Should new research on under-fives reshape our approach to development?