Long-Term Natural Cycle Momentum and Housing Price Changes in Israel, 1995–2013

Measures of price momentum, typically a moving average of previous price changes, are known to improve the predictive power of empirical price models. We introduce polynomially-smoothed price-trend derivatives as a refined momentum measure and apply them to a database of 216 consecutive months of Israeli housing prices. Over the entire housing price cycle and for phases characterized by prices that increase or decrease at an accelerating or decelerating rate, our models perform better than models with only economic fundamentals or those that add a moving average of past prices. We also show that some predictor variables that are significant over the entire housing cycle are not key predictors in all four phases; moreover, momentum is a key explanatory variable in some but not all phases of the housing price cycle.