对全球风险的预测敏感性：BVAR分析2025

11132 Produced by the Research Support TeamAbstractThe Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about developmentissues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry thenames of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely thoseof the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank andits affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent.Policy Research Working Paper11132Developing countries face uncertainties driven by globalmacroeconomic variables over which they have little to nocontrol. Key exogenous factors faced by most developingcountries include interest rates in high-income countries,commodity prices, global demand for exports, and remit-tance inflows. While these variables are sensitive to commonglobal shocks, they also exhibit idiosyncratic fluctuations.This paper employs a Bayesian Vector Autoregression modelto capture interdependencies of global variables and sim-ulates global risks using the empirical joint distribution ofglobal shock as captured by joint Bayesian Vector Autore-gression errors. The simulated shocks are then integratedThis paper is a product of the Economic Policy Global Department. It is part of a larger effort by the World Bank toprovide open access to its research and make a contribution to development policy discussions around the world. PolicyResearch Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contactedat hruberl@worldbank.org or rtercioglu@worldbank.org@worldbank.org. into the World Bank’s macro-structural model to assesshow a range of potential global disturbances could impacteconomic outcomes across countries. The methodologyis applied to 115 countries, using the World Bank’s fall2024 edition of the Macro-Poverty Outlook forecasts as abaseline. Although the individual country results are het-erogeneous, the aggregate distribution of gross domesticproduct outcomes across the 115 countries suggests thatglobal factors influence gross domestic product levels inindividual developing countries by less than plus or minus2 percent in most years, but by between 2 and 4 percent inabout 3 in 10 years. Forecast Sensitivity to Global Risks: A BVARAnalysis∗Heather Ruberl1, Remzi Baris Tercioglu2, Adam Elderfield1,2Economic Policy Global Department, World BankJEL Classification:C10, C50, E17Keywords:Economic modeling, forecasting, global risks, macroeconomic shocks, macroeconomic modeling andstatisticsWe thank Andrew Burns, Fernando Giuliano and Charl Jooste who provided valuable feedback and advice on this 1IntroductionMacroeconomic forecasting is an exercise undertaken in the face of considerable uncertainty. Indeed,forecasting can be seen as an exercise in reducing uncertainty – in part by exploiting empirical regular-ities in the way economies operate and in part by making artificial assumptions about what the state ofthe world may be. For developing countries, a key source of uncertainty lies in global macroeconomicconditions. Forecasts typically rest on assumptions about variables such as oil prices, trade policies ofmajor economies, and demand from key trading partners. Each of these reduces the range of possibleoutcomes, and helps forecasters develop a conditional projection of what economic outcomes mightlook like if these global assumptions are realized. Typically these projections are presented as pointforecasts, what the world will look like assuming these global conditions hold. Of course, in the realworld the assumptions are never realized exactly (and sometimes are very far from expectations).This paper proposes a methodology to move beyond these point estimates conditional on specificglobal assumptions by generating a range of outcomes around a given point estimate by runninghundreds of simulations each based on the underlying uncertainty of different external-to-the countryglobal economic variables. These include the world prices of oil and other commodities, monetarypolicy conditions in high-income countries, the level of import demand in the countries to which acountry exports, and the level of remittance inflows to the country (in many developing countriesthese remittances can be as high as 25% of GDP). A Bayesian Vector Autoregression (BVAR) acrossexternal variables allows controlling for the endogeneity between external variables (typically whenoil prices are high so are those of other commodities, while oil prices, monetary policy, and importsare sensitive to the economic cycle in high-income countries). By randomly drawing from thedistribution of historical errors of the BVAR, an infinite range of potential outcomes can