An application of the transformed kernel density estimation to labor earnings in Spain

Most economic data are difficult to explore with the global window width kernel density estimator, because they require different amounts of smoothing in different locations. When data are right-skewed one possible approach to this problem is based on transformations of the data. We propose a simple method based on the skewness shown in the original data set. We illustrate the potential of our approach with two examples and we evaluate its usefulness when working with microeconomic data. The estimation algorithm is described. We discuss the application of our results to the study of labor earnings in the Spanish population using information from the Spanish Family Expenditure Survey.