Friday, February 01, 2008

When Five Percent is More Than Half

"Canadian artists also released more music in 2005 compared with the previous survey year. New releases of musical recordings by Canadian artists rose 8.8% to 521 releases. On the other hand, the overall number of new releases (including those by non-Canadian artists) fell 5.6% to just over 3,900 releases."
-Statistics Canada, "New Annual Survey of Sound Recording and Music Publishing, The Daily, November 7, 2007

Nice, new Canadian artist releases are up 8.8%.

Too bad that around 400(?) of the new Canadian artist releases are missing. That close to 1700(?) new recordings are missing elsewhere in terms of new releases of all stripes.

Tragically, its the small "Canadian controlled" sector missing.

This is the small Canadian music recording producers and acts with their own labels. It includes small new acts or former bigger names who may be taking control, end to end, from their label. It includes producers who also take Indie records from other countries and promote them here to help cover their costs of pushing out new Canadian Indie recordings.

The huge Canadian controlled segment of small producers is missing. The big tail of the long tail, is not counted.

Why do I say this?

The small players were not included in Statistics Canada's 2007 publication!

This statistical and maybe very political story started in October 2005 when Statistics Canada published their 'occasional' Sound Recording Industry survey for the 2003 year. Then for the first time in "recent" memory, new Canadian artist releases in Canada had fallen below the 1,000 mark, to 904 (total new releases were 5,619).

This was big news at the time, hitting the MSM. There was the commentary of the day that varied from Canada was an infringing nation and this showed that we needed new stronger Copyright law, to the large role of the Canadian controlled industry in breaking Canadian acts had, raising the opposite doubt on whether any such strong law was needed anytime soon.

Two years and a month go by, and in the November 2007 Survey new Canadian artist releases hit 521, a dramatic fall from the former comfort level of "1,000" releases but oddly up by 8.8% from the "previous survey year." The "previous survey year" in this 2007 report was 2003. New Canadian artist releases for that year are shown in the survey as 479.

2005 Survey, Census type, 2003 year, 904 new Canadian artist releases

2007 Survey, Sample type, 2003 year, 479 new Canadian artist releases.

2005 Survey, Census type, 2003 year, 5,619 new releases.

2007 Survey, Sample type, 2003 year, 3,903 new releases.
What happened to the 425 New Canadian artist releases now missing from the 2003 year, in the report published in 2007? (1,716 new releases of all stripes?)

And what does it say about 2007 reports' of such releases at just 571 in 2005?

And the soon to come 2006 year "survey"?

StatCan did change their methods, upon reading their advisories and methodology, significantly, a) from a census of every firm to surveying by a sample of 95% of the "industry" by revenues, b) by expanded the "industry" definition to include publishers (plus 2 others), and c) by estimating the 5% unspecified by their sampling to project from "administrative data."

They caution boldly and rightfully:
"Data for 2005 should not be compared with previously published data on sound recording due to significant changes in coverage and methodology in the new survey."
But their new methods "dropped" the lowest 5% by revenue because it seemed to be hard - read expensive - to do otherwise. Read it yourself. Its very lame.

Who is in the lowest 5% group of total recording industry revenues, including now publisher and sound recording studio revenue?

This segment of the industry is the Canadian controlled sector from the bottom of the revenue pile on up, is a very large number of very small companies. They each do work that was not apparently in "administrative data" but that was included in the "old" 2003 year industry census type data.

The industry census type of survey published in 2005, for 2003 showed the Canadian controlled sector had just 15.3% of recording industry revenues, yet it had an 88.9% share of the new Canadian artist releases, 804 in number. In comparison, the foreign controlled companies had just 100 new Canadian artist releases, 2.7% of all their releases, yet 84.7% of industry revenue.

In 2003, there were 287 companies producing sound recordings controlled by Canadians, per the industry census, including Nettwork and 5 other Canadian companies who were members in the CRIA at the time. The CRIA in fact did claim 95% of the industry in 2005.

So what happened to the output of the other 281 Canadian controlled record companies, now missing in the 2007 publication for this 2005 year?

In 2003, the 287 Canadian controlled companies including Nettwerk Records, True North Records, Aquarius Records, the Children's Group, Anthem Records and Linus Entertainment produced 804 new Canadian artist releases. The "missing" 281 produced 425 such recordings in 2003 that were not in the "administrative data" per the 2007 StatCan publication (unless any of those companies were remaining in the CRIA: there is no evidence of that to be found).

What did the 281 companies in the Canadian recording industry not included in your reports, the vast majority of the Canadian companies in this business, do in 2005, Statistics Canada? HUH?

And what will be the impact of their omission in the future?

Is it going to be something like this in the near future?
"File sharing is way up. See, the disaster Statistics Canada shows. There used to be over 1,000 new Canadian artist release a year. We need stronger protection."
Here are some pictures, one prepared from the latest Statistics Canada "2007" report, and the second one carefully prepared from the "census type" Statistics Canada 2005 report, from work I published here at Northworthy (click on them for better views).

I have more than a few questions I want answered in addition to the ones asked above.

Why is the Canadian controlled company and the non-Canadian controlled company data no longer released by Statistics Canada? Was it a top down decision? Was it a political funding decision?

If it was important to know before, and it still appears to be in the national interest to know this type of information, to know the health of Canadian industry and its development, how is the Canadian controlled recording industry actually doing these days, Statistics Canada? The 6 CRIA defectors accounted for about 2/3 of the Canadian company share of revenue and 379 new Canadian artist releases in 2003 (thanks to your new survey information). The 281 other Canadian controlled companies account for 1/3 of that revenue and 425 new Canadian artist releases in 2003. This bottom 1/3 does not matter anymore? They produce more new Canadian artist releases than the companies you, Statistics Canada, now includes in your survey.

Now I am not intimate with how these things are done but it is clear the former census type of data was available to the makers of the 2007 report; the prior survey report was certainly. How could Statistics Canada publish something so off the marks established by the prior industry census?

In 2003, the non-Canadian controlled sector published just 100 new Canadian artist releases. Did they get to a 100 in 2005? The trend (202 in 1998, 166 in 2000) would suggest about 76 in 2005. Maybe a bit less. Maybe a bit more. Should we not know what this foreign segment, actually does for Canadian artists, when likely close to only 2% of their new releases in 2005 in our markets were by Canadian artists? Why should we even be giving them notice on their pushes for a political stake, for Canadian Copyright reform or for anything else that might matter to Canadians?

Yes Statistics Canada might be right on with respect to the revenues and profits of the group of companies in the top 95% and their production. But they exclude the results of 281 Canadian companies. Is the Department of Canadian Heritage no longer interested in the work of all the Canadian controlled companies in this industry?

The Chief Statistician should be embarrassed as much as I am as a Canadian for this apparently very shoddy, cheap work. I do not care a whit about a caution not to use the backcasted survey data. By doing so, it clearly exposes respectfully that the data that is not backcasted, your claimed "gold", is "fools gold" from the prior census type survey. Why did you bother publishing it? Your workers clearly had a problem with a big chunk of Canadian industry releases missing. Why not a note about that? Or a forecast of what might have been the production of this group? A difference of 425 recordings from the 2003 year industry census result, now showing as 479 reported, is rather too large a difference not to notice. Reading your report in the Daily, it said nothing to this, instead it talks as if these companies no longer exist. Why?

Worse, between albeit only two data points, 2000 and 2003, the industry census shows a trend in the new Canadian artist releases that is negative: 1,034 in 2000 to 904 in 2003, down 13%. The new survey shows 445 in 2000 and 479 in 2003, up 8%. Did the industry actually get better back then when looking at its history??? The report in its cautions said you could still discern the trends. How? By using the wrong ones?

Now in exploring the data further, the census and the survey two years overlap, 2000 and 2003. There is a caution not to use backcasted data, but if one does not use financial numbers which are of non-release related firms, we can likely use the raw release numbers, ipsofacto these are the same firms in the survey data as we suspect.

If we take that and the fact that the CRIA claimed 95% of this industry, then we have the picture of the CRIA, the 6 Canadian firms that defected in 2006, and the 5% who Statistics Canada does not report on in 2007 for the 2005 year:

(Click on sheet for larger view)

I got a bad hunch that it does not matter here what I dug out. I hope I am wrong (I am going to email it to some folks and see what they have to say). I equally hope that the people in this industry, more intimate with this type of information than I, who know this 2007 report for what it is, who have been silent, speak out.

(Ed. note: The 6 2006 CRIA defectors, were in the 95% of the industry claimed by CRIA press releases of the 2005 year. 5 of them can be found on the member lists that CIRPA maintains. The CRIA does not recently show its member companies on its current website. In the past it did under the former Presidency of Brian Robertson that ended November 15, 2004: see this link from The CRIA included companies like Cinram that manufacture sound recordings for retail distribution, in whatever format of the day, and distributors; neither having any role in creation of new artist releases. I noted the "easy" 95% figure of the CRIA in this post. I do have doubts of their claimed share but take them at their word, carefully. It has also been stated as 96% varying by year prior to the defections in 2006. The CRIA "members" page BTW was last seen December 30, 2004)

No comments: