Estimation Methods for Business Statistics Variables That Cannot Be Obtained from Administrative Data Sources
Pete Brodie
Methodology Directorate, Office for National Statistics, United Kingdom

Statistics offices are increasingly in the position where they wish to, or are required to, replace surveys with administrative data. For multivariate surveys, it is often the case that some of the variables required are available directly from administrative sources, whereas others are not. This paper presents results on work to assess the applicability of estimation methods for a number of variables that are not generally available without using a dedicated survey. The work is part of the European Statistical System Network (ESSnet) Project on the uses of Administrative and Accounts Data for Business Statistics. The aim of this part of the project is to provide statistical offices with the tools to be able to replace surveys by a combination of administrative data and estimated variables.

The paper will present the results of investigating estimation methods for the first set of variables considered by the project – New orders, Payments for agency workers, components of Change in stocks of goods, Purchases of goods for resale in the same condition as received, and Number of employees in full time equivalent. There will also be discussion of the remaining variables due to be analysed and of the anticipated final outcome of the project.

Keywords: quality estimates; administrative data; business statistics; survey methodology

Biography: Pete Brodie is a senior principal methodologist at the Office for National Statistics in the UK, with responsibility for ONS methods for editing, imputation, sampling and estimation in business surveys.