We discuss the search for optimal stocking regimes in semi-arid grazing lands. We argue that ‘one size does not fit all’ and that different stocking regimes are appropriate under different conditions. This paper is an attempt to move beyond polarization of the current debate towards a more integrative and flexible approach to grazing management. We propose five different conditions as major influences on grazing regimes: environmental variability and predictability; degradation and thresholds; property right regimes; discount rates; and market stability and prices. We suggest a lack of connection between the micro-economics literature and natural science and social-anthropological literature. It is timely to achieve greater integration around some key questions and hypotheses, and recognize that policy prescriptions at national or even regional levels are likely to have limited value due to context specificity.