In economic and business data, the covariance or correlation matrix of a random vector often fluctuates with time and exhibits seasonality. The most widely-used approaches for estimating and forecasting the correlation matrix (e.g., multivariate GARCH) often are hindered by estimation and inference difficulties, especially in high dimensions, and require strong assumptions. In this talk, we propose a new method for modeling and forecasting correlation matrices that allows the correlation to be driven nonlinearly by common factors. The nonlinear common factor approach simplifies estimation in high-dimensional settings and provides more flexibility than previous factor-based methods. This talk will introduce our method and illustrate its use on natural gas and power prices in Boston, which are related to the common factors of daily temperature and humidity.