3. The model and empirical evidence
We discuss the model and then comment on the empirical evidence,
addressing waste generation (section 3.1), landfill diversion drivers (section 3.2),
and investigate the relevance and nature of spatial phenomena (section 3.3).
Finally (section 3.4), we present a convergence analysis aimed at showing whether
Italy, although showing drastically different waste performance in the South,
might be characterised by the catch up of the poorer, less well performing
regions.
We estimate the empirical non-structural model by specifying the following
general panel-based reduced form that is usual in the Environmental Kuznets Curve
(EKC). The EKC is a non-theoretical approach (McKitrick 2007).
Given the nature of data, we opt for fixed effect LSDV (least square dummy
variable) estimations which account for individual fixed effects by including N-1
dummies (Hausman test outcomes are not shown). Linked to the comments above,
we tackle the potential endogenous nature of waste management factors
instrumenting the policy variable with its own first lag. The model is: