2013年-IMF国际货币组织全球_GPM6_The_Global_Projection_Model_with_6_Regions_79页_1mb

Summary of GPM6 - The Global Projection Model with 6 Regions

Core Content

The Global Projection Model (GPM6) is a small quarterly macroeconomic model that incorporates real and financial linkages across six major regions: the United States, the euro area, Japan, emerging Asia, Latin America (LA5), and a remaining countries grouping. This model is designed to improve the forecasting and policy simulation capabilities of economists and central banks by providing a more comprehensive and coherent view of global economic dynamics.

GPM6 is part of a series of studies aimed at developing a full global projection model, with each paper extending the model to include more regions and financial variables. The sixth paper in the series extends the previously estimated G3 (US, euro area, Japan) model by adding three more regions, thus covering approximately 85% of global GDP. The model is estimated using Bayesian techniques, which allow for the incorporation of prior judgments and improve the forecasting properties and dynamic consistency of the model.

Main Views

• GPM6 is an intermediate model between fully micro-founded DSGE models and purely time series models, combining theory, calibration, and Bayesian estimation.
• It is designed to be easy to use and comprehensible, allowing for internal consistency in projections and facilitating policy simulations.
• The model includes real and financial linkages, making it more suitable for analyzing cross-country spillovers and global shocks.
• It allows for "what if" scenarios and provides confidence intervals around forecasts, enhancing the robustness of the model.

Key Information

1. Model Structure

• The model is a small quarterly projection model with 6 regions.
• It includes 5 observable variables for each G3 economy: real GDP, unemployment rate, CPI inflation, policy interest rate, and exchange rate.
• For emerging economies, the unemployment rate is not used due to data limitations.
• The model includes stochastic processes for variables such as potential output, NAIRU, equilibrium real interest rate, and real exchange rate.

2. Stochastic Processes

• Potential Output: Defined as the sum of its lagged value, a quarterly growth rate, and a disturbance term. It allows for both permanent level shocks and persistent growth shocks.
• NAIRU: The equilibrium unemployment rate is modeled with a growth term and a disturbance term, capturing both level and growth shocks.
• Equilibrium Real Interest Rate: A function of the steady-state rate and a disturbance term, allowing for stochastic divergence from the steady state.
• Real Exchange Rate: Defined as the log of the nominal exchange rate adjusted for inflation differentials between the US and the country. It includes trend appreciation for emerging economies.

3. Behavioral Equations for G3 Economies

• Output Gap: Influenced by domestic effects (lagged output gap, medium-term real interest rate), external effects (real exchange rate gaps, trade spillovers), and financial-real linkages.
• Inflation: Modelled as a function of domestic and external factors, including the CPI inflation rate and exchange rate movements.
• Policy Interest Rate: A function of the medium-term interest rate and inflation expectations.
• Medium-term Interest Rate: A function of the policy interest rate and financial conditions.
• Uncovered Interest Parity: Links interest rates between countries and the exchange rate.
• Unemployment Rate: Modelled as a function of the output gap and structural shocks.

4. Behavioral Equations for Emerging Economies

• Output Gap: Similar to the G3 model but with different coefficients and no unemployment rate due to data constraints.
• Uncovered Interest Parity: Adjusted for trend appreciation in the real exchange rate.
• Unemployment Rate: Not used for emerging economies.

5. Model Estimation and Results

• The model is estimated using Bayesian methods, which help in imposing restrictions and generating plausible impulse response functions.
• Calibration and estimation are used to derive coefficients for the additional regions.
• The model includes global demand shocks and common demand shocks across multiple countries.
• It provides root mean squared errors, variance decompositions, and impulse response functions to evaluate its performance.

6. Model Applications

• GPM6 is intended for use by central banks and IMF country desks.
• It improves multilateral consistency in the World Economic Outlook (WEO) projections.
• It allows for alternative policy simulations and confidence intervals around forecasts.
• A companion guide (GPM+) is being developed to help other countries integrate their models into GPM6.

Conclusion

GPM6 represents a significant step towards a truly global macroeconomic model by incorporating six major regions and financial-real linkages. It balances economic theory with empirical calibration, enabling forecasting and policy analysis in a multi-country context. The model is expected to enhance the accuracy and consistency of global economic projections and support international policy coordination.