Cohort Longitudinal Analysis of the Income Composition of Demand for New Houses, 1960-2000

Rachel E. Dwyer, Ohio State University

This paper argues that shifts in the population of new house buyers underlie the increasing size of new houses in the US at the end of the twentieth century. Following Myers (1990, 1999), I draw on a revised “filtering theory” to conceptualize historical change in the housing stock within a demographic framework. Since income trends are key period effects for cohorts of households, I also analyze the income composition of demand for new houses. I compare the age and income profile of households living in 1990s vintage houses to those living in 1950s and 1960s houses when new. I use US Census microdata for 1960 to 2000 to develop a cohort longitudinal estimation of historical change, and employ both analytic graphs and regression techniques. I find significant shifts between cohorts of households in the age and income composition of demand for new houses, with implications for the entire housing market.