2014年-IMF国际货币组织全球_Debt_and_Growth_Is_There_a_Magic_Threshold__19页_674kb

Summary of "Debt and Growth: Is There a Magic Threshold?"

Core Content

This paper investigates the relationship between government debt and economic growth, focusing on whether there exists a specific debt threshold that significantly impairs medium-term growth prospects. The authors, Andrea Pescatori, Damiano Sandri, and John Simon, use a novel empirical approach and a large dataset from the IMF Fiscal Affairs Department to analyze this question.

Main Viewpoints

• No Clear Debt Threshold: The paper finds no evidence of a specific debt-to-GDP threshold above which medium-term growth prospects are dramatically compromised. This contrasts with earlier studies such as Reinhart and Rogoff (2010), which suggested a threshold around 90% of GDP.
• Debt Trajectory Matters: The trajectory of debt (whether it is increasing or decreasing) is as important, if not more so, than the debt level in predicting future growth. Countries with high but declining debt have historically grown at similar rates to those with lower debt.
• Output Volatility: While high debt does not necessarily lead to lower growth, it is associated with higher output volatility. This suggests that even if debt does not directly reduce growth, it can introduce instability into the economy.
• Reverse Causality and Endogeneity Concerns: The paper acknowledges that reverse causality (i.e., weak growth causing high debt) and endogeneity issues remain challenges in interpreting the relationship. However, by analyzing longer-term data, the authors attempt to mitigate these concerns.

Key Findings

• The average debt-to-GDP ratio in the sample is 55%, with an average real output per capita growth rate of 2.25%.
• For the short run, countries with debt-to-GDP ratios above 90% experience lower GDP growth in the following year, with growth averaging around 2% for those below 90% and -2% for those above.
• Over medium-term horizons (5, 10, and 15 years), the negative association between debt and growth weakens. Countries with high debt on a declining trajectory show similar or better growth performance compared to those with lower debt.
• Relative growth performance (Figure 5) shows that countries with high debt but declining trajectories grow at a rate close to their peers, with differences less than 0.5% per year.
• GDP volatility (Figure 7) increases with higher debt levels, suggesting that while debt may not reduce growth directly, it can increase economic instability.

Methodology

• The authors use a comprehensive dataset of gross government debt-to-GDP ratios for nearly all IMF members from 1875 to 2011.
• They analyze the long-term relationship between debt and growth, focusing on debt trajectories rather than just levels.
• They exclude overlapping episodes and control for peer growth performance to reduce bias.
• They also adjust for reverse causality by examining the relative growth performance and volatility rather than absolute growth rates.

Robustness Checks

• The authors test the robustness of their findings by comparing episode growth with the average growth of all economies over the same period.
• They remove the first five years of each episode to reduce the influence of short-term reverse causality effects.
• The results remain consistent even after these adjustments, reinforcing the idea that debt levels alone are not a strong predictor of growth, but debt dynamics are.

Conclusion

The paper concludes that while there is no clear debt threshold that severely undermines growth, high debt can still be associated with increased output volatility. This implies that although debt may not be the direct cause of poor growth, it can contribute to economic instability. The authors caution against drawing strong policy implications due to endogeneity and reverse causality concerns, but their findings suggest that fiscal policy should be evaluated more holistically, considering both the level and trajectory of debt.

Figures Overview

1. Figure 1: Short-run growth performance following a debt threshold crossing (h = 1).
2. Figure 2: Medium-term growth performance (h = 5, 10, 15).
3. Figure 3: Debt dynamics over the medium term.
4. Figure 4: Growth performance based on debt trajectory.
5. Figure 5: Relative growth performance across debt levels.
6. Figure 6: Growth performance from 5 to 15 years after crossing a debt threshold.
7. Figure 7: GDP volatility during the 15 years after crossing a debt threshold.

Data Appendix

• The dataset includes real per capita GDP and debt-to-GDP ratios for a range of advanced economies.
• Data gaps are addressed by linear interpolation for GDP and debt during wartime periods.
• The data covers a long historical period, including events such as two world wars and the Great Depression, which may influence growth patterns.

References

• Key references include studies by Reinhart and Rogoff (2010, 2012), Herndon, Ash, and Pollin (2013), Kumar and Woo (2010), and Cecchetti et al (2011), among others. These provide the theoretical and empirical foundation for the current analysis.