Modeling Co-Movement among Different Agricultural Commodity Markets: A Copula-GARCH Approach
Round 1
Reviewer 1 Report
About the submission with the title "Modeling co-movement among different agricultural commodity markets: A copula-GARCH approach" I have the following comments:
The literature review must to be significantly improved and updated. There are many documents about these topics.
The methodology chosen should be clearly explained, namely, in terms of options made and must be supported by scientific literature. The adequacy of these methodologies for the objectives designed should be, also, explained.
The kind of data used need to be identified (time series, panel data, ...).
There are many statistical tests related with the regressions presented do not present, as those related, for example, with autocorrelation, multicollinearity, adequacy of the model (RESET test)....
The linkages between the methodologies described and the results must to be showed.
In the implications and suggestions section the Authors should be more specific about the practical implications for the several stakeholders and policy instruments.
Author Response
The literature review must to be significantly improved and updated. There are many documents about these topics.
Answer: According to another reviewer comment, I have cited more reference.
The methodology chosen should be clearly explained, namely, in terms of options made and must be supported by scientific literature. The adequacy of these methodologies for the objectives designed should be, also, explained.
Answer：Thank you very much for pointing out this very important point. I have revised it, please check the page 3.
The kind of data used need to be identified (time series, panel data, ...).
Answer: The kind of data used is time series.
There are many statistical tests related with the regressions presented do not present, as those related, for example, with autocorrelation, multicollinearity, adequacy of the model (RESET test).
Answer: Thank you for your suggestion. The log-return of spot prices have autocorrelation and ARCH effect, so this paper first apply ARMA-GJRGARCH model. Before I put the data to copula mode, I have test the by LB test and KS test. The test show that all of the first four moments are serially independent, while the results of the KS test indicate that all series are distributed as uniform(0,1).
The linkages between the methodologies described and the results must to be showed.
Answer: I don't understand this opinion. What are the linkages between the methodologies described and the results?
(6)In the implications and suggestions section the Authors should be more specific about the practical implications for the several stakeholders and policy instruments.
Answer：Thank you very much for pointing out this very important point. I have revised it, please check the page 16-18.
Reviewer 2 Report
I suggest to replace "agricultural commodities" with "agricultural grain commodities", because the first is too broad, including many other commodities as bananas, meat, etc.; grain commodities regards both cereals and soya.
There are still some reference lacks: the existing international warning services must be cited.
Author Response
First Referee's Comment：
I suggest to replace "agricultural commodities" with "agricultural grain commodities", because the first is too broad, including many other commodities as bananas, meat, etc.; grain commodities regards both cereals and soya.
Answer: Many thanks for your suggestion. I check some reference, most of papers are "agricultural commodities", not "agricultural grain commodities" in the title. If you think "agricultural grain commodities" is better, I will revise them. I have replaceed "agricultural commodities" with "agricultural grain commodities" in many places.
There are still some reference lacks: the existing international warning services must be cited.
Answer: According to another reviewer comment, I have cited more reference.
Round 2
Reviewer 1 Report
The Authors need to explain the following questions referred in my previous report: " There are many statistical tests related with the regressions presented do not present, as those related, for example, with autocorrelation, multicollinearity, adequacy of the model (RESET test).
The answer presented in not enough "Answer: Thank you for your suggestion. The log-return of spot prices have autocorrelation and ARCH effect, so this paper first apply ARMA-GJRGARCH model. Before I put the data to copula mode, I have test the by LB test and KS test. The test show that all of the first four moments are serially independent, while the results of the KS test indicate that all series are distributed as uniform(0,1).
The linkages between the methodologies described and the results must to be showed.".
Because the answer is only for the autocorrelation (however I did not find any explanation about the autocorrelation in the paper) and about the other statistical problems (multicollinearity, heteroskedasticity, ...)?
Without a clarification about these aspects your results and conclusions may be compromised.
About "Answer: I don't understand this opinion. What are the linkages between the methodologies described and the results?", my suggestions is you explain your results considering your methodology widely described. For example, it could be interesting you presnt the model/equations described in the methodology section for each regression table. It could be important, also, explain the expected values by the respective model and the obtianed and benchmark this with already published literature.
Author Response
Answer： Thank you for your comment. Before I put the data to ARMA-GJRGARCH model, I have done the ACF and PACF figure to check the autocorrelation. This paper has 12 series, so I did not include the pictures of ACF and PACF in the article. The RESET test is regression specification error test, is it? I I downloaded some literature of GARCH-copula model, such as Reboredo(2012), J et al. (2018), Bedoui et al. (2019). (There's a lot of literature, and I'm not going to list it all here.) I read them carefully and do not find that these two tests (multicollinearity, and RESET test) need to be done.
I'd like to discuss this with you. If you have any questions or suggestions, please tell me.
Reboredo. (2011) .How do crude oil prices co-move? A copula approach, Energ. Econ. 33, 948–955.
Q.,Ji, E., Bouri, D.Roubaud, S.J.H., Shahzad (2018). Risk spillover between energy and agricultural commodity markets: A dependence-switching CoVaR-copula model.Energy Economics,75, 14-27.
R., Bedoui, S., Braiek, K., Guesmi, J., Chevallier (2019).On the conditional dependence structure between oil, gold and USD exchange rates: Nested copula based GJR-GARCH model. Energy Economics, 80, Pages 876-889.
