Actually there is. The statement was true in maybe January. Since then theres more and more evidence of immunity. 4 million plus cases and zero reinfections also adds weight to it. As do the other 4 coronaviruses that affect people every year where most people get immunity for 2-3 years.
Also, if there is no immunity whats your solution? Permanent house arrest for the entire population?
Yes, there's evidence of *some* immunity, but what we don't currently know is how long it lasts. Could be permanent, could be a few months, could be the 2-3 years you suggest - which is probably more likely, given the type of virus and how others of type play out.
My solution? Gradual relaxation of restrictions, see what works and what doesn't based on infection rates, backout those changes that don't. Each country (or state/region) trying slightly different things actually works well on a global scale for this - we basically have a few hundred tests running in parallel. All the while we should obviously continue to capture data, develop and refine your procedures and models, and improve your medical capabilities. The crux, and entire purpose of restrictions, is preventing medical services from being overwhelmed until you either get to something close to herd immunity or develop a vaccine (they are the only two real "endgames") - and avoiding another situation where hospitals were overrun like in Lombardy at all costs.
I don't see any reason why we can't go back to work, start enjoying the countryside etc., but I do think that there needs to be continued restrictons and additional measures, especially invovling things like mass transit and sporting events where large numbers of random people congregate. That does, however, need to be balanced against the available medical resources; if you have a lot of beds and ventilators then you have a lot more latitude to relax restrictions than if you don't. It's also a personal matter - everyone's circumstances are different, so it's also very important that if an employer wants to resume operations then they should not be penalising any employees who may not wish to take unnecessary risks if they have good reason - being in, or caring for someone in, a high risk group, for instance.
Stockholm showed roughly 30% exposure. NYC had 24% in some studies for example. Actual figures. The ONS is just guesswork based on the same proven flawed imperial modelling. The same day other studies estimated 20% based on actual tests (albeit small sample sizes).
Diamond Princess 74% of positives were asymptomatic. CdG aircraft carrier showed 50%. US carrier roughly similar.
Chinese roughy 78% (BMJ 369). Iceland over 50%.
US Federal prison study showed 96%.
The ONS study is based on antibody test kits sent out to randomly selected people.
So, we have 24% (NYC), 30% (Stockholm) - both general, and very representative urban populations - and 50% (Iceland) - roughly 50-50 between urban (mostly Reykjavik) and extremely rural - none of which support your 70-75% assertion. China's figures are in the ballpark, but either false (as the US keeps claiming), a demonstration of how effective a draconian lockdown can be on keeping mortality down, or some combination of the two - everyone will need to make up their own mind on that point (I'm in the "combination" camp). On the otherhand we have captive audiences where the virus is going to run rampant like ships and prisons ranging from 50-96%. The latter does bode quite well, but those are not representative population samples; the ships all have a population generally biased towards younger crew and I don't have any data on the prison survey demographics, but I believe prison populations are more biased towards youth as well. Either way, there are some awfully error bars in there - larger than those from Ferguson's "model", in fact - so we're back to Russian roulette.
We need data, we don't need to be reckless.