Smith, J.

forecast model setar interval evaluation density forecasting journal setar model clement regime table error experiment statistic number testing sample parameter setar models ar model level value performance point non-linearity hansen threshold density forecasts intercept power p-value process coverage rejection normality result series period independence 0.000 alternative rejection rates diebold point forecasts gaussian heteroscedasticity wiley potter forecast accuracy forecast evaluation techniques change degree forecast performance smith copyright accuracy method 359–375 comparison 0.062 point forecast performance regime-speci fic heteroscedasticity .062 0.070 0.3 0.066 0.001 0.002 .080 0.042 0.060 0.080 0.054 slope fi ndings 1- step forecasts erasmus university rotterdam .058 uncond