Mortgage Interest Rates, Rents, and Local House Price Movements in New Zealand
Shi, Song, Journal of Real Estate Portfolio Management
Using monthly data from New Zealand housing markets, this paper examined the long-run relationship between mortgage interest rates, rents and local house price movements in an error correction model. It was found that house prices, rents, and interest rates were cointegrated and local house prices mean revert to the fundamentals, as indicated by the present value model. During this dynamic price adjusting process, the effect of interest rates on housing prices differed significantly across local markets. In general, interest rates had a smaller impact than rents on house price movements and the speed of adjustment of house prices to their long-run equilibrium was slow. Finally, the paper demonstrates how analysts can use a short time series to compare fundamental value to market price in a way that may lead to some predictability of local house price movements.
(ProQuest: ... denotes formulae omitted.)
Housing is the single largest asset held by many households worldwide. In New Zealand, house value represented nearly 75% of household net worth in 2005 (Bollard, Hodgetts, Phil, and Mark, 2006). Traditionally people viewed housing just as a consumption good. Today many people also view housing as an asset class. For example, in 2006 the Chicago Mercantile Exchange commenced trading housing futures contracts. However, the best methodology for pricing houses still remains unclear in the literature, due to both the consumable and economic nature of housing.
One obvious asset pricing method for housing is the present value model. This model relates the price of an asset to its expected future cash flows discounted to the present by using an expected discount rate. Campbell and Shiller (1988a, 1988b) suggested a log-linear present value model, called the dividend-ratio model, which has been widely used in finance and real estate investment analysis. Shiller (2006) argued that in theory house prices should be the present discounted value of future rents. By using the housing market data in the United States since 1995, he found there was a significant divergence between real house prices, interest rates, and rents. He concluded there was an irrational overpricing (bubble) for house prices in general.
Several studies have focused on testing the predictive power of the rent-price ratio as an explanation for house price movements (e.g., Clark, 1995; Capozza and Seguin, 1996; and Gallin, 2008). Very few studies, however, directly test the relationship between house prices, interest rates, and rents.
Méese and Wallace (1994) found house prices, rents, and the cost of capital are cointegrated, but they did not further examine the long-run equilibrium implied by the cointegration. One reason for the lack of studies directly testing the relationship between house prices, interest rates, and rents is due to the problem of small sample sizes in empirical research. Gallin (2004, p. 11) in his research of the long-run relationship between house prices and rents concludes that "using an error-correction model at a quarterly frequency may be asking too much of the data."
This study utilized local housing market monthly data to test the dynamics between house prices, mortgage rates, and rents by using an error correction model. The data period is from January 1994 to December 2009, which covers the interesting time period of global credit crunch in 2007. The benefit of using a short time series (monthly vs. quarterly) is it significantly increases the degrees of freedom in the proposed vector autoregressive (VAR) test and is more likely to reveal both the short-run and long-run relationship between variables. This paper contributes to the existing asset pricing literature in several ways. Firstly, it successfully applies the error correction model at monthly frequency to show that local house prices converge to their long-run fundamentals, as suggested by the present value model. …