I think it is so you can compare year to year, if they keep changing the reference point, you could not compare. They always use the same elevation, gas, temp, air pressure, etc.
The last time I checked, several years ago, the auto makers could use any fuel blend for the tests. EPA provided a formula to translate the results back to a standardized blend of pure gasoline. Do remember that there are differences in volumetric energy content even among 'pure gasoline' blends due to differing regional and seasonal requirements. In fact, the fuel economy tests don't even measure fuel volume input at all. Instead, they measure the tailpipe emissions. As for bellyaching about E0 vs E10, the later is not yet universally required. But the 2008 'adjustment' of the EPA labels was much larger than the E0 vs E10 difference, so there is really no reason to complain. It is easier to achieve the post-2008 EPA ratings with E10 than it was to match the pre-2008 numbers with E0.
Besides what's mentioned at FAQs - Transport Canada , I recently learned of Canadian vs. American fuel efficiency: Why the big discrepancy? - AutoNorth - Auto Industry News - Canada.