Great questions. Well, the coauthors on the project, myself included, are doing a lot of the work, and we're not going to take salaries. We have four research assistants from my institution currently doing various things, including some of the replicating. Also, we started this project in 2017, paid RAs to simply do the computational robustness (check data and code to make sure we could replicate) for about six months of the AER, and then got busy with other projects. We pay the institutional wage, which is pretty low these days with the dollar strong. Probably I could have been more ambitious, but, yes, $7,000 would in fact cover roughly our RA expenses and some rewards for prediction. Note some of our RAs may also get coauthorship if they do well, plus letters if need be.
We haven't yet decided 100% on what form the poll & prediction market would take. However, my coauthors still have access to what they used for their Science article which we could alter, and, of course, I also run a prediction market (Insight -- we'd have to make some changes, and not sure it would be wise to use it for this project).