Robin Youll’s research while affiliated with Office for National Statistics and other places

There has been considerable attention paid in recent months to Office for National Statistics early estimates of Gross Domestic Product (GDP). These estimates provide a timely indication of the evolution of the economic cycle, and there has been particular interest in their reliability, given the potential for theestimates to be revised.This article describes the context in which the offi cial estimates of GDP are produced, and examines the reason for revisions to early estimates. It goes on to discuss how these revisions can best be analysed over different maturities of the estimates and points out some potential pitfalls. Based on this discussion, itoffers guidance on the best approach to understanding the revisions cycle and for interpreting the published information on revisions.The article then presents the results of analysis of a new real time database for GDP growth in the UK, which contains a time series of estimates back to 1961. Over the period since the mid-1990s, revisions have been much smaller than previously, perhaps reflecting the relative stability of the economy over this time.The article concludes by discussing the issues surrounding the use of bias adjustments for GDP.

The Office for National Statistics set out its plans for the modernisation of the UK’s National Accounts in June 2008. Since then the programme of work which supports this development has been given sharper focus, following consultation in late 2008 with key stakeholders. This article sets out the background to the modernisation work, describes the current scope and benefits of the new programme, and its key milestones.

Considers why this has occurred and the ONS response to reducing the potential for such bias in futureHistorically, Office for National Statistics (ONS) early estimates of the growth of the volume of gross domestic product have tended to be revised up on average. This article considers why this has occurred and ONS's response to reducing the potential for such bias in future. It concludes that there is insufficient evidence for ONS to make an explicit aggregate level adjustment to these early estimates to anticipate potential future bias, and that there are many more fundamental reasons why it should not make such an adjustment. Instead ONS should continue to research the reasons for any bias, and seek to reduce or remove it altogether. Economic & Labour Market Review (2008) 2, 48–52; doi:10.1057/elmr.2008.105

Describes the above, resulting from concerns about quality of many detailed estimates and a wider reprioritisation of ONS's businessThe Index of Production (IoP) published on 10 March 2008 was based on improved methods, as described in an article published in the January 2008 edition of Economic & Labour Market Review. These methodological changes were made primarily in response to concerns in the Office for National Statistics (ONS) about the quality of many of the very detailed published seasonally adjusted estimates. The change was also a response to a wider reprioritisation of ONS's business which led to a 20 per cent reduction in the number of businesses sampled in the Monthly Production Inquiry, as used in the IoP. This article describes the impact of these methodological changes on the published results. Economic & Labour Market Review (2008) 2, 43–46; doi:10.1057/elmr.2008.89

Examines the relationship between the two measures, highlighting contributions of different products to differences in growth rates over timeThis article examines the relationship between the implied price deflator derived from the Retail Sales Index and the change in prices calculated from a comparable price index constructed using components of the Consumer Prices Index. A decomposition approach is used to highlight the contribution of different products to the difference in growth rates over time. Economic & Labour Market Review (2008) 2, 35–39; doi:10.1057/elmr.2008.74

Sets out the reasons for these changes and their impact on the levels of detail ONS will publishIn January 2008, as part of a wider reprioritisation of the Office for National Statistics? (ONS) business, the sample size for the Monthly Production Inquiry (MPI) will be reduced by 17 per cent. Of itself, the reduced sample size would lead to lower-quality estimates of the change in production output as published in the Index of Production (IoP) First Release. However, ONS will introduce a number of methodological changes in March 2008, including a reoptimisation of the MPI sample. These changes will maintain, and in some cases improve, the quality of the aggregate IoP indices. At the same time, the level of detail published for the IoP will be reduced, providing greater focus on the aggregate series. This article sets out the reasons for the methodological changes and their impact on the levels of detail which ONS will publish. Economic & Labour Market Review (2008) 2, 30–37; doi:10.1057/elmr.2008.7

Sets out the results of recent work on the accuracy of IoP estimates, based on an analysis of sampling error.In recent times the key measure of quality used for the Index of Production (IoP) has been the revisions performance of key aggregates. This is published as a set of revisions triangles alongside the monthly IoP release on the National Statistics website. Additionally, the IoP homepage on the website has a link to a Summary Quality Report which describes other aspects of quality of the series, for example, their timeliness, punctuality and relevance. This article sets out the results of recent further work on another dimension of the quality of these series: the accuracy of the estimates, based on an analysis of their sampling error. Additionally the article introduces an approach to defining quality bands for each series, to allow users to compare the relative quality of different IoP components. Economic & Labour Market Review (2007) [1], 24–28; doi:10.1057/palgrave.elmr.1410121