ralslm_breaks#

Purpose#

Computes the Augmented Dickey-Fuller unit root test with 1 or 2 breaks and the RALS technique for non-normal errors.

Format#

{ LM_min, tb, p, cv_lm, RALS_lm, p2, cv_ralslm } = ralslm_breaks(y, model, nbreaks[, pmax, ic, trimm])#
Parameters:
  • y (Nx1 matrix) – Time series data to be tested.

  • model (Scalar) –

    Model to be implemented.

    1

    Level break (Meng, Im, Lee, & Tieslau (2014))

    2

    Level and trend break (Meng, Lee, & Payne (2017))

  • pmax (Scalar) – Optional, the maximum number of lags for \(\Delta y\). Default = 8.

  • ic (Scalar) –

    Optional, the information criterion used for choosing lags. Default = 3.

    1

    Akaike.

    2

    Schwarz.

    3

    t-stat significance.

  • trimm (Scalar) – Optional, trimming rate. Default = 0.10.

Nbreaks:

Number of breaks (1 or 2).

Returns:
  • LM_min (Scalar) – Minimum LM statistic without RALS terms.

  • tb (Vector) – Location of estimated breaks.

  • p (Scalar) – Optimal number of lags selected by determined information criterion.

  • cv_lm (Vector) – 1%, 5%, 10% critical values for LM test based on response surfaces.

  • RALS_lm (Scalar) – LM statistic based on RALS procedure and breaks.

  • rho2 (Scalar) – The estimated rho square.

  • cv_ralslm (Vector) – 1%, 5%, 10% critical values for RALS-LM test basen the estimated rho2.

Examples#

library tspdlib;

// Load date file
y = loadd(getGAUSSHome() $+ "pkgs/tspdlib/examples/TSe.dat");

// With constant
model = 1;
{ LM_min, tb, p, cv_lm, RALS_lm, rho2, cv_ralslm } = RALSLM_breaks(y, model, nbreaks);

Source#

rals_lm_breaks.src

See also

Functions lmkpss(), lm_1break(), lm_2breaks(), ralslm()