Marrying Up by Marrying Down: Status Exchange between Social Origin and Education in the United States

Abstract

Intermarriage plays a key role in stratification systems. Spousal resemblance reinforces social boundaries within and across generations, and the rules of intermarriage govern the ways that social mobility may occur. We examine intermarriage across social origin and education boundaries in the United States using data from the 1968–2013 Panel Study of Income Dynamics. Our evidence points to a pattern of status exchange—that is, persons with high education from modest backgrounds tend to marry those with lower education from more privileged backgrounds. Our study contributes to an active methodological debate by pinpointing the conditions under which the results pivot from evidence against exchange to evidence for exchange and advances theory by showing that the rules of exchange are more consistent with the notion of diminishing marginal utility than the more general theory of compensating differentials.

Mismatches in the Marriage Market

“…For cohabiting couples, a good job is typically a requirement before committing to marriage or for making marriage financially feasible (Smock et al, 2005). The implication is clear: Mismatches in the marriage market in the form of shortages of economically attractive men may exacerbate uncertainty and heighten disincentives to marriage, especially at a time of rising education and growing financial independence among American women (Gibson-Davis et al, 2005;Schwartz, Zeng, & Xie, 2016;Watson & McLanahan, 2011).…”

Partner choice and parameter estimates: modelling the effect of preferences

“…Extending the simulations to address other theoretical issues would also be attractive, status exchange being an obvious candidate. The simulations currently take account of education and an unspecified attribute uncorrelated with education; it would be straightforward to extend them to incorporate other dimensions, and to explicitly model the impact of relative preferences on behaviour, and on outcomes, relating to work in the tradition of Schwartz et al (2016).…”

“…Loglinear models lose the individual dynamic focus, but control for margins. Schwartz et al (2016) make an interesting contribution to a debate on status exchange (Rosenfeld, 2005;Kalmijn, 2010;Gullickson & Fu, 2010;Rosenfeld, 2010) demonstrating convincingly why loglinear models are necessary to detect the presence of status exchange, by controlling not only for the marginal distributions, but also for correlation between the two status factors, and that simpler indices of association are misleading. This is a valuable paper in that it tackles in detail how to relate complex specifications of loglinear models to theoretical arguments, how to estimate very specific quantities (in particular, the extent to which there are excess cases in status exchange cells, over and above what would be observed without status exchange preferences).…”

“…Assortative mating remains a fertile base for development of loglinear models of square tables: Bucca and Urbina (2019) present a Lasso regularisation approach to selecting models, which they show to be more sensitive (particularly with small or sparse data sets) to the data generating process than conventional goodness of fit statistics like BIC. Similarly, Schwartz et al (2016) demonstrates very clearly that loglinear models are necessary to detect evidence of status exchange by controlling not only for the marginal distributions but also the association between the various status dimensions. Bouchet-Valat (2014) presents a single index, the Mean Absolute Odds Ratio, a transformation of loglinear coefficients, which can be interpreted as a measure of the strength of homogamy.…”

“…Currently there is significant activity on issues such as the interplay between homogamy or heterogamy and factors such as • the income distribution (e.g., Salverda and Brals, 2012;Ravazzini, Kuhn and Suter, 2019) • inter-and intra-generational inequality (Schwartz, 2013) • migration, ethnicity and status exchange (Choi & Mare, 2012;Schwartz, Zeng & Xie, 2016;Telles & Esteve, 2019;Azzolini & 22:20:25, 2013),…”