Working Paper: NBER ID: w1701
Authors: David E. Bloom; Neil G. Bennett
Abstract: This paper analyzes cohort marriage patterns in the United States in order to determine whether declining rates of first marriage are due to changes in the timing of marriage, the incidence of marriage, or both. Parametric models, which are well-suited to the analysis of censored or truncated data, are fit separately to information on age at first marriage derived from three data sets which were collected independently and at different points in time. Extended versions of the models are also estimated in which the parameters of the model distributions are allowed to depend on social and, economic variables.The results provide evidence that the incidence of first marriage is declining and that there is only a slight tendency for women to delay marriage. In addition, education is the most important correlate of decisions about the timing of first marriage whereas race is the most important correlate of decisions about its incidence.
Keywords: marriage; demography; socioeconomic factors
JEL Codes: J12; J13
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| incidence of first marriage is declining (J12) | proportion of women marrying (J12) |
| education influences timing of first marriage (J12) | age at first marriage (J12) |
| race correlates with marriage incidence (J12) | probability of ever marrying (J12) |