Abstract
Despite a huge popularity of convertible bonds and a variety of different pricing models, very little empirical research on their valuation has been undertaken. This paper investigates the pricing performance of arguably the most popular among practitioners, the Tsiveriotis and Fernandes approach implemented within a trinomial-tree model with exogenous credit spread using daily market prices on the largest available historical database (∼1500 CBs). We demonstrate that using a constant credit spread within that framework would lead to significant mispricing in average. By contrast, we present a practical approach for valuing convertible bonds with moving-average credit spread, which leads to acceptable results. We prove that historical volatility and implied credit spread are cointegrated for the majority of bonds in our sample. This allows us to model credit spread based on cointegrated relationship with volatility and extend our model to the cases with limited market information available. The mispricing is proven to be statistically insignificant in the researched convertible bonds.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have