Robust Inference for Measures of Persistence in Singapore Sectoral Property Price Indexes

This paper examines the pattern and persistence of changes in price indexes of Singapore private housing, office space, shop space and industrial properties. Unit root tests find convincing evidence of non‐stationarity in all four sectors' quarterly prices. However, more detailed analysis of the autoregressive coefficient estimate, using recently developed methods that are robust to near unit‐root possibilities, reveals clear differences between residential and the non‐residential property price indexes. Three distinct types of dynamic behaviour can be discerned, characterized by the shape of the empirical impulse response function. Only private residential prices are clearly non‐stationary. For shop space, a stationary model is indicated, while the remaining two categories -- namely office space and industrial property -- are more likely to be near‐unit root processes depicting a high degree of persistence. Further analysis reveals hump‐shaped empirical impulse responses in the office space and industrial property price series.