The research priorities defined by the Population Dynamics and Economic Growth Agenda can be summarized as follows:
(1) Given the projected trends in fertility and mortality changes, what are the implications for economic growth and income distribution and the incidence of poverty?
(2) How does investment in reproductive health affect economic conditions at the household level, including the productivity, labor force participation and savings behavior of women, children and households?
(3) How do different types of investments in reproductive health affect the health of women and children? How does the type and organization of services affect their effectiveness, including ability to reach poor and vulnerable populations?
The Working Group identified four priorities for data collection, each of which would benefit from higher levels of investment:
- Cross-sectional household data: There is a need for improvement in current methods to collect current period and retrospective data through cross-sectional household data. The main area for improvement is in ensuring that, where possible, both demographic and economic information is collected. In particular, the Working Group suggested that the Demographic and Health Surveys supported by USAID could work toward inclusion of economic variables; and the Living Standards Measurement Surveys supported by the World Bank could refine their collection of health and health service use information.
- Panel Data Sets: Establishing one or more African sites for panel data collection – including demographic, health, fertility histories, education, labor market, migration and household structure variables – would constitute a resource for current and future researchers. Importantly, panel data collection should be designed to follow those who migrate out of the research area, because these individuals are likely to have different economic (and other) outcomes than those who remain.
- Sub-national Data Sets: Within-country analysis has the advantage of controlling for time-invariant fixed effects. Ideally, these data would include information about program exposure and program characteristics, population characteristics and demographic behavior, and economic conditions (labor force participation rates by sector, savings rates, income per capita, wage rates, by gender and age, and others).
- Random Assignment Evaluation Methods: Experience has shown that the quality of evidence that can be derived from a well designed evaluation can have more value to the policy and program design process than dozens of weakly designed retrospective evaluations that provide no way to estimate the true impact of a particular program.