Volume?related Heteroskedasticity and Liquidity Premium in Hedonic Pricing Model

Volume of transaction plays two roles in a market, first, it is related to liquidity of assets; second, it affects information cost in the price discovery process. The former results in a liquidity premium on price, which is a first moment effect of hedonic price; the latter produces price dispersion, which causes heteroskedasticity of error terms (the second moment) of hedonic price. This study tests empirically on these volume?related liquidity premium and heteroskedasticity in a panel sample of housing. The sample consists of more than 1600 transactions in 1999 of various housing developments in a small area in Hong Kong is studied. The results agree with our postulations that (1) relative liquidity imposes a positive premium on housing price; and (2) the magnitude of error terms in the hedonic pricing analysis is negatively related with the transaction volume in the previous 30?day in the same estate. The hedonic pricing model is re?estimated by an iterative generalized least squares (GLS) approach with volume?effect weighting. The efficiency of the estimation is greatly improved and a homoskedastic estimation is obtained.