Analysis of Revisions to Quarterly GDP-A Real-Time Database

Article excerpt

Gross Domestic Product (GDP) is one of the key data series used by the Reserve Bank to inform monetary policy decisions. The measures of GDP, published by Statistics New Zealand (SNZ), are estimates rather than exact figures and may be revised in subsequent releases. Analysis of the most recent measures of GDP should incorporate the extent of uncertainty that surrounds these estimates. To enable a more detailed examination of revision patterns, the Reserve Bank has constructed a real-time database containing each quarterly release of Expenditure GDP (GDP(E)) and its components. The database is available to users on the Reserve Bank's website and will be regularly updated. This article provides an introduction to the database and, by way of example, presents a basic analysis of the revisions made to GDP(E) and its components.

1 Introduction

Gross Domestic Product (GDP) is a measure of the value of economic activity within a country over a given period. SNZ publishes measures of GDP for New Zealand on both a quarterly and annual basis. As it is neither possible to observe all forms of economic activity nor calculate their values precisely, the published measures of GDP are estimates rather than exact figures.

The Reserve Bank uses the measures of quarterly GDP both to gauge the current pace of economic growth and as a basis on which to forecast future growth. These measures may be substantially revised by SNZ in later quarters to incorporate additional and improved data. It is helpful to have some idea of the extent to which the latest GDP measures are reliable indicators of the revised and more accurate measures that will come to be associated with the quarters in later years.

The Reserve Bank has formed a 'real-time' database that contains each quarterly release of Expenditure GDP (GDP(E)) and its components published by SNZ. (2,3) The first constant price measures of GDP(E) were released in June 1990 and the first current price measures were released in July 1994.

For a selected series, the database provides the complete set of estimates that have been associated with each quarter. The database can be used to analyse the size and dispersion of the revisions made to each series, the results of which may assist in analysing the latest unrevised data. (4)

It is important to recognise that the revisions analysis undertaken in this article provides an assessment of the data's reliability, but not of its accuracy. The IMF's Data Quality Assessment Framework (DQAF) distinguishes the two concepts as follows (Carson and Laliberte, 2002, p.4):

* Accuracy refers to the closeness of the estimated value to the (unknown) true value that the statistic is intended to measure. In practical terms, there is no single overall measure of accuracy; accuracy is evaluated in terms of the potential sources of error.

* Reliability refers to the closeness of the initial estimated value to the subsequent estimated values. Assessing reliability involves comparing estimates over time. Data that are revised more frequently are not necessarily less accurate.

2 Measures of Gross Domestic Product

GDP can be measured in three ways, by estimating the value of the production, expenditure or income of an economy. SNZ publishes annual estimates of each measure and quarterly estimates of the production and expenditure measures. (5) The three measures are defined by SNZ as follows: (6)

Production GDP

This approach measures the value added by producers and is calculated by deducting the value of goods and services used up in production from the total value of goods and services produced. Production GDP is SNZ's headline measure of GDP and is released with a breakdown of value added by industry.

Expenditure GDP (GDP(E))

This technique directly calculates the value of goods and services produced for final use by measuring consumer purchases. …