Don't (fully) exclude me, it's not necessary! Causal inference with semi-IVs (SSRN) (arXiv) 

-- new version! 

R Package: semiIVreg - [vignette]

Presented at: Duke Microeconometrics Class (2023), NESG 2023, Bristol Econometric Study Group 2023, ESWM 2023, ASSA 2024, Oxford Encounters in Econometrics 2024, BSE Summer Forum 2024, IAAE 2024, ESEM 2024.

Abstract: This paper proposes semi-instrumental variables (semi-IVs) as a practical alternative to instrumental variables (IVs) to identify the causal effect of a binary (or discrete) endogenous treatment. A semi-IV is a less restrictive form of instrument that is relevant but excluded only from one, not necessarily both, potential outcomes. Having two semi-IVs, one excluded from the treated and the other from the untreated potential outcomes, is sufficient to nonparametrically point identify the local average treatment effect (LATE) and marginal treatment effect (MTE). In practice, semi-IVs provide a solution to the problem of finding valid IVs, because they are easier to find: most selection-specific shocks, policies, costs, or benefits are valid semi-IVs. As an application, I estimate the returns to manufacturing using sector-specific semi-IVs.

Identification with possibly invalid IVs (SSRN) (arXiv) 

joint with Jad Beyhum

Presented at: EEA 2024, EC^2  2024. 

Abstract: This paper proposes a novel identification strategy relying on quasi-instrumental variables (quasi-IVs). A quasi-IV is a relevant but possibly invalid IV because it is not completely exogenous and/or excluded. We show that a variety of models with discrete or continuous endogenous treatment, which are usually identified with an IV - quantile models with rank invariance additive models with homogenous treatment effects, and local average treatment effect models - can be identified under the joint relevance of two complementary quasi-IVs instead. To achieve identification we complement one excluded but possibly endogenous quasi-IV (e.g., "relevant proxies'' such as previous treatment choice) with one exogenous (conditional on the excluded quasi-IV) but possibly included quasi-IV (e.g., random assignment or exogenous market shocks). In practice, our identification strategy should be attractive since complementary quasi-IVs should be easier to find than standard IVs. Our approach also holds if any of the two quasi-IVs turns out to be a valid IV.

Discrete-Continuous Dynamic Choice Models: Identification and Conditional Choice Probability Estimation (SSRN) 

Presented at: DSE 2021, ESEM 2021, BSE Summer Forum 2024

Abstract: This paper develops a general framework for models and games, static or dynamic, in which individuals simultaneously make both discrete and continuous choices. The framework incorporates a wide range of unobserved heterogeneity. I show that such models are nonparametrically identified. Based on constructive identification arguments, I build a novel two-step estimation method in the lineage of Hotz and Miller (1993) and Arcidiacono and Miller (2011) but extended to simultaneous discrete-continuous choice. In the first step, I recover the (type-dependent) optimal choices with an expectation-maximization algorithm. In the second step, I estimate the primitives of the model taking the estimated optimal choices as given. The method is especially attractive for complex dynamic models because it significantly reduces the computational burden associated with their estimation compared to alternative full-solution methods.

Housing Prices Propagation: A Theory of Spatial Interactions 

joint with Guillaume Chapelle, Jean-Benoît Eyméoud and Etienne Wasmer

Abstract: Price-to-rent ratios in the housing market vary a lot in time and space, and this cannot be explained solely by differences in local discount rates or rent growth differences. We study a variant of asset pricing equations for housing markets that include a price gradient in space, analogous to the transport equation in physics. The rationale for it is the existence of spatial search frictions for housing. A parametrization, backed by data, implies that a 10% increase in the price of an adjacent city raises the price of the city by approximately 0.05%. These effects cumulate over time and space and lead to a 15% increase in prices in the periphery after 30 years. More complex asset price equations such as heat diffusion equations are also derived from the search friction model and briefly discussed but second order terms are not found in the data.