Revisions to Macroeconomic Data: Ireland and the OECD

Abstract

This paper examines revisions to Irish quarterly macroeconomic data focusing on growth rates of real GDP as well as the main expenditure components. A real-time database is constructed from the Central Statistics Office’s Quarterly National Accounts release. This is used to measure the extent of data revisions while also formally testing the presence of statistical bias in Irish data. Although estimates of GDP are found to be unbiased, the same cannot be said for some of the subcomponents, most notably investment spending. Using data from the OECD, we compile an additional data set for 25 OECD economies to assess the relative scale of revisions to Irish data. Finding that revisions to Irish real GDP growth rates are among the largest observed – even when allowing for differences in growth rates – we examine a number of factors that may indicate why cross-country differences emerge. The scale of data revisions, particularly in Ireland, appears to bear some relation to the structure of the traded sector. More generally the paper also highlights the potential for weaker forecasting performances in countries where data revisions are likely to be relatively high.