Survival Analysis for the Adjustment Phase Following Investment in Swiss Dairy Sheds

Round 1

Reviewer 1 Report

Excellent work and nice results. I have some small suggestions.

L9: space missing after ‘ : ‘

L13: missing word or ‘ , ‘ , probably you need to write: the family capital.

L27 and further: why use both a number [1] and (Schick & Hartmann, 2005); one of the two is superfluous. If this a journal demand, I will not do any review for this journal again, because this is confusing for the reader.

L48: ‘recover’ does seem the best verb choice, perhaps just regain, or reattain as used on L139.

L187: ‘during the time in the panel.’ does seem the best word choice, perhaps write: during the study. or: during the period of observation.

L191: insert a blank line.

L192: Insert (Table 2) after ‘significant.’

L192: Please delete: ‘The results can be interpreted as follows:’ 

L192 to 195, could better be integrated in L198 to 205.

L195-197: Could better be moved to the Discussion.

L234-248: Why do authors not discuss and consider the natural increase of cattle through reproduction; this might be a reason for the low value (2,200) as an opportunity cost for the cows from the own farm.

L250-253: Delete these sentences; these are for the Abstract and not a Conclusion. E.g.; you didn’t test whether the survival analysis was better than other methods; you assumed this.

L253: Superfluous: ‘The result that’ ; afterwards rewrite the remaining sentence.

Author Response

First, we would like to thank you for your comments. Hereinafter your comments are shown in italic letters. We respond to them point by point. All changes in and supplements to the article are shown by means of the Word correction function.

1) Excellent work and nice results. I have some small suggestions.

Thanks for your encouraging feedback.

2) L9: space missing after ‘ : ‘

We inserted the missing space.

3) L13: missing word or ‘ , ‘ , probably you need to write: the family capital.

We tried to follow your suggestion. Please have a look if our rewording matches your intention.

4) L27 and further: why use both a number [1] and (Schick & Hartmann, 2005); one of the two is superfluous. If this a journal demand, I will not do any review for this journal again, because this is confusing for the reader.

Number and reference are both used to ease the review process. We skipped the references except for the cases where want to mention the authors explicitly.

5) L48: ‘recover’ does not seem the best verb choice, perhaps just regain, or reattain as used on L139.

We amended our sentence according to your suggestion.

6) L187: ‘during the time in the panel.’ does not seem the best word choice, perhaps write: during the study. or: during the period of observation.

We followed your suggestion and replaced the wording by “during the period of observation”.

7) L191: insert a blank line.

We inserted a blank line.

8) L192: Insert (Table 2) after ‘significant.’

We amended the paragraph as suggested.

9) L192: Please delete: ‘The results can be interpreted as follows:’

We amended this as suggested by you.

10) L192 to 195, could better be integrated in L198 to 205.

This paragraph explains how to evaluate the results. We feel that this should be noted ahead of describing the results.

11) L195-197: Could better be moved to the Discussion.

Intuitively, the reader might think that a coefficient larger than 1 corresponds to a shorter time to re-attain calculated profit. For the reader it is important to know that this only holds true if solely time-invariant variables were included. Thus, we put the statement into the results section. We now amended our text in order to highlight this point to the reader.

12) L234-248: Why do authors not discuss and consider the natural increase of cattle through reproduction; this might be a reason for the low value (2,200) as an opportunity cost for the cows from the own farm.

We added a small paragraph on why this was not done.

13) L250-253: Delete these sentences; these are for the Abstract and not a Conclusion. E.g.; you didn’t test whether the survival analysis was better than other methods; you assumed this.

We deleted the statement.

14) L253: Superfluous: ‘The result that’ ; afterwards rewrite the remaining sentence.

We amended the sentence as suggested.

Reviewer 2 Report

Major comments

The scientific value of the study should not be overestimated, although  “survival analysis” applied for farming related investments is a strong point of the paper. There are however a few weak points to be mentioned:

A shortcoming of the analysis is a relatively small sample, but my main doubts relate to a small set of independent variables. This is to some extent probably a consequence of basing the analysis on FADN database, which does not provide all details. Regarding variables (used or missing in my opinion): “animal purchases” – what animals were purchased – calves, heifers or cows? “stocking rate” – LU per hectare of  grassland or agricultural land? if the latter – what is the share of dairy production in generating farm incomes (profits)? what about milk-yielding use of concentrates for cows? I would suggest to expand a theoretical part, elaborating more on investments and conceptual framework. Some doubts regarding chapter 4 (discussion): Paragraph in rows 218 – 225. It seems that statement “Off-farm income has a negative impact on the probability of attaining pre-investment profits. One possible reason could be that farms with combined incomes are less productive…. making it more difficult to take full advantage of technological progress ” is inconsistent with an initial statement on technical progress and less labour as an effect. Paragraph which starts in the row 234 and the statement “the purchase of additional animals has a negative effect….”. Firstly, it should be added what is negatively affected – imputed profit? Secondly, impact of animal purchases on productivity or profits depends largely on what animals are purchased (e.g. heifers or cows improving genetic potential of the herd or not), at what age (the younger the greater is a delay with generation of positive effects) and so on. What is more – explanations following the first sentence of the paragraph are not quite convincing and written with the use of hardly understandable language. A radical change is required here.  

Minor remarks

 

Rows: 25/26 – “…effect of an investment typically a technical progress takes place”.

“typically” probably is not the best word – may be commonly or, better in my opinion, likely; “technical progress” is too general. First of all, in my opinion investment as such is an element of the technical progress (not an effect) assuming new solution is better in some respects. Reducing labour input may be considered an “effect”.

38 -  I suggest to replace “hesitantly” with “marginally” or similar.

46 – “imputed or calculated profit” – is “calculated” suggested as another expression for “imputed”? This is confusing. If these terms have a different (what?) meaning, I would say that “imputed” is also calculated. Please make it clear, eventually find a better way to distinguish.

 

61/62 – “…the valley and hill regions”, allowing for the analysis of specialized dairy farms with comparable natural conditions.

This is probably a writing in English problem, but certainly conditions in the valley and hills areas re very different.

 

80 – “all earnings” – I would suggest to replace with “revenues” or may be “receipts”.

standard deviations.

 

123 – text in brackets should be moved to a footnote

 (Swiss francs; average 123 exchange rate in 2017; 1 CHF = 0.90 Euro = 1.02 USD; https://data.snb.ch, retrieved on 29 January 124 2018).

134 – “Opportunity costs should also be included in the model because they are perceived less strongly than financial accounting costs [20] (Kahnemann et al., 1991).  Something is wrong (missing) with this sentence.

135 – “If a farm is generally satisfied…”. I understand the mental shortcut  but “satisfaction” may attributed to the farmer only, not farm.

141 – “… two directions of effect…” – may be two types of effects (categories). If the word “directions” is left then the sentence should be reformulated.

148-150 – “Combining agricultural and non-agricultural income sources is not ideal from an economic perspective and, ……”. I disagree with such a statement. Firstly – how the Authors understand expression “not ideal” (it is not a professional term in the economic paper used in this context). Similar question regarding “economic perspective”.

Taking farm economics and farmers perspective possibility of combining incomes from different sources, supported by the way by agricultural policies in many countries (including EU) may be considered “ideal” in a colloquial sense.

Please clarify.

Table 1. To me this is a strange way of expressing units (10,000 CHF) – we use rather thousands or million in order to reduce values. Not clear what the value 211 for “milk produced” indicates . Also, please clarify what is meant by “ratio of offspring”, which I am afraid is not properly used if the mean value is 0,25. What do you mean by “the wage approach”?

Author Response

First, we would like to thank you indeed for your very helpful comments. Hereinafter your comments are depicted in italic letters. We respond to them point by point. All changes and supplements in the article are depicted with the Word correction function.

 

A shortcoming of the analysis is a relatively small sample, but my main doubts relate to a small set of independent variables. This is to some extent probably a consequence of basing the analysis on FADN database, which does not provide all details. Regarding variables (used or missing in my opinion): “animal purchases” – what animals were purchased – calves, heifers or cows? “stocking rate” – LU per hectare of  grassland or agricultural land? if the latter – what is the share of dairy production in generating farm incomes (profits)? what about milk-yielding use of concentrates for cows? I would suggest to expand a theoretical part, elaborating more on investments and conceptual framework.

We address the relatively small sample in the conclusion section.

Unfortunately, in the FADN data some desirable characteristics of the dairy herd are not included. E.g. within purchased animals, it is not possible to distinguish between cows, heifers and heifer calves. It should be noted though that the amount of money spent is substantially higher for cows and heifers than for calves.

In materials and methods, we describe stocking rate as expressed in livestock units per ha of agricultural land. We added a sentence to explain that arable land and grassland are included in this term. 

The two farm types included in the analysis are dairy farms and combined farms dairy/arable farming. Both are types that obtain the main share of their income from dairy farming. The farm types are defined by their share of grassland and proportion of dairy cows to total animals. They are defined as having more than 75% of their livestock as cattle and more than 25% of those as dairy cows. In the Swiss context, these are very specialized dairy farms.

Milk yielding concentrates have a different importance in Switzerland compared to other countries. Given high import tariffs on cereals and high prices on domestic cereal, concentrates are relatively expensive, hence less attractive to increase milk yield. Furthermore, concentrates might also be used with other cattle such as future dairy cows, not only with dairy cows. This might further distort the picture.

 

Some doubts regarding chapter 4 (discussion): Paragraph in rows 218 – 225. It seems that statement “Off-farm income has a negative impact on the probability of attaining pre-investment profits. One possible reason could be that farms with combined incomes are less productive…. making it more difficult to take full advantage of technological progress” is inconsistent with an initial statement on technical progress and less labour as an effect. Paragraph which starts in the row 234 and the statement “the purchase of additional animals has a negative effect….” (260). Firstly, it should be added what is negatively affected – imputed profit? Secondly, impact of animal purchases on productivity or profits depends largely on what animals are purchased (e.g. heifers or cows improving genetic potential of the herd or not), at what age (the younger the greater is a delay with generation of positive effects) and so on. What is more – explanations following the first sentence of the paragraph are not quite convincing and written with the use of hardly understandable language. A radical change is required here.   It is not inconsistent if we assume less labour per animal (row 26) as an effect and a concurrent increase in animals. The literature suggests (Mann and Mittenzwei, 2017) that farms with a combination of farm-income and off-farm-income are less efficient. The negative effect is indeed with respect to imputed profit which we added in the text. Our additions also address that age of purchased animals cannot be analyzed due to lack of adequate data. We extended the discussion about the negative impact of animal purchases and split it into several paragraphs for, hopefully, easier comprehension.

3) Rows: 25/26 – “…effect of an investment typically a technical progress takes place”.

We tried to modify the sentence as you suggested.

4) “typically” probably is not the best word – may be commonly or, better in my opinion, likely; “technical progress” is too general. First of all, in my opinion investment as such is an element of the technical progress (not an effect) assuming new solution is better in some respects. Reducing labour input may be considered an “effect”.

 

We tried to modify the sentence corresponding to your suggestions.

5) 38 -  I suggest to replace “hesitantly” with “marginally” or similar.

In order to emphasize the temporal aspect, we replaced ‘hesitantly’ by ‘with some delay’.

6) 46 – “imputed or calculated profit” – is “calculated” suggested as another expression for “imputed”? This is confusing. If these terms have a different (what?) meaning, I would say that “imputed” is also calculated. Please make it clear, eventually find a better way to distinguish.

Thank you for your question. “Calculated” and “imputed” were used as a synonym. However, this might irritate the readers because they might assume different meanings. Therefore, we now have switched to the term “imputed” throughout the paper.

7) 61/62 – “…the valley and hill regions”, allowing for the analysis of specialized dairy farms with comparable natural conditions. This is probably a writing in English problem, but certainly conditions in the valley and hills areas are very different.

Farming regions in Switzerland are divided into valley, hill and mountain region. The mountain region is very different to the hill and valley region in terms of natural conditions. Farming in mountain region requires special machinery. In order to distinguish between valley and hill region, we analyzed a model with a dummy variable for one region, which showed no significant result for this dummy.

8) 80 – “all earnings” – I would suggest to replace with “revenues” or may be “receipts”.

We replaced earnings by revenues.

9) standard deviations.

We changed to the singular, if this is what your remark related to.

10) 123 – text in brackets should be moved to a footnote (Swiss francs; average 123 exchange rate in 2017; 1 CHF = 0.90 Euro = 1.02 USD; https://data.snb.ch, retrieved on 29 January 124 2018).

As MDPI does not allow inserting footnotes, we opted for a sentence in brackets.

11) 134 – “Opportunity costs should also be included in the model because they are perceived less strongly than financial accounting costs [20] (Kahnemann et al., 1991).  Something is wrong (missing) with this sentence.

We replaced “strongly” with “important”.

12) 135 – “If a farm is generally satisfied…”. I understand the mental shortcut, but “satisfaction” may attributed to the farmer only, not farm.

We modified the sentence according to your observation.

13) 141 – “… two directions of effect…” – may be two types of effects (categories). If the word “directions” is left then the sentence should be reformulated.

We amended the sentence following your line of thought.

14) 148-150 – “Combining agricultural and non-agricultural income sources is not ideal from an economic perspective and, ……”. I disagree with such a statement. Firstly – how the Authors understand expression “not ideal” (it is not a professional term in the economic paper used in this context). Similar question regarding “economic perspective”.

We tried to formulate this in a more scientific and stringent way.

15) Taking farm economics and farmers’ perspective, the possibility of combining incomes from different sources, supported by the way by agricultural policies in many countries (including EU) may be considered “ideal” in a colloquial sense. Please clarify.

We agree that an income combination can be seen as ideal strategy. Focusing on investment, we won’t discuss income strategies here in more detail. Rather, we think it would be an interesting subject for a further analysis.

16) Table 1. To me this is a strange way of expressing units (10,000 CHF) – we use rather thousands or million in order to reduce values. Not clear what the value 211 for “milk produced” indicates. Also, please clarify what is meant by “ratio of offspring”, which I am afraid is not properly used if the mean value is 0,25.

The ratio is defined by the ratio in livestock units. The ratio is still reasonable taking in account that younger animals represent a lower amount of livestock units, variations in replacement rates and some farms might have outsourced rearing cows.

With respect to the change of units, we would like to keep them as they are. Our main argument is that for all respective variables, the reflected change of one unit as per the Cox proportional hazard model represents a meaningful change. This can best be seen in the last sentence of section 3: “To illustrate the interpretation of the hazard ratio, we use the example of imputed profit before investment: For every CHF 10,000 of imputed profit, the probability of re-attainment is reduced by 14%.” If we chose different units, say kCHF, the resulting coefficients of the model would be very close to 1, and a bit harder for the reader to interpret by inspection, as basically all coefficients would vary very slightly around 1. (E.g. for every CHF 1,000 of imputed profit, the probability of re-attainment is reduced by 1.4%. This would correspond to roughly a value of the corresponding coefficient of 0.986.) All of the changes by one unit, reflected by the current coefficients of the model based on 10,000 CHF, are also reasonable in the sense that they are still below a change by one standard deviation of the respective variable, i.e. not extreme changes reflected by the coefficient.

Also, we corrected kg of milk per year to ton per year. Thank you for pointing this out.

17) L229: What do you mean by “the wage approach”?

We reformulated this sentence: “Since, as for opportunity a costs, the remuneration of family labour surpasses clearly the remuneration of capital [27], it is not surprising that the use of family labour has a more significant influence than equity capital.”

 

Mittenzwei, K., & Mann, S. (2017). The rationale of part-time farming: empirical evidence from Norway. International Journal of Social Economics, 44(1), 53-59. doi:http://dx.doi.org/10.1108/IJSE-10-2014-0207

Round 2

Reviewer 2 Report

Acceptable with the improvements introduced.