This paper explores estimation of parameters within large general equilibrium models, given limited data availability and without calibration. This leads directly to consideration of possible measures of the robustness of the model vis-à-vis the historical record (as represented by available data series), and the use of specification tests for comparison and testing of different generalequilibrium based theoretic paradigms. Such an approach also offers an opportunity for assessment of general equilibrium effects (like the impact of trade on wages or on the location of industry) currently highlighted in the policy literature.