Software
This page lists program packages that I wrote for my research. I work on models including unobservable heterogeneity and these packages provide a useful toolkit to estimate such models.
The packages are available from my Github repository @yujunghwang and the CRAN.
R package
factormodel
The package factormodel provides functions to estimate a factor model using either discrete or continuous proxy variables.
For example, if you use discrete subjective attitude data to proxy for true attitude (e.g. variables asking whether to agree/disagree), an alternative to using a z-score index is to estimate measurement (stochastic) matrices in Hu (2008) and Hu (2017). (See the references in the vignette). This method handles potentially nonlinear measurement error better than the z-score index method. You can easily do this by using 'dproxyme' function in the package.
If you estimate a linear factor model (e.g. Cunha, Heckman, Schennach (2010)), you can use a 'cproxyme' function in the package.
bndovb
The package bndovb implements a Hwang(2021) estimator to bound omitted variable bias. It provides an alternative bounding estimator to Altonji, Elder Taber (2005) and Oster (2019). For further detail, check Hwang, Yujung (2021), "Bounding Omitted Variable Bias Using Auxiliary Data."
R Code
SeminarOrganizer
This repository includes functions to automate weekly seminar scheduling and announcement. You can set up the automation once at the beginning of the semester and forget about sending emails to the seminar speakers / to your seminar group. The code will send the customized email at the scheduled time.