Synthetic panel methods

To end extreme poverty, policy makers need insights into the dynamics of poverty – whether those currently in poverty are likely to be stuck there over extended durations, whether they are able to climb out of poverty, or whether those no longer living in poverty are vulnerable to falling back into it. Typically we rely on panel data that can follow individuals and household over time to shed light on these questions. But such data is rarely available due to cost and complexity.

DEEP’s programme of work on synthetic panels provides a valuable solution. Rather than relying on panel data, we use cross-sectional data that is readily available in most low- and middle-income countries to generate pseudo panels. The synthetic panel methods involve prediction techniques to estimate likely trends and changes in household or individual income or consumption levels over time. The methods have been found to generate useful insights in settings where panel data is not feasible and/or is prohibitively expensive.