Granger Causality Analysis of Land Supply and Housing Market Performance: Implications for Long Term Urban Land Policy in Hong Kong

Principal Investigator: L.H. LI (PI), S.K. WONG (Co-PI)
Funding body: PPR


Property prices have since the SARS attack in 2003 risen to a record high level in 2012 in Hong Kong . This unprecedented and sustained increase of housing price is attributable to a lot of factors, among them the allegation of “high land price policy” adopted by the government. Despite the political controversies behind this allegation, it is a fact that land supply in Hong Kong is limited and that impacts on the housing production in the city. However, for a long term and sustainable urban land policy to be formulated, it is imperative to establish the causal relationship between the land supply mechanism and the overall housing market performance (in terms of price movements and housing supply) first. This project looks at this causal relationship since 1980 by means of a Granger Causality model . We need to understand how land supply by the government via the public land sale programmes affects general housing prices, as well as specific sub-sector in the housing market, in terms of magnitude and time lag effect. Since the government has direct control over government land sale, this examination allows the authority to estimate their land supply policy on impact of market housing price level more effectively.


  1. This proposal examines the relationship between the overall land supply mechanism and the housing market. The examination concentrates on the correlation between land supply from government and housing price.
  2. The project intends to establish the causality correlation between land supply from government channel and housing price changes.
  3. This project intends to propose urban land policies that can deal with housing supply and affordable housing issues.


In this project, we use housing units completed on land supplied via government land sale programme as a proxy for land supply. Our dataset consists of the new residential housing supply estimated from the Government land sale and the private housing prices in Hong Kong. The dataset runs from the first quarter (Q1)1987 to the last quarter (Q4) of 2012. Land sale refers to the sale of land to developers from the government for residential development.

Over the 25-year period, the new housing supply generated from government land sale did not Granger-cause housing prices or vice versa in Hong Kong. Our results of an insignificant relationship between new housing supply and housing prices for the entire sample period from 1987 to 2012 is in line with our hypothesis that Government land sales comprise only a small portion of the housing stock.

On the other hand, a significant unidirectional causality result from housing prices to new housing supply is identified in subsample time periods, in particular after the Application List system was formally adopted and fully operational as the official channel for selling land in Hong Kong.


Hence, ceteris paribus whilst the supply of new land sites via the Government land sale programme are more responsive to housing price movements subsequent to the implementation of Application List due to the demand-oriented nature of the system which induces developers to apply for more land when housing prices are increasing. We find that higher housing price does Granger-cause an increase in housing units supplied from government land sale programme during that period.

The analyses of the correlation provide government with adequate empirical data support to streamline different sources of land supply directly or indirectly controlled by the authority. Urban land policy priority can then be formulated so that sustainable land development can be achieved.