One of the interesting things about gathering more data would be to compare traditional versus P3 models on similar projects, to see which ones perform better over the long term in terms of costs and completion rates and so on.
An infrastructure bank like this could do that. It's very much skewed toward the P3 side, but it could also offer traditional financing at very low rates. It would be interesting to see in the real world—not retrospectively—how those projects perform against each another.