nag_stat_prob_dickey_fuller_unit (g01ew) returns the probability associated with the lower tail of the distribution for the Dickey–Fuller unit root test statistic.

Syntax

[pvalue, state, ifail] = g01ew(type, n, ts, 'method', method, 'nsamp', nsamp, 'state', state)

[pvalue, state, ifail] = nag_stat_prob_dickey_fuller_unit(type, n, ts, 'method', method, 'nsamp', nsamp, 'state', state)

Description

If the root of the characteristic equation for a time series is one then that series is said to have a unit root. Such series are nonstationary. nag_stat_prob_dickey_fuller_unit (g01ew) is designed to be called after nag_tsa_uni_dickey_fuller_unit (g13aw) and returns the probability associated with one of three types of (augmented) Dickey–Fuller test statistic:

τ

τ_{μ}

τ_{τ}

, used to test for a unit root, a unit root with drift or a unit root with drift and a deterministic time trend, respectively. The three types of test statistic are constructed as follows:

To test whether a time series,

y_{t}

, for

t = 1, 2, \dots, n

, has a unit root the regression model

\nabla y_{t} = β_{1} y_{t - 1} + \sum_{i = 1}^{p - 1} δ_{i} \nabla y_{t - i} + ε_{t}

is fit and the test statistic

τ

constructed as

τ = \frac{{\hat{β}}_{1}}{σ_{11}}

where

\nabla

is the difference operator, with

\nabla y_{t} = y_{t} - y_{t - 1}

, and where

{\hat{β}}_{1}

and

σ_{11}

are the least squares estimate and associated standard error for

β_{1}

respectively.

To test for a unit root with drift the regression model

\nabla y_{t} = β_{1} y_{t - 1} + \sum_{i = 1}^{p - 1} δ_{i} \nabla y_{t - i} + α + ε_{t}

is fit and the test statistic

τ_{μ}

constructed as

τ_{μ} = \frac{{\hat{β}}_{1}}{σ_{11}} .

To test for a unit root with drift and deterministic time trend the regression model

\nabla y_{t} = β_{1} y_{t - 1} + \sum_{i = 1}^{p - 1} δ_{i} \nabla y_{t - i} + α + β_{2} t + ε_{t}

is fit and the test statistic

τ_{τ}

constructed as

τ_{τ} = \frac{{\hat{β}}_{1}}{σ_{11}} .

All three test statistics:

τ

τ_{μ}

and

τ_{τ}

can be calculated using nag_tsa_uni_dickey_fuller_unit (g13aw).

The probability distributions of these statistics are nonstandard and are a function of the length of the series of interest,

n

. The probability associated with a given test statistic, for a given

n

, can therefore only be calculated by simulation as described in Dickey and Fuller (1979). However, such simulations require a significant number of iterations and are therefore prohibitively expensive in terms of the time taken. As such nag_stat_prob_dickey_fuller_unit (g01ew) also allows the probability to be interpolated from a look-up table. Two such tables are provided, one from Dickey (1976) and one constructed as described in Further Comments. The three different methods of obtaining an estimate of the probability can be chosen via the method argument. Unless there is a specific reason for choosing otherwise,

method = 1

should be used.

References

Dickey A D (1976) Estimation and hypothesis testing in nonstationary time series PhD Thesis Iowa State University, Ames, Iowa

Dickey A D and Fuller W A (1979) Distribution of the estimators for autoregressive time series with a unit root J. Am. Stat. Assoc. 74 366 427–431

Parameters

Compulsory Input Parameters

1: $type$ – int64int32nag_int scalar

The type of test statistic, supplied in ts.

Constraint:

type = 1

2

3

2: $n$ – int64int32nag_int scalar

n

, the length of the time series used to calculate the test statistic.

Constraints:

if $method \neq 3$ , $n > 0$ ;
if $method = 3$ and $type = 1$ , $n > 2$ ;
if $method = 3$ and $type = 2$ , $n > 3$ ;
if $method = 3$ and $type = 3$ , $n > 4$ .

3: $ts$ – double scalar

The Dickey–Fuller test statistic for which the probability is required. If

$type = 1$: ts must contain $τ$ .
$type = 2$: ts must contain $τ_{μ}$ .
$type = 3$: ts must contain $τ_{τ}$ .

If the test statistic was calculated using nag_tsa_uni_dickey_fuller_unit (g13aw) the value of type and n must not change between calls to nag_stat_prob_dickey_fuller_unit (g01ew) and nag_tsa_uni_dickey_fuller_unit (g13aw).

Optional Input Parameters

1: $method$ – int64int32nag_int scalar

Default:

1

The method used to calculate the probability.

$method = 1$: The probability is interpolated from a look-up table, whose values were obtained via simulation.
$method = 2$: The probability is interpolated from a look-up table, whose values were obtained from Dickey (1976).
$method = 3$: The probability is obtained via simulation.

The probability calculated from the look-up table should give sufficient accuracy for most applications.

Constraint:

method = 1

2

3

2: $nsamp$ – int64int32nag_int scalar

Default:

100000

method = 3

, the number of samples used in the simulation; otherwise nsamp is not referenced and need not be set.

Constraint: if

method = 3

nsamp > 0

3: $state (:)$ – int64int32nag_int array

method = 3

, state must contain information on the selected base generator and its current state; otherwise state is not referenced and need not be set.

Output Parameters

1: $pvalue$ – double scalar
2: $state (:)$ – int64int32nag_int array: If $method = 3$ , state contains updated information on the state of the generator otherwise a zero length vector is returned.
3: $ifail$ – int64int32nag_int scalar: $ifail = 0$ unless the function detects an error (see Error Indicators and Warnings).

Error Indicators and Warnings

Errors or warnings detected by the function:

Cases prefixed with W are classified as warnings and do not generate an error of type NAG:error_n. See nag_issue_warnings.

$ifail = 11$: Constraint: $method = 1$ , $2$ or $3$ .

$ifail = 21$: Constraint: $type = 1$ , $2$ or $3$ .

$ifail = 31$: Constraint: if $method \neq 3$ , $n > 0$ .

Constraint: if $method = 3$ and $type = 1$ , $n > 2$ .

Constraint: if $method = 3$ and $type = 2$ , $n > 3$ .

Constraint: if $method = 3$ and $type = 3$ , $n > 4$ .

$ifail = 51$: Constraint: if $method = 3$ , $nsamp > 0$ .

$ifail = 61$: On entry, $method = 3$ and the state vector has been corrupted or not initialized.

W $ifail = 201$: The supplied input values were outside the range of at least one look-up table, therefore extrapolation was used.

$ifail = - 99$: An unexpected error has been triggered by this routine. Please contact NAG.

$ifail = - 399$: Your licence key may have expired or may not have been installed correctly.

$ifail = - 999$: Dynamic memory allocation failed.

Accuracy

When

method = 1

, the probability returned by this function is unlikely to be accurate to more than

4

5

decimal places, for

method = 2

this accuracy is likely to drop to

2

3

decimal places (see Further Comments for details on how these probabilities are constructed). In both cases the accuracy of the probability is likely to be lower when extrapolation is used, particularly for small values of n (less than around

15

). When

method = 3

the accuracy of the returned probability is controlled by the number of simulations performed (i.e., the value of nsamp used).

Further Comments

When

method = 1

2

the probability returned is constructed by interpolating from a series of look-up tables. In the case of

method = 2

the look-up tables are taken directly from Dickey (1976) and the interpolation is carried out using nag_interp_2d_scat (e01sa) and nag_interp_2d_scat_eval (e01sb) . For

method = 1

the look-up tables were constructed as follows:

(i)	A sample size, $n$ was chosen.
(ii)	$2^{28}$ simulations were run.
(iii)	At each simulation, a time series was constructed as described in chapter five of Dickey (1976). The relevant test statistic was then calculated for each of these time series.
(iv)	A series of quantiles were calculated from the sample of $2^{28}$ test statistics. The quantiles were calculated at intervals of $0.0005$ between $0.0005$ and $0.9995$ .
(v)	A spline was fit to the quantiles using nag_fit_1dspline_auto (e02be).

This process was repeated for

n = 25, 50, 75, 100, 150, 200, 250, 300, 350, 400, 450, 500, 600, 700, 800, 900, 1000, 1500, 2000, 2500, 5000, 10000

, resulting in

22

splines.

Given the

22

splines, and a user-supplied sample size,

n

and test statistic,

τ

, an estimated

p

-value is calculated as follows:

(i)	Evaluate each of the $22$ splines, at $τ$ , using nag_fit_1dspline_auto (e02be). If, for a particular spline, the supplied value of $τ$ lies outside of the range of the simulated data, then a third-order Taylor expansion is used to extrapolate, with the derivatives being calculated using nag_fit_1dspline_deriv (e02bc).
(ii)	Fit a spline through these $22$ points using nag_interp_1d_monotonic (e01be).
(iii)	Estimate the $p$ -value using nag_interp_1d_monotonic_eval (e01bf).

Example

See Example in nag_tsa_uni_dickey_fuller_unit (g13aw).

Open in the MATLAB editor: g01ew_example

function g01ew_example


fprintf('g01ew example results\n\n');

% Test type
type = int64(1);

% Order of the AR process
p = int64(1);

% Time series
y = [ -217; -177; -166; -136; -110;  -95;  -64;  -37;  -14;  -25;
       -51;  -62;  -73;  -88; -113; -120;  -83;  -33;  -19;   21;
        17;   44;   44;   78;   88;  122;  126;  114;   85;   64];

% Calculate the Dickey-Fuller test statistic
[ts,ifail] = g13aw(type,p,y);

% The p-value routine can issue a warning, but still return
% sensible results, so save current warning state and turn warnings on
warn_state = nag_issue_warnings();
nag_issue_warnings(true);

% Get the associated p-value
n = int64(size(y,1));
[pvalue,~,ifail] = g01ew(type,n,ts);

% Reset the warning state to its initial value
nag_issue_warnings(warn_state);

% Print the results
fprintf('Dickey-Fuller test statistic     = %6.3f\n', ts);
fprintf('associated p-value               = %6.3f\n', pvalue);

g01ew example results

Dickey-Fuller test statistic     = -2.540
associated p-value               =  0.013

PDF version (NAG web site, 64-bit version, 64-bit version)

Chapter Contents

Chapter Introduction

NAG Toolbox