As Duncan notes, in his subsequent press release which is now on the link, the issue is whether the requisite economic data can be gathered in less intrusive ways (without presumably sacrificing in a material way accuracy or quality) then a census survey, that if you refuse to answer, results in a 5K fine potentially.
If there is, then it seems to me that Duncan has the upper hand in the balancing test between the needs for the polity to have the data for policy making purposes, and concerns about individual privacy and the burdens imposed on such individuals.
We seem to be putting up a lot of threads lately where there is more to the story than what the article source (or headline), or the spin placed thereupon here, would imply. I don't bother with most of them, but on this one, I thought I would dig a bit deeper since it all seemed so absurd, as to be implausible.
Winning