Estimating the Income shares of a grouped frequency distribution of Incomes

Several techniques of income distribution analysis assume that the income shares of the classes into which the income range has been partitioned are known. These shares are often omitted from the income data and must be interpolated. The purpose of this article is first, to review three procedures for piecewise interpolation of income shares; and second, to compare their performances with a set of observed distributions. The results suggest that a spline function of loglogistic densities provides plausible estimates of income shares also with a very limited amount of information