The new S language: a programming environment for data analysis and graphics
The new S language: a programming environment for data analysis and graphics
| Hi-index | 0.00 |
It is important for financial institutions to develop methods to predicttheir exposure and keep their risk under control. Portfolio managers caninsure themselves against the value (of a diversified stock portfolio)dropping below a certain level, by holding in conjunction with the stockportfolio, an index option derivative security. The work reported in thispaper is concerned with the study of non-parametric methods for estimatingthe pricing formula of option derivative securities. Two non-parametricapproaches, the projection pursuit method (PPR) and the local polynomialapproach (LOESS), are studied and compared to a benchmark parametricBlack–Scholes (B–S) approach. The practical relevance of theseapproaches is tested, when applied topricing and hedging of real-world LIFFE FTSE 100 index options from April 1997to November 1997. We compare the two methods by means of constructing ariskless portfolio of stocks, bonds and option derivatives securities. Theportfolio is then delta-hedged on a daily basis using a dynamic tradingstrategy in stocks and bonds during the lifetime of the option instrument.The tests carried out show that both methods generate similar responses,although each method can outperform the others depending on marketconditions, such as, time to maturity of the option instrument.