CREEstimator class

class pytwoway.cre.CREEstimator(data, params=None)

Bases: object

Uses multigrid and partialing out to solve two way Fixed Effect model.

Parameters

• data (Pandas DataFrame) –

  cross-section labor data. Data contains the following columns:

  i (worker id)

  j1 (firm id 1)

  j2 (firm id 2)

  y1 (compensation 1)

  y2 (compensation 2)

  t1 (last period of observation 1)

  t2 (last period of observation 2)

  w1 (weight 1)

  w2 (weight 2)

  m (0 if stayer, 1 if mover)

  cs (0 if not in cross section, 1 if in cross section)

• cre_params (ParamsDict or None) – dictionary of parameters for CRE estimation. Run tw.cre_params().describe_all() for descriptions of all valid parameters. None is equivalent to tw.cre_params().

fit(rng=None)

Run CRE solver.

Parameters

rng (np.random.Generator or None) – NumPy random number generator; None is equivalent to np.random.default_rng(None)