世界银行-对全球风险的预测敏感性_BVAR分析（英）-2025_135页_87mb

Summary of "Forecast Sensitivity to Global Risks: A BVAR Analysis"

Core Content

This paper explores how global macroeconomic variables influence the economic outcomes of developing countries, using a Bayesian Vector Autoregression (BVAR) model to analyze the sensitivity of forecasts to global risks. The study is part of the World Bank's broader effort to enhance macroeconomic forecasting and policy analysis by incorporating uncertainty from external global shocks.

Main Points

• Global Risks and Forecast Uncertainty: Developing countries are exposed to global macroeconomic variables such as interest rates in high-income countries, commodity prices, global demand for exports, and remittance inflows. These variables are subject to both common global shocks and idiosyncratic fluctuations, which affect economic outcomes.

• Methodology: A BVAR model is used to capture the interdependencies among global macroeconomic variables. The model is applied to 115 developing countries using the World Bank's Macro-Poverty Outlook (MPO) forecasts as a baseline. The BVAR is calibrated using historical data from 2005 to 2023 and incorporates a multivariate normal distribution of shocks derived from the model's residuals.

• Simulation Process: The paper conducts 1,000 simulations for each country, drawing from the joint error distribution of the BVAR model. These simulations generate a distribution of potential outcomes for GDP, current account, inflation, and fiscal variables, allowing for a probabilistic assessment of future economic conditions.

• Impact of Global Shocks: The aggregate GDP outcomes across 115 countries show that global factors influence GDP levels by less than ±2% in most years, but by ±2% to ±4% in about 30% of the years. This indicates that while global shocks have an impact, they are not the dominant factor in GDP fluctuations for most developing countries.

• Key Transmission Channels: The paper identifies several channels through which global shocks affect the domestic economy:

  • Demand-Side: Changes in import and oil prices influence domestic inflation and consumer prices, especially in sectors reliant on imports.
  • Monetary Policy: Central banks monitor global price and economic indicators to adjust domestic monetary policy.
  • Exchange Rates: Exchange rate regimes (pegged, floating, mixed) affect the responsiveness of domestic currencies to global shocks.
  • Fiscal Balance: Global shocks can affect government spending and tax revenue, thereby influencing the fiscal balance.
  • Current Account Balance: Export/import prices, volumes, and exchange rates determine the current account balance, which is sensitive to global conditions.

Key Variables and Their Impact

The following variables are included in the BVAR model and have distinct impacts on the domestic economy:

• Import Price: Influences overall inflation and government revenues from import taxes.
• Export Price: Affects export volumes and government revenues from export taxes.
• Oil Price: Impacts both import and export prices, and is a key variable for oil-producing countries.
• Export Market Demand: Reflects the demand from major trading partners, affecting export volumes.
• US Interest Rate: Influences global borrowing costs and exchange rates.
• EU Interest Rate: For countries with a currency pegged to the Euro, this variable is critical.
• Remittances Inflow: Affects household disposable income and current account balance.

Applications and Results

• Simulation Process: For each of the 115 countries, the paper follows a three-step simulation process: 1) generating 1,000 shocks from the BVAR error distribution, 2) introducing these shocks into the macro-structural model (MFMod), and 3) analyzing the resulting distributions of GDP, inflation, fiscal balance, and current account outcomes.

• Sample Countries: Detailed results are shown for four sample economies:

  • Poland: A high trade openness economy.
  • Angola: An oil-producing economy.
  • The Gambia: An economy with significant remittance inflows.
  • Philippines: A relatively diversified economy with limited exposure to external shocks.

• Shock Correlations: The correlation matrix shows strong relationships between oil prices and export/import prices, as well as between global interest rates and export market growth. These correlations highlight the interconnectedness of global and domestic economic variables.

• Outcome Distributions: The paper presents forecast intervals for GDP, growth rate, inflation, fiscal balance, public debt, and current account balance. These distributions show the range of possible outcomes, with the median GDP growth around 3%, and a wide dispersion between the 10th and 90th percentiles.

Limitations and Future Work

• Data Constraints: The analysis is limited to the most recent two decades due to structural changes in the global economy post-2000s and data availability issues for earlier periods.
• Model Assumptions: The BVAR model assumes that shocks are random and not predictable, and thus cannot capture unforeseen events such as the U.S. mortgage crisis or the COVID-19 pandemic.
• Further Research: The paper suggests that future work could explore more detailed modeling of global interactions, better data sources for earlier periods, and improved methods for capturing non-linear relationships and structural breaks.

Conclusion

The study provides a framework for understanding the impact of global shocks on developing countries through a probabilistic approach. By integrating the BVAR error distribution into the World Bank's macro-structural model, it enables a more comprehensive assessment of forecast uncertainty, offering insights into the potential variability of economic outcomes. The results suggest that while global factors play a role, their influence on GDP levels is relatively modest in most years, with significant effects occurring in about 30% of cases.