摘要
本文原创地提出了基于偏最小二乘回归(PLS)的可转债定价模型,将基于PLS的美式期权定价方法拓展到了可转债的定价;与传统模型相比,可以更好地解决多因素扰动条件下的可转债定价问题和可转债条款中的路径依赖问题。利用上述定价模型,本文计算了2004.8.1-2005.8.1期间在沪深两市交易的31只可转债的理论价格。实证结果显示,模型较好地模拟了可转债实际价格运动路径,价格估计误差在5%以下。说明该方法在实际中是可操作的,可为实际投资决策提供理论依据。
This paper creatively develops the convertible bond pricing model based on Partial Least Square Regression (PLS) . We formulate this model from the American option pricing model based on PLS. Compared with the traditional models, this model originated from a new thought, and can solve the convertible bond pricing problem under multi - factors and path- dependence. Using the data from Aug. 1st 2004 to Aug. 1st 2005 in Chinese capital market, we calculate the market prices of 31 convertible bonds during this period. The results show that the theoretical price of PLS model has a good fit with the actual price of convertible bond. The ratio of price error is lower than 5 %. This testifies that the convertible bond pricing model based on Partial Least Square Method is practical and can provide theoretical support for investment decision.
出处
《中国管理科学》
CSSCI
2006年第4期81-87,共7页
Chinese Journal of Management Science
基金
国家自然科学基金资助项目(70271010)
关键词
可转债
偏最小二乘回归
美式期权
定价模型
实证研究
convertible bond
Partial Least Square
binomial tree
American options
options pricing