Abstract: This paper discussed the data-driven pricing method of toll road asset-backed securities. Firstly, the ARIMAR-GARCH model reflecting the risk characteristics of traffic volume was constructed, then the issuance scale and coupon interest rate of different grades were calculated based on the waterfall repayment structure and the NelsonSiegel model. Finally, the simulation analysis was carried out by combining the real data of Chengyu and Yurong Expressways. The results show that the revenue in the mature period had a sharp peak and thick tail, and the revenue in the growth period had a significant fluctuation and aggregation effect. The data-driven pricing model can describe the risk characteristics of revenue and reasonably estimate the proportion of securities, which was better than the cash flow discount method which only considers interest rate risk. From the perspective of issuing method, the financing structure of the combined issue was significantly better than that of the separate issue. The advice is to change the franchise financing based on a single expressway project and optimize the financing structure through the merger and issuance of projects with high performance and poor performance.

Key words: transportation economy, toll road, asset securitization, structured pricing, financing structure, ARIMAGARCH model

摘要： 从数据驱动视角出发，本文探讨收费公路资产支持证券的结构化定价方法。首先，构建反 映交通量风险特征的收费公路通行收入自回归积分滑动平均模型；其次，基于瀑布式偿付结构和 到期收益率Nelson-Siegel模型计算不同等级证券发债规模与息票利率；最后结合成渝高速和渝 蓉高速现实数据进行仿真分析。研究发现，高速公路成熟期的通行收入具有尖峰厚尾特征，成长 期的通行收入更兼具显著的波动聚集效应。数据驱动的定价模型能够刻画通行收入的风险特 征，合理估计优先级的证券融资比例，比只考虑利率风险的现金流折现法更具有优势；从发行方 式来看，合并发行的融资结构显著优于单独发行。管理启示是改变以单个高速公路项目为基础 的特许经营权融资，通过绩优绩差项目的合并发行优化融资结构。

关键词: 交通运输经济, 收费公路, 资产证券化, 结构化定价, 融资结构, ARIMA-GARCH模型