|
Activity Number:
|
173
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Business and Economic Statistics Section
|
| Abstract - #304070 |
|
Title:
|
Multiple Imputation for Top-Coded Wages in German Social Security Register Data
|
|
Author(s):
|
Thomas Büttner*+ and Susanne Rässler
|
|
Companies:
|
Institute for Employment Research and Otto-Friedrich University Bamberg
|
|
Address:
|
Regensburger Str. 104, Nuremberg, International, 90478, Germany
|
|
Keywords:
|
top-coding ; missing data ; censored wage data ; multiple imputation
|
|
Abstract:
|
Top-coding or right-censoring of wages is a common problem with administrative data sets of economic interest, like the German IAB employment sample (IABS), which is based on the register data of the German unemployment insurance. We treat this problem as a missing data problem and use multiple imputation approaches to impute the censored wages by draws of a random variable from a truncated distribution, based on MCMC techniques. Here, we additionally use uncensored wage information from a survey (German structure of earnings survey, GSES) to confirm the validity of these approaches and to further improve the imputation quality. We perform simulation studies to compare different imputation approaches (e.g. considering heteroscedasticity vs. assuming homoscedasticity or based on a tobit model vs. using external information) and strategies (e.g. different imputation models).
|
