Advice on research project.

Posted by Joseph Wang on Jan 28, 2005; 11:26pm
URL: http://quantlib.414.s1.nabble.com/Advice-on-research-project-tp3553.html

I have a small research project that I am trying to do with quantlib,
and before I start I'd like some advice on modifying the code.

In the Journal of Equity Investment (9/1998), Tsveriotis and Fernades
come up with a way of modelling default risk in convertible bonds via
two coupled Black-Scholes equations.  One that models the convertible
bond and one hypothetical security that models the cash only part of the
convertible bond.  Fortunately the coupling occurs does not involve any
of the derivatives and I think I can fit the system into the current
finite difference engine.

I'm curious if anyone has any thoughts on what to do, and if there are
any other parts of quantlib which could be modified to value convertible
bonds.  I'm starting with Tsveriotis and Fernades models in part because
it involves parts of the code that I'm reasonably familar with.

In case anyone is wondering, the mini-research project involves the
value of convertible bonds on Shanghai.  Using the assumption of zero
credit risk, there are some papers in Chinese which argue that the value
of convertible bonds in Shanghai are undervalued.  What I'd like to do
is to take the prices and derive an implicit credit risk (which may or
may not have anything to do with the actual credit risk.

My sense is that what is happening is that the market is taking into
account interest rate and inflation risk.  These normally aren't taken
into account in convertible bond models.  What happens in the West is
that most convertible bonds tend to be used to finance high risk
ventures, and so they seem to start out very much out of the money,
whereas in China convertible bonds are used as stock substitutes in part
to prevent dilution of ownership of state owned enterprises, and they
tended to be issued with the conversion price very close to the stock
price.  There is also very little chance of credit risk for that one
company since these tend to be large state owned enterprises.  (There is
however the risk that the entire Chinese economy could tank, which I
think is part of what is being factored in.)

My gut feeling is that in this regime, interest rate risk plays a much
more important role that would be the case in the West, and I'm trying
to use quantlib to see if this is the case or not.