The 1993 SNA does not recognize the output of R&D as capital formation, despite the fact it is thought to be a major contributor to future economic growth. If the SNA were to be changed, should all expenditure on R&D, or only some, be recorded as capital formation? Can all the practical difficulties of deriving satisfactory estimates be overcome, for example by using expenditure data collected in accordance with the Frascati Manual, and obtaining appropriate deflators and service lives?