Chapter

CHAPTER 8 Moving Ahead with Structural Reform

Author(s):
Susan Schadler, and Hugh Bredenkamp
Published Date:
June 1999
Share
  • ShareShare
Show Summary Details
Author(s)
Jorg Decressin, Zia Ebrahim-zadeh, Louis Dicks-Mireaux and Ali Ibrahim1 

Large, inefficient public enterprise sectors and distressed banking systems existed in virtually all of the 36 ESAF countries covered by this review and weighed heavily on the finances of the budget and private sector entrepreneurship. They thus hampered efforts to achieve durable macroeconomic stability.2

During the 1980s, the public enterprise sector accounted for a sizable share of production, investment, and formal sector employment in most of the ESAF countries under review (Table 8.1)—this snapshot of the public enterprise sector is based on fragmentary data that is quite partial in its coverage of public enterprises. Public enterprises were involved in a wide range of activities, typically with a dominant presence in key sectors such as mining, manufacturing, utilities, agricultural marketing boards, and trading monopolies. Moreover, they often benefited from preferential treatment through monopoly or monopsony rights, pricing and marketing controls, import and export licensing, quota systems together with preferential access to rationed foreign exchange, and subsidized bank credit. The evidence points to a sector with significant financial losses, low productivity, and dissaving that was a burden on the budget and banking sector, and on the economy in general.

Table 8.1Role of Public Enterprises in Economic Activity, 1978–911(Unweighted averages)
Non-ESAF-Eligible Developing Countries2ESAF-Eligible Countries3
In percent
Share of public enterprises
GDP9.313.3
Gross domestic investment17.126.7
Gross domestic credit (including gross credit to government)7.913.7
External debt15.412.1
Employment44.516.9
In percent of GDP
Overall balance of public enterprises before transfers5-1.5-1.8
Net financial flows from government5-0.80.6
Source: World Bank (1995a).

Complete time series for each variable are not available for many countries; three-year moving-average estimates were used to replace missing observations. For many countries, the data cover only part of the public enterprise sector.

Low- and middle-income developing countries. The number of countries differs for each variable: GDP—24 countries; gross domestic investment–32 countries; gross domestic credit–16 countries; external debt–34 countries; employment–11 countries; overall balance–29 countries; and net financial flows–24 countries.

The number of countries differs for each variable: GDP—16 countries; gross domestic investment—23 countries; gross domestic credit—20 countries; external debt—40 countries; employment—10 countries; overall balance—17 countries; and net financial flows—13 countries.

For some countries, mainly African, expressed as a percentage of formal sector employment.

The overall balance is defined as revenues less expenditures. Net financial flows include government loans, equity, and subsidies as well as public enterprise tax and dividend payments. These data do not take account of indirect flows between the government and public enterprises, such as quasi-fiscal subsidies including preferential credit terms and the forgiveness of taxes. In cases where evidence of such flows exists, they have been found to be large. Thus, the figures reported in the table likely give a positively biased picture of the financial performance of public enterprises.

Source: World Bank (1995a).

Complete time series for each variable are not available for many countries; three-year moving-average estimates were used to replace missing observations. For many countries, the data cover only part of the public enterprise sector.

Low- and middle-income developing countries. The number of countries differs for each variable: GDP—24 countries; gross domestic investment–32 countries; gross domestic credit–16 countries; external debt–34 countries; employment–11 countries; overall balance–29 countries; and net financial flows–24 countries.

The number of countries differs for each variable: GDP—16 countries; gross domestic investment—23 countries; gross domestic credit—20 countries; external debt—40 countries; employment—10 countries; overall balance—17 countries; and net financial flows—13 countries.

For some countries, mainly African, expressed as a percentage of formal sector employment.

The overall balance is defined as revenues less expenditures. Net financial flows include government loans, equity, and subsidies as well as public enterprise tax and dividend payments. These data do not take account of indirect flows between the government and public enterprises, such as quasi-fiscal subsidies including preferential credit terms and the forgiveness of taxes. In cases where evidence of such flows exists, they have been found to be large. Thus, the figures reported in the table likely give a positively biased picture of the financial performance of public enterprises.

Since the early 1980s, virtually all of the countries in this review have experienced significant banking sector problems.3 In at least one-third of the nontransition countries, more than half of the banking system’s loan portfolio was nonperforming at the time of the distress (see Table 4.6 in Chapter 4)—in large part, often a result of the weak finances of public enterprises. Although the proximate causes of distress were typically macroeconomic, they also reflected long-standing poor governance that, in turn, stemmed from government interference in bank operations, weak managers and management systems, and an inadequate regulatory and supervisory framework. The poor health of the banking system was a drain on the budget, complicated monetary policy, and was an impediment to efficient financial intermediation.

The need for reform was clear, and public enterprise reform (from the beginning) and systemic bank restructuring (when widespread distress became apparent) were integral parts of most SAF/ESAF-supported programs. However, the last ESAF review (Schadler and others, 1993), which examined the experience under the first six years of the SAF/ESAF, concluded that considerably less progress (compared with other areas of reform) had been made in reforming public enterprises; a more forceful effort was needed to address this issue in close collaboration with the World Bank; and there was a strong need to press ahead with bank restructuring.4 That review also attached importance to improving the consolidated data on the financial position of the public sector so that the origins of public sector imbalances could be accurately identified and corrected, and progress in addressing the problems of public enterprises could be assessed, at least in terms of their financial flows.

The assessment of the last review still holds true. While reforms have moved ahead, public enterprise reform has fallen short of expectations, data on the financial performance of the public enterprise sector are still inadequate, and banking sector reform has continued to be a drawn-out process, with bank distress persisting for several years after the onset of reform in most countries. Further reform is clearly needed. The distortions and interrelated financial weaknesses of the public enterprise and the banking sectors have impeded private entrepreneurship by preempting resources and hampering competition, as well as impairing efficient financial intermediation, thereby constraining business investment. The unhealthy state of these sectors has had important adverse macroeconomic effects, especially through large budgetary costs and by complicating monetary management. Prospects for both sectors are interwoven through their financial relations: the health of the banking sector is a reflection of its clientele, of which public enterprises are a large share, and sound banking practices, based on commercial criteria, would serve to impose financial discipline on public enterprises.

In light of their macroeconomic importance and the protracted pace of reform, this chapter examines—through detailed country case studies—possible reasons for the slow progress in public enterprise reform and banking system restructuring.5 The aim is to identify what could be done in the context of ESAF-supported programs to help to accelerate reform. Given the lead role of the World Bank (and in some countries, other multilateral development banks) in advising on the design and implementation of reforms in both areas, the focus is on those aspects of policy—the linkages between structural reform and financial policies—that fall within the IMF’s area of responsibility. Related issues of how the IMF might better complement the efforts of the Bank and how to improve IMF-Bank collaboration are also examined.

The analysis is based on the experience of a small group of countries: Bolivia, Ghana, Mongolia, Senegal, and Zimbabwe for public enterprise reform; and Bolivia, Ghana, Lao P.D.R., Senegal, and Tanzania for bank restructuring. While a case-study approach offers the advantage of examining the reform process within its specific institutional setting, it raises questions about the generality of the conclusions. For this reason, the countries were selected to be representative of the problems faced and strategies followed in the countries that have used ESAF resources; to include stronger and weaker performers; and to reflect the geographic diversity of the countries under review. Both studies identify government commitment and domestic support for reforms to be a key factor. Issues concerning national ownership of reform programs and the design of social policies in programs to mitigate the costs of adjustment, which are related to the strength of commitment and support for reforms, are examined in detail in the external evaluation of the ESAF (IMF, 1998).

Back to Business: Reforming Public Enterprises

Bolivia, Ghana, Mongolia, Senegal, and Zimbabwe pursued a strategy that initially targeted the divestment of small public enterprises but favored restructuring for large, strategic public enterprises. Since the early 1990s, as the limited success of restructuring became apparent, reform efforts for larger, strategic public enterprises have shifted toward privatization. In general terms, the evidence suggests that progress in improving the efficiency of public enterprise activities and in alleviating the sector’s burden on the budget and banking system was limited. Furthermore, setbacks in some countries had an adverse macroeconomic impact. Overall, the role of the public enterprise sector in the economy was substantially reduced in Bolivia and Mongolia but was little changed in Ghana, Senegal, and Zimbabwe.

The following sections review the initial setting of public enterprise reform programs in the five countries; examine the objectives and strategies of the reforms; assess the achievements and the monitoring of reforms; review collaboration with the World Bank; consider the need for better data; discuss the role of domestic support for public enterprise reform; and conclude by drawing lessons for the design of future ESAF-supported programs.

The Setting

In the five countries, the evidence suggests that before the SAF/ESAF period, public enterprises’ productivity was low, and their activities were plagued by serious inefficiencies and political interference that limited their profitability. The public enterprise sector typically made financial losses that were limited by moderate profits stemming from a few key monopolies—for example, the cocoa marketing board (COCOBOD) in Ghana, the petroleum company (YPFB) in Bolivia, the copper concern (Erdnet) in Mongolia, and the petroleum company (NOCZIM) in Zimbabwe. Notwithstanding the profitability of these public enterprises, the rationale for public ownership in their field of activity was not compelling, since significant market failures were not evident, and often their productivity was low compared with private companies.6 At times, public enterprises fueled financial imbalances that threatened macroeconomic stability: in Bolivia in the early 1980s, in Ghana intermittently during the 1980s and 1990s, in Zimbabwe until 1995–96, and in Mongolia as recently as 1996–97. Also, the weak financial position of the public enterprise sector contributed in important ways to banking system distress in Ghana, Mongolia, and Senegal.

The underlying weakness of the public enterprise sectors in the five countries prevented them from adjusting to a number of developments, notably liberalization measures taken at an early stage of SAF/ESAF-supported programs (such as the dismantling of protectionist barriers), sharp falls in world commodity prices, and droughts. In Bolivia and Ghana, sharp declines in world tin and cocoa prices brought to the fore inefficiencies in the state-run mining sector and COCOBOD, respectively. In Mongolia, the dissolution of the Council for Mutual Economic Assistance (CMEA) trading system exposed public enterprises to a marked change in relative prices. In Ghana, Senegal, and Zimbabwe, droughts exacerbated the difficulties of agricultural production and marketing public enterprises.

The Reform Program

Objectives and Strategies

The primary objectives of public enterprise reform were to improve the efficiency and financial performance of the public enterprise sector. From a macroeconomic perspective this was critical to raising public saving, curbing inflationary pressures, and strengthening banking systems. Also, by reducing the size of the public enterprise sector, more room would be made for private sector activity, which was increasingly seen as the engine of growth. In Mongolia, public enterprise reform was part of the transformation from a centrally planned to a market-driven economy. Appendix 8.1 provides a summary history of public enterprise reforms in the five countries.

Lack of consistent and comprehensive data precluded a full analysis of the public enterprise sector’s economic and financial performance at the outset of the reforms. With the notable exception of Bolivia, the accounts of both small and large strategic public enterprises often did not exist, were of questionable quality, or were incomplete or outdated.7 Even when the accounts did exist, little progress was made in reviewing and consolidating them to build up a picture of the whole public enterprise sector or key parts of it. Thus, in many cases, it was difficult to assess the sector’s role in the economy, and this hampered the design of financial programs and the reform strategy. Hence, in the five countries plans to improve data on the activities of public enterprises, notably their relations with the government budget, were part of reforms supported under the ESAF.

The first step in reform was the preparation of studies to identify problems and solutions. An important element of this was auditing public enterprises, especially large strategic public enterprises—a time-consuming task that had been neglected for many years in most of the countries. The public enterprise reform strategies developed by the five countries in collaboration with the World Bank envisaged the following measures:

  • hardening budget constraints by reducing the provision of budgetary subsidies as well as curbing public enterprises’ access to bank credit;
  • restructuring public enterprise operations in the context of performance or management contracts designed to introduce commercial management standards,8 or divestment of public enterprises through liquidation or privatization;
  • introducing market discipline through regulatory reforms aimed at creating a more competitive environment for public enterprises—an appropriate legal framework and regulatory institutions were also essential for the divestment of large strategic public enterprises; and
  • building domestic support for public enterprise reform and establishing social safety nets.

To assist in the design of these often complex reforms, technical assistance was to be provided by the World Bank and others.

Hardening Budget Constraints

Hard budget constraints on public enterprises serve both to safeguard the fiscal and monetary programs as well as to reinforce the reforms undertaken by the public enterprises. In principle, they can comprise limits on the losses, or access to financing, of the public enterprise sector as a whole, or subgroups or individual public enterprises. In practice, the ESAF-supported programs of the five countries included budget constraints on broad groups of public enterprises, reflecting a focus on the macroeconomic implications of public enterprise reform, rather than on enterprise-specific reform strategies. The “hardness” of these constraints ranged from explicit limits on credit flows to a less rigid monitoring of subsidies within overall budget targets; significant unexpected developments would trigger discussions with the authorities. In general, the indicators monitored were not comprehensive, reflecting the inadequacy of data, and left scope for leakage in the budget constraint.

ESAF-supported programs of all five countries aimed at curtailing budgetary subsidies and net lending to public enterprises. With the exception of programs under Senegal’s first ESAF arrangement (1988–91), where quantified annual reductions in public enterprise subsidies were targeted, programs contained general commitments (rather than specific quantitative targets) to phase out or reduce direct subsidies over the short to medium term. These commitments related either to all public enterprises (Bolivia, Ghana, Mongolia, and Senegal) or a subset (Zimbabwe).9 Although IMF staff monitored direct operating subsidies, they typically relied on the World Bank’s assessment of the public investment program for the appropriateness of capital transfers and net lending (Ghana, Mongolia, Senegal, and Zimbabwe). The ability to monitor, let alone constrain, fiscal and quasi-fiscal financing of public enterprise activities through channels other than direct subsidies, capital transfers, or net lending was virtually precluded by lack of data and transparency in public sector accounts. Such types of financing would include the subsidy implicit in government lending at below-market interest rates, tax and debt-service arrears, and the forgiveness of tax liabilities and outstanding loans, all of which are typically noncash, unrecorded transactions. The granting of government guarantees—a future contingent liability—for domestic borrowing of public enterprises also was not monitored by the authorities and IMF staff in some instances. The presence of these alternative channels effectively softened the constraints on subsidies and net lending.

Limits on domestic credit outstanding to the public enterprise sector as a whole were not set in the five countries; however, developments in bank credit to public enterprises were kept under review in all countries except Senegal, where data were not available.10 Limits on credit to specific public enterprises were set in Ghana. In Mongolia, because of the poor quality of credit data there was a general, rather than a quantified, commitment to end directed credits and ensure their repayment. In Ghana, Mongolia, and Senegal, the extent of directed credit and preferential interest rates was to be reduced. Borrowing by public enterprises from domestic nonbank sources was typically neither monitored nor subject to ceilings, whereas foreign borrowing was monitored under the standard ceilings on external debt.

In Bolivia, the presence of consolidated public sector accounts allowed for a more comprehensive and aggregate approach, and ceilings were set on the deficit and domestic financing of the combined public sector, thus not specifically targeting public enterprises.11 In Zimbabwe, the combined losses of the nine largest loss-making public enterprises were subject to a ceiling; this narrower focus was necessary because of constraints on data availability.12

Restructuring and Divestment13

On theoretical grounds, the principal economic factor that should influence the extent of government intervention in a market, be it through ownership or regulation, is the scope for market failures and, hence, inefficient resource allocation. Typically, this is a particular concern for public utility and transportation companies. The possible benefits from such intervention, however, need to be weighed against the efficiency losses that arise when performance incentives and managerial accountability under public ownership are inadequate; this is a key concern in countries with limited administrative capacity. Recent experience has demonstrated that there can be considerable benefits from, and scope for, reducing government intervention through privatizing significant parts of the activities of public utility and transportation companies, provided that transparent and credible regulatory reforms and institutions are in place (World Bank, 1995a; and Galal and Shirley, 1994). For example, economies of scale in electricity transmission and distribution grids, distribution of water, or local phone services may favor public ownership, but operations for which scale economies are less important—electricity generation, railroad shipping, the retail marketing of electricity, water, and long-distance phone services—could merit privatization (World Bank, 1992).14

In practice, the five countries initially opted to divest only small and medium-sized public enterprises (Table 8.2). For large, strategic companies, reforms aimed to improve performance by restructuring their operations through measures such as performance contracts (Bolivia, Ghana, Senegal, and Zimbabwe), labor force reductions, as well as price and market reforms. At a later stage, the approach shifted toward divestment and private sector management contracts. The reforms were focused on public enterprises that were considered to be key bottlenecks to growth and, in a number of cases, imposed a significant drain on the budget; they accounted for a sizable share of output and formal sector employment.15 The early preference for restructuring over divestment of large public enterprises stemmed from a reluctance of governments to cede control of public enterprises that were considered to be strategically important for development, particularly to foreign investors. Restructuring was also administratively less burdensome and complex to implement than privatization, which required the creation of an appropriate legal and regulatory framework. However, more recently—as the limited success of restructuring, particularly performance contracts, in improving operations has become apparent—the five countries have begun to undertake or plan sales of ownership or management of some strategic companies. Bolivia stands apart for having forcefully adopted such an approach: virtually all strategic enterprises have been divested, and a suitable regulatory framework and supervisory institutions have been established for former public enterprises that enjoy monopoly or near-monopoly status.

Table 8.2Strategies to Reform Public Enterprises1
Monitor or RestructureDivestment
LiquidateCapitalize2Privatize
Bolivia
Small public enterprises++
Large strategic public enterprises+/–+
Ghana
Small public enterprises++
Large strategic public enterprises++/–
Mongolia
Small public enterprises+
Large strategic public enterprises
Senegal
Small public enterprises++
Large strategic public enterprises++/–
Zimbabwe
Small public enterprises+/–
Large strategic public enterprises+
Source: IMF staff estimates.

The signs “+” = yes, and a large number of public enterprises; “+/–” = yes, but only a few public enterprises; and “–” = hardly any or no public enterprises.

Private investors acquire a public enterprise in return for a commitment to undertake capital expenditure.

Source: IMF staff estimates.

The signs “+” = yes, and a large number of public enterprises; “+/–” = yes, but only a few public enterprises; and “–” = hardly any or no public enterprises.

Private investors acquire a public enterprise in return for a commitment to undertake capital expenditure.

Barriers to Entry and Market Discipline

Reforms typically included liberalizing prices of key goods, dismantling regulatory obstacles to private sector participation, streamlining investment approval procedures, opening strategic sectors to foreign participation (Bolivia, Ghana, and Mongolia), and initiating labor market reform (Senegal). Pricing and regulatory reforms to improve the competitive environment of public enterprises were part of the program in varying degrees across a range of sectors in all five countries. In Mongolia, the liberalization of prices, the overhaul of the legal and regulatory framework, and the opening up of the economy to competition were part of the country’s transition to a market economy.

Building Consensus and Social Safety Nets

Consensus building is a complex political process, and the need to build broad support for public enterprise (and other) reform initiatives was addressed in four ways:

  • Policy framework papers (PFPs), prepared jointly with World Bank and IMF staff, served as a vehicle for establishing a consensus within the government for policy reform in a medium-term perspective.
  • Dialogue among IMF staff missions and the authorities, parliaments, labor unions, and other interested parties helped to foster consensus.
  • Some programs made use of popular divestment modalities—such as capitalization16 and the sale of shares at advantageous prices to workers in Bolivia, and voucher schemes in Mongolia.17
  • Social safety nets were put in place to protect laid-off workers. All five countries provided some form of assistance to employees laid off as a result of reforms.18 In most cases, this was limited to severance benefits, but retraining was also provided in Senegal. An important consideration was the budgetary costs of such benefits, especially since public enterprises were often unable to meet their financial obligation under existing end-of-service and pension agreements.

Conditionality and Reform Priorities

The public enterprise reform component of IMF-supported programs was largely drawn from, and supportive of, the strategy and measures contained in World Bank-supported reform packages. Conditionality was attached to measures that were selected to signal the priority given to reforms and to ensure implementation of the program; also it fixed the timing of important measures, for which the Bank-supported program required implementation only prior to the release of the next loan tranche.19 IMF staff relied on World Bank expertise for assessing the realism and appropriateness of the measures and the timetable of reform, as well as monitoring the implementation of specific measures. IMF staff also set financial targets consistent with reforms to harden budget constraints, and in Ghana and Zimbabwe negotiated structural reforms, designed with input from the World Bank, that fell outside ongoing Bank operations. Discussions with the authorities were often held in the context of joint or parallel IMF-Bank missions.

Reform measures covered by performance criteria (PCs), structural benchmarks (SBs), or prior actions (PAs) fell into four broad groups (see Appendix 8.1, Table 8.12):20

  • Measures related to restructuring and divestment.
  • Measures to introduce market discipline, such as the removal of barriers to private competition and the conclusion of performance or management contracts.
  • Financial targets to harden budget constraints, but only in Bolivia, Ghana, and Zimbabwe. The limited use of SBs and PCs in this area reflected, in large part, the spotty availability or absence of financial data, such as bank credit to the public enterprise sector in Mongolia and Senegal. Specific PCs or SBs on budgetary subsidies and net lending to public enterprises were eschewed in favor of more general commitments. In Bolivia, consolidated public sector accounts made it possible to target a comprehensive budget constraint that indirectly constrained the financing options of public enterprises. In Zimbabwe, the combined operating balance of the nine key public enterprises was subject to a PC. In Ghana, a much more selective approach was adopted, and separate PCs were put on bank credit to the cocoa board and the petroleum corporation (GNPC).21
  • Commitments to improve data on public enterprises’ financial performance. However, notwithstanding the poor quality and coverage of data on financial flows between public-enterprises and the budget and banking system, specific conditionality on data collection was used only in Bolivia (1988) and Senegal (1995).

For the first two categories of measures above, IMF staff generally relied on the World Bank to assess compliance.

Achievements

Public enterprise reform has been protracted in Bolivia, Ghana, and Senegal, extending over more than a decade. In Mongolia and Zimbabwe, reforms began more recently, in the early 1990s. It is difficult to reach a definitive assessment of the achievements of public enterprise reform, for a variety of reasons. Lack of adequate data, with the exception of Bolivia, is commonplace. In addition, public enterprise reform programs had multiple objectives, which often differed among countries or changed over time within a single country and often were not expressed in terms of well-defined targets against which to measure progress. Drawing on the evidence does, however, suggest very little progress overall until 1994: restructuring under public ownership had limited success in redressing public enterprises’ financial imbalances and inefficiencies. There was minimal divestment of public enterprises, except in Mongolia; although there was some reduction in the public enterprise sector’s pressure on the budget and banking system in Bolivia, Ghana, and Zimbabwe, it was marred by significant financial slippages in the latter two countries. Since 1994–95, Bolivia, Ghana, and Senegal have begun selling ownership or management stakes in large, strategic public enterprises; but only in Bolivia has this been done in a forceful and successful manner.

Operational Efficiency of Public Enterprises

One of the main objectives of reforms was to improve the efficiency of public enterprise activities. There are several yardsticks by which operational efficiency can be measured, but the data needed to assess adequately progress in strengthening efficiency are hard to come by. Direct measures of operational efficiency such as productivity and unit labor costs are rarely available for the five countries, and indicators of profitability and labor cost shares that do exist must be interpreted with caution because they were affected by changes in administered prices and (sometimes) the assumption of public enterprise liabilities by the government.

The limited evidence available on public enterprise profitability and labor costs in Bolivia, Ghana, and Zimbabwe (Box 8.1) suggests that operational efficiency improved, at least within a sizable number of the larger public enterprises; data are too meager to reach an assessment for Mongolia and Senegal. In Zimbabwe, the improvement in financial performance was linked, to a considerable extent, to the liberalization of prices and the takeover of public enterprise debt by the government. Performance contracts aimed at strengthening operational efficiency were used in all countries except Mongolia, but with mixed results in Ghana and Bolivia and poor results in Senegal;22 in Zimbabwe, the evidence is inconclusive. The effectiveness of these contracts was weakened because they did not include measures to improve the transparency of public enterprise operations needed to monitor adequately managers’ performance, rarely specified performance-based rewards and penalties for managers, and failed to address systematically the structural problems to be tackled. In addition, governments frequently demonstrated little commitment to their role under the contracts.23 In Bolivia, the contracts provided perverse incentives and were quickly abandoned in the face of poor public enterprise performance.

Box 8.1Operational Efficiency of Public Enterprises

In Bolivia, the ratio of personnel costs to operating revenues fell by about 5 percentage points during 1988–94, suggesting an increase in efficiency. At the same time, annual operating profits declined from 2.8 percent of GDP during 1988–91 to about 2.0 percent in 1992–95. However, this drop owed much to severance payments related to public enterprise restructuring—amounting to 2.0 percent of GDP during 1993–95. Also, the drop in hydrocarbon prices, from about $22 to $16 a barrel during 1991–94, cut profits of one of the largest, most profitable public enterprises (YPFB).

In Gbana, the operational efficiency of key public enterprises with performance contracts appears to have improved modestly. The combined profits of 11 of them increased from 2.2 percent of GDP in 1987 to 2.7 percent in 1993; investments climbed from 2.5 percent to 36.8 percent of GDP, including 19.4 percent of GDP from foreign sources for Ghana Airways; and the share of wages and salaries in revenues in 5 of 13 companies under performance contracts declined from 12.1 percent in 1987 to 7.8 percent in 1991; reliable data are not available for the other 8 companies. At least one key company (GNPC) experienced serious financial trouble during 1994–96.

Little information is available on key public enterprises in Mongolia. A World Bank (1996a) study noted that, since the initial wave of divestments (1992–94), the burden of public enterprises on the economy has been increasing because of excessive wages and continued overstaffing. The worsening performance reflects the absence of a reform program for public enterprises kept in the public sector.

In Senegal, lack of data precludes definitive conclusions on the efficiency gains resulting from public enterprise reforms. With the exception of SONATEL, there appears to have been no major improvement in the financial situation of public enterprises under performance contracts (World Bank, 1994b and 1995a). A key contributory factor was that neither public enterprise management nor the government respected obligations under the contracts.

In Zimbabwe, the operating losses of eight of the nine major public enterprises improved considerably, from a combined deficit of 2.0 percent of GDP in 1990/91 to a deficit of 0.5 percent of GDP in 1994/95; losses of the Grain Marketing Board (GMB), however, rose. Of the nine companies, two were profitable in 1990/91 and five in 1994/95. Price adjustments contributed to this turnaround in performance. The restructuring of GMB, including the takeover of liabilities by the government and the liberalization of the maize market, allowed it to earn a modest profit in 1995/96.

The effectiveness of public enterprise restructuring under public management with performance contracts was minimal at best in the five countries. To achieve more progress, adequate incentives for sound management based on commercial criteria need to be introduced in these contracts. Alternatively, greater involvement of the private sector may be desirable. A detailed, comprehensive World Bank study of this issue (World Bank, 1995a) reached a similar conclusion, finding that the larger the involvement of the private sector in reforming enterprise operations, the more performance improved. In particular, the study concluded that management and regulatory contracts with the private sector did a better job of improving performance than contracts with public managers.24

Size of the Public Sector

In most of the five countries, large numbers of public enterprises were divested, but through 1996 these operations were of limited economic significance except in Bolivia (see Appendix 8.1).25 Under Zimbabwe’s ESAF-supported program, no significant divestment has taken place so far because of political concerns—cumulative privatization revenues amounted to a mere 1 percent of GDP up to 1994/95. In the remaining countries (except Bolivia), divestiture has only very recently encompassed large, strategic public enterprises. In Senegal, about 40 small and medium-sized public-enterprises, accounting for 10 to 15 percent of the government’s equity in the public enterprise sector, were divested by end–1995; two strategic public enterprises were privatized by end–1996 through contracting out management (water) and by sale (telecommunications); also, the groundnut marketing board (SONACOS) was brought to the point of sale. Divestiture was somewhat more comprehensive and of greater economic significance in Ghana: out of about 347 active public enterprises at end–1988, divestiture was under way or completed in 101 by end–1995, and approval has been granted or is planned for another 146. These divestments and approvals included four strategic public enterprises, notably the government’s remaining holdings in the Ashanti Goldfields Company. Bolivia stands out for having made by far the most significant strides in privatization, largely since 1994, with most small and large, strategic public enterprises having been divested or capitalized; almost all of the few remaining public enterprises are planned to be divested or capitalized during 1997, leaving the public enterprise sector with a small role in the economy.

In Mongolia, under the central planning system virtually all enterprises of economic significance were publicly owned. The transition to a market economy, which began in 1990, involved a comprehensive divestment effort, including sales under a voucher scheme of 44 percent of state assets held in about 4,500 public enterprises; cash sales of more than 400 public enterprises are planned for the near future. However divestment in strategic sectors has been limited to a 49 percent stake in the telecommunication company. In 1995, an estimated 36 percent of GDP still originated from public enterprises (World Bank, 1996a). An important question mark over the economic impact of Mongolia’s privatization program is that a vast change in public enterprise ownership has not been accompanied by a commensurate change in managers and business practices.

Overall the economic significance of the public enterprise sector changed little in Ghana, Senegal, and Zimbabwe. In Mongolia, the public enterprise sector was substantially cat back, and in Bolivia it was reduced to a small role. In some countries (Senegal and Ghana), programs emphasized for a number of years the sale and liquidation of small and medium-sized enterprises with a limited macroeconomic impact. More recently, in light of the limited improvements achieved with restructuring and the headway made in improving the regulatory environment for domestic activity and foreign direct investment, governments in Bolivia, Ghana, and Senegal have increasingly opted (or begun to opt) for privatizing management or ownership of large public enterprises.

Impact on the Banking System and the Budget

Public enterprise reform contributed to a reduction in the crowding out of the private sector’s access to domestic credit in all five countries except, possibly, Senegal (Table 8.3).26 However, in part reflecting the delayed or incomplete implementation of reforms, all of the countries (except Bolivia and, to a lesser extent, Senegal) experienced episodes in which the public enterprise sector compromised the stabilization objectives of programs. In Ghana (1994) and Mongolia (1994–96), a worsening of the financial position of the public enterprise sector that was accommodated by an expansion of bank credit fueled a surge in inflation well above targeted levels.27 In Mongolia, commitments to phase out and repay directed credits to public enterprises were not fulfilled, contributing to a deterioration of banks’ portfolios and the eruption of a banking crisis in 1996 with adverse financial implications for the budget; the provision of loans to public enterprises had been facilitated by the presence of state-owned banks and government intervention, including the granting of guarantees. In Zimbabwe, the pricing policies of the agricultural marketing boards led to losses that were financed by a sizable accumulation of government-guaranteed domestic nonbank debt. As a result, as part of their restructuring in 1994/95 the government took over loans equivalent to about 6.9 percent of GDP.

Table 8.3Changing Scope of Public Enterprises in the Economy Before and After Reforms
Prereform YearMost Recent Year
Bolivia198819941
Operating revenue (percent of GDP)17.219.1
Bank credit to public enterprises (percent of domestic credit)21.12.3
19881993
Share of public enterprise debt in total external debt9.510.6
Ghana19871994
Operating revenue (percent of GDP)214.011.8
Bank credit to public enterprises (percent of domestic credit)326.618.4
19871993
Share of public enterprise debt in total external debt8.03.3
Mongolia19901993
Estimated share in GDP95.036.0
Bank credit to public enterprises (percent of domestic credit)425.3
19911993
Share of public enterprise debt in total external debt12.14.5
Senegal19881994
Operating revenue (percent of GDP)
Bank credit to public enterprises (percent of domestic credit)
19881993
Share of public enterprise debt in total external debt6.43.5
Zimbabwe19901994
Operating revenue (percent of GDP)20.356
Bank credit to public enterprises (percent of domestic credit)11.35.3
19901993
Share of public enterprise debt in total external debt9.77.9
Sources: IMF staff estimates; and World Bank, Debtor Reporting System, for external debt data.

Data for 1996 are not comparable to the earlier data because several of the largest public enterprises were divested in 1995–96. As a result, operating revenue dropped to 12.4 percent of GDP in 1996.

Includes only 11 key companies (including COCOBOD) with a performance contract and excludes revenue from cocoa export duty.

Comprises cocoa board financing and claims on public entities.

No comparable figure available.

Data refer to fiscal year 1989/90 and to nine key public enterprises only.

Data for revenues after 1990/91 are not available.

Sources: IMF staff estimates; and World Bank, Debtor Reporting System, for external debt data.

Data for 1996 are not comparable to the earlier data because several of the largest public enterprises were divested in 1995–96. As a result, operating revenue dropped to 12.4 percent of GDP in 1996.

Includes only 11 key companies (including COCOBOD) with a performance contract and excludes revenue from cocoa export duty.

Comprises cocoa board financing and claims on public entities.

No comparable figure available.

Data refer to fiscal year 1989/90 and to nine key public enterprises only.

Data for revenues after 1990/91 are not available.

A complete picture of developments in the fiscal burden of the public enterprise sector, in particular with respect to unrecorded quasi-fiscal transactions such as preferential lending activities and unpaid government-guaranteed liabilities of public enterprises, is clouded by data deficiencies. Data on taxes, profits transferred to the budget, and direct subsidies paid from the budget suggest that reforms may have contributed to an improvement inthefiscal situation in Ghana and Zimbabwe but not in Senegal and Mongolia (Table 8.4). In Bolivia, an underlying reduction in the burden of public enterprises on the budget is masked by temporary restructuring costs: preparatory costs for the capitalization of public enterprises, severance payments to redundant employees, and reduced tax payments and transfers of dividends to the government because of reform costs borne by the public enterprises. However, these observed improvements should be tempered by the scattered evidence that, following the curtailment of direct subsidies, indirect subsidies and payment arrears typically rose and were often large.28 To date, sales of large, profitable public enterprises have been rare,29 and privatization revenues have not made a significant contribution to the budget;30 the exception is the sale by the government of Ghana of a large portion of its shares in the Ashanti Goldfields Company.31

Table 8.4Direct Budgetary Flows to and from Public Enterprises
Prereform YearMost Recent Year
Bolivia198819941
Transfers and taxes to budget8.98.2
(percent of general government revenue)(46.0)(35.6)
Subsidies received from budget0.21.5
(percent of general government expenditure)(0.8)(5.1)
Ghana219871993
Transfers and taxes to budget0.00.2
(percent of central government revenue)(0.7)(1.0)
Subsidies received from budget0.00.0
(percent of central government expenditure)(0.3)(0.0)
Mongolia19931995
Transfers and taxes to budget34.311.2
(percent of general government revenue)(43.1)(33.4)
Subsidies received from budget45.37.4
(percent of general government expenditure)(10.7)(16.5)
Senegal19881994
Transfers and taxes to budget
(percent of central government revenue)(…)(…)
Subsidies received from budget1.61.9
(percent of central government expenditure)(9.5)(13.1)
Zimbabwe51990/911994/95
Transfers and taxes to budget60.61.3
(percent of central government revenue)(1.8)(4.7)
Subsidies received from budget2.60.8
(percent of central government expenditure)(6.1)(2.0)
Sources: IMF staff estimates; for Mongolia, World Bank (1996a).

Available data for 1996 are not comparable to the earlier data because several large public enterprises were divested in 1995–96. The 1996 data show transfers and taxes to the budget dropped to 6.4 percent of GDP and 34 percent of general government revenue, and subsidies declined to 0.3 percent of GDP and 1.4 percent of expenditure.

Only 11 public enterprises; does not include GNPC and Ghana Railways because of lack of data. Contributions to the budget exclude the cocoa export duty, which amounted to 3.6 percent of GDP in 1987 and 3.7 percent of GDP in 1993.

Share of total corporate income taxes in GDP and revenue—the bulk of corporate taxes is paid by public enterprises. In 1994, 23 large public enterprises accounted for two-thirds of total corporate income taxes.

Includes indirect subsidies.

Only the nine public enterprises with a performance contract.

Comprises budget item, “Parastatal interest, dividends, and other.”

Sources: IMF staff estimates; for Mongolia, World Bank (1996a).

Available data for 1996 are not comparable to the earlier data because several large public enterprises were divested in 1995–96. The 1996 data show transfers and taxes to the budget dropped to 6.4 percent of GDP and 34 percent of general government revenue, and subsidies declined to 0.3 percent of GDP and 1.4 percent of expenditure.

Only 11 public enterprises; does not include GNPC and Ghana Railways because of lack of data. Contributions to the budget exclude the cocoa export duty, which amounted to 3.6 percent of GDP in 1987 and 3.7 percent of GDP in 1993.

Share of total corporate income taxes in GDP and revenue—the bulk of corporate taxes is paid by public enterprises. In 1994, 23 large public enterprises accounted for two-thirds of total corporate income taxes.

Includes indirect subsidies.

Only the nine public enterprises with a performance contract.

Comprises budget item, “Parastatal interest, dividends, and other.”

The fiscal costs of public enterprise reforms, notably revenue losses resulting from divestiture of profitable public enterprises and restructuring costs, including for severance payments and the development of social safety nets, have deterred more rapid progress. These concerns were in some cases compounded by the uncertainty of the magnitude of costs: when large unanticipated costs emerged, there was little scope within the budget to accommodate them. For example, in Ghana, public enterprises did not have the resources to pay workers end-of-service entitlements, and this constrained the reduction of overstaffing in public enterprises during 1987–90. In 1991, the government stepped in to ensure that retrenched workers received their legal entitlement, but mindful of the cost it placed a cap on severance payments, which were to be paid in installments over five years out of divestment proceeds. By contrast, Bolivia’s 1994 divestment plan for the large, profitable enterprises included at the outset a detailed assessment of the impact on the budget, especially the cost of severance benefit payments and the takeover of public enterprise debt. A relatively robust fiscal position, together with strong external support, allowed Bolivia to move apace with divestment. This also enabled it to opt for a divestment plan, namely capitalization, that was more likely to attract public support while forgoing divestment revenues to cover the costs of reform.

Data deficiencies make it difficult to get a complete picture of developments in the financial burden of the public enterprise sector on the economy hi the countries reviewed. Indications are that progress in relieving the pressure of public enterprises on the banking system and budget was limited and marred by significant slippages in Ghana, Mongolia, and Zimbabwe; and that headway was also minimal in Senegal. In Bolivia, by contrast, improvements in public enterprise operations contributed to raising their net contribution to the budget. However, this impact was partially offset by the short-run costs of restructuring operations that preceded the capitalization program. In Ghana, such costs at times deterred public enterprise reform.

Introducing Market Discipline

Advances in opening markets and strengthening competitive forces in the five countries were uneven: although reforms were considerable in some areas, their impact was diluted by the absence of complementary reforms elsewhere. Only recently have reforms become more comprehensive. Slow progress in reforming the regulatory environment and strengthening the role of market forces constrained the sale of large, strategic public enterprises, the development of competing firms that would expose public enterprises to market discipline, and the replacement of jobs lost as large public enterprises were reformed. By contrast, the successful comprehensive divestment of strategic public enterprises in Bolivia was supported by a complete overhaul of the regulatory environment.

In Bolivia, public enterprise prices and utility tariffs have been set on a cost-recovery basis since the mid–1980s, and comprehensive regulatory changes in strategic sectors (particularly in mining and hydrocarbons, but also in telecommunications) were initiated in the early 1990s with a view to stepping up domestic and foreign private participation as well as competition. A second broad wave of reforms to deregulate further the electricity, telecommunications, mining, and hydrocarbons sectors was introduced in the mid–1990s to support the successful capitalization of the large public enterprises that dominated these sectors.

In Ghana, Senegal, and Zimbabwe, considerable market and regulatory barriers existed when public enterprise reform began, but only modest, and often piecemeal, progress was made in removing them until recently. At an early stage of ESAF-supported programs, producer prices and private sector participation in marketing activities for major agricultural commodities were liberalized to encourage the atrophy of state-owned marketing boards. However, in Ghana and Zimbabwe, the reforms were insufficiently comprehensive to elicit a vigorous private sector response. In Ghana’s cocoa sector, private traders held back because the export of cocoa remained under a public monopoly, thereby perpetuating the de facto monopoly status of the marketing board. In Zimbabwe, despite the abolition of domestic marketing monopolies, production and marketing incentives have been dampened by the continued presence of public monopolies for the import and export of maize, wheat, cotton, and beef, which prevented producer prices from rising to world levels. In other strategic sectors, especially utilities and petroleum, regulatory reform in Ghana and Senegal to allow greater private sector activity and divestment has been introduced since 1994–95.32 In Senegal, although public enterprise reforms commenced in the mid–1980s, major strides in industrial deregulation were achieved only after the CFA franc devaluation in 1994. Reforms included the abolition of special agreements that shelter numerous industries, the opening up of maritime transportation to the private sector, and the overhaul of the regulatory framework for the telecommunications, electricity, and water sectors to permit privatization plans to move ahead in these sectors.33 Among the five countries, labor market regulation that severely limited the scope for layoffs and thereby hampered public enterprise reforms was a source of concern only in Senegal, but here, following initial proposals in 1985, the reform of labor laws has been minimal.

Headway was made by all five countries in reforming laws and procedures governing foreign direct investment. However, progress has been protracted; sizable regulatory barriers remain in Zimbabwe; new procedures have been applied illiberally in Mongolia and Senegal; and the full liberalization of investment codes was implemented at a late stage of reform in Ghana.34 Wary of “recolonization,” some governments have been reluctant to offer assets or management of large, strategic public enterprises for sale to foreigners (Senegal) or even to nonindigenous parts of the local population (Zimbabwe).

The ability to attract foreign direct investment was important for privatization, especially for large, strategic public enterprises. In Bolivia and Ghana, the deregulation of foreign direct investment was an essential precondition for the recent acceleration in the divestment of public enterprises: significant inflows of foreign direct investment are supporting the capitalization of the major public enterprises in Bolivia,35 and in Ghana, about one-half of divestment transactions through 1995—most prominently the Ashanti Goldfields Company—involved foreign investors. The experience of Ghana also suggests that attracting foreign direct investment by divesting larger profitable public enterprises can catalyze broad foreign investor interest in the economy.36

The Record on Conditionality

The majority of SBs (structural benchmarks) and PCs (performance criteria) attached to public enterprise reform measures or budget constraints were met (see Appendix 8.1, Table 8.12). When they were not met, they were typically subsequently implemented at a later stage as PAs (prior actions) for a new arrangement or completion of a review (Bolivia, Ghana, and Senegal). In Zimbabwe, the only one of the five countries in which PCs were not met, the poor record of meeting public enterprise objectives was a key factor in the interruption of the programs. Bolivia was the only case in which a review was held up for a long period pending the divestment of a key public enterprise.

Notwithstanding the relatively good record in meeting PAs, SBs, and PCs in Ghana, Mongolia, and Senegal, the record of public enterprise reform in these countries was weak. The serious financial slippages on account of GNPC during 1994–95 in Ghana, the banking crisis that erupted in Mongolia, and the limited progress in assessing and reducing direct and indirect. subsidies in Senegal suggest that conditionality was not set on the most relevant measures.

IMF-World Bank Collaboration

The World Bank had the lead role in assisting the country authorities with the design and implementation of public enterprise reforms under ESAF-supported programs. The IMF’s involvement was focused on the macroeconomic impact of the public sector reforms—in particular on the budget and banking system—and its implications for the design of stabilization policies.

The reforms were supported by the World Bank under structural and sectoral operations (see Appendix 8.1, Tables 8.13, and 8.14). In Bolivia, Senegal, and Zimbabwe, the early Bank-designed public enterprise reform programs under SAF/ESAF-supported programs were comprehensive in nature. In Bolivia and Senegal, they were accompanied and followed by a broad range of sector-specific World Bank operations. In Ghana, the Bank initially supported public enterprise reform primarily through sectoral operations; a comprehensive program began in 1995. In Mongolia, the Bank’s lending operations have thus far consisted of a few sectoral operations. In Bolivia, Ghana, and Senegal, the Bank provided technical assistance for building domestic capacity for public enterprise management and reform.

The World Bank’s sectoral operations tended to focus on enterprises considered to be key bottlenecks to growth (public utilities in Bolivia, Ghana, Senegal, and Zimbabwe; or public enterprises involved in key exportables such as mining in Bolivia, cocoa in Ghana, and groundnuts in Senegal). The choice of operations was primarily guided by the scope for efficiency gains in the specific public enterprises covered under the operations, rather than by the size of the sectors’ financial burden on the budget and banking systems, although the two objectives were interlinked. The comprehensive reform programs were by design systemic in nature, concerned with, among other things, the financial performance of the public enterprise sector as a whole.

In several respects, collaboration between the IMF and World Bank was close and constructive. Financial programs under ESAF-supported programs were designed to be consistent with Bank-supported public enterprise reform operations, and IMF conditionality was supportive of structural measures contained in them.

SAF/ESAF-supported programs were able to play a relatively effective complementary role to Bank-supported reform programs in Bolivia and Zimbabwe by monitoring the financial performance of public enterprises. The coordination of the World Bank’s and IMF’s operations in the context of Bolivia’s comprehensive capitalization program was founded on a detailed Bank study of the medium-term financial impact of the program. The impact study served as an input to the design of fiscal policy and more broadly to the medium-term macroeconomic framework of Bolivia’s 1994 ESAF arrangement. The impact of public enterprise reforms on the financial program could be seen to be large, and the IMF put high conditionality on the financial bottom line of the public sector and on key structural measures. When completion of a review was held up largely because of nonobservance of PCs set on public enterprise reforms in 1995–96,37 IMF staff consulted with the World Bank in deciding on an appropriate course of remedial action.

In Zimbabwe, Bank-IMF coordination was centered around targets for the combined losses of key enterprises. Under ESAF-supported programs, these losses were closely monitored and subject to a quantitative performance criterion. The specific ceilings on these losses were set in consultation with the World Bank to ensure consistency with the Bank’s policy advice on price and marketing reforms. Even though monitoring covered only part of the public enterprise sector, it proved to be effective in highlighting the need for corrective policies for the Grain Marketing Board (GMB), whose losses contributed to repeated breaches of the performance criterion and a delay in completing a review.

Consistency and complementarity in World Bank and IMF policies were not always possible for a variety of reasons, stemming from inadequate data on the financial performance of public enterprises and interruptions in Bank-supported programs.

In practice, the public enterprise reform strategy—the selection of public enterprises, as well as the modalities and timetable of reforms—was designed by the World Bank and the authorities, with the FMF playing little, if any, active role except in a few instances. The IMF might have been expected to focus on the financial implications of the reform programs. However, lacking comprehensive or timely data, IMF staff could offer little or no advice on the selection of public enterprises for reform in Ghana and Mongolia; nevertheless, in Ghana, the need to reform COCOBOD—a major public enterprise—was clearly apparent early on and was tackled. In Bolivia and Zimbabwe, and even in Senegal (where detailed data on public enterprises’ financial performance were not available), the most egregious lossmakers were well known and were included in the reform programs.

In some cases, because of poor data on the financial accounts of public enterprises, the IMF was not able to identity mounting problems at an early stage and provide advice aimed at ensuring that key lossmakers were covered by reform programs. This shortcoming was most apparent when reforms were centered on sectoral, as against comprehensive, operations, where the risk of major lossmakers falling outside the scope of reforms was greatest, and the focus on financial performance, per se, was less direct. In Ghana, the monitoring of the financial performance of COCOBOD, for which data were available, was the basis for effective coordination with the World Bank: performance criteria on COCOBOD’s recourse to bank financing were consistent with the Bank’s advice on restructuring and producer prices. However, the narrow scope of public enterprise monitoring meant that the reform needs of GNPC went unnoticed (see discussion below). An earlier recognition of the vulnerability of GNPC in Ghana could have made it the target of reforms before the deterioration of its finances put at risk the stabilization objectives of IMF staff-monitored and ESAF-supported programs. In Senegal, the scope of coordination through the monitoring of financial flows was also limited. Here, quantitative limits on direct budgetary subsidies to public enterprises under the 1988 ESAF arrangement were consistent with the Bank’s SAL (Structural Adjustment Loan) III and IV operations.

In Ghana and Zimbabwe, the modalities of the World Bank’s involvement in public enterprise reform did not in two instances provide the Bank with the capacity to respond to new, unexpected adverse developments, either because the events lay outside its ongoing lending operations or because its own reform operations had stalled. In Ghana, inappropriate pricing policies by GNPC led to major losses in 1994 that were financed by the central bank, GNPC was not covered by a Bank-supported reform program; hence the Bank was unable to tackle the needed reforms in the context of a lending operation. In the event, because of the serious macroeconomic implications of the slippages in GNPC’s performance, IMF staff took the lead in negotiating GNPC reforms, relying on World Bank input for their design. In Zimbabwe, the Bank had suspended disbursements of the second tranche under its SAL II operation, in part because of policy disagreements concerning the reform of the steel company (ZISCO). Hence, its engagement in the reform of other major public enterprises, including GMB, was also reduced. As a result, when GMB’s losses continued to accumulate because of the authorities’ reluctance to adjust key prices, IMF staff gave advice on restructuring GMB’s operations during 1995—96, including on pricing policy and debt restructuring as well as the establishment of a strategic grain reserve (with input from the Bank) and Staff retrenchment (based on a U. S. Agency for International Development reform program). IMF staff shared their policy analysis with the World Bank; Bank staff also attended IMF missions’ discussions with the authorities.

The Need for Better Data

The inadequacy of information on public enterprise performance has long been recognized, and both the IMF and World Bank have sought improvements in this area. In Bolivia, consolidated public sector accounts were available since the onset of public enterprise reform. In the other countries, only fragmentary data on public enterprises have been available, and improvements have been modest at best (Table 8.5). In Ghana, Mongolia, and Senegal, repeated commitments were made to improve data and strengthen the capacity for monitoring public enterprise operations. However, little progress has been made in Mongolia and Senegal; in Ghana there has been a modest improvement, but data on public enterprises’ performance remain partial and unreliable and are produced with substantial delay.38 In Zimbabwe, data on the combined financial performance of the nine largest public enterprises were available before the first ESAF arrangement (1992). Under that arrangement, the authorities committed to strengthening the unit of the Accountant General’s office in charge of monitoring public enterprises, with a view to improving the quality of these data. Only limited progress was made in this regard, but in early 1997 the construction of a combined financial position was expanded to cover more public enterprises.

Table 8.5Availability of Public Enterprise Financial Data1
BoliviaGhanaMongoliaSenegalZimbabwe
Overall balance of public enterprise sectorCS*NNS
Operating balance of public enterprise sectorCS*NNS
Budgetary operating subsidiesCCCCC
Budgetary capital grants and net lendingCC**C*C**C**
Indirect budgetary subsidiesCNNNN
Credit from banking systemCCC*NC
Credit from domestic nonbank sourcesCNNNN
Foreign debtCCCCC
Sources: IMF staff estimates; and World Bank reports.

C = comprehensive coverage of the public enterprise sector; S = coverage of key public enterprises only; N = hardly any or no available data; * = data published with significant delays or that are of poor quality; ** = data that are not readily available for IMF missions, but which Bank staff identify in the context of public investment program reviews.

Sources: IMF staff estimates; and World Bank reports.

C = comprehensive coverage of the public enterprise sector; S = coverage of key public enterprises only; N = hardly any or no available data; * = data published with significant delays or that are of poor quality; ** = data that are not readily available for IMF missions, but which Bank staff identify in the context of public investment program reviews.

Why have improvements in information on public enterprises’ financial performance and relations with the budget and banking system been so limited? The compilation and consolidation of public enterprise data are complex and resource-intensive tasks, and countries have been reluctant to allocate the requisite manpower from their limited supply of skilled civil servants. Indeed, the quality of public enterprise data in Bolivia has benefited from World Bank technical assistance-and financial support to improve the data collection and recruitment of skilled, well-paid civil servants for this purpose. Perhaps even more important, governments have encountered difficulties in exposing the financial activities of sometimes politically influential public enterprises that might draw attention to poor governance (Ghana and Senegal). Mindful of their own limited resources, IMF missions in Ghana and Zimbabwe tended to focus their attention on gathering data on the most egregious loss-making public enterprises—usually well-known even in the absence of detailed information—that could threaten the fiscal and monetary program. However, such an approach risks leaving programs open to unpleasant surprises for which more comprehensive monitoring could have provided early signals. Even when attention was focused on smaller groups of public enterprises, information has not been comprehensive. Conditionality, in the form of PCs and SBs, was used to achieve data improvements only in Bolivia (1988–89) and Senegal (1995–96).

The experience of the five countries suggests that better information on public enterprises” performance could have strengthened the identification of the most important problems and helped to guide reform priorities on the basis of the financial impact of public enterprises. More complete data would also have enhanced the ability to broaden and effectively monitor budget constraints, as well as to forewarn of significant worsening in public enterprises’ financial performance, such as occurred in Ghana.

The Role of Domestic Support

Domestic support, including the government’s commitment, played a major role in the extent of progress in public enterprise reform in the five countries. This finding is supported more broadly by a recent study of World Bank public enterprise reform operations (World Bank, 1996b), which concluded that government commitment was a key, if not the most important, influence on the success of public enterprise reforms, but that Bank-supported programs had been over/optimistic in assessing this factor.

The wide scope of public enterprise reform in Bolivia and early on in Mongolia owed in part to strong political and popular support. In Mongolia, this support stemmed from popular dissatisfaction with the socialist planning system. In Bolivia, it grew out of the disenchantment with the hyperinflation and serious economic upheavals in the mid–1980s;39 more recently, the personal commitment of the Head of State has been a driving force behind the divestment program. Among the other countries, government support for public enterprise reform was constrained by social and political concerns: for example, a reluctance in a period of drought to adjust prices for certain basic foods (such as maize) in order to strengthen the financial performance of public-enterprises in Zimbabwe; public enterprise support for the dominant political parry in Senegal (World Bank, 1995a); and antipathy toward the sale of assets or transfer of management of large, strategic public enterprises to foreigners (Senegal) and nonindigenous elements of the population (Zimbabwe).

Reform programs in the five countries contained a number of measures that contributed to budding consensus for reform, notably the development of social safety nets. A major component of these social safety nets was the provision of severance payments. Such payments proved costly, and in some instances were financed by donors.40 In Ghana, the government was unable to afford negotiated severance payments and in 1991 scaled back public enterprise employee layoffs to reduce the severance bill. In this regard, a strategy of cash privatization was advantageous: it generated revenues that could be used to pay for the costs of social safety nets (Bolivia, Ghana, and Mongolia).

Policies that increased the visibility of, and broadened participation in, the benefits of public-enterprise reforms also helped to strengthen domestic support for reforms. Measures of this sort included divestment strategies with widespread popular participation (Bolivia and Mongolia), the flotation of shares on the domestic stock exchange (Ghana), and greater transparency in public enterprise operations (including the publication of annual performance audits) and the reform process. In Bolivia, a public awareness program provided the basis for public debate and ultimately broad acceptance of the reforms, while weakening the position of entrenched interest groups. However, in other countries public knowledge of the weak performance of public enterprises—the amount of resources they were absorbing and the opportunity cost (forgone benefits) of inaction on reforms—was minimal and qualitative, at best. In Ghana, lack of transparency in privatization procedures—shares tended to be sold directly rather than through the stock exchange—fueled criticisms that the process was heavily influenced by vested political and financial interests (World Bank, 1997). Steps to improve information could provide the foundation for more public support of public enterprise reform and divestment.

Lessons for Program Design

Significantly alleviating the economic encumbrance of the public enterprise sector will require the introduction of effective market and financial discipline, as well as appropriate management accountability. Public enterprise reform in the five countries points to several areas for improvement in program design. In general, to strengthen the IMF’s contribution to public enterprise reform and better complement the work of the World Bank (within the overall framework of Bank-IMF coordination), the IMF should focus on the links between structural reforms and financial policies.

  • The limited success in restructuring under public management—leaky budget COB straints and poor results from performance contracts—suggests that greater private involvement is the surest—and possibly the only—way to improve performance durably. The observed shift in this direction should be encouraged.
  • Absent a wholesale disengagement of the state from the public enterprise sector, a decisive reduction in the burden of enterprises kept under public ownership is unlikely until reliable information is secured on the financial performance of the sector. This would help to impose harder budget constraints and greater management accountability. The IMF should take a more proactive role, in cooperation with the World Bank, in improving the transparency of public enterprises’ financial relations with the budget and banking system, by helping to marshal donors’ technical and financial assistance for what is a complex and sizable task. In the short run. much could be gained by focusing on a subset of key public enterprises; coverage of public enterprises should be expanded systematically to reduce the vulnerability of programs to unpleasant surprises, and more generally to promote good governance. Improvements in public enterprise financial information should be included as specific policy targets in program documents, and program conditionality—PAs, SBs, and PCs—should be focused on the requisite measures; progress in this area should also be pursued in the context of pre-and post-program surveillance activities.
  • Programs should ensure more comprehensive monitoring, including through conditionality, of financial flows as part of the effort to impose market discipline, harden budget constraints, and safeguard the macroeconomic objectives of programs. Particular attention should be paid to monitoring off-budget (noncash) and quasi-fiscal support of public enterprises, which appear to have been significant. One such area is the issuance of government guarantees for domestic borrowing by-public enterprises,41 which have encouraged lax lending practices by the banking system and have often burdened the budget.
  • Programs should be more selective in the application of PAs, PCs, and SBs in an effort to strengthen their impact on the success of reform. Such monitoring should be supportive of the World Bank’s reform operations but should be focused on the implementation of measures that are critical to the success of the financial programs—for example, divestment of a particular public enterprise that would have important financial implications or institutional measures that directly pave the way for substantive reforms. In view of the importance of the government’s commitment, PAs should be used more widely.
  • Programs should reinforce the implementation of reforms, planned under the aegis of the World Bank, that are aimed at paving the way for divestment and disciplining the activities of restructured public enterprises by enhancing market forces. Conditionality should be used to locus attention on systemic measures such as the liberalization of investment codes and foreign direct investment regulations as well as the abolition of monopoly rights.
  • When the IMF’s priorities, guided by a budgetary and macroeconomic perspective, are different from that in ongoing World Bank activities, and the macroeconomic consequences of inaction are serious (for example, in Ghana, 1994, and in Zimbabwe, 1995), the IMF should take an active role in advising the authorities on those aspects of policies toward public enterprises that are particularly important for the financial program, in consultation with Bank staff, until the Bank can be more actively engaged.
  • Popular support is a key factor for successful public enterprise reform; hence, consensus building is critical. An important aspect is to bring potentially vulnerable groups such as labor unions into the process of decision making, to address social concerns through the development of social safety nets, and to identify financing for them. Steps to improve information, and greater transparency in the reform process, can broaden public awareness of the costs of inaction and benefits of reform, and thereby strengthen support for them. Divestment strategies could be tailored to provide the public with a tangible stake in reforms.

The Need for Financial Order: Restructuring the Banking System

Since the mid–1980s, Bolivia, Ghana, Lao P.D.R., Senegal, and Tanzania have experienced bouts of systemic bank distress. All five countries addressed the large “stock” (bad loan) problem of distressed banks, often at a high fiscal cost. However, less success was achieved in establishing an appropriate set of incentives and constraints, and strengthening the legal and judicial backing for property rights, to support sound banking practices. Overall, Bolivia, Ghana, and Senegal restored to varying degrees a substantial measure of banking system soundness. In Lao P.D.R. and Tanzania, reforms are still measurably incomplete.

This study focuses on the progress achieved in establishing the structural foundations of a strong banking system. Thus, it does not examine the macroeconomic effects and causes of bank distress, which can be prominent over shorter periods.42 The following sections describe the scope and causes of bank distress in the five countries; examine the strategy—including World Bank-IMF collaboration—and costs of banking sector reforms; assess the major achievements under the bank restructuring programs; review the use of IMF conditionality; and conclude with lessons for the design of future ESAF-supported programs.

The Scope and Causes of Bank Distress

In the mid–1980s, serious financial distress in the banking system became apparent in most of the five countries in the context of ongoing adjustment programs (Table 8.6).43

Table 8.6Extent of Banking Sector Distress
CountryYears of Bank DistressExtent of Bank Distress
Bolivia1985–95Thirty percent of banks’ loan portfolio was nonperforming at end–1985. The bulk of bad loans were held by the three state-owned banks, which accounted for about one-third of the banking sector’s loans: at end–1988, the proportion of bad loans in the three state-owned banks was over 50 percent and in private commercial banks was 10 percent. All major public sector banks were insolvent.
Ghana1985–91About 60 percent of banks’ loan portfolio was nonperforming at end–1988; annual losses of banks were estimated at 4.8 percent of total assets. Seven out of the 11 main banks needed substantial restructuring.
Lao P.D.R.1992-presentThe seven state-owned banks needed recapitalization estimated at 22 percent of the banking sector’s loan portfolio at end–1993.
Senegal1986–91Fifty percent of the banks’ loan portfolio was nonperforming at end–1988. Eight (of 15) banks were severely distressed, with over 70 percent of their loan portfolio nonperforming.
Tanzania1986-presentThe banking system, which comprised five state-owned banks, was insolvent (but functioning) at end–1989, with 50–90 percent of banks’ loan portfolios nonperforming. The National Bank of Commerce—the dominant bank, with over 80 percent of total assets of the banking system—had about 70 percent of its loan portfolio nonperforming.
Source: World Bank country reports.
Source: World Bank country reports.

A sizable share—ranging from 20 percent to over 60 percent—of the banking system’s loan portfolio was nonperforming.44 Although most banks were affected, state-owned banks, which had a dominant presence (as was true in most ESAF countries) except in Bolivia, were the most seriously distressed. Despite the severity of banking distress, it did not degenerate into an open crisis with runs on problem banks. This may have reflected the government’s substantial involvement in banks, which was perceived as an implicit guarantee on depositors’ claims.

In Bolivia and Ghana, the central bank was also burdened by significant quasi-fiscal losses and needed substantial organizational restructuring. In Bolivia, the central bunk suffered annual losses that rose from 0.2 percent to 0.6 percent of GDP during 1988–90, reflecting lending to the public sector at below-market rates, compensation paid to depositors at banks that were closed, and growing administrative expenditures. In Ghana, the central bank had revaluation losses estimated at some 19 percent of GDP at end–1989, stemming from the impact of the large depreciation of the cedi since 1983 on the bank’s negative net foreign asset position.

The proximate causes of banking problems included unstable macroeconomic policies that fueled high inflation (Bolivia and Ghana); trade liberalization that rendered inefficient enterprises uncompetitive and unable to service their debts (Ghana, Senegal, and Tanzania); and exogenous shocks, including droughts (Ghana and Senegal) and declines in the terms of trade (Bolivia, Ghana, and Senegal). The impact of these factors on bank performance was significant because of the narrow production base of the real economy (and limits on banks’ foreign investments) and hence also the loan portfolios of banks. The underlying causes of the difficulties of the banking system, however, were long-standing structural weaknesses: poor bank governance, stemming in turn from substantial government interference in banks’ operations; weak managers and management systems; and an inadequate regulatory and supervisory environment.

Government directives to provide credit at preferential rates to specific sectors or borrowers, notably public enterprises and the agricultural sector,45 played a major role in the buildup of nonperforming loans in all five countries. Such intervention was perhaps greatest in Tanzania, where official annual plans directed banks to extend credits to public enterprises regardless of there creditworthiness and to open branches and agencies throughout the country irrespective of marketing or cost considerations. Also, in Lao P.D.R. the state-owned banks inherited a large number of nonperforming loans to public enterprises and cooperatives from the former monobank, the State Bank. In Bolivia, banks did not lend to public enterprises, and some two-thirds of the banking system’s had loans were held by the public development banks in which government intervention in lending was considerable.46 Banking system profitability also suffered from interest rate controls in the five countries and from high reserve requirements with below-market remuneration (Bolivia and Ghana).

Government ownership and intervention in the banking system provided little incentive for managers (often political appointees) to base their decision making on commercial standards. In all five countries, banks were also suffering from weak managerial and banking skills, high operational costs, and poor accounting and internal control systems. These shortcomings resulted in poor risk assessment, inaccurate reporting on banks’ condition, and misuse or mismanagement of funds.

Inadequate supervision and banking regulations provided little restraint on poor banking practices. In Senegal, inspection and control of banking operations suffered from the division of responsibility between the Central Bank of West African States (BCEAO) and the national authorities. In Tanzania, official banking oversight was not designed to monitor credit quality, capital adequacy, or management performance; instead it focused on the implementation of the annual foreign exchange and credit plans. Also, in Bolivia, Ghana, Lao P.D.R., and Senegal, capital adequacy requirements and guidelines on risk exposure, loan classification, and provisioning were clearly inadequate.

The Bank Restructuring Program: Strategy and Costs

A Medium-Term Strategy

The main objective of bank restructuring reforms was to restore a solvent and efficient banking system capable of efficiently mobilizing and allocating scarce financial resources.47 In all of the countries, the first step was to undertake preparatory studies to assess the health of individual banks and the problems facing the banking sector as a whole. These studies were initiated in most of the countries before or shortly after the start of SAF arrangements. They benefited from substantial technical assistance from the IMF, the World Bank, and major donors.48

The restructuring of individual banks began a year or two later, and was typically supported by a Financial Sector Adjustment Credit (FSAC) from the World Bank, which coincided with the first ESAF arrangement in each of the countries except Senegal. In Bolivia, restructuring was initiated in 1987, in Ghana in 1988, in Tanzania in 1991, and in Lao P.D.R., with Asian Development Bank support, in 1994. In Senegal, bank restructuring began in 1989 (under the second annual ESAF arrangement) within the framework of a broad reform program of the then West African Monetary Union (WAMU).49

The medium-term reform strategy comprised three major components:50

  • financial restructuring designed to address the problem of the stock of nonperforming loans and accumulated losses of distressed banks, involving the cleanup of their loan portfolios, and their recapitalization (including the central bank where necessary);
  • operational restructuring aimed at stemming the flow of losses and improving the banks’ income position, typically requiring a downsizing of banks (closures of unprofitable branches and retrenchment of excess staff) and management reforms;
  • improvements in the operating environment of banks designed to ensure sound banking practices and prevent the recurrence of banking distress, through changes in the regulatory framework, banking supervision, and the structure of the banking system (liquidation, mergers, or privatization).

Although all five countries followed this general scheme, the reform strategy varied according to the particular circumstances of the country, notably the structure of the banking system, the fiscal position, and implementation capacity (Table 8.7).

Table 8.7Main Elements of Banking Reform Strategy
CountryRecapitalizationPortfolio CleanupPrivatizationLiquidationDownsizingChange in ManagementChange in Regulations
Bolivia
Central bankYesYesYesNoYes
State-owned banksYesYesYesYesYesYesYes
Other banksYes1Yes1NoYesYesYesYes
Ghana
Central bankYesNoNoNoYes
State-owned banksYesYesYesYesYesYesYes
Other banksYesYesNoYesYesYes
Lao P.D.R.
Central bankNoNoNoNoYes
State-owned banksYesYesNoNoNoNoYes
Other banksNoNoNoNoNoNoYes
Senegal
Central bankNoNoNoNoYes
State-owned banksYesYesYesYesYesNoYes
Other banksYesYesYesYesYesYesYes
Tanzania
Central bankYesYesYesNoYes
State-owned banksYesYesYesYesYesYesYes
Other banksNoNoNoYesYesYesYes
Sources: IMF staff estimates; and World Bank country reports.

Applies only to 1996 bank restructuring operations.

Sources: IMF staff estimates; and World Bank country reports.

Applies only to 1996 bank restructuring operations.

Financial Restructuring

It was clear from the outset that the government would bear the brunt of the costs of bank recapitalization. In principle, the costs of bank restructuring—arising from closures, recapitalization, and administering the restructuring—should be borne by those who are responsible for the banking sector problems. The allocation of these costs should take account of moral hazard effects on future behavior, the probability of a run on banks, and the costs to the budget. On the grounds of responsibility, there was good reason For the government in the five countries to assume at least the bulk of these costs: it was the major shareholder in most of the distressed banks; it had guaranteed many of the banks’ loans (especially to public entetprises); and, more generally, it had often interfered in lending decisions either through explicit controls on credit allocation and interest rates or through moral suasion. Also, governments were concerned about the potential political and economic repercussions of imposing the costs on depositors and triggering a loss of confidence in the banking system.

Reflecting the weak fiscal position in the five countries, banks were recapitalized in a manner that sought to minimize the immediate cash cost to the budget. Thus, rather than inject new cash, preference was given to soft loans to banks through the central bank and to replacing banks’ nonperforming assets with government and central bank bonds. In Lao P.D.R. and Tanzania, recapitalization was effected mostly by replacing nonperforming loans with government bonds and, in Ghana, by central bank bonds.51 A relatively small amount of new cash was provided to some banks in Lao P.D.R. and Tanzania. In Bolivia, financial support—soft central bank loans—was granted only to the state-owned banks, which were the most seriously distressed; in Senegal, the government provided a soft loan (financed by a highly concessional external loan) to one major private bank. In Ghana, the government assumed a portion of banks’ non-performing loans to public enterprises as payment for past government loans to these banks. Centralized loan recovery agencies were established in all five countries to recoup the fiscal costs of restructuring and to reduce moral hazard effects by imposing some of the burden on borrowers.

Several countries included bank liquidation as an option, if the bank was assessed to be unviable in the future. However, only Bolivia and Senegal actually closed banks at an early stage of reforms. In Bolivia, three private and two private-state joint-venture banks were closed during 1987—88; the central bank compensated most depositors at substantial cost. For the remaining private banks, in which the extent of bad loans was on a smaller scale than that of the state-owned banks, the focus of reform was on strengthening regulation and supervision—the banks would recapitalize without official financial support. In Senegal, the government decided at the outset to liquidate banks because the cost of rehabilitation was considered to be excessive. Thus, five state-owned banks were liquidated, and their performing loans together with an equivalent amount of private sector deposits were distributed to stronger banks. Nonperforming assets were transferred to a loan recovery agency, and counterpart liabilities—mostly to the BCEAO—were assumed by the government.52 The authorities in Senegal also encouraged burden sharing in two private banks by insisting, upon threat of closure, on a cash injection by all shareholders (including the government).

Operational Restructuring

The rehabilitation programs (except in Lao P.D.R.) contemplated substantial operational restructuring aimed at improving banks’ efficiency and restoring profitability.53

Measures typically included:

  • downsizing through the rationalization of the branch network, retrenchment of labor, and other cost-cutting measures;
  • bringing in new management, in some cases from foreign banks; and
  • improving banks’ operating systems, particularly the accounting system, internal control procedures, and risk assessment methods—these measures were supported by significant technical assistance.

In Lao P.D.R., bank restructuring began only in 1993 and initially focused on the financial rehabilitation of the state-owned banks’ balance sheets as part of establishing a two-tier banking system after the breakup of the old State Bank. Hence, operational restructuring and improvements in banks’ operating environment were not started until 1996.

Improvements in the Operating Environment of Banks

Improvements in the regulatory environment and in banking supervision were needed to enhance the soundness of the banking system and to ensure that the financial and organizational restructuring of distressed banks would have a lasting impact. A strong supervisory framework would also strengthen the ability of the authorities to detect banking problems at an early stage. Except in Lao P.D.R., the restructuring programs placed a considerable and early emphasis on improving the regulatory environment and banking supervision. Key reforms in this area included:

  • stricter prudential regulations (notably in Bolivia, Ghana, Senegal, and Tanzania), such as increased capital adequacy requirements, higher standards for loan classification and provisioning, and limits on risk exposures and connected lending;
  • enhanced bank supervision through the creation of independent supervision units (Bolivia and Senegal)54 or strengthening of the supervision capabilities and authority of the central bank (Ghana, Lao P.D.R., and Tanzania), and increased reporting and information disclosure obligations for banks; and
  • increased competition in the banking sector through liberalizing the entry of new private banks and, in particular, foreign banks (Bolivia, Ghana, and Tanzania).55

The institutional capacity to supervise was very weak in these countries, and substantial technical assistance, including the use of resident experts, was provided. This assistance was designed not only to address the immediate need of the program, but also to build up long-term domestic capacity. Most notably, the reform programs in Bolivia and Ghana included training and professional development courses for commercial and central bankers, particularly in accounting, and a strengthening of existing training institutes.

The effectiveness of supervision also depends on the availability and credibility of sanctions. In this regard, with the exception of Bolivia and Senegal, none of the reform programs envisaged the closure of banks at the outset of the program, although the option was left open in Ghana. Even when state-owned banks were clearly not viable from the beginning, as in Bolivia and Tanzania, the authorities attempted first to rehabilitate them. In both these countries, the reluctance to close banks outright reflected concerns about the adverse impact of a closure on employment and on the private sector’s access to bank services, particularly when the bank had a significant branch network. In Tanzania, the dominant state-owned bank (NBC) was considered “too big to fail.” In Bolivia, the risk of undermining confidence in the banking system also played an important role.

Collaboration with the World Bank

The World Bank played the lead role in advising on the design and implementation of bank restructuring reforms, except in Lao P.D.R. where the Asian Development Bank was the primary institution.56 The IMF’s involvement comprised four elements, in line with the overall framework of collaboration with the World Bank (Table 8.8):

Table 8.8IMF Involvement in Bank Restructuring
CountryEnsure Consistency with Financial PoliciesStrengthen Supervision/ RegulationDevelop Instruments of Monetary PolicyAssist in the Reorganization of the Central Bank
BoliviaYesYesYesYes
GhanaYesNoYesNo
Lao P.D.R.YesNoYesYes
SenegalYes(Yes)1(Yes)1No
TanzaniaYesYesYesYes
Source: IMF staff estimates.

IMF involvement in these areas was channeled through the Central Bank of West African States (BCEAO) rather than the national authorities.

Source: IMF staff estimates.

IMF involvement in these areas was channeled through the Central Bank of West African States (BCEAO) rather than the national authorities.

  • to incorporate the fiscal costs of bank restructuring into the budget and overall macroeconomic framework and to ensure that the financial program took account of the monetary implications of the restructuring;57
  • to assist central banks in strengthening the regulatory framework of banking activity and in enhancing the capacity to supervise the banking system;
  • to assist in introducing indirect instruments of monetary policy; and
  • to assist in the reorganization of the central bank and introduction of a new Central Bank Act redefining the responsibilities of the bank (in Bolivia, Lao P.D.R., and Tanzania).

Issues concerning the last two of these areas are not dealt with in this study.

In each of these areas, the IMF’s involvement was complementary to, and devised in collaboration with, World Bank and Asian Development Bank reform programs. In Ghana and Lao P.D.R., the absence of IMF support for strengthening supervision and regulation reflected the provision of assistance by the World Bank and Asian Development Bank.58 In all of the countries, Bank-IMF collaboration benefited from joint or parallel IMF-Bank missions and discussions with the authorities, and also from the participation of IMF experts (from the Monetary and Exchange Affairs Department) and resident representatives.

The reform strategy for bank restructuring—the identification and selection of individual banks, the modalities of bank restructuring, and the timetable—was formulated by the World Bank and Asian Development Bank. While the IMF was well informed about the strategy, thereby ensuring a consistent formulation of the IMF’s financial program and the Bank’s structural program, IMF staff were not actively involved in the design of the restructuring exercise per se, except in Tanzania and Bolivia (in 1995–96).

In Tanzania, since 1994 IMF staff have taken a more active role in the restructuring of NBC—the dominant state-owned bank, whose weak performance bad major fiscal implications. Following the recapitalization of its balance sheet in 1992, NBC continued to generate sizable losses because of a failure to strengthen the efficiency of its operations. Recognizing the likely need for a further recapitalization, at considerable and increasing cost to the budget, IMF staff pressed for the timely implementation of reform actions under consideration by the Bank and authorities. As a result, IMF staff-monitored programs during 1994—96 contained benchmarks on institutional measures that paved the way for a reorganization of NBC’s operations and performance-based targets—namely, a freeze on lending operations until NBC had begun to break even or move into surplus. A target date was specified for achieving profitability. These measures and their timing were chosen in collaboration with the World Bank to ensure their feasibility, and were aimed at having an immediate impact on the financial performance of NBC. At the time, the Bank was in the process of negotiating a successor FSAC, and thus the use of IMF conditionality served a complementary role in sustaining the pace of reform.

In Bolivia, IMF staff (particularly from the Monetary and Exchange Affairs Department) were actively involved during 1995–96 in the World Bank’s program for strengthening the banking system in response to the recent liquidity and solvency problems in four private banks. IMF staff advised on the size and modalities of official support, seeking to minimize the cost to the authorities and ensure a meaningful stake of shareholders in the reforms through an adequate, up-front capital contribution.

IMF staff involvement in banking sector reform can have a considerable payoff, and can complement the World Bank’s policy advice and technical assistance, by injecting a sense of urgency based primarily on a macroeconomic perspective.

Public Enterprise Reform and Bank Restructuring

The nonfinancial public enterprises had been a drain on the profitability of the banking sector in all of the countries except Bolivia. In Bolivia, public enterprises held accounts only with the central bank, and their borrowing was financed by on-lending of external loan disbursements authorized by the government. At the outset of the restructuring programs, public enterprises accounted for at least half of the nonperforming loans held by banks in the four other countries. The financial health of the public enterprise sector affected bank activities for three reasons.

First, directed credit policies favored the public enterprise sector (and did not discriminate adequately between profitable and loss-making public enterprises). As a result, a large share of banks’ lending was absorbed by public enterprises, and the banking system was unable to exercise discipline on public enterprises. Directed lending and preferential interest rates were largely eliminated before or at the outset of the bank restructuring exercises.59 Nevertheless, even after the start of restructuring, government interference in bank lending continued in several countries through informal channels or moral suasion, most notably in Tanzania until 1992.60 The availability of government guaranrees (explicit or implicit) may also have encouraged banks to continue lending to lossmakers; in particular, in Lao P.D.R. Public enterprises were unincorporated and fully owned by the state, and thus any bank loan was in effect a loan to the government. Second, the large role of the public enterprise sector in the economy meant that, even in the absence of credit directives, public enterprises would likely account for a sizable part of banks’ clientele; also, banks’ limited capacity for credit analysis, together with pressures to improve loan appraisals, tended to make it more attractive to lend to familiar borrowers rather than to newer and smaller borrowers. Third, the rate of recovery of non performing loans depended in part on how badly public enterprises ailed.

In light of the close financial links between the two sectors, to what extent were the reform components of the banking and public enterprise sectors coordinated? Public enterprise reform began prior to bank restructuring in Senegal and Lao P.D.R., at about the same time in Ghana, but only a few years after banking sector reform began in Tanzania.61 While many of the reforms undertaken in each sector were complementary, and the two reform programs progressed in parallel in some countries, there does not appear to have been any specific coordination.62 However, even though public enterprise reform had only a limited impact on the performance of the public enterprise sector as a whole, and hence also on banks (in all the countries except Bolivia), measures were taken that helped to insulate the banking system from the problems of public enterprises.

In Senegal, ONCAD (the groundnut marketing company), which was a major source of balance sheet problems for banks, was liquidated in 1982, and SONACOS (the groundnut oil processing company) was restructured in the late 1980s to improve its profitability. These two measures, combined with the reform of the system of crop credits within the WAMU and the government’s decision to abstain from issuing guarantees for domestic public or private borrowing, played an important role in relieving pressures on banks to lend to public enterprises and state marketing boards. In Ghana, two major public enterprises (COCOBOD, the cocoa marketing board, and GNPC, the Ghana National Petroleum (Corporation) that had sizable credit demands and periodically experienced large financial losses underwent restructuring and have adopted more appropriate pricing policies since 1988 (COCOBOD) and 1995 (GNPC). Concomitantly, their credit demands were subject to performance criteria. In Lao P.D.R., a decree was issued in the context of the privatization program that limited state-owned enterprises’ access to bank credit. In Tanzania, NBC continued to lend to loss-making public enterprises until its lending to nonperforming borrowers (both private and public sector) was frozen under the 1994 IMF staff-monitored program.

In sum, the large role of public enterprise liabilities in bank distress in some countries and the limited, if any, coordination of public enterprise and bank reform suggests considerable scope for improvement in this area. Such coordination would address one of the key sources of bank distress and ensure that costly financial support and reforms are not wasted on situations where the problems underlying the banks’ distress are not eliminated.

Fiscal and Quasi-Fiscal Costs of Bank Restructuring

The full costs of bank restructuring (including the absorption of banking system losses and administrative costs) were difficult to predict with precision at the outset of the reform program, in large part because of uncertainty about the size of the losses of distressed banks and of their nonperforming loans. More precise estimates of the budgetary costs of reform were possible only as the restructuring program unfolded and the extent and modalities of the restructuring of balance sheets were better known, at which time they were incorporated in the budgetary accounts.63 An indication of the budgetary costs of bank restructuring, albeit incomplete, is given by the overall up-front cost of the restructuring reform package, which was met by the government (in large part through external financial assistance), and by estimates of the annual debt-service costs associated with bank restructuring based on the terms of the bonds issued by the government and external financing (Table 8.9). These estimates are conservative because they do not take into account the costs of running loan recovery agencies, or of retrenching labor in problem banks, and the loss of claims by some depositors in Senegal and Bolivia when banks were closed; on the other hand, loan recovery revenues, albeit small, also are not included.

Table 8.9Partial Estimates of the Costs of Bank Restructuring(In percent of GDP, unless otherwise noted)
Country1989199019911992199319941995
Bolivia
Cost of reform package15.20.01.70.0
Annual budget debt service0.30.40.3
(in percent of budget revenues)(1.3)(1.6)(1.0)
Ghana
Cost of reform package14.419.02
Annual budget debt service3
Lao P.D.R.
Cost of reform package11.60.0
Annual budget debt service0.10.1
(in percent of budget revenues)(0.8)(0.8)
Senegal
Cost of reform package116.1
Annual budget debt service0.70.60.51.20.90.8
(in percent of budget revenues)(3.6)(3.1)(2.7)(7.4)(6.4)(5.3)
Tanzania
Cost of reform package110.17.60.00.7
Annual budget debt service1.01.61.21.1
(in percent of budget revenues)(7.8)(10.7)(8.3)(7.3)
Source: IMF staff estimates.

The costs are reported in the year in which the reform program began.

Revaluation losses of the central bank which were assumed by the government.

Bank restructuring bonds were also issued by the central bank, which was responsible for interest payments on the bonds; these payments were of the order of 0.4–0.5 billion cedi annually (on average about 0.1 percent of GDP, or 1 percent of budgetary revenues, annually during 1991–95).

Source: IMF staff estimates.

The costs are reported in the year in which the reform program began.

Revaluation losses of the central bank which were assumed by the government.

Bank restructuring bonds were also issued by the central bank, which was responsible for interest payments on the bonds; these payments were of the order of 0.4–0.5 billion cedi annually (on average about 0.1 percent of GDP, or 1 percent of budgetary revenues, annually during 1991–95).

Even on the basis of these conservative estimates, bank restructuring was clearly an expensive operation, notably in Ghana, Senegal, and Tanzania, where the costs surpassed 15 percent of GDP; annual debt service costs during 1993–95 ranged from about 0.3 percent of GDP in Bolivia to 1.3 percent in Tanzania. To help meet these costs, the five countries received external financial assistance. In principle, the high cost of bank restructuring should have been offset, at least partially, by the recovery of bad loans. In practice, however, the recover)’ rate typically was low.64 A large share of nonperforming loans was considered irrecoverable (85 percent in Senegal and Tanzania, and 70 percent in Ghana). In Ghana and Senegal, the modest recovery targets for the remaining bad loans were met, but with delays, over a period of five years; in Tanzania, only a small fraction of loans has been recovered. The slow pace of recovery and extensive writing off of loans reflected to some extent a weak legal and judiciary framework, as well as inadequate loan collaterals.

Notwithstanding that a large share of nonperforming loans was written off as irrecoverable, a strengthened legal and judicial system to support property rights and a more aggressive approach to the recovery of the remaining loans could have mitigated the burden of bank restructuring on the budget. It could also have played an important signaling role to delinquent debtors and potential borrowers and could have distributed the costs of bunk restructuring more equitably.

The presentation of restructuring costs in the budget varied considerably among countries in terms of the classification of costs and the transparency with which restructuring costs were shown.65 Most of the budgetary costs were incorporated in the fiscal targets. The exchange of bonds for bad loans was recorded either as net lending (Tanzania) or under net domestic financing (Bolivia, Senegal, Ghana), with counterpart items in external financing; in Lao P.D.R., the issuance of government bonds was not included in the fiscal accounts. Interest costs were recorded in expenditures but were not separately identified or discussed in IMF staff reports. Estimates of the quasi-fiscal costs, such as the subsidy element of solvency support from the central bank, are not available. In view of the uncertainty and large size of the costs of banking reforms, adjustors were placed on the fiscal targets and performance criteria so that any shortfall in costs compared with the programmed level of financial assistance would not provide room for a loosening of the fiscal position. To this end, programs in Lao P.D.R., Senegal, and Tanzania included asymmetric adjustors on net credit to government, while in Ghana financing operations related to bank restructuring were excluded from the definition of bank credit to government.

Achievements

Although comprehensive measures of the effectiveness pf bank restructuring are not available,66 existing data suggest that improvements in banking system soundness varied considerably (Box 8.2). Among the five countries, Bolivia, Ghana, and Senegal have restored a substantial measure of soundness to their banking system. Such progress reflected (in Bolivia in the later rounds of reforms) the authorities’ strong ownership of the reform program, their decisiveness in moving to rehabilitate or close problem banks, and the substantial technical preparation of the operations. This assessment of banking soundness nevertheless requires some qualification; in Bolivia, the most recent round of reforms has yet to experience the test of time; in Ghana, several cases of fraudulent lending—hut not with systemic implications—were uncovered in 1995; and in Senegal, the national agricultural credit bank (GNCAS) has been kept in operation despite continuing difficulties.

In Ghana and Senegal, steady progress was made in implementing reforms. There have been sustained improvements in banks’ balance sheets and clear signs of an improvement in profitability. In Ghana, the ratio of bank operating expenses to income of the “distressed” banks fell from 64 percent to 45 percent between 1990 and 1993,67 and their combined financial position moved from loss to profit. Whereas in 198” most banks did not satisfy a 6 percent minimum capital adequacy ratio, by 1992 this was more than met by must banks, and loan-loss provisions had fallen (as a percent of assets). Data for the two largest banks for 1994 confirm the positive trend in banks’ financial performance. In Senegal, in 1995 the five major remaining banks were assessed to be solvent, liquid, and profitable (World Bank, 1995e). In Bolivia, the path of reform was uneven. After an initial round of reforms aimed at private banks, state-owned banks were closed in 1992, and the profitability of many private banks improved. In 1995, however, several experienced liquidity problems, stemming from lax lending practices. This triggered a second wave of reforms, including a substantial strengthening of official oversight of the banking system that appears to have arrested the deterioration in banks’ performance.

In Tanzania, bank restructuring has been protracted and costly, largely as a result of the authorities’ reluctance to put the operation of the major state-owned banks on a strict commercial basis, and continued lending by these banks to loss-making public enterprises and cooperatives. However, on a positive note, the authorities intended to implement in 1997 a fundamental reform of NBC (the dominant bank). In Lao P.D.R., the recapitalization of state banks in 1994 was an important first step toward restoring the health of the banking system, but the subsequent reemergence of bad loans indicates that operational restructuring of banks and stronger bank supervision are needed.

Financial Restructuring

In Ghana, Lao P.D.R., and Senegal, the recapitalization of problem banks was completed at an early stage of the reform program. The resulting improvement in the balance sheets of banks in Ghana and Senegal, as measured by the ratio of nonperforming loans to total bank loans, appears to have been sustained (Table 8.10); in Ghana, the share of bad loans is still relatively high but is fully provisioned, and banks are operating at a profit. In part, this was a product of strengthened banking supervision combined with the reform of the system of crop credits and restriction on the use of government guarantees (Senegal),68 and limits on borrowing by major public enterprises (Ghana). However, in Lao P.D.R., where recapitalization took place in 1994, there are renewed signs of had loans because banking supervision remained weak and the operational restructuring of banks has yet to be initiated.

Table 8.10Summary Indicators of Bank Restructuring
Ratio of Nonperforming Loans in Bank Credit (percent)Number of Banks
Postreform
CountryPrereform (year)Latest (year)Before reformsRestructured1PrivatizedClosedNew (of which foreign)Proportion of Staff Retrenched (percent)
Bolivia2252830110
(1987)2(1996)2
Ghana73291172402(2)35
(1989)3(1993)3
Lao P.D.R.22108007(7)5
(1993)
Senegal5014152365040
(1988)(1995)
Tanzania70….51115(5)66
(1989)
Source: World Bank project reports.

Includes banks that underwent both financial and operational restructuring.

The prereform data include state-owned banks (with a ratio of about 50 percent) and private banks (with a ratio of 10 percent). The latest year data comprise only private banks—state-owned banks had been shut down.

Figures refer to six of the seven distressed banks that were restructured.

Privatization of the two banks is partial. In one, the government is still the majority shareholder pending sale of 30 percent of the shares to a strategic investor.

Also, two new domestic banks—joint ventures with local and foreign capital―were opened in 1989 and 1993.

The government reduced its shareholding from 35 to 3 percent in one bank by not participating in subsequent capital increases; from 42 to 25 percent in a second bank through sales of shares; and from 45 to 24 percent in a third bank by selling 4 percent and requesting the West African Development Bank to manage 17 percent.

Source: World Bank project reports.

Includes banks that underwent both financial and operational restructuring.

The prereform data include state-owned banks (with a ratio of about 50 percent) and private banks (with a ratio of 10 percent). The latest year data comprise only private banks—state-owned banks had been shut down.

Figures refer to six of the seven distressed banks that were restructured.

Privatization of the two banks is partial. In one, the government is still the majority shareholder pending sale of 30 percent of the shares to a strategic investor.

Also, two new domestic banks—joint ventures with local and foreign capital―were opened in 1989 and 1993.

The government reduced its shareholding from 35 to 3 percent in one bank by not participating in subsequent capital increases; from 42 to 25 percent in a second bank through sales of shares; and from 45 to 24 percent in a third bank by selling 4 percent and requesting the West African Development Bank to manage 17 percent.

In Bolivia, the anticipated restructuring of the state-owned banks in the late 1980s was delayed because of a lack of political consensus: the authorities decided to reevaluate the role of the state-owned banks, resulting in an increase in the costs of banking reform. During the delay, the central bank provided soft loans (liquidity support) to the banks. In the event, the banks were closed in 1991—92 as prior actions under the World Bank’s SAC; the government assumed the net liabilities by issuing long-term securities to the central bank. In the first restructuring program (1987), the private banks that were not closed were expected to recapitalize themselves through improved profits on the basis of strengthened operating practices—bad loans accounted for about 10 percent of the banks’ loan portfolio. However, during the 1990s the loan portfolios of several banks deteriorated significantly. Thus, two banks were dosed in late 1994, and during 1995—96 official loans were extended to help recapitalize four banks69 subject to strict conditions, including the formulation and implementation of a bank strengthening program and a minimum capital contribution by shareholders.70

In Tanzania, the initial recapitalization of NBC (the dominant bank) in 1992 was smaller than planned because the government did not issue the full programmed amount of bonds to replace non-performing assets of the bank (nor did it fully service the limited amount issued). Moreover, the benefits of the financial restructuring soon eroded, reflecting the lack of commitment of NBC’s management to reform. NBC continued to lend to defaulting public enterprises and cooperatives and reversed some of the earlier reforms, including branch closures and staff retrenchment. As a result, there was a succession of recapitalization exercises in 1993–95 that involved a cumulative issuance of government bonds equivalent to 8 percent of 1995 GDP, and some cash injection. Despite these actions, NBC continued to incur substantial losses until late 1995. Following the start of a more aggressive restructuring plan in 1995, with assistance from the World Bank, progress toward profitability has been made. The plan includes a freeze in lending by NBC, substantial branch closures and staff retrenchment, and increases in fees and lending rates.

The financial restructuring of the central bank also featured importantly in the bank restructuring strategy in Bolivia (1992) and Ghana (1990). In Ghana, the government assumed the Bank of Ghana’s foreign revaluation losses, equivalent to 19 percent of GDP, in exchange for long-term government bonds.71 In Bolivia, the government issued securities equivalent to 2 percent of 1992 GDP to the central bank to cover its quasi-fiscal losses.72

Efforts to strengthen balance sheets cannot be sustained without a distinct break from past inadequate operating practices: financial restructuring should be accompanied by forward-looking improvements in banks’ operations aimed at improving their efficiency and reducing their vulnerability to poor loans. A strong government commitment to the reform process and the installation and enforcement of a regulatory and supervisory system for sound banking are essential prerequisites to this process. Failure to implement operational improvements has resulted in the dissipation of previous financial support and emergence of new fiscal costs.

Operational Restructuring

The operational restructuring of distressed banks through cost-cutting measures and a strengthening of management helped to improve the efficiency and profitability of banks in all five countries except Lao P.D.R., where such reforms have not yet been initiated.73 Staff reduction played a major role in most of the countries: in Ghana and Senegal, cuts of 35-40 percent were implemented; in Tanzania, two major public banks (including NBC) cut their staff by over 65 percent; and in Bolivia, the staff of one of the development banks was reduced by almost 30 percent. In Tanzania, NBC also closed 23 loss-making branches. More recently, under Tanzania’s 1995 restructuring plan, NBC cut its work force by 3,000 employees and closed 48 branches and regional offices.

Box 8.2Highlights of Bank Restructuring

Bolivia

Under the FSAC (1988–91), new banking regulations and a supervision agency (the Superintendence of Banks) were established. A draft Banking Law was submitted to Congress in 1989 but was passed only in April 1993, The new regulations and the improved banking supervision helped to stem the deterioration of the financial position of private banks, although their situation remained fragile. Reform of the three state-owned banks (BANEST, BAB, and BAMIN) was disappointing, because the government presented pro forma restructuring plans to meet second-tranche conditions under the FSAC but never effectively implemented them, hi fact, the government decided to reevaluate the role of the public sector in the financial sector before commencing any restructuring.

The SAC (1991–95) focused, inter alia, on the reform or liquidation of the three state-owned banks targeted under the FSAC and the Mineral Exploration Fund (FONEM), as well as continued improvements in prudential regulations and enforcement. As a prior action for the SAC, BAB was closed in July 1991; BAMIN and FONEM were closed in 1992. The Banking Law was approved in April 1993, and a new Central Bank Law was approved in October 1995. A special fund (FONDES1F) was set up in September 1995, to help restructure and capitalize viable banks. Two private banks facing liquidity problems were closed in 1994 by the Superintendent of Banks and the central bank, and three others were restructured in 1996 with support from FONDESIF. Recent problems in the banking system reflected excessive lending and inadequate prudential controls during 1991–94. Supervision and prudential regulation were strengthened further in 1995.

Ghana

For the most part, the objectives of FINSAC-I (1988—91) were achieved. Six of the seven distressed banks were restored to solvency and underwent substantial managerial and organizational changes resulting in a return to profitability and improvement in most financial operating ratios. Banking regulation and supervision were also strengthened, and the technical capacity of the Bank of Ghana and skills and training of the banking and accounting sectors were upgraded.

Under FINSAC-II (1991–94), a further strengthening of the financial position of the Bank of Ghana, development of indirect monetary policy instruments, and improved institutional capacity (a revised Bank of Ghana Act, reorganization, and training) of the Bank of Ghana was achieved. Efforts to improve competition in the banking system fell short, in part reflecting the slow pace of privatization. Privatization of the state-owned banks was delayed significantly; thus far only the largest of these banks has been privatized (in 1997). The high intermediation costs of the largest commercial bank (privatized in 1997), relative to the other two major commercial banks, have kept interest rate spreads high throughout the banking sector. Following incidents of fraudulent lending in 1995—96, bank supervision is being strengthened further.

Lao P.D.R.

In 1988, the monobanking system was replaced by a central bank and eight state-owned banks. Under the Asian Development Bank financial sector credit (1994), the state-owned banks were recapitalized in 1994 to relieve them from bad loans inherited from the old State Bank. Reforms of the regulatory and prudential framework were modest, and the operational restructuring of the state banks is yet to be initiated. Barriers to entry had been reduced previously, including the opening of foreign bank branches in 1992. Seven foreign and two local-foreign joint ventures (in 1989 and 1993) were licensed, but their activities were restricted to the capital city. There are indications that new bad loans were accumulating again in the state-owned banks toward the end of 1996, reflecting the lack of operational restructuring and shortcomings in bank supervision.

A second Asian Development Bank financial sector reform credit was approved in 1996, focusing on strengthening the management, accounting, and credit practices of the state banks. Starring in 1996, banks are required to have annual external audits. The banking supervision capacity of the Bank of Lao and the framework for prudential control are being enhanced, with technical assistance from the IMF and the International Development Bank. None of the eight state-owned banks is to be privatized as part of this operation, but the government has left open the option of merging some of the banks.

Senegal

In late 1989, reforms of the West African and Monetary Union (WAMU) addressed various obstacles to sound banking and provided a strong basis for restructuring the banking sector. Reforms included the abolition of the preferential discount rate, simplification of the interest rate structure, granting greater flexibility to banks in determining their rates and fees, abolition of the policy of sectoral credit allocation, and strengthening banking supervision.

The FSECAL (1989–92) built on these achievements to carry out the restructuring of 10 distressed banks (out of 15 banks). Two private banks were successfully restructured; two private banks, whose shareholders failed to recapitalize, were closed; five public sector banks were closed; and one—the agriculture bank (CNCAS)—is still operating but has serious problems. The government has precluded closure on the grounds that it would deprive the rural population of essential banking services. The government also reduced its shareholding to 25 percent or less in most banks. Banking supervision was strengthened with the creation of the Banking Commission in October 1990. The five main commercial banks—all with significant foreign participation.—are solvent, liquid, and profitable. Recovery of bad loans was disappointing, reflecting an inadequate legal and judiciary framework.

Tanzania

The main objectives of the reforms under the FSAC (1991–93) were to strengthen banking regulations and supervision, create a competitive environment and introduce private participation in the banking sector, and restructure distressed public banks (NBC, CRDB, PBZ, TIB, and THB). Banking regulations and supervision were improved with the approval by parliament in April 1991 of a New Banking and Financial Institutions Act and an amendment to the Bank of Tanzania Act. The Bank of Tanzania has begun to develop most of the basic skills needed for examinations and has conducted a series of on-site bank inspections. However, professional training and skills among the existing staff are still weak; off-site monitoring is yet to be fully developed; and the Bank of Tanzania has not yet exercised its enforcement powers.

Progress in creating a competitive environment was mixed. Five new foreign private banks were licensed, but competition remained limited because they confined their activities mostly to foreign exchange transactions and to the capital city. NBC continues to dominate the banking system, despite some downsizing in recent years. Restructuring of the state-owned banks was minimal, reflecting weak government commitment, continued lending to inefficient public enterprises, and inadequate banking supervision. Despite substantial recapitalization, staff retrenchment, and branch closure, the financial position of NBC continued to deteriorate through end–1995, and the viability of the bank is not assured, CRDB was privatized but continued to face serious difficulties; PBZ remained undercapitalized and burdened with nonperforming loans; TIB has ceased all lending pending its restructuring; and THB was liquidated in mid–1995.

A Financial Institutions Development credit (1995) aims to downsize and restructure NBC more aggressively, restructure PBZ, and strengthen banking supervision.

The scope of improvements in banks’ management varied across countries. In Ghana and Senegal, the authorities moved quickly to replace the top management and boards of directors of distressed banks. In Ghana, twinning arrangements were used in several banks to bring in foreign managers with the requisite skills and experience. In Senegal, some of the banks were affiliates of major French banks, from which new expatriate managers were drawn. As well as strengthening banking skills, the introduction of external management (and new owners) created an arms-length distance from possible government interference. By contrast, initially there were no significant changes in the management of Tanzanian and Bolivian banks. Subsequently, in response to strong pressures from donors, expatriate management teams were installed in Tanzania. In the interim, significant delays were incurred in the restructuring program, as the former managers of distressed banks were not genuinely committed to reforming the banks. In Bolivia, the introduction of new management was part of the 1995—96 restructuring of private banks.

The strengthening of accounting systems and the training of the staff in problem banks featured importantly in Bolivia, Ghana, and Tanzania, involving significant technical assistance from the IMF, the World Bank, and bilateral donors. In Ghana, considerable emphasis was placed on building a long-lasting improvement in domestic capacity in this area: a training college for bankers was established, and accounting courses were offered through the Ghanaian Institute of Chartered Accountants.

Operating Environment: Regulation, Supervision, and Market Discipline

A substantial strengthening of banking regulations was undertaken in all five countries. Minimum capital requirements were raised in 1987 and then again in 1995 to 10 percent of assets in Bolivia, and to the Basle standard of 8 percent of risk-weighted assets in Lao P.D.R. and Senegal. Stricter guidelines were established for loan classification and provisioning, as well as for the treatment of accrued unpaid interest (Bolivia and Ghana). Risk exposure was also tightened with the imposition of limits on total credit to any single borrower (Bolivia and Ghana) or specific sector (Ghana). Uniform accounting systems were introduced in Ghana, Lao P.D.R., and Senegal. Also, in Bolivia and Ghana a comprehensive (multiyear) effort was made to address weaknesses in implementation capacity: training was provided to bank accountants, and the existing training institute was strengthened.

The impact of these changes, as well as existing rules, was strongly influenced by the extent to which they were enforced. All of the countries undertook changes in their systems of banking supervision, but only in Ghana and Senegal does the evidence suggest a marked improvement in the effectiveness of supervision. Substantial efforts, including considerable technical assistance, were made to improve existing supervision systems in the five countries (but only since 1996 in Lao P.D.R.), and Bolivia and Senegal moved forcefully to establish a new and strong banking supervision agency. However, in Bolivia, Lao P.D.R., and Tanzania, enforcement was weak because of a number of factors. Despite improvement, the level of supervisory skills and staffing in the central bank was inadequate, and penalties for non-compliance were weak or were not being applied. In these three countries, the need for effective supervision applied particularly to state-owned banks (eventually shut down in Bolivia), but the record of enforcement was poor. In Bolivia, the authorities strengthened significantly the supervision and regulation of banks in 1995.

The importance of effective enforcement of prudential regulations for successful and sustained restructuring places a premium on providing adequate technical support for strengthening the domestic capacity in this area and on the execution of significant sanctions for non-compliance. To this end, reform programs should focus on the “delivery” of prudential supervision systems.

Reforms of the legal and judiciary system needed for the collection of loans or realization of collaterals were modest at best. This situation was an important factor in the low recovery rates of loans deemed to be recoverable in Senegal and Tanzania. The higher recovery rate (of loans not written off) in Ghana owed in part to the granting of special judicial powers to speed up the process of loan workouts.

Measures were taken in most of the countries to improve the competitive environment of banks, notably through a more liberal policy for the establishment of foreign banks. Lao P.D.R. (before the start of the 1994 reform program) and Tanzania, in particular, introduced fundamental changes in their banking laws that opened up banking activities to the private sector, notably foreign banks.74 Also, Bolivia and Ghana streamlined procedures for the entry of new banks. However, despite the entry of several new foreign banks in Ghana, Lao P.D.R., and Tanzania (see Table 8.10), a handful of state-owned banks retained a dominant position. In Lao P.D.R. and Tanzania, new banks have restricted their activities to the capital area and have focused mostly on foreign exchange transactions. In part, the narrow focus of their activities reflected legal shortcomings concerning the ability to secure adequately loans with collateral. Tangible progress in the divestment of government shareholdings in, or outright privatization of, state-owned banks to foreign and domestic buyers was achieved only in Ghana and Senegal. In Senegal, privatization was undertaken at an early stage of the reform process. In Ghana, the goal (set in 1991) was to reduce government shares in banks’ capital to below 40 percent by end–1993, but the process was not initiated until 1995: by 1997, 42 percent of the shares of the largest bank was held by domestic owners and 30 percent by a foreign strategic investor.

Financial Intermediation

One of the objectives of bank restructuring is to strengthen financial intermediation by the banking system, which should contribute to financial deepening at least over time. However, it is difficult to identify the separate effect of bank restructuring on financial deepening, which is influenced by several other factors: the range of available saving instruments, access to bank credit (particularly in rural areas), the adequacy of property rights and the judicial system and the stability of the macroeconomic environment.

Bearing these caveats in mind, selected indicators of financial intermediation in the pre- and postreform periods point to a mixed performance, with improvements in the rate of monetization and financial saving in Bolivia, Lao P.D.R., and Tanzania (Table 8.11). The continued low rates of financial saving (except in Bolivia) may in part reflect a persistent lack of confidence in the banking system as a result of lingering bank distress (for example, in Tanzania), or the adverse impact of the closure of bank or bank branches on the mobilization of saving, particularly in rural areas (Ghana, Senegal, and Tanzania). In Bolivia, financial intermediation was also positively affected by the presence of a more stable macroeconomic environment since the mid–1980s, the liberalization of interest rates, and the substantial margin of interest rates on foreign currency deposits and loans over international rates. Nevertheless, improvements in the viability of the private commercial banks since 1988 also contributed to a strengthening of confidence in the banking system. One indicator of the efficiency of financial intermediation—the interest rate spread—worsened in most countries, reflecting several factors: the impact of interest rate liberalization (Lao P.D.R. and Tanzania) together with banks’ efforts to recapitalize and restore profits, limited competition in the banking sector (Ghana, Lao P.D.R., and Tanzania), and the cost of holding required reserves (Ghana).

Table 8.11Indicators of Financial Intermediation
Countryt - 3t - 2t - 1tt + 1t + 2t + 3Latest Year
Bolivia (t = 1991)
Monetization rate116.218.523.229.134.642.646.541.7
Financial saving rate210.213.417.721.725.932.131.526.6
Interest rate spread38.68.37.87.76.97.76.66.5
Ghana (t = 1988)
Monetization rate116.416.917.718.016.913.913.217.6
Financial saving rate23.43.74.14.13.42.93.64.4
Interest rate spread35.43.07.99.19.08.79.414.6
Lao P.D.R. (t = 1994)
Monetization rate17.57.510.212.512.813.313.3
Financial saving rate23.24.97.79.49.09.0
Interest rate spread35.811.012.012.012.011.711.7
Senegal (t = 1989)
Monetization rate125.524.225.124.623.924.323.920.5
Financial saving rate28.28.68.09.49.57.56.86.5
Interest rate spread37.48.28.38.79.09.09.0
Tanzania (t = 1991)
Monetization rate122.921.420.824.129.032.434.231.2
Financial saving rate24.56.97.58.88.68.67.8
Interest rate spread33.03.03.55.07.015.013.019.0
Source: IMF staff estimates.

Ratio of broad money to GDP in percent; includes domestic and foreign currency deposits.

Ratio of saving and time deposits to GDP in percent.

Defined as the difference between domestic currency lending and deposit (saving) rates, except in Bolivia where the rates are those on dollar-denominated loans and deposits.

Source: IMF staff estimates.

Ratio of broad money to GDP in percent; includes domestic and foreign currency deposits.

Ratio of saving and time deposits to GDP in percent.

Defined as the difference between domestic currency lending and deposit (saving) rates, except in Bolivia where the rates are those on dollar-denominated loans and deposits.

Reform Priorities and Monitoring: The Use of Conditionality

The importance of bank reform, and of specific restructuring measures, was signaled in ESAF-supported programs by, among other things, the use of PAs (prior actions—in 5 instances), SBs (structural benchmarks—in 8 instances), and structural performance criteria (SPCs—in 24 instances) in all five countries except Senegal (see Appendix 8.1, Table 8.15). Such conditionality was attached to measures that were part of the broad strategy contained in World Bank-supported reform programs and were not restricted to areas of direct IMF involvement. Hence, typically the measures were drawn from the Bank-supported reform program or were selected in collaboration with the Bank; IMF staff relied on the Bank for monitoring their implementation. In many cases, PAs, SBs, and SPCs were attached to measures to fix the timing of actions that otherwise “floated” as prior conditions for the release of Bank loan tranches; they also added to the urgency of implementing measures after previous delays. For the most part, SBs, SPCs, and PAs were attached to institutional measures—such as a new Central Bank Law (Bolivia) or the implementation of a new accounting system for commercial banks (Lao P.D.R.)—and the completion of plans or approval of measures that were prerequisites for sound banking practices, such as the completion of external audits and approval of bank restructuring plans (Ghana, Lao P.D.R., and Tanzania). By comparison, the application of higher conditionality on the implementation of restructuring measures, such as the liquidation (Bolivia) and recapitalization of banks (Ghana, Lao P.D.R., and Tanzania) or strictures on lending (Bolivia and Tanzania), was rare.

The record in meeting commitments covered by PAs, SBs, and SPCs was generally good. In Ghana, there were long delays in the divestment of state-owned banks, but these, in part, owed to a timetable that subsequently was assessed to have been unrealistic (World Bank, 1995d); also they occurred during a more general hiatus in adjustment efforts associated with the run-up to elections in 1992. Delays were also encountered in Bolivia, with respect to the revisions to the Central Bank Law and the liquidation of distressed state-owned banks. Slippages in the implementation of reforms (except in Bolivia, these concerned benchmarks rather than SPCs) did not, however, trigger any delays in completing a program review. There does not appear to be a clear link between meeting PAs, SBs, and SPCs and the progress in bank restructuring. Although not conclusive, this would tend to suggest that PAs, SBs, and SPCs did not give sufficient emphasis to monitoring progress in terms of operational achievements toward sound banking practices.

Lessons for Program Design

Systemic bank restructuring is a complex, multiyear process, and in some cases the initial timetable of reform proved to be unrealistic. The case studies point to several reasons for the faltering progress in restructuring and reforming banking systems.

The most important impediment was the authorities’ reluctance to relinquish an interventionist role (especially through state-owned banks) in the allocation of credit, as was evident in Bolivia and Tanzania. Concerns about the scale of redundancies and cutbacks in rural bank networks, when large public banks were to be reformed, also played a role.

Two elements of the reform strategy appear to have helped to minimize such problems in Senegal and Ghana. First, the implementation of reforms was preceded by a comprehensive joint review by the authorities and the World Bank of the problems faced and appropriate strategies to address them. Thus, from the outset domestic ownership of the reform program was strong. Second, the implementation of the reform program was closely linked to tranche loan releases under the Bank-supported programs: in Senegal, many key actions were taken prior to loan approval, in part because of the extensive technical groundwork that had been undertaken. In Ghana, FINSAC-I contained three, rather than the more typical two, loan tranche releases, which appear to have kept the policy dialogue focused on the reform agenda: following an eight-month delay in the release of the second tranche, the third tranche was released three months earlier than originally expected.

From a systemic macroeconomic perspective, there are a number of steps that the IMF could take that would complement the efforts of the World Bank and could improve performance under ESAF-supported programs.

  • The fiscal costs of banking system reform are massive. Thus, to ensure that costs do not hold up reforms, it is important to identify reform strategies that minimize costs as well as the means by which those costs will be met. IMF staff should engage more actively in discussions with the authorities and the World Bank (and other involved multilateral development banks) from an early stage, to ensure that all potential fiscal costs are assessed. This assessment should be as comprehensive as possible—for example, including the quasi-fiscal operations of loan recovery agencies as well as contingent liabilities associated with government guarantees. Medium-term fiscal scenarios could be used to examine the impact of uncertainties about the costs of reform on the budget. A sharper focus on the link between the pace of the reform strategy and its costs might suggest more ambitious, and typically cheaper, approaches. Furthermore, it could encourage greater emphasis on design features—such as exit (bank closure) policies and the introduction of new management and ownership—that would both promote substantive reform and save budgetary resources. Greater consideration could be given to a broader sharing of costs with shareholders, creditors, and depositors. This would also reduce the moral hazard consequences of reform.
  • A comprehensive, forward-looking reform plan from the outset is essential to achieve sustained improvements in banking system health. Piecemeal implementation weakened the impact of reforms: for example, proceeding with financial restructuring (to tackle the “stock” problem) of banks without reforming banks’ operations or enforcing an adequate regulatory framework and strong supervision system tended to result in recurrences of bank distress. The IMF can contribute to a more comprehensive implementation of reform measures by giving more emphasis in its conditionality to operational achievements that would represent effective progress in terms of results, in addition to establishing “rules” or “laws.” Hence, programs should monitor the adequacy of policies affecting the banking system and the ability of the authorities to enforce best practices.75 Key aspects would include licensing and exit policies, lender-of-last-resort facilities, rules for loan classification and provisioning, capital standards, and the legal authority and capacity of supervisors to implement prudential regulations and impose penalties. Performance in these areas, as well as in relation to the broader longer-term goals of banking reform, should be assessed in the context of program reviews.
  • Lack of progress in public enterprise reform undermined bank restructuring efforts. Programs should focus on the financial relationships between banks and public enterprises and hence expand the coordination of reforms in these two sectors. To this end, the IMF should take a more proactive role, in cooperation with the World Bank, in improving the transparency of the public enterprises’ financial relations with the banking system. When specific public enterprises appear to pose a threat to bank restructuring, these should be targeted for reform as a priority; programs could also adopt stopgap measures—such as specific credit limits on individual public enterprises—to protect banks’ portfolios.
  • Programs should encourage the establishment of a legal and judicial system that supports the exercise of property rights. Insufficient support for property rights contributed to low rates of loan recovery that added to the costs of bank restructuring, weakened incentives for prudent borrowing practices, and dampened the expansion of lending activities by banks, especially new entrants.
  • Banking expertise is scarce, and institutional capacity for enforcing an effective prudential framework is in some instances rudimentary. Building up the requisite skills will require a long-term technical assistance effort, and the continuation of extensive technical assistance from the IMF and other institutions.
Appendix 8.1. Summary History of Public Enterprise Reform Programs

Public enterprise reforms in Bolivia, Ghana, Mongolia, Senegal, and Zimbabwe are described below. For additional details, see Tables 8.128.15.

Table 8.12Monitoring Structural Public Enterprise Reforms1
CountryPerformance Criteria (PC)/Structural Benchmarks (SB)Observance: On Time (OT)/After Short Delay (SD)/After Long Delay (LD)/No (N)/Yes (Y)Prior Actions (PA) for New Arrangements (A) and Reviews (R)
Bolivia
ESAF I.1. (1988)
Establish SAFCO accountsSBN
Reorganize YPFBSBSD
Ceiling on indebtedness of nonfinancial public sectorPCOT
Ceiling on CBB credit to nonfinancial public sectorPCOT
ESAF I.2. (1989)
Approval of revised codes for mining, hydrocarbons, and foreign direct investment2SBLDPA (R)
Approval of SAFCO lawSBLDPA (R)
Ceiling on indebtedness of nonfinancial public sectorPCOT
Ceiling on CBB credit to nonfinancial public sectorPCN
ESAF I.3. (1991)
Issue decree with principles, procedures, and public enterprises for divestmentSBOT
Send Congress legislation for public enterprise salesPCOT
Ceiling on combined deficit of public sectorPCOT
Ceiling on indebtedness of nonfinancial public sectorPCOT
Ceiling on CBB credit to nonfinancial public sectorPCOT
ESAF I.4. (1992)
Offer 31 public enterprises for salePCOT
For 20 public enterprises award sale or closeSBOT
Ceiling on combined deficit of public sectorPCOT
Ceiling on CBB credit to nonfinancial public sectorPCN
ESAF II.1. (1994)
Capitalize ENDEPCSDPA (R)
Capitalize ENTELSBSD
Sell or close 10 public enterprisesSBSDPA (R)
Sell or close another 30 RDC public enterprises3SBSD
Ceiling on combined deficit of public sectorPCOT
Ceiling on domestic financing of public sectorPCOT
ESAF II.2. (1996)
Sell or close 7 public enterprisesSBY
Capitalize YPFBPCSDPA (R)
Capitalize EMVSBPending
Sell or close remaining RDC public enterprisesSBPending
Ceiling on combined deficit of public sectorPCY
Ceiling on domestic financing of public sectorPCY
Ghana
ESAF I.1.(1988)
Divest 6 joint ventures4SBN
Adjust cocoa priceSBOT
Ceiling on BoG credit for cocoaSBOT
Ceiling on bank credit for cocoaPCOT
ESAF I.2. (1989)
Divest 5 joint ventures4PA (R)
Divest 6 joint ventures5SBSD
Divest another 5 joint ventures5SBOT
Adjust cocoa priceSBOT
Ceiling on bank credit for cocoaPCOT
Sign performance contractsOTPA (A)
ESAF I.3. (1991)
Divest 3 public enterprisesPA (A)
Divest 12 public enterprisesSBSD
Divest another 13 public enterprisesSBOT
Adjust cocoa priceSBOT
Introduce flexible petroleum pricing6SBN
Ceiling on bank credit for cocoaPCOT
ESAF II.1. (1995)
Divest 10 public enterprisesSBOT
Divest another 14 public enterprisesSBOT
Adjust cocoa priceSBOT
Remove cocoa export monopolySBN
Submit review of GNPC operationsSBSD
Remove oil import restrictionSBLDPA (R)
Allow oil prices to fully reflect international pricesSBLDPA (R)
Ceiling on BoG credit for cocoaPCOT
Ceiling on BoG credit for GNPCPCOT
Mongolia
ESAF I.1. (1993)
Enact securities law for secondary market tradingSBSD
Complete voucher privatizationSBOT
New procedures for corporate governance of public enterprisesSBOT
ESAF I.2. (1994)
Terminate BoM lending to public enterprisesPA (A)
Timetable for cash salesPA (A)
Schedule of repayments of loans to budget by public enterprisesPA (A)
Sell one quarter of public enterprises earmarked for cash privatizationSBOT
Establish regulatory commission to review tariffsSBOT
Senegal
ESAF I.1. (1988)
Plan of action for groundnut sectorPCOT
ESAF I.2. (1989)
Renegotiate protocol governing financial relations between government and SONACOSPCOT
Complete plan of action for public enterprise reformPCOT
ESAF I.3. (1991)
Submit draft labor law to National AssemblyPA (R)
ESAF II.1. (1994)
Amend draft labor lawPA (A)
Adopt law aimed at letting public enterprise freely adjust their labor forceSBOT
Offer SONACOS for saleSBN
Deregulate rice pricesSBOT
ESAF II.2. (1995)
Offer SONACOS for salePA (A)
Withdraw CPSP from rice via auditing and liquidationPCOT
Liberalize imports and prices of wheat flourSBOT
Preparation and periodic communication of data on public enterprisesSBSD
Zimbabwe
ESAF I.1. (1992)
Ceiling on operating losses of 9 key public enterprisesPCN
ESAF I.2. (1994)
Draft legislation allowing change of status of several large public enterprises to Companies Act and restructure balance sheet of ZISCOSBN
Establish pricing policy to limit temporary trading loss of GMB to amount consistent with budgetPCN
Ceiling on operating losses of 9 key public enterprisesPCN
Source: IMF staff estimates.

Year of approval of arrangement is shown in parentheses. Short delays (SDs) are defined as less than six months, and long delays (LDs) are defined as 6 months or more.

This measure was to have been implemented under ESAF I.1 but was not. Hence, it was made a SB under ESAF I.2.

Five public enterprises were sold on schedule; the remaining eight public enterprises were to be sold or liquidated under ESAF II.2 (but not subject to a SB).

This SB was not implemented under ESAF I.1. and hence was made a PA under ESAF I.2.

None was sold on time: the SB was then reset for a later data during the midterm review. However, its scope was reduced to cover only five out of the six original enterprises.

Only partially implemented.

Source: IMF staff estimates.

Year of approval of arrangement is shown in parentheses. Short delays (SDs) are defined as less than six months, and long delays (LDs) are defined as 6 months or more.

This measure was to have been implemented under ESAF I.1 but was not. Hence, it was made a SB under ESAF I.2.

Five public enterprises were sold on schedule; the remaining eight public enterprises were to be sold or liquidated under ESAF II.2 (but not subject to a SB).

This SB was not implemented under ESAF I.1. and hence was made a PA under ESAF I.2.

None was sold on time: the SB was then reset for a later data during the midterm review. However, its scope was reduced to cover only five out of the six original enterprises.

Only partially implemented.

Table 8.13Bank Lending Operations Supporting Cross-Sectoral Public Enterprise Reform, 1985–96
CountryAmountApproved (month/year)Completed (month/year)UndisbursedLong DelaysPublic Enterprise Content
Bolivia
Structural Adjustment Credit$40m9/9112/95YesRegulatory reform, including new investment, mining and hydrocarbon laws.

Restructuring under performance contracts.

Elimination of transfers to public enterprises and arrears.

Divesting small public enterprises and privatizing national airline (LAB).
Capitalization Program Adjustment Credit$50m6/95Pending$38mNoRegulatory reform in telecommunications, electricity/mining, and hydrocarbon sectors.

Capitalization of ENTEL, ENDE, EMV, YPFB.

Financial sector reform, including pension reform.
Senegal
Structural Adjustment Loans II$64m2/866/87YesNoRestructuring key public enterprises under performance contracts.
III$85m5/876/89YesNoReducing public enterprise subsidies to specific targets and eliminating arrears.
IV$80m2/90NoFinal tranche canceledYesDivesting numerous small public enterprises; privatize SONACOS.

Reform labor laws.
Zimbabwe
Structural Adjustment Loans I$125m1/9212/93YesNoAdjusting and liberalizing pricing and marketing of key goods provided by public enterprices.
II$130m6/936/97Second tranche not disbursedYesRestructuring operations of all key public enterprises under performance contracts.

Reduce operating losses of nine public enterprises undergoing restructuring to specified levels monitored by IMF.
Ghana
Private Sector Adjustment Credit 1$70m5/95Pending$23mNoPrivatization of public enterprises accounting for 40 percent of public enterprise employment; includes privatization of four large, strategic public enterprises (Ghana Telecom; State Insurance Corporation; State Housing Corporation; Mim Timber).

Build ownership and consensus for measures to liberalize cocoa sector exports.
Structural Adjustment Credit 1$81m4/876/90Clear cross debts and arrears.
II$120m4/893/91Privatize or liquidate selected small public enterprises.

Conclude performance contracts; adjust cocoa prices.
Mongolia
Banking and Enterprise$10mPendingPendingRestructuring and corporatizing two public enterprises.
Sector Adjustment CreditElimination of budgetary subsidies to all public enterprises except for urban transport.

No new budgetary loans to public enterprises, except for the two enterprises undergoing restructuring.

Reduce public enterprises’ arrears on loan repayments.
Source: World Bank documents and http://conan.worldbank.org. Data as of February 28, 1997. TA = technical assistance.
Source: World Bank documents and http://conan.worldbank.org. Data as of February 28, 1997. TA = technical assistance.
Table 8.14Bank Sectoral and Technical Assistance Operations Affecting Public Enterprises, 1985–96
CountryAmountApproved (month/year)Closing Date (month/year)UndisbursedRemarks
Bolivia
Rural water and sanitation$20m1/9612/00$17m
Power sector reform TA project$5m11/9512/98$4mCapitalization of ENDE
Hydrocarbons sector reform and capitalization TA$llm6/9512/98$4mCapitalization of YPFB
Regulatory reform and capitalization TA$15m6/946/98$3mCapitalization of ENTEL and LAB
Public financial management II$llm6/9112/97$4mPutting in place SAFCO System
Water and sewage$35m12/906/97$5m
Mining sector rehabilitation operations$35m5/896/97$6mRehabilitation of COMIBOL
Economic management strengthening operations$10m12/886/95
Power sector rehabilitation$8m6/8712/94
VG gas enclave$15m6/8612/90
Ghana
Public enterprise and privatization TA$27m6/9612/01$25mTechnical assistance for privatization under PSAC, development of a comprehensive public enterprise database, and an improved performance contract system
Thermal (P-VII)$176m2/956/01$139m
Private sector development$13m12/946/00$llm
Community water and sanitation$22m4/9412/99$18m
Agricultural sector investment$22m12/9311/99$6m
Urban transport$76m5/9312/98$56m
National electrification$80m3/939/98$57m
Agricultural sector adjustment$80m3/9212/95Reforming COCOBOD; liberalizing marketing of agricultural inputs and outputs
Economic management support$15m3/9112/97$2m
Transport rehabilitation II$96m12/9012/97$9m
VRA/sixth power$20m3/906/98$15m
Fifth power (ECG)$40m8/8912/96$lm
Water sector rehabilitation$25m6/896/97$8m
Telecommunications II$19m7/886/96
Mining sector rehabilitation$40m6/8812/96$3m
Cocoa rehabilitation$40m12/876/96Reforming COCOBOD
Transport rehabilitation 1$60m12/8712/95
Public enterprise TA$llm10/876/96Building capacity of State Enterprise Commission in charge of negotiating performance contracts and divestment
Refinery and distribution rehabilitation$15m6/8712/94
North grid extension$6m2/876/92
Ports$25m3/864/93
Industrial sector adjustment credit$25m3/8612/90
Power$28m9/856/91
Senegal
Agricultural sector adjustment credit$45m6/956/97Privatizing SONACOS
Water sector$100m6/956/01$87mPrivatization of management of water distribution
Private sector adjustment credit$40m2/956/95Reform of regulatory and market environment
Transport sector SECAL$65m6/9112/97$18mRestructuring railways and divesting airlines
TA development management$17m5/886/97$3m
Energy IIRestructuring and divestment of SENELEC
Industry sector$33m12/876/98$13m
Telecommunications 1$22m6/866/93
Energy sector$20m6/8612/92
Water supply II$24m3/856/93
Mongolia
Mongolia coal project$35m5/9612/01$33m
Transport rehabilitation$30m5/946/99$18m
Zimbabwe
Power III$90m1/9412/97$62mRestructuring ZESA
Railways II$39m12/906/98$20mReforming National Railways
Power II$44m1/886/94Reform of ZESA incorporated during project implementation
Source: http://conan.worldbank.org. Data as of February 28, 1997. TA = technical assistance.
Source: http://conan.worldbank.org. Data as of February 28, 1997. TA = technical assistance.
Table 8.15Summary of Banking Reform Conditionally
Prior Actions (PAs), Structural Benchmarks (SBs), and Structural Performance Criteria (SPCs)Status
Bolivia
Issue decree to stop lending by three state-owned banks (BAB, BAMIN, and FONEM) establish plan to liquidate BAB and FONEM, and privatize BANIM; July 1991 (SPC)Observed
Submit new Central Bank Law to Congress; July 1991 (SB)Implemented with delay, April 1992
Submit new Banking System Law to Congress; July 1991 (SB)Observed
Restructure Central Bank’s Financial System Department; July 1991 (SB)Implemented with delay, March 1992
Complete recapitalization of the central bank; December 1992 (SPC)Observed
Congressional approval of Central Bank and Banking System Laws; December 1992 (SB)Observed partially
Liquidate public banks (BAB and BAMIN) and FONEM; December 1992 (SB)1Implemented with delay, 1994
Curtail operations of state-owned BANEST; December 1992 (SB)Observed
Implement Central Bank Law and complete central bank restructuring; March 1995 (SPC)Implemented with delay, October 1995
Establish NAFIBO and end central bank on-lending of development funds; June 1995 (SB)Implemented with delay, October 1995
Start on-lending by NAFIBO and begin transfer of on-lending portfolio from the central bank; May 1996 (SB)Observed partially
Restructure balance sheets of selected banks; June 1996 (SPC)Observed partially
Ghana
Government approval of restructuring plans for three major banks; April 1990 (SB)Observed
Replace nonperforming private sector loans with long-term bonds; March 1991 (SB)Observed
Complete and publish audited accounts of Bank of Ghana; June 1991 (SB)Implemented with delay, July 1991
Cut government equity in one public bank to less than 40 percent; end–1992 (SB)2Not observed
Select consultants to implement the first phase of privatization of public banks; preparation of consultant’s report on potential buyers; September 1994 (SB)3Observed
Offer 60 percent of shares in Social Security Bank for sale; September 1995 (SPC)Delayed
Offer 60 percent of shares in National Investment Bank for sale; December 1995 (SB)Observed
Offer 60 percent of shares in Ghana Commercial Bank for sale; March 1996 (SB)Observed
Lao P.D.R.
Complete recapitalization of commercial banks; December 1993 (SB)Implemented with delay, September 1994
Adopt plan to enhance bank regulation and supervision; April 1996 (SB)Observed
Implement new accounting system for commercial banks; June 1996 (SPC)Observed
Adopt plan to restructure state-owned banks; June 1996 (SB)Observed
Complete review of agricultural lending policies; June 1996 (SB)Implemented with delay, December 1996
Complete external audits of commercial banks; December 1996 (SB)Delayed
Senegal
None
Tanzania
Complete financial and management audits of financial institutions; June 1992 (SB)Observed
Complete government bond issue to banks for loans transferred to LART; (PA)4Observed
Freeze NBC lending to nonperforming borrowers; February 1994 (PA)5Observed
Reconstitute NBC’s Board; March 1994 (PA)5Observed
Reduce NBC staff to 6,500; April 1994 (PA)5Observed
Cabinet approval of legislation to increase NBC’s autonomy and to register NBC under the Companies Act; April 1994 (PA)5Observed
Submit law to Parliament to replace NBC Act; June 1994 (PA)5Implemented with delay, August 1994
Complete audit of the NBC; June 1994 (PA)5Implemented with delay, October 1994
NBC to break even financially and be allowed to resume normal lending; June 1996 (PA)5Not observed
Approve terms of reference for a review of NBC’s restructuring; December 1996 (SPC)Observed
Determine a restructuring plan for PBZ; December 1996 (SB)Not observed
Source: IMF staff estimates.

These banks had been previously shut down in the second half of 1991.

Under IMF-monitored program for 1992.

Under informal monitoring by IMF staff for 1994.

Prior action for second annual ESAF arrangement approved in November 1992.

Structural benchmarks under the IMF staff-monitored program during the first half of 1994, and for June 1996 under the staff-monitored program during the first half of 1996. These SBs were effectively PAs for the approval of an ESAF arrangement.

Source: IMF staff estimates.

These banks had been previously shut down in the second half of 1991.

Under IMF-monitored program for 1992.

Under informal monitoring by IMF staff for 1994.

Prior action for second annual ESAF arrangement approved in November 1992.

Structural benchmarks under the IMF staff-monitored program during the first half of 1994, and for June 1996 under the staff-monitored program during the first half of 1996. These SBs were effectively PAs for the approval of an ESAF arrangement.

Bolivia

In the early 1980s, about 30 nonfinancial public corporations were operational in Bolivia, but two public enterprises in the petroleum (YPFB) and mining (COMIBOL) sectors, whose performances had been worsening steadily, accounted for more than 60 percent of the consolidated overall deficit of public enterprises in 1984.

Under Bolivia’s SAF arrangement (December 1986), the government initiated a medium-term public enterprise reform program that focused on restructuring YPFB and COMIBOL; reorganizing several other major public enterprises; substantially increasing prices and tariffs of goods and services provided by public enterprises to cost-recovery levels; introducing a new regime of public enterprise taxation; and tightening control of public enterprise expenditures. One corporation (CBF) was disbanded, and its enterprises were transferred to the Regional Development Corporation (RDC). The reform program contributed to a strengthening of public enterprises’ financial performance.

Under the first ESAF arrangement (July 1988), the restructuring and streamlining of strategic enterprises—YPFB, COMIBOL, ENDE (electricity), and ENFE (railroad)—was continued. Regulatory changes were initiated to allow joint ventures and operating contracts with the private sector (particularly in the mining and hydrocarbons sectors), and legislative changes were sought to allow privatization. There were substantial delays in implementing these changes, and hence in developing a privatization program.

Under the third annual arrangement (March 1991), the government committed to developing an action plan for divesting the national airline (LAB) as well as 50 small RDC companies. In addition, performance contracts were concluded with the 12 largest public enterprises but were soon abandoned, since their impact was mixed. In the first half of 1992, Congress finally approved legislation allowing the sale of state assets; LAB was offered for sale in April. As part of a fourth annual arrangement (September 1992), 20 of the RDC companies of limited economic significance were sold or liquidated. Restructuring measures and regulatory changes to support the reform of large public enterprises were implemented.

The second three-year ESAF arrangement (December 1994) included the capitalization (private acquisition in return for a commitment to invest) of the major public enterprises in the electricity (ENDE), telecommunications (ENTEL), and railroad (ENFE) sectors—all completed in 1995; in hydrocarbons in 1996 (YPFB); and in mining (EMV) in 1997. The last four companies alone accounted for about 75 percent of the combined operating revenue of public enterprises in 1994. LAB was also capitalized in 1995, and the remaining small RDC companies were expected to be divested by June 1997. Once completed, the government will have pulled out almost completely from business that could be performed by the private sector, except for considerable residual elements of YPFB. Levels of employment in these residual elements are excessive, in part because they absorbed surplus labor shed by YPFB during preparations for its capitalization.

Ghana

It is estimated that the public enterprise sector in Ghana included about 347 companies at end- 1988 and employed some 240,000 people—about 51 percent of formal sector employment or 5 percent of the total labor force. Budgetary transfers to the public enterprise sector averaged some 10 percent of government expenditure during 1979–83. Public- enterprises accounted for 50 percent of nonperforming assets in the banking system. Inappropriate pricing policy, poor management, and a lack of adequate supervision were key to the sector’s poor performance.

Building on preparatory work initiated in 1985 under the Economic Recovery Program, Ghana embarked on a divestiture program for small enterprises under its first three-year ESAF arrangement (November 1988). Also, no new public enterprises were to be created, except under specific conditions. About 62 small public enterprises and a number of cocoa plantations were privatized; the divested companies did not account for a significant share of GDP. Performance contracts were concluded with the 14 largest public enterprises, including the cocoa marketing board (COCOBOD); the operating revenue of these public enterprises was equivalent to about 14 percent of GDP. Restructuring aimed at eliminating government subsidies by the end of the ESAF arrangement. Bank of Ghana financing for COCOBOD was phased out, and total bank financing was strictly controlled. Concomitantly, COCOBOD was restructured, and the cocoa producer price was raised in stages.

Reassessing the results of these reforms, the authorities noted that the lack of wide-ranging reform for the larger public enterprises was holding back investment in critical sectors, such as cocoa. Accordingly, in the context of an IMF staff-monitored program, a milestone decision was taken in 1993/94 when the government divested its shares in seven of the most profitable enterprises, by flotation on the Ghana stock exchange, and sold half its shares in the Ashanti Goldfields Company. The divestment of small public enterprises continued. Further performance contracts were concluded, covering 46 public enterprises by end–1994. The World Bank (1995b) assessed that while the system of performance contracts needed strengthening, it had been successful in reintroducing systematic corporate planning and improving financial viability through better control over administrative and staffing costs. In 1994, problems emerged in Ghana National Petroleum Corporation (GNPC) that were initially accommodated by providing large amounts of central bank credit. As a result, inflation, at 34 percent, considerably exceeded the end–1994 target of 15 percent. The problems of GNPC were subsequently addressed through restructuring measures and cutting off credit from the Bank of Ghana.

The second ESAF arrangement (June 1995) included a list of 114 public enterprises targeted for divestment established under a World Bank PSAC (1995). The list covers public enterprises accounting for a total of 42,000 employees and includes four large, strategic public enterprises. Strategic divestments include the sale of the remaining government stake in the Ashanti Goldfields, Ghana Telecom, State Insurance, State Housing, and Mim Timber. So far, about half of all the transactions have involved foreign investors. A review of the performance of the privatized public enterprises revealed that outright sales or joint ventures had yielded the best results, while the sale of shares was less effective in improving operations. As of end–1995, out of an estimated 347 public enterprises operating at end–1988, 79 had been fully divested and 31 partially divested; 85 had been approved for divestiture; and another 61 divestitures were to be approved in the context of the PSAC.

Mongolia

Mongolia initiated the transformation from a centrally planned to a market economy in 1990; public sector output was about 95 percent of GDP in 1990. As a first step in public enterprise reform, restrictions on private ownership of livestock were rescinded in mid–1990. New regulatory and legal frameworks were passed by mid–1991 in support of a voucher privatization program. Vouchers were distributed to citizens free of charge in view of the low level of domestic financial saving, the lack of a well-developed capital market, and the absence of adequate means of valuing public enterprise assets. For the privatization of small public enterprises, individuals used red vouchers they received or acquired in the secondary market to bid for shares of public enterprises at auctions. For the privatization of large public enterprises, individuals used blue vouchers to bid for blocks of shares on the Mongolian Stock Exchange. Agricultural units were privatized by distributing assets to farmers and herdsmen.

Under a Stand-By Arrangement (October 1991), the government sought to improve the efficiency of public enterprises that were to remain in the public sector by hardening budget constraints and raising public enterprise autonomy in regard to pricing, hiring and firing, and production. Barriers to entry were removed, and prices were deregulated except for public services, utility tariffs, rents on public housing, and selected medicines. Utility prices were to be adjusted regularly.

In the context of an ESAF arrangement (June 1993), voucher privatization was completed by end–1994, covering about 44 percent of state-owned assets. A second phase of privatization was initiated in June 1994, focusing on cash privatization of more than 400 public enterprises, including 20 percent of the equity of the Erdnet copper concern. The 1995 cash privatization target of ½ of 1 percent of GDP was exceeded, owing to the sale of a 49 percent stake in the telecommunications corporation to a foreign investor. A farther important step was the start in August 1995 of the secondary market for shares.

By the end of 1995, only 36 percent of GDP originated from the public sector. In agriculture and trade, almost all activity was in private hands; an estimated 60 percent of construction, 45 percent of industry, and 35 percent of transport were private. However, public enterprises continue to be important, dominating various sectors and enjoying monopoly status in several key industries. Also, the public enterprise sector undermined macroeconomic stabilization efforts because of public enterprises’ continued large-scale access to bank credit and direct or indirect budgetary subsidies; a failure to adjust utility prices fully to cost-recovery levels; and inadequate internal managerial capacity and incentive structures, including diffuse ownership that allowed the government, with only a minority stake, to interfere in company decisions. These problems were compounded by the failure of the vast shift in ownership to engender a commensurate change in managerial practices. As a result, newly privatized enterprises as well as existing public enterprises defaulted on loans, triggering a banking crisis in 1996.

Senegal

In 1987 the public enterprise sector in Senegal comprised 87 public enterprises with at least a majority government shareholding and about 100 public enterprises with a minority holding. The sector accounted for about 29 percent of total investment in the economy, 17 percent of formal sector employment, but only 7 percent of GDP. Public enterprises often had duplicating and overlapping operations, and excessive administrative controls limited the flexibility with which they managed their operations. Public enterprises had incurred financial losses estimated at some 2 percent of GDP annually, covered primarily through budgetary transfers and bank borrowing, which contributed to a weakening of the banking sector’s financial position.

The government’s strategy to reform the public enterprise sector was embodied in the “New Policy for the Parapublic Sector,” adopted in 1985, and aimed to privatize public enterprises that the private sector could run better and to rehabilitate strategic public enterprises that would remain in the public domain. Efforts were reinforced in the context of an ESAF arrangement (November 1988) and by the adoption of a more comprehensive reform program in 1989 that was supported by the World Bank. This program continued to focus on the rehabilitation of the strategic enterprises; it also targeted the privatization and liquidation of 45 public enterprises that accounted for only about 12 percent of the government’s equity in the public sector (World Bank, 1994c). Furthermore, it included the settlement of cross-payment arrears, the reduction in direct and indirect subsidies to public enterprises, and the termination of government guarantees and subsidies on foreign borrowing by public enterprises.

By 1993, 20 public enterprises had been liquidated; 24 privatized, either through the sale of shares or assets, or long-term leases; and a few reorganized. Prices of all but 24 consumer goods had been decontrolled and quantitative import restrictions eliminated; the prices of staple foods (rice, sugar, and flour) and important producer prices (groundnut, lettuce, and rice) remained regulated. The divested public enterprises were of minimal importance, and public enterprises and public agencies remained dominant in strategic sectors. At end–1993, 63 public enterprises were still operational.

In 1994—95, under a Stand-By Arrangement (March 1994) and a new ESAF arrangement (August 1994), two public enterprises were liquidated and one privatized. Consumer prices for 13 commodities and the producer price of rice were liberalized, and prior authorization for imports of 18 products and exports of 6 products was eliminated. In 1996 a divestment program of selected strategic public enterprises was launched with the privatization of SONATEL (national telecommunications) and the management of water production and distribution. Also, several public enterprises were put up for sale, and a plan to privatize the national electricity company (SENELEC) was adopted. The authorities targeted the privatization or liquidation of the strategic groundnut company (SONACOS) and 16 other public enterprises by end–1997.

Zimbabwe

The government of Zimbabwe owns all or part of a large number of companies listed under the Companies Act and designated public enterprises, of which 22 have commercial activities, including at least 7 with some monopoly rights. Public enterprises incurred substantial losses, reaching 4.4 percent of GDP at the end of the 1980s. In principle, public enterprises are compensated from the budget for their audited losses; in practice, however, subsidies have often been less than the losses, with the remainder financed by government-guaranteed borrowing from nonbank financial institutions.

Reducing the budgetary burden of public enterprises was a key objective under the government’s adjustment program adopted in early 1991 and supported by an ESAF arrangement (September 1992). Restructuring plans were to be devised and implemented by selected major public enterprises whose sales and operating revenue were equivalent to about 20 percent of GDP in 1989/90. A timetable was established to reduce direct subsidies and transfers and was introduced as part of a phased action program aimed at increasing the autonomy of public enterprises, raising the flexibility of price setting, and stepping up private sector competition through deregulation. No significant divestments were undertaken.

The financial performance of public enterprises improved: five out of nine key public enterprises recorded operating surpluses in 1994/95, up from two in 1990/91. However, improvements in eight of the public enterprises were offset by an increase in the financial losses of the Grain Marketing Board (GMB), stemming from inappropriate price setting for maize. Following a major maize price adjustment in August 1995, GMB’s finances improved significantly. While budget subsidies have been sharply curtailed since 1991 and direct government lending to public enterprises appears to have declined, continuing public enterprise losses have been financed with government-guaranteed loans. As of March 1994, about one-third of total outstanding loans to public enterprises—about 25 percent of GDP—was government-guaranteed external debt.

Progress was made in raising the autonomy of public enterprises: four out of the nine key public enterprises, including three agricultural marketing boards other than the GMB, were registered under the Companies Act in 1994/95. Under this measure, pricing and marketing were decontrolled, and the companies were subject to private sector competition. In a major financial restructuring of the agricultural boards and GMB, the government took over loans in 1995—equivalent to about 7 percent of GDP—of which more than 80 percent were owed by the GMB, and committed to cease providing loans or loan guarantees. More notably, in 1996 an important restructuring of the GMB separated its commercial operations from the management of the subsidized Strategic Grain Reserve and significantly liberalized the domestic grain market. These changes allowed the GMB to begin functioning on a commercial basis, and in 1995/96 GMB made a profit.

Beginning in 1994/95, a program of government asset sales (through publicly traded shares) was instituted: sales have been erratic, albeit substantial—cumulatively reaching $100 million by mid–1996. In late 1996, the Cabinet issued a strategy paper on the privatization and commercialization policy for 28 targeted public enterprises. Progress on privatization has been slower than expected because of conflicting views on divestiture policy and delays in developing plans to use privatization as a vehicle for promoting indigenous ownership of productive assets. In 1997 the steel company (ZISCO), a major public enterprise, was put on sale.

Appendix 8.2. The Fiscal Program and the Treatment of Privatization Proceeds

The costs of reform and privatization receipts were incorporated in the fiscal projections of the programs, with particular attention paid to insulating deficit targets from uncertainties surrounding the realization of these costs and revenues. In principle, the latter consideration—as well as the loss of future dividend revenues when profit-making public enterprises were privatized—argued against relying on receipts to finance expenditures in order to safeguard the adjustment objectives of the fiscal program (see also Hemming and Miranda, 1991).

In practice, this standard was relaxed to finance pressing general expenditure needs. Some programs allowed all the divestiture proceeds to be used to finance general government expenditure (Ghana, ESAF I; Senegal, ESAF I; and Zimbabwe), and in Mongolia revenues were earmarked for social expenditures or the public investment program. More recent programs stipulated that divestment proceeds could be used to finance general government expenditure up to a specified level, with any excess over this level to be used in line with understandings reached with IMF staff (Ghana, ESAF II; and Senegal, ESAF II).76 Under Ghana’s staff-monitored program (1993—94), receipts from the sales of shares in the Ashanti Goldfields were to be paid into an investment trust fund; in the event, a large portion of the receipts was used for larger-than-programmed general government expenditure. Only in Bolivia did programs include an explicit understanding that receipts could be used to finance privatization costs. The use of revenues in excess of these costs was constrained.77 Since all programs allowed at least a fraction of projected divestment receipts to be used for purposes other than retiring public debt, shortfalls in divestment receipts would have required an offsetting compression of expenditures. However, this never became an issue because such proceeds were conservatively projected.

References

    AlexanderWilliam E. and otherseds.1997Systemic Bank Restructuring and Macroeconomic Policy (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    Basle Committee on Banking Supervision1997Core Principles for Effective Supervision: Consultative Paper (BasleApril).

    BlanchardOlivier and RichardLayard1992“How to Privatize,” in The Transformation of Socialist Economies: Symposium 1991ed. byHorstSiebeck (Tübingen, Germany: Mohr) pp. 2744.

    • Search Google Scholar
    • Export Citation

    BoyckoMaximAndreiShleifer and Robert W.Vishny1994“Voucher Privatization,”Journal of Financial EconomicsVol. 35 (April) pp. 24966.

    • Search Google Scholar
    • Export Citation

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

    • Search Google Scholar
    • Export Citation

    FischerStanley1991“Privatization in East European Transformation,” NBER Working Paper 3703 (Cambridge, Massachusetts: National Bureau of Economic ResearchMay).

    • Search Google Scholar
    • Export Citation

    Folkerts-LandauDavid and Carl-JohanLindgren1998Toward a Framework for Financial Stability (Washington: IMFJanuary).

    GalalAhmed and Mary M.Shirleyeds.1994Does Privatization Deliver? Highlights from a World Bank Conference EDI Development Studies (Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    GamarraE.1990“The Privatization Debate in Bolivia,” in Privatization and Deregulation in Global Perspectiveed. byDennis J.Gayle and JonathanGoodrich (Westport, Connecticut: Quorum) pp. 198212.

    • Search Google Scholar
    • Export Citation

    HemmingRichard and KennethMiranda1991“Privatization,” in Public Expenditure Handbook: A Guide to Public Policy Issues in Developing Countriesed. byKeyoungChu and RichardHemming (Washington: International Monetary Fund) pp. 13945.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund1998External Evaluation of the ESAF: Report by a Group of Independent Experts (Washington).

    LindgrenCarl-JohanGillianGarcia and Matthew I.Saal1996Bank Soundness and Macroeconomic Policy (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    MallonR.D.1994“State-Owned Enterprise Reform Through Performance Contracts: The Bolivian Experiment,”World DevelopmentVol. 22 (June) pp. 92534 (Washington).

    • Search Google Scholar
    • Export Citation

    MathiesonDonald J. and Richard D.Haas1994“Establishing Monetary Control in Financial Systems With Insolvent Institutions,”IMF Paper on Policy Analysis and Assessment 94/1 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    SaderFrank1995“Privatizing Public Enterprises and Foreign Direct Investment in Developing Countries 1988—93” Foreign Investment Advisory Service Occasional Paper 5 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation

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

    • Search Google Scholar
    • Export Citation

    ShengAndrewed.1996Bank Restructuring: Lessons from the 1980s (Washington: World Bank).

    World Bank1992“Bolivia: Restructuring for Growth,”Draft No. 11075-BO (Washington: World Bank).

    World Bank1994a“Bolivia: Structural Reforms, Fiscal Impacts, and Economic Growth,”Report No. 13067-BO (Washington: World Bank).

    World Bank1994b“Senegal: Completion Report SAL IV,”Report No. SECM94–1207 (Washington: World Bank).

    World Bank1994c“Senegal: Private Sector Assessment,”Report No. 11317-SE (Washington: World Bank).

    World Bank1995aBureaucrats in Business (New York: Oxford University Press).

    World Bank1995b“Ghana: Draft Public Sector Management Review” (unpublished; Washington: World Bank).

    World Bank1995c“Bolivia: Capitalization Program Adjustment Credit,”Report No. P–6332-BO (Washington: World Bank).

    World Bank1995d“Ghana—Project Completion Report: Financial Sector Adjustment Credit I,”Report No. 14158 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    World Bank1995e“Performance Audit Report—Senegal: The Fourth Structural Adjustment Credit and the Financial Sector Adjustment Credit,”Report No. 14818 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    World Bank1996a“Mongolia: Public Enterprise Review,”Report No. 15875-MOG (Washington: World Bank).

    World Bank1996b“Review of Public Enterprise Reform and Privatization Operation” (Washington: Private-Sector Development Department, World Bank).

    • Search Google Scholar
    • Export Citation

    World Bank1997“Privatization in Africa” (unpublished; Washington: World Bank).

1Contributions by the authors of this chapter were as follows: Jorg Decressin and Zia Ebrahim-zadeh, public enterprise reform; and Louis Dicks-Mireaux and Ali Ibrahim, systemic bank restructuring. The analysis has benefited from comments by the World Bank staff.
2See Chapter 4 for a survey of policies and achievements with public enterprise and banking system reform in all of the 36 ESAF countries.
3For a survey and discussion of the experience of a broad range of countries that encountered serious systemic banking sector problems in the 1980s and early 1990s, see Caprio and Klingebiel (1996) and Sheng (1996).
4A broader study (World Bank, 1995a) of public enterprise reform found that limited progress was typical of most developing countries.
5The focus of the study of the banking sector is on those reforms directed at establishing sound banking practices; it does not examine the full range of financial sector reforms undertaken in ESAF-supported programs.
6A study of Bolivia’s hydrocarbon sector (World Bank, 1992) found that YPFB, even though it was profitable and in 1990 generated 50 percent of the treasury’s current revenue, was chronically delinquent in paying private companies and compared poorly with other companies in staff productivity and operational efficiency.
7The label “strategic” is given to public enterprises that dominate a sector of major importance to the economy and includes public enterprises producing and marketing key exports; agricultural marketing boards; and public enterprises operating in communications, transportation, marketing, and power supply.
8Under a management contract, management of a public enterprise is contracted out against a success fee, while under a performance contract it is still appointed by the government. Under management contracts, the government typically remains responsible for fixed investments and has majority ownership (World Bank, 1995a).
9In Zimbabwe, subsidized government loans were also to be reduced. The scale of direct subsidies was contained by progressively lowering ceilings on the losses of the nine largest loss-making public enterprises.
10In 18 of the 36 countries covered by the ESAF review, separate data on bank credit to public enterprises were not available.
11In the countries under review, this approach was followed only in Bolivia, Guyana, and Honduras.
12Data on bank credit to public enterprises were available in Zimbabwe, but no targets were set; public enterprises typically did not have major recourse to the banking system, relying instead on government-guaranteed nonbank financing.
13Divestment, as referred to in this study, comprises liquidation, transfer of ownership (for example, through voucher schemes), and the sale of a minority or majority stake in public enterprises’ assets. The latter includes sales against cash or injection of new capital (capitalization). Buyers could include domestic and foreign investors, management, or employees.
14A recent study (Galal and Shirley, 1994) evaluated the benefits of privatizing large companies on the basis of 12 monopolies and oligopolies in Chile, Malaysia, Mexico, and the United Kingdom. In 11 cases the benefits were substantial, averaging about 2.5 percent of GDP.
15In Ghana, restructuring and monitoring covered 13 core public enterprises, estimated to have accounted for 60 percent of total public enterprise employment, 72 percent of sales value, and 67 percent of value added. Their operating revenue was equivalent to about 14 percent of GDP in 1987 (World Bank, 1995a). In Zimbabwe, operating revenue of the nine core public enterprises amounted to 20 percent of GDP in 1989/90. Restructuring operations in Bolivia focused on hydrocarbon and mining public enterprises and key public utilities. In Senegal, performance contracts were introduced in the public utilities, transportation, agricultural production and marketing, and housing sectors.
16Under Bolivia’s capitalization scheme, private investors (in practice, typically foreign) acquire a 50 percent stake in, and management control of, the public enterprise in return for a commitment to undertake capital expenditure. The remaining shareholding is retained by the government and used to finance a minimum pension for all Bolivian citizens who are now at least 21 years old, to be received at age 65.
17Voucher schemes and capitalization appear more applicable to situations in which the government can afford to forgo the immediate proceeds of an outright sale of existing assets, and in the latter approach there are clear needs for large new investments. See Fischer (1991), Blanchard and Layard (1992), and Boycko, Shleifer, and Vishny (1994) for assessments of mass privatization strategies in transition economies.
18More than 100,000 positions (in part filled by “ghost” workers) were eliminated from the payroll in the restructuring of COCOBOD in Ghana. Roughly 40,000 employees were retrenched in Ghana during 1989–93 in sectors other than cocoa production. In Bolivia, employment in the public enterprise sector has fallen sharply from a peak of 32,400 in 1991 to 16,000 in 1995, in part as a result of restructuring and divestment.
19Typically, the interval between tranche releases was greater than one year, or their expected timing did not coincide with that of annual arrangements. Hence, there was a need to pin down the timing of specific reforms (with input from the World Bank) that were important for the design of the IMF-supported program.
20Progress in public enterprise reform was assessed at the time of each review, but only in Senegal (under the 1988 ESAF arrangement) was it the subject of a specific review clause.
21The PC on GNPC was introduced in the three-year ESAF arrangement approved in 1995, following major macroeconomic slippages under an earlier staff-monitored program on account of this enterprise.
22See also World Bank (1994a and 1995c) for Bolivia; World Bank (1995b) for Ghana; and World Bank (1994b) for Senegal. In Ghana, two key public enterprises (COCOBOD and GNPC) did not adhere to the performance contracts. World Bank (1995a) provides a detailed assessment of selected performance contracts in various countries, including Ghana and Senegal. Of the four contracts reviewed for these two countries, none was associated with an improved return on assets.
23In Bolivia, after the national railway company awarded incentive payments prematurely, the Ministry of Planning did not reverse the payments or hold company managers accountable (Mallon, 1994). In Senegal, the government did not pay the telecommunications company for services rendered, and in Ghana it did not support the collection of revenue arrears and timely approval of tariffs for the utility companies—see World Bank (1995a).
24See World Bank (1995a), which reviewed in depth a wide variety of performance, management, and regulatory contracts (the last of these specify the relationship between the government and owners of a private regulated monopoly).
25See also World Bank (1997) for a review of the scope of privatization efforts in Africa that reaches a similar conclusion.
26Data on bank credit to public enterprises are not available for Senegal. However, estimates by the World Bank (1994c) indicate that public enterprises’ share in total credit and term loans has remained stable at about one-third and about one-half, respectively, during 1987–93, despite the elimination of government interference in lending decisions since 1989; data after 1993 are not available. According to the report, several reasons contributed to the continued large share of public enterprises in bank credit: banks’ belief that the government’s stake in public enterprises constituted an implicit loan guarantee; the limited credit analysis capacity of the banks, which, coupled with limited differences in the interest rate spread charged to borrowers of different sizes, made lending large sums to established and familiar borrowers more attractive than making many smaller loans; and public enterprises held large deposits with the banks.
27In Ghana, this involved a single major enterprise, the GNPC.
28In Mongolia, it is estimated that indirect subsidies from net lending at below-market interest rates and nonrepayment of loans from the budget rose from 2.6 percent of GDP in 1993 to 4.0 percent of GDP in 1995 (World Bank, 1996a). Figures for Ghana suggest that through mid–1994, 12 core public enterprises carried an outstanding liability for unpaid taxes of about 0.8 percent of 1993 GDP (World Bank, 1995b), Senegal began to end direct operating subsidies in 1985 and stopped guaranteeing public enterprise borrowing in 1990, but public enterprises still have access to indirect subsidies such as cheap overdrafts as well as arrears, and cross-debts with the government periodically build up and are written off (World Bank, 1995a). The curtailment of direct public enterprise subsidies in transition economies, including Mongolia, often led to the accumulation of tax and enterprise payments arrears, the latter particularly on payments for energy imports.
29Note that Bolivia capitalized rather than sold its major public enterprises.
30See Appendix 8.2 for a discussion of the treatment of these receipts under programs.
31In 1994 alone, the proceeds from a sale of a 30 percent stake amounted to 379 million cedis, or 7.7 percent of 1994 GDP. Privatization receipts recorded in the budget amounted to 5.3 percent and 1.5 percent of GDP, respectively, in 1994 and 1995. In Mongolia, receipts amounted to 0.8 percent of GDP in 1995—prior to that, cash sales were very limited. In Zimbabwe significant sales took place for the first time in 1994/95, with receipts reaching 1 percent of GDP.
32In Ghana, the import and domestic marketing of petroleum is still being liberalized, and the introduction of new regulatory frameworks for the telecommunications, electricity, and water sectors started in 1995; also, the protected position of Ghana Airways has been dismantled, and, as a result, it no longer operates on domestic routes. In Zimbabwe, Air Zimbabwe was exposed to private competition by its conversion from a statutory body to a commercial company under the Companies Act in 1995.
33In addition, revisions to the mining code and the liberalization of imports and marketing of petroleum are ongoing.
34In Ghana, a fully liberalized investment code was introduced only in 1994. In Mongolia, reforms of foreign direct investment laws were initiated in the early 1990s, and foreign firms were soon allowed to invest in practically any sector. However, further reforms were needed in 1996 to make clear that majority ownership of certain strategic public enterprises would be permitted. In Senegal, it was not clear that the one-stop investment approval window was effective (World Bank, 1994c). In Zimbabwe, the limit on foreign ownership in domestic companies has been raised only gradually, and as of September 1996 it stood at 40 percent.
35The capitalization of the major public enterprises in Bolivia has attracted foreign investor commitments equivalent to 3 percent of GDP a year for the period 1996–2002, or a total of $1.7 billion.
36To the extent that there are considerable sunk costs involved in information gathering, attracting foreign interest in divestment can have positive spillover effects. Reportedly this has been the case in Ghana, where flotation of the government’s shareholding in the Ashanti Goldfields Company on the Ghana stock exchange attracted foreign interest to other companies, although the spillover effect has been less than originally hoped for. Sader (1995) found that privatization in a wide group of developing countries has generally attracted foreign direct investment above and beyond the amounts involved directly, with spillover effects lasting over time.
37The capitalization of ENDE in 1995 and of YPFB in 1996.
38In 1991, only 5 of the 17 public enterprises monitored by the State Enterprise Commission in Ghana had current audited accounts. The most recent audited accounts for the others were one to three years old (World Bank, 1995a).
39See Gamarra (1990) for details on the privatization debate in Bolivia.
40The costs of these schemes were financed from a variety of sources: off-budget privatization receipts in Ghana and Mongolia, directly by donors in Zimbabwe but outside the budget, or by external assistance channeled through the budget in Ghana and Senegal. In Bolivia, where the scope of divestment and restructuring was the largest among the five countries, severance payments for retrenched public enterprise employees were substantial, reaching almost 1 percent of GDP a year on average during 1993—94; these were financed from public enterprises’ general revenues and, for employees retrenched from smaller enterprises, from privatization receipts.
41Government-guaranteed external borrowing is monitored through existing external debt ceilings.
42For a discussion of these issues see Lindgren, Garcia, and Saal (1996) and Mathieson and Haas (1994).
43The exception was Lao P.D.R., where the difficulties of the banking system became apparent only in the early 1990s, following the breakup of the monobank system.
44Loans were generally poorly classified and provisioned, reflecting deficient prudential regulations and their weak enforcement. Thus, until banks had been adequately audited, the magnitude of bank distress and costs of bank restructuring were uncertain.
45In Bolivia, one of the development banks lent to the agricultural sector on preferential terms; in Ghana, Senegal, and Tanzania a large share of agricultural loans were to state marketing boards; and in Lao P.D.R., banks had to lend at least 10 percent of their deposit resources to the agricultural sector.
46The two largest development banks were the agriculture bank (BAB), which lent at preferential rates to private agricultural cooperatives, and BANEST, which was mandated to provide banking services in those parts of the country poorly served by private commercial banks and whose loans were often politically motivated.
47Other reforms that were important for promoting efficient financial intermediation, such as the introduction of indirect instruments of monetary policy, were also implemented but are not taken up in this study.
48These studies were conducted in the context of a World Bank Public Financial Management Loan in Bolivia and a third Structural Adjustment Loan in Senegal; in Tanzania, the review was carried out by a Commission of Enquiry reporting to the president, with substantial technical assistance from the IMF and the World Bank; and in Ghana, the authorities and the Bank undertook a broad joint review of the financial sector. In Lao P.D.R., initial reforms during the SAF arrangement (1989–92) that aimed primarily at establishing a two-tier banking system were supported by technical assistance from the Asian Development Bank, the IMF, and the World Bank.
49In 1994, the WAMU became the West African Economic and Monetary Union (WAEMU).
50Bank restructuring is still going on in most of the countries. Following the first FSAC, the Bank supported a second round of financial sector reforms under a SAC in Bolivia (1991), and a second FSAC in Ghana (1991) and Tanzania (1995). The Asian Development Bank approved a second financial sector credit for Lao P.D.R. in 1996.
51In Ghana, the government was responsible for the redemption and servicing of central bank bonds issued with respect to banks’ claims on public enterprises. The domestic currency counterpart of external assistance under FINSAC-I was held in a blocked account at the Bank of Ghana as backing for the restructuring bonds.
52These liabilities included crop credits rediscounted with the BCEAO and private deposits. The deposits were frozen but earned interest, and they were to be unfrozen with loan recoveries.
53The central bank was also subject to significant organizational restructuring in Bolivia, Ghana, and Tanzania.
54In Senegal, supervision became the responsibility of the Supranational Banking Commission of the WAMU, created in 1990. In Bolivia, a Superintendency of Banks was established, as well as an independent resolution mechanism for banking distress—the Financial System Department—within the central bank. This department was supposed to assist problem banks when liquidation was not desirable, or to assist in the liquidation of banks in collaboration with the Superintendency of Banks. However, the department was never able to fulfill its functions effectively, largely because of a lack of clear definition of its mandate and rivalry with the Superintendency of Banks.
55In Lao P.D.R., foreign banks have been allowed to open branches since 1992; two domestic joint-venture banks opened in 1989 and 1993 prior to the 1994 reform program.
56More generally, among all the countries in the ESAF review the World Bank and other multilateral development banks had the lead role in bank restructuring operations, except for Mauritania, where the IMF has taken the lead in proposing and implementing banking system restructuring measures since 1993. Going beyond the sample of countries covered by the ESAF review, the IMF has taken the lead recently in a few ESAF countries (Armenia, 1996, Georgia, 1995, and Mongolia, 1997).
57For example, through adjustments to credit ceilings on account of the removal of nonperforming loans from balance sheets, sterilization operations to offset the liquidity impact of central bank support to problem banks, and advising on the maturity of government bonds issued to replace nonperforming loans, to mitigate their influence on liquidity.
58In Senegal, a large amount of technical assistance to the national authorities and the BCEAO was provided by France.
59The only notable exception was Lao P.D.R., where banks were required to have at least 10 percent of their deposit base used for lending to the agriculture sector. This requirement was abolished in the second half of 1996.
60By its nature, such influence is difficult to document.
61Initially, the World Bank’s banking reform operation in Tanzania was intended to coincide with an industrial sector restructuring adjustment operation that included a comprehensive public enterprise reform program. However, as of end- 1996 the latter has not been initiated. The authorities launched their own public enterprise reform program in 1992, which began to take effect in 1993.
62At the outset of Ghana’s bank restructuring program, a study was undertaken to assess the magnitude and nature of distress in the corporate (both public and private) enterprise sector and to recommend a program to facilitate the restructuring of potentially viable enterprises. However, after initial studies, the project was abandoned.
63At the beginning of the reform, the costs were relatively well known in Lao P.D.R. and, to a lesser extent, in Ghana and Senegal.
64Information on recovery targets and success rates is not available for the earlier restructuring exercises in Bolivia and Lao P.D.R. In Bolivia’s 1994–96 restructuring program, some 40 percent of loans of the banks closed in 1994 had been recovered as of end–1996.
65For a full discussion of current practices in the recording of bank restructuring costs in the fiscal accounts, and a proposal—an “augmented balance”—for a more uniform treatment, see Alexander and others (1997).
66Information on the health of the banking sector is particularly sensitive because of concerns about public confidence, and forward-looking assessments of banks’ financial performance are typically held in confidence by national authorities.
67First nine months of 1993.
68Since 1989, crop credits were no longer automatically eligible for rediscounting at the central bank and were included under the overall credit ceiling. Since 1990, government-guaranteed bank loans to public enterprises are also included under the credit ceiling of the government.
69During 1995, the share of past-due loans in total loans of the four banks rose from 5 percent to 17 percent, and they lost about 29 percent of their deposit base.
70A special fund, FONDESIF (the Fund for the Development of the Financial System and Productive Sector), was created in 1995 and charged with supporting and monitoring the implementation of action plans for individual banks that would merit official support.
71These bonds were issued with an interest rate of 4 percent a year.
72A mixture of securities was issued in 1993, with fixed interest rates set at slightly below the level of market rates at that time.
73The closure of banks in Bolivia and Senegal also contributed to improvements in the performance of the banking sector.
74In Lao P.D.R., foreign banks have been allowed entry since 1992, and two joint-venture banks with state, local, and foreign participation were opened in 1989 and 1993.
76In Ghana, the excess could be used to purchase foreign investments or could be saved.
77In Bolivia and Ghana (under ESAF II), programs included adjusters to financial performance criteria and benchmarks that specified clearly how proceeds in excess of projected amounts had to be used.

    Other Resources Citing This Publication