Identification with sharp difference-in-discontinuities and discontinuity-in-differences

Submitted

This note revisits identification for difference-in-discontinuities (DIDISC) and discontinuity-indifferences (DISCID) when treatment assignment is sharp. I show that both estimands equal the average treatment effect at the cutoff (ATC) under two assumptions. The first assumption is the continuous trend assumption. The second assumption is no anticipation at the cutoff. 

Split incentives, asymmetric information and energy efficiency subsidies  With Sven Damen & Stef Schildermans

We examine whether adverse selection due to asymmetric information reduces the cost-effectiveness of subsidy programs for energy efficiency investments. In 2009, the Flemish government introduced a reduction in the property tax for newly built houses whose energy efficiency level is below a specific threshold. We find that the response of houses built by a developer is four times smaller than the response of houses built by the owners themselves. This makes the program only half as cost-effective for developer-built housing. Additional data suggests information concerning the efficiency level of housing units was asymmetric between developers and buyers. It usually takes several years before the efficiency level is officially determined and it is not always correctly reported in listings. 

Being revised

Housing market responses to the mortgage interest deduction. with Sven Damen

Regional Science and Urban Economics, 2025, Accepted

Reconstructing cities: stimulating redevelopment through the tax code  

Regional Science and Urban Economics, 2023, 99, Article 103880. 

Do market rents reflect user costs? with Erik Buyst

Journal of Housing Economics, 2019, 44, 112-130.

Building codes: theory and evidence from energy efficiency requirements  (With Stef Schildermans)

Keeping investors at bay: the impact of uniform and preferential reductions in the transaction tax (With Joren Vandenbergh)

Taxing vacant land (With Thomas Boogaerts & Stef Schildermans)

See how the land lies (With Sven Damen & Jacqueline Seufert)

Do taxpayers look beyond the statutory incidence? (With Thomas Boogaerts)

Reducing tax avoidance through information provision (With Thomas Boogaerts)

Local (synthetic) difference-in-differences  (With Jacqueline Seufert)

Hold out in land assembly for residential development 

With Thomas Boogaerts

Using the universe of construction permits, sales and land use in Flanders (Belgium), we present new evidence on frictions in the land market. First, developed parcels for assembly projects are bought at a premium of 17%. This premium is not the result of differences in land use restrictions, local prices, reservation prices nor the quality of the existing structure. Second, parcels in assembly projects are less likely to be (re)developed than similar parcels in single-plot projects. The results from three tests suggest that holdout by landowners might be one of the frictions: i) the last acquired plot is sold at an additional premium of 9%, ii) smaller parcels receive a higher premium and are less likely to be (re)developed and iii) assembly projects with more parcels face a higher average premium. Lastly, we document an important difference between developed and undeveloped parcels. While undeveloped parcels in assembly projects are also less likely to be (re)developed, they are bought without a premium.