How Does Oil Future Price Imply Bunker Price—Cointegration and Prediction Analysis

Round 1

Reviewer 1 Report

The paper attempts to investigate whether there is a cointegration relationship between MGO bunker oil prices and crude oil futures contract prices. The authors estimate ARMA, ARMAX, VAR, and VECM models and attempt to analyze the forecasting strength of these models. The paper needs to be written concisely. The econometric strength of the current paper is weak. The following comments are given to help improve the paper.  

1. The introduction of the paper is poorly down. I do not see the research gap explored. It is not clear what the research objectives of this paper are. The organization of the paper is missing in the introduction.
2. The current organization is poor. Section 3 and the sections and subsections that follow need to be reorganized.
3. The methodology needs to be positioned well with econometric substance. The authors intended to use cointegration which leads also to the specification of the VECM. The concept is missing. For instance, what is cointegration and why is the need for it? I expected a much better elaborative explanation leading to the specification of the VECM method. The authors should reference Mensah, E. K., Triacca, U., Bondzie, E. A., & Fosu, G. O. (2016). Crude oil price, exchange rate and gross domestic product nexus in an emerging market: A cointegration analysis. Opec energy review, 40(2), 212-231.
4. The test for stationarity is carried out with the ADF test only. This present only one stop no further option as against other tests like the Philip Peron test and the KPSS test. At least, two tests for comparison are needed for the series. Additionally, the hypothesis and critical values for which the trend in the series are against tested must be specified.
5. The test of cointegration is empiric and I suggest must come after the methodology and concept are well explained. It is poorly done how the cointegration test was done. What test was used? Is it Johansen cointegration rank test? This must be specified with details on the hypothesis for the trace statistic and the maximum eigenvalue statistic and the rank value on which the cointegration result relied on.
6. How was the lag length chosen or specified for the ARMA/ARMAX/VAR/VECM model? Did the authors rely on the information criterion for specification? What does mean the lag = 4 in Table 3. The cointegration test at different rank in correspondence with the lag length is spurious and seem incorrect.
7. The model estimation sub-section should be the main section as empirical analysis/results. As the title of the paper suggests, it is intriguing how the authors perform the cointegration test and estimate VECM without performing short-run and long-run analyses. Where is the error correction term in Table 4. What does it say? This again begs the lack of hypothesis and the concept of cointegration.
8. The theoretical framework for the impulse response function should be given. This can be included in the methodology. In this context, the authors should explain how over time, the endogenous variable will react to exogenous impulses or shocks. The sentence “The focus of impulse response analysis is how one endogenous variable in the model will impact the other variables” is not correct. Authors can reference Jin, X., Lin, S. X., & Tamvakis, M. (2012). Volatility transmission and volatility impulse response functions in crude oil markets. Energy economics, 34(6), 2125-2134.
9. Section 6 is an empiric rather than a discussion. I will suggest restructuring this in the organization.
10. It appears to me the paper is structured to focus on comparing the forecastability of the listed models. The title of the paper speaks differently.

Author Response

If you are not satisfied our responses, we can revise our paper in a further step.

Author Response File: Author Response.docx

Reviewer 2 Report

Introduction

1. The section of introduction should include: the context of the study, the gap in literature that the present paper intends to cover, which is the used methodology, which are the main results presented in short, which is the originality of this paper,  the main implication policy of these results and a short plan of the rest of the paper.
2. The current form of the introduction doesn’t correspond with these general rules. In addition the plan of the rest of the paper should be put at the end of Introduction, not at the end of literature review.

 

Literature review

1. The authors should pay attention to the way in which realized the revised on the literature on the topic in order to avoid enumeration of researches but rather to have a critical review of literature. The section of literature review should be reorganized in order to get a critical review of the research field.

Results and discussion

2. I suggest to highlight two separate parts such as: Results and discussion
3. The Sections 4 and 5 ( Model specifications and estimation and 5. Out-of-sample Prediction Performance) should be joined under a single section titled Results
4. I suggest to put the robustness checks within the section of Results (in this view to highlight 2 sections: main results and robustness checks)
5. The findings should be aligned with other findings from literature

 

Conclusion

This section should be completed with the limits of the research and the way in which these limits will be addressed in the future.

 

Other:

2. From where are the statistical data in rows 31 to 37 on the percentage change in prices obtained? Must be very clear for any reader -  even unfamiliar with the subject
3. The same question for rows 171 – 177
4. The same question for rows Figure 1, Table 1 and Table 3
5. Table 2 is staggered on two pages. To avoid such format issues, I recommend grouping Figure 1, Table 1 and Table 2 and Table 3 in Annex 1 at the end of the article, with reference to the Annex 1 in the body of the article / chapter.
6. Even if the difference between the ARMA / ARMAX models is explained in subchapter 4.1, I recommend the introduction of a single sentence at the beginning of the chapter 4, with reference to established sources (referring to the two models
7. I recommend placing Table 4 in Annex 2, with reference to the Annex 2 in the body of the article / chapter
8. I recommend placing Table 5 and Table 5 in Annex 3, with reference to the Annex 3 in the body of the article / chapter
9. I recommend placing Figure 2, Table 7, Table 8 and Table 9 in Annex 4, with reference to the Annex 4 in the body of the article / chapter
10. I recommend formulating (at least one) concrete proposal regarding the result of the study - related to addressability (to whom is the study addressed - for whom is it useful?)

References

1 The reference in the text don’t address the rules of the Energies journal. They should be aligned with the requirement of the journal.

Author Response

If you are not satisfied with our responses, we can revise our paper in a further step.

Author Response File: Author Response.docx

Reviewer 3 Report

The introduction must be rewritten, highlighting the study background, aim and structure. The literature review should be apart, being necessary to be extended.

The conclusions can be improved.

Author Response

If you are not satisfied with our responses, we can revise our paper in a further step.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The authors have responded to several of my earlier comments and made a substantial improvement to the manuscript. I still believe the paper must be improved and the following comments are further given to guide the authors.

1. I emphasize again that this paper is about cointegration and the concept must be explained otherwise, it is difficult for the reader to understand the paper. Authors must give conceptual meaning to cointegration from the econometric perspective like “Cointegration exists among two or more non-stationary time series if they have the same order of integration, and the non-trivial linear combination of their series is stationary” or “The test for cointegration among the variables is necessary to see if the variables share common trend including long term relationship”. Put differently, the integration (I) of the variables is I(1) and I(0) or stationary after differencing. Authors must explain and differentiate with ARDL model which is applicable irrespective of the order of integration of the underlying variables: I(0) and/or I(1) except for the presence of I(2) or the Toda and Yahamoto model. Reference should be made to Mensah, E. K., Asamoah, L. A., & Ahiadorme, J. W. (2021). On the impact of exchange rate uncertainty on private investment in Ghana. International Journal of Finance & Economics, 26(1), 208-217. On this where cointegration and ARDL strategies have been used and this differentiation on the order of integration is made. I suggest this to be provided as a beginning explanation in Section 3.2. In addition to this, the last paragraph of Section 2 provides a sublime introduction to cointegration. This can be moved to Section 1, line 38 before “At present, 38 there is no literature…” and before the paragraph in line 41.
2. The sentence “There is no literature to study the bunker price considering other energy markets.” Can be phrased as “there is limited or not enough literature to …”
3. The early paragraph of the impulse response analysis in Section 5.1 is not a discussion but a methodology. Until line 373, I suggest this part be moved to the methodology section under a similar subsection title.
4. One objective of employing the cointegrated vector error correction model (VECM) was to study the short-run dynamics and long-run effects of a change in MGO and oil future contract variables. This is not explained yet results are provided in Table 5. I recommend to the authors to provide a sentence or two to explain short and long-term dynamics.
5. Much as the authors emphasize that there is no literature to study the bunker price considering other energy markets, there is a vast literature on oil price and other economic variables. Besides limited current literature is provided. For instance, only three papers are current from 2018 and 2019 in this paper. I suggest the authors include additional current relevant literature studies on the topic from particular oil future price dynamics in the energy market.

Author Response

If your have any further suggestions, please let we know and we can make this paper better.

Author Response File: Author Response.docx

Reviewer 2 Report

The authors have included all adressed suggestions. I am satisfied with the present form of the paper. Good job!

 

Author Response

Thank you for reviewing our manuscript!

Author Response File: Author Response.docx

Reviewer 3 Report

I consider the paper was improved, so it can be considered for publication in the actual form.

Author Response

Thank you for reviewing our manuscript!

Author Response File: Author Response.docx