Recent work on the relationship of productivity growth to expenditures on R&D is discussed with special reference to ""unfinished business"": unsolved conceptual and econometric problems. The standard ""R&D as a capital"" model raises a number of difficult conceptual issues, especially the meaning and measurement of the depreciation of ""knowledge"" and of the associated economic rents created by it. In addition, empirical implementations of this model have to struggle with the possibility of simultaneity and reverse causality and the implication of imperfect markets for knowledge-based products. An empirical example of some of these issues is presented using data from a panel of U.S. R&D -performing firms, showing both the difficulties alluded to above and evidence for the continued importance of the R&D process for the growth of productivity."