Juniorprofessorship (Assistant Professor) in Economics and BusinessGoethe University Frankfurt Contact Information
E-mail : Choi@econ.uni-frankfurt.de |

__Publications__

Lee, M.J. and J.Y. Choi, 2012,

**Bounding Endogenous Regressor Coefficients Using Moment Inequalities And G****eneralized Instruments**, Statistica Neerlandica, 66, 161-182 (Impact Factor: 0.564).Choi, J.Y. and M.J. Lee, 2017,

**Regression Discontinuity Essentials: a Review with Extensions**, Statistical Papers, 58, 1217-1246 (Impact Factor: 0.813).Choi, J.Y. and M.J. Lee, 2018

**, Minimum Distance Estimator for Sharp Regression Discontinuity with Multiple Running Variables**, Economics Letters, 162, 10-14 (Impact Factor: 0.588).Choi, J.Y., 2018

**, Prejudice Matters in Elections: An Estimator for Binary-Outcomes with Sample-Selection**, Oxford Bulletin of Economics and Statistics, forthcoming (Impact Factor: 1.247).Choi, J.Y. and M.J. Lee, 2018,

**Regression Discontinuity with Multiple Running Variables Allowing****Partial Effects**, Political Analysis, forthcoming, (Impact Factor: 4.655).

__Working Papers__**Choi, J.Y. and M.J. Lee, 2017, Are Twins the Same? : Heavily Censored Semiparametric Quantile Estimation with Fixed Effects**(R&R, Journal of the Royal Statistical Society: Series A, Impact Factor: 1.702).

Twin studies are popular, because twins are believed to be the same/similar in genes and environmental exposures. It is well documented, however, that the fi
rstborns are healthier at birth. This paper uses the US records of twin births during 1995-2000 to show that the survival duration parameters differ between twins depending on the birth order: even if the twins were exactly the same in their health at birth, still their survival patterns would differ systematically depending on the birth order. We also fi
nd that wiser (e.g. older or educated) or married (thus resource-richer) mothers take a more care of the weaker; this compensating behavior is opposite to the differentiating behavior seen in animals. The survival pattern difference and the mothers intervention against the nature send cautions to twin studies that regard twins homogeneous on average. Since the survival duration in our data is 97% right-censored in one year, we devise a quantile-based
fixed-effect semiparametric estimator that can handle a heavy censoring, which is our methodological contribution.

**Choi, J.Y. and M.J. Lee, 2017, Finding Quantile Gains of Movers with Selection Correction under Heteroskedasticity and Hetero-Correlation**(under review).

The usual remedy for a sample selection problem is selection correction approach of adding a term to the model to negate the bias due to the selection; the selection correction term per se is not of interest, but a nuisance. If the selection is an individual choice of moving out of a state, however, the selection correction term is of interest on its own, because it represents a gain in selection/move. This paper
nds quantiles of the gain distribution for the movers from one state to another, using quantile selection correction terms. Our approach is fully parametric, but we allow heteroskedasticity and hetero-correlation between the error terms in the selection and outcome equations, whereas existing methods allow only highly restricted forms of those. More importantly, our approach enables comparing the movers to the stayers, whereas the existing methods can compare the movers only to the population consisting of the movers and stayers; the latter dilutes the mover effect, and thus becomes possibly misleading. We apply our approach to a real data set to demonstrate that heteroskedasticity and hetero-correlation matter greatly, and comparing the movers to the stayers is far more informative than comparing the movers to the population.

**Choi, J.Y. and M.J. Lee, 2017, Relaxing and Testing Local Average Treatment Effect Assumptions for Fuzzy Regression Discontinuity**(under review).

In fuzzy regression discontinuity (RD) with a running variable S and a cutoff c, the identi
ed treatment effect is the effect on compliers at c. This local average treatment effect (LATE)interpretation is based on the instrumental variable estimator (IVE) where the treatment D is instrumented by the dummy variable for S crossing c. We revisit the assumptions for the LATE in RD to do a couple of tasks. First, the LATE interpretation requires the independence of the potential variables from the instrument, which we replace with much weaker moment continuity conditions, as the independence assumption is unlikely to hold in reality. Second, IVE needs an exclusion restriction which may be, however, violated for certain RD cases, and we thus relax the exclusion restriction and impose treatment exogeneity instead, under which ordinary least squares estimator (OLS) is valid. Third, we propose three tests to determine which is correct between the IVE allowing treatment endogeneity under the exclusion restrictionand the OLS needing no exclusion restriction under treatment exogeneity. We provide an empirical illustration for the well-known retirement-consumption puzzle.

**Lee, M.J. and J.Y. Choi, 2017, Feasible Procedures to Estimate Individual and Time Effects with Big Square Panel Data for Binary Response.**

** **

**Lewbel, A., J.Y. Choi, and Z. Zhou,
2018, General Doubly Robust Identification and Estimation. **

** **

**Johansson, P., M.J. Lee, and J.Y. Choi, 2017,
Compensating or Reinforcing Behavior of Parents.**