In the long-term, starting from a non-uniform spatial distribution of assets, people
in better off districts are largely net contributors to social infrastructure projects,
while people in poorer districts are mostly receivers of net transfers ([26]). However,
in the medium term and assuming budget-constrained maximisation of average
access rates, marginal improvements in access rates (fi
′(Ei)) will reflect the above
patterns, i.e. fw
′(Ew) < fb
′(Eb) for healthcare as long as no sufficient infrastructure
network is in place, and fw
′(Ew) > fb
′(Eb) for primary schools. With the marginal
impact of spending in poorer districts being relatively lower for healthcare, and
higher for schools, public authorities will target points on the production possibility
frontiers (Fig. 1) to the left of the 45 ◦ line (where Sw = Sb) in the former sector,
and to the right in the latter (transformation curves are elongated in the direction
of respective returns on investment).