The harmony of programs package: Quasi-experimental evidence on deworming and canteen interventions in rural Senegal

This paper uses a unique and large-scale quasi-experimental data to
study the effect of deworming and school meals programs as a package on
educational outcomes (pupils' test scores: aggregate, French or math;
enrollment, promotion or dropout rates) in rural Senegal. We extend the
endogenous selection model à la Heckman to incorporate a double-index
selection mechanism. We also generalize the Roy model accordingly. We
develop estimation strategies based on the full information maximum
likelihood and the two-step method. We derive a wide and rich collection
of treatment effects ranging from exclusive to relative effects
including sequential and substitution effects. The results show that the
combination of deworming and school meals programs is more beneficial to
pupils' achievements than taking programs separately. The sequence of
implementation does matter. The two programs are complementary in
increasing scores and promotion rates. However, they are substitutes in
reducing dropouts. The cost-effectiveness analysis shows the deworming
program is by far cheaper than the meals intervention. Implementing
meals program before deworming is more cost-effective than the reverse.
Lastly, unlike the deworming, meals program and the package (deworming
and meals) have a welfare-enhancing effect on households.

Key words: Deworming programs; school meals programs; double-index selection;
complementarity vs. substitutability; educational outcomes; quasi-experiment; social welfare; Africa; Senegal