A Research Agenda for Development
by Nicholas Stern

To help the development community be more effective in reducing poverty in developing countries, development-oriented research should be derived from a strategy that is based on two pillars: improving the investment climate and empowering and investing in poor people. The research agenda should rely on both empirical surveys and analytical understanding of the dynamics of the investment climate, of preferences, and of political reform. Research and action must reinforce each other.

The World Bank’s development strategy can be described in terms of two basic pillars. The first pillar focuses on improving the investment climate. It involves more than just the levels of investment and also includes the environment for entrepreneurship and the investment process itself. The aim is to create a good investment climate, that is, one that encourages private firms to invest, create jobs, and increase productivity through appropriate government policy and behavior. The three important elements of a good investment climate are macroeconomic stability and openness to trade; good governance and institutions, including control of bureaucratic harassment and crime, as well as effective financial and legal institutions; and adequate transportation, power, and communications infrastructure.

The second pillar focuses on empowering and investing in poor people by ensuring that they have opportunities for education and health care, avenues for risk reduction and mitigation, and mechanisms for participating in the key decisions that affect them and their families. Empowerment is both an end in itself and an avenue to development. Indeed, it is key to scaling up to achieve development results because it not only fosters social inclusion and participation in growth, but also promotes growth by mobilizing poor people’s assets and energies (box).

Thus the processes embodied in the two pillars are mutually reinforcing in their effects on growth and poverty reduction. This two-pillar strategy suggests a program for research that can take forward both the theoretical and the empirical investigation of public policy.

Appropriate Data

The two-pillar strategy suggests a need to gather and analyze two new types of micro data: first, data from firms to provide detail on the quality of the investment climate (including questions such as "How often are you visited by the authorities and how much management time is spent with them?" "What fraction of your turnover do you spend on bribes?"); and second, data on the functioning of various services that are key elements of empowerment, such as health, education, and social protection. For both datasets going beyond inputs and outputs and asking how processes actually work and how people are involved is fundamental. In recent years, we have begun to accumulate data on both these important areas. As in the past, we will still need our third major type of micro dataset, information on households and individuals, if we want to assess outcomes.

These three datasets will provide powerful research tools for analyzing development processes. They are far richer and deeper than the cross-country macro datasets that, driven by the aggregate approach to growth, have formed the basis of so many cross-country regression analyses. For example, systematic investigation of the investment climate using firm-level surveys started at the World Bank and the EBRD during the 1990s has gathered pace in the last year or two. We now have comparable surveys completed, under way, or planned for the next year in 30 countries that cover large, random samples of firms, including some 1,200 firms in India and 1,500 in China.

These surveys collect the usual firm information on sales, outputs, inputs, and costs, but they also include specific quantitative questions about the investment climate. The results can be striking. For India we can now compare the investment climate at the state level, for instance, firms in Uttar Pradesh reported twice as many visits from officials as firms in Maharashtra and twice as many firms in Maharashtra than in Uttar Pradesh owned power generating facilities. In Uttar Pradesh, the largest and one of the poorest states in India, the electric power utility is less reliable than in Maharashtra, which explains why—despite the higher capital intensity—the growth rate is lower than in Maharashtra. These findings are strong and easily communicated and can have a powerful impact on policymakers.

The empirical information on empowerment is no less rich, although to date is somewhat less structured. The 2004 World Development Report, entitled Making Services Work for Poor People, will analyze much of the evidence, drawing on a large number of examples in areas ranging from basic services to economywide issues, such as community involvement in school management in El Salvador and Nicaragua; citizen report cards on services in Ukraine; legal and judicial reform and property rights in slum areas in Ecuador, Guatemala, Peru, and Venezuela; and corruption surveys in Albania, Georgia, and Latvia.

The challenge is to generate more structured data on empowerment so that we can carry out further systematic empirical investigation. The research agenda should be based on both empirical surveys and analytical understanding of the dynamics of the investment climate, of preferences, and of political reform.

Investment Climate

Small beginnings can lead to substantial accelerations of growth. Some changes can generate strong and sustained increases in growth rates and large transformations in such major sectors as education and health. The following are three analytical examples of accelerating and self-reinforcing change in the investment climate:

• Models that involve the diffusion of ideas or knowledge. Griliches demonstrated that the adoption of new technologies like hybrid corn was not a single event, but a series of developments that occurred at different rates across geographical space (Zvi Griliches, Hybrid Corn: An Exploration in the Economics of Technological Change, 1957).

• Models that generate multiple equilibria (Kevin Murphy, Andrei Shleifer, and Robert Vishny, Industrialization and the Big Push, 1989). A relatively small disturbance (institutional change) improves the investment climate, boosts investment or the productivity of capital, and initiates a movement from a lower-level to a higher-level equilibrium.

• Political economy models that focus on crucial historical turning points (as analyzed in Daron Acemoglu, Simon Johnson, and James A. Robinson, Reversal of Fortune: Geography and Institutions in the Making of the Modern World Income Distribution, 2001).

Preferences

Much of economics takes preferences as given and works with a simple model of an optimizing individual who understands the relevant constraints on decisionmaking. This has been a fruitful approach, especially for public economics, that relates social welfare to individual welfare. Increases in the latter can then be associated with a bundle of commodities that the individual prefers. However, some philosophers, including Mill and Keynes, who emphasized objectives in terms of expanded freedoms, also recognized that preferences adapt through development and discussion, and as the following example illustrates, development is in large measure about fundamental changes in behavior driven by shifts in preferences.

Suppose the Pakistani government recognizes the importance of increasing the enrollment of girls in school. How should it go about convincing a girl’s parents, particularly her father, to allow her to attend school? The standard economists’ answer, which would mention merit goods and externalities, would be to subsidize girls’ attendance by eliminating school fees, or even by giving the family a small grant.

However, we could go about expanding attendance in at least two other ways, each supported by a different model of the world. The first of these would take preferences as given, but would recognize that people making decisions do not always have all the relevant information. The policy response to this lack of information would be to provide the parents with good data about the advantages of schooling. These data could be used to point out, for example, that more educated girls are healthier as adults and that their own children will be healthier and better educated, and the father will allow his daughter to attend school even without any change in his preferences. That is, while his objective was always to ensure a good life for his daughter and for his grandchildren, he now understands more clearly that education will increase the chances that will happen.

By contrast, a second policy approach recognizes that preferences are not immutable. Part of the policy challenge is to persuade fathers to think differently about the kind of life they would like their daughters to lead. Note that this preference transformation can happen for reasons other than direct efforts to change preferences and can also result from either of the two policy interventions described earlier. If a subsidy or better information gets the girl into school, then her father’s preferences may subsequently change once he sees how valuable education can be for the life of his daughter. Of course, as with most real-world issues, the problem of girls’ education in Pakistan has a number of other important dimensions. The point is, preference change is a widespread issue in development and one that bears much further study.

Understanding changes in preferences will require new modeling approaches. The following three examples are intimately linked with empowerment and central to the development process:

• Most aspects of empowering people involve changing attitudes and behaviors.

• Education changes values and preferences in ways that are likely to be hard to predict, yet students and parents make decisions about education in the knowledge—or indeed the hope—that preferences are likely to be changed;

• Values can change in response to examples set by leaders. Many would argue that individual behavior in Russia deteriorated in the 1990s in part because of the example of large-scale looting of the state by individuals and groups close to the sources of political power.

Thus a move toward more systematic analysis of changing preferences is hard to avoid.

Political Reform

Further research seems warranted in three related areas: the political economy of growth, the effective participation of poor people in the growth process, and the role of building constituencies for change in order to advance economic and social reforms.

If economic growth creates large inequalities within a society—as opposed to shared growth—it can undermine the potential for future development. Winners can use their influence with the government to "remove the ladder," that is, to limit competition from other groups in society. In their paper mentioned earlier, Acemoglu, Robinson, and Johnson compare North and South America. Initial features of the political, economic, and physical landscape led to much greater inequality in Latin America, while in North America an improvement in the investment climate favored the establishment of more small and medium firms, thereby creating a constituency for change.

The challenge of ensuring the inclusion and empowerment of poor people is, in large measure, one of assuring the availability, and delivery of basic services as analyzed in the World Development Report 2004. In 1513 Machiavelli already saw that "one change leaves the way open for the introduction of others." But he also warned, referring to the relationship between prospects for change and constituencies for change: "There is nothing more difficult to plan, more doubtful of success, nor more dangerous to manage, than the creation of a new system. For the initiator has the enmity of all who would profit by the preservation of the old institutions and merely lukewarm defenders in those who would gain by the new ones." Events in Russia in the 1990s demonstrate Machiavelli’s points.

Accordingly, research should investigate how to carry out reforms within some given rules of the political game and what institutions should be created to minimize political opposition to innovation and adaptability, that is, research should find the best rules of the political game at every given stage of development.

Research and Action

Research and action must reinforce each other. Changing institutions and governance is at the heart of the agenda for action. To translate knowledge into action to promote change means

• Collecting evidence and information and carrying out rigorous analysis of programs and approaches.

• Establishing institutions and building capacity for taking knowledge forward. For example, a World Bank project to build a ring road around Shanghai also helped create local institutions that were then able to apply the techniques for construction, tendering, and finance across the country.

• Encouraging political action by "policy entrepreneurs." Like markets for private goods, markets for policy and institutional innovation are responding to well-chosen combinations of inputs.

• Entrusting the media to goad policymakers to change. Both the World Development Report, 2002: Building Institutions for Markets, and the World Bank Institute’s recent study, The Right to Tell: The Role of Mass Media in Economic Development, include a host of examples of the media’s role in catalyzing change, not only of policies, but also of institutions.

Many more examples of how one can trigger institutional change are available. We are beginning to go beyond an understanding of the importance of institutions to an analysis of how to change them, but we have a long way to go. If we researchers want to move forward on a sound basis, then we must have rigorous evidence and well-founded ideas. As academics, it is our responsibility to provide them.

The author is chief economist and senior vice president of the World Bank. This article is based on his presentations at a workshop on Measuring Empowerment: Cross Disciplinary Perspectives in February 2003 and his keynote speech at the World Bank Economic Forum in April 2003. j