I regress real GDP growth rates on the IMF’s growth forecasts and find that IMF forecasts behave similarly to those generated by overfitted models, placing too much weight on observable predictors and underestimating the forces of mean reversion. I identify several such variables that explain forecasts well but are not predictors of actual growth. I show that, at long horizons, IMF forecasts are little better than a forecasting rule that uses no information other than the historical global sample average growth rate (i.e., a constant). Given the large noise component in forecasts, particularly at longer horizons, the paper calls into question the usefulness of judgment-based medium and long-run forecasts for policy analysis, including for debt sustainability assessments, and points to statistical methods to improve forecast accuracy by taking into account the risk of overfitting.
and, potentially, to poor policy design.
This paper has explored the trade-off for IMFgrowthprojections, which are largely based on human judgment, and has found strong symptoms of overfitting, particularly at longer horizons. Forecasters incorporate too much information that is irrelevant at the margin. The fact that, particularly for low-income countries, judgment based long-term forecasts are not significantly more accurate than forecasts that ignore any information beyond the historical average has important implications for policy analysis. The level of
Sustainability Framework for Low-Income Countries
2. Examples of the Use of the Flexibility Inherent in the DSF
3. IDA’s Non Concessional Borrowing Policy (NCBP) Country examples
4. Asset-Creating Nature of Debt-Financed Public Investment
5. Assessing IMFGrowthProjections in Countries with High Levels of Public Investment
6. Assessing the Growth Dividend of Public Investment in the DRC
7. Have remittances been incorporated in DSAs when they are large?
8. Impact of CPIA Fluctuations on the Risk of Debt Distress Classification
9. Analytical Underpinnings of the
welcome the second upward revision of the IMFgrowthprojection, from 4% to 4.8% for end-2022. The authorities also consider that the growth potential remains elevated in the coming years, based on gross fixed capital formation, mainly through an increased absorption of European funds, which would help develop local infrastructure and create job opportunities.
The expected inflows of EU funds come from three channels: the 2014–2020 structural funds (which can be spent by end-2023), the 2021–2027 structural funds, and the RRF. As the report clearly
Mr. Charalambos Christofides, Mr. Atish R. Ghosh, Ms. Uma Ramakrishnan, Mr. Alun H. Thomas, Ms. Laura Papi, Mr. Juan Zalduendo, and Mr. Jun I Kim
help identify specific growth bottlenecks facing a country.
Box 3.6 .
Medium-Term Growth Projections Using Cross-Country Growth Models
IMFgrowthprojections tend to be overly optimistic as documented in a number of reports and forums; for example, the PRSP/PRGF reviews and PRSP progress report, the IEO’s reports on prolonged users of IMF resources and on fiscal adjustment in IMF-supported programs, and reports on IMF projections prepared outside the institution (such as the U.S. General Accounting Office (2003) and the Heritage Foundation (see Beech and
International Monetary Fund. Middle East and Central Asia Dept.
where unemployment is barely responsive to GDP growth like some GCC economies), growth would need to reach close to 10 percent ( Figure 2.12 , panel 1)—to reduce unemployment, growth would need to exceed this threshold. Similarly, in some CCA countries, unemployment-stabilizing growth would need to rise above 8 percent. Such growth rates are much higher than what was realized on average over the pre-pandemic period and, for many countries, exceed IMFgrowthprojections, which implies that if these projections materialize, they will be insufficient to create enough
Ms. Olessia Korbut, Mr. Gonzalo Salinas, and Cheikh A. Gueye
lowest growth rates after becoming stable/liberalized 17 ) demonstrating that the sound performance of stable/liberalized SSA economies is not determined by only a few countries. We reach a similar conclusion in the third column, which shows average growth rates since 1995, thus excluding previous years in which only a few SSA countries (Botswana, Ghana, Mali, Mauritius, Uganda) had stabilized/liberalized . The fourth column presents the averages by country group including IMFgrowthprojections up to 2014, showing that the superior performance of SSA stabilizers
10. The IMF’s growth forecasts in past years have been below those of the authorities, as well as the eventual outcomes . While the main reason appears to have been the expectation that the growth costs of excessive government intervention and strongly expansionary policies would manifest themselves imminently, other factors have also played a role. First, IMFgrowthprojections were also lower in several other CIS countries, including Russia and Ukraine in recent years. Second, Belarus’s close ties with Russia—notably, Russian support
Default Premium ”, mimeo , Indiana University .
Hellwig , Klaus-Peter ( 2018 ), “ Overfitting in Judgment-based Economic Forecasts: The Case of IMFGrowthProjections ”, IMF Working Paper No. 18/260 .
IMF ( 2020 ), “ Review of the Debt Sustainability Framework for Market Access Countries ”, forthcoming.
Krishnamurthy , Arvind and Tyler Muir ( 2017 ), “ How Credit Cycles Across a Financial Crisis ”, NBER Working Paper No. 23850 .
Krugman , Paul R. ( 1988 ), “ Market-Based Debt-Reduction Schemes ”, NBER Working Paper No. 2587