SMS Surveys of Selected Expenditures
- Author(s): Lang, Megan;
- Ligon, Ethan
- et al.
High-frequency measures of economic well-being can allow policymakers and researchers to understand and quickly respond to dynamic problems, but collecting such data is expensive. Can short message service (SMS) surveys enable researchers and policymakers to measure household welfare and firm performance at a high frequency in low-income countries? We detail the implementation of two SMS surveys and evaluate their efficacy for gathering high-frequency data. One measures consumption expenditures in Rwanda and the other measures microenterprise revenues in Uganda. We successfully calculate a measure of household welfare for households that respond to the SMS survey in Rwanda and track changes in revenues over time for microenterprises in Uganda. Our SMS surveys are substantially less costly than equivalent in-person surveys; however, nonresponse is a significant problem. We propose combining SMS surveys with in-person data collection to compute weights that correct for nonresponse bias, then evaluate the performance of our method using the revenues data from Uganda.