This paper examines the relationship between house prices, rents, and user costs of housing in the United States from January 2009 to March 2022. We first use the time-varying coefficient cointegration model to explain the long-run relationship and adopt an error correction model with endogenous regime switching, which turns out to fit the data better than existing models. Our results show that, following the sub-prime mortgage crisis, the U.S. housing market has switched between a strong or a weak house price momentum state. In the strong momentum regime, house price returns have been more persistent and error correction has been slower. The degree of house price momentum is estimated to be very high and explosive in the strong regime while it is moderate in the weak regime. It is estimated that 74% of the data remains in the strong momentum regime. The extracted latent factor determines the regime of the housing market, and we run the adaptive lasso on the FRED-MD to identify the link between house price momentum and the macroeconomic and financial variables.