[In recent decades U.S. households have experienced residential house prices moving persistently, that is, returns being positively serially correlated. We set up a realistically calibrated life cycle model with slow-moving time variation in expected housing returns, showing that not only age, labor income, and pre-existing housing wealth but also the state of the housing market significantly affect household decisions. Consistently with the data, the model predicts that in good states of housing market cycles (1) homeownership rates increase, (2) households buying homes invest a larger share of their net worth in their home, and (3) these households lever up more.]

During the last decades households in the U.S. have experienced that residential house prices move in a persistent manner, i.e. that returns are positively serially correlated. Since an owner-occupied home is usually the largest investment of a household it is important to understand how households act when they base their consumption and investment decisions on this experience. We show in a setting with housing market cycles and households who can decide whether they rent or own the home, that -besides the consumption and the precautionary savings motive -serial correlation in house prices generates a new speculative motive for homeownership. In particular, we show how good and bad housing market cycles affect homeownership rates, leverage, stock investments and consumption and can explain empirically observed household behavior during housing market boom and bust periods.