摘要： The empirical P-martingale simulation (EPMS) is a new simulation technique to improve the simulation efficiency for derivatives pricing when a risk-neutral model is not conveniently obtained. However, since the EPMS estimator creates dependence among sample paths to reduce its estimation variance, the corresponding standard error is not available from using only one simulation path. The asymptotic normality of the EPMS estimator is derived for piecewise smooth payoffs, which could be either continuous or discontinuous. Simulation results indicate that the asymptotic distribution serves as a persuasive approximation for samples consisting of as few as 500 simulation paths, which helps to reduce the computational costs. 出版者： Elsevier B.V 出版日期： 2014-05-01 出處： Computational statistics & data analysis, 2014-05, Vol.73, p.129-145 資源來源： Elsevier ScienceDirect Freedom Collection 版權： 2013 Elsevier B.V. 識別號： ISSN: 0167-9473 識別號： EISSN: 1872-7352 識別號： DOI: 10.1016/j.csda.2013.12.001
