In some US states, policy makers, pressed by local and regional industrial interests, are debating how to "reform" technology transfer at public universities. "Reform" in this context is generally understood to mean redirecting university technology transfer activities to increase the benefits of state-funded research to local industries. Progress towards this goal is often constrained by federally mandated laws applicable to technology transfer at universities (such as the Bayh-Dole Act) and by university policies that have been placed by state legislatures outside the purview of policy making state officials. Calls for change have also been countered by the view of many universities that the system is not broken. Suggested reforms range from the abolition of the Bayh-Dole Act, which gives universities the flexibility to transfer ownership of federally funded inventions to local industries, to structural or management changes in universities that will promote innovation and/or expedite the licensing of new ideas. This article proposes a new paradigm: instead of measuring the success of technology transfer by counting numbers of patents and licensing deals, the authors suggest measuring knowledge flows between state universities and their localities. This approach should produce a more accurate picture of the full impact of universities on their regions. (Contains 2 tables and 15 notes.)