Calculate Anderson-Rubin Confidence Intervals for fixest IV objects (Compatible with Fixed Effects)

This function calculates Anderson-Rubin (AR) confidence intervals for a specified endogenous variable from an IV regression estimated with fixest::feols. It is robust to weak instruments and compatible with models containing fixed effects.

Usage

ar_ci(model, param, level = 0.95, grid_range = NULL, grid_length = 1000)

Arguments

model

A 'fixest' object resulting from an IV estimation.

param

A character string specifying the name of the endogenous variable. If the model has only one, this can be omitted.

level

The confidence level required (e.g., 0.95 for a 95% CI).

grid_range

A numeric vector of length 2, c(min, max), specifying the range of coefficient values to test. If NULL, a default is chosen.

grid_length

An integer specifying the number of points in the grid to test.

Value

Returns an object of class 'ar_ci' containing the confidence interval, the grid of tested values, and their corresponding p-values.