Annotation Imposing cointegration on a forecasting system, if cointegration is present, is believed to improve long-horizon forecasts. Contrary to this belief, at long horizons nothing is lost by ignoring cointegration when the forecasts are evaluated using standard multivariate forecast accuracy measures. In fact, simple univariate Box-Jenkins forecasts are just as accurate. Our results highlight a potentially important deficiency of standard forecast accuracy measuresthey fail to value the maintenance of cointegrating relationships among variablesand we suggest alternatives that explicitly do so
Bibliography
Includes bibliographical references (pages 29-30)
Notes
Master and use copy. Digital master created according to Benchmark for Faithful Digital Reproductions of Monographs and Serials, Version 1. Digital Library Federation, December 2002. http://purl.oclc.org/DLF/benchrepro0212 MiAaHDL
Print version record
digitized 2010 HathiTrust Digital Library committed to preserve pda MiAaHDL