短期课程：Bayesian VAR models and forecasting
时间：2019年4月8日-26日
主讲人：Professor Sune Karlsson,Örebro University, School ofBusiness
课程大纲：Vector autoregressions(VARs) have become the workhorse model for macroeconomic forecasting. Theinitial use in economics was to a large degree motived by Sims (1980) critiqueof the ”incrediblerestrictions” used by the large macroeconometric modelsdeveloped inthe 1970s and much effort was put into tools for policy analysisbased on VAR models.This role of the VAR model has to some degree been taken over bythe current crop ofDSGE models, a new generation of theory based models which are – at times – ill at easewith the data. The role of the VAR model as the baseline, serious,model for economicforecasting is, however, unchallenged. The popularity stems inpart from it’s relativesimplicity, flexibility and ability to fit the data but, ofcourse, also from it’s success as aforecasting device.
This short course surveys Bayesian approaches to inference in VAR models with afocus onforecasting. Oneimportant feature of the chapter is that it gathers many algorithms forsimulating from theposterior distribution of the parameters, some of which have not beenclearly statedpreviously. This provides the necessary tools for analyzing the posteriorand predictivedistributions and forecast with the models and priors that are studied inthe course.
时间 |
主题 |
地点 |
第一部分：Introduction to Bayesian VARs |
4月8日（星期一）10:00-12:35 |
Bayesian Vector autoregressive (VAR) models and forecasting |
沙河东区三号楼323 |
4月22日（星期一）10:00-12:35 |
Model selection and model averaging, Large Bayesian VARs |
沙河东区三号楼323 |
4月29日（星期一）10:00-12:35 |
Time varying parameters and stochastic volatility |
沙河东区三号楼323 |
第二部分：Advanced topics on Bayesian VARs |
4月12日（星期五）16:00-18:35 |
Bayesian Vector autoregressive (VAR) models and forecasting, modelselection (II) |
沙河东区主教301 |
4月26日（星期一）14:00-16:00 |
Large Bayesian VARs, time varyingparameters and stochasticvolatility (II) |
沙河东区主教104 |
参考书目
Karlsson,Sune. "Forecasting with Bayesian vector autoregression." In Handbookof economic forecasting, vol. 2, pp. 791-897. Elsevier, 2013.https://doi.org/10.1016/B978-0-444-62731-5.00015-4
主讲人简介：ProfessorSune Karlsson obtained his PhD in economics from Purdue University in 1989. Hewas appointed as Associate Professor in Stockholm School of Economics.Professor Karlsson was appointed as Professor of Statistics, Örebro UniversitySchool of Business. He served as Dean, Faculty board of Business, Science andEngineering in Örebro University.Professor SuneKarlsson’s research interests includes econometrics, economic forecasting andBayesian methodology. He has published peer-reviewd papers in top journals likeInternational Journal of Forecasting, Journal of Econometrics, ComputationalStatistics and Data Analysis, and Econometric Reviews.ProfessorSune Karlsson co-founded RePEc (Research Papers in Economics, http://repec.org/) inMay 1997 and participated in development of underlying protocols and datastructures. He also develops software implementing core RePEc functionalities,edits the nep-ecm mailing list which provides information on new econometricsworking papers in RePEc. Nep-ecm is part of New Economic Papers (NEP,http://nep.repec.org/) current awareness service and operates the EconPapersservice (http://econpapers.repec.org) which provides access to the RePEc dataset, the largest collection of on-line working papers and journal articles inEconomics.
由中央财经大学引智项目支持
