bmark gwu edu dave costello office of energy markets and end use emeu information administration us department 202 586 1468 e mail eia doe gov abstract the short term integrated forecasting system stifs generates monthly forecasts demand supply prices using some forecast that is incorporated into generated by other models do not run in an framework with this includes macroeconomic projections for certain variables there no direct feedback between components members team can attempt to coordinate iterations two however desirable main reasons first it suffers from specification problems richness complexity dynamic interactions because directly estimated second iteration process requires staff time resources are limited project tests experimental model interaction natural gas consumption industrial sector activity strategies followed involves a simple var capturing series dynamics testing granger causality examining impulse response functions