I'm a huge fan of using simulations to ground qualitative arguments. While the sims usually need to be fine-tuned so extensively as to leave them open to claims of 'overfitting', the benefit is that it nails the assumptions in your argument to the church door.
Abstract:
Patterns of political unification and fragmentation have crucial implications for comparative economic development. Diamond (1997) famously argued that “fractured land” was responsible for China's tendency toward political unification and Europe's protracted political fragmentation. We build a dynamic model with granular geographical information in terms of topographical features and the location of productive agricultural land to quantitatively gauge the effects of “fractured land” on state formation in Eurasia. We find that either topography or productive land alone is sufficient to account for China's recurring political unification and Europe's persistent political fragmentation. The existence of a core region of high land productivity in Northern China plays a central role in our simulations. We discuss how our results map into observed historical outcomes and assess how robust our findings are.
https://sci-hub.tw/https://www.nber.org/papers/w27774
