Titanium isopropoxide works wonders, if your substrate can tolerate it (and a lot can). Theres a paper by Seebach somewhere (Helv Chim Acta?) with the details. But if its neutral catalysts you want, you can't go past Oteras distannoxane catalysts, you can make them yourself, and they're active at really small loadings (at least 0.1 mol% from memory). They do contain two atoms of tin/molecule catalyst, but its only present in trace amounts.