On generating adequate counterfactuals for national policy evaluations. A response to Mundt et al.

Abstract

We appreciate the opportunity to respond to the comments from Mundt, Contreras & Ibanez [1] regarding our paper on cannabis legalization among school students in Uruguay [2]. In their letter, they raise several concerns regarding the use of Chilean secondary school students’ data as a comparison group. We believe that their concerns are, in fact, the main argument in favor of using Chile as comparator. First, the authors mention that Chile had the highest past-year prevalence of cannabis use in secondary school students in the Americas, which is true [3]. Uruguay is the country with the second highest prevalence of cannabis use among secondary students in South America [3]. In addition, the difference between both countries in the years prior to legalization was negligible (see Figure 1). Source: Rivera-Aguirre A et al. [2] Secondly, the authors mention that the large increase in past-year cannabis use from 2009 to 2015 in Chile could affect the comparison and, ultimately, the validity of our results [4]. Uruguay, however, also experienced an increase in past-year cannabis use in this same period, which peaked in 2016 [5]. Further, the key threat to the validity of a difference-in-difference analysis such as ours would be non-parallel pre-legalization trends in the two countries. Instead, past-year and past-month cannabis use followed similar pre-legalization trends in the two countries [2]. Thirdly, Mundt et al. also argue that Chile's 2005 cannabis decriminalization is another reason why Chile is not a good comparison group for Uruguay. We argue that this, in fact, makes Chile a better comparison group, because in Uruguay, cannabis use was already decriminalized [6]. In other words, decriminalization in Chile would resemble more closely what would have happened in Uruguay in the absence of legalization. Finally, in our effort to produce reliable estimates of the legalization policy in Uruguay, we recognize that we worked with imperfect counterfactuals and with limited opportunities to isolate the mechanisms that may or may not affect population levels of cannabis use. We agree with Mundt, Contreras & Ibanez [1] that increases in use (in Chile and in Uruguay) may be the result of changes in perceptions and social norms towards cannabis, probably influenced by decades of debate and more recent changes in policy in the Americas and beyond. In fact, we noted this in the Discussion section of our paper, as one of our findings was a transitory increase in cannabis use in Uruguay in 2014, prior to plausible changes in cannabis availability through legal sources. In conclusion, the information currently available supports our conclusions that allowing and implementing legal access to cannabis in a highly regulated market, as is the case of Uruguay [7], did not produce short-term changes in use among adolescents. This work was supported by the National Institute on Drug Abuse, R01DA040924-01 (Cerdá). ARA, AC and RQ received funding from ANID-Millennium Science Initiative Program, NCS2021003. None. Ariadne Rivera-Aguirre: Conceptualization; investigation. Alvaro Castillo-Carniglia: Conceptualization; investigation. Hannah S. Laqueur: Conceptualization; investigation. Kara E. Rudolph: Conceptualization; investigation. Silvia S. Martins: Conceptualization; investigation. Jessica Ramírez: Conceptualization; investigation. Rosario Queirolo: Conceptualization; investigation. Magdalena Cerdá: Conceptualization; funding acquisition; investigation.