The Structural Gravity Model and Its Implications on Global Forest Product Trade

Round 1

Reviewer 1 Report

Dear authors,

I appreciate your effort to apply the theoretical economic model to the forest products sector. Forest products trade and market have their specifics which follow either from the nature of supply or the nature or demand for forest products. You tried to consider the influence of these factors by defining and selecting a range of variables. Even if proxy variables were defined and justified, markets for individual forest products or their groups are very specific and driving forces behind these markets differ significantly. Therefore, your findings can be interpreted onky at a very genera level. In this regard, instead of going deeper to sectoral level (which, as you state, might be difficult) I suggest you to elaborate a little bit more on the following issues:

Extend/supplement Introduction by a text/subchapter dealing with the description of specifics of markets and trade with individual products categories. There are already known features of trade patterns of specific products/groups (e.g. coniferous sawnwood trade to countries with high proportion of wooden dwellings on started constructions, relatively short distances in trade with particle boards due to their weight and related transport costs, etc.) As part of Introduction, shortly describe the actual main global patterns in trade with individual forest product categories and use this underlying analytics when discussing your findings. In Discussion please define the limitation of your findings following from your assumptions and model definition In more detail explain the meaning of variable “forest rents” Discussing your results at the level of products groups (e.g. sawnwood, wood pulp+newsprint+paper and paperboard) instead of products themselves (e.g. coniferous sawnwood, non-coniferous sawnwood) leads to generalisation of findings as the products groups are used for reporting purposes only and not for actual trading.

In Methodology, please explain “PPML” for the first time you mention it in the text (line 67). The same applies for other abbreviations e.g. CEPII or FAO (lines 131,132)

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Review of the manuscript “The structural gravity model and its implications on global forest product trade”. Manuscript ID: forests-695556.
General comments
The topic and analyses carried out in the manuscript seem to be useful and publishable. However, the manuscript needs to be better improved before it can be accepted for publication. I see a need to improve introduction, methods and the results and discussion section as follows:
Introduction
There is a need to make introduction section stronger (currently, it looks weak) by 1) better clarifying the structural gravity model the authors are using/estimating, and 2) better describing the limitation of the traditional gravity model in explaining trade and the need for the use of the structural gravity mode (e.g., describe why structural gravity model of trade is superior to traditional gravitational model. Provide your case supported by published studies (use studies from non-forest sector if none is existing for forest sector).
Methodology
Methodology section also needs to be improved by better explaining the structural gravity equations. Needs a brief standalone explanation rather than just referring to Head et al. (2004). Also, it will be helpful to show the traditional gravity model in equation form.
Results and discussion
Finally, the results and discussion section need to be improved by 1) referring appropriate tables to the text, 2) providing estimated numbers where necessary (e.g., what is the estimated value in the past literature and what is the estimated value in your study? To be able to say that past studies overestimated the impacts of GDP on forest sector trade, you should show and compare the estimated numbers in your study and in past studies), and 3) better explaining why some of the results are contrary to the prior expectation (of the structural gravity theory).
Some of the questions/comments, answering/explaining which may improve the authors’ cases in the introduction section are: 1) why can't one estimate a traditional gravity model at a disaggregated level (see lines 40 to 41, where you state that estimates at the aggregated level include aggregation bias if used for disaggregated sector level estimations), 2) is the model you are using (structural gravity model) a new concept? If not, clarify it upfront and refer to some past studies that estimated the structural gravity model (in non-forest sector, if none existing for forest sector). You could emphasize that you are using the structural gravity approach in the forest sector, which hasn't been carried out by previous researchers (if the statement is true), 3) what are the important research gaps and why is your study important? (e.g., no past study have evaluated the effect of various socio-economic and geographic factors in explaining forest products trade and that you are trying to trying to understand the importance of those factors in explaining forest products trade), and 4) consider moving lines 53 to 75 to the introduction section.
Specific comments
1. Line 39. Consider replacing the phrase “Here, the trade will be influenced…” with the phrase “According to the tradition definition of the gravity model, trade is influenced……”
2. Lines 41 to 44. Suggest rephrasing for better clarity. Also explain why one can’t estimate a gravity model at a disaggregated level? If one can estimate it, then it’s not a research challenge/problem/gap.
3. Which research gap? You need to better outline research gap. See the general comments above.
4. Line 46 and throughout the manuscript. Consider replacing the confusing phrase “the appearance of trade” with the phrase “the likelihood of the occurrence of trade” for better clarity. Also consider the replacing the word “intensity” with the phrase “magnitude”.
5. Line 46. Are you introducing (as a new concept) or applying the already existing structural gravity approach? I think you are applying this approach, which has been already introduced by other others (e.g., Head et al. 2004). You could emphasize that you are using this approach in the forest sector, which hasn't been carried out by previous researchers (verify and add to the text if the statement is true).
6. Line 47. Briefly introduce the structural gravity approach in the intro section before explaining in more detail in the methods section.
7. Lines 53 to 75. Consider moving the text to introduction section.
9. Lines 69 to 70. Not clear. Size of what? Correct and/or rephrase for clarity.
10. Lines 72 to 73. What do you mean by structural framework of gravity model? Clarify. Also, provide a citation here (e.g., Head et al. 2004) to clarify that structural gravity model was proposed by other authors.
11. Line 76. Before explaining structural gravity model, it will be helpful to explain traditional gravity model too in equation form. Also, provide a citation when describing both traditional and structural gravity model. (e.g., insert a citation after the phrase “The structural gravity approach…..”
12. Line 82. How do you measure degree of competition? Insert in the parenthesis “e.g., consumption of forest products in exporting country, as described below”. Similarly, how do you measure bilateral accessibility? Insert in the parenthesis “e.g., geographical distance between two countries, as described below”.
13. Footnote 1. Although you can refer to any citations for more details, the methods you are describing should be standalone in to understand the model and its derivation. Better describe the model by elaborating each components/variables and how are they derived.
14. Equation 3. Delete the typo “m” in the subscript of β4.
15. Line 107 and everywhere. Replace the confusing phrase “appearance of trade” with the phrase “likelihood of occurrence of trade”.
16. Line 116. Replace the symbol H0 with the phrase “null hypothesis (H0)”. Use superscript for zero.
17. Lines 114 to 115. Clarify why, if coefficients for the PPML significantly differ, you need to apply ML. Does this mean that you do not need to apply ML, if the coefficients for the PPML significantly differ?
18. Line 119. Just testing for heteroscedasticity doesn’t automatically lead to PPML. It would lead to PPML if heteroscedasticity is found. Therefore, correct the text accordingly.
19. Line 139. Clarify how you tested for possible combinations of various variables and chose the most fitting ones.
20. Line 170. Consider rephrasing the phrase “trade is attracted by the near distance between the Russian and German capital” with the phrase “closer the distance, the greater the chance that trade could occur between the partner countries”.
21. Results and discussion section. Refer to the appropriate tables when describing the results. Results can then look at the particular table you are referring to. Also it will be helpful to explain the results by reporting the estimated sign and the statistical significances (e.g., significance at 1%, 5% or 10% ) of the estimated coefficients.
Also, interpret the results. Describe the marginal effects, if possible? Can you apply the traditional interpretation of the long-log model, where you interpret the results as elasticities (i.e., 1% change in explanatory variable leading to X% change in response variable)?
22. Lines 190-192. Report the values of the estimated GDP coefficients in your study and in past studies to support the statement that past study overestimated the effects of GDP.
23. Line 198. Consider replacing the phrase “one model if it is not in the other model” with the phrase “both models” for better clarity. If you mean something else, then rephrase for better clarification.
24. Line 200. Consider replacing the phrase “contrary to the export value” with the phrase “except for the quantity of export”
25. Line 201. Consider replacing the phrase “Vice versa” with the phrase “ In contrast”.
26. Line 202 and throughout manuscript. Consider replacing the phrase “value of trade” with the phrase “quantity of trade” for better clarity. The phrase “value” can be confusing with to monetary value of export or import (export or import quantity X price).
27. Line 202. Replace the word “negative” with the correct word “negatively”.
28. Line 205. Provide the table number and title at the top of the table (not at the buttom).
29. Line 214. Remove the phrase “our” in the phrase “our theoretical classification” to avoid any potential confusion.
30. Line 218. Insert the phrase “, based on our results” after the phrase “positive way” for better clarity.
31. Replace “trade value” with “trade quantity”.
32. Lines 235 to 241. To avoid confusion, remove all the text to section 3.1 where you describe the results for the Industrial roundwood.
33. Lines 245-247. This is also true with previous results. So, there is no need to present FE and PPML with zero trade flows for previous results too.
34. Line 253 and throughout the results and discussion section. If your results are contrary to what theory would suggest, you need to explain the possible reasons behind such opposite outcomes.
35. Line 255. Replace the phrase “Regards” with the phrase “For”.
36. Line 256. Replace the phrase “trade value” with the phrase “trade quantity”.
37. Line 257. How do you define bigger or smaller country? If you mean to say high income countries, then correct it accordingly.
38. Line 260 to 262. This results is contrary to the expectation. Provide possible reasons for the discrepancy with the expected.
39. Lines 262 to 263. Rephrase for better clarity. It's confusing at present.
40. Line 267. Replace the phrase “trade value” with the phrase “trade quantity”.
41. Line 268. Add the phrase ", based on the ML results", after the phrase “countries”.
42. Line 270. Replace the phrase “appearance” with the more appropriate phrase “likelihood”.
43. Line 274. Replace the phrase “remarkable that” with the more appropriate phrase “contrary to the expectation that". Also explain possible reasons for the opposite outcome.
44. Line 279. Replace the phrase “trade value” with the phrase “trade quantity”.
45. Line 280. Replace the phrase “appearance” with the more appropriate phrase “likelihood of occurrence”. Also remove the phrase “our” to avoid any confusion.
46. Line 281. How do you measure slowing down of trade quantity? Do you mean negatively affect? If so rephrase.
47. Lines 281 to 283. Do you mean to say that both the magnitude and the likelihood of trade is reduced if only one partner is EU member? If so rephrase to better clarify.
48. Line 299. Replace the phrase “trade” with the phrase “quantity”.
49. Lines 301 to 303. How can you conclude this? Elaborate.
50. Line 304. Add the phrase “(other than GDP and distance only, as in the traditional gravity theory)” after the phrase “additional factors” for better clarity.
51. Line 306. What do you mean by hidden laws? If you meant other factors that may affect forest sector trade, then clarify accordingly. Law is different than factors affecting something.
52. Lines 306 to 308. The points you are trying to make here is not clear. Rephrase for better clarity.
53. Lines 308 to 309. Here, means where? In your study or in past study? Since you are estimating structural gravity, I guess you are referring to your estimated coefficients of GDP and distance with coefficients estimated in past studies? Clarify. Any conclusions has to be supported by the results.
54. Line 315. Replace the phrase “different” with the correct phrase “differently”.
55. Line 314-317. What could be the possible reasons for such differences in findings?
56. Consider replacing the phrase “differences between determinants for the occurrence and the magnitude of trade flows” with the phrase “the factors influencing both the likelihood and the magnitude of trade flows”.
57. Lines 322 to 324. This is contrary to expectation of the structural gravity theory. More explanation is needed in the discussion section or here as why you found such an outcome.
58. Table A1. Why do you need to include table A1? Is it any helpful? I don’t see why this table is important. Consider removing.
59. In table 2, you described the significance levels as ***, **, and * are significant at the 0.1%, 1%, and 5% level, respectively but in other tables you used *** p < 0.001, ** p < 0.01, * p < 0.05. Be consistent.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

The article is well written, the method is clearly described. The results support the conclusions. 

However, the data section (chapter 2.2) should be checked for potential inconsistencies. According to the eq.2 and the sentence at lines 158-159, the degree of competition in the importing market Φn should be represented by production of a good in the importing country. Nevertheless, "Consumption (exporter)" is given in the table 1 as Φn variable, which is inconsistent with eq.2 and text at 158-159 lines. 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Authors accepted and explained all my comments and quaetions.