Abstract
The percent of households headed by married couples has recently fallen below 50%, while the percent of unmarried couples (both heterosexual and homosexual) continues to rise. Nationally-representative estimates of unmarried couples which first appeared in the 1990 and 2000 decennial Censuses are now available on an annual basis through the American Community Survey. In this paper we use state-level panel data from 2000–2006 on the percent of households headed by married couples, same sex couples and opposite sex unmarried couples to assess widespread claims in the popular press of causality across living arrangements. Based on Granger causality tests we can reject claims that an increase in same sex couples has caused either a decline in marriage or (except in one case) an increase in heterosexual cohabitation. There is mixed evidence whether or not opposite sex couples may have Granger caused same sex couples, but stronger evidence that marriage and heterosexual cohabitation are interrelated.