If those who discover new ideas could appropriate all the benefits, then, at least in principle, market mechanisms could efficiently determine investments in science. But private and collective incentives diverge in the presence of externalities. We just do not know in advance where, when, or for whom research results will become valuable. Because predicting or charging for such applications can be difficult, markets tend to underallocate and misallocate support for basic research. This grant funds research by a team led by Jason Owen-Smith to examine the return to investments in basic science by tracking how research grants eventually do and do not result in gainful applications. To collect, process, and study the detailed data necessary for carrying this out, Owen-Smith and his colleagues will establish an Institute for Research on Innovation and Science (IRIS) based at the University of Michigan.  Foundation funds will support data infrastructure at the University of Michigan, as well as infrastructure at the University of Chicago and Ohio State.