a methodology change in the early years?
Good catch.
The Montréal series starts in 1998 while most others begin in 1995, which creates an apparent gap at the beginning of the comparison period. Rather than excluding early years, should we consider whether the Montréal definition was different in those years, or whether we're measuring something that simply didn't exist before the 1998 institutional reform?
Hugo Kobayashi The trend looks right.