Diversification in Indian agriculture towards high value crops: Multilevel determinants and policy implications

•This paper partitions the variation in area share of high value crops at different geographical levels to assess their contextual effects over and above the compositional effects.•Along with household levels differences, contextual effects of higher geographical levels are important in shaping the geographical pattern of diversification.•The contextual effects differ across crop categories and farm classes.•Ignoring one or more geographical levels results in inaccurate estimates and erroneous conclusions. Employing a multilevel model, this paper demonstrates the importance of contextual effects, over and above the compositional effects, in shaping the geographical pattern of agricultural diversification towards high value crops. The results reveal that, besides household-level differences, the contextual effects of higher geographical levels, especially states and villages, explain significant variation in land-use under high value crops, but these do not apply to all types of crops and farm classes in a similar manner. Further, the findings also ascertain that ignoring one or more geographical or administrative levels may yield biased estimates, leading to erroneous conclusions and an incomplete understanding of the geography of agricultural diversification. From a policy perspective, these evidences provide an important feedback for the correct targeting of diversification strategies, their coordination and monitoring at different geographical or administrative levels for efficient, sustainable and inclusive growth of agriculture.