|Title:||Data and development: synergies between social science and astronomy|
|Speaker:||Dr Dieter von Fintel (Department of Economics, Stellenbosch University)|
|Date:||Thursday, 14 September 2017|
|Time:||11:00 - 12:00|
Since the advent of the World Bank’s Living Standards Measurement Surveys, the study of household socio-economic conditions has expanded dramatically into the developing world. Micro level data allow researchers to understand how households escape poverty, and which interventions can be put in place to improve the livelihoods of individuals. The work of Nobel laureate, Angus Deaton, attests to this massive emphasis towards understanding the micro-level effects of policies and household-specific drivers of development. These data have also assisted in setting clear Sustainable Development Goals, and monitoring their effectiveness. Understanding “development”, however, is more multi-faceted than only focusing on incomes and Gross Domestic Product – Amartya Sen pioneered the capability approach, which emphasizes that people should not only have access to financial resources, but to education, health and food security. This list could easily be extended to other (sometimes intangible) characteristics such as mental health, subjective happiness, crime-free environments and safe climatic conditions. Household surveys have become the work horse for measuring many of the micro-level facets of socioeconomic development. However, many low income countries still do not enumerate regular surveys, and even in better resourced economies, household surveys contain substantial biases. These include misreporting, operation outside of the experimental ideal, and missing important components of economic development. Morton Jerven laments this as Africa’s “statistical tragedy”. Not only are the data messy, but development economists operate outside laboratory settings. Economists and other social scientists are therefore more concerned about selection bias, omitted variables and endogeneity than researchers from many other disciplines. Longitudinal household data can – in some circumstances – assist in circumventing these problems. However, this type of data is rare in most African countries. To assist in modelling dynamics and solving statistical biases, economists are increasingly turning to data sources more commonly used in other disciplines – in particularly remote-sensed satellite data. Night lights data and vegetation coverage are known to correlate well with a host of development outcomes. Instead of following households, we follow the regions in which they reside at regular time intervals. Not only is data less noisily measured, but longitudinal estimates approach unbiased results. Economists are, however, open to learn about other approaches to measure economic development more effectively.