The relationship between research activity and economic well-being are examined within the U.S. manufacturing and agricultural sectors for the periods 1950--1982 and 1964--1986. State crop, livestock, and aggregate agricultural total factor productivity (TFP) data, 1950--1982, are examined for evidence of convergence to a single TFP level (sigma-convergence) or to a steady state rate of growth (beta-or conditional convergence). Empirical results do not support sigma-convergence but do support beta-convergence. The rate of beta-convergence is variable and depends on research and development (R&D) spillins from other states, private R&D, and farmers' schooling. U.S. aggregate manufacturing and agricultural data, 1964-1986, are examined for evidence that R&D spillovers do exist. The empirical results support the notion that research activity in one sector does impact resource allocation decisions in another sector. Own-sector R&D stocks are shown to have strong positive impacts on output supply and input demand decisions. Public agricultural research stocks were shown to have stronger impacts on the manufacturing sector's output supply and input demand decisions than the private manufacturing R&D stocks on the agricultural sector's decisions. Private manufacturing R&D and public agricultural research was shown to be complements. The finding of strong positive interstate and inter-sector spillover effects have policy implications. First, independent state planning of agricultural research is inefficient and cooperation across states boundaries including establishment of new political jurisdictions for financing public agricultural research can enhance efficiency. Second, public planning of agricultural research should capture all positive externalities (own-sector plus the externalities that exist outside of the sector).