AIM : Crop variable inputs (CVI's) are critical to successful crops. So we here ask: "What are the marginal returns to crop variable inputs?" And explore whether observed CVI levels maximise economic returns to farmers. We compare results to national aggregates in India. DATA: Analysed Farm Business Survey 2004-2012, where crop gross margins and input spending are available, for conventional winter wheat and oilseed in England and Wales. RESULTS: Marginal spending on variable inputs (e.g. seed, fertiliser, crop protection) returns in economic product significantly less than GBP£1 per marginal pound spent. Therefore, expenditure allocation on those inputs could be quite far from economic optima. However marginal physical products (yields) are positive, but small, and significantly different from zero. These conclusions hold across a wide range of alternative economic models and subsets of the data. The same conclusions are observed, in estimations for Indian grain production, and for maize in China where lower national rates of fertiliser application appear optimal. DISCUSSION: Unknowns, including yield, quality and price, make it difficult to optimise ex ante input levels. Tied advice could reduce the efficiency in the farm sector - owing to possible perverse incentives. And the preferences of farmers, may be to avoid risk, or to maximise yields. Farmers may also be biased - relative to full information and perfect competition. All of which might distort prices from the neoclassical equilibria with perfect information and perfect competition. Thus, one could ask "How useful are the prices seen in practice, for allocation in the context of the farmer behaviour reported here?"