2023-07-12-IMF-主权财富基金会降低财政政策的顺周期性吗_使用非参数方法的新证据_31页_2mb

Summary of IMF Working Paper WP/23/133

Paper Title

Do Sovereign Wealth Funds Reduce Fiscal Policy Pro-cyclicality? New Evidence Using a Non-Parametric Approach

Authors

Ali J. Al-Sadiq and Diego Alejandro Gutiérrez

Date

June 2023

Key Objectives

• To assess whether stabilization Sovereign Wealth Funds (SWFs) reduce fiscal policy pro-cyclicality.

Approach

• Uses an unbalanced panel dataset for 182 countries from 1980-2019.
• Employs two econometric methods to address the self-selection (endogeneity) problem:
  1. Regression-based treatment effect model.
  2. Propensity Score Matching (PSM) using a non-parametric approach.

Methodology

• Measures discretionary fiscal policy volatility through changes in government consumption.
• Considers commodity price volatility as a key driver of fiscal pro-cyclicality.
• Accounts for selection bias (non-random SWF establishment).

Key Findings

• Stabilization SWFs are associated with reduced fiscal policy volatility.
• Countries with stabilization SWFs experienced:
  • About 14% lower volatility in government consumption compared to non-SWF countries.
• The results are robust across different model specifications and econometric methods.

Implications

• Stabilization SWFs can help smooth budget resources and mitigate the impact of commodity price shocks.
• Fiscal policy volatility is still influenced by other factors (institutions, macroeconomic stability), not just SWFs.

Contribution to Literature

• The study provides empirical evidence supporting the role of SWFs in smoothing fiscal policy.
• Highlights the methodological importance of addressing endogeneity when evaluating SWF effects.

Data/Variables Used

• Countries: 182 advanced/developing countries.
• Period: 1980-2019.
• Outcome: Volatility of discretionary fiscal policy (measured via government consumption).
• Treatment: Dummy for stabilization SWF presence.
• Controls: Real GDP, inflation, government size, openness, commodity terms of trade, institutions, etc.