Basically, the whole thing is a bit more complicated. However, it's useful to require pre-registration and sharing of data etc for studies you plan to fund. I

Basically, the whole thing is a bit more complicated. However, it's useful to require pre-registration and sharing of data etc for studies you plan to fund. Ideally, you give researchers an in-principle approval, but then hold the funds in escrow and only disburse them after they published their data etc. (Financial markets can provide the bridge financing between in-principle approval and actual disbursal.) There's many journals that have open data requirements, but more often than not they are flouted. The above suggestion would give them real teeth.