Assessing GDP and sub-component revisions

Gross Domestic Product (GDP) is a measure of economic activity published by Statistical Agencies (SAs). It is one of the key data series followed by Central Banks (CBs) to inform monetary policy decisions. Official estimates of economic growth serve as a basis on which CBs forecast future growth. They are regularly revised and therefore forecasts for GDP growth are done on the basis of ever-changing data. The revisions would ideally respond to the incorporation of additional and improved data over time. But is it really so? Users of data understand the uncertainty surrounding the early GDP announcements. In order to make a good use of them, they need to understand how reliable subsequent revisions of GDP data are. The economic literature has intensively studied the properties of GDP and other macroeconomic aggregate revisions. These revisions feed into the various forecasting models with significant implications.

However, it is much less known about the reasons why Statistical Agencies (SAs) decide to revise their estimates beyond the timeliness of their data collection. Our project sheds more light on this topic and hypothesizes that SAs also behave as risk managers who have an implicit interest (loss function) in not revising their GDP estimates after their initial releases while being much more open to revising GDP sub-components over time.

Our research is going to test the properties of GDP and their sub-component revisions. We aim to shed light on how Statistical Agencies behave when they produce their initial quarterly announcements and subsequent revisions. The outcome of this study will equip academics, policy-makers and professionals with better knowledge of how these estimates change over time and thus construct better and more accurate forecasts.