Microdata Imputation and Macrodata Implications: Evidence from the Ifo Business Survey

• Activity Number: 471
• Type: Contributed
• Date/Time: Wednesday, August 3, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Section on Survey Research Methods
• Author(s): Christian Seiler*+
• Companies: Ifo Institute for Economic Research
• Keywords: Business survey ; Longitudinal data ; micro data ; Imputation ; Nonresponse ; macro data

Abstract

Surveys are commonly affected by nonresponding units which can produce biases if these missing values can not be regarded as missing at random (MAR). As many papers examine the effect of nonresponse in individual or household surveys, only less is done in case of business surveys. This paper analyses the missing values in the Ifo Business Survey, which most prominent result is the Ifo Business Climate Index, a leading indicator for the businss cycle development in Germany. Covering the period from 1994 to 2009 (with about 7,000 responding firms per month) a considerable number of observations is available to compare various imputation approaches for longitudinal data to evaluate the method which reflects the underlying latent data generating process best. After this, the imputed data sets will be aggregated as usual and compared with the original indices to evaluate their implications to the macro level. We analyse the indicators for different aggregation levels and show that the bias is minimal.