Recent studies of the effects of campaign spending by political parties and candidates at elections in Canada and elsewhere have established the importance of local constituency campaigns. However, particular claims to measure the effects of campaign spending on the vote have been questioned on methodological grounds. This article revisits the question of whether local spending matters in Canadian federal elections. Responding to some criticisms of earlier work, this analysis presents the results of two parallel regression analyses (the first employing two–stage least squares estimation, the second using three–stage least squares techniques) of the effects of local spending in the 1993 and 1997 elections. The results offer strong confirmation that comparatively greater local spending by candidates enhances their vote shares, and diminishes that of rivals, albeit to different degrees for different parties and elections.