Additional Topics

Smoothing

Smoothing methods tend to greatly over-estimate labor income at young ages. The most reliable method is to exclude children who have not yet reached working age. This may be determined empirically as the first age for which a positive value is estimated or on a priori grounds based on the survey instrument, e.g., all those under age 15 or under age 12. We then smoothed the labor income for those who had reached working age using Supsmu with a relatively small bandwidth, e.g., 0.05 or 0.10. To assess the results it is a good idea to calculate the smoothed relative to the unsmoothed value to judge the percentage error.

Social benefits for public workers

Cohort data

Age profiles can be converted to cohort data. An example using labor income profiles for Taiwan, Labor income by cohort. For further information on converting cross-sectional data to cohort, refer to the Methods page on Cohort Data.