This paper proposes a broad approach to the study of criminal behavior through time using states as the units of analysis. In doing so, it argues that when a specific crime is targeted in one state, a criminal displacement effect in neighboring states can be expected. This is illustrated by the case of direct state intervention to combat kidnapping in Mexico City. A multivariate GARCH model is estimated in order to incorporate the effects of such intervention into the behavior of the conditional correlation between the kidnapping cases of Mexico City and two neighboring states. Results show a reduction in the number of kidnapping cases in Mexico City following the implementation of the security policy, along with a crime displacement effect. The findings suggest the need to redesign local and federal security policies in order to incorporate coordination mechanisms between the different levels and entities involved in combating kidnapping.

本文提出了一种广泛的方法，以状态为分析单位，通过时间来研究犯罪行为。这样做时，它认为当一个特定的犯罪以一个州为目标时，可以预期在邻国产生犯罪转移效应。墨西哥城国家直接干预打击绑架的案例说明了这一点。估计多变量 GARCH 模型，以便将这种干预的影响纳入墨西哥城和两个相邻州的绑架案件之间的条件相关行为中。结果表明，在实施安全政策后，墨西哥城的绑架案件数量有所减少，同时还产生了犯罪转移效应。