Modeling Feedback between Economic and Biophysical Systems in Smallholder Agriculture in Kenya: The Crops, Livestock and Soils in Smallholder Economic Systems (CLASSES) Model

We investigate natural resource-based poverty traps using a system dynamics model of smallholder farms in highland Kenya. System dynamics modeling is well-suited to examining the complex interactions and feedback between farm-household economic decision making and long-term soil dynamics which may be at the source of persistent poverty among smallholders in this region. We examine the effects of changing initial endowments of land, labour and stocks of on-farm soil organic matter on the long-term welfare of these households. We find that larger farms are better able to cope with both labour shocks and deteriorating natural capital than smaller farms, with smaller farms remaining poor and unable to invest into more diversified agricultural activities, like livestock. This suggests locally increasing returns to various combinations of economic and biophysical assets. Information obtained through such simulation model experiments may lead to better targeting of poverty alleviation programs as well as suggest a broader array of strategies that play off of the complex interactions between economic and biophysical assets. The flexibility to examine these different leverage points is partially provided by our use of system dynamics, rather than other modeling techniques, to develop a descriptive rather than prescriptive model of farm behaviour.