If the market was homogeneous worldwide, then yes, BCN would be a good source of information to extrapolate from. But it isn't, at all*, and so people make wrong assumptions because they base them on incomplete data. What's worse, though, is that what BCN publishes is very partial information. They make sales rankings for each month of the year, but never list the number of units or market share percentages, and their EOY rankings are extremely sparse in data too, only giving general market share percentages, so that it's impossible to build a complete picture even if both data sets are combined.
Canon can certainly assign more importance to Japan or whatever market they prefer. That says absolutely nothing about how their products are selling in other regions of the world. In East Asia, for example, there are wild differences from country to country - China is crazy about 35mm DSLRs, in South Korea Sony is king, in Thailand Fujifilm have the #1 spot, whereas Vietnam reflects the Japanese market somewhat. Base any far-reaching conclusions out of any of these countries and you'll get a completely unrealistic picture of worldwide sales.
When it comes to accurate numbers, it's CIPA or bust. One can certainly use other sources to complement this, but CIPA reports still constitute the only authoritative sales information available to non-insiders.
*Just look at Olympus - they were #1 in Japan for a couple of years with their PEN series, but could only gain a foothold in Europe and the US with the OM-D's... which, E-M10 excepted, sell in much lower volumes in Japan Add DSLRs and different formats into the mix, and things get complicated very quickly.
it's not their job to be all comprehensive and do all the extensive research. they report news as it comes in. it's not even their job to verify whether any of the information is correct. the only verification is if the source is a legitimate organization. you can't expect them to verify whether that organization measured it properly.