HADRI Procedure |
@HADRI implements the LM tests for unit roots in panel data from Hadri(2000). This has stationarity as the null, and rejects (in favor of some unit roots) if the detrended or de-meaned data are too persistent. It is, in effect, an aggregation of KPSS test statistics, using large N asymptotics applied to the average of the KPSS statistics. Alternative tests for unit roots in panel data are @Breitung, @LevinLin and @IPShin.
@HADRI( options ) series start end
Parameters
series |
series to analyze |
start, end |
range of series to use. By default, the defined range of series. |
Options
DET=NONE/[CONSTANT]/TREND
Individual-specific deterministic components to remove. Note NONE is an error and is provided simply to keep the form of the option similar to related procedures.
SMPL=standard SMPL option[not used]
VARIANCE=[HOMOGENEOUS]/HETEROGENEOUS/ROBUST
LWINDOW=NEWEY/BARTLETT/FLAT/PARZEN/QUADRATIC
LAGS=# of lags for the lag window (or bandwidth for LWINDOW=QUADRATIC) [Schwert's]
If VARIANCE=HOMOGENEOUS, the variances are assumed to be the same in each cross section. HETEROGENEOUS is for no serial correlation, but differing variances. ROBUST allows for heterogeneous serial correlation, with the long-run variance computed according to the LWINDOW and LAGS options.
TITLE="title for output" ["Hadri Unit Root Test: Series ..."]
[PRINT]/NOPRINT
Variables Defined
%CDSTAT |
the Hadri test statistic (REAL) |
%SIGNIF |
the significance level of %CDSTAT as a one-tailed N(0,1) (REAL) |
%NGROUP |
number of individuals (INTEGER) |
Example
This does the Hadri test on the real exchange rate from OECD data. The data have had time effects removed. The calculation uses HAC covariance estimates for the variance with a Bartlett (Newey-West) window of width 5.
open data pennxrate.dta
calendar(panelobs=34,a) 1970
data(format=dta) 1//1970:01 151//2003:01 year xrate ppp id capt realxrate lnrxrate oecd g7
panel(entry=1.0,time=-1.0,smpl=oecd) lnrxrate / cxrate
@hadri(smpl=oecd,det=constant,var=robust,lwindow=bartlett,lags=5) cxrate
Sample Output
This is the output from the example. This rejects stationarity in favor of at least some unit roots.
Hadri LM Unit Root Test: Series CXRATE
Test has large N, large T asymptotics
Null is Stationary. Alternative is Some Unit Roots
Individual Specific Components: Constant
Robust to Serial Correlation, Bartlett(Newey-West)(5)
N 27
T 34
Z 8.835604
Signif 0.000000
Copyright © 2025 Thomas A. Doan