This study examines how well the recent demand for various categories of consumer durables can be predicted for the short-run period. Two main types of model specification are computed. The first is based on traditional economic determinants, whereas the other uses new variables derived from surveys measuring households' financial expectations. It is found that this latter model explains and predicts the observed demand series as efficiently as the more popularly utilised type of equation. The rapid availability of expectations information renders this data potentially useful in generating short-term predictions of typically volatile demand series.