The rise of large-scale farms in modern-day agricultural frontiers has frequently displaced and fragmented smallholder farming due to an unbalanced competition for land and water resources. A better understanding of land-use decision making of heterogeneous farming systems and farmer behavior is needed for tailoring policy instruments that help shaping these frontiers towards more efficient, equitable, and sustainable outcomes. Here we analyze a survey of 235 land users, ranging from large-scale agribusiness enterprises to small-scale subsistence-oriented farmers, across the Northern Argentine Dry Chaco. We use the survey data and auxiliary geospatial data in a Bayesian network to quantify the key determinants of gross farm revenues for the surveyed farms and simulate how the farm revenues respond to alternative policy scenarios. If farms are crop or cattle farms as well as the level of intensity of farming explained a large proportion of the variation in revenues for commercial and semi-subsistence farms. Revenues in commercial farms are mainly formed by the size of the farm and the level of capital input. In contrast, access to credit and land tenure are substantially more important in shaping gross revenues in the semi-subsistence farms. The simulation of a scenario with a hypothetical land titling program combined with better access to credits suggests that unlocking these barriers could generate higher revenues for semi-subsistence farmers without compromising the revenues that can be attained by the large fully commercialized farms. Our finding hence underpins claims that secure land titles can crucially benefit particularly the economically marginalized and smaller land users and, in that way, may also contribute to more inclusive rural development.