Building a Culture of Accountability in Service Delivery: Conclusions from the GDN Project on Varieties of Governance and Service Delivery
Guillermo Perry and Ramona Angelescu Naqvi
The GDN Project on Varieties of Governance in Service Delivery
This chapter summarizes the main conclusions from a Global Development Network (GDN) research project that was designed and directed by the authors. It gives detailed references to the three Latin American case studies published in this volume, which examined the impact of decentralization reforms on quality of education and water supply in Colombia, the effects of citizen participation in municipal budgeting on the quality of water supply in Peru and the consequences of a wide set of reforms—decentralization, teacher
G. Perry (*)
Universidad de Los Andes, Bogota, Colombia Center for Global Development, Washington, DC, USA
R. Angelescu Naqvi
Global Development Network, New Delhi, India
© The Author(s) 2017 1
G. Perry, R. Angelescu Naqvi, Improving Access and Quality of Public Services in Latin America, Latin American Political Economy,
incentives, school autonomy—on the quality and equity of basic education in Chile and Uruguay. It also makes occasional references to some of the other 13 country or multi-country case studies of the GDN project (four in Asia, six in Africa and three in Central and Eastern Europe), covering 30 countries in total. All of these case studies were selected through an open international competitive call on the basis of both the quality of the proposal and the intrinsic interest of the topics under study. The GDN project, as well as this chapter, benehted from a background paper on political economy issues related to service delivery, which is published as Chapter 2 in this volume, in which the authors analyze the role of incentives, institutions and politics in markets and hierarchies in public service provision in Latin America, and three sectoral background papers that surveyed the theoretical and empirical literature on governance and institutional issues in the three sectors selected at the beginning of the project (basic education, water supply and roads).1
The case studies focused on governance reforms of particular importance for one or more of these sectors, usually related to decentralization, community participation or private sector participation and the role of quasi-markets in service delivery. Authors were encouraged to analyze the role of different actors in the reform process and its implementation, the incentives they faced and the impact of the reform on coverage and, especially, on quality and equity of access. Emphasis was put on identifying the role of formal and informal accountability process, information flows and financing structures. Case studies used both econometric and qualitative (surveys, structured interviews) techniques as complementary tools.
notwithstanding, many of its case studies also analyzed quantitative service outcomes in these sectors.
Quality and equity of access problems are only partially due to scarcity of financial resources. To a large extent they are due to the lack of adequate incentives along the delivery chain, which, in turn, are associated with inadequate governance, institutions3 and policies. Several previous studies, starting with the 2004 World Bank Development Report,4 have pointed out to the lack of adequate mechanisms of accountability of service providers (and of policy makers) to users and citizens as the main reason behind poor incentives and poor service delivery performance.
The case studies in the GDN project illustrate why accountability mechanisms are especially weak with respect to quality and equity outcomes, reflecting the fact that principal/agent problems are particularly severe for such outcomes. This fact explains why most developing countries perform particularly badly with regard to quality and equity of access in service delivery.
To begin with, information asymmetries and lack of effective enforcement mechanisms, which are at the core of principal/agent problems,5 are much more pronounced with respect to quality than to quantity of services. Quantity is more easily observable than quality. When coupled with policy makers’ traditional short horizons (maintaining power and being re-elected), this creates a major bias against investments in and monitoring of quality. Public sector providers may also have relatively short horizons, if there is high turnover of managers and weak administrative careers. Private providers, without adequate regulation, may be willing to compromise on quality in order to increase short-term profits, as discussed in the second chapter in this volume6 and found in some of the empirical case studies.7
A few examples from the GDN project case studies and background papers illustrate these problems in the selected sectors. Parents can easily observe that a school is being constructed or is functioning, that their children are attending school and that teachers are showing up in the school. But neither the parents nor the policy makers observe what happens in the classroom: how teachers are teaching and how much children are learning. Thus, they may become aware of poor quality only much later, when their kids fail to make it to university.8
Similarly, users (and policy makers) can easily check if water is available and with what regularity. It is more difficult for them to gauge quality. citizens know how the water looks and tastes, but they may only learn much later that their children were sick because of poor quality of water,9 In the same vein, neighbors and users know if a road is constructed and how it looks like, but they may only know after some years, just when it breaks down and needs major repairs or reconstruction, that it was built with sub-standard materials or was poorly maintained,10
In all these cases, citizens may become aware of quality problems only when it is too late to punish those policy makers or providers that were responsible for such poor performance, Consequently, policy makers and providers around the world are much more interested in building schools, roads and water systems, because their voters can immediately observe them, than in making sustained efforts to improve quality, that will only be recognized in the longer term, In addition, construction contracts can be used to beneht campaign financiers, and construction companies are, partially because of this reason, a powerful lobby behind public investment in most countries (see the Castelar and Schneider paper in this volume), No equally strong private lobbies normally advocate for maintenance budgets, In a similar fashion, policy makers may put more effort in making sure the children and teachers are in school, or that water is available most of the time, than in setting up and administering complex quality assurance and information dissemination systems, As the effects of poor quality may only be fully known over time, they may not affect citizens’ satisfaction levels in the shortterm and, hence, policy makers’ re-election chances, Citizens’ awareness of quality problems will likely be a problem for future governments and public managers, which may be different from those taking present decisions,
Several institutions and policies can partially correct these problems, Earmarking or setting apart funds for maintenance and quality enhancement may partially correct this bias, Many countries use such earmarking and special funds set by law, for this reason, Administrative monitoring and information systems are normally used in every education system, including, in some cases, inspectors randomly attending classes, periodic testing of children’s progress in learning abilities is often a complement to reduce these information gaps, Similarly, policy makers can organize periodic testing for water quality and an effective system of supervision of road quality and maintenance, hopefully carried out by independent and competent agents, However, such information systems may only affect performance if they lead to effective sanctions and rewards for providers and policy makers, and to policy evaluation and change, Several GDN project case studies found that internal administrative information system results are rarely known to parents, users and citizens, and thus often become pure formalities, with minor consequences for providers. Hence, they have, at most, marginal effects on actual performance (more on this in Sect. 4 below). To affect incentives they must be widely disseminated (so that parents and citizens can use them to demand improvements) and should lead to effective rewards or punishment of providers (e.g. improvements in student test results should be widely publicized and used as a basis for promotions or bonuses for principals and teachers and to more school autonomy, as in Chile). They should also lead to policy evaluation and adaptation, either by action of hierarchical superiors (policy makers) or through citizens’ pressure, including higher participation or voting with their feet, when options for voice or choice are available.
The so-called short-route mechanisms11 can partially mitigate some of these problems, provided that technical information on quality is amply provided to citizens and citizens are able to use it effectively to influence policy makers’ or service providers’ decisions. However, this is not normally the case, limiting the effectiveness of institutionalized participatory systems, as the Peruvian case shows. Exercising voice (in participatory mechanisms) or choice (when there is competition among providers) requires access to good information on quality, which, as mentioned, is not normally available to citizens. They also require a culture of participation, whereby citizens actively seek this kind of information and use it as a basis for their decisions. In nascent democracies, such as most of the ones covered in this project in Central and Eastern Europe and the Caucasus (ten countries covered in three comparative case studies), that is not often the case, and institutionalized citizen participation channels are found to have little impact.
Private and community participation in provision may also help in some cases in this regard. Such providers normally have longer horizons than politicians and public managers, and thus are likely to care more for maintenance and quality. If they do not, their enterprises or cooperatives may eventually become non-viable or bankrupt and cannot rely on public rescue, as public utilities normally do. Well-structured water and road concession contracts may help to reduce the bias against maintenance and quality which is typical of purely public provision.
However, these alternative modes of delivery should be well-regulated and complemented by public testing and dissemination of information on quality, and corresponding rewards and sanctions to providers, to avoid the temptation to maximize short-term revenues by sacrificing quality, as found in one of the GDN case studies in Central Europe.12 And regulation of quasi-markets for competitive provision of basic services is even more intensive in institutional and informational requirements than hierarchical systems of service delivery, as discussed in the Castelar and Schneider paper in this volume. Finally, as discussed below, private providers of non-regulated water often charge excessively high prices and offer poor quality in poor neighborhoods.
Similarly, accountability channels normally work better for the rich and the middle classes than for the poor. Poor families or regions, or rural areas, usually have less access to information and less political and economic power than richer urban households. Policy makers normally care more for the more powerful voice and mobilization potential of the urban rich and middle-class populations and for their swing vote. Service providers also care more for these larger and more concentrated consumers, which are easier and more profitable to serve. Poorer and rural families are normally in a disadvantaged position to exercise their voice, due to lower access to the media and policy makers, or lower opportunities to choose, as they rarely have access to alternative providers. As an example, the Peruvian case study found that poor rural communities do not participate in the municipal budget participatory process set by law, unless authorities make a special effort to reach out to them, which they seldom do.
As a consequence, and especially in highly unequal societies, access to services, and especially to quality services, is highly uneven. Children of poor families normally have lower enrolment rates and higher repetition and dropout rates. Further, they normally attend lower quality schools. In some countries, especially in rural environments, girls still have less access than boys to schooling and other public services. Poor neighborhoods and rural populations normally have less access to regulated water grids (either run by public or private providers) and have to rely on unregulated, poor-quality and high-cost private providers.13 These poor neighborhoods and rural centers are usually not as well connected—both in a physical and a political sense—as rich or middle-class urban neighborhoods and large cities.
To counteract such biases, there is normally the need to establish special institutions and policies in charge of leveling the playing field. Otherwise, normal accountability channels, whether through the long route or shorter routes, are likely to behave lopsidedly in favor of the richer, more educated and urban middle classes, hence maintaining unequal access to basic services that reproduce inequality along and across generations. The Peru case study illustrates this last point well. In contrast, the Indian case study in the GDN project found that political reservations and quotas based on gender, caste and tribe for elected ofhcials do appear to contribute to higher equality of access, as well as to improved quality of services. There is a much stronger and deep-rooted culture of participation in India than in the rest of the countries covered in the project, especially the newer democracies in the post-communist region or in Latin America.