An Economic Model of Intended and Realized Fertility

• In Event: Lightning Session: Family and Child Policy & Population and Migration Issues

Abstract

We use temporal and geographic variation in economic conditions and housing costs in the United States, the United Kingdom, Russia, and Israel to predict whether and when people have children. We use data from the US Panel Study of Income Dynamics 1968-2023, UK Understanding Society 2009-2024, British Household Panel Study 1991-2008, Russia Longitudinal Monitoring Study 1997-2023, and Israel Central Bureau of Statistics administrative data that include all cohorts born between 1950 and 1995. We merge to our individual respondents, time-series housing cost data for the small geographic areas where they reside in each year. Because these data follow households and household members over time, we exploit rich longitudinal data to model first and higher-order births. Our multi-prong approach exploits country specific features of social norms and social conditions, labor market conditions, and housing market features. For example, in the US, we use zoning restrictions that limit whether and in what form developers can build new housing units and recent relaxation of zoning restrictions. (https://www.opb.org/article/2024/02/17/the-hottest-trend-in-u-s-cities-changing-zoning-rules-to-allow-more-housing/). We exploit cross-country and time-series variation in economic factors that constitute part of the costs of having and raising a child. The factors include the cost of buying or renting a house, the prevailing fixed mortgage interest rate on a 30 year loan, prevailing adjustable mortgage interest rate (with a five year balloon), the prevailing median wage in manufacturing,  and the local unemployment rate. We also include in the empirical model each respondent's religious affiliation to capture the influence of normative beliefs as proposed by the Theory of Planned Behavior (Ajzen; 1991, 2005, 2012, 2013). We incorporate these measures into a model along the lines of Becker and Barro' 1988 dynastic utility model that predicts decisions to have a child. Our results are not yet available because we are using restricted data from each source that identifies the subnational areas of residence. These geographies allows us to exploit within-country cross-area variation in the factors mentioned above. To get results, we must request disclosure reviews from the respective institution that controls each data.

Author

• Dean R Lillard, The Ohio State University
  Presenting Author