The basis of my model and its calculation of income distribution was gone into very carefully in my previous report, which contained a chapter about that and detailed comparisons with the best official surveys. To some extent, I stick to my guns about the robustness of the income distribution side of my work.