Payin ’ it Backward: Migration and Democratic Diffusion in Latin America

1. Payin’ it Backward: Migration and Democratic Diffusion in Latin America David Crow División de EstudiosInternacionales Centro de Investigación y DocenciaEconómicas (Mexico City) david.crow@cide.edu Clarisa Pérez-Armendáriz PoliticalScienceDepartment Bates College cperezar@bates.edu

2. Migration and Democratic Diffusion: Theory Our Work on Mexico: • “Social remittances”  Transmission of values, attitudes, practices: here, migrants spread values from more democratic host countries to less democratic (Levitt, 1998) • Democratic Diffusion  democracy spreads across international borders (Brinks and Coppedge 2006, Przeworski et al. 2000, Starr 1991). BUT no theory of agency • Policy Diffusion  supplies theory of agency (functionaries, social elites, members of organized civil society, etc.) applied to policy change (Kapur and McHale 2005, Keck and Sikkink 1998, Tarrow 2005) New Theoretical Element: • Interpersonal Communication  people are more receptive to and put more stock in ideas they receive from people they know (e.g., “two-step flow” in which “primary reference groups” and friends and family, Lazarsfeld, Berelson, Katz, 1944, 1955)

3. Migration and Democratic Diffusion: The Case of Mexico • Three transmission “paths” • Return migration: Mexicans absorb values in host country (U.S.) and retain them back home  item: “Have you ever lived outside Mexico?” • Cross-border communication: Mexicans abroad communicate values to friends and family back home  item: “Do you have family members or friends who live outside Mexico?” • Information networks in high-volume sending communities:  item: CONAPO, Migration Intensity Index - Data: Original survey, DesencantoCiudadano , June, 2006 (N=650)

4. Attitudes and Behaviors • Attitudes: • Satisfaction with Democracy • Government Respect for Rights • Composite Tolerance Indicator (political, religious, sexual orientation) • Behaviors: • Non-electoral individual participation (contact authorities, handed out flyers, signed complaint, etc.) • Organizational participation (parties, neighborhood associations, civic organizations, etc.) • Protest

5. Findings • Migration increased democratic attitudes and behaviors, but effects are differentiated • Return migration • Increased tolerance • Negative evaluations of government respect for rights (“critical citizens”) • Friends and Family Abroad • Less satisfied with democracy • More individual and organizational participation, protest • High-intensity migration community • More organizational participation • No Effect for Remittances

6. Extending the Research to Latin America: Challenges • Different migratory destinations: Central and South American migration is much more varied than Mexican migration (e.g. Peruvians go to Argentina, Chile, Spain, and U.S., Nicaraguans go to Costa Rica and U.S., etc.) Do Nicaraguans in Costa Rica learn the same things as Peruvians in Spain and Mexicans in the U.S.?  Probably not BUT, migration generally occurs from less to more democratic countries  so, the general expectation is for some democratic learning

7. Extending the Research to Latin America: More Challenges • Different reasons for migrating: Mexican migration has been almost exclusively economic and familial, but Central and South American migration has also been political and related to natural disasters Do economic migrants, political migrants, family reunification migrants and refugees from natural disasters have the same predisposition toward political learning  Again, probably not

8. Extending the Research to Latin America: Yet More Challenges! • Different national contexts: Political institutions, histories, cultures, and economic development also influence political attitudes and behaviors Study on Mexico holds constant not only sending country context, but (given 95% in U.S.) host country context Ideally, the different national contexts could be measured and included in the model as variables (“turn proper names into variables”)  easier said than done

9. Data: Latin American Public Opinion Project (2008)

10. LAPOP 2008: Data Limitations • Return migration: poor measure • Time window for living abroad (5 yrs. ago) too narrow  only 1.3% of Latin Americans (and 0.71% of Mexicans) • No item on destination (i.e., host country) • Question only asked in 10 (of 22) countries • Family members’ host countries: little information  just coarse categorization (U.S. only, U.S. and other, other) • Communication: frequency, not content • Reasons for migrating: no information • Remittances/Communication: does receiving remittances count as communication?  measurement error for communication

11. LAPOP 2008: Dependent Variables • Evaluations of Democracy - Satisfaction with Democracy (Mucho, Algo, Poco, Nada) - Political system’s protection of “basic rights” (7-pt. scale, “Nada” to “Mucho”) • Political engagement - Convince others to vote for party or candidate (4-pt. “Never” to “Frequently”) - Meet with neighbors to solve community problem (4-pt. “Never” to “Frequently”) - Protest in last 12 months (3-pt. “Never”, “Almost Never” and “A Few Times”

12. LAPOP 2008: Independent Variables – Migration • Household members residing abroad • In U.S. only (binary) • U.S. and Elsewhere (binary) • Elsewhere (binary) • Frequency of communication (5-pt. ordinal “Never”, “Rarely”, “Once or Twice a Month”, “Once or Twice a Week”, “Every Day”) • Household receives remittances (binary)

13. Distribution of Family Members Living Abroad

14. Frequency of Communication with Family Members Abroad

15. Countries by % of Households that Receive Remittances

16. Countries by Dependence on Remittances Among Receiving Households % of Households (Among Receptors) that Depend “A Lot” or “Somewhat” on Remittances for Monthly Income

17. LAPOP 2008: Controls • Sociodemographic • Sex • Age • Income • Education • Employment • Retrospective economic evaluations • Pocketbook • Sociotropic

18. Expectations • Having family members abroad will • Make citizens more critical in their evaluations of their country’s democracy • Increase different forms of civic engagement • Differentiated effects among host country categories  US > US + Other > Other • Greater communication should also increase critical citizenship and civic engagement • Remittances • Increase criticism • Decrease engagement (following Goodman and Hiskey)remittances may substitute for goods capital, etc., one would normally get leveraging local networks, obviate need for community participation

19. Results: Attitudes

20. Results: Engagement

21. Findings • Household Members Abroad • Makes citizens more critical, increases engagement • BUT, order different than predicted; in most cases having household members in the U.S. and elsewhere has an effect of greater absolute magnitude • Frequency of Communication • Some direct effects, but important as suppressor variable i.e., effects of household members abroad become apparent (or sharper) only when comparing within each level of communication frequency.

22. More Findings • Remittances: • Insignificant for attitudes toward democracy • Positive for electoral persuasion • Negative for civic involvement and protest  Some evidence of remittance-induced alienation

23. Next Steps • What’s going on in the individual countries? • Better statistical model: random effects model not good enough for exploring heterogeneity between countries. • How do the effects of household members abroad, communication, etc., differ across countries  i.e., random slopes as well as intercepts • Explicitly include aggregate-level variables to account for differences in institutions, culture, history, etc. • Case studies • Effects differentiated by host country • Better theory: a more coherent account of why the the implantation of democratic values/behaviors should vary according to host country • Auxiliary data: more precise idea of who’s going where, possibly imputing host countries • Communication: frequency and content  do people really talk about politics? • Modify survey items • Qualitative interview data • Return migrants: ???