> For example, instead of a cap on a neighborhood of three stories, the cap could be 33% more than the tallest structure in the adjacent two blocks. Or 50% more than the median height of the adjacent two blocks.

I was just thinking about this a little more and I think this is actually quite bad.

Because rents falling to construction costs is relativistic. Suppose it costs a given amount to build a 20 story building and you can recover at least that much in net present value from it in rent. Then you're going to do it if it's permitted and you have an empty lot. But not if you have an existing 10 story building. Then it would have to generate double the construction cost in rents, because you're getting half that amount of rent from the status quo.

Which means that to minimize that inefficiency, you don't want to be replacing 10 story buildings with 20 story buildings. What you want is to be replacing 4 story buildings with 20 story buildings. Because then you only have to beat the construction cost by 20%. Which doesn't work well if you have to wait until the median building is 10 stories tall before you can build a 20 story building.

But it's not completely wrong if you use different numbers. Say, have three classes of zoning. One allows buildings up to 4 stories, the next buildings up to 20 stories, the last unlimited. You automatically transition from the first to the second when the median building height reaches 3 stories, and from the second to the third when it reaches 6 stories.