We use low-order projection methods to compute numerical solutions of the basic neoclassical stochastic growth model. We assess the quality of the obtained solutions, and compare them to numerical approximations derived with first and second-order perturbation techniques. We show that projection...