【摘要】 基于资产定价的视角,本文通过将标准的动态戈登增长模型和传统的住房使用成本模型相结合,建立了一个关于住房市场租金收益率的动态住房使用成本模型。该模型将租金收益率分解为购房的预期资金成本、预期购房相对于租房的风险溢价和预期未来租金增长率三个部分的现值之和。进一步,本文将该模型应用于京沪广深四大城市的季度数据,并使用方差分解方法来考察国内住房市场动态波动的影响因素及其相对重要性。本文结果表明,资金成本变动在四大城市的住房市场波动中为最重要的影响因素,而租金在住房市场波动中虽然存在着一定的影响作用,但并不如资金成本显著。此外,本文还发现,不能直接观测得到的购房相对于租房的风险溢价也是影响国内住房市场的一个不可忽视的重要因素。值得注意的是,近年四大城市居民租房面临的风险相对于购房正日益上升。更多还原

【Abstract】 Housing prices in China’s major cities have been surging since 2003. Given that this may be one of the main macro risks to the real economy in China,great effort has gone into establishing the economic forces driving the large swings in China’s housing market. Asset pricing theory typically tries to answer this question by relating an asset’s price to its future cash flow uncertainty,usually in the form of a present-value statement.Among the various measures of valuation,the rental yield,equivalent to the dividend-price ratio in the stock market,is of particular importance in assessing the housing market because it reveals an agent’s expectations of future returns and rental growth in the housing market. Moreover,given the rapid rise in rental prices in major cities over the past few years,the rental yield has triggered a fresh wave of attention and public discussion regarding the forces driving the fluctuation in rental yield.From the perspective of asset pricing,researchers usually use the dynamic Gordon growth model,which is originally developed by Campbell and Shiller to decompose the stock market dividend yield in order to relate rental yield to future rental growth and housing returns. However,in the literature of real estate economics,a large body of work on housing market fluctuations applies the user cost of housing model. This model is in fact a no-arbitrage condition in which the marginal benefits( rental price) equal the marginal cost of housing including the cost of capital,the potential capital gain or loss,and the risk premium of owning relative to renting a house.Therefore,we incorporate the user cost of housing model into the standard Campbell-Shiller present-value model of rental yield to decompose the rental yield into three components: the expected cost of capital,which measures the cost of capital of buying a house; the expected risk premium of owing versus renting,which gauges the premium of house tenure as a hedge against the risk of renting a house; and the expected future rent growth,which captures the value of housing service flows in the future. Based on our theoretical model,we exploit a unique matched dataset of the sale prices and rents for the four Chinese first-tier cities compiled by the DTZ Company,the base-year rental yield from CICC,and interest rates from the PBC’s website. We then construct empirical proxies for the relevant expectations in our present-value model using the vector autoregressive method. We apply the variance decomposition approach to examine quantitatively how much of the variation in rental yield comes from the three components mentioned above. Our results show that the cost of capital plays the most vital role in all four major cities,followed by changes in the risk of owning versus renting.While innovation in rental growth also plays a part in the fluctuation of rental yield,it is not as significant as the cost of capital. Recently in particular,the risk premium of renting relative to owning a house in most first-tier cities seems to be rising.Our contribution is three-fold. First,to our knowledge,no prior study has applied the dynamic Gordon growth model in China’s four major cities to examine the driving forces in the housing market by exploiting a unique sample of matched transaction data on housing and rental prices. This exercise extends the research perspective and framework of the existing literature. Second,and more importantly,in contrast to the common practice of decomposing return into the risk-free rate and risk premium in the stock market,and the static relation between rental yield and housing user cost in the real estate economics literature,this study combines the standard present-value model with the classical user cost of housing model to build a dynamic user cost model to identify the forces driving the fluctuations in rental yield in China’s housing market. Third,our approach provides a new model-based measure of the risk premium of owning versus renting,which is not directly observable,and gauges its impact on housing market fluctuations.更多还原