I’m sure we can come up with contrived examples of things, but be open-minded to how things could work.
Anyway, it’s not that property taxes are 1:1 with population, however it does broadly correlate with density. A farmer next to a subdivision is a more extreme version of the same thing that only represents which farmer happened to sell their land. Meanwhile a farm is less value per acre than a housing development. Just as high rises represent more value per acre than garden style apartments and garden style apartments beat single family homes.
Land value taxes on the other hand ignore all such differences by design. The farmer pays the same amount per acre as the subdivision next door. LVT can capture increases in value as regions develops, but those increase don’t line up with political borders.
So sure, it’s of course possible for properties next to each other to have significantly different values per acre, but nothing says the higher density one must have a higher or lower value so let’s just assume their identical for simplicity. Any way you slice it it’s clear LVT doesn’t care about the actual population on the land.
Local population (incl. the density thereof) is arguably the driver of value, moreso than any other factor besides sheer size and maybe location (in exceptional cases, e.g. adjacent to a coastline). If you could build a store or a restaurant anywhere, then all else being equal, would you build it where there are 10,000 potential customers within walking distance or where there are merely 1,000? Likewise, are you going to move where there are lots of places to eat and shop or where there are few? Generally people want to be where other people are, and this produces a positive feedback loop between local population and demand for land; there are exceptions, of course (some people like to live away from population centers instead of in the middle of them), but they ain't numerous enough to meaningfully suppress that feedback loop.
