Intervention (Hidden)
We will conduct an individually randomized control trial in Oaxaca, Mexico. Not only one of the largest States in Mexico, but its more diverse, Oaxaca encompasses several indigenous ethnic groups, including the Mixtecs, Zapotecs and Mixe and a majority of municipalities that retain traditional indigenous governance institutions, called usos y costumbres. Oaxaca is also a historic center for the production of cochineal, a domesticated insect that was produced by indigenous communities and was Mexico’s most valuable processed export during the colonial period, second only to gold and silver. Micro-climatic differences in suitability for cochineal provides exogenous variation to historic market exposure among indigenous communities that we will stratify upon in our study. The experiment involves visiting close to 4,000 voting age Mexicans in 66 municipalities of Oaxaca, explaining that we are surveying people on their opinions and we may also ask them to participate in some games. Upon providing signed informed consent, we will allocate them to 5 main arms.
The objective of the arms is to understand both, how perceptions about corruption and government economic underperformance as well as whether nation-building activities—both through priming shared identity and providing shared exposure to the economy—may affect stated preferences for democracy, civic participation, and mutual trust. Of particular interest is how these effects differ among historically marginalized indigenous groups and those with different histories of exposure and assimilation, as well as by their political sympathies.
Two videos of corruption: We will randomly expose different to videos showing official information about how prevalent and costly to the pockets of citizens certain kinds of corruption are, using data from INEGI, while also showing them footage of corruption scandals of major politicians in Mexico in the last 10 years. One video involves politicians of the current ruling party, and a second video of those in the opposition. We hypothesize that compared to those in the control who were not already aware of the videos, these effects will diminish support, particularly for the culpable group, but also for the democratic system and for trust more broadly. We will test for differential effects among ex ante supporters of these parties (due to motivated reasoning/ cognitive dissonance or greater informational content). In particular, we anticipate that while many people think the opposition parties are corrupt, the incumbent party –Morena-- is viewed as much less so, so there could be more updating with the Incumbent video.
Economic performance: A third video summarizes some of the (unfortunately mostly negative) economic results for Mexico in recent years (economic growth, poverty, and the handling of the COVID19 crisis, compared to other Latin American countries). The aim of this video is to test whether retrospective information about economic performance – i.e. failures in outcomes— have different effects than corruption—i.e. failures in processes. We again hypothesize that since performance has been relatively poor, these effects will diminish support for incumbent party, for the democratic system, and mutual trust. We will again test for differential effects among ex ante supporters of the ruling and opposition parties.
Common identity video: we have designed (with the help of a professional creative designer) a 3 minute motivational video that primes individuals to recall many aspects of Mexican history and culture that forge its common identity (including food, music and soccer), and including a common past with indigenous people, and how it took centuries to have a democracy with power alternating parties. It also discusses the importance of getting involved to shape a common future, and the importance of protecting democracy. For a small sub-sample, we combine the video with a narrative exercise on why one should vote and participate in civic life (following Broockman and Kalla Science 2016). Here we expect heterogeneous treatment effects by indigenous identity to be particularly important, with indigenous people experiencing a broader sense of community with other Mexicans as a result of the video, reflected in more trust. A modification we will attempt if feasible is to also develop subsets of the broader video more tailored to priming the indigenous communities’ contributions to Mexico and another tailored to mainstream Mexican identity.
Stocks: sharing the future (cross randomized). After these treatments, we will give a subset of people in the urban sample the opportunity to trade a financial asset that tracks the Mexican stock exchange (200 pesos). Each week they will be able to invest or divest 25 pesos. We will use by one of the authors in the Israeli context in Jha and Shayo “Valuing Peace: The Effects of Financial Market Exposure on Votes and Political Attitudes” (2019, ECMA). Unlike those studies, which recruited individuals participating in an online panel in a richer country specifically for a financial study, in this study we will be recruiting door-to-door for a broader study. We anticipate low compliance and thus will focus on treatment effects on the treated. In addition to testing for the outcomes and heterogeneous treatment effects above, we will also test for financial literacy effect differences by gender.
Control group: this group will receive no information/videos about corruption and no stock assets.
Stratification: In the urban sample we will stratify in electoral precincts. In the rural sample, we will sample communities that were historic cochineal producers because they were located in the optimal micro-climate for cochineal prior to 1870 (following Diaz-Cayeros and Jha 2016 and Diaz-Cayeros, Espinosa-Balbuena and Jha 2021), compared with otherwise similar nearby municipios that had sub-optimal micro-climates and did not produce cochineal. As Diaz-Cayeros and Jha show, cochineal towns were exposed more historically to global markets and were more likely to develop entrepreneurial cultures. We will test for differences both on average between cochineal and non-cochineal municipalities in their baseline levels of the outcomes above, but also for heterogeneous treatment effects of the arms above. In the urban sample, we also examine whether the effects differ for migrants from rural municipalities.
Implementation Notes:
1) We will present results anticipating differential effects among those who ex ante are already aware of the informational content of the information videos.
2) Given the face-to-face nature of the intervention, and the possible temptation to the surveying team of the cash involved in some of the games, we will present results controlling for surveyor fixed effects, and exclude observations when the surveyor has demonstrated failure to follow the research protocol.
3) We will also use precinct fixed effects, and control for demographics and prior opinions about corruption and democracy to increase power.
4) To correct for potential imbalances ex post, if necessary, we will use double-lasso procedures as suggested by Belloni, Chernozhukov and Hansen 2014 and Duflo et al. 2020.
5) Some of the outcomes are measured on a 4 or 5 points Likert scale. We will either convert these into dichotomous variables by grouping answers, or use ordered probit/logit models to use the whole range of variation.