1. Centre d’Estudis Demogràfics Educational homogamy of Mexicans in Mexico and the USA: gender, generation, ethnicity and educational attainment * * * Albert Esteve and Robert McCaa Centre d’Estudis de Demogràfics (Barcelona) University of Minnesota Population Center aesteve@ced.uab.es rmccaa@umn.edu aesteve@ced.uab.es

2. Centre d’Estudis Demogràfics Outline 1. Theory: educational homogamy and ethnic/racial intermarriage » Homophily: like marry like » Increasing role of schooling in conjugal choices » Becker (1974): assymetry, female hypergamy » Oppenheimer (1988): economic independence of women leads to homogamous marriage markets 2. Educational assortative mating in Mexico: a point of departure 3. Data and methods 4. Models of educational homogamy and interaction parameters 5. Conclusions

3. Centre d’Estudis Demogràfics Theory: educational homogamy and ethnic/racial intermarriage 1. An expected increase in educational assortative mating » Homophily: like marry like » Increasing role of schooling in conjugal choices » Mating preferences: from gender assymmetry to gender symmetry 2. How educational assortative mating interplays with ethinc/origin assortative mating » Marital assimilation (intermarriage): » Least vs most educated » First vs later generations » Evidence suggests that Mexicans seems to follow the classical pattern of assimilation

4. Centre d’Estudis Demogràfics Educational assortative mating in Mexico: a point of departure (data from IPUMS-International) (Esteve and McCaa, Latin American Research Review, forthcoming) 1. Context 1970 - 2000: » Strong expansion of schooling in Mexico » Progressive reduction in differences by gender in educational attainment 2. Educational assortative mating in Mexico » Homogamous unions have diminished » Homogamy remains the basic rule at all levels » Increases among the most educated » Remains strong among the least educated » Female hypergamy is nearing its end

5. Centre d’Estudis Demogràfics Data and Methods 1. 5% IPUMS-USA census microdata samples 1980, 1990, 2000 2. We selected unions in which: » Both of the spouses belong to one of these categories: Mexican Born, Mexican Origin, Non Hispanic White » At least one of the spouses a. is aged 30-39 at the time of the census b. resided in the US before the age of 20 3. Years of schooling completed: < 9, 9-11, 12, 13-15, 16+ 4. Loglinear models

6. Centre d’Estudis Demogràfics N HusbandsWives 1980 1990 2000 198019902000 Mexican born 4,515 10,957 24,814 2,900 7,491 20,918 Mexican origin 15,591 20,053 18,675 17,001 22,694 22,563 Non-Hispanic White613,294688,987586,231613,500689,812 586,238 Table 1. USA co-resident spouses, at least one age 30-39 resident in USA before age 20 % < 9 years of school (of all years of schooling) % < 9 years of school (of all years of schooling) HusbandsWives HusbandsWives 1980 1990 2000 1980 1990 2000 Mexican born54.549.940.246.442.638.5 Mexican origin22.1 6.7 4.321.8 7.6 4.8 Non-Hispanic White 5.9 1.9 1.2 3.5 1.2 0.8 % >=16 years of school % >=16 years of school HusbandsWives HusbandsWives 1980 1990 20001980 1990 2000 1980 1990 2000 1980 1990 2000 Mexican born 5.7 4.8 4.2 3.1 3.4 3.3 Mexican origin10.512.114.5 4.9 8.914.2 Non-Hispanic White28.429.831.318.324.531.4

7. Centre d’Estudis Demogràfics Table 2 (abbreviated). Marital unions by year, ethnicity/origin of the spouses and by type of educational assortative mating % Mexican Born - Mexican Born Mexican Origin - Mexican Origin Non Hispanic White - Non Hispanic White 198019902000198019902000198019902000 Homogamy57,852,948,142,746,947,448,551,053,2 Female Hypogamy20,025,526,523,424,127,619,321,525,6 Female Hypergamy22,221,725,3 33,929,025,0 32,227,521,3

8. Centre d’Estudis Demogràfics Table 3: Likelihood-ratio Chi-Square Statistics for selected models of educational homogamy and ethnicity/origin Basic ModelsG2G2 dfBIC 1.YEAR*ETH*EDH + YEAR*ETW*EDW + YEAR*ETH*EDW + YEAR*ETW*EDH 2963087,25401983121,0 2.1 + ETH*ETW + EDH*EDW 11496,45203478,7 Diagonal Models 3.2 + YEAR*D et + YEAR*D ed 8229,8504527,2 4.2 + YEAR*D et + EDH*D et + EDW*D et + YEAR*D ed + ETH*D ed + ETW*Ded 2877,2468-3850,8 5.2 + YEAR*D et + EDH*EDW*D et + YEAR*D ed + ETH*ETW*D ed 1621,1460,0-5076,3 6.2 + YEAR*EDH*EDW*D et + YEAR*ETH*ETW*D ed 1004,8408-4921,4 Symmetry Models 7.2 + YEAR*S et + YEAR*S ed 8292,4503570,4 8.2 + YEAR*S et + EDH*EDW*S et + YEAR*S ed + ETH*ETW*S ed 2049,8458-4639,0 9.8 + YEAR*A ed + ETH*ETW*A ed 1530,8446-4971,7 Note: YEAR = census year, ETH = ethnic group husband, ETW = ethnic group wife, EDH = education husband, EDW = education wife, D et = diagonal specification ethnicity, D ed = diagonal specification education, S et = symmetry specification ethnicity, S ed = symmetry specification education, A ed = asymmetry specification education.

9. Centre d’Estudis Demogràfics Table 4: Log-odds ratios for educational homogamy by year and ethnicity

10. Centre d’Estudis Demogràfics Table 4: Log-odds ratios for educational homogamy by year and ethnicity (2000) -3 -2 0 1 2 3 4 5 Schooling +- HomogamyHomogamyHomogamyHomogamy NHW-NHW MEX O – MEX O MEX B – MEX B

11. Centre d’Estudis Demogràfics Table 4: Log-odds ratios for educational homogamy by year and ethnicity (2000), including Mexico -3 -2 0 1 2 3 4 5 Schooling +- HomogamyHomogamyHomogamyHomogamy NHW-NHW MEX O – MEX O MEX B – MEX B MEXICO

12. Centre d’Estudis Demogràfics Table 5 (abbreviated): Log-odds ratios for endogamous unions by census and years of schooling completed 198019902000 <9=> 16<9=> 16<9=> 16 Mexican Born / Mexican Born <92,962,713,533,283,863,61 => 162,982,363,552,943,883,26 Mexican Origin / Mexican Origin <9-0,75-0,62-1,24-1,11-1,52-1,39 => 16-0,76-0,45-1,25-0,94-1,53-1,22 Non Hispanic White / Non Hispanic White <93,313,733,383,803,423,85 => 163,274,32 3,354,39 3,394,43

13. Centre d’Estudis Demogràfics Summary 1. 4 patterns of educational homogamy: Mexico: Residents; USA: Mex-Born, Mex-Origin, NHW 2. Compared with Mexican couples resident in Mexico: a. Mex-Born in USA are more educated than Mex-Resident b. Gender differences are not as marked in USA c. Hypergamy in Mexico vs. homogamy in USA d. Strong homogamy at extremes of educational hierarchy both in Mexico and, for NHW, in the USA. 3. Education has an unequal role according to the ethnic pairings of spouses, especially among the lower grades of schooling 4. Mexican ancestry supports the hypothesis of assimilation …..

14. Centre d’Estudis Demogràfics Summary 5. Endogamy varies in function of the educational attainment of the spouses. >> At higher levels of education, endogamy decreases among Mexican Born >> and increases among Non Hispanic White

15. Centre d’Estudis Demogràfics Thank you!! * * * * * * Contact: aesteve@ced.uab.es aesteve@ced.uab.es