Motivation: Statistics Canada reports population projections by age for Canada, but at an annual frequency with a mid-year benchmark (1 July). Our prior quarterly estimation method did not align with this.
Old methodology
For age brackets, forward-fill the annual estimate of population by five-year age bracket using the corresponding Statistics Canada's population projection. Sum to 10-year and broad brackets. Convert from annual to quarterly by assigning each annual value to Q2 and interpolating over the remaining periods.
For all-ages population, benchmark to total population estimated and break-adjusted by Moody's Analytics using Statistics Canada's quarterly population. Convert from annual to quarterly using the method above. Forward-fill our population estimate.
To ensure additivity, squeeze using a top-down approach.
New methodology: Convert from annual to quarterly by assigning each annual value to Q3 and interpolating over the remaining periods. Convert quarterly to daily as START, i.e., by assigning each quarter's value to the first day and interpolating over the remaining periods. Convert daily back to quarterly by END, i.e., assign the value of the last day of each quarter to the whole quarter.
What you will see: All periods are revised. For the 10-year and broad brackets, one period of history has been lost; the new start date is 1971Q3. The all-ages projection start date is unchanged at 1946Q1.
The series reside in the historical catalog (Canada » Demographics » Population » Moody's Analytics Estimates » Population projection by age) and include, for example:
- RJPOPQ.ICAN = Population - All ages, (Ths. #, SA)
- RJPOP1564Q.ICAN = Ages 15 to 64 [broad bracket]
- RJPOP2534Q.ICAN = Ages 25 to 34 [10-year bracket]
Because catalog locations are subject to change, the upper-right search box on DataBuffet.com provides a "find in catalog" mode that accepts a mnemonic.
See also
|