Ultimately, I offer a simple tool which economists can use to check if their models are suitable for their housing counterfactual.

The Elasticity/Baseline framework can help pick a plug-and-play model, or suggest you may have to design your own!

Remember: quantitative implications can be large!

Also, not all specifications work for all counterfactuals.

Duranton and Puga (2023) permit cost shock works fine, but other counterfactuals do not.

Greaney, Parkhomenko, Van Nieuwerburgh (2025) shock TFP because elasticity effects are small. But is it always true? How to map into real world?

10/n

These assumptions have quantitative implications.

I introduce a more flexible price function, and calibrate it using D'Amico et al. (2025) and Parkhomenko (2023) estimates of how regulations affect TFP.

In a Hsieh and Moretti (2019) model, this function estimate x6-x10 larger policy effects.

9/n

These implicit assumptions impose all kind of problems. From lack of identification, to assuming that no change in regulations can make prices fall if demand is fixed.

Some are not necessarily issues, but require reflection on whether they are suitable for the policy question of the paper.

8/n

Turns out that popular house price functions do not allow to separately characterise the Elasticity and Baseline effects of regulations.

When matching the former from data (e.g. with Saiz (2010)), implicit assumptions are made on the latter.

Big issue for counterfactuals on elasticities!

7/n

This decomposition is *extremely* useful to understand the implicit assumptions we are making in our models!

Have we accounted for both channels of house prices?

Can they be independently characterised?

What is their inter-dependence/correlation?

So... what are the assumptions in the lit?

6/n

I formalise this argument into seeking to estimate the "Elasticity" and "Baseline" effects of policies.

For example, reducing minimum plot sizes not only reduces the price-demand elasticity (Elasticity effect), but also the cost of supplying the existing units (Baseline Effect).

5/n

Indeed, we can. When we perform a counterfactual on housing regulations or supply, we want to know:

1) How we change the elasticity with which house prices respond to demand

2) How we change the price of satisfying the existing (or "reference") demand for housing, everything else equal

4/n

Problematically, not all house price functions, estimation techniques and counterfactuals go well together! Other papers (see this excellent note by Brian Greaney) have noticed this in specific contexts. drive.google.com/file/d/1iNdQ...

But can we generalise the argument?

3/n

Theoretically, estimating the aggregate effects of housing supply policies (think land use regulations, plot sizes, ...) requires a location choice model. And, crucially, a house price function.

Once we have these ingredients, we estimate the model parameters and run a counterfactual.

2/n