Estimating Business Statistics by integrating administrative and survey data: an experimental study on small and medium enterprises

The paper deals with the problem of estimating structural business statistics by exploiting already existing administrative information integrated with survey data. In particular, the aim of the study is to verify the possibility of estimating key structural variables which are not directly available from administrative sources: this implies the need of using either estimation or imputation models to derive the required estimates. In the present paper, the attention is focused on the variables relating to changes in stocks of goods and services investigated in the annual survey on small and medium enterprises (Small and medium enterprise survey -SME): different imputation strategies are experimentally evaluated depending on the different scenarios corresponding to the various response patterns determined by the availability of the analysed variables in one, more or none of the considered administrative archives.