A large body of literature in real estate has been devoted to the correlation between prices and changes in liquidity of the market (i.e. the ability to sell and buy more or less rapidly). However the specific role of liquidity risk ñ uncertainty in time to sell ñ has not been explored yet. Due to the lack of centralized information on current demand and supply, a household wishing to sell its house is confronted with two sorts of risk: (i) pricing risk (or uncertainty in the price level), (ii) liquidity risk which is damaging since keeping its house unsold on the market is costly. We complement the empirical analyses that have illustrated the influence of price risk on price dynamics by considering the role of liquidity risk proxied by a measure of volatility in sales volumes. Our empirical approach is based on a panel VAR model with GARCH in means effects estimated on data from several areas in and around Paris. In line with microeconomic arguments explaining risk-adverse buyers and sellers' strategies, uncertainty in liquidity negatively affects price changes and sales volumes, but price risk affects negatively price changes but positively sales volumes. The spatial dimension associated to the panel structure decomposes shocks into a local and a common component. Common shocks explain in the long-run more than 75% of the variability in prices and its larger part is related to the consequences of uncertainty. The situation is reversed when turning to the variance in sales volumes. Moreover, persistence of uncertainty is short lived for the common shocks and long lived at a local level. This gives a contrasted picture whose economic properties are analyzed in detail with help of a structural identification scheme in terms of demand and supply shocks.