I hear all the concerns and agree with many of them.
Just one point in the interest of balance though;
Those working in public health would point out that the information helps long term health planning in so many ways. How many in patient beds an area will need versus access to community based care, or anticipated life expectancy, all of which help plan NHS services effectively, (even when successive governments like to interfere with the structures.)
Mind you, it would be cheaper and possibly more effective, from health perspective, to buy into the databases used by the large supermarkets. Tesco club points etc help supermarkets plan what to stock in a particular area. It also tells you which post codes buy fresh fruit & veg and which have Micro Chips and crisps. Knowing that helps plan for the heart disease care tomorrow. (And the reality of how it would help is far more complex and useful than that sweeping example.)
If you could make The NHS only 1% more efficient and targeted from the census, you have recovered the cost twice over. That's before planning for education, roads, infrastructure....
I have issues with the cost of the census. After all, technology exists to get a reply from 95% of the population just by putting the form on the internet at (if you keep consultants out of it) the cost of designing a form similar to ones that Amazon etc knock out each and every day. The remaining 5% of the population can therefore be targeted at a much lower cost.
Right, next problem?