With few exceptions, induced innovation theories give little consideration to the role of distortions or externalities as determinants of the commodity or factor biases of innovations demanded by farmers. Nor has the theory devoted much attention to the influence of technical progress, with or without distortions, on the sectoral structure of production. This analysis identifies the demand for innovations as a function of a specific policy setting which conditions and is in turn conditioned by the sectoral structure of production. In this context, when some sectors contribute more than others to land degradation and soil erosion externalities, the capacity for divergence between privately optimal and welfare- maximizing allocations of research resources -calculated at market and shadow prices respectively - is substantial. In some circumstances it may be optimal to employ research budget allocations as second-best substitutes for Pigouvian taxes.