Research and policy in international development: introduction

JULIUS COURT, INGIE HOVLAND and JOHN YOUNG

Background

The challenge of evidence-based policy making

Better utilization of research and evidence in development policy and practice can help save lives, reduce poverty and improve the quality of life. For instance, the results of household disease surveys informed processes of health service reform which contributed to 43 and 46 per cent reductions in infant mortality between 2000 and 2003 in two districts in rural Tanzania. A recent study by the UK Department for International Development (DFID), as part of its effort to develop a new research policy, forcefully catalogues the value of research for development (Surr et al. 2002). However, researchers frequently do not see influencing activities as a priority once the research is completed. Policy makers and other stakeholders often do not know what research exists, which policies are most suitable, or how they can best be implemented in different contexts. Or, as is illustrated (with devastating consequences) by the lack of response to HIV/AIDS in some countries, policy makers may be unwilling to act on the evidence.

Major investment goes into research on international development – from stakeholders in the South (government and non-government) and from governments in the North and international agencies. While an exact figure is difficult to arrive at, we estimate that Northern and international sources provide around US$3 billion for international development research. If we are to reduce poverty in the South and meet the Millennium Development Goals (MDGs), the key questions are: When and why does development research make a difference? Why do some research findings influence policy and practice while others do not? And how can we promote more informed international development policy?

Despite the importance of research–policy questions in the international development sector, very little relevant academic work exists on the topic. Studies on evidence-based policy have mainly focused on policy processes in the OECD (Organisation for Economic Cooperation and Development) countries, paying particular attention to the national medical and educational sectors in these countries. In the USA, the 1980s has been termed the golden age of studies on knowledge utilization, following Weiss's (1977) seminal research on 'knowledge creep' and 'percolation'. In the UK, evidence-based policy is the central theme of the Centre for Management and Policy Studies, established by the Cabinet Office in mid-1999. In the same year, the Economic and Social Research Council (ESRC) established the Evidence Based Policy and Practice Initiative, a collaborative network of seven research units aiming to bring social-science research nearer to the decision-making process.

Analytical reflection on the many different contexts and factors influencing research–policy processes in countries outside the OECD is much more rare. This is no small challenge, as the massive diversity of cultural, economic and political contexts makes it especially difficult to draw valid generalizations and lessons from existing experience and theory. In addition, international actors have an exaggerated impact on research and policy processes in several countries in the South. Alongside these analytic gaps, there is a lack of relevant case-study material to illustrate and highlight successful and unsuccessful practices.

There has been some interest in these questions in the international development sector over the past few years. In 1998, the International Institute for Environment and Development (IIED) identified a six-point programme for improving the policy impact of research (Garrett and Islam 1998). The Overseas Development Institute (ODI) has been researching research–policy linkages since 1999, with an early report providing a 21-point checklist of what makes policies happen (Sutton 1999). Impact assessments by the International Food Policy Research Institute (IFPRI) have focused increasingly over the last few years on measuring the policy impact of its research programmes, and how it can be improved. The link between research and policy has been a key issue for the Global Development Network (GDN) since its inception in 1999. The International Development Research Centre (IDRC) (Canada) is currently completing a strategic evaluation of the influence of IDRC-supported research on public policy.

This volume aims to enhance understanding of how research can contribute to pro-poor policy and practice. We believe it is a useful contribution for a number of reasons. First, it presents a cohesive framework that can be used to understand and analyse research–policy linkages. Second, it provides evidence from four systematic and in-depth case studies. The approach is innovative in that each case details instances of policy change and then does a 'tracer' study backwards in time to chart the different factors – including research – that led to the policy shift. We believe the approach is valuable in that it highlights the range of factors that affect policy uptake of research. Third, through comparison across the four cases, we draw out themes that will be relevant to those involved in development research or policy and who wish to use evidence to contribute to pro-poor policy making. In particular, we provide suggestions for how researchers might enhance the policy impact of their work.

The global context

An understanding of research–policy linkages in international development needs to take into account both the wide range of different contexts that development policies appear in, and the key challenges of the macro-level political context. There are a number of notable global trends here.

Democratization has been one of the striking developments of the past decades. In 1901, there were no nations that could be considered democratic – even according to the most basic definition. Now there are estimated to be more than 121 electoral democracies (Freedom House 2003). From 1976 to 1999 the number of democratic regimes has more than doubled, from fewer than 40 to more than 80 (Gurr et al. 2001). This provides new entry points to the policy-making process. It is thought that democratic contexts would better enable research to be conducted and communicated (due to freedoms), and that there would be greater incentives for policy makers to use research (due to accountability mechanisms).

But many developing countries remain undemocratic and many countries have deficits in these areas even if they are seen as democratic in form (Hyden et al. 2004). Political systems are less representative and have weaker structures for aggregating and arbitrating interests in society, even in countries seen as democratic (Grindle 1980; Hyden et al. 2004). Policy-making processes tend to be more centralized and thus often remote and inaccessible with limited scope for wider input or participation except at implementation stage (Grindle 1980: 15). Often political leaders may view input from civil society as illegitimate or inefficient. The result is to increase policy makers' power whilst simultaneously isolating them from society. Policy formulation becomes responsive to the needs of élites rather than the majority – or the poor.

Markets are spreading and economies are increasingly open. Twenty years ago, only 2.9 billion people lived in a market economy, compared to 5.7 billion people today. What does this mean for research–policy linkages? Most obviously, it brings new actors into political processes. It also creates increased demand for research that can be accessed by those outside academia. There is demand for research by private companies; and there is demand for research by governments – so that they are able to cope with technical issues around economic policy formulation and regulation.

Civil society is generally thought to be blossoming in an increasing number of nations. The number of non-governmental organizations (NGOs) is growing. The role played by civil society (NGOs, the media, and think-tanks) in shaping national priorities is expanding. A theme common to many countries is the increasing importance of non-state organizations as actors in the governance realm (Edwards 2004). An apt example is the role played by civil society through consultations and follow-up in national Poverty Reduction Strategy Paper (PRSP) processes. Nevertheless, many challenges remain (Hyden et al. 2004). The input of civil society into public policy is still quite limited. Civil society and state often live rather separate lives, with governments continuing to set the policy agenda largely on their own. There is still a tendency for governments to arrest or intimidate citizens who propagate views different from those in power.

The rapid development of information and communication technology (ICT) has incredible potential to transform the generation and sharing of information. The poor are poor not just because they are marginalized in economic and political processes, but also because they are marginalized in information flows in society. Widespread dissemination through ICTs can make information accessible to a far wider audience than before, and this information can potentially act as a catalyst for action – even for groups that have traditionally been marginalized. However, ICTs are no miracle cure. The information gap is widening, and the need to be 'plugged in' to information networks in order to be able to influence policy may also exclude groups that fall outside the information society.

There is increasing concern that the 'war on terror' is leading to an increased politicization of aid and is diverting attention from poverty reduction (Christian Aid 2004). It is too soon to confirm the implications for research–policy links, but it appears that there is less emphasis on ensuring freer political contexts (important for collecting and communicating evidence) and on using evidence to encourage pro-poor policies in developing countries. The worry is that progress may stall in terms of the increasingly favourable context for pro-poor policy making that was evolving in many countries over the past two decades. These factors provide a broad outline of the context in which we wish to situate this study.

Conceptual framework

This global context raises the question of research–policy linkages. How can we understand research and policy in international development today? In this section, we provide a brief introduction to our approach to these concepts, and the way we have used them in the conceptual framework that guided the case studies.

Definitions: research and policy

In preparing the case studies, the project decided to use relatively open definitions of research and policy. This was important given the preliminary nature of the work, the diversity and complexity of the study topics, and the relative lack of existing case studies.

Like others, we thought it was difficult, and often unhelpful, to provide an overly specific definition of research since the exact meaning will depend on the context. For the case studies in the Research and Policy in Development (RAPID) project we considered research as 'any systematic effort to increase the stock of knowledge'. This included therefore any systematic process of critical investigation and evaluation, theory building, data collection, analysis and codification related to development policy and practice. It includes action research, self-reflection by practitioners oriented towards the enhancement of direct practice.

We suggest that policy is defined as a 'purposive course of action followed by an actor or set of actors' (Anderson 1975). Our suggestion is to focus on public policy with the understanding that policy is not restricted to government. International organizations, bilateral agencies and NGOs also all have policies. What makes policies public is not that they are adopted and implemented by government, but also that they affect the public or are visible to the public. The cases explored how research can influence policy makers' horizons, policy development, declared public policy regimes and policy implementation or practice (Lindquist 2001).

An integrated framework: context, evidence, links

Traditionally, the link between research and policy has been viewed as a linear process, whereby a set of research findings is shifted from the 'research sphere' over to the 'policy sphere', and then has some impact on policy makers' decisions. At least two of the assumptions underpinning this traditional view are now being questioned. First, the assumption that research influences policy in a one-way process (the linear model); second, the assumption that there is a clear divide between researchers and policy makers (the two-communities model).

Recent reviews of the literature on research–policy links move away from these assumptions, towards a more dynamic and complex view that emphasizes a two-way process between research and policy, shaped by multiple relations and reservoirs of knowledge (see, for example, Garrett and Islam 1998; Keeley and Scoones 2003; Lindquist 2001; Neilson 2001; RAWOO 2001; Stone 2001; Sutton 1999).

Following Weiss (1977), current research–policy reviews usually recognize that although research may not have direct impact on specific policies, the production of research may still exert a powerful indirect influence through introducing new terms and shaping the policy discourse. Weiss describes this as a process of 'percolation', in which research findings and concepts are gradually filtered through various policy networks. Thus, while the traditional question could be phrased: 'How can research be transported from the research to the policy sphere?', the new question to be addressed concerns research uptake pathways: 'Why are some of the ideas that circulate in the research-policy networks picked up and acted on, while others are ignored and disappear?' The answer to this seems to lie in a combination of several determining influences. The integrated framework developed by RAPID (Crewe and Young 2002) has broadly divided these influences into three areas, as illustrated in Figure 1.1 (see page 9).

First, research–policy linkages are shaped by political context. The policy process and the production of research are in themselves political processes: from the initial agenda-setting exercise through to the final negotiation involved in implementation. Political contestation, institutional pressures and vested interests matter greatly. So do the attitudes and incentives among officials, their room for manoeuvre, local history and power relations (Clay and Schaffer 1984; Kingdon 1984). In some cases the political strategies and power relations are obvious, and are tied to specific institutional pressures. In other cases the pressures are more vague, and are tied to broad discourses or paradigms that may exert a powerful influence over which ideas are noticed and which are ignored.

Second, the RAPID framework proposes that the nature of the evidence is important for policy uptake. The literature suggests that significant here are research approach and methodology, credibility of the evidence, simplicity of the message, how it is communicated and the degree it challenges received wisdom. The sources and conveyors of information may be as influential as the content; for example, people accept information more readily from those they trust. But it is our hypothesis that good-quality research, local involvement, accurate messages, and effective dissemination strategies are all important in aiming for more evidence-based policy making.

Another key set of issues concern how evidence is communicated. The way new messages are packaged (especially if they are couched in familiar terms) and targeted can make an enormous difference to how a policy document is perceived and utilized. For instance, marketing is based on the insight that packaging rather than content (in and of itself ) often determines people's reaction to a new product or idea (Williamson 1996). The key message is that communication is an extremely demanding process and that it is most effective to take an interactive approach (Mattelart and Mattelart 1998). Continuous interaction leads to greater chances of successful communication than a simple or linear approach.