Measuring a boom and bust: The Sydney housing market 2001–2006

Abstract

a b s t r a c tThe Sydney housing market peaked in 2003. The period 2001-2006 is, therefore, of particular interest since it captures a boom and bust in the housing market. We compute hedonic, repeat-sales and median price indexes for five regions in Sydney over this period. While the three approaches are in broad agreement regarding the timing of the turning point in the housing market, some important differences also emerge. In particular, we find evidence of sample selection bias in our hedonic and repeat-sales data sets (with the former focusing more on better quality dwellings and the latter more on lower quality dwellings). These sample selection biases could in turn cause bias (in opposite directions) in our hedonic and repeat-sales indexes. Median indexes may likewise be biased as a result of an apparent decline in the average quality of dwellings sold in the latter part of the sample. We also find evidence of convergence in prices across regions during the boom and divergence in the subsequent bust.

Hedonic Regressions and the Decomposition of a House Price Index into Land and Structure Components

“…Assume that these prices are equal to the sum of the land and structure costs plus error terms t n which we assume are independently normally distributed with zero means and constant variances. 6 This leads to the following hedonic regression model for period t, where t and t are the 2 For some recent literature, see Crone et al (2009, Gouriéroux and Laferrère (2009), Hill (2011), and Hill et al (2009. One variant of the hedonic technique regresses the logarithm of the selling price of the property on its price determining characteristics and on time dummy variables for all periods (except the base period).…”

Bubbles in the Taiwan housing market: The determinants and effects

“…Zhou and Sornette (2006) examined whether the price bubble burst in the US, and predicted that the turning point of the bubble would occur around mid-2006. Since the Sydney housing market peaked in 2003, Hill, Melser, and Syed (2009) became interested in the market covering the period 2001e2006, which captured a boom-and-bust cycle. When prices of homes in the US rose rapidly from the fourth quarter of 1999 to the fourth quarter of 2005, Diewert and Nakamura (2009) developed a new accounting approach for housing in the determination of the consumer price index (CPI).…”

Managing hedonic housing price indexes: The French experience

“…23 16 It also assumes that the mean and variance of the error term are identical across all time periods, unless the model is corrected for heteroskedasticity. 17 Hill et al (2009) provide an insightful comparison of the bias induced by using repeat sales and time dummy hedonic method in the city of Sydney. 18 It is important to estimate the models on homogeneous price zones, where prices are not too different, and move in the same way over time.…”