Robust estimation of nonlinear structural models

The problem of the identification of nonlinear errors-in-variables models with large observational errors in the explanatory variable is considered. On the basis of robust estimation methods, we propose a development of the algorithms of adjusted and total least squares is suggested. This enabled us to get a better precision of the reconstruction of the response in the presence of outliers in the sample. The proposed algorithms are used in constructing the Engel curve from the data of a budget survey. As a result we managed to make more correct conclusions about the behavior of households with income variation.