Scribd
Upload
24 views3 pages

ARDL Cointegration

Research Paper

Uploaded by

Agha
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
24 views3 pages

ARDL Cointegration

Research Paper

Uploaded by

Agha
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 3

ARDL cointegration

The Autoregressive Distributed Lag (ARDL) model is a popular econometric approach for
analyzing both long-term and short-term relationships among variables, especially when dealing
with cointegration in small samples. Unlike traditional cointegration tests, the ARDL approach is
flexible because it can be used regardless of whether the variables are integrated of order zero
(I(0)), integrated of order one (I(1)), or a mix of both (though it cannot include I(2) variables).

Key Features of ARDL Cointegration

1. Flexibility with Different Orders of Integration: ARDL can be used with variables that
are a mix of I (0) and I (1), which is useful because many cointegration techniques
require that all variables be I (1).
2. Separate Long- and Short-term Estimation: The ARDL model can provide separate
estimates for long-run relationships (cointegration) and short-term dynamics (error
correction), allowing you to see both immediate and persistent effects of explanatory
variables.
3. Small Sample Suitability: ARDL cointegration tests are known for being effective even
with small samples, making them preferable in cases where sample size is limited.

ARDL Model Specification

For two variables Y (dependent) and X (independent), an ARDL model can be specified as:

where:

 Yt is the dependent variable at time t,


 Xt is the independent variable at time t,
 p and q are the lag lengths for Y and X, respectively,
 βi and δj are coefficients to be estimated,
 ϵt is the error term.

1
Steps for ARDL Cointegration Analysis

1. Test for Stationarity: Although ARDL can handle a mix of I (0) and I (1) variables, it is
essential to confirm that no variable is I(2) (integrated of order 2) to avoid invalid results.
Stationarity tests like the Augmented Dickey-Fuller (ADF) or Phillips-Perron (PP)
tests can be used.
2. Choose Lag Lengths: Select appropriate lag lengths for both the dependent and
independent variables. Criteria such as the Akaike Information Criterion (AIC) or the
Schwarz Criterion or Bayesian Information Criterion (BIC) are commonly used for this
purpose.
3. Estimate the ARDL Model: Estimate the ARDL model using the chosen lags for each
variable.
4. Make sure that the errors of this model are serially independent.
use the LM test to test the null hypothesis that the errors are serially independent, against
the alternative hypothesis that the errors are (either) AR(m) or MA(m), for m = 1, 2,
3,...., etc.
5. Make sure that the model is "dynamically stable".
check that all of the inverse roots of the characteristic equation associated with our model
lie strictly inside the unit circle.

6. Bounds Test for Cointegration: Conduct the Bounds test for cointegration to check if a
long-term relationship exists between the variables. This test compares an F-statistic to
two critical values (lower and upper bounds).
o If the F-statistic exceeds the upper bound, there is evidence of cointegration
(long-run relationship).
o If it falls below the lower bound, there is no evidence of cointegration.
o If it lies between the bounds, the test is inconclusive.
7. Long-run and Short-run Coefficients: If cointegration is confirmed, calculate the long-
run coefficients from the ARDL model and the short-run dynamics using an Error
Correction Model (ECM).
8. Use the results of the models estimated in step 7 to measure short-run dynamic effects,
and the long-run equilibrating relationship between the variables.

2
In the ECM, the error correction term (ECT) measures the speed at which deviations
from the long-run equilibrium are corrected. A significant ECT coefficient (usually
negative) implies that the system returns to equilibrium after a shock.

ARDL Error Correction Model (ECM)

Once cointegration is established, the ARDL model can be re-parameterized into an ECM to
capture the short-term adjustments back to the long-run equilibrium:

where:

 Δ represents the first difference operator (indicating short-term dynamics),


 ECTt−1 is the error correction term derived from the long-run relationship,
 λ is the adjustment coefficient, which shows the speed of adjustment back to the long-run
equilibrium. A negative and significant λ indicates that the variables revert to their long-
run relationship.

Advantages of the ARDL Approach

 Simplicity and Flexibility: Works with a mix of I(0)I(0)I(0) and I(1)I(1)I(1) variables.
 Separate Short- and Long-run Effects: Allows separate estimation of long-term and
short-term effects.
 Effective in Small Samples: Performs well with limited sample sizes.

Limitations of the ARDL Approach

 Not Suitable for I(2) Variables: Variables integrated of order two can invalidate the
results.
 Model Complexity: Choosing appropriate lags for each variable can be complex, and the
results can be sensitive to lag selection.
 Bounds Test Ambiguity: The results of the Bounds test can be inconclusive when the F-
statistic falls between the bounds.

You might also like