F-CUR Project

JACKKNIFE INSTRUMENTAL VARIABLE INFERENCE 

F-CUR project 2274-2022-CF-PSR2021-FCUR 001

SUMMARY OF THE PROJECTInstrumental variable (IV) methods are part of the standard researcher’s toolbox in a number of disciplines ranging from economics (and social sciences in general), statistics and epidemiology. This project aims at introducing a set of test statistics for linear regression models in the presence of endogeneity, heteroskedasticity and many, potentially weak, IVs. The starting point of the study is the objective function of the symmetric jackknife instrumental variable (SJIVE) estimator of Bekker and Crudu (2015, link). The test statistics are derived in analogy with standard (quasi) likelihood theory. The project is mainly theoretical with two main facets: first, the study of the limiting distribution of the proposed statistics under certain, practically relevant, null hypotheses; second, the analysis of the power properties of the tests. The finite sample behaviour of the proposed tests is investigated via a set of Monte Carlo simulation on models that are commonly found in applications.

BUDGET: 17290 Euros

DURATION OF THE PROJECT: 2022-2023 (18 months)

OUTPUT, WORK IN PROGRESS AND RELATED PUBLICATIONSJackknife Instrumental Variable Inference (with G. Mellace and Z. Sándor), coming soon.The Gray Zone: How not Imposing a Strict Lockdown at the Beginning of a Pandemic Can Cost Many Lives (.pdf, 2022, with R. Di Stefano, G. Mellace and S. Tiezzi), 2024, Labour Economics. In the press: lavoce.info.Narrow and Wide Replication of Chalfin and McCrary (REStat, 2018) (with A. Moharir), 2024Journal of Applied Econometrics.

WORKSHOPWorkshop on "Econometric Theory and Applications" 22-23 of June 2023 in Siena (link).