the overall balance employed.

Countries that had auctions for government or central bank paper and had freed bank interest rates during programs or by 1992. Slow reformers comprise the residual group of countries. Real interest rates, defined in Section IV of this volume, Behavior of Nominal and Real Interest Rates, are measured as of the year preceding the first arrangement.

Costa Rica, Ecuador, Egypt, El Salvador, Nigeria, Trinidad and Tobago, and Venezuela.

Jamaica, Mexico, Morocco, Papua New Guinea, the Philippines, Tunisia, and Uruguay.

Algeria, Haiti, Jordan, Madagascar, and Pakistan.

Cameroon, the Congo, Côte d’Ivoire, Gabon, and Mali.

Costa Rica, Côte d’Ivoire, Ecuador, Haiti, Jamaica, Madagascar, Mali, Mexico, Morocco, the Philippines, and Uruguay.

Algeria, the Congo, Egypt, Gabon, Nigeria, and Tunisia.

Cameroon, El Salvador, Jordan, Pakistan, Papua New Guinea, and Trinidad and Tobago.

Bulgaria, Czechoslovakia, Hungary, Poland (1991 only), and Romania.

Initial Conditions

Excluding Central Europe and countries for which data on financing was not comprehensive (see Table 2-6 for countries included), the average overall borrowing requirement in the year before the first arrangement reviewed for each country was 7 percent of GDP (Table 2-6).23 On average, more than half was financed externally, although there was a wide range. More than half of the domestic financing component was met by the banking system. Eight countries, four in the CFA franc zone, were accumulating budgetary arrears—domestic and external—on current transactions.

Countries that started with negative real interest rates tended to have a similar pattern of financing whether they were to proceed rapidly or slowly with reform: on average, a domestic borrowing requirement of about 4 percent of GDP, most of which was financed through the banking system.24 By contrast, countries with positive real interest rates prior to their arrangements, on average, met nearly all their domestic borrowing requirements outside the banking system.25 For the CFA franc zone, domestic borrowing requirements were small and most of the difference between overall borrowing and external finance was arrears accumulation. In Central Europe, with little external financing available prior to programs, overall borrowing requirements were low except in Bulgaria and Poland and mostly financed by banks.

Table 2-7.Financing Targets: Change from Pre-Program Year1(Average, differences of ratios to GDP)
OverallExternalDomestic BorrowingArrears
TotalNonbankBank
All countries excluding Central Europe–2.51.2–2.2–0.5–1.8–1.5
Of which:
Advanced reformers with:2
Negative real interest rates3–4.3–0.1–3.30.1–3.4–1.0
Positive real interest rates4–2.01.2–3.1–2.1–0.9–0.1
Slow reformers5–1.70.6–2.10.1–2.1–0.3
CFA countries6–2.13.7–0.20.4–0.6–5.6
Countries with several previous arrangements7–2.21.1–2.2–1.4–0.8–1.1
Countries with one previous arrangement8–2.51.0–0.31.0–1.3–3.1
New users9–3.41.6–4.2–0.3–4.0–0.7
Central European countries10–0.92.0–1.70.2–1.9–1.2
Source: IMF staff estimates.

Average of each country’s program target; differences between program targets or target average if multiyear program and estimates for the year preceding the first arrangement under review for each country. Data exclude countries for which satisfactory program financing figures are unavailable (Argentina, Brazil, Guyana, Honduras, Yugoslavia, and Zaїre).

Countries that had auctions for government or central bank paper and that had freed interest rates in the banking system either during programs or by 1992. Slow reformers comprise the residual group of countries. Real interest rates as of the year preceding the first arrangement under review.

Costa Rica, Ecuador, Egypt, El Salvador, Nigeria, Trinidad and Tobago, and Venezuela.

Jamaica, Mexico, Morocco, Papua New Guinea, Philippines, Tunisia, and Uruguay.

Algeria, Haiti, Jordan, Madagascar, and Pakistan.

Cameroon, the Congo, Côte d’Ivoire, Gabon, and Mali.

Costa Rica, Côte d’Ivoire, Ecuador, Haiti, Jamaica, Madagascar, Mali, Mexico, Morocco, the Philippines, and Uruguay.

Algeria, the Congo, Egypt, Gabon, Nigeria, and Tunisia.

Cameroon, El Salvador, Jordan, Pakistan, Papua New Guinea, and Trinidad and Tobago.

Bulgaria, Czechoslovakia, Hungary, Poland (1991 only), and Romania.

Source: IMF staff estimates.

Average of each country’s program target; differences between program targets or target average if multiyear program and estimates for the year preceding the first arrangement under review for each country. Data exclude countries for which satisfactory program financing figures are unavailable (Argentina, Brazil, Guyana, Honduras, Yugoslavia, and Zaїre).

Countries that had auctions for government or central bank paper and that had freed interest rates in the banking system either during programs or by 1992. Slow reformers comprise the residual group of countries. Real interest rates as of the year preceding the first arrangement under review.

Costa Rica, Ecuador, Egypt, El Salvador, Nigeria, Trinidad and Tobago, and Venezuela.

Jamaica, Mexico, Morocco, Papua New Guinea, Philippines, Tunisia, and Uruguay.

Algeria, Haiti, Jordan, Madagascar, and Pakistan.

Cameroon, the Congo, Côte d’Ivoire, Gabon, and Mali.

Costa Rica, Côte d’Ivoire, Ecuador, Haiti, Jamaica, Madagascar, Mali, Mexico, Morocco, the Philippines, and Uruguay.

Algeria, the Congo, Egypt, Gabon, Nigeria, and Tunisia.

Cameroon, El Salvador, Jordan, Pakistan, Papua New Guinea, and Trinidad and Tobago.

Bulgaria, Czechoslovakia, Hungary, Poland (1991 only), and Romania.

Program Objectives

On average, the programmed decline in overall borrowing requirements relative to GDP, and small increases in external financing relative to GDP, were to permit lower domestic borrowing and a repayment of arrears (domestic and external) on current transactions (Table 2-7).26 In about three quarters of the countries, lower domestic borrowing requirements were to allow a significant reduction of borrowing from banks; for about half of the countries, even a repayment to banks. The clearance of arrears, principally though financing relief from arrears holders, featured particularly in the CFA countries.

Little change was envisaged in the ratio to GDP of borrowing from nonbanks—except in the countries where real interest rates were positive prior to their arrangements. For these countries, where, on average, repayment to banks equivalent to 1 percent of GDP was envisaged, nonbank borrowing was to be reduced from 3 percent of GDP in the pre-program year to 1 percent on average during programs. By contrast, in advanced reformers that started with negative real interest rates, nonbank borrowing was expected to remain equivalent, on average, to about 1 percent of GDP while bank borrowing—previously about 3 percent of GDP—was to be eliminated. Similarly, in the slow reformers, the CFA franc zone, and Central Europe, bank borrowing was to be curtailed, and, reflecting limited opportunities in undeveloped financial markets, nonbank financing was to remain small.

Table 2-8.Financing of the Fiscal Balance: Deviations from Targeted Changes1(Average differences of changes in ratios to GDP)
OverallExternalDomestic BorrowingArrears
TotalNonbankBank
All countries excluding Central Europe0.1–0.60.4–0.40.7
Of which:
Advanced reformers with:2
Negative real interest rates3–1.1–1.9–0.11.0–1.00.9
Positive real interest rates40.1–0.40.50.40.1
Slow reformers51.11.0–0.7–0.2–0.50.8
CFA countries61.0–0.10.10.4–0.31.1
Countries with several previous arrangements7–0.20.1–0.5–0.50.3
Countries with one previous arrangement8–0.8–1.9–0.7–0.1–0.61.8
New users91.0–0.51.313–0.10.4
Central European countries104.5–0.83.60.33.31.7
Source: IMF staff estimates.

Differences between targeted and actual changes. Positive numbers indicate excesses in financing requirements over targets. Program years weighted by country. Data exclude countries for which comparison is unavailable (Argentina, Honduras, Zaїre) or where programs went off track immediately after approval (Brazil, the Congo, Guatemala, Madagascar, Uruguay, and Yugoslavia). Overall balances may differ from those reported in Table 2-4 owing to differences in the countries covered and in the definition of the overall balance.

Countries that had auctions for government or central bank paper and that had freed interest rates in the banking system either during programs or by 1992. Slow reformers comprise the residual group of countries. Real interest rates measured as of the year preceding the first arrangement.

Costa Rica, Ecuador, Egypt, El Salvador, Nigeria, Trinidad and Tobago, and Venezuela.

Jamaica, Mexico, Morocco, Papua New Guinea, the Philippines, and Tunisia.

Algeria, Haiti, Jordan, and Pakistan.

Cameroon, Côte d’Ivoire, Gabon, and Mali.

Costa Rica, Côte d’Ivoire, Ecuador, Haiti, Jamaica, Mali, Mexico, Morocco, and the Philippines.

Algeria, Egypt, Gabon, Nigeria, and Tunisia.

Cameroon, El Salvador, Jordan, Pakistan, Papua New Guinea, Trinidad and Tobago, and Venezuela.

Bulgaria, Czechoslovakia, Hungary, Poland (1991 only), and Romania.

Source: IMF staff estimates.

Differences between targeted and actual changes. Positive numbers indicate excesses in financing requirements over targets. Program years weighted by country. Data exclude countries for which comparison is unavailable (Argentina, Honduras, Zaїre) or where programs went off track immediately after approval (Brazil, the Congo, Guatemala, Madagascar, Uruguay, and Yugoslavia). Overall balances may differ from those reported in Table 2-4 owing to differences in the countries covered and in the definition of the overall balance.

Countries that had auctions for government or central bank paper and that had freed interest rates in the banking system either during programs or by 1992. Slow reformers comprise the residual group of countries. Real interest rates measured as of the year preceding the first arrangement.

Costa Rica, Ecuador, Egypt, El Salvador, Nigeria, Trinidad and Tobago, and Venezuela.

Jamaica, Mexico, Morocco, Papua New Guinea, the Philippines, and Tunisia.

Algeria, Haiti, Jordan, and Pakistan.

Cameroon, Côte d’Ivoire, Gabon, and Mali.

Costa Rica, Côte d’Ivoire, Ecuador, Haiti, Jamaica, Mali, Mexico, Morocco, and the Philippines.

Algeria, Egypt, Gabon, Nigeria, and Tunisia.

Cameroon, El Salvador, Jordan, Pakistan, Papua New Guinea, Trinidad and Tobago, and Venezuela.

Bulgaria, Czechoslovakia, Hungary, Poland (1991 only), and Romania.

Outcomes Relative to Target

On average, outside Central Europe, the six countries where programs went off track immediately after approval, and countries with data deficiencies, changes in overall financing requirements were broadly as planned (Table 2-8).27 About half the countries experienced shortfalls relative to targeted increases in external financing of at least 20 percent of total programmed financing requirements. (Frequently, these shortfalls were not reflected in shortages in economy-wide external financing because other levels of government or the private sector obtained larger-than-targeted increases in external financing.) In some of these, the shortfall in increases in external financing were partially, fully, or even more than fully matched by increases in domestic or external arrears. Thus, overruns in the change in domestic financing requirements were greater than 1 percent of GDP in only six countries and on average were small. In Central Europe, overall as well as domestic financing requirements rose considerably more than targeted in Bulgaria, Hungary, and Poland but were about as planned in Czechoslovakia and Romania.

Changes in the structure of domestic financing often differed markedly from those programmed, and patterns of these deviations were related to the stage of development of financial markets. For the advanced reformers with positive real interest rates prior to programs, greater-than-planned shifts from bank to nonbank financing in a few countries were more or less matched by opposite shifts in other countries and on average changes were close to those programmed. For the advanced reformers that started with negative real interest rates, however, the switch of funding from banks to nonbanks was much greater than planned (Chart 2-8). Why did this happen? Explanations range from a heavier-than-intended reliance on treasury bill auctions in order to maintain tight monetary conditions to the effects of sterilization (conducted with treasury bills) in response to private capital inflows. Several of these countries, which began with negative real interest rates, saw interest rates rise temporarily to positive levels that raised concerns. (See the paper Behavior of Nominal and Real Interest Rates in this volume.)

Chart 2-8.Nonbank Versus Bank Financing

(In percent of GDP)

Source: IMF staff estimates.

1 Countries that had introduced auctions for treasury bills and liberalized interest rates by 1992 and that had entered programs with negative real interest rates, excluding Central Europe.

2 Countries that had either not introduced auctions or not decontrolled interest rates by 1992, excluding the CFA zone and Central Europe.

In countries where financial reform was slow, the scope for financing from nonbanks remained limited, except in Pakistan. It is therefore not surprising that, except in Pakistan, nonbank financing tended to be negligible and no higher than programmed. In Pakistan, excesses relative to targets in the overall financing requirement were met by increasing tap issues of bonds despite their relatively high cost. It is notable that, apart from Jordan, these countries generally failed to raise domestic real interest rates to positive levels. In Jordan, sharply lower than targeted domestic bank borrowing (in part reflecting resort to external arrears) facilitated an increase in domestic real interest rates to positive levels.

Financing Through 1992

On average, outside Central Europe, the countries where arrangements went off track immediately after approval, and those for which financing data are incomplete, the progress in reducing overall borrowing requirements relative to GDP during programs was continued through 1992 (Table 2-9). On average, reductions in overall financing requirements relative to GDP translated about evenly into lower external financing and lower domestic borrowing relative to GDP. The latter was about evenly split between lower bank and nonbank borrowing. There were, however, a variety of changes in financing patterns, related to countries’ success in reducing finance requirements, their access to external financing, and the development of their domestic financial markets.

Most of the advanced reformers (whether starting with positive or negative real interest rates) significantly reduced domestic financing requirements relative to GDP by cutting their overall deficit and, in about half, by increasing external financing. In almost all these countries, most of the drop in the domestic financing requirement was reflected in lower bank borrowing or even repayments of previous bank borrowing, while nonbank borrowing either was constant or rose relative to GDP. In general, the shift from bank to nonbank financing was greatest for the countries that started with negative real interest rates.28 The slow reformers not only reduced their deficits by less but also increased their recourse to bank financing. In these countries, nonbank financing, which was small from the start, fell relative to GDP. Banks also remained the chief means of financing in Central Europe, where, despite financial reforms, nonbank markets were not yet developed by 1992.

The countries that had accumulated arrears on government current transactions prior to arrangements had mixed success in clearing them. These arrears were concentrated in the CFA countries (Chart 2-9). There was more success in clearing, or at least slowing the accumulation of, domestic arrears than external arrears. This was in part because external arrears were often on interest payments to commercial banks, which in Bulgaria, Ecuador, and Nigeria were put on hold pending reschedulings. Except in Morocco and Poland, arrears resumed after the programs under review.

Program Monitoring: The Role of Performance Criteria and Adjusters

Quantitative fiscal performance criteria typically had two objectives: (i) to cap the overall deficit of the government or public sector and (ii) to constrain access to bank credit, both to leave room for bank credit to the private sector and to curtail money creation. These objectives were pursued through several different constructions of performance criteria involving ceilings on the government or public sector borrowing requirement, the domestic borrowing requirement, and net bank credit to the government (Appendix Table 2-A2). Often these ceilings were supplemented with “adjusters”—clauses that specified changes in the level or structure of financing that should occur when revenues, expenditures, or specific forms of financing deviated from targets. In fact, some of these adjusters were equivalent to an additional, separate performance criterion: for example, a ceiling on domestic financing with a downward adjuster for excesses in external financing would be equivalent to a ceiling on the total borrowing requirement.

In broad terms, there were three approaches to setting performance criteria to meet these objectives (Table 2-10). In almost half of the arrangements, the focus was on strictly capping the fiscal deficit: accordingly, a ceiling was placed on the government or public sector borrowing requirement without adjusters or on net bank credit with downward adjusters for excesses in external or nonbank financing. In another third of the arrangements, the focus was on limiting the government or public sector borrowing requirement, while allowing symmetric deviations from the target in specified conditions—mostly when oil prices, specified revenue items, or external financing differed from projections.29 Performance criteria in these countries consisted of ceilings on domestic financing with no adjusters or, equivalently, ceilings on the government or public sector borrowing requirement with adjusters for deviations from projected external financing. Performance criteria for the remaining countries focused on net credit to the government or public sector alone—some with adjusters for oil revenues or excesses in external financing.

Table 2-9.Financing of the Fiscal Balance: Pre-Program Year Through 19921(In percent of GDP)
Domestic Borrowing
OverallExternalTotalNonbankBankArrears
Pre-program yearProgram average1992Pre-program yearProgram average1992Pre-program yearProgram average1992Pre-program yearProgram average1992Pre– program yearProgram average1992Pre-program yearProgram average1992
All countries excluding Central
Europe8.25.14.93.73.91.93.50.72.01.81.51.21.7–0.80.80.90.51.0
Of which:
Advanced reformers with:2
Negative real interest rates38.13.14.23.21.81.34.51.02.61.72.21.62.9–1.21.00.40.40.4
Positive real interest rates45.73.41.81.92.92.13.80.6–0.33.51.21.70.3–0.6–1.9–0.1–0.1–0.1
Slow reformers58.88.87.04.66.22.74.21.25.41.21.1–0.13.00.15.50.6–1.0
CFA countries611.59.08.86.67.42.10.5–0.11.10.50.91.10.1–1.1–0.14.31.85.7
Countries with several previous arrangements76.94.03.72.53.82.12.5–0.60.62.40.70.70.2–1.3–0.91.90.81.0
Countries with one previous arrangement88.74.25.55.41.7–0.63.21.74.82.02.81.71.2–1.03.10.10.81.3
New users99.27.05.94.05.43.34.91.62.01.11.60.73.71.30.30.7
Central European countries102.26.17.1–0.31.21.61.53.35.20.20.80.31.32.54.91.01.70.3
Source: IMF staff estimates.

Latest estimates. Data exclude countries for which data through 1992 are unavailable (Argentina, Czechoslovakia, Zaїre) or where programs went off track immediately after approval (Brazil, the Congo, Guatemala, Madagascar, Uruguay, and Yugoslavia). Pre-program data may differ from those in Table 2-2 because of the inclusion of different countries and differences in coverage.

Countries that had auctions for government or central bank paper and that had freed bank interest rates either during programs or by 1992. Slow reformers comprise the residual group of countries. Real interest rates are measured as of the year preceding the first arrangement.

Costa Rica, Ecuador, Egypt, El Salvador, Honduras, Nigeria, Trinidad and Tobago, and Venezuela.

Jamaica, Mexico, Morocco, Papua New Guinea, the Philippines, and Tunisia.

Algeria, Haiti, Jordan, and Pakistan.

Cameroon, Côte d’Ivoire, Gabon, and Mali.

Costa Rica, Ecuador, Haiti, Jamaica, Mali, Mexico, Morocco, and the Philippines.

Algeria, Egypt, Gabon, Nigeria, and Tunisia.

Cameroon, El Salvador, Honduras, Jordan, Pakistan, Papua New Guinea, Trinidad and Tobago, and Venezuela.

Bulgaria, Hungary, Poland (1991 only), and Romania.

Source: IMF staff estimates.

Latest estimates. Data exclude countries for which data through 1992 are unavailable (Argentina, Czechoslovakia, Zaїre) or where programs went off track immediately after approval (Brazil, the Congo, Guatemala, Madagascar, Uruguay, and Yugoslavia). Pre-program data may differ from those in Table 2-2 because of the inclusion of different countries and differences in coverage.

Countries that had auctions for government or central bank paper and that had freed bank interest rates either during programs or by 1992. Slow reformers comprise the residual group of countries. Real interest rates are measured as of the year preceding the first arrangement.

Costa Rica, Ecuador, Egypt, El Salvador, Honduras, Nigeria, Trinidad and Tobago, and Venezuela.

Jamaica, Mexico, Morocco, Papua New Guinea, the Philippines, and Tunisia.

Algeria, Haiti, Jordan, and Pakistan.

Cameroon, Côte d’Ivoire, Gabon, and Mali.

Costa Rica, Ecuador, Haiti, Jamaica, Mali, Mexico, Morocco, and the Philippines.

Algeria, Egypt, Gabon, Nigeria, and Tunisia.

Cameroon, El Salvador, Honduras, Jordan, Pakistan, Papua New Guinea, Trinidad and Tobago, and Venezuela.

Bulgaria, Hungary, Poland (1991 only), and Romania.

Chart 2-9.Fiscal Accounts: Accumulation of Domestic and External Arrears1

(In percent of GDP)

Source: IMF staff estimates.

1 Arrears on current transactions of the nonfinancial public sector or central government. This represents a much narrower coverage of arrears than reported in the balance of payments, which includes other public sector entities, the central bank, and the private sector and which covers capital account as well as current account transactions.

2 Country with no domestic arrears.

3 CFA country.

4 Country with no external arrears.

Table 2-10.Fiscal Financial Performance Criteria
1.Countries with ceilings on total borrowing and no upward adjustments (or equivalent)1
Argentina

Brazil

Bulgaria2

Costa Rica (1991)

Côte d’Ivoire2

Hungary (1990, 1991)2

Papua New Guinea2

Philippines (1991)

Poland (1991, 1992)2

Trinidad and Tobago (1989, I990)2

Uruguay

Zaїre2

With downward adjuster for shortfalls in external financing

Czechoslovakia2

Madagascar2

Romania2
2.Countries with ceilings on total borrowing subject to adjustments (or equivalent)3
With adjuster for oil price or other revenue

Algeria2

Mexico2

Morocco (1988, 1989)2

Venezuela

With adjuster for external financing

Costa Rica (1989)

Guatemala

Guyana2

Haiti

Honduras

Jamaica (1988, 1990, 1991)

Philippines (1989)

Yugoslavia
3.Countries with ceilings on bank financing only
Congo4

Jordan

Pakistan

Tunisia



With adjuster for oil prices



Ecuador

Nigeria (1989, 1991)



With adjuster for excess in external financing



Cameroon

Egypt

El Salvador

Gabon4

Mali4
Source: IMF reports.

Ceiling on bank credit with asymmetric adjusters for nonbank and external financing for Bulgaria, Czechoslovakia, Madagascar, Romania, and Zaїre. For Costa Rica, ceiling on domestic financing with symmetric adjuster for external financing.

With subceilings on bank credit financing.

Ceiling on total borrowing subject to symmetric adjuster for external financing except for Costa Rica, Guatemala, Haiti, Honduras, and Yugoslavia, where ceiling was on domestic financing with no adjuster.

With ceiling on domestic arrears.

Source: IMF reports.

Ceiling on bank credit with asymmetric adjusters for nonbank and external financing for Bulgaria, Czechoslovakia, Madagascar, Romania, and Zaїre. For Costa Rica, ceiling on domestic financing with symmetric adjuster for external financing.

With subceilings on bank credit financing.

Ceiling on total borrowing subject to symmetric adjuster for external financing except for Costa Rica, Guatemala, Haiti, Honduras, and Yugoslavia, where ceiling was on domestic financing with no adjuster.

With ceiling on domestic arrears.

The reasons for these differences in the structure of fiscal performance criteria are not always clear, although judgments about several types of risks to the program did influence the choice. First, the design of some performance criteria was influenced by the availability, reliability, and timeliness of data. Second, many countries that had explicit adjustments for deviations from target of external finance gave more favorable treatment to project finance on concessional terms. Deviations in such financing could be reflected in a higher borrowing requirement without unmanageable effects on the debt-service burden. Furthermore, 13 of the 36 countries recorded significant external aid/grants, and these were classified as part of revenues in calculating the deficit for monitoring purposes. In programs for these countries, external aid was therefore effectively subject to a symmetric adjuster.

Third, most oil producers had adjusters for unexpected movements in oil prices, reflecting the importance of revenue from oil exports. The producers of other primary commodities, however, did not have equivalent adjusters for changes in their commodity prices.30 On the whole, this reflected the fact that oil tended to account for a much larger share of budgetary revenues for oil producers than did other primary commodities for other exporters. In virtually all cases, these adjusters were symmetric—up to a specified upper ceiling on borrowing—and were related to changes in the price of oil (the upward adjusters for Mexico and Venezuela were partial only).

Fourth, in some instances, ceilings on bank credit alone were used because alternatives to bank credit for financing were believed to be limited. In most of the CFA franc zone countries this was supplemented by a ceiling on domestic arrears—a major financing source in those countries in the recent period. This narrower focus reflected the absence of significant nonbank sources of financing in the CFA franc zone countries, inadequate data to monitor other nonbank financing, or a willingness to accept higher-than-targeted nonbank financing as long as it was arranged at market interest rates.31

Performance criteria on above-the-line items in the fiscal accounts remained rare. Cameroon and the Congo had a performance criterion on non-oil revenues—in the Congo this was combined with a performance criterion on government expenditure—and Mali on total revenues. Seven countries had a performance criterion on all or a portion of government expenditures. Typically, these were included when rising expenditures were the principal source of fiscal pressures and particularly close monitoring of them was needed. In another seven countries, there were performance criteria on credit to, or financial balances of, off-budget public enterprises or special funds.

There was no obvious correlation between the approach to monitoring fiscal performance and the achievement of fiscal targets. In general, however, performance criteria succeeded in capturing deviations from the intent of the program. The coverage of financing items subject to performance criteria appears to have been an issue in two countries (Ecuador and Pakistan), where ceilings only on bank credit meant that fiscal deficits in excess of targets but financed without bank credit were not captured fully in a timely way by the net of quarterly performance criteria: the overrun in the deficit was, however, eventually captured by performance criteria in Ecuador and in a review in Pakistan.

As financial markets develop and, in particular, nonbanks become a more important source of financing for the public sector, performance criteria on bank credit alone (without adjusters for other forms of domestic financing) will provide increasingly inadequate checks on the overall deficit. The lack of a criterion on bond issues to the nonbank sector may also compromise the achievement of external reserve targets, if the bonds prove to be close substitutes for money. This situation appears to have occurred in Tunisia. There, the larger-than-planned issue of treasury bills seems to have substituted to some degree for money in private portfolios; with net domestic assets rising in line with targets, external reserves were below program projections.32 If, however, domestic bonds are closer substitutes for foreign bonds than for domestic money, then the issuance of domestic bonds might actually increase reserves by inducing capital inflows.

Public finances were an important area for structural reforms, and all the countries under review established objectives and priorities in this domain. Furthermore, programs in 21 countries explicitly monitored progress in such reforms. The reforms monitored most frequently related to public enterprise privatization or liquidation (principally of state banks) or to tax reform. On the whole, programs relied far more on reviews than on performance criteria for monitoring progress on structural reforms. This seems appropriate given the inevitable uncertainties in the timing of such reforms and the difficulty of specifying the criteria for assessing observance.

Fiscal Policies in Programs That Went Off Track

Of the 45 arrangements under review, countries were eligible for all purchases in only 17 (Appendix Table 2-A3): 19 programs were on track for the first purchase at the first test date but not for all purchases, and 9 went off track immediately after the initial purchase. In most cases, the critical factor that pushed programs off track was fiscal performance. For the nine programs that went off track immediately after the first purchase, all were affected by deviations from fiscal performance criteria. Of the 19 arrangements that were on track beyond the first purchase but not for all purchases, 14 were affected at various times by deviations from fiscal performance criteria.

Table 2-11.Slippages in Expenditures and Revenues in Programs That Went Off Track1
Expenditure OverrunFailure to Adjust Administered Prices
Revenue ShortfallWagesPensionsPublic enterprisesOther
Algeria 1991X
Argentina 1989XXXX
Cameroon 1988XX2
Côte d’Ivoire 1989XX
Ecuador 1989XX
Gabon 1989X
Haiti 1989X
Hungary 1991X
Jordan 1989XX3
Morocco 1990XX
Pakistan 1989X4
Philippines 1989XX5
Venezuela 1989XXX
Zaїre 1989X
Source: IMF staff estimates.

Above-the-line fiscal deviations from targets during arrangements eligible for more than initial purchase but less than total access.

Goods and services other than wages.

Extrabudgetary expenditures related to the Middle East conflict.

Defense expenditures.

Higher-than-targeted domestic interest payments.

Source: IMF staff estimates.

Above-the-line fiscal deviations from targets during arrangements eligible for more than initial purchase but less than total access.

Goods and services other than wages.

Extrabudgetary expenditures related to the Middle East conflict.

Defense expenditures.

Higher-than-targeted domestic interest payments.

In light of the fact that targets for the overall balance were on average met, this might seem a large role for fiscal slippages. Of course, most countries that were not eligible for all purchases were ones that missed their targets on average for the overall deficit. For the other cases, in which average targets for the overall balance were met but fiscal slippages prevented all purchases, it must be remembered that the relationship between compliance with fiscal performance criteria and the meeting of annual targets for the overall balance is imperfect because performance criteria are usually quarterly and apply to aggregates other than the overall balance. Thus, in two countries (Ecuador and Gabon), programs went off track initially because of fiscal slippages (which subsequently were more than reversed) but remained off track owing to slippages in other policies. Venezuela initially overperformed strongly on the fiscal accounts, only to underperform by smaller but critical amounts in the subsequent program year. Haiti and Zaїre overperformed in the first program year, but went off track early in the second year before fiscal targets for the year as a whole could be set. In Algeria, initial fiscal targets were revised to be more ambitious; the revised target was missed but the initial target was met.

For many countries, remaining on track required waivers of performance criteria on the basis of a positive assessment of countries’ efforts to bring programs back on track. Three quarters of the arrangements eligible for all purchases involved waivers, compared with only half of those that were not eligible for all purchases. Program modifications and extensions were also about twice as common in arrangements eligible for all purchases compared with those that were not. Six countries were eligible for all purchases in all arrangements even though average fiscal performance was worse than targeted: four of these had program modifications (Honduras, with waivers; Mali; Papua New Guinea; and Tunisia) and two set performance criteria while revising fiscal targets as the program proceeded (Bulgaria and Czechoslovakia).

In the 14 countries where fiscal developments were chiefly responsible for programs being off track through the end of the arrangement period, the nature of the main slippage varied (Table 2-11). Revenue shortfalls were the principal reason for missing fiscal performance criteria during arrangements for eight countries. Two countries (Ecuador and Venezuela) failed to pass through higher international fuel prices to domestic petroleum prices, a measure that would have both improved the allocation of resources and raised revenues.

On the expenditure side, wage increases were the most common reason for overruns—occurring in six countries and always in the context of civil unrest or strong union pressure. Apart from wages, government expenditure overruns occurred in Argentina (pensions), Cameroon (goods and services), Jordan (outlays related to the Middle East conflict), Pakistan (defense), and the Philippines (domestic interest payments in 1989). A number of countries also had long-standing problems of expenditure control in public enterprises and related organizations, which re-emerged during their programs (Argentina, Côte d’Ivoire, Ecuador, and Venezuela).

Conclusions

  • In general, the fiscal adjustment strategy was tailored to the different initial conditions and structures of imbalances in the countries reviewed. For many of the countries, initial primary fiscal positions were relatively strong, but long-term fiscal viability required debt relief, which in turn meant that measures were needed to ensure the orderly servicing of debt in the future. For other countries, programs addressed large primary imbalances stemming from high levels of spending (often capital as well as current) or low revenues.

  • Fiscal adjustment during the arrangements was on average close to that targeted. While it is not possible to account quantitatively for the separate roles of discretionary policy changes, exogenous influences, and cyclical factors, it appears that each contributed to some degree. However, the fact that improvements in fiscal balances typically persisted through 1992 (after most of the arrangements under review were completed, and in many countries the effects of favorable terms of trade movements had in part dissipated) suggests that fundamental policy changes had been significant.

  • On average, fiscal adjustment was programmed to rely evenly on revenue increases and expenditure restraint. However, actual expenditure restraint was somewhat greater than planned, and actual revenue increases were lower than planned. Proportionately—that is, relative to initial levels of expenditure—restraint of capital spending was somewhat greater on average than that of current spending. While some of the spending concerned may not have been of the highest quality, it is impossible to be confident that the cuts in capital spending were not damaging to growth prospects. In general, program documents were less forthcoming on the nature and effect of lower-than-targeted spending than was desirable, in large part because at the macroeconomic level it is often difficult to discern such effects. Reviews of public expenditure programs should be encouraged so as to improve the assessment of expenditure restraint.

  • In Central Europe, substantial progress has been made in implementing fiscal reforms and related structural measures, particularly on the revenue side, but tax administration and expenditure control remain a challenge. On the expenditure side, pressures have stemmed in large part from the operation of pensions and other social safety nets. Particularly in light of the unusually high level of government spending relative to GDP and the potential demands on budgets (for enterprise and bank restructuring and environmental cleanup, for example), the restructuring of social safety nets is critically important.

  • Many of the countries reviewed successfully shifted a sizable portion of government financing from banks to the nonbank private sector. Typically, to achieve such a shift, countries had to have (i) recently liberalized interest rates and created auctions for government paper; and (ii) reduced overall financing requirements so that newly issued government paper would not lack credibility. In general, this process appears to have contributed to increases in real interest rates, especially in countries that started their arrangements with real interest rates at negative levels.

  • For the most part, the design of performance criteria appears to have been appropriate to ensuring the twin objectives of controlling the overall borrowing requirement and controlling bank credit to the public sector. In only a few countries were performance criteria too narrowly defined to ensure compliance with the objectives for the overall balance. As financial markets and sources for finance other than the banking system develop, it will be increasingly necessary to ensure adequate comprehensiveness of performance criteria on domestic financing of deficits.

  • While evidence shows that fiscal targets were on average met, it has not been possible to assess the adequacy of the targeted adjustment because only some program documents evaluated fiscal sustainability. Assessments of sustainable fiscal positions are, of course, fraught with difficulty: data required, particularly for domestic debt stocks, are often not easily available; standards for fiscal sustainability are typically highly judgmental; and any concrete analysis is subject to enormous changes as conditions evolve. Nevertheless, greater attention to assessing sustainable fiscal positions in the context of medium-term scenarios would help focus attention on needed structural changes and the magnitude of the overall adjustment effort.

Bibliography

    BlejerMario and AdrienneCheasty“The Measurement of Fiscal Deficits: Analytical and Methodological IssuesJournal of Economic LiteratureVol. 29 (December1991) pp. 164478.

    BlinderAlan S. and Robert M.Solow“Analytical Foundations of Fiscal Policy” in The Economics of Public Financeed. by Alan S.Blinder (Washington: Brookings Institution, 1974).

    EasterlyWilliam and others “Good Policy or Good Luck? Country Growth Performance and Temporary ShocksJournal of Monetary EconomicsVol. 32 (December1993) pp. 45983.

    EasterlyWilliam and SergioRebelo“Fiscal Policy and Economic Growth: An Empirical InvestigationJournal of Monetary EconomicsVol. 32 (December1993) pp. 41758.

    El-KuwaizAbdullah“Oil Exports and Public Finance in Arab Countriesin Fiscal Policy in Open Developing Economiesed. by VitoTanzi (Washington: International Monetary Fund1990).

    FischerStanley and WilliamEasterly“The Economics of the Government Budget ConstraintWorld Bank Research ObserverVol. 5 (July1990) pp. 12742.

    IzeAlain“Measurement of Fiscal Performance in IMF-Supported Programs: Some Methodological Issues” in How to Measure the Fiscal Deficited. by M.Blejer and A.Cheasty (Washington: International Monetary Fund1993).

    NsouliSaleh M. and others The Path to Convertibility and Growth: The Tunisian Experience IMF Occasional Paper No. 109 (Washington: International Monetary Fund1993).

    ParkerKaren E. and SteffenKastnerA Framework for Assessing Fiscal Sustainability and External Viability with an Application to India IMF Working Paper No. WP/93/78 (Washington: International Monetary FundOctober1993).

    RobinsonDavid J. and PeterStella“Amalgamating Central Bank and Fiscal Deficitsin How to Measure the Fiscal Deficited. by M.Blejer and A.Cheasty (Washington: International Monetary Fund1993).

    TanziVito“The Budget Deficit in Transition: A Cautionary NoteStaff PapersInternational Monetary FundVol. 40 (September1993) pp. 697707.

Appendix Tables
Table 2-A1.Coverage of Fiscal Accounts
Overall and Primary BalancesFinancing of the Fiscal Balance
AlgeriaCentral governmentCentral government and Rehabilitation Fund balance
ArgentinaCentral government__1
BrazilNonfinancial public sector (NFPS)__1
BulgariaGeneral governmentGeneral government
CameroonCentral governmentCentral government
CongoCentral governmentCentral government
Costa RicaNFPSNFPS
Côte d’IvoireCentral governmentCentral government
CzechoslovakiaGeneral governmentGeneral government
EcuadorCentral governmentNFPS
EgyptGeneral government2General government2
El SalvadorNFPSNFPS
GabonCentral governmentCentral government
GuatemalaCentral governmentNFPS
HaitiCentral governmentNFPS
HondurasCentral governmentNFPS
HungaryConsolidated State BudgetConsolidated State Budget
JamaicaCentral governmentCentral government
JordanCentral governmentCentral government
MadagascarCentral governmentCentral government
MaliCentral governmentCentral government
MexicoNFPS before privatization receiptsNFPS after privatization receipts
MoroccoCentral governmentCentral government
NigeriaCentral governmentCentral government
PakistanGeneral governmentGeneral government
Papua New GuineaCentral governmentCentral government
PhilippinesCentral governmentNFPS
PolandCentral governmentConsolidated general government
RomaniaGeneral governmentGeneral government
Trinidad and TobagoCentral governmentNFPS
TunisiaCentral governmentCentral government
UruguayCentral governmentNFPS
VenezuelaNFPSNFPS
YugoslaviaGeneral government__1
ZaїreCentral government__1
Source: IMF reports.

Not included in sample because of data inconsistencies.

Including investment by public enterprises.

Source: IMF reports.

Not included in sample because of data inconsistencies.

Including investment by public enterprises.

Table 2-A2.Financial Fiscal Performance Criteria and Adjusters
Performance Criteria
Government or Public Sector Borrowing Requirement (PSBR)

  • Argentina, 19891

  • Brazil, 1988

  • Jamaica, 1988, 1990, and 19911

  • Philippines, 1989 EFF and 1991

  • Uruguay, 19911

  • Venezuela, 1989 EFF1

Government/PSBR and credit to the government

  • Algeria, 19912

  • Côte d’Ivoire, 19892

  • Guyana, 1990

  • Hungary, 1990 and 1991 EFF

  • Mexico, 1989 EFF

  • Morocco, 1988 and 1990

  • Papua New Guinea, 19902

  • Poland, 1990 and 1991 EFF

  • Trinidad and Tobago, 1989 and 1990

Domestic financing

  • Costa Rica, 1989 and 1991

  • Guatemala, 1988

  • Haiti, 1989

  • Honduras, 1990

  • Yugoslavia, 1990

Credit to the government

  • Bulgaria, 1991

  • Cameroon, 1988

  • Congo, 1990

  • Czechoslovakia, 1991

  • Ecuador, 1989

  • Egypt, 1991

  • El Salvador, 1990

  • Gabon, 1989

  • Jordan, 1989

  • Madagascar, 1988

  • Mali, 1988

  • Nigeria, 1989 and 19913

  • Pakistan, 1988

  • Romania, 1991

  • Tunisia, 1988 EFF

  • Zaїre, 1989

Domestic arrears
  • Congo, 1990

  • Côte d’Ivoire, 1989

  • Gabon, 1989

  • Mali, 1988

  • Morocco, 1988 and 1990

External arrears
  • Argentina, 1989

  • Brazil, 1988

  • Cameroon, 1988

  • Congo, 1990

  • Costa Rica, 1989 and 1991

  • Côte d’Ivoire, 1989

  • Ecuador, 1989

  • Egypt, 1991

  • El Salvador, 1990

  • Gabon, 1989

  • Guatemala, 1988

  • Guyana, 1990

  • Haiti, 1989

  • Honduras, 1990

  • Jamaica, 1988, 1990, and 1991

  • Jordan, 1989

  • Madagascar, 1988

  • Mali, 1988

  • Morocco, 1988 and 1990

  • Nigeria, 1989 and 1991

  • Tunisia, 1988 EFF

  • Venezuela, 1989 EFF

  • Zaїre, 1989

Adjusters
Countries with full downward adjusters to ceilings on credit or domestic financing for excesses in external financing
  • Cameroon, 1988

  • Czechoslovakia, 1991

  • Ecuador, 1989

  • Egypt, 1991

  • El Salvador, 1990

  • Gabon, 1989

  • Madagascar, 1988

  • Mali, 1988

  • Nigeria, 1989 and 1991

  • Romania, 1991

Countries with full downward adjusters to ceilings on credit or domestic financing for excesses in nonbank domestic financing

  • Bulgaria, 1991

  • Czechoslovakia, 1991

  • Madagascar, 1988

  • Romania, 1991

  • Zaїre, 1989

Countries with two-way adjusters to performance criteria for deviations in external financing

  • Bulgaria, 19914

  • Costa Rica, 19915

  • Guyana, 19906

  • Jamaica, 1988, 1990, and 19917

  • Philippines, 1989 EFF5

  • Zaїre, 19898

Countries with two-way adjusters to performance criteria for other deviations
  • Algeria, 19919

  • Brazil, 198810

  • Ecuador, 198911

  • Mexico, 1989 EFF12

  • Morocco, 199013

  • Nigeria, 1989 and 199114

  • Poland, 1990 and 1991 EFF15

  • Venezuela16

Source: IMF reports.

Including central bank.

Ceiling on overall central government balance measured above the line.

In 1989 SBA, net credit to the public sector included central bank losses on account of interest payments of the central government.

Full adjustment for excess foreign borrowing and partial adjustment for shortfalls in external financing.

Full adjusters.

Ceiling on PSBR of the nonfinancial public sector adjusted fully for excess external project financing. Ceiling on net domestic financing adjusted fully for nonpayment of external debt service and partially for shortfalls in external financing.

Full adjustment for excesses and partial adjustment for shortfalls in foreign project disbursements.

Full adjustment for nonpayment of external debt service and partial adjustments for shortfalls in external aid.

Partial adjustment for deviations in the exchange rate from baseline projections and negative deviations in hydrocarbon prices and full adjustment for positive deviations in hydrocarbon prices.

Additional performance criterion on operational balance of the nonfinancial public sector.

Partial adjustment for deviations in oil prices.

Additional performance criteria on primary and operational balances of the nonfinancial public sector. Adjustment for deviations beyond threshold in oil price and LIBOR from baseline projections.

Revenues from privatization and tax amnesty.

Full adjustment for higher-than-projected oil prices.

In 1991 EFF, credit ceiling adjusted symmetrically for deviations in debt service and cost of debt reduction.

Symmetric adjusters for deviations in oil revenues.

Source: IMF reports.

Including central bank.

Ceiling on overall central government balance measured above the line.

In 1989 SBA, net credit to the public sector included central bank losses on account of interest payments of the central government.

Full adjustment for excess foreign borrowing and partial adjustment for shortfalls in external financing.

Full adjusters.

Ceiling on PSBR of the nonfinancial public sector adjusted fully for excess external project financing. Ceiling on net domestic financing adjusted fully for nonpayment of external debt service and partially for shortfalls in external financing.

Full adjustment for excesses and partial adjustment for shortfalls in foreign project disbursements.

Full adjustment for nonpayment of external debt service and partial adjustments for shortfalls in external aid.

Partial adjustment for deviations in the exchange rate from baseline projections and negative deviations in hydrocarbon prices and full adjustment for positive deviations in hydrocarbon prices.

Additional performance criterion on operational balance of the nonfinancial public sector.

Partial adjustment for deviations in oil prices.

Additional performance criteria on primary and operational balances of the nonfinancial public sector. Adjustment for deviations beyond threshold in oil price and LIBOR from baseline projections.

Revenues from privatization and tax amnesty.

Full adjustment for higher-than-projected oil prices.

In 1991 EFF, credit ceiling adjusted symmetrically for deviations in debt service and cost of debt reduction.

Symmetric adjusters for deviations in oil revenues.

Table 2-A3.Eligibility for Purchases1(In percent of GDP)
Eligible Purchases as Percent of Total Scheduled PurchasesOverall Fiscal Balance
Pre-arrangement yearTargeted changeActual changeActual change to last year of arrangement
1. Countries eligible for initial purchase only
Brazil 1988 (1)14–30.6–6.0–18.5–18.5
Congo 1990 (2)14–10.8–3.3–1.8–6.9
Costa Rica 1989 (1)25–0.60.8–2.6–2.6
Guatemala 1988 (1)17–2.5–0.1–1.2–1.2
Madagascar 1988 (1)25–5.01.2
Nigeria 1991 (1)17–5.90.3–3.8–3.8
Poland 1991 (1)251.0–1.4–5.6–5.6
Uruguay 1990 (1)170.2–0.90.40.4
Yugoslavia 1990 (1)141.5–0.7–1.8–1.8
Average19–5.9–1.1–3.9–4.4
2. Countries eligible for more than initial purchase but less than total access
Gabon 1989 (2)29–9.45.15.67.6
Ecuador 1989 (2)43–1.82.84.25.8
Haiti 1989 (1)50–1.30.10.30.3
Jordan 1989 (2)50–23.76.74.25.3
Philippines 1989 (2)50–3.31.0–0.3–0.8
Hungary 1991 (3)560.6–2.4–6.5–7.7
Pakistan 1989 (2)57–8.52.11.42.0
El Salvador 1990 (2)60–5.92.82.41.5
Poland 1990 (1)60–8.17.36.76.7
Zaїre 1989 (1)60–22.36.210.710.7
Costa Rica 1991 (2)67–3.13.93.13.4
Cameroon 1988 (2)71–5.22.9–0.4–2.2
Jamaica 1988 (2)71–0.9–0.6–2.6–0.6
Algeria 1991 (1)753.61.21.71.7
Morocco 1990 (1)75–5.72.92.42.4
Romania 1991 (1)80–0.3–1.2–0.6–0.6
Argentina 1989 (1)83–16.114.910.310.3
Côte d’Ivoire 1989 (1)83–12.45.04.64.6
Venezuela 1989 (3)83–7.85.98.26.0
Average262–6.42.92.52.6
3. Countries eligible for total access
Bulgaria 1991 (1)100–13.09.5–2.5–2.5
Czechoslovakia 1991 (1)1000.40.5–2.1–2.1
Egypt 1991 (2)100–20.910.612.212.2
Guyana 1990 (2)100–49.5–13.3–3.96.3
Honduras 1990 (2)100–8.73.21.82.8
Hungary 1990 (1)100–0.91.41.61.6
Jamaica 1990 (1)100–0.7–1.23.43.4
Jamaica 1991 (2)1001.7–0.41.01.1
Mali 1988 (2)100–10.30.70.40.7
Mexico 1989 (4)100–12.99.610.213.7
Morocco 1988 (2)100–6.11.60.6–0.1
Nigeria 1989 (1)100–11.43.07.67.6
Papua New Guinea 1990 (1)100–7.00.2–3.1–3.1
Philippines 1991 (2)100–3.92.11.82.4
Trinidad and Tobago 1989 (1)100–6.82.72.62.6
Trinidad and Tobago 1990 (1)100–4.10.73.33.3
Tunisia 1988 (5)100–3.6–0.3–0.41.3
Average3100–6.82.72.42.8
Source: IMF staff estimates.

Number of program years covered by arrangement(s) in parentheses. Targets and actual developments recorded up to 1992 only.

Excluding Argentina, where the program data and the outturn data are not comparable.

Excluding Guyana, where GDP is subject to measurement problems.

Source: IMF staff estimates.

Number of program years covered by arrangement(s) in parentheses. Targets and actual developments recorded up to 1992 only.

Excluding Argentina, where the program data and the outturn data are not comparable.

Excluding Guyana, where GDP is subject to measurement problems.

See Fischer and Easterly (1990) for a summary of the theory of public finance, particularly the effects on the economy’s savings-investment balance and the effects of deficit financing on inflation, crowding out, and external debt.

The seminal discussion of fiscal sustainability was Blinder and Solow (1974), but Parker and Kastner (1993) have proposed a practical framework for assessing fiscal sustainability in IMF-supported programs. Recent staff reports for developing countries have incorporated analyses of fiscal sustainability. For many countries, adjustments to domestic interest rates and changes in patterns of financing have significant implications for domestic debt service and thereby fiscal sustainability in the future.

See Blejer and Cheasty (1991) for an extensive review of the literature on the measurement of fiscal deficits.

Tanzi (1993) points out the difficulties of measuring fiscal deficits when the public sector represents a large portion of the economy and the risks that strict fiscal targets in such a setting can promote inefficient transfers among levels of the public sector.

The financing of the Mexican fiscal deficit is considered after privatization receipts.

Guyana, where the deficit of the nonfinancial public sector at the start of the arrangement was 50 percent of GDP, is excluded from the analysis because of likely inaccuracies in the GDP data.

Countries with several previous arrangements comprise Argentina, Brazil, Costa Rica, Côte d’Ivoire, Ecuador, Haiti, Hungary, Jamaica, Madagascar, Mali, Mexico, Morocco, the Philippines, Uruguay, Yugoslavia, and Zaїre. Countries with one previous arrangement comprise Algeria, the Congo, Egypt, Gabon, Nigeria, and Tunisia. New users of Fund resources, outside Central Europe, comprise Cameroon, El Salvador, Guatemala, Guyana, Honduras, Jordan, Pakistan, Papua New Guinea, Trinidad and Tobago, and Venezuela. Central European countries comprise Bulgaria, Czechoslovakia, Hungary, Poland, and Romania. See also Chart 1-1.

The oil producers with arrangements under review were Algeria, Cameroon, Ecuador, Gabon, the Congo, Mexico, Nigeria, Trinidad and Tobago, and Venezuela. El-Kuwaiz (1990) details the fluctuations in fiscal revenues of oil exporters and considers the implications of declining oil prices for development programs.

See Easterly and Rebelo (1993) for a cross-country analysis of fiscal revenues, which shows that low-income countries rely relatively heavily on trade taxes and generally have small tax bases.

The frequent revisions to past data made subsequent to the formulation of programs means that the objectives for fiscal adjustment must be considered in terms of targeted changes relative to the situation believed to prevail at the inception of the program, rather than relative to the most recent estimate for the preprogram year.

This relationship between targets and starting positions was tested using a simple bivariate cross-section regression. On the assumption that the targeted level of a variable would be related to some long-run objective with a partial adjustment toward this objective, targeted changes in the overall deficit were related to preprogram levels and a constant. The resulting regression for the overall balance showed a strong relationship, with an R2 of 0.52, implying a relationship significant at the 95 percent confidence level (for 33 countries, excluding Argentina, Brazil, and the Congo).

Some countries envisaged a deterioration of deficits within program periods, if not for the program period as a whole. These were Jamaica (for cyclical reasons in 1990–91), Mali (restructuring of public enterprises and banks in 1989), and Venezuela (program modification in 1991).

The impact of reschedulings or debt reduction operations was not factored into program projections when negotiations were not complete, lest delay or failure to complete invalidate all the program assumptions.

As with the overall balance, a partial adjustment model was postulated (for the 32 countries which targeted the primary balance). Regressing the targeted change in noninterest expenditures on the initial level produced an R2 of 0.56, while regressing the targeted change in revenues on the initial level gave an R2 of 0.40. Both implied a relationship significant at the 95 percent confidence level.

The six countries where programs went off track immediately after approval (Brazil, the Congo, Guatemala, Madagascar, Uruguay, and Yugoslavia) are excluded from the rest of the analysis in fiscal outcomes. The average deviation in the overall balance for these six countries was –2.2 percent of GDP, a large part of which is accounted for by the significant deviation from target for Brazil. Three countries—Argentina, Jordan, and Guyana—had major revisions to or distortions in GDP data that prevented a clear measurement of deficits relative to GDP.

An unpublished review of the problems encountered during 22 programs approved between 1985 and mid-1988 in meeting fiscal targets found that while technical factors such as the ability to monitor and control expenditures had an impact on the budgetary outcome, other factors such as delays in policy implementation and expenditures associated with emergencies had a greater impact. Also, an important source of slippage was that original estimates for expenditures (wages, debt service, and transfers to enterprises) proved unrealistic.

Easterly and others (1993) discusses the relative importance of terms of trade shocks versus other factors in influencing economic performance.

Deviations from targeted capital expenditures in relation to GDP were negatively correlated with the size of targeted reductions in capital spending. Regressing the former on the latter produced an R2 of 0.53, implying a statistically significant relationship. Deviations in noninterest current spending were similarly related to the size of targeted cuts, but rather less strongly with an R2 of only 0.35.

Arrangements encompassed 1992 for five countries. All these countries had either multiple or extended arrangements, so the comparison of 1992 with the program period includes progress made during arrangements.

Cameroon, Czechoslovakia, Ecuador, El Salvador, Hungary, and Romania.

Argentina, Bulgaria, Côte d’Ivoire, Gabon, Jordan, Morocco, and Papua New Guinea had successor stand-by arrangements in place in 1992. Honduras had an ESAF. Also, 1992 was included in multiyear arrangements for Costa Rica, Egypt, Jamaica, Mexico, the Philippines, and Tunisia. The only country that did not have an arrangement in 1992 was Trinidad and Tobago.

Averages for financing differ from above-the-line deficits discussed above because different countries are excluded owing to inadequate data and because data are reported for different levels of government.

Advanced reformers outside Central Europe that began their arrangements with negative real interest rates comprised Argentina, Brazil, Costa Rica, Ecuador, Egypt, El Salvador, Guyana, Honduras, Nigeria, Trinidad and Tobago, and Venezuela. Argentina, Brazil, and Guyana are excluded from the averages in Tables 2-6 through 2-8 owing to problems with the data. Slow reformers outside the CFA zone (almost all of which also began with negative real interest rates) comprise Algeria, Haiti, Jordan, Madagascar, Pakistan, Yugoslavia, and Zaїre. Yugoslavia and Zaїre are excluded from the averages owing to incomplete data.

Advanced reformers that had positive real interest rates in the year prior to their arrangements comprise Guatemala, Jamaica, Mexico, Morocco, Papua New Guinea, the Philippines, Tunisia, and Uruguay.

For the purposes of measuring the financing of the fiscal balance, arrears have been considered only for current transactions. Arrears on capital repayments are assumed to net out from amortizations below the line. This definition of arrears represents a much narrower coverage of arrears, and concerns fewer countries, than those reported in the balance of payments. Arrears in the balance of payments include those of other public sector entities not covered in the accounts considered here, as well as of the central bank and private sector, and cover capital as well as current account transactions. The problem of arrears on fiscal current transactions, external and domestic, was concentrated in the CFA zone, but was notable also in Bulgaria, Costa Rica, Ecuador, Jordan, Morocco, Nigeria, and Poland.

The remainder of the discussion of fiscal financing excludes the six countries where programs went off track immediately after approval (Brazil, the Congo, Guatemala, Madagascar, Uruguay, and Yugoslavia) and where data on financing were not completely available (Argentina, Honduras, and Zaїre). Among slow reformers, only Pakistan had access to nonbanks (via tap issues of bonds) but because interest rates on these bonds were unusually high, programs aimed to reduce reliance on such bonds and ultimately to introduce treasury bill auctions.

The reduction in the average ratio of nonbank financing to GDP between the program average and 1992 for the advanced reformers starting with negative real interest rates reflects the large influence on the average of Nigeria, where the retraction of financial reforms in 1991 was accompanied by an increase in the deficit and a sharp shift from nonbank to bank financing.

A few countries had adjusters for other items such as deviations from projected receipts from privatization and tax amnesty (Morocco) and from projected foreign interest rates (Mexico). Many adjusters had caps or thresholds so they were purely symmetric only within specific ranges of deviations from targets.

See Ize (1993) for a discussion of the merits of allowing adjustments to fiscal targets in response to external shocks.

All countries in this third group had targets for the overall borrowing requirement within which these performance criteria were to be observed, but these targets were not subject to performance criteria.

This process was significant even though the elasticity of substitution between domestic bonds and money is estimated to have been low—only about 0.1; see Nsouli and others (1993).

    Other Resources Citing This Publication