PurposeThis paper aims at analyzing the asymmetries created by the Great Recession in the US real estate sector.Design/methodology/approachThis paper uses a Markov-switching dynamic regression model in which parameters change when the housing market moves from one regime to the other.FindingsThe results show that the effect of real estate loans, interest rate, quantitative easing and working age population are asymmetric across bull and bear regimes. It is also found that the estimated parameters are larger in bull regime than bear regime, indicating a tendency to create house price bubbles in bull market.Practical implicationsSince three of those asymmetric variables (real estate loans, interest rate and quantitative easing) are related to monetary policy, the Fed can mitigate their impact on an interest-sensitive sector such as housing by engaging in a countercyclical monetary policy.Originality/valueThe estimated intercept and the variance parameter both vary from one regime to the other, thus justifying the use of a regime-dependent model.