The Determinants of Competitiveness in Global Palm Oil Trade
Robert Ackrill
Round 1
Reviewer 1 Report
The Determinants of Competitiveness in global Palm Oil Trade
The paper investigates countries’ competitiveness in palm oil trade using the symmetric RCA index and the Trade balance index. It provides a mapping of comparative advantages and test the stability of the indicators. Finally, it estimates the determinants of palm oil competitiveness using a regression analysis. The results find both positive and negative determinants of palm oil competitiveness.
General Comment
The paper investigates an interesting topic since palm oil is one of the most used commodities in the food industry and in the production of biofuels. It uses a standard analysis of revealed comparative advantages and a basic regression framework. The innovation of the paper lies only in the country coverage and time (from mid nineties to 2018) of the analysis.
There are several issues with the paper, both in terms of style and substance. The main problem refers to the empirical analysis which seems to suffer serious identification problems (endogeneity of some regressors) and overparameterization. The mapping of comparative advantage is ok but is mostly a descriptive analysis with little or no innovation. The text is quite row and should be substantially improved.
In what follows, you’ll find specific comments.
Introduction
It is ok, but there are several unclear sentences
Survey
The survey should not only provide a mere list of results but also a bit of explanation, at least for the less intuitive ones (for example FDI negative affecting beer competitiveness). In addition, a brief description of samples and empirical methodologies used should be added.
Methodology
- This sentence should be clarified:
“However, the RCA index has a problem that the RCA value is asymmetric or various (Mirzaei, Mostafa Mostafavi, 149 and Yazdani 2012). Therefore, it ignores the several implications of agricultural policies.”
How do policy implications relate to this problem?
- Provide a justification of the variables included in equations (5) and (6). Do other major studies on the issue use the same set of variables? Some of them might be endogenous, like GDP with respect to the TBI. Further, you should better explain why you introduce the Asia variable. The definition of the RSPO variable should also be better explained. To my understanding it is a dummy=1 if a country participates in this roundtable. In addition, the suffixes in the equations are confusing:
- It is not clear the type of sample you are using. I see i,t as the only suffixes, thus it seems to be a panel of exporting countries to the World. But after the equations you write “Where RSCAijt and TBIijt is the measurement of palm oil trade competitiveness in country i with year t.” with i,j, and t. Same applies to the error term. Please correct the mistake and give a better explanation of the panel structure.
- This is partial. “FDI is a factor that indirectly influences a country's competitiveness 241 through government tax regulation (Domazet and Marjanovic 2017).” In general FDIs affect countries competitiveness through several channels, most importantly technology transfer.
- The most important issue relates to the empirical strategy. With a sample of 19 countries over 22 years, you should check for stationarity in the series and use a proper methodology. In the case of non-stationary regressors you should run a cointegration analysis (see for example Esposito et al., 2017; Tsionas, 2018). However, given the many variables involved in the analysis and the relatively small sample size, you are unlikely to get good testing and estimation results. The alternative is to estimate the model in first differences. In addition, you should add country specific fixed/random effect and test which is the most suitable model using the Hausman test. Time dummies should also be added to control for Cross Sectional Dependence (which, by the way, should also be tested). Further, lacking a proper way to control for endogeneity of some regressors (namely GFDPpc, FDI, Trade openness), at least add all regressors with a lag. This would reduce the bias of estimates although not their inconsistency. Finally, the sample size for the estimates should be explained in the methodology section. I hade to read until the end of section 4 to understand that you are using only 19 countries.
Results
- Most figures are quite hard to read, I’d suggest to make summary tables with variations among subperiods to show information more efficiently.
- Please justify why the mapping of RSCA-TBI is done for all countries in the world while the determinants of competitiveness are investigated for 10 countries only. It is obvious that most countries are not produces of palm oil and thus belong to group D.
- The signs of most coefficients are against expectations and the explanation provided are not really satisfactory. I believe there are serious identification issues due to endogeneity and non-stationarity of the series.
- Why do you introduce the fixed effect estimator only in Table 5? It seem you are doing a cherry picking of the models. You should have clearly stated the reasons for that in the Methodology section. Collinearity seems to be also an issue in spite of the fact that you overlook it.
Discussion
There are several inconsistencies, for example “Domestic palm oil consumption is positively influenced by competitiveness, showing that the leading exporters must focus on the domestic market to increase competitiveness.” Isn’t it the other way round? I mean, competitiveness is the dependent variable.
This part, however, seems to be redundant as it could be integrated within the results part.
Language
The overall language of the paper should be reviewed carefully. There are many sentences that just do not sound right although the meaning can be understood. There are also other problems like lack of consistency between plural and singular words/verbs as well as several typos. A professional proof check is highly recommended.
Author Response
Dear Reviewer 1
Thank you for your valuable comments, We have already revised and confirmed through the PDF file with name "Economies_Answer for Reviewer 1"
Kind Regards
Author
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments for author File: Comments.pdf
Author Response
Dear Reviewer 2
Thank you for your valuable comments, We have already revised and confirmed through the PDF file with name "Economies_Answer for Reviewer 2"
Kind Regards
Author
Author Response File: Author Response.pdf
Reviewer 3 Report
Overall, I am afraid that the manuscript is at a very early stage and reveals serious flaws in many aspects. Below are some comments that I would like to share with the author(s) to improve the manuscript.
Above everything else, the oil palm industry requires very special natural conditions: plantation of palm oil needs a rainforest climate and a lot of land. If today Indonesia and Malaysia produce almost 90% of the world’s supply, it should be because, among others, these countries best meet such conditions. It does not make sense to talk about the competitiveness of the oil palm industry for other countries that do not at all meet the basic conditions.
The estimated regression models only include country-wide variables as explanatory variables without any consideration of the industry-specific characteristics, and do not provide any insights regarding their interactions. Why country-wide more FDI inflows, GDP, population, etc. affect negatively the country’s competitiveness of the oil palm industry? Without any control for the industry characteristics, all the proposed results seem not to be meaningful.
Furthermore, I should say that I find other previous publications written exactly in the same format and by just changing the industry data. The manuscript needs a substantial revision in motivation and methodological design, as well as in writing, to be considered for publication in international quality journals.
Author Response
Dear Reviewer 3
Thank you for your valuable comments, We have already confirmed on these comments through the PDF file with name "Economies_Answer for Reviewer 3"
Kind Regards
Author
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The revised version of the paper addressed some of the points i previously raised, however there are still some issues than need further work.
There is still a problem with the style of the text, especially in the introduction and the survey. The new sentences are often just embedded in the old text without harmonizing the structure. I recommend to ask rewrite the whole part and use a professional editor to help you restructure the sentences.
Survey:
you start talking about general studies on comparative advantage and then move to studies on Palm Oil, but within this part, the study on beer competitiveness seems to be misplaced. You should restructure this section by showing in a single paragraph studies on other products and main determinants, then focus on studies analyzing palm oil trade.
Regression analysis:
1) thanks for showing the unit root tests, however, since you introduce variables in levels, the relevant test is the on in level, thus POP, POC, EUV and TO are non-stationary. In addition, there si no test for GDPpc which is very likely to be non stationary. You should introduce the test table in the paper and introduce first differenced variables when levels are non stationary.
2) With are you stuck using the OLS? in a panel framework pooled ols mix between and within group variability and do not contro for unobserved heterogeneity. I recommend to use a starting model either a random or fixed effect model (and test which one is the most suitable using the Hausman test) and be consistent in the subsample of advanced and developing countries. Your expalation for the use of FEM in the subsample might be ok but this is a further reason to be consisten and use the same methodology for the whole sample, otherwise the comparability of the results is redcued.
3) As I said in the previous review, contemporaneous regressors might be biased, thus introduce alla variables with a lag to, at least, reduce the endogeneity bias. This might lower the significance of your results but it would make the analysis more scientifically sound. Otherwise you should talk about correlations only and not causation.
Author Response
Dear Reviewer 1
Thank you for your comment, we have already revised our manuscript. On the other hand, we also fixed our manuscript with revision in language and statistics. There are two files to submitted: 1) revised manuscript and 2) Answer to Reviewer (round 2).
Kind Regards
Author
Author Response File: Author Response.pdf
Reviewer 2 Report
Comments for author File: Comments.pdf
Author Response
Dear Reviewer 2
Thank you for your comment, we have already revised our manuscript. On the other hand, we also fixed our manuscript with revision in language and statistics. There are two files to submitted: 1) revised manuscript and 2) Answer to Reviewer.
Kind Regards
Author
Author Response File: Author Response.pdf
Reviewer 3 Report
The manuscript still requires a careful editing of English language and style.
Author Response
Dear Reviewer 3
Thank you for your comment, we have already revised our manuscript. On the other hand, we also fixed our manuscript with revision in language and statistics. There are two files to submitted: 1) revised manuscript and 2) Answer to Reviewer.
Kind Regards
Author
Author Response File: Author Response.pdf
Round 3
Reviewer 1 Report
The paper addressed my previous comments, in particular, the empirical analysis is now more convincing.
There are however few minor issues having to do with typos and wrong interpretations and use of sentences/words. I’m suggesting how to revise these sentences. After doing that, the paper, in my opinion, is suitable for publication.
There is a typo in the abstract
Line 98 “The current study on comparative advantage focuses on various commodities.” This might be misleading as it suggest that your study do that, you should use “Recent studies” instead of “The current study”
Line 102 “Indonesian cocoa beans are inelastic” you might wanna replace inelastic with “price-inelastic” otherwise it seems to refer to the chemical properties of coca beans
Line 236 “year t EUV” add a comma after “t”
Line 237 “Asiaait “ one too many “a”
Lines 245-246 “Furthermore, using the Levin-Lin and Chu test, we derived the stationary test for
analyzing cross-sectional dependency in time-variant variables.” This is confusing, rewrite as “…using the Levin-Lin and Chu procedure, we tested for the stationary of the series. The test allows to control for the effect of cross-sectional dependence. Non stationary variables are introduced in first differences.”
Lines 255-256 “…two economic circumstances, particularly in advanced economies.” Circumstances is not the correct word and you have to add developing countries, so please rewrite as: “…two main groups: advanced and developing countries.”
Lines 505-509 “Table 10 shows that the null hypothesis (no CD in residuals) may
be strongly rejected at the 5% level for a number of variables, including RSCA, TBI, FDI,
GDPC, IMPAVFO, TO, and OEXR. According to these characteristics, all panel time
series give considerable evidence for cross-sectional dependency. Hence, POP and EUV 5
are non-stationary at level with probability above 5%.” I think you have no clear understanding about the test you are using. The null assumption is that series are non-stationary, the tests (if applied to demeaned data) is robust to the bias induced by cross sectional dependence, but IT DOES NOT test for it. So please rewrite the above sentence in the proper way. I recommend something like “Table 10 shows that the null hypothesis (series have a unit root) may
be strongly rejected at the 5% level for a number of variables, including RSCA, TBI, FDI,
GDPC, IMPAVFO, TO, and OEXR. According to these characteristics, all panel time
series give considerable evidence for stationarity. The only exceptions being POP and EUV
which are non-stationary at level with probability above 5%.”
Line 571 “calculated” if think you mean “discussed”
Author Response
Dear Reviewer 1
Thank you for your revision. We have already revised our manuscript based on your suggestions. Furthemore, we also attached the two files including: 1) Update Manuscript and 2) Answer to Reviewer
Kind Regards
Authors
Author Response File: Author Response.pdf
