## Can the IMF's Medium-Term Growth Projections Be Improved?

_IMF Working Papers, October 1, 2004_

## Source details

**Canonical URL:** [Can the IMF's Medium-Term Growth Projections Be Improved?](https://www.imf.org/en/publications/wp/issues/2016/12/31/can-the-imf-s-medium-term-growth-projections-be-improved-17601)

## Other formats

- [Markdown version](/en/publications/wp/issues/2016/12/31/can-the-imf-s-medium-term-growth-projections-be-improved-17601/index.md)
- [Structured JSON version](/en/publications/wp/issues/2016/12/31/can-the-imf-s-medium-term-growth-projections-be-improved-17601/index.json)
- [Bundle manifest](/en/publications/wp/issues/2016/12/31/can-the-imf-s-medium-term-growth-projections-be-improved-17601/bundle-manifest.json)

## Bibliographic details
- Authors: Juan Zalduendo, Catia Batista
- Published: October 1, 2004
- Series: IMF Working Papers
- DOI: https://doi.org/10.5089/9781451874488.001

---

### Summary and main findings
- Numerous reports have noted that the IMF's medium-term growth projections are overly optimistic.
- The authors estimate a growth model and examine its out-of-sample forecasting performance relative to IMF projections.
- The model's projections outperform those of the IMF:
  - Projections are less biased: one-quarter of the IMF bias.
  - Projections have smaller standard errors: 20 percent lower root mean squared errors.
- Outperformance holds in all regions and among most income groups.
- Results persist even after controlling for the IMF's macroeconomic assumptions.
- The paper does not attempt to address broader criticisms of the empirical growth literature, but suggests benefits from systematic analysis of cross-country information.

### Methodology (as described)
- Estimation of a growth model constructed to produce out-of-sample projections.
- Comparative evaluation of model projections versus IMF staff medium-term projections.
- Controls included for IMF macroeconomic assumptions when assessing relative performance.

### Performance metrics and quantitative results
- Bias comparison: model bias equals one-quarter of the IMF bias.
- Precision comparison: model root mean squared errors are 20 percent lower than IMF projections.
- Geographic and income coverage: outperformance documented in all regions and among most income groups.

### Policy implications and interpretation
- Systematic, cross-country empirical analysis can improve medium-term growth projections relative to current IMF staff projections.
- Incorporating model-based forecasting methods could reduce optimism bias and improve projection precision in IMF medium-term work.

---

## Content in this bundle

- **Wp04203**
  - [Wp04203 (Markdown version)](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04203.pdf.md){rel="alternate" type="text/markdown"}
  - [Wp04203 (PDF)](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04203.pdf){rel="external" type="application/pdf"}

---

_Source: https://www.imf.org/en/publications/wp/issues/2016/12/31/can-the-imf-s-medium-term-growth-projections-be-improved-17601_