This paper presents a first step towards a new theory of housing market fluctuations. We develop a life-cycle model where agents face credit constraints and their housing consumption is restricted to a discrete set of possibilities. The market interaction of young credit constrained agents climbing the property ladder with old agents trading down, generates co-movements of aggregate house prices, volume of transactions and income, consistent with the patterns observed in the U.S. and the U.K. Under plausible assumptions, the model reproduces the slight lead of transaction volume over the other two series as documented in the data. Our theory asserts that the fluctuations in housing prices depend crucially on fluctuations in the current income of young households (the first-time buyers). Thus, it sheds light on why housing prces are more volatile than GDP, and why they exhibit some degreee of predictability in a market where households optimize over the timing of their transactions.