This study addresses for the first time systematic evaluation of a wid
ely used class of forecasts, regional economic forecasts. Ex ante regi
onal structural equation model forecasts are analysed for 19 metropoli
tan areas. One- to ten-quarter-ahead forecasts are considered and the
seven-year sample spans a complete business cycle. Counter to previous
speculation in the literature, (1) dependency on macroeconomic foreca
sting model inputs does not substantially erode accuracy relative to u
nivariate extrapolative methodologies and (2) stochastic time series m
odels do not on average, yield more accurate regional economic predict
ions than structural models. Similar to findings in other studies, cle
ar preferences among extrapolative methodologies do not emerge. Most g
eneral conclusions, however, are subject to caveats based on step-leng
th effects and region-specific effects.