kpss

Purpose

Test for stationarity using a Lagrange Multiplier score statistic.

Format

{ tstat, crit } = kpss(y, max_lags[, trend, qsk, auto, print_out])
Parameters:
  • y (Nx1 vector) – data.
  • max_lags (scalar) – Optional input, if max_lags <= 0, maximum lag set using Schwert criterion; if -1, Schwert criterion = 12; if 0, Schwert criterion = 4; else if max_lags > 0, maximum lag = max_lags. Default = 0.
  • trend (scalar) – Optional input, 0 no trend, 1 trend. Default = 0.
  • qsk (scalar) – Optional input, if nonzero, quadratic spectral kernel is used. Default = 0.
  • auto (scalar) – Optional input, if nonzero, automatic maxlags computed. Default = 1.
  • print_out (scalar) – Optional input, if nonzero, intermediate quantities printed to the screen. Default = 1.
Returns:
  • tstat (matrix) – test statistic for each lag.
  • crit (matrix) – Elliot, Rothenberg and Stock (1996) critical values for the GLS detrended unit root test at the 1%, 2.5%, 5%, and 10% significance level.

Example

new;
cls;
library tsmt;

// Load data
npdb = loadd( getGAUSSHome("pkgs/tsmt/examples/nelsonplosser.gdat") );
yt = packr(npdb[., "lrgnp"]);

// Test using basic KPSS testing: Trend stationary
// Step One: Set-up testing parameters
// Maximum lags to include
max_lags = 5;

// Include trend
trend = 1;

// Use quadratic spectral kernel
qsk = 1;

// Automatic maxlag computation
auto = 1;

// Print results to screen
print_out = 1;

// Running KPSS test
{ mat, crit } = kpss(yt, max_lags, trend, qsk, auto, print_out);

print "The tstats for all possible lags:";
mat;

print "Critical values:";
crit;

Library

tsmt

Source

kpss.src