Illegal activity, such as deforestation for illicit crops for cocaine production, has been inferred as a cause of land change. Nonetheless, illicit activity is often overlooked or difficult to incorporate into causal inference models of land change. Evidence continues to build that narcotrafficking plays an important, yet often unreported, role in forest loss. This study presents a novel strategy to meet the challenge of estimating the causal effect of illicit activity in land change using consolidated news media reports to estimate the relationship between drug trafficking and accelerated forest loss in Central America. Drug trafficking organizations engage in illegal land transactions, money laundering, and territorial control that can manifest as forest conversion to agriculture or pasture land uses. Longitudinal data on 50 sub-national units over a period of 16 years (2001-2016) are used in fixed effects regressions to estimate the role of narcotrafficking in forest loss. Two narcotrafficking activity proxies were developed as explanatory variables of forest loss: i) an “official” proxy from drug seizures data within 14 sub-national units; and, ii) an “unofficial” proxy developed from georeferenced news media accounts of narcotrafficking events. The effect of narcotrafficking was systematically compared to the other well-known causes of forest loss, such as rural population growth and other conventional drivers. Both proxies indicate narcotrafficking is a statistically significant (p<0.01) contributor to forest loss in the region, particularly in Nicaragua (p<0.05, official proxy), Honduras (p<0.05, media proxy), and Guatemala (p<0.05, media proxy). Narcotrafficking variables explain an additional 5% (media proxy) and 9% (official proxy) of variance of forest loss not captured by conventional models. This study showed the ability of news media data to capture the signal of illicit activity in land use changes such as forest loss. The methods employed here could be used to estimate the causal effect of illicit activities in other land and environmental systems. Our results suggest that current drug policy, which concentrates drug trafficking in remote areas of very high cultural and environmental value, has helped to accelerate the loss of Central America's remaining forests.

非法活动，例如为生产可卡因的非法作物砍伐森林，被认为是土地变化的原因。然而，非法活动经常被忽视或难以纳入土地变化的因果推论模型。越来越多的证据表明，贩运毒品在森林流失中起着重要的作用，但往往没有报道。这项研究提出了一种新颖的策略，可以利用合并的新闻媒体报道来估计中美洲毒品贩运与森林加速流失之间的关系，以应对估计土地变化中非法活动的因果关系的挑战。贩毒组织从事非法土地交易，洗钱和领土控制，这些可能表现为森林向农业或牧场土地利用的转变。在固定效应回归中，使用了16年（2001-2016年）内50个国家以下单位的纵向数据，以估算麻醉品贩运在森林流失中的作用。开发了两种麻醉品贩运活动代理作为森林损失的解释变量：i）来自14个国家以下部门的缉获毒品数据的“官方”代理；ii）从地理参考新闻媒体关于贩运人口事件的报道中开发的“非官方”代理。系统地将麻醉品贩运的影响与其他众所周知的森林流失原因进行了比较，例如农村人口增长和其他常规驱动因素。这两个指标均表明，贩运毒品是造成该地区森林损失的统计显着因素（p <0.01），特别是在尼加拉瓜（p <0.05，官方代理），洪都拉斯（p <0.05，媒体代理）和危地马拉（p < 0.05，媒体代理）。麻醉品贩运变量解释了传统模型未捕获的森林损失方差的另外5％（媒体代理）和9％（官方代理）。这项研究表明新闻媒体数据能够捕获土地用途变化（例如森林流失）中非法活动的信号。这里采用的方法可用于估计其他土地和环境系统中非法活动的因果关系。我们的结果表明，当前的毒品政策将毒品贩运集中在具有很高文化和环境价值的偏远地区，这有助于加速中美洲剩余森林的流失。这项研究表明新闻媒体数据能够捕获土地用途变化（例如森林流失）中非法活动的信号。这里采用的方法可用于估计其他土地和环境系统中非法活动的因果关系。我们的结果表明，当前的毒品政策将毒品贩运集中在具有很高文化和环境价值的偏远地区，这有助于加速中美洲剩余森林的流失。这项研究表明新闻媒体数据能够捕获土地用途变化（例如森林流失）中非法活动的信号。这里采用的方法可用于估计其他土地和环境系统中非法活动的因果关系。我们的结果表明，目前的毒品政策将毒品贩运集中在具有很高文化和环境价值的偏远地区，这有助于加速中美洲剩余森林的流失。