| Summary: | The Survey of Well-being via Instant and Frequent Tracking (SWIFT) program was created in 2014 to produce poverty statistics cost-effectively, timely, and in a user-friendly manner. Under the SWIFT program, poverty rates are estimated by (i) training a poverty rate projection model on a previous household budget survey, (ii) collecting data in the field on identified poverty correlates, and (iii) applying the model to collected data to produce poverty projections. The SWIFT program has expanded since 2014 and now includes a wide variety of activities, namely, aiding in estimating official poverty statistics and monitoring poverty outcomes of lending operations and investment projects in over 50 countries. One key advantage of the SWIFT methodology is the significantly shorter interview time compared to the surveys used in traditional poverty data collection. In 2015, the SWIFT program prepared a guideline for using the SWIFT methodology to ensure the quality and precision of poverty estimation under the program. The guideline includes a set of recommendations on data collection and poverty projection. This report documents the performance evaluation and possible modifications of the SWIFT poverty projection methods since 2015. It critically evaluates the SWIFT poverty projection method proposed by the 2015 guideline, identifies limitations, and introduces several remedies - SWIFT Plus and SWIFT 2.0 - with their evaluations. This paper is organized as follows: Section 2 reviews the literature, Section 3 describes the SWIFT methodology, Section 4 discusses challenges for SWIFT and the solutions, Section 5 explores future research topics, and Section 6 concludes
|
|---|