Reproducibility package for Rural Labor and Long Recall Loss

Lack of availability of administrative records on informal and seasonal labor participation dictates reliance on self-reported measures in many rural settings in developing countries. A common approach relies on annual recall to capture labor activities. In a panel of rural households in Malawi, this study uses a survey experiment to test the effect of annual recall on reported labor supply relative to quarterly interviews using ninety-day recall windows. Annual recall reduces reported labor participation and months worked by 20%. These reductions are greater for activities further in the past and when labor supply is reported by proxy. The profile of households’ primary respondents, predominantly male and older, can induce meaningful biases against women and youth. The study additionally shows that recall window choice can significantly alter estimates of the relationship between labor supply and education. These findings are especially relevant for impact evaluations and studies with constrained sample sizes.