TFG has built a method to value options with a non-standard pay-out function and Knock-Out barriers. As this calculation is integrated into a real-time system, performance and stability are critical. This has been achieved by using implicit finite differences where more detailed grids are used when the estimate of the error exceeds the required tolerance.
We describe a variety of VaR models in terms of their key attributes and differences, e.g., parametric vs. nonparametric, historical sampling vs. Monte Carlo simulation. We start with the simpler, well known models and then describe randomized historical simulation and filtered historical simulation, highlighting the features and benefits of these alternative methods. Filtered historical simulation has some unique attributes that could make it a better alternative for managing risk.
A standard technique used in both market and credit risk management for determining the risk of a portfolio of complex financial transactions is to value those transactions on a large number of possible market scenarios to see how changes in market prices could affect the value of the portfolio. A simulation model is used to produce a large number of possible future scenarios that represent how the markets could evolve. Validating these scenarios, given the large amounts of data involved, has typically been a cost and computational challenge. However, technologies such as Hadoop allow this to be accomplished at a much lower cost.
The objective of financial risk management systems is to estimate how much the value of a portfolio could potentially change, and in the case of credit risk, how that change in value impacts the exposure to a counterparty. A standard technique for doing this, called Monte-Carlo simulation, involves the generation of scenarios of the future possible values of the numerous market rates and prices that impact a portfolio’s value, valuing the portfolio on each possible scenario, then comparing the possible value to the current value. For statistical accuracy a large number of scenarios needs to be considered and the element of time included, as a result the total number of scenarios can easily reach 500,000 that the portfolio has to be valued over.
The formally humble discount curve has had an exciting time over the last few years. In this short paper we aim to show you where best practice for curve construction is heading, why it is necessary, and what its effects will be.
In this short paper we detail the underlying process for using radial basis functions, show a simple example, and then show an application of the method to simplify the estimation of the volatility of interest rate swaps when the supplied data may be sparse.