learn_nickell_bias

source

learn_nickell_bias(n_units=200, rho=0.6, periods=(3, 4, 6, 10, 20, 40), seed=0)

Show the Nickell bias in dynamic-panel fixed-effects estimates.

The within estimate of a lagged dependent variable is biased downward in short panels and the bias shrinks as the panel lengthens. Simulates a dynamic panel y_it = rho*y_{i,t-1} + alpha_i + e and estimates rho by within (fixed-effects) regression at several panel lengths T: for small T the estimate sits well below the truth; as T grows it converges up to rho.

Parameters

Name	Type	Description	Default
n_units	int	Number of units (kept large so the bias is the dominant signal).	`200`
rho	float	True autoregressive parameter.	`0.6`
periods	Sequence[int]	The panel lengths `T` to estimate at.	`(3, 4, 6, 10, 20, 40)`
seed	int	Random seed.	`0`

Returns

Name	Type	Description
	SandboxResult	`df` (`periods_T`, `fe_rho`, `bias`), `fig`, `summary` and `topic`.

Examples

This sandbox simulates its own dynamic panels, so the call needs no DataFrame:

import expdpy as ex

res = ex.learn_nickell_bias()
res.fig