Geert Goeyvaerts

Postdoctoral Researcher @LeuvenEconomics | KU Leuven.

[Geert Goeyvaerts @GGoeyvaerts]

Happy to see my note out in Applied Economics Letters.

To choose between RD, DID and difference-in-discontinuities, you need to know the minimum assumptions each design requires. For sharp DIDISC and DISCID, two suffice.

https://t.co/byT8OfNY4E

[IEB @FundacioIEB]

📣 #Call for Expressions of Interest!
The IEB and @ubeconomics invite applications for a 4-year full-time #PhD #Fellowship in #UrbanEconomics.

The successful candidate will join the research by IEB researchers @eviladecans & @RosaSanchisG

🔗 Information: https://t.co/cd2qfndieM

[Geert Goeyvaerts @GGoeyvaerts]

The second assumption was hiding in the notation. In earlier formulations it is built into how potential outcomes are defined, so it cannot fail by construction.

It's one of DID's two core assumptions. It should be stated, not assumed away.

SSRN: https://t.co/btDLq5C6XS

[Geert Goeyvaerts @GGoeyvaerts]

Happy to see my note out in Applied Economics Letters.

To choose between RD, DID and difference-in-discontinuities, you need to know the minimum assumptions each design requires. For sharp DIDISC and DISCID, two suffice.

https://t.co/byT8OfNY4E

[Geert Goeyvaerts @GGoeyvaerts]

This has an important implication for applied researchers: the average untreated outcome can jump at the cutoff, for any reason, driven by any number of confounders. You never have to identify them, model them, or name them. Only the change over time has to be continuous.

[IEB @FundacioIEB]

📣 #Call for Expressions of Interest!
The IEB and @ubeconomics invite applications for a 4-year full-time #PhD #Fellowship in #UrbanEconomics.

The successful candidate will join the research by IEB researchers @eviladecans & @RosaSanchisG

🔗 Information: https://t.co/cd2qfndieM

[Geert Goeyvaerts @GGoeyvaerts]

New version of our comment on Borusyak & Hull (2023)!

In dynamic settings, recentering may remove an important part of the total treatment effect.

In BH's Chinese HSR application, a novel instrument yields a ~4x larger employment elasticity.

https://t.co/zCY8877Tlv

[Geert Goeyvaerts @GGoeyvaerts]

The takeaway: transaction tax design involves real distributional trade-offs. Sellers benefit more from uniform cuts (more sales, higher prices), while buyers and tenants benefit more from preferential cuts (stable prices, lower rents). (6/6)

[Geert Goeyvaerts @GGoeyvaerts]

Does it matter whether transaction tax cuts benefit all buyers or just homeowners?

In a new paper with Joren Vandenbergh, we find the design matters a lot! Uniform vs preferential cuts produce strikingly different outcomes. https://t.co/AQRp7rSOHC (1/6)

[Geert Goeyvaerts @GGoeyvaerts]

The mechanisms differ too. Uniform cuts increase sales of owner-occupied homes but not rentals—expanding the rental stock. Preferential cuts increase sales of both and reduce the BTR probability of both, shifting properties from rental to owner-occupied sector. (5/6)

[Sven Damen @damensven]

Wil je baanbrekend onderzoek combineren met beleidsrelevante inzichten? Wij zijn op zoek naar een voltijdse doctoraatsbursaal in woningmarkteconomie aan @UAntwerpenFBE

Solliciteren kan tot 26 februari 2026 https://t.co/Rrzd6SDVkJ

[Sven Damen @damensven]

How does early disclosure of flood risk affect housing prices and who buys risky homes?

In a new paper, we study a reform that required flood risk to be displayed directly in real estate listings, indicating whether a property lies in a high- or medium flood risk area (1/5)

[Geert Goeyvaerts @GGoeyvaerts]

Happy to see this paper published in RSUE. @damensven and I find that subsidies for homeowners largely capitalize in house and land prices, but do not significantly affect sales of existing houses and undeveloped land, nor permits for new houses. https://t.co/UgVwQY50lt

[Geert Goeyvaerts @GGoeyvaerts]

1) continuous trends and 2) no anticipation at the cutoff. You can find the paper here: https://t.co/btDLq5C6XS. Comments always welcome!

[Geert Goeyvaerts @GGoeyvaerts]

Wrote a note in which I revisit identification for difference-in-discontinuities (DIDISC) and discontinuity-in-differences (DISCID) when treatment assignment is sharp. It shows that both estimands equal the average treatment effect at the cutoff (ATC) under two assumptions: