The Poverty Stoplight: A New Metric For Microfinance
The Poverty Stoplight is a tool that has been implemented in Paraguay since 2010 to measure poverty. It is a self-diagnostic visual survey to assist poor families to assess their level of poverty across the 50 indicators and to develop personalized poverty elimination plans. The tool uses stoplight colors (red, yellow, and green), illustrations, maps, electronic tablets, and simple software to create dashboards and indexes. Although it can be used in a wide variety of settings, it was created in order to fill a gap that exists among poverty measurement tools that are used by the microfinance industry. Most of these tools are focused on monetary poverty, and only one uses a constructivist approach to understand poverty. Despite trends in academic literature to consider poverty a multidimensional phenomenon and to measure poverty through hybrid positivist and constructivist methods, the Poverty Stoplight is the only tool used by the microfinance industry that attempts to accomplish this. The aim of this dissertation is to contribute to the academic literature by analyzing the practical benefits and difficulties that measuring multidimensional poverty through a combination of epistemological paradigms entails. To do so, in this dissertation I evaluate the robustness of a specific implementation of these two trends: the metric aspect of the Poverty Stoplight. In order to do this, I seek to answer four research questions: is the Poverty Stoplight (1) reliable, (2) valid, (3) practical and (4) does it have discriminatory power"u2014from a positivist and constructivist point of view. My analysis is based on data I collected through four methods: (1) application of the visual survey tool, (2) focus groups, (3) semi-structured interviews, and a (4) participatory wealth ranking. While results suggest that there is test-retest reliability, consequential validity, content validity and criterion-related validity, problems related to generalizability compromise internal consistency reliability and construct validity. Taken as a whole, the Poverty Stoplight has limited robustness. I end this dissertation with recommendations to make it a more robust tool, such as separating the Poverty Stoplight metric from the coaching methodology or reformulating indicators and dimensions in order for these to better represent poverty.