Persistence in the Political Economy of Conflict: The Case of the Afghan Drug Industry

Links between licit and illicit economies fuel conflict in countries mired in irregular warfare. We argue that in Afghanistan, cultivating poppy and trading drugs bring stability to farmers who face the unintended consequences of haphazard development efforts while lacking alternative livelihoods and security necessary to access markets. Drug trafficking funds the crime-insurgency nexus and government corruption, in turn foiling attempts to establish a unified governance body. We show how individual rationality, market forces, corruption and opium stocks accumulated at different stages in the supply chain counteract the effects of poppy eradication. To that end, we use initial results from a multiagent model of the Afghan drug industry. We define physical, administrative, social and infrastructural environments in the simulation, and outline objectives and inputs for decision making and the structure of actor interactions. Statement of the Problem We start by casting the post-2001 insurgency in Afghanistan as a complex adaptive system (CAS) and outline a countryscale, multiagent model that captures the persistence of the Afghan drug industry when outside forces intervene and its resilience against government policy perturbations. Post-2001 Conflict in Afghanistan as a CAS The current insurgency in Afghanistan, like that in 1979– 1989 against the Soviet occupation, is rooted in resistance to external intervention, unlike conflicts before the Soviet invasion that generally stemmed from internal ethnic and economic divisions. The country is populated with rural and urban households, a medley of opportunist armed groups and criminal gangs led by local strongmen, the government and International Security Assistance Force (ISAF), and insurgents. Adaptive and “strategic” interactions among actors that have determined the emergent course of the current conflict are informed by individual attitudes, kinship considerations and tribal affiliations on multiple scales through weakly-coupled, nested systems enabled by the remoteness of population settlements, lack of transportation and communication infrastructure and scarcity of manpower. Copyright © 2010, Association for the Advancement of Artificial Intelligence (www.aaai.org). All rights reserved. Questions We have built a multiagent model of the Afghan drug industry that can answer the following policy, operational and analytic questions inspired by (Flynn, Pottinger, and Batchelor 2010) and (Hartley 2009): Policy How important is drug trade to the livelihood of the Afghan population? How can we decrease drug trade in Afghanistan? How can we decrease its significance to the population? Operations How many men and for how long are needed to eradicate poppy fields, interdict drug shipments in a given province and prevent poppy cultivation from shifting to neighboring or other provinces? How do wemanage moneyfor-work programs such that they do not disappear due to corruption and end up fueling organized crime and insurgency? Which roads and how much will have to be improved so farmers in a given district can think of alternative crops? Analysis How much bribes and protection money is gathered from drug traders frequenting a given route? By whom? Addressing these questions requires modeling links between farmers’ livelihood and the quality of governance. It also constrains the scalability, flexibility and taxonomy of entities and environments in the model. Moreover, since many players in the Afghan drug industry cannot be modeled as fully rational, understanding the short-term dynamics of the industry that are shaped by boundedly rational agents is key to answering our questions; hence a multiagent model (Axtell 2000). Objectives We use our multiagent model of the Afghan drug industry to show that engaging in poppy cultivation and drug trade provides stability to farmers who lack alternative livelihoods and security necessary to access markets and face the unintended consequences of haphazard development efforts. In particular, we will demonstrate the following empirical patterns that highlight the resilience of the Afghan drug industry to external intervention: 86 Complex Adaptive Systems —Resilience, Robustness, and Evolvability: Papers from the AAAI Fall Symposium (FS-10-03)