Predicting prepayment risk of China’s RMBS

Categories: Macroeconomics, Industry/Sector Analysis, General Market Analysis, Fundamental Analysis, Technical Analysis, Other, Fixed Income, Economics, Financial Analysis, Investor Education

Country or region: China

A new semi-parametric prepayment prediction model of China's residential mortgage-backed securities (RMBS) incorporates the country's market and mortgage characteristics. Many factors have highly non-linear effects on the prepayment rates.

This report qualifies for 1 CE credit under the guidelines of the CFA Institute Continuing Education Program. 
We encourage CFA Institute members to log in to the CE tracking tool to self-document these credits.

Reference URL:


In this research, we develop a semi-parametric prepayment prediction model that incorporates China’s market and mortgage characteristics. Exploiting the flexibility of this semi-parametric model, we provide evidence that many factors have highly non-linear effects on the prepayment rates. Our empirical findings show that
  • Partial prepayments and full prepayments exhibit different behavioral patterns in China’s mortgage prepayment data. In general, the patterns of total prepayments are mainly determined by those of full prepayments.
  • Loan age (or seasoning) shows different relationships with partial and full prepayments. For example, in China Construction Bank’s (CCB) prepayment data, the loan age effect on the partial prepayment rate first increases, reaches a peak around the 23rd month, and then declines. In contrast, the impacts on the full prepayment rate keep increasing as loan age increases.
  • Both partial and full prepayments have an apparent seasonal pattern. The prepayment rates are higher in February, March and April than in other months, which are possibly driven by the seasonal pattern in home sales or borrowers’ available income.
  • The difference between the rate on the mortgage and the current market rate has obvious influences on both partial prepayment and full prepayment behaviors.
  • The borrowers’ prepayment behaviors are significantly influenced by housing market activities and macroeconomic factors.
Our findings about the relationships between the factors and prepayments shed new light on how the driving factors affect borrower’s prepayment decisions. The prediction results based on historical prepayment data and RMBS performance data support the model specification and demonstrate the model’s superior predictive power. The applications of the model include cash flow projections as well as valuation and risk management for mortgages and RMBS.

Date of original publication:


Total Views: 584
Total Downloads: 20

Share Article

Reader Comments

No comments made on this post yet


If you have any copyright and other associated infringements related to this item, please click on the Terms and Conditions link where you will be directed to the Digital Millennium Copyright Act (DCMA) that will outline the procedure for raising your concern.

If you have any concerns with the content of the item [e.g., offensive language and/or material, inappropriate material] then please proceed to utilize the Contact Us form. Remember that when using the Contact Us form, please ensure you reference/cite clearly the item in question (e.g., name of article, author(s) of article) and the nature of the complaint.