Eg. I buy a property that I want to use as a factory. Some people who work for me move their own accomodation to be close to my factory. I spend absolutely tons of time and money filling my factory with machines, some of which are bolted to walls or have to be constructed inside the warehouse.

The price I'm willing to accept for this property isn't the price of the property, it's the cost to me of moving all my people and goods and rebuilding things. Plus the cost of the interruption to my production flow etc.

Caveat I know nothing about manufacturing but the point is more general, property isn't even close to fungible.

Tesla wouldn't sell their hyperfactory thing for the price they bought it for!

Everyone would be overpaying tax to the extent that their reliance on property isn't fungible, I'm not sure what the implications of that are but property is so fundamental and so much money that they'd be huge.

If I understand, your argument is that self-assessed taxes would eliminate investment efficiency.

The property owner should factor transaction costs into their own value of the building. This would increase their assessment of the land-value, and increase their taxes. This works because, keeping land valuations constant, the one with the highest transaction costs is the one who can derive the most efficiency from the land.

Then maybe the rule is that you will be forced to sell to anyone who pays you N × the value you self-assessed at, where N > 1. E.g. if transaction costs average 20%, you have to sell to anyone who will pay 1.2 times your self-assessed value.

I think the real issue is whether is promotes certain uses of land over others, and whether that's problematic.

Everyone overpays, but they overpay in proportion to how sticky their business is to a specific property. It would take someone who knows more about commercial property to dissect what businesses would get destroyed by that.