Gains from Combining the Anderson-Moore Algorithm and Julliard's Stack Algorithm

This paper describes an extension of the stack algorithm of Julliard, which makes it possible to (a) use any linear terminal condition, and (b) compute fixed points with the algorithm. Numerical results are presented for solving three different macroeconomic models including the Canada Model. The results show that employing the linear constraints generated by the Anderson-Moore algorithm significantly reduce the computational burden. The author can provide potential users with either a C or Mathematica version on request.