Statistical analysis of revisions to Irish national accounts

Abstract

This paper examines the effects of data revisions on the major expenditure components in the Irish national accounts. The various tests used to establish their extent confirm that revisions affect levels more seriously than growth rates. A strong tendency is found in most series for preliminary estimates to understate the final (true) figures; and while most of the revision occurs in the first period, there is occasionally some over-correction in later periods. The immediate problems such revisions pose for the econometrician and policy maker are discussed.