The World Bank's Unified Survey projections : how accurate are they? an ex-post evaluation of US91-US97

Since 1984, the Unified Survey has been the World Bank's principle mechanism for gathering quantitative macroeconomic information from country teams on Bank member countries. After gathering annual data those teams also do most-likely-scenario projections. The author examines the numerical projections of macroeconomic indicators carried out by World Bank country teams for Unified Surveys for fiscal years 1991-97. He studies the accuracy of short-term projects (for the current year, first year, and three years ahead) for 23 countries in the different World Bank regions. He also compares the Unified Survey projections with the International Monetary Fund's (IMF's) projections for its fall World Economic Outlook (WEO). He finds that: 1) The Unified Survey projections are inaccurate when evaluated over the whole period investigated (1990-96). However, their accuracy has improved over time. 2) Improvements are notable in projections for investment, GDP inflation, and government deficit. Projections of external indicators - such as import and export growth - are still substantially inaccurate and should be greatly improved. 3) The Unified Survey projections are as accurate as - or more accurate than - the WEO projections. 4) One cannot characterize the United Survey projections as optimistic. This is the first systematic attempt to evaluate the accuracy of country team macroeconomic projections over time and the first to compare these with the IMF's WEO projections.