December 5, 2007

CBC BC Almanac

Hello Mark Forsyth,

Regarding your conversation, December 4, 2007, with the Director of BC Stats, Don McRae, it is important to realize that BC Stats does not work closely with Statistics Canada on Census releases nor on population change estimates including migration.

In fact, Statistics Canada warns users of their Annual Demographics release (the Demographic Estimates Compendium http://www.statcan.ca/english/ads/91-213-SCB/index.htm) that BC’s numbers do not add up, because BC has it’s own statistical agency that uses it’s own methods for estimating population.

"With the exception of British Columbia, preliminary estimates for July 1, 2006 were produced using the component method. The population estimates for this province were created by its respective agency. As a result, the sum of components does not equal the population growth for 2005-2006."

BC Stats uses the change in hydro and telephone (landline) hookups to estimate population change. Unfortunately, the use of Telus data drives down the population in urban centres, particularly the GVRD, and particularly, Richmond and Burnaby. Mr. McRae wanted telephone hookups to be used as an indicator of population change because, as he said "two indicators look better than one". When the GVRD complained to BC Stats about the low numbers, I was told, as the provincial expert in migration to the sub-provincial level, to remove people from high growth areas such as Comox/Courtenay and Kelowna and put them into the Lower Mainland. I refused and was subjected to a constructive dismissal.

Also, the Ministry of Finance arbitrarily lowered the number of migrants expected to come to the province by 7,000 people. I was told, net migration would still be positive and the government would look good if more people arrived than expected. Unfortunately for Health and Education planner’s decisions are made based on the unreliable population numbers provided by BC Stats. People attending meetings about opening or closing facilities should be aware of the strengths and weaknesses of the population methods.

Dubious methods are used to create population numbers for various boundaries including Provincial Electoral Districts (PEDs). I was told to review the results of the method used to estimate and forecast PED population and found that all of the lower mainland suffered a decline in population since 1996, which was obviously incorrect. I corrected the method but this was rejected and instead population numbers were made up, typed in by hand rather than using a statistical model. I informed Mr. McRae of the non-statistical method used but my concern was not addressed, nor was it addressed by the Deputy Minister whom I also informed. Instead, as mentioned, I was subjected to a constructive dismissal.

BC Stats divides municipalities into two groups based on population change between 1996 and 2001. This method was implemented without proper testing. Municipal population estimates are used to establish insurance rates. Statistics Canada is currently working towards estimating municipal population within BC annually. British Columbians will be paying twice for two different sets of population estimates, but at least one set will be reliable. If BC Stats population section did not exist, not only would costs be cut, but British Columbians would already be using population numbers generated using proper statistical methods by Statistics Canada.

There are many more examples of problems in BC Stats including the mismanagement of data and the use of perks and intimidation to coerce analysts. British Columbians would benefit from an open debate on the reliability of the population numbers created by BC Stats.


Sincerely,


Warren Munroe

Return to British Columbia Population Data Warning