Very interesting!

I've got a paper on small-scale (buy-to-let) housing investors. The results pointing in the same direction as the papers you mention. Taxing investors increases homeownership and total welfare, while increasing rents.
papers.ssrn.com/sol3/papers....

Isn't the colors on the housing part wrong? The Haig-Simons should be light?

Absolutely! Getting at the total welfare effects, including increased demand for workers, changing ameneties etc. is really complicated. But the benefit likely differs. Even if the total effect of Airbnb is positive, the effect on those struggling with housing costs might be negative.

Allowing more housing to be build is the most important solution, by all means.
But there is also convincing research that Airbnb can cause significant increases in housing prices and rents for cities with large tourist demand, like Barcelona. www.sciencedirect.com/science/arti...

The dataset described here (on Norwegian residential housing auctions) is very nice, though unfortunately not publicly available.
housinglab.oslomet.no/wp-content/u...

I wonder here about the construction outcome: % growth in housing. Is it a good supply measure?

Imagine City A has high regulation, B low. Both have 100 inhabitants, but A has 80 houses while B has 100.
Following an inflow of 10, A builds 8 houses, B 10. Same % increase, but not similar results.

I have a recent WP with colleagues about how that is done in Norway.
bsky.app/profile/erle...

That's interesting. I do however think that the capacity for implementing such a tax (particularly the valuation) has increased since the early 20th century. Though the political pressure to attach loopholes is of course still there.

While Norway has few inhabitants, it has a lot of variation both in geographic features and in density of housing. The simple model has a reasonabe fit even in rural areas with few transactions. In more densely populated countries, a similar model would likely have an even better fit. 5/5

The existing system in Norway functions quite well. It is not very data intensive, though does require digitalization of property register and transaction data.
We also present an improved system which better utilizes available geographic information. 4/5

We’re arguing that a simple, hedonic regression based system, as existing in Norway, can be accurate and cost-effective. An automatic system also avoids assessors corrupting the process for their own ends. 3/5 bsky.app/profile/nber...