Making education reform work for the poor: accountability and decentralization in Latin America

Abstract

We build upon and examine critically the framework for analysing accountability set forth in the World Bank’s World Development Report 2004 (WDR04) through a review of selected literature studying accountability‐focused reforms in three Latin American countries – Brazil, Colombia and Chile. We examine the successes and pitfalls of the three accountability‐related reforms (which also involve some type of decentralization) in fostering institutional environments in which the key actors involved in provision are held responsible for fulfilling their roles in providing education services. We apply the WDR04 framework in our analysis of the process of implementation and impact of different types of accountability‐focused reforms in the three countries. The framework was conceptualized by analysing case studies such as ours for a broad range of sectors, and now has widening influence in the policy‐making community, thus justifying a critical assessment. We discuss common challenges impeding implementation of the reforms, most notably political and technical challenges and discuss how well the framework and report capture and emphasize such challenges. In studying the impact of the reforms, we find that their effects are generally more positive in less poor communities and discuss the need for a strong central state to monitor and address such issues.

Keywords:

• accountability
• decentralization
• education reform
• World Bank
• Latin America

Acknowledgements

The authors thank Elizabeth Carlton, Margarita López and Margarita Peña for helpful comments.

Notes

1. The El Salvadoran EDUCO (Educación con Participación de la Comunidad) reform, for example, a school‐based management‐focused reform, is mentioned 50 times in the report.

2. We do not claim that the writers of the WDR04 were unaware of these issues, and they are indeed mentioned in the report.

3. It is important to note that there is a very important distinction between the two terms. Referring to citizens as clients may infer that they are simple recipients of goods and services, as in a market, while in the public sector the relationship between citizens and government is much more complex. Citizens, for instance, generally express preferences for what they consume through the democratic process and they value not only the private value of the goods, but also consider the broader public value (Alford 2002).We use the term client throughout the report in order to be consistent with the terms used in the WDR04.

4. The response to the problem of teacher absenteeism is provided as an example of how the accountability relationships might function in an effective system and not as a way of supporting this specific policy.

5. Policy‐makers create the policy framework that the organizational providers work in and leaders of organizational providers make internal policies for the organization. Occasionally, as the WDR04 points out, the distinction between the leader of the organizational providers and policy‐makers is not always clear.

6. For an excellent review on the subject of teacher incentives, see Vegas and Umansky’s (2005) report, ‘Improving teaching and learning through effective incentives: What can we learn from education reforms in Latin America?’

7. The concept of voice was first developed by Hirschman (1970) in his book Exit, Voice, and Loyalty. For Hirschman voice is discussed as one option, along with exit, open to consumers as a product with which they have become dissatisfied.

8. Murillo (1999), in her review of the role teachers’ unions played in decentralization reforms in Mexico and Argentina shows, in the case of Mexico, the strong connection between the centralized monopolistic national union and the central government inhibited the motivation of the union to represent teachers. Union leaders were given political and administrative positions within the government and ministry as a reward for loyalty. This was clientelism writ large.

9. The WDR04 discusses the accountability problems that result when reforms are implemented due to donor conditions rather than client influence stating, ‘There is ample evidence today that conditions based on promises do not work well, because they undermine ownership of the reform program’ (209).

10. Some have theorized that the gap in coverage and quality is related to income inequality in Latin America, which is higher than any other country in the world (see Reimers 2000; Engerman and Sokoloff 2002).

11. Works by Fischman, Ball, and Gvirtz (2003) and Torres and Puiggrós (1995) argue that ‘neoliberal’ reforms in Latin America may have been accompanied by a decrease in concern for promoting equity and education due to the focus on global competition.

12. Lamounier (2003) argues, however, that Brazil is not nearly as diverse and difficult to govern as some countries. It is less ethnically and geographically diverse than the USA, for example.

13. At the time of Birdsall, Bruns, and Sabot’s publication (the early 1990s), there was great variation in per pupil spending between municipal and state‐run schools within the same municipalities, with state‐run school sometimes receiving as much as 10 times as municipal schools.

14. Although the reforms were later rolled back in that state after a change in government.

15. More recent reforms in Brazil may have been implemented to address these inequities.

16. The first step of decentralization was political decentralization, and mayors and governors were elected by popular vote beginning in 1985. Later, legislation was passed that increased the amount of money transferred to municipalities and in education, transferred to them the responsibilities of school construction and maintenance and later human resource responsibilities including teacher assignment, transfer and firing (Fiske 1996).

17. Beginning in 1968 funding was transferred to the departmental level. Curriculum development took place at the central‐level beginning in 1976 (Reyes 2001).

18. It is important to point out, however, that larger Colombian cities, specifically Bogotá, face their own unique challenges in the administration of education including dramatic increases in population, especially among the poor and the displaced and the challenge of providing education services in underdeveloped and often violent areas of the city.

19. In Bogotá in 2003, the combination of payroll and required recurrent costs amounted to 93% of the total central transfer, leaving 7% for spending in other categories (Peña and Rodríguez 2005). Also in Bogotá, local education funding made up a significant portion (30%) of total basic education funding in the city in 2003 (Peña and Rodríguez 2005).

20. While data early in the reform indicates that municipal schools were more likely to serve poor students, more recent analysis shows that both types of schools serve a wide range of students from different socioeconomic levels. According to a national socioeconomic survey conducted by the Ministerio de Planificación in 1990, 70% of students enrolled in municipal schools were made up of the country’s poorest inhabitants, compared to 50% in subsidized private schools (Espinola 1994). Mizala, Romaguera, and Ostoic (2004) find, however, that subsidized private schools ‘have not “specialized” in a specific socioeconomic group but provide educational services to broad representation of the population. This result does not support the claim of school choice critics that the new schools will concentrate children of parents with high socioeconomic status’ (15). Parry (1997) discusses high enrolment in municipal schools in Santiago’s wealthy areas.

21. Also part of the reform was a system of standardized tests which was created in order to provide information to clients regarding school performance. Scores were not made public, however, until 1995 (Mizala, Romaguera, and Ostoic 2004).

22. Both municipal schools and state‐subsidized private schools are financed like the subsidized private schools by a per‐student subsidy. Some subsidized private schools, however, receive additional funds from parent fees in the mid 1990s (Mizala, Romaguera, and Ostoic 2004).

23. Although Tokman (2004) finds that Chile has lower than expected schooling for preschool and post‐secondary school and lower than expected quality.

24. As stated previously, state‐subsidized schools maintained greater flexibility in hiring and dismissing teachers as well as negotiating wages. Teachers in private subsidized schools tend to by younger and were more likely to have studied in private universities (Mizala and Romaguera 2004).

25. Mizala and Romaguera (2004), in an analysis of a national teacher survey, did not observe significant differences in teacher satisfaction with the teaching profession and in opinion regarding the working climate between teachers in state‐subsidized private schools and teachers in municipal schools.

26. About the issue of the inability of most education systems to provide incentives, the WDR04 states, ‘a total lack of connection between incentives and performance allows excellent teachers working in adverse circumstances and those who never show up to be paid the same amount. This undermines the morale of good teachers and drives them out of the profession’ (124).

27. Moreover, Fischer, González, and Serra (2006) state the increase in salaries had by the late 1990s led to a slow improvement in the quality of new teachers entering the workforce. The authors, however, critique the ED discussing how it made it difficult to firex underperforming teachers. In 2005, the teachers’ union accepted an evaluation programme for teachers that provide a remedial programme and the potential for dismissal with severance for underperforming teachers (Fischer, González, and Serra 2006).

28. In Colombia, the national teachers’ union accepted a salary incentive programme after 80% of funding for the programme was targeted towards general teacher salary increases (Herrán and Uribe 2004).