This article uses regime-switching models of the threshold type to analyze the adjustment process of rental prices for three U.K. commercial real estate sectors over the period 1974–2008. The nonlinear models outperform their linear counterparts in in-sample fit. Their out-of-sample forecasting ability is better whenever the corresponding linear models contain a significant amount of neglected nonlinearity. Regime switches are triggered when the growth rates of rental price exceed certain threshold levels. For the industrial and retail sectors such regime switches occur in situations of strong excess demand, for the office sector they occur when there is strong excess supply.