This article proposes the generalized lambda distribution as a tool for modeling nonlognormal security price distributions. Known best as a facile model for generating random variables with a broad range of skewness and kurtosis values, the generalized lambda distribution has potential financial applications, including Monte Carlo simulations, estimations of option‐implied state price densities, and almost any situation requiring a flexible density shape. A multivariate version of the generalized lambda distribution is developed to facilitate stochastic modeling of portfolios of correlated primary and derivative securities. © 2001 John Wiley & Sons, Inc. Jrl Fut Mark 21:213–236, 2001