Sample design for producer price indices

Abstract

The collection of price information is complicated because of the structure of prices charged by businesses. It is only practical to sample businesses at the first stage, rather than sampling prices directly, because this is the only frame information available. This leads to a multistage design, with the additional challenge that knowledge of which products an individual business makes is alsoneeded, so that the sample size of price quotes can be controlled at the product level. There are different ways to operationalise such a sampling procedure, depending on whether there is already a source which contains some information on the products originating from a business.In this paper we review the development and implementation of the UK’s PPI sampling strategy, which changed to a probabilistic design in the 1990s, and document the challenges in setting up a path to develop the CSO’s Wholesale Price Index to follow a similar probabilistic approach. The development of a probabilistic sample will be a long‐term project, because of challenges around the availability of product‐level data, and the response rates to business surveys in Ireland. The results provide an interesting case study of implementing a business price survey in a country with a smaller economy, notably including small numbers of large businesses concentrated in specific industries and thereforewith production of particular products.

More information

Identifiers

ORCID for Paul A. Smith: orcid.org/0000-0001-5337-2746

Catalogue record

Export record