The devil's in the details. It's hard to come up with measures and targets that are both fair and easily communicable.
If you go with number of people, well, that's the wrong measure (and hard to enforce // easy to fake). Number of bedrooms is also incorrect, as is square footage. Number of rentable units is also weird, since it encourages studio apartments and the like at the expense of splitting bigger units and sharing areas.
Another option is using a baseline of "however many people currently live there", but that favors parking lot owners and the like. "Average density of the local region" is unfair in that it suddenly imposes a new cost on the density hold-outs.
You can calculate a target based on something like distance-to-transit and local zones, but that gets complicated and political in a hurry. Land-value would be a good choice, since the price conveys really useful information - but that is likely illegal due to proposition 13.
My next-best idea is collective punishment: charge zoning districts based on the difference between how their residential and commercial/industrial construction. Mountain View is much less likely to vote to let Google expand their campus while blocking commensurate housing construction if it means they have to start subsidizing the places those workers start commuting from.
"Once it reaches a certain threshold, the process of institutionalization becomes counter-productive."
- Ivan Illich
This is an example of why planning doesn't work. And the unforeseen side-effect of any policy is an example of why central control is not just inefficient but often counter-productive -- why hiring processes to identify high-capacity creative individuals select for the opposite; why the school system produces not learning but limits it so as to produce ignorance; why housing policies aim to improve cities and lower inequality makes cities like San Francisco and Stockholm some of the most unequal in the world.
You could tie minimum wage to median home value. It'd push places out of business, lowering labor demand. Sort of an indirect way of making people leave the city. And let those that can stay afford whatever ridiculous amount of money they have to pay for a 300sq ft studio.
The goal isn't to push poor people out of the area and replace them with rich people. The goal is to price out landowners who fail to develop housing.
To that end, it's probably better to somehow align the tax consequences with the people who can successfully NIMBY up development projects. That is, subsidizing living next to a high-density apartment complex, rather than subsidizing building a high-density apartment complex. Or in addition.
Like, if building a tall apartment complex in the Mission meant that all the neighbors got $50/month in density subsidies, I'd imagine that a lot of opposition would dissolve.
It's generally called a 'land value tax'. Most of the value of land derives from the amenities provided around it. By taxing the land you encourage development and discourage speculation: excess land value will be extracted through taxation.
Land-value is a better measure of how much housing "should" be built in an area. Ideally, it'd be a land-value tax with a housing capacity subsidy, making the entire thing revenue-neutral.
Unfortunately, Prop 13 makes a LVT infeasible to implement. Maybe a national LVT program would work, but that's replacing a smaller problem with a more difficult one.
Instead of a subsidy you could just allow the write-off of the structure's depreciation. Structures, unlike land, lose value over time since they have to be maintained and suffer wear and tear.
The bigger the structure, the bigger the tax write-off.
LVTs make perfect sense and won't ever be implemented because we just started off on the wrong path. We're stuck with Prop 13.
