Potential housing bubble with Chinese characteristics : analysis and policy design through an operational model

Abstract

This thesis analyzes the potential housing bubble in the Chinese urban housing market. Using an operational model built on system dynamics (SD), it explores the unique housing market structures, information flows, key agents and their decision-making processes, and the system constraints that may contribute to the bubble forming phenomenon. Its SD models differ from intensive data-driven economic models. They are structural, operational, and focus on causal relationships. They can more easily accommodate non-market features and unique institutional components, where long-range historical data are not readily available. As the Chinese government heavily controls its housing markets, housing prices greatly depend on political decisions. The models incorporate the incentives and decisions of key agents in the market and have predictive power through simulation and scenario analysis.

They can thus help decision-makers transform decisions, actively manage risks and opportunities, timely design and implement policies, and consequently change the system from within. This thesis specifically discusses three Chinese-specific features: 1. Rising housing price and the cap rate change; 2. High vacancy rates caused by speculators purchasing multiple housing units as money-storage investments they are unwilling to rent; 3. Land financing schemes where local governments rely on income from land sales to support their budgets which leads them to use their monopolistic position to short supply the land. The research design starts from the DiPasquale-Wheaton model (D-W) which Western urban economics theory has validated for analyzing housing cycles. It operationalizes this by: 1. Converting the D-W to a basic SD model; 2. Augmenting this with additional generic features; 3. Incorporating unique Chinese market features to create China-specific models; and then 4. Creating an integrated overarching model for conducting a case study using historical data of Nanjing, China.

The result is one of the first operational models for the Chinese housing market that has the explanatory mechanisms and somehow overcomes the data availability issues. It provides an intuitive and transparent structure that we can easily modify to address complex issues. This thesis analyzes the potential housing bubble in the Chinese urban housing market. Using an operational model built on system dynamics (SD), it explores the unique housing market structures, information flows, key agents and their decision-making processes, and the system constraints that may contribute to the bubble forming phenomenon. Its SD models differ from intensive data-driven economic models. They are structural, operational, and focus on causal relationships.

They can more easily accommodate non-market features and unique institutional components, where long-range historical data are not readily available. As the Chinese government heavily controls its housing markets, housing prices greatly depend on political decisions. The models incorporate the incentives and decisions of key agents in the market and have predictive power through simulation and scenario analysis. They can thus help decision-makers transform decisions, actively manage risks and opportunities, timely design and implement policies, and consequently change the system from within. This thesis specifically discusses three Chinese-specific features: 1. Rising housing price and the cap rate change; 2. High vacancy rates caused by speculators purchasing multiple housing units as money-storage investments they are unwilling to rent; 3 . Land financing schemes where local governments rely on income from land sales to support their budgets which leads them to use their monopolistic position to shortsupply the land.

The research design starts from the DiPasquale-Wheaton model (D-W) which Western urban economics theory has validated for analyzing housing cycles. It operationalizes this by: 1. Converting the D-W to a basic SD model; 2. Augmenting this with additional generic features; 3. Incorporating unique Chinese market features to create China-specific models; and then 4. Creating an integrated overarching model for conducting a case study using historical data of Nanjing, China. The result is one of the first operational models for the Chinese housing market that has the explanatory mechanisms and somehow overcomes the data availability issues. It provides an intuitive and transparent structure that we can easily modify to address complex issues.