robust_solve_XtX

This function computes the least squares solution beta = (X'X)^(-1) X'Y
in a numerically stable way using QR decomposition, handling rank-deficient
matrices gracefully.

Estimates Two-way Fixed Effects difference-in-differences/event-study models using the approach proposed by Gardner (2021) <doi:10.48550/arXiv.2207.05943>. To avoid the problems caused by OLS estimation of the Two-way Fixed Effects model, this function first estimates the fixed effects and covariates using untreated observations and then in a second stage, estimates the treatment effects.

Kyle Butts

did2s

Two-Stage Difference-in-Differences Following Gardner (2021)

John Gardner

Grant McDermott

Laurent Berge

robust_solve_XtX function

<dl><dt>X</dt>
<dd>Design matrix (sparse or dense)</dd>
<dt>Y</dt>
<dd>Response matrix/vector (can be X'Y if already computed)</dd></dl>

Arguments

Robust solve for X'X beta = X'Y using QR decomposition — robust_solve_XtX

<dl>

<dt>X</dt>
<dd>Design matrix (sparse or dense)</dd>


<dt>Y</dt>
<dd>Response matrix/vector (can be X'Y if already computed)</dd>

</dl>

robust_solve_XtX: Robust solve for X'X beta = X'Y using QR decomposition

Description

Usage

Value

Arguments