Relative to spatial diffusion effects, shorter-range diffusion seems to predominate
for primary schools (Table 2: lnsch(sp-min) vs. lnsch(sp)). Relative to healthcareinstead, the cut-off distance near an average connectivity level yields spatial parameter
estimates with relatively higher statistical significance (lnhca(sp) vs. lnhca(spmin)).
Regardless of values of the eligibility dummy, worse off districts are found to
register additional gains in access over the medium term, with results being nearly
equivalent for the two sectors. By contrast, the numbers effect turns out to be either
irrelevant, as a criterion for primary schools (with a negative parameter sign for the
variable Nsch), or not to play a definite relevant role, as highlighted by mixed evidence
in terms of statistical significance of the positive parameters associated with
the variable Nhca (Table 2).8
If regression results based on a 4