https://en.wikipedia.org/wiki/Henry_George_theorem
IE, the aggregate value of public spending in an area tends to soak into the land value -- give everybody $X,000 a year and the landlords will dutifully raise annual rents by $X,000.
Rather than getting $50 a day as cash, you might get 4 nutritionally balanced but generally unpalatable meal rations, two payment vouchers for one day in one state-owned 250 sq.ft. housing capsule, 30 kWh of metered electricity, 100 gallons of potable water, 10 Gb of network bandwidth, etc. But you put redemption limits in place so that you can't use vouchers to pay for more than 1000 sq.ft. of capsules, 60 kWh of electricity, 200 gallons of water, or 20 Gb of electricity per day. It places limits on the amount of benefit any single person could suck out of the system at the expense of everyone else in it.
The state would also have to be a major wholesale buyer for those things, and sometimes also set up and spin off new competitive suppliers, otherwise the incumbent suppliers might be able to jack up the price and capture a portion of the benefits intended for the end consumer. Taxes and subsidies have a different mechanism of action economically than the entry of a new supplier firm.
See: Food stamps.
The only caveat is that there are other possible "rentiers" that can capture the rent before it winds up in land. Such as patents, barriers to entry, etc.
But exclusive use of land is the first and most basic barrier to entry.
It also is being given people, not directly to the area, so if rents are raised in one area, people can move away from it. This isn't the case with infrastructure improvements such as road/rail/schools/electricity improvements.
The likely effect will be that cheap housing will increase in price, but it will be bounded by moderate housing (i.e. for those people who don't see a real terms change in their post-tax/benefits income).
The other point, is that that the sector most dramatically affected by this will be the homeless and unemployed.
An interesting question is to what extent the basic income should be linked to inflation and differential cost of living (e.g. should it be more for people who live in San Francisco as San Francisco is expensive, or should it be equal because we don't care where people live).
The CoL problem is currently "solved" in Unemployment and Social Security by tying it to past wages (which probably correlate with this). I think we do need to replicate that in some way.
A similar metric might be to tie it to the minimum wage of a region, which will hopefully capture inflation and CoL.
The theorem's theoretical basis is pretty spherical-cow: a government that governs land centered on a single point with people trying to minimize distance to the space. (Which is to say, it models best to the idea of a city-state rather than a federal government.)
