It demonstrates that the adaptive agent framework can be used to extend traditional models of comparative advantage in international trade, showing that the presence of increasing returns to scale in some industries shifts the basis of comparative advantage arguments, making room for industrial policy and the regulation of trade.
Next, the dissertation demonstrates that the size distribution of cities within nations, generally thought to approximate the "Zipf" distribution, can be reproduced using a simple agent-based model. This model produces insights into the evolution of the distribution as well as departures from it especially in France and Russia. This understanding of urban dynamics has implications for easing the structural transition of the Russian economy and for designing policies to reduce the size of megacities in the developing world.
The dissertation goes on to examine individual level data from the Guatemalan civil war from an adaptive agent modeling perspective. It finds several novel patterns in the data which may serve as benchmarks for adaptive agent modeling efforts and suggests avenues by which existing conflict models might be brought into closer accord with the data.
The dissertation concludes that adaptive agent modeling is useful in a policy context because it allows quantitative work to be done while relaxing some of the unrealistic assumptions which are often required to gain analytical traction using traditional methods. The method is found to be particularly useful in situations where path dependence, heterogeneity of actors, bounded rationality, and imperfect information are significant features of the system under examination. The individual based nature of the method is also found to be well suited to assessing distributional impacts of changes in process or policy.