Summary
Angrist (2001) uses U.S. immigration quota legislation enacted in 1921 and 1924 as a natural experiment to identify the causal effect of sex ratios on marriage markets and labor outcomes. By exploiting exogenous variation in immigrant sex ratios across 11 ethnic groups and three census years, the study shows that higher male-to-female ratios among the foreign-born causally increase female marriage rates, reduce female labor force participation (LFP), and raise male wages for second-generation Americans. The endogamy structure of early 20th-century immigrant communities — over 50% within-group marriage for most groups — validates the marriage-market channel. Effects extend to the third generation: children of the affected cohorts show higher paternal income scores and per-capita family income, consistent with the hypothesis that more stable family formation improves intergenerational outcomes.
Key Claims
- The Immigration Acts of 1921 (Johnson Act) and 1924 (Johnson-Reed Act) capped national-origin immigration quotas, exogenously changing the sex composition of immigrant flows by favoring family reunification over male labor migration — prior to the acts, immigrant streams were male-dominated.
- Instrument: sex ratio of arrivals by ethnicity and decade from Ferenczi and Willcox (1929), linked to second-generation outcomes in Integrated Public Use Microdata Series (IPUMS) 1910, 1920, and 1940; 33 year-ethnicity cells covering 11 groups (British, Irish, Italian, Canadian, Mexican, Nordic, German/Austrian, Hungarian/Romanian, Russian/Polish, Central/Eastern European Jews, not elsewhere classified (NEC)).
- Ordinary least squares (OLS) estimate: a 0.25-point increase in the sex ratio raises female marriage rates by approximately 3 percentage points; instrumental variable (IV) estimates are similar in sign.
- Female LFP falls by roughly two-thirds the size of the marriage rate increase, consistent with home-production substitution following marriage in the Becker (1973) specialization model.
- Male marriage rates rise: two-stage least squares (2SLS) coefficient of +0.036 for second-generation men (marginally significant); +0.048 for all foreign-stock men.
- Male log wages show a small positive effect at the 10% level, consistent with male "efficiency" investment when marriage prospects improve.
- Endogamy is high and validates the channel: over 80% within-group marriage for Italians and Central/Eastern European Jews; effects are stronger in high-endogamy groups, confirming marriage-market spillovers operate through co-ethnic networks.
- Third-generation outcomes: a 1-standard deviation (SD) increase in the sex ratio is associated with approximately +1.5% per-capita family income for children of the affected cohort; paternal income-score effects are also positive and significant.
- The exclusion restriction holds: quota-induced changes in the sex composition of immigrants affected second-generation marriage markets, not directly the second-generation's own labor market productivity.
- First-stage F-statistics are strong; the instrument is not weak.
Concepts Introduced or Extended
- Marriage Market — provides historical IV evidence that sex ratios causally shift female marriage rates and female LFP via the Becker specialization mechanism; complements Autor, Dorn, and Hanson (ADH) (2019) trade-shock evidence
- Nonmarital Fertility — sex-ratio variation shapes marriage-market conditions that determine nonmarital childbearing; historical complement to Wu (2008) cohort trends
- Intergenerational Income Mobility — third-generation income effects of parental marriage-market conditions confirm downstream welfare consequences of sex-ratio shocks
Entities Mentioned
Quotes
"The results suggest that variation in sex ratios in the United States leads to substantial changes in marriage rates and labor supply in a manner consistent with the predictions of economic models of marriage."
"The identification strategy exploits variation in sex ratios due to the Immigration Acts of 1921 and 1924, which imposed national-origins quotas on immigration for the first time in American history."
My Take
The paper's greatest contribution is methodological: one of the cleanest IV designs in family economics, exploiting legislation that is unambiguously exogenous to individual marriage and labor market decisions. The finding that female LFP falls by ≈2/3 of the marriage rate increase provides clean historical support for the Becker specialization model — two decades before ADH (2019) demonstrated the same mechanism using trade shocks. The endogamy interaction is particularly convincing: if the mechanism runs through marriage markets, effects should concentrate in groups that actually marry within the group, which is precisely what the data show. The main limitation is sample size: 33 year-ethnicity cells provide limited statistical power, especially for third-generation outcomes. The exclusion restriction is credible for marriage outcomes but is more debated for wages, since quota-induced immigration could also affect labor market competition — Angrist controls for immigration count but the residual variation may not be fully clean.