Business survey,which starts from the microeconomic level,is a widely used short-term forecasting tool in practice.In this study,the authors examine whether foreign trade survey data collected by China's Ministry of Commerce would provide reliable forecasts of China's foreign trade.The research procedure is designed from three perspectives including forecast information test,turning point forecast,and out-of-sample value forecast.First,Granger causality test detects whether survey data lead exports and imports.Second,business cycle analysis,a non-model based method,is per-formed.The authors construct composite indexes with business survey data to forecast turning points of foreign trade.Third,model-based numerical forecasting methods,including the Autoregressive Inte-grated Moving Average Model with Exogenous Variables(ARIMAX)and the artificial neural networks(ANNs)models are estimated.Empirical results show that survey data granger cause imports and ex-ports,the leading composite index provides signal for changes of trade cycles,and quantitative models including survey data generate more accurate forecasts than benchmark models.It is concluded that trade survey data has excellent predictive capabilities for imports and exports,which can offer some priorities for government policy-making and enterprise decision making.

Academy of Mathematics and Systems Science,Chinese Academy of Sciences,Beijing 100190,China; School of Economics and Management,University of Chinese Academy of Sciences,Beijing 100190,China; Center for Forecasting Science,Chinese Academy of Sciences,Beijing 100190,China

[1]BAI Yun;WANG Shouyang;ZHANG Xun-.Foreign Trade Survey Data:Do They Help in Forecasting Exports and Imports?)[J].系统科学与复杂性学报（英文版）,2022(05):1839-1862

Imports,microeconomic,granger

Foreign,Trade,Survey,Data,Do,They,Help,Forecasting,Exports,Business,survey,which,starts,from,level,widely,used,short,term,forecasting,tool,practice,this,study,authors,examine,whether,foreign,trade,data,collected,by,China,Ministry,Commerce,would,reliable,forecasts,research,procedure,designed,three,perspectives,including,information,test,turning,out,sample,value,First,Granger,causality,detects,exports,imports,Second,business,analysis,formed,construct,composite,indexes,points,Third,numerical,methods,Autoregressive,Inte,grated,Moving,Average,Model,Exogenous,Variables,ARIMAX,artificial,neural,networks,ANNs,models,are,estimated,Empirical,results,show,that,cause,leading,provides,signal,changes,cycles,quantitative,generate,more,accurate,than,benchmark,It,concluded,has,excellent,predictive,capabilities,can,offer,some,priorities,government,policy,making,enterprise,decision

0.611785