|previous article in this issue||next article in this issue|
Document Details :
Title: A Note on the Normal Power Approximation
Author(s): RAMSAY, Colin M.
Journal: ASTIN Bulletin
Volume: 21 Issue: 1 Date: April 1991
The normal power (NP) approximation essentially approximates the random variable X as the quadratic polynomial X̄ ≈ Y+γ(Y2- 1)/6 where , X̄ = (X - μ)/σ is the standardized variable, Y ~ N(0, I), and μ, σ γ are the mean, variance skewness of X respectively. The coefficients of this polynomial are not determined by equating the lower moments. It is shown that matching these moments does not improve the overall accuracy of the approximation.