How to calculate the long run causality between exports and economic growth?
i am working on export-led growth policy for Pakistan. For short run causal link i have used the VAR model but i am little confused how the calculate the long run causality. please guide me in this matter
thanks and best regards
Sayed Irshad Hussain
https://www.researchgate.net/home
Vector Error Correction model (VEC) can be used to check on the longrun relationship. You can also use Impulse Response Functions (IRFs) to check on the same. We have done a similar study for Zimbabwe, you can check our paper "Export Sector Contribution to Economic Growth in Zimbabwe: A Causality Analysis"
http://ssrn.com/abstract=2687414
Dear Sayed Irshad Hussain
Greetings,
Please find the attached file about causality between exports and economic growth , may be help you ,
Best Regards,
If the variables display a mixed order of intergration (i.e. a mixture of I(1) and I(0) variables), then I would advice to run an ARDL bounds testing cointegration model (see Pesaran et al., 2001). Although correlation does not imply causation, cointegration does. I hope this helps.
Dear
Glauco De Vita
Then how we obtain the causal relationship from ARDL model
thanks and best regards
Sayed Irshad Hussain
If all the variables are non stationary at level but become I(1) ( same order of integration) on differencing, then its best to run a Johansen test and if cointegration is established , run a VECM to determine the short run and long run causality.
Good morning;
My comments are more on the underlying economic model, to complement the excellent suggestions on the econometrics.
From a macro perspective (according to the Post-Keynesian school), the contribution of the external sector to growth may be derived from the Thirwall Law. Id est, a tension between exports (which have an accounting impact of GDP on the short term, but --more importantly-- provide much needed hard currency) and imports (required in particular for investment goods, so needed for long term growth). The tension comes from the negative sign of imports in the GDP identity and the net effect boils down to comparing long-term semi-elasticity, the econometrics is similar to what has been indicated in the other comments.
In Thirwall, exports and imports are technically independent (even if they may be positively correlated through changes in the Balance of Payments constraint). More recently (since the 1990s), we know that exports (esp. of manufacture) are increasingly intensive in imported intermediate goods ("Trade in Value-Added" along Global Value Chains).
So the right approach, using a simple export-led model instead of the more complete Thirwall model, would be to measure the net impact of exports (gross exports, minus the part of imports required for producing the exports) instead of only gross exports. How to do it, I do not know. Personally, I'd use the full Thirwall model where imports are explicitly specified and provides a better economic background.
My two cents,
Best regards,
Hubert
I would suggest you look at the following work by Jim Feyrer, who has a well-founded identification strategy for answering precisely this question using time-varying measures of geography as instrumental variables:
http://www.nber.org/papers/w14910.pdf
http://www.nber.org/papers/w15557.pdf
Best of luck!
You can examine the long run equilibrium relation through vector error correction model (VECM). Before running this model, you have to conduct Johansen cointegration test.
In addition to VECM, you can examine the causality through Granger causality test.
Before conducting all these tests, you have to test the stationarity of data.
The ARDL bounds testing approach establishes whether is a LR cointegrating relationship, and the direction of that relationship. Unlike simple correlation, that cointegration does prove causation!
I also suggest you to use VAR model and cointegration technique to estimate the long-run relatoinship between exports and economic growth.
For short term causal relationships , the best method is the Granger Causality Test.
For Long term equilibrium relationship, the best method is VAR . But first you need to run unit root tests in order to check on the performing conditions.
- Badges
- Science topic
- Similar topics
- Geoscience
- Asia
- India