This paper reviews economic developments in Ethiopia during 1995–99. It provides an update on macroeconomic performance and structural reforms during FY96–FY99 (fiscal year ended July 7), when Ethiopia—as Africa’s second most populous country and one of the world’s poorest nations—continued to make strides in transitioning to a market-based economy and alleviating widespread poverty. The paper also highlights the major challenges in the areas of financial sector liberalization, civil service reform, and privatization.

Abstract

This paper reviews economic developments in Ethiopia during 1995–99. It provides an update on macroeconomic performance and structural reforms during FY96–FY99 (fiscal year ended July 7), when Ethiopia—as Africa’s second most populous country and one of the world’s poorest nations—continued to make strides in transitioning to a market-based economy and alleviating widespread poverty. The paper also highlights the major challenges in the areas of financial sector liberalization, civil service reform, and privatization.

III. Civil Service Reform in Ethiopia34

34. When the Transitional Government of Ethiopia (TGE) replaced the totalitarian Derg regime in 1991, it inherited a civil service that suffered from administrative complexity, excessive centralization, and poorly trained staff. These factors, together with low wages, a compressed salary structure, and promotions based on seniority rather than merit, resulted in poor morale, evasion of responsibility, and corruption. As the TGE focused on reducing the role of the government and introducing market elements in the economy (via macroeconomic liberalization and private sector development), the reform of the civil service became an important part of the government’s program. The overall objective of the reform in Ethiopia was to provide a lean, efficient, and ethical civil service that would also support the development of critical social areas, such as education and health care, and promote private sector growth.

35. Although some ad hoc measures aimed at improving civil service performance were adopted in the early 1990s, a comprehensive and more systematic reform only started in 1994, with the establishment of a task force to identify the main weaknesses and propose solutions. After a 16-month study, the task force published a report from which the Civil Service Reform Program (CSRP) was developed.

36. This study reviews progress in implementation of the reform during the 1990s and flags some of the remaining problems. The study is organized as follows: The first section describes the evolution of the civil service before 1990; the second section assesses the results of the reform during the 1990s; and the final section summarizes the experience with the demobilization and reintegration of ex-combatants during the early 1990s—which may be applicable again once the current border conflict with Eritrea is resolved.

A. Pre-Reform Period (Pre-1990)

37. During the Derg regime (1971-90), the government became the largest formal sector employer through nationalization and the growth of the central government.35 The number of ministries increased from 16 in 1974/75 (July-June) to 20 in 1989/90, and 34 completely new government agencies were established; the number of civil servants jumped by 134 percent, from 101,000 to over 236,000 during this period (Figure 4). Despite this increase, however, the size of the civil service (normalized by the size of Ethiopia’s population) remained lower than in most sub-Saharan countries, raising a broader issue of lack of capacity.36 Furthermore, the majority of new hires were support staff. In addition, the capacity problem was compounded by the background of the professionals hired with university degrees, few of which were trained in economics, law, or other fields relevant for solving the country’s economic and social issues.37

Figure 4.
Figure 4.

Ethiopia: Number of Civil Servants, 1974/75 - 1995/961

Citation: IMF Staff Country Reports 1999, 098; 10.5089/9781451812626.002.A003

Source: Ethiopian authorities; and Fund staff calculations.1Fiscal year ended July 7.

38. The rapid growth of the civil service and its organizational complexity made supervision and control more difficult. In addition, performance incentives were diminished by appointments to higher positions based on political affiliation rather than merit. Furthermore, since the salary scale introduced in 1972 (Scale Regulation No. 2, Legal Notice No. 419) was not adequately adjusted over time to the increasing cost of living, in particular for professional staff, real wages declined.38 The emphasis on hiring support staff whose wages tended to be adjusted more than those of the professional staff contributed to a compression of the wage structure. Consequently, the quality of staff and its performance deteriorated significantly.

B. Reforms During the 1990s

Initial reforms: 1990/91-1994/95

39. The measures undertaken by the TGE in the early 1990s focused exclusively on the quantitative, “first-generation” aspects of the reform, such as salary reforms and a limited retrenchment of workers during the restructuring of ministries.39 Little attention was paid to “second-generation” aspects, such as improving the quality of staff, providing incentives for better performance, and fighting corruption.

40. The government attempted to streamline the organizational structure by redeploying some civil servants from the center to the regions, introducing an early retirement scheme for workers aged over 45 with 20 or more years of service, and retrenching redundant workers. For example, information available for 14 ministries and government organizations restructured before June 1994 shows that 2,325 (36 percent) out of the 6,343 employees submitted by various ministries for review were redundant. Consequently, 516 civil servants were transferred to the regional governments, 465 were sent to early retirement, and 1,553 were retrenched.40 However, the retrenchment was not aimed at improving the quality of staff: most retrenched workers lost their jobs because their positions were eliminated, not because they performed poorly. Furthermore, as part of the restructuring of ministries, some positions were being eliminated (and workers retrenched) at the same time that creation of new positions and new hiring were taking place, such that the size of the civil service increased to 293,452 employees by 1994/95 (a 36 percent increase since 1990/91).

41. Salaries were also increased in addition to the employment and organizational changes adopted in the early 1990s. The government announced an increase in the minimum wage for the first time since 1975 from Br 50 to Br 105 in 1992/93, but by an amount (110 percent) that proved insufficient to prevent a decline in the real minimum wage (as noted below, the increase only came into effect in 1993/94).41 Increases for other categories were also announced. However, since the increases granted to workers earning more than the minimum wage in absolute terms were almost identical for all grades, the decline in real wages was correspondingly larger for the higher grades (Table 10), further compressing the salary distribution (Figure 5). More specifically, before the salary increase, the bottom 70 percent of employees would have received about 49 percent of the total wage bill and the top 10 percent received 11 percent; after the increase, the share of the bottom 70 percent would have risen to 54 percent while that of the top 10 percent would have fallen to 9 percent.42

Table 10.

Ethiopia: Civil Service Salary Distribution, Before and After 1992/93 Wage Increase1

(In birr, unless otherwise indicated)

article image
Source: Ethiopian authorities; and Fund staff estimates.

For fiscal year ended July 7.

Figure 5.
Figure 5.

Ethiopia: Civil Service Salary Distribution Before and After the 1992/93 Reform1

Citation: IMF Staff Country Reports 1999, 098; 10.5089/9781451812626.002.A003

Source: Ethiopian authorities; and Fund staff calculations.1Fiscal year ended July 7.

42. The decision to raise salaries was not implemented until 1993/94, resulting in a significant further decline in real remunerations. In November 1994, the government introduced a revised salary scale for civil servants, including a career structure for teachers to enhance motivation. Except for professional and scientific staff, the changes in the ranges for starting salaries had the effect of narrowing the range for each category, as larger percentage increases were accelerated at the lower ends of the range for each category (Table 11).

Table 11.

Ethiopia: Monthly Starting Salaries of Civil Service in 1994/95 (By Employment Category)

(In birr, unless otherwise indicated)

article image
Sources: Ethiopian authorities; and Fund staff estimates.

43. Since the distribution of salaries by profession does not necessarily tell much about the returns to human capital (education, experience, etc.), a breakdown of starting salaries by education is provided in Table 12. It shows that the November 1994 salary adjustment provided the largest percentage increases to employees with the lowest educational levels.

Table 12.

Ethiopia: Monthly Starting Salaries of Civil Service in 1994/95 (By Level of Education)

(In birr, unless otherwise indicated)

article image
Sources: Ethiopian authorities; and Fund staff estimates.

44. Although the federal government increased employment in the higher-salary categories, the regional governments hired more support staff in the lower grades, such that the impact of the 1994 wage adjustment on the wage distribution within the civil service was only minimal (Figure 6). In addition, implementation of the new wage structure proceeded much more slowly than expected. While the 1994/95 budget allocated almost Br 200 million (14 percent) for wage increases owing to the new salary structure (not including the budgeted increase due to other factors), the total increase in the wage bill for that year was only 9 percent, contributing to a decline in the share of wages and salaries in GDP to 5.6 percent, from 6.2 percent in 1993/94. Table 13 also highlights a typical feature of civil service wage policy in Ethiopia: cost of living adjustments to wages were introduced infrequently and in amounts that generated a decline in real wages over time.

Figure 6.
Figure 6.

Ethiopia: Civil Service Salary Distribution Before and After the 1994/95 Reform1

Citation: IMF Staff Country Reports 1999, 098; 10.5089/9781451812626.002.A003

Source: Ethiopian authorities; and Fund staff calculations.1Fiscal year ended July 7.
Table 13.

Ethiopia: Total Employment and Wages in Civil Service, 1986/87-1994/95

article image
Source: Ethiopian authorities; and Fund staff estimates.

45. To summarize, the government of Ethiopia’s pay and employment policies in the early 1990s followed a trend similar to those of other sub-Saharan African countries during the 1970s and 1980s:

  • As Table 13 shows, real wages declined, and the anecdotal evidence suggests that the public-private sector wage differential narrowed during the early 1990s in favor of the private sector; however, according to Mengistae (1997), average wages in the public sector are still higher than those for comparable jobs in the private sector. He argues that the expectation of high public sector wage premiums creates an implicit queue of private sector workers for public sector jobs.

  • As the Ethiopian public sector moved from a low-employment/high-wage situation to a high-employment/low-wage situation (Figure 7), the average wage per civil service worker in percent of GDP converged to the ratio observed in other sub-Saharan African countries.43 In 1994/1995, the Ethiopian government wage bill (5.6 percent of GDP) was only slightly below the sub-Saharan average of 6.1 percent. At the same time, the number of civil servants per 1,000 population was 5.2 in Ethiopia, compared with an average of 10.5 in sub-Saharan Africa (Lienert and Modi, 1997), suggesting that wages of civil servants in Ethiopia are high relative to average income.

  • The wage structure became significantly compressed as a result of the government’s policy of protecting wages in the lower grades of the pay structure at the expense of the higher grades, further reducing incentives to acquire skills.

Figure 7.
Figure 7.

Ethiopia: Real Wages and Employment in Civil Service During 1986/87-1996/971

(1986/87=100)

Citation: IMF Staff Country Reports 1999, 098; 10.5089/9781451812626.002.A003

Source: Ethiopian authorities; and Fund staff calculations.1Fiscal year ended July 7.

Reforms since 1995/96

Civil Service Reform Program

46. In 1994/95, the government established a task force (consisting of 23 senior officials) to undertake a comprehensive evaluation of the civil service, based on the premise that pay and employment reforms alone would not lead to the desired improvement in civil service performance. The task force’s final report identified weaknesses and proposed solutions in the following areas: (i) expenditure management and control; (ii) human resources management, (iii) services delivery; (iv) top management systems; and (v) ethics. A theme unifying all these areas was the need for capacity building at the federal, regional, and subregional levels. Based on the recommendations of the report as well as a major contribution from the World Bank, particularly in the area of expenditure management and control, the government approved the Civil Service Reform Program (CSRP) in March 1996.

47. Given the weak implementation capacity in the civil service, the government decided to carry out the CSRP in increments. So far, most progress has been achieved in the area of expenditure management and control. A legal framework for management of financial resources by the federal government has been issued, and its implementation has begun. Accounting procedures have also been improved, in particular the speed with which the public accounts are closed.44 Strategies have been designed in the areas of human resources management and services delivery, but implementation has not yet started. With respect to human resources management, a decision has been made to improve job classification, performance appraisals, and the recruitment and promotion process. In the services delivery area, a decision has been made to encourage an outcome-oriented behavior by linking, whenever possible, the allocation of resources to the attainment of objectives in the provision of public services. At the beginning of every fiscal year, each part of the civil service would state its objectives for the year, explain how these activities would be funded, and establish indicators of services delivery. Strategies for the areas of ethics and top management systems still need to be formulated.

Decentralization

48. Although decentralization was never explicitly stated as an objective of the CSRP, it permeates all of the program’s components. The main reasons behind the government’s push for decentralization were as much economic as political, in that it gives local governments decentralization and various regional and ethnic groups increased authority to handle their own affairs.45 However, the regions remain financially dependent on the federal government—according to the 1995/96-1997/98 budgets, the regions financed on average only about 30 percent of their expenditures from their own sources. The implementation of regional administrative authority started in the early 1990s with the transfer of employees from the federal government to the regions. As regions were given the right to conduct their own hiring, employment at the regional level soared. Regional civil service employment increased to 265,198 employees by 1996/97, constituting about 85 percent of the permanent civil service. The trend of hiring mainly workers in administrative positions continued, and the share of professional workers was less than 6 percent at the regional level in 1995/96 (for women, that share less than 2 percent) (Figure 8).

Figure 8.
Figure 8.

Ethiopia: Distribution of Employment by Type of Service and by Regions, 1994/951

Citation: IMF Staff Country Reports 1999, 098; 10.5089/9781451812626.002.A003

Source: Ethiopian authorities; and Fund staff calculations.1Fiscal year ended July 7.

49. Decentralization and the recruitment for professional positions at the regional level have been complicated by a severe shortage of skilled workers; the majority of workers have at most an elementary school education (Figure 9).46 So far, the government has undertaken mainly an ad hoc approach to training, driven to a large extent by the availability of donor funding. There are only two training institutions (the Civil Service College, established in 1995, and the Ethiopian Management Institute), and neither of them provides a systematic approach to training civil servants. In addition, both institutions lack capacity themselves. The Civil Service College, which awards degrees in economics and law and emphasizes long-term training, graduates only a small number of students per year and cannot be expected to bridge the skills gap in the near future. Short term (one-week) training is conducted by the Ethiopian Economic Institute, but the quality of education suffers from the lack of a professional staff.

Figure 9.
Figure 9.

Ethiopia: Education in Selected Regions in 1994/951

Citation: IMF Staff Country Reports 1999, 098; 10.5089/9781451812626.002.A003

Source: Ethiopian authorities; and Fund staff calculations.1Fiscal year ended July 7.

50. In addition to the shortage of skilled workers, obstacles to successful decentralization include poor communication and coordination between federal and regional governments (caused to a large degree by poorly functioning infrastructure, such as telephones, roads, computers, and fax machines), and the regional governments’ lack of accountability to the center for the delivery of services. Plans to improve all these areas are specified in the CSRP but will require long-term effort and substantial investment.

C. Demobilization in the Early 1990s and the Impact on Civil Service

51. Civil service reform took into account the demobilization of troops after the defeat of the previous regime in 1991. Demobilization and reintegration into society involved half a million soldiers. Former combatants were divided into four groups: (i) rural, (ii) urban, (iii) those who had served less than 18 months, and (iv) disabled veterans. During the demobilization phase (July 1991-January 1992), veterans reported to discharge centers, from which they were transferred to their original communities. These centers received basic humanitarian assistance from donors, but lacked food and drinkable water and were overcrowded. The immediate needs relating to the transfer of veterans to their communities were addressed by a transitional safety net package that provided them with financial and in-kind assistance, according to the location of settlement and duration of service. The strategy during the reintegration phase (1991-95) aimed at identifying the minimal necessary assistance for each category, so that veterans could achieve the same status as average civilians in that category.47 The strategy is regarded as having succeeded in rural areas (i.e., ex-combatants have achieved the living standard of the average civilian), where living standards are low. In urban areas, however, the process of demobilization and reintegration was hampered by lack of skills on the part of veterans and the generally depressed labor market (Dercon and Ayalew, 1998).

52. The direct cost of demobilization and reintegration during 1991-95 was estimated at an annual average of 0.4 percent of GDP during the period (Colletta, Kostner, and Wiedenhofer, 1996). At the same time, military expenditures declined from an annual average of 9.6 percent of GDP during 1988/89-1990/91 to an average of 2.4 percent during 1991/92-1995/96. The decline in military expenditures was accompanied by a fall in tax revenues from 14.1 percent of GDP in 1988/89 to 7.8 percent in 1991/92,48 but altogether there was a net “peace dividend” that freed resources for increased social expenditures.49

53. In summary, the reforms in the early 1990s focused on quantitative (first generation) changes in the civil service, such as salary increases and retrenchment. Failure at that stage to take into account qualitative aspects, such as repercussions on employee performance, led in some instances to a worsening in the incentive system (for example, through wage compression and retrenchment of workers based on positions they occupied rather than performance). These shortcomings were addressed in the design of the CSRP, which focused on the quality of performance. Substantial progress has since been achieved in some areas, such as expenditure management, legal framework, and accounting, although the outbreak of a border conflict with Eritrea in May 1998 has posed new challenges regarding expenditure management.

References

  • Bevan, David, and Sanjay Pradhan, 1994, “Fiscal Aspects of the Transition to Peace: with illustrations from Uganda and Ethiopia,” in Some Economic Consequences of Transition from Civil War to Peace, World Bank Policy Research Working Paper No. 1392, by Jean-Paul Azam and others (Washington: World Bank).

    • Search Google Scholar
    • Export Citation
  • Colletta, Nat, Markus Kostner, and Ingo Wiederhofer, 1996, “Case Studies in War-to-Peace Transition: The Demobilization and Reintegration of Ex-Combatants in Ethiopia, Namibia, and Uganda,World Bank Discussion Paper No. 331 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation
  • Collier, Paul, 1995: “Civil War and the Economics of the Peace Dividend,Centre for Study of African Economies at Oxford University Working Paper WPS/95-8 (Oxford: University of Oxford).

    • Search Google Scholar
    • Export Citation
  • Dercon, Stefan and Daniel Ayalew (1998), “Where Have All the Soldiers Gone: Demobilization and Reintegration in Ethiopia,World Development, Vol. 26 (September), pp. 16611675.

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 1998, Uganda - Selected Issues and Statistical Appendix, SM/98/61 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Lienert, Ian, and Jitendra Modi, 1997, “A Decade of Civil Service Reform in Sub-Saharan Africa,IMF Working Paper 97/179 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Mengistae, Taye, 1999, “Wage Rates and Job Queues: Does the Public Sector Overpay in Ethiopia?World Bank Policy Research Working Paper No. 2105 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation
34

Prepared by Zuzana Brixiova.

35

A survey undertaken in 1983 by the Ministry of Labor and Social Affairs, which included establishments employing ten or more people, put the public sector employment at 516,864 (out of a total of 708,565 employees in the formal sector), i.e., 73 percent of employment in the formal sector. Of this, 183,405 were civil service workers. The total labor force in 1983 was estimated at slightly below 14 million, indicating that employment in the formal sector represented only 5 percent of the labor force.

36

Owing to the large population growth during this period, the number of civil servants per 1,000 population increased only from 4.8 in 1984/85 to 5 in 1989/90. According to data in Lienert and Modi (1997), only the following sub-Saharan African countries had fewer civil servants per 1,000 population in 1990: Chad (4.6), Mali (4.8), and Niger (4.6). The average for sub-Saharan Africa in 1991 was 12.7 civil servants per 1,000 population (the average for 1990 is not available).

37

Like other command economies, the Derg regime had biased the education system toward the hard sciences and engineering, neglecting the social sciences, law, and public policy.

38

For example, the civil service minimum wage was increased only once between 1972 and 1990, by 100 percent (from Br 25 to Br 50 in 1975), even though the price level during the same period increased by about 450 percent. Furthermore, the two-year step increments in wages were frozen for all employees earning more than Br 285 per month from 1975 to 1982.

39

In spite of this limited retrenchment, the total number of civil servants increased.

40

Retrenched workers with one year of service received as severance payments three months’ salary, and those with more than one year of service received one-third of their monthly salary for each year of service, up to twelve months. Severance payments in 1994, estimated at less than 0.1 percent of GDP, had little budgetary impact.

41

The price level had increased by 330 percent during the same period. Furthermore, the increase of the minimum wage to Br 105 effectively eliminated the first nine grades of the previous wage scale.

42

For comparison, in Uganda in 1994/95, the bottom 70 percent received 59 percent, and the top 10 percent received 18 percent of the total wage bill. As in-kind benefits in Uganda were monetized and included in wages in 1996, the wage structure became less compressed and the bottom 70 percent received 49 percent while the top 10 percent received 25 percent of the total wage bill, making Uganda’s wage structure less compressed than Ethiopia’s (IMF, 1998).

43

There is no clear criterion to determine which wage/employment combination would best achieve the various objectives of the public sector. In Ethiopia, the high growth of employment in the public sector, together with the decline in real wages, has probably led to a decrease in efficiency, along with lower morale and a deterioration of skills in the civil service.

44

Until 1997, past accounts were closed with delays of three-four years, but this lag has since been significantly reduced. The goal is to reduce the delay to seven-eight months, in line with international standards.

45

District governments are expected to formulate their own budgets and be responsible for delivery of services in crucial areas, such as health, education, and soil conservation. In addition, they are expected to enforce human rights and law compliance.

46

The shortage of skilled female workers is particularly serious. In 1995, women with master’s degrees or higher education constituted less than 1 percent of permanent employees at both the federal and regional levels; the figures for males were 6 percent at the federal and 1 percent at the regional government levels.

47

This was a challenging task, as the majority of the 455,000 veterans were less than 25 years old and illiterate and had virtually no skills. Participants in the rural reintegration program received technical and material support to engage in small farming. The urban reintegration program initially consisted of employment, education, and training; eventually, a Revolving Credit Fund was established to provide concessional loans to cooperatives of urban ex-combatants.

48

According to Bevan and Pradhan (1994, p. 91), this drop in tax revenue can partly be attributed to the end of the war and demobilization process. As they put it, “this fall partly reflects the operation of a relatively primitive and yet somewhat unreformed tax system when excessive coercion is removed.”

49

In particular, expenditures on both public health and on education and training increased significantly from 1988/89-1990/91 to 1991/92-1995/96, namely from an average of 0.9 percent of GDP to 1.1 percent in the case of health expenditures, and from 2.7 percent of GDP to 3.5 percent for education and training.

Ethiopia: Recent Economic Developments
Author: International Monetary Fund
  • View in gallery

    Ethiopia: Number of Civil Servants, 1974/75 - 1995/961

  • View in gallery

    Ethiopia: Civil Service Salary Distribution Before and After the 1992/93 Reform1

  • View in gallery

    Ethiopia: Civil Service Salary Distribution Before and After the 1994/95 Reform1

  • View in gallery

    Ethiopia: Real Wages and Employment in Civil Service During 1986/87-1996/971

    (1986/87=100)

  • View in gallery

    Ethiopia: Distribution of Employment by Type of Service and by Regions, 1994/951

  • View in gallery

    Ethiopia: Education in Selected Regions in 1994/951