The South African government intends to develop the Mzimvubu Water Project (MWP) which includes the construction of two dams (the Ntabelanga and Laleni dams) in the Tsitsa River, Eastern Cape province, South Africa. This investment is believed to be important to unlock the economic potential of this rural, poor and underdeveloped area. We consider a range of variables to ascertain what the realistic economic lifespan of each of the dams are, subject to uncertainty and complexity, as the benefits from the development of the dams are closely linked to such lifespans. The selected sites are prone to soil erosion thereby potentially jeopardising the derived benefits from the investment. We therefore develop a catchment-wide system dynamics model, incorporating both technical and behavioural dimensions, in an effort to determine which factors influence the economic lifespan of the dams and to what extent. The results suggest that, without taking any restorative or behavioural effects into consideration, the lifespan of the Ntabelanga and Laleni dams are 55–68 years and 26–33 years, respectively. Introducing business-as-usual type behavioural patterns could reduce the dams’ lifespans by between 35% and 44% due to the increased erosion. The increased soil erosion and resulting sedimentation is the direct result of increased human and animal movements due to the economic investment and associated land use change. Restorative and desirable behavioural action that will mitigate these impacts have the potential to reduce this loss to only 4%–9%. This would imply investing heavily both in restoration and in capacity-building and community development programmes to facilitate desirable behavioural change. In sum, this suggests that investment in social and natural capital has to coincide with the investment of financial capital in manufactured capital to make economic development last. Investment in social and natural capital could therefore be viewed as a type of insurance policy, and thus an insurance investment, against the risk of very plausible losses of the benefits derived from the manufactured capital (the dams). Alternatively, the opportunity cost of not investing in social and natural capital, is the much reduced lifespans of the dams and all associated economic benefits.