Wait, I don't get it. Why are published population numbers such as the total population of Canada (over 34 million versus 33.5 million), Toronto etc. much higher than the numbers presently released by statistics canada even though the growth rates 2006-2011 exceed those of 2001-2006?
Intercensal estimates are always higher than official preliminary census releases. The census just makes a physical count (well, it goes to like 99% or whatever, and then extrapolates the final 1% just so enumerators don't spend six months arguing with people to get the last two houses on a street to fill out their form). The off-year estimates take uncountable people into account like the homeless or unregistered illegal immigrants, as well as people like students or migrating workers that are hard to situate as of Census Day, etc. They also look at housing starts and completions, utility connections, and so on, so they are invariably higher than the first census numbers - it's a complete estimate, not just a static count of countable people on a certain day.
Anyway, by basing estimates partially off dwellings, though, you run into the problem of overestimating how many people actually live in them. Most of the condo units downtown have a low people-per-household rate, yet are typically assumed to have the city average rate. It also doesn't account for declines like divorces, deaths, kids moving out, etc. So for a few years the city has been assuming that it got shafted in the 2006 census when the 'official' growth was lower than expected. "But we have all these condos!" Yeah, but there's only like 1.2 people living in each of them.
Meanwhile, a block away, a house broken up into 4 apartments was renovated back into one dwelling for a couple without kids. The single mother of two living on one floor of the house maybe upped and moved in with her sister in a 905 house. The old city of Toronto lost people between 2001 and 2006 despite a condo boom. The Double Cohort probably didn't help the city's numbers either since twice as many young people left high school at once and began scattering - how many of them ended up in Waterloo or Calgary or Markham? There's lots of little factors, some of which you can get a peek at by looking at data for census tracts or lower.
Come 2011, the total number of condos bumps growth into positive territory through sheer brute force. Also, that Gen X couple that renovated the house might finally have a kid now. As long as Echo Boomers have kids before their grandparents die, the numbers will creep forward.
edit - oh, and I forgot to add that the preliminary census releases will be adjusted eventually to try to capture some of the gap between who was counted (the census numbers), and who should have been counted (the running estimates). But by the time they've done this, the preliminary numbers will be published and will have become canon (possibly being used in projecting tax revenues for the next few years, redistributing electoral ridings, changing "Welcome to X-ville, population Y,Z00" signs, etc., etc.?). The adjusted numbers are kind of useless because everyone's focus has moved to the preliminary figures for the *next* census, leaving the adjusted numbers good for little else than tweaking the running estimates two or three years down the road.
There's always going to be a gap because we're increasingly compelled to generate excessive information and fuss over keeping it updated, though I wonder if our government wouldn't prefer to abandon the quinquennial census completely and just keep the population clock:
http://www.statcan.gc.ca/ig-gi/pop-ca-eng.htm