Sebastián Jaén Isaac Dyner

Proceedings of the 008 Winter Simulation Conference S. J. Mason, R. R. Hill, L. Mönch, O. Rose, T. Jefferson, J. W. Fowler eds. CRIMINAL CYCLES IN THE ILLEGAL DRUG INDUSTRY: A SYSTEM DYNAMICS APPROACH APPLIED TO COLOMBIA Sebastián Jaén Isaac Dyner Cra 80 No 65 3, M8307 School for Systems, Centre for Complexity CeiBa Universidad Nacional de Colombia Medellín, Colombia ABSTRACT The coca farming dynamics seem to partially follow the patterns of a long war against organized crime in Colombian. Since the early 80s, the cocaine market in the US and the rest of the world has been mainly supplied by Colombian cartels. Consequently, these illegal organizations have been targeted by Colombian and American law enforcement agencies. Our work argues that such policy has had a counterintuitive effect which has contributed to increases in coca farming and reductions in cocaine prices. This paper hypothesizes that such situation was the consequence of the way that drug cartels were dismantled thus the Colombian paradox. The consequences of the war against cartels may be assessed with the support of a dynamic theoretical framework and SD. This research assesses alternative policy for dismantling drug cartels. INTRODUCTION The history of the world cocaine market, especially in American, has been tightly associated with the most renowned Colombian cartels during the last two decades. In the most recent years this association has been mainly with the guerrillas and paramilitary forces. Most scholars in the field have focused their research almost exclusively on business costs, some on the associated business risk, and a small number on industry structure. Abundant literature addresses the American and Sicilian mafia cases (e.g., Rossemberg 973, Reuter 983, Fiorentini and Peltzman 997, and Fiorentini 999) but the literature on the Latin American organized crime and specifically on Colombia is still limited. Up until 993, Peru and Bolivia were the main cocaleaf providers worldwide while Colombia was just a marginal producer (Vargas 004). The Colombian cocaine boom started during the early 90s and developed complex organizations which, in many cases, operated autonomously (Thoumi 006). It has been acknowledged that the Colombian drug trafficking activity coincided with an accelerated demandrise in the US, which started an unusual process of accumulation of enormous amounts of money in the hands of the few involved in this activity (LópezRestrepo y CamachoGuizado 003). Two big cartels were conformed (Medellin and Cali) and many other small ones initiated operation across Colombian, Central America and Mexico (Thoumi 006). The growth and expansion of these criminal groups coincide with increases in homicide rates in Colombia. Continuous increases in violence and a terrorist threat shocked Colombians who become aware of their long connivance with the fast enrichment of illegal actors. Society and government reacted and got support from the US in terms of military training and technology (Sauloy and Bonniec 994). In 9993 most outstanding members of the Medellín cartel were in jail or dead. The Colombian government then went ahead to fight the Cali cartel. In the mid 90s the two main members of the Cali cartel were taken to prison but immediately the Norte del Valle cartel incorporated the Cali cartel s routes and social networks (Vicepresidencia 006). The Cali cartel played an important role as mediator between guerrillas and paramilitaries (Duncan 006). Once this cartel was dismantled, small traffickers were coopted by paramilitary groups who by 996 had already controlled many remote and central regions of Colombia. The association that was established between traffickers and paramilitaries increase their possibilities for bribing and buying firearms but created frictions among themselves for market control. The guerrilla forces got also involved in the cocaine business. Given the rural nature of these groups, it is not an oddity that the new cartels focused on cocaleaf production, gaining considerably knowledge to improve both crops and profits. The war against cocaine has been complex problem given the multidimensional characteristics that this involves. Some triumphs have soon turned loose battles, co 9784447086/08/$5.00 008 IEEE 49

Jaén and Dyner operation has not always been consistent and policy has been erratic at times. There have been unexpected consequences in this war. These emerge as a consequence of the complexities involved, given the nonlinear interactions between the main issues of concern. We thus developed a model for better understanding this criminal activity, specifically aiming to assess alternative policy for dismantling cartels. The paper is organized as follows: section two shows the evolution of coca crops and it argues that growth might be partially explained by the policy of dismantling big cartels, the cocaine supply chain and the role played by organized crime. Section three shows the model that has been build to evaluate the dismantling policy that was applied as well as alternative dismantling strategies and their consequences. Section four discusses validation issues and conclusions. BACKGROUD. The Evolution of Coca Farming Colombia has experienced two outbreaks of coca farming during the last twenty years. Figure, depicts the series of cocaleaf farming by the State Secretary (INCSR). The first outbreak starts in the early 80s were followed by heavy increases of coca production during almost ten years. After five years of apparent stagnation (99995), in the mid 90s, coca farming reached high levels until eradication policies reverted this tendency in 000. The stagnation period does not seem to be explained by any special eradication policy during those years as indicated in Figure. An event that coincides with the stagnation period, perhaps the only one directly linked to illicit drug trade, is the dismantling of the Medellin and Cali cartels. What happened during the period that the Medellin and Cali cartels were being dismantled (99995)? As expected: stagnation, moderate fall, and initial slow recovery in coca farming. But following the dismantling of the Cali cartel (soon after the disappearance of the Medellin Cartel) something that seemed unexpected happened: a new outbreak occurred. An important number of Colombian scholars in the field (e.g. LopezRestrepo and CamachoGuizado 003; Vargas 004; Thoumi 006; Duncan 006) argue that the cartel dismantling policy gave way to a new generation of criminals. They indicate that the fall of the cartels promoted the surge of many small gangs, which were poor and atomized compared with big cartels, as well as vulnerable to police forces and other traffickers. What Colombian and foreign scholars do not acknowledge is that such a breakup into smaller divisions is the result of a cartel dismantling policy and the approach used to dismantle such cartels. In other words, the fight against big criminal structures have contributed to split these into a large number of smaller units, bringing more people into this illegal industry given that such policy have destroyed the costly entry barriers to this unlawful business. 80.000 60.000 40.000 0.000 00.000 80.000 60.000 40.000 0.000.983.987.99.995.999.003.98.985.989.993.997.00.005 Años Figure : Coca bush hectares in Colombia. Source: INCSR (US 008) and Rocha 997. Almost all Colombian scholars in the field (e.g. LopezRestrepo and CamachoGuizado 003; Vargas 004, Thoumi 006; Duncan 006) describe the cartel dismantling policy as a starting point for a new criminal generation. They indicate that the fall of the cartels gave way to the surge of many small gangs, which were poor compared with the big cartels, vulnerable to police forces and other traffickers and atomized. What Colombian and foreign scholars do not acknowledge is that such break into bits is the consequence of a cartel dismantling policy and the way that such cartels were dismantled. In other words, the fight against big criminal structures have contributed to split these into a number of smaller units, bringing more people into this illegal industry given that such policy have destroyed the costly enter barriers to this illegal business. It is not obvious to establish the causality between fewer cartels or a concentrate industry and more coca crops. However, would be a theory that might light the possibility of such causality. Figure, depicts the Buchanan s Defense of organized crime theory. The horizontal axis represents the resources devoted to law enforcement while the vertical axis the resources used by criminal activity. C curve, draws the criminal response. Z, is the initial equilibrium. Under a criminal monopoly, Cm represents the supply curve (criminal response) and Zm a new equilibrium. Buchanan 973, states: Monopoly in the sale of ordinary goods and services is socially inefficient because it restricts output or supply. The monopolist uses restriction as the means to increase market price which, in turn, provides a possible source of monopoly profit. This elementary argument provides the foundation 430

for collective or governmental efforts to enforce competition. If monopoly in the supply of goods is socially undesirable, monopoly in the supply of bads should be socially desirable, precisely because of the output restriction. Figure : Criminal monopolies equilibrium. Source: Buchanan 973. Figure 3: Drug price and supply after a cartel dismantling policy. Jaén and Dyner As law enforcement impedes monopolization and concentration, fragmentation and competence emerges. Effect on market, after a cartel dismantling policies resembles a proportional tax effect on the supply curve such as Figure 3 shows. Under monopoly the initial relationship between price and quantity is P M and Q M (Figure 3). A cartel dismantling policy would change the curve s slope to a less inclined trajectory, where the new relationship between price and quantity is now P F and Q F (F denotes a fragmented industry). It is, more quantity by less price.. A Concentrated Industry During the early days (about 980s) the Colombian cocaine industry was characterized by its high degree of concentration and integration (Krauthausen 998, Thoumi 006). Coca base was imported from Bolivia or Peru, processed in remote laboratories an exported mainly to the US. Those activities were very expensive as there was a need to supplying remote laboratories with the coca leaves and then transport the cocaine powder by fast boats and airplanes to clandestine and legal airstrips and docks in the US. During the period (8095), the main power position in the market became the route owner, a trafficker which controlled the logistic route to export cocaine from an isolated laboratory to the market place. The cocaine market was and still is a free market with relative low barriers to enter. Many illegal businessmen might buy coca base in Bolivia or Peru and resell it in Colombia. Many others could process coca into cocaine powder as outsourcing business; nevertheless, very few of them carry it from Colombia to US defying the huge risk associated to such venture (Krauthausen 998). The real vocation of the infamous big Colombian traffickers until 995 was having, controlling and managing their routes. The route owner is on a main position: He coordinates the network and decides who his partners are he is the main money winner (Krauthausen 998). There was a natural network monopoly where the route owner has the privilege to decide who go with him into the illegal business. Routes are fiercely owned and collected by traffickers who have power to do it. A market with several route owners would guarantee the presence of negative externalities whether any neihbor were under law enforcement push. In addition no many traffickers can manage risk associated to smuggling cocaine across several countries. The best route is the safest one. There exist indeed several economic barriers in order to become a big route owner. During the 80s and 90s coca prices declined: It is difficult to account for the continued sharp decline until 989 and the more modest declines in the 990s (Reuter 00). Figure 4, depicts high wholesale prices during the late 80s and early 90s, the blossom years of cartels, compared with slightly lower average prices during the mid and late 90s, when cartels were being dismantled. 43

Jaén and Dyner Figure 4: Cocaine wholesale prices in the USA. Source Washington Office on Latin America (004). As cartels were broken into many small operational units along the productiondistribution chain, the new units incorporate agents coming from the former cartels as well as newcomers into the business. Thus, it is possible to infer that the dismantling policy brought about an interesting dynamic to the industry. This issue is discussed ahead and it is incorporated into our model structure. 3 MODEL 3. Criminal Firm's Basic Structure The head of an illegal firm has under his command several functional areas and a large number of members. There is a well defined barrier between members and heads. The Heads of the firm establish monopolistic and economic barriers (identified as the variable promotion barriers in the causalloop diagram, Figure 5) which obstruct members to easily reach the Heads level. The monopolistic barrier works by force. Economic barriers influence the Capacity to hire, which establishes a well defined hierarchical structure. Cartel resources come from the initial Heads contribution; but as any firm, the very dynamic of reinforcing sources must come from the business itself. The model assumes that the rate by which cartel resources increase (Fraction profit) depends on the firm size (Figure 6). Figure 7, indicates that illegal firms outsource some services (i.e. coca production, drug supply to endcustomer and information services). As previously discussed monopolies (cartels) have no incentives for growing very big given the risks of information filtering or the possibility of loosing operational control (Krauthausen 998). Our model suggests that the criminal firm incorporates basic structures to conduct illegal actions. Competition between small firms (those that render outsourcing services) brings criminal price down, benefiting cartels. ected Capacity to Hire R Promotion Barriers B Promoted Heads Illegal Activities Firm's Resources Figure 5: The criminal firm s basic causal loop. FIrm_capacity Ideal_members_per_capo _to_hire Time_under_oversight _hired Capacity_to_hire Posibility_to_promote _promoted Max_profit_to_reach Fraction_profit Initial_capital Economic_barriers Monopoly_barriers Heads Figure 6: Cartel Capacity and Profit. 3. Simulation basecase scenario Firm_incom Firm_resourc Firm_loses The simulation of the basecase scenario shows the evolution of the firm s initial configuration. Figure 8, draws a situation in which the Firm s heads start raising (i), hiring enough members (ii). When the firm reaches a determined size, it increases the use of outsource services (iii). Small suppliers grow as they are still profitable with low entry barriers. Firm s operations profitability (iv) continuously grow at diminishing returns, given the firm position in the market. As Outsourcing Services increases coca crops farming increases as well (v). In general, growth is initially accelerated and then only increases at diminishing return rates, as is the case with monopolies. 43

Hiring expected rate R Outsourcing Services Capacity to Hire Small firms B R Promotion Barriers B Promoted Competition for supplying criminal services Heads Illegal Activities Firm's Resources Cost of criminal services Figure 7: Causal loop of illegal workforce and criminal services. Jaén and Dyner total impunity until the dismantling policy is implemented (this operates as a switch function in period 5). To model such policy, law enforcement initially succeeds by eroding the firm s members and later striking bosses after some delay. Figure 9, shows two trajectories: Number one () indicates the basecase scenario and number two () indicates the dismantling policy. There are reductions in members and heads of cartels. Cartel resources fall as well as firm lacks its operative capacity. However, many of the low level criminal services (outsource services) take charge at lower costs, due to competition. As law enforcement persists members and heads disappear as well as profitability ( iv). A very interesting issue is that as outsourcing increases (iii) new entrants emerge due to the dismantling policy and hiring requirements. As the dismantling policy prevails, avoiding recruitment, workforce migrates to smaller gangs that act as outsourcing firms of criminal services. Firm's Heads 9 6 3 0 0 0 0 30 40 50 60 (i) Firm's 50 00 50 0 0 0 0 30 40 50 60 (ii) Outsourcing Services 40 30 0 0 0 0 0 0 30 40 50 60 (iii) Firm's Profitability Reached 0,5 0,4 0,3 0, 0, 0 0 0 30 40 50 60 (iv) Figure 8: Simulation base: Firm s initial configuration 3.3 Policy against Medellin cartel The Medellin cartel was the first big Colombian criminal organization to be dismantled. A bloody and long war against such group lasted almost ten years (98493) until the Colombian and American law enforcement agencies killed or imprisoned an important number of the cartel members. Our model replicates this policy using a probability of capture which triggers exit flows for both criminal categories: Heads and. The model assumes Coca hectares 40.000 30.000 0.000 0.000 0 0 0 0 30 40 50 60 (v) Figure 9: Cartel dismantling policy against Medellín cartel. 3.3 Policy against the Cali cartel The new small firms fiercely rival among themselves to gain power as observed after the disappearance of the Medellin and Cali cartels. The attack on the Cali cartel was apparently successful as by 995 their main head members 433

were imprisoned (Duncan 006; Thoumi 00). This was not a long and bloody war against this criminal group, simply the heads of the cartel disappeared (only heads) as they were taken to prison. Figure 0, illustrates the policy s impact: (i) heads fall, and (ii) members fall (had to look elsewhere for jobs). thus migrate away or establish their own firms. Such outbreak of new small firms renders criminal services at a very low cost. Jaén and Dyner Services level register lower numbers that the basecase and firm s profit shows a sharp fall. However after period 30, Heads increase at an important rate, now supported on outsource services, obtaining limited profits but increasing cocaleaf plantation (Figure ). Figure : Alternative policy against a cartel Next section discusses initial model evaluation and validation. Figure 0: Cartel dismantling policy against Cali cartel Thoumi 00, describes the members migration from big cartels (the most integrated) to small firms small cartels or cartelitos (Duncan 006). Simulations show similar migration patters (Figure 0 (iii)). 3.4 Alternative policies against cartels We now simulate an alternative policy to the one that was actually implemented. This consists on dismantling members (leaving the heads in place) this opposes the strategy implemented against the Cali cartel which mainly dismantled the heads of the organization. Figure, shows three main trajectories. The first one is the simulation of basecase scenario. In the second one, law enforcement is applied against firm s members and heads, and the third one, consists on an exploratory policy of law enforcement on members only. Simulation results (until period 30) seem to reveal a promising behavior: Firm members decrease, Outsourcing 4 CONSIDERATIONS AND CONCLUSIONS The purpose of the model has been to explore the dynamics of criminal organizations and to assess alternative policy, aiming to reduce their impact on society. As observed, the model largely reproduces the developed theoretical framework. However, is there a statistical evidence for this? Figures and 3, show simulation results and reported coca plantations (hectares) during two important periods of the war against cartels. Both model results and real data show similar patterns: constant increase at the beginning followed by a period of stagnation. Discrepancies still exist. Analysing the model during the war against the Cali cartel, we observe some basic coincidences (Figure 3). Both model and real data have some similar patterns: constant increase at the beginning and few periods of stagnation followed by several years of accelerated increases. Although there is not enough validation evidence the model seems promising. Our preliminary prototype provides useful insights for a more elaborated model. How 434

Jaén and Dyner ever we will first assess if more model detail provides sufficient gains in learning about policy and its effect on the system. Coca Bush Hectares 50.000 40.000 30.000 0.000 0.000 0.980.98.984.986.988.990.99.994 Period (i) Coca_actual coca_simulate Figure : Simulated and actual coca farming hectares during the war against the Medellin cartel. Coca Bush Hectares 50.000 00.000 50.000 0.988.990.99.994.996.998.000 Period (i) Coca_actual coca_simula Figure 3: Simulated and actual farming hectares during the war against the Cali cartel. Previous simulation results help us assessing Buchanan s theory with respect to the impact that a cartel dismantling policy has on cocaine trafficking. Simulations show that a cartel dismantling policy may bring about counter intuitive results as it may induce higher levels of criminal activity at a lower cost. Simulations show that a policy that focuses on parts of the criminal productiondistribution chain usually induces side effects (higher levels of criminality in the long run). Although a criminal antitrust strategy shows unexpected and undesired consequences, policy against heads seems to be the most harmful strategy to confront the criminal monopoly. Nonetheless, simulations reveal that it is more beneficial, at least in the short run, to focus on member heads. Such policy reveals to be the most expensive one. This however might be improved if complemented by declaring a war against lowkey members of cartels as soon as heads are either in jail or dead. This research confirms theoretical issues and learns from alternative policy in the war against cartels. There is still much work ahead. As this is on going research, we will next consider side effects and reinforcing issues regarding homicide and inmate increases and their effects on communities. REFERENCES Buchanan, James M. 973. A Defense of Organized Crime? in Rottenberg, Simon (ed.), The Economics of Crime and Punishment. American Enterprise Institute for Public Policy Research. Blumstein, A. 00. Crime modeling. Operations Research 50(): 64. Duncan, G. 006a. Historia de una subordinación: Cómo los guerreros sometieron a los narcotraficantes? Revista Foro. 57: 467. (In Spanish). Duncan, G. 006B. Los señores de la guerra: de paramilitares, mafiosos y autodefensas en Colombia. Editorial Planeta Colombia S.A. Third edition. Bogotá, D.C. (In spanish). Fiorentini, G., y Peltzman, S. 995. The Economics of Organized Crime. Cambridge University Press. ISBN 0 5 4748. Fiorentini, G. 999. Organized Crime and Illegal Markets. Dipartimento di Scienze Economiche. Università di Bologna. Frontline. 006. Interview to Juan David Ochoa. Available via <http://www.pbs.org/wgbh/pages/frontline/shows/drug s/interviews/ochoajdo.html> [accessed August 8, 008] (In Spanish) Krauthausen, C. 998. Padrinos y mercaderes: Crimen organizado en Italia y Colombia. Editorial Planeta. Bogotá, D.C. 475 p. LópezRestrepo, A., CamachoGuizado, A. 003. From smugglers to druglords to traquetos : Changes in the Colombian illicit drug organizations. Canadian Journal of Latin American and Caribbean Studies. 8(55 56): 4976. Reuter, P. 983. Disorganized Crime: The Economics of the Visible Hand. London, MIT, 33 p. En: Fiorentini, G. 999. Organized Crime and Illegal Markets, Dipartimento di Scienze Economiche. Università di Bologna. Reuter, P. 00. The Need for Dynamic Models of DrugsMarkets. Bulletin on Narcotics. LIII ( & ). Available via <http://www.unodc.org/unodc/en/dataand analysis/bulletin/bulletin_000 0 page003.html http://www.unodc.org/unodc/en/bulletin/bulletin_ 435

0000 page003.html> [accessed August 8, 008] Rocha, R., 997. Aspectos económicos de las drogas ilegales en Colombia, Dirección Nacional de Estupefacientes, PNDU y Planeta. In: Thoumi, F. Drogas ilícitas en Colombia: su impacto económico, político y social. Rottemberg, S. 973. The Economics of Crime and Punishment. American Enterprise Institute for Public Policy Research. 3 p. Rubin, H. Paul. 973. The Economic Theory of the Criminal Firm. In Rottenberg, Simon (ed.), The Economics of Crime and Punishment. American Enterprise Institute for Public Policy Research. P. 5578. Sauloy, M., Bonniec, Y.L. 994 A quién beneficia la cocaína? TM Editores. ISBN: 958605068. (In Spanish). Thoumi, F. 006. From drug lords to war lords: the development of the illegal drug industry and the unintended consequences of antidrug policies in Colombia, Centro de Estudios y Observatorio de Drogas y DelitoCEODDFacultad de Economía, Universidad del Rosario. UNODC. 006. World Drug Report. United Nations. Available via <http://www.unodc.org/unodc/en/dataandanalysis/wdr006.html> [accessed August 8, 008] AUTHOR BIOGRAPHIES SEBASTIAN JAEN is an PhD candidate in Systems Engineering at the National University of Colombia. He received a master s degree in System Engineering and a BS in Management Engineering at the same university. His current research is about crime and narcotrafic, using simulation models. He has publicated some works in Operational Research, mainly in areas such as bounded rationality and multiobjective optimization. His email address is <jsjaenp@unal.edu.co>. ISAAC DYNER is a Titular Professor specialising in energy policy at the Faculty of Mines, National University of Colombia, Medellin, and PhD in Science Decisions. He holds advanced degrees from the Universities of Warwick and Southampton, and from the London Business School, and has published a number of papers using system dynamics to analyze policy, strategic behavior, conservation and trading in the energy sector. His email address is <idyner@unal.edu.co>. Jaén and Dyner UNODC. 007. World Drug Report. United Nations. Available via <http://www.unodc.org/unodc/en/dataandanalysis/wdr007.html > [accessed August 8, 008] US. State Secretary. 008. International Narcotics Control Strategy Report. Available via <http://www.state.gov/p/inl/rls/nrcrpt/008/ > [accessed August 8, 008] Vargas, R. 004. Drogas, conflicto armado y seguridad global en Colombia. Nueva sociedad (9): 73. Available via: <http://www.nuso.org/upload/articulos/3_.pdf > [accessed August 8, 008] (In Spanish) Vicepresidencia. Colombia. 006. Dinámica reciente de la violencia en el Norte del Valle. Observatorio del Programa Presidencial de Derechos Humanos y DIH. Available via <http://www.derechoshumanos.gov.co/observatorio/04 _publicaciones/nortedelvalle.pdf> [accessed August 8, 008] (In Spanish) Washington Office on Latin America (WOLA). 004. Cocaine wholesale prices in the USA. Available via < http://www.wola.org/index.php?option=com_content &task=viewp&id=397&itemid=8 > [accessed August 8, 008] 436