We study the special local content rules of the NAFTA for automotive goods. We present two variants of Benders decomposition to solve the non-linear optimization problems. We show that Benders decomposition outperforms secants approximation. Through the right EA-choice supply chain costs can be reduced significantly.