Statistical offices are increasingly in the position where they wish to, or are required to, replace surveys with administrative data. As a result of this trend, many European countries base turnover estimates on the administrative data of the VAT (Value Added Tax) register. For short term statistics, however, it is often the case that the VAT-data are still incomplete when the turnover estimates are required. One of the work packages of the European Statistical System Network (ESSnet) project on the use of administrative data in business statistics aims to address this issue. In particular, it aims to provide recommendations on possible estimation techniques for monthly and quarterly turnover estimates based on incomplete VAT-data.
Incompleteness of VAT-data for STS-estimates is generally caused by a combination of two factors. The data are incomplete, because they may be reported later. The data are incomplete, because they may be reported for a longer reference period. This ESSnet project revealed that the VAT data are – in practice - fairly complete when the quarterly turnover estimates have to be made. Hence, these estimates can be based on VAT-data alone in most European countries. The situation is more complex for monthly turnover estimates, especially in the case of the early estimates for the retail trade (according to European regulations first turnover estimates for the retail trade have to be provided thirty days after the end of the month). By far most European countries have only few VAT-data available when these 't+30 days' estimates have to be provided. As a consequence, these estimates cannot be based on VAT-data alone. In this paper, it will be shown that VAT-data from previous periods might be used as auxiliary information for these estimates. Two techniques will be described: a) using VAT as auxiliary information for weighing a mini-survey with a regression estimator, b) extrapolation the VAT-data with a mini-survey. There will also be discussion of other techniques due to be analysed and of the anticipated outcome of the project.
Keywords: Administrative data; Short term statistics; VAT; Turnover
Biography: Pieter Vlag has worked at Statistics Netherlands since 1998. He has contributed on various project on the edge of statistical production and pure methodology. Pieter was one of the first users of administrative data at Statistics Netherlands. Since 2009 he is leading a work package about the use of administrattive data for short term statistics within an european ESSnet “AdminData” project.