We study the interactions among geographical mobility, unemployment, and home-ownership in an economy with heterogeneous locations, endogenous construction, and search frictions in the markets for both labour and housing. The decision of home-owners to accept job offers from other cities depends on how quickly they can sell their houses ( i.e. the houses' liquidity ), which in turn depends on local labour market conditions. Consequently, home-owners accept job offers from other cities at a lower rate than do renters, generating a link between home-ownership and unemployment both at the city level and in the aggregate. When calibrated to match aggregate U.S. statistics on mobility, housing, and labour flows, the model predicts that the effect of home-ownership on aggregate unemployment is small. When unemployment is high, however, changes in the rate of home-ownership can have economically significant effects.

The dynamics of house prices, sales, construction, and population growth in response to city-specific income shocks are characterized for 106 US cities. A dynamic model of search in the housing market in which construction, the entry of buyers, house prices, and sales are determined in equilibrium is then developed. The theory generates dynamics qualitatively consistent with the observations and a version calibrated to match key features of the US housing market offers a substantial quantitative improvement over models without search. In particular, variation in the time it takes to sell induces transaction prices to exhibit serially correlated growth. (JEL D83, R21, R23, R31)