If your analysis is correct, why is purchasing exactly ONE Abbey the correct decision? If meeting venues in general earn money and appreciate in value, then every dollar people donated to philanthropic causes would be better spent on more meeting venues, until the market is saturated with meeting venues and they cease to earn returns that make it worthwhile.
For that matter, why invest in real estate at all? A diversified portfolio of stocks and shares would give a higher rate of return and hedge against real-estate specific risk, such as another pandemic. I'm fairly sure you could get better than 2.4% if you were prepared to lock your money up for 8522 days.
It seems to me that your argument proves far too much, and is missing a very important ethical consideration that EA is primarily about doing altruism - there are critically important causes that need funding NOW, not funded at a higher rate in 20 years. The competing use of money is these causes, not real estate investment returns
You might be misunderstanding some of the dynamics here - it is really hard to successfully conceive a child if you are currently starving, because a very low BMI will interrupt a woman's menstrual cycle. However, a lot of people live extremely marginal lives - for example one failed harvest away from going into calorie deficit, or reliant on transitional forms of employment. So people might conceive a child in relatively good circumstances, and then find that they hit an economically challenging period before the child is five. Children are very susceptible to starvation that an adult can survive too, unfortunately - so even quite a small income shock could be lethal for a small enough neonate
On your comment about what exactly the 0.66 QALY means, there is extensive public discussion about how to assign a QALY weighting to moderate-to-severe depression in the NICE guidance on esketamine
https://www.nice.org.uk/guidance/ta854/history
(Download the 'Committee Papers' published on 28th Jan 2020)
I'm not sure if any of that is helpful, but it might give some useful upper and lower bounds