Can the two new aid-growth models be replicated?

Recent aid effectiveness literature centers on two competing models from the family of conditional models: The Good Policy Model, where the key feature is policy times aid, and the Medicine Model, where it is aid squared. Both models were reached on a sample of 1/3 of the available data. The models are simplified to be replicatable on more of the data. Within-sample the Good Policy Model proves fragile, while the Medicine Model is more robust. Both models fail in out-of-sample replications. A semi-parametric technique is used to test for an unknown functional form of the aid-growth term. It rejects that aid is statistically significant.