Inferring Expectations from Observables: Evidence from the Housing Market

We assess the role of price expectations in forming the U.S. housing boom in the mid-2000s by studying the dynamics of vacant properties. When agents anticipate price increases, they amass excess capacity. Thus, housing vacancy discriminates between price movements related to shocks to demand for housing services (low vacancy) and expectation shocks (high vacancy). We implement this idea using a structural vector autoregression with sign restrictions. In the aggregate, expectation shocks are the most important factor explaining the boom, immediately followed by mortgage rate shocks. In the cross-section, expectation shocks are the major factor explaining price movements in the Sand States, which experienced unprecedented booms.