This paper discusses the relative strengths and weaknesses of alternative solution methods in applied general equilibrium (AGE) analysis. The particular focus is not on the technical properties of solution algorithms, but instead deals with the general issue of using linearised approximation methods, more popularly referred to as the Johansen class of AGE models, relative to models which are solved in level form via solution of a system of non-linear equations. Particular attention is given to the practical aspects of AGE empirical work with special emphasis on problems encountered in multi-country modelling.