A machine learning approach to find the determinants of Peruvian coca illegal crops

Débora Belén Cipriano Romero, Yadira Gina Melo Estrella, María Isabel Zambrano Laureano, Rubén Ángel Ruiz Parejas, Jimmy Alberth Deza Quispe

Producción científica: Artículo CientíficoArtículo originalrevisión exhaustiva


The current study analyzed the determinants of the Peruvian coca illegal plantations in the period 2003-2019. Hence, the DEVIDA database variables were gathered at first. Then, a machine learning-based technique is employed to select the most relevant variables for the study. That technique, Lasso, selected as accurate variables eradication of coca plantations and pasta base. Both OLS and VAR are employed to analyze the relevance of the selected variables. OLS finds that eradication was negatively related to the dependent variable. Nonetheless, pb confiscation had a positive relationship with illegal coca crops. Furthermore, VAR encounters that only pb confiscation affected the dependent variable. Supplementary tests are carried to ensure the accuracy of the results. In consequence, it is concluded that eradication policies by themselves were not enough to discourage the coca plantations. Farmers should get instruction about alternative crops and financial help. Furthermore, it has been claimed that pb confiscation generates scarcity of the drug, which elevates its price. Thus, coca farmers are more motivated to plant coca because of the higher prices. Therefore, as long as the international demand, which is disposed to pay high prices, the coca illegal crops and its illicit products will exist.

Idioma originalInglés estadounidense
Páginas (desde-hasta)127-136
PublicaciónDecision Science Letters
EstadoIndizado - abr. 2022

Nota bibliográfica

Publisher Copyright:
© 2022 by the authors; licensee Growing Science, Canada.


Profundice en los temas de investigación de 'A machine learning approach to find the determinants of Peruvian coca illegal crops'. En conjunto forman una huella única.

Citar esto