cdTest¶
Purpose¶
Runs cross-sectional dependence, CD, tests for panel data. The test statistics constructed from the mean t-statistic has an asymptotic standardized normal distribution and tests the null hypothesis that all series are I(1) against the alternative that all series are I(0)
Format¶
-
cd =
cdTest(res, model, grp_vector)¶ Parameters: - res (Nx1 matrix) – residuals from panel data regression.
- model (scalar) – indicates which test to run: 1 for Pesaran’s test, 2 for Friedman’s test, 3 for Frees’ test. Note: the only model appropriate for unbalanced data is model 1. This model will run by default if unbalanced panel data is detected.
- grp (Nx1 matrix) – group indicator variable.
Returns: cd (matrix) – test statistic
Examples¶
Example One: Pesaran Test¶
new;
cls;
library tsmt;
/*************************************
//Generate data
**************************************/
// Set panel size parameters
N = 48;
T = 17;
// Simulate balanced group vector
grp = vec(reshape(seqa(1, 1, N), 17, 48));
// Simulate random stacked residuals
rndseed 90782;
y = rndn(T*N, 1);
/*************************************
/Run tests
**************************************/
// Run pesaran test [model = 1]
z1 = cdTest(y, 1, grp);
This prints the following results to the Command Window:
Test: Pesaran Test (2004)
Timespan: Unknown
Ho: Cross-sectional independence
N. Obs: 17
N. Groups: 48
Panel Type: Balanced
============================================================
CD-stat 0.780
P-val 0.218
============================================================
Cannot reject the null hypothesis of cross-sectional
independence.
Example Two: Friedman Test¶
// Run Friedman test [model = 2]
z2 = cdTest(y, 2, grp);
// Run Frees test [model = 3]
z3 = cdTest(y, 3, grp);
The results printed to screen are
Test: Friedman (1937)
Test Variable:
Timespan: Unknown
Ho: Cross-sectional independence
Number of breaks: None
N. Obs: 17
N. Groups: 48
Panel Type: Balanced
============================================================
CD-stat 21.771
P-val 0.151
============================================================
Cannot reject the null hypothesis of cross-sectional
independence.
Library¶
tsmt