Checks the convergence of a feols estimation by computing the first-order conditions of all fixed-effects (all should be close to 0)
Usage
check_conv_feols(x)
# S3 method for fixest_check_conv
summary(object, type = "short", ...)Arguments
- x
A
feolsestimation that should contain fixed-effects.- object
An object returned by
check_conv_feols.- type
Either "short" (default) or "detail". If "short", only the maximum absolute FOC are displayed, otherwise the 2 smallest and the 2 largest FOC are reported for each fixed-effect and each variable.
- ...
Not currently used.
Note that this function first re-demeans the variables, thus possibly incurring some extra computation time.
Value
It returns a list of N elements, N being the number of variables in the estimation
(dependent variable + explanatory variables +, if IV, endogenous variables and instruments). For
each variable, all the first-order conditions for each fixed-effect are returned.
Examples
base = setNames(iris, c("y", "x1", "x2", "x3", "species"))
base$FE = rep(1:30, 5)
# one estimation with fixed-effects + varying slopes
est = feols(y ~ x1 | species[x2] + FE[x3], base)
# Checking the convergence
conv = check_conv_feols(est)
# We can check that al values are close to 0
summary(conv)
#> Maximum absolute value of the first-order conditions:
#>
#> FE FE[[x3]] species species[[x2]]
#> y 1.598727e-15 1.154394e-15 2.246661e-07 1.502477e-07
#> x1 9.769908e-16 5.367730e-16 4.490874e-07 3.018886e-07
summary(conv, "detail")
#> Smallest and largest values of the first-order conditions:
#>
#> y
#> | FE: -1.60e-15, -1.07e-15, ..., 7.11e-16, 7.11e-16
#> | FE[[x3]]: -8.50e-16, -7.77e-16, ..., 9.09e-16, 1.15e-15
#> | species: -1.91e-07, -3.34e-08, 2.25e-07
#> |species[[x2]]: -3.45e-08, -8.55e-09, 1.50e-07
#> __|
#>
#> x1
#> | FE: -4.44e-16, -3.55e-16, ..., 6.22e-16, 9.77e-16
#> | FE[[x3]]: -5.37e-16, -5.33e-16, ..., 3.18e-16, 3.29e-16
#> | species: -3.83e-07, -6.57e-08, 4.49e-07
#> |species[[x2]]: -6.90e-08, -1.69e-08, 3.02e-07
#> __|