Loading...
3 results
Search Results
Now showing 1 - 3 of 3
- Pulled-to-par returns for zero-coupon bonds historical simulation value at riskPublication . Beleza Sousa, João; Esquível, Manuel L.; Gaspar, R. M.Due to bond prices pull-to-par, zero-coupon bond historical returns are not stationary, as they tend to zero as time to maturity approaches. Given that the historical simulation method for computing value at risk (VaR) requires a stationary sequence of historical returns, zero-coupon bonds' historical returns cannot be used to compute VaR by historical simulation. Their use would systematically overestimate VaR, resulting in invalid VaR sequences. In this paper, we propose an adjustment of zero-coupon bonds' historical returns. We call the adjusted returns "pulled-to-par" returns. We prove that when the zero-coupon bonds' continuously compounded yields-to-maturity are stationary, the adjusted pulled-to-par returns allow VaR computation by historical simulation. We firstly illustrate the VaR computation in a simulation scenario, and then, we apply it to real data on eurozone STRIPS.
- Brownian bridge and other path-dependent Gaussian processes vectorial simulationPublication . Beleza Sousa, João; Esquivel, M. L.; Gaspar, R. M.The iterative simulation of the Brownian bridge is well known. In this article, we present a vectorial simulation alternative based on Gaussian processes for machine learning regression that is suitable for interpreted programming languages implementations. We extend the vectorial simulation of path-dependent trajectories to other Gaussian processes, namely, sequences of Brownian bridges, geometric Brownian motion, fractional Brownian motion, and Ornstein-Ulenbeck mean reversion process.
- Machine learning Vasicek model calibration with Gaussian processesPublication . Beleza Sousa, João; Esquivel, M. L.; Gaspar, R. M.In this article, we calibrate the Vasicek interest rate model under the risk neutral measure by learning the model parameters using Gaussian processes for machine learning regression. The calibration is done by maximizing the likelihood of zero coupon bond log prices, using mean and covariance functions computed analytically, as well as likelihood derivatives with respect to the parameters. The maximization method used is the conjugate gradients. The only prices needed for calibration are zero coupon bond prices and the parameters are directly obtained in the arbitrage free risk neutral measure.