The Impact of Agricultural Digitization on Land Productivity: An Empirical Test Based on Micro Panel Data

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Digitization and land productivity are important topics in the world. This paper explores how agricultural digitization affects land productivity based on farm household data, and talks about the mechanism. After reading the paper, I still have some questions.

1. The dependent variable is crop production per unit and the data comes from Jiangsu province. We know that crop includes wheat, corn, rice, ect. For Jiangsu province, there is a significant difference in planting between the southern and northern regions. Using crop production per unit as dependent variable, the dependent variable is very general and may affect the accuracy of the results. If the data is available, please try to analyze the impact of digitization on different varieties.

2. For the model, I think PSM is more suitable, why haven't you tried using PSM?

3. About the model on line 314, fertilizers, pesticides, machinery, etc. are important factors that affect land productivity. You did not take these factors into account, but instead used them as mechanism variables. I think it is inappropriate to not consider these factors in the model.

4. This paper does not have descriptive statistics, so I am unable to know the basic information of some key variables.
5. This paper has a very interesting result, digitization has a greater impact on households with a higher share of elderly workers. Could you give me more information about the elderly workers. How to identify the elderly worker? How old?

Author Response

Question 1. The dependent variable is crop production per unit and the data comes from Jiangsu province. We know that crop includes wheat, corn, rice, ect. For Jiangsu province, there is a significant difference in planting between the southern and northern regions. Using crop production per unit as dependent variable, the dependent variable is very general and may affect the accuracy of the results. If the data is available, please try to analyze the impact of digitization on different varieties.

Author's Response: Thank you for pointing this out. The author fully agrees with and accepts the reviewer's suggestions. Therefore, based on the available data, we have decided to modify the explanatory variables in the robustness analysis, incorporating an analysis of the major crop varieties (wheat, corn, and rice). According to the data, 87.58% of households grow rice, 73.05% grow wheat, and 21.40% grow corn. The percentage of households growing other crops such as soybeans and potatoes is quite small, with the highest at just 7% (some households grow multiple crops, so the total exceeds 100%). The estimation results show that agricultural digitization can improve land productivity for households growing different types of crops. The revised detailed content can be found in lines 366-373 of the main text.

 

Question 2. For the model, I think PSM is more suitable, why haven't you tried using PSM?

Author's Response: Thank you for pointing this out. The author fully agrees with and accepts the reviewer's suggestions. Therefore, we have added the PSM (Propensity Score Matching) test in the robustness check. Since the quality of PSM matching largely depends on the accuracy of the model, unsuccessful matching may lead to bias, which would affect the estimation results. Based on this, we did not replace the entire analysis with PSM, but only incorporated the PSM test in the robustness check. The estimation results show that under the PSM method, the estimated coefficients remain significantly positive, which validates the robustness of the previous findings. The empirical results can be found in Table 5 of the revised version, and the detailed modifications are presented in lines 388-398 of the main text.

 

Question 3. About the model on line 314, fertilizers, pesticides, machinery, etc. are important factors that affect land productivity. You did not take these factors into account, but instead used them as mechanism variables. I think it is inappropriate to not consider these factors in the model.

Author's Response: Thank you for the reviewer’s insightful comments. The author has carefully considered this issue and ultimately decided not to include the mechanism variables in the baseline regression for the following reasons:

The primary goal of the baseline regression is to estimate the direct effect of the core independent variable on the dependent variable while controlling for key confounding or potential influencing factors. Mechanism variables are typically not the focus of the study but rather are part of an extended analysis. Including mechanism variables in the baseline regression may weaken the estimated effect of the core independent variable on the dependent variable, as mechanism variables might be highly correlated with the core independent variable.

Mechanism variables are used to explain the specific pathways or underlying mechanisms through which the core independent variable influences the dependent variable. Their main purpose is to answer the "why" or "through what channels" question regarding the impact of the core variable on the outcome variable. Not including mechanism variables helps avoid issues such as multicollinearity or unclear effect decomposition between variables, thus enabling a more accurate capture of the total effect of the core independent variable on the dependent variable in the baseline regression.

In summary, if the study primarily focuses on the total effect of the core variable, mechanism variables do not need to be included in the baseline regression. However, if the study aims to explore mechanisms or explanatory pathways, mechanism variables can be gradually introduced in subsequent models.

Finally, to test the impact of mechanism variables on the estimation results, the author has included mechanism variables such as fertilizer and pesticides in the baseline regression. The regression coefficient was 0.128, which remains significantly positive at the 5% significance level. This alleviates concerns about a significant impact of the mechanism variables on the estimation results.

 

Question 4. This paper does not have descriptive statistics, so I am unable to know the basic information of some key variables.

Author's Response: Thank you for the reviewer’s comments. The manuscript has been revised accordingly. The descriptive statistics and variable descriptions are now presented in the same table, Table 1, and we have added information on the maximum and minimum values of the variables. The detailed information can be found in line 322.

 

Question 5. This paper has a very interesting result, digitization has a greater impact on households with a higher share of elderly workers. Could you give me more information about the elderly workers. How to identify the elderly worker? How old?

Author's Response: Thank you for your attention to the research findings of this study. The relevant information obtained from the data is as follows:

（1）Jiangsu Province is the most densely populated province in China, with a relatively high level of population aging and a large elderly population.（2）By the end of 2023, the population aged 60 and above in Jiangsu Province had reached 20.429 million, with an aging rate of 26.02%. It is projected that by the middle of the 21st century, the aging rate in Jiangsu will reach 40%. （3）This study uses the proportion of elderly individuals participating in farming within a household to represent the aging of the agricultural population.（4）The elderly population refers to individuals aged 60 and above, with no difference in aging standards between men and women. This standard is based on the Law on the Protection of the Rights and Interests of the Elderly of the People's Republic of China.

The supplementary information added in the revised manuscript can be found in lines 475-479.

 

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

-        Introduction and background:

The introduction effectively highlights the importance of digitizing agriculture in addressing global food security and productivity challenges. However, consider briefly elaborating on how this study builds on or departs from previous research. This might provide a clearer context for readers unfamiliar with the topic. In addition, including examples from other countries or regions could emphasize the wider applicability of your results.

 

-        Methodology:

The methodological framework is robust, but could benefit from greater clarity for non-professional readers.

 

-        Specifically:

Provide more details on the construction and calculation of mismatch indices, including their theoretical basis and practical significance.

Explain the reasons for selecting particular control variables and how they align with the research objectives.

Consider adding a visual representation of the research design (eg, a flowchart) to guide readers through the key steps of the analysis.

 

-        Results and visual presentation:

The results are well organized and comprehensive. However, the inclusion of visual aids, such as graphs, maps or trend charts, could make the key results more accessible and interesting.

For example:

Visualize the impact of agricultural digitization at different land scales or demographic groups to highlight heterogeneity findings.

Include comparative visuals to show how digitization impacts productivity versus traditional practices.

 

-        Discussion and implications:

The discussion is insightful but could be improved by exploring the broader implications of your results.

For example:

How might these results affect policy decisions outside of China? Addressing this issue could broaden the document's relevance to an international audience.

Discuss the potential obstacles to the implementation of digitization in less developed regions and propose solutions to overcome them.

 

-        Policy recommendations:

The recommendations are practical and grounded. However, strengthening the link between quantitative findings and policy proposals would increase their credibility.

For example:

Give specific examples of policies or programs that could support vulnerable groups, such as older farmers or those with limited education.

Discuss the extent to which your results align with the global Sustainable Development Goals (SDGs) and what this implies for international agricultural strategies.

 

-        Heterogeneity analysis:

The heterogeneity analysis provides valuable insight into the different impacts of digitization. Extending these results could make them more effective.

For example:

Discuss targeted interventions that could maximize benefits for small farmers or regions with aging populations.

Elaborate on the equity implications of your findings, particularly in terms of reducing the digital divide in rural areas.

 

-        Sustainability and environmental impact:

While the study touches on sustainability, consider more detailed research on how the digitization of agriculture contributes to long-term environmental benefits, such as reducing pollution and improving resource efficiency. Linking these results to global efforts in sustainable agriculture could strengthen the impact of the research.

 

-        Future directions of research:

Conclude the paper by identifying areas for future research, such as investigating the cost-effectiveness of digitization initiatives or studying their long-term impact on rural livelihoods and environmental sustainability. This could provide a guide for further research in the field.

Comments for author File: Comments.pdf

Author Response

Revision Notes (II)

Question 1. Introduction and background:

The introduction effectively highlights the importance of digitizing agriculture in addressing global food security and productivity challenges. However, consider briefly elaborating on how this study builds on or departs from previous research. This might provide a clearer context for readers unfamiliar with the topic. In addition, including examples from other countries or regions could emphasize the wider applicability of your results.

Author's Response: Thank you for your valuable suggestions on the introduction section of this paper. Your comments have been instrumental in improving the quality of the paper. We have made revisions and enhancements in the following two aspects:

（1）Relationship with Existing Research: We have revised and supplemented the third and fourth paragraphs of the introduction to highlight the following points: Existing research primarily focuses on the impact of agricultural digitization on the development of business entities, the efficiency of agricultural input use, and green development, often using macro-level data at the provincial or municipal level. This study innovatively uses micro-level survey data and, by constructing indicators for the distribution of village-level digitalization projects, systematically analyzes the impact mechanisms of agricultural digitization on land productivity for the first time. This approach not only continues the research focus on the economic effects of agricultural digitization found in the literature but also breaks new ground in terms of research perspective and data precision. Moreover, in the discussion section, we employed a comparative literature review approach to illustrate how this study both draws from and deviates from previous research. Please refer to lines 525-560 for detailed revisions/amendments

（2）International Cases: In the second paragraph of the introduction, we have added references to relevant international practices: The European Union's Common European Agricultural Data Space (CEADS) project, Japan's Agricultural Data Collaboration Platform (WAGRI), and India's Digital Agriculture Mission plan. These practices highlight the global significance of agricultural digitization and provide a practical foundation for the application of the findings of this study. Please refer to lines 31-34 for detailed modifications

 

Question 2. Methodology: The methodological framework is robust, but could benefit from greater clarity for non- professional readers. Specifically: Provide more details on the construction and calculation of mismatch indices, including their theoretical basis and practical significance. Explain the reasons for selecting particular control variables and how they align with the research objectives. Consider adding a visual representation of the research design (eg, a flowchart) to guide readers through the key steps of the analysis.

Author's Response: Thank you for the reviewer's valuable suggestions regarding the methodology. We have made the following revisions:

（1）Mismatch Index: We have provided more detailed explanations of the mismatch index in terms of its theoretical foundation, practical significance, and complete calculation details. Please refer to lines 255-263.

（2）Control Variables: We have added explanations for the selection of control variables to ensure clarity and rigor in the empirical analysis. Please refer to lines 302-307

（3）Theoretical Framework: We have included Figure 1 to present the theoretical analysis framework of the study, which clearly illustrates the mechanism through which agricultural digitization impacts land productivity. The subsequent empirical analysis is conducted based on this framework. Please refer to lines 215.

 

Question 3. Results and visual presentation: The results are well organized and comprehensive. However, the inclusion of visual aids, such as graphs, maps or trend charts, could make the key results more accessible and interesting. For example: Visualize the impact of agricultural digitization at different land scales or demographic groups to highlight heterogeneity findings. Include comparative visuals to show how digitization impacts productivity versus traditional practices.

Author's Response: We appreciate the reviewer's suggestions regarding results presentation and visualization. We fully agree with and accept these comments. Accordingly, in Section 3, we have created a schematic diagram illustrating the mechanism of how agricultural digitalization affects land productivity. Additionally, we have visualized the heterogeneous effects of agricultural digitalization by creating bar charts to facilitate intuitive comparisons. However, considering that tables can present the impacts of agricultural digitalization more precisely, and given the space constraints of the article, we decided not to include these bar charts in the main text.

Appendix Figure 1. Heterogeneous Effects of Agricultural Digitalization on Land Productivity

 

Question 4. Discussion and implications: The discussion is insightful but could be improved by exploring the broader implications of your results. For example: How might these results affect policy decisions outside of China? Addressing this issue could broaden the document's relevance to an international audience. Discuss the potential obstacles to the implementation of digitization in less developed regions and propose solutions to overcome them.

Author's Response: Thank you for the reviewer’s valuable comments regarding the discussion and impact sections. The author fully agrees with and accepts these suggestions. We have supplemented the manuscript with relevant content, and the main revisions are as follows:

（1）Addition of a “Discussion” Section: We have specifically added a "Discussion" section in the manuscript. This section explores the theoretical contributions, policy implications, and practical value for international application based on the key findings and innovations of this study, while also integrating insights from other relevant research.

（2）Enhancement of Policy Recommendations: We noticed that the original policy recommendations section already provided a preliminary mention of the study’s international significance. Based on your suggestions, we have further strengthened this section to make it more detailed and in-depth. Specifically, we emphasized the implications of our findings for other developing countries, discussed potential barriers to digitalization implementation in less-developed regions, and proposed possible solutions to overcome these challenges.

 

Question 5. Policy recommendations: The recommendations are practical and grounded. However, strengthening the link between quantitative findings and policy proposals would increase their credibility. For example: Give specific examples of policies or programs that could support vulnerable groups, such as older farmers or those with limited education. Discuss the extent to which your results align with the global Sustainable Development Goals (SDGs) and what this implies for international agricultural strategies.

Author's Response: Thank you for your valuable suggestions. Based on your feedback, we have made more comprehensive revisions and expansions to the discussion and policy recommendation sections. Specifically, we have explored support policies targeting vulnerable groups in light of the study's findings and added a detailed explanation of the international policy implications. Furthermore, we highlighted the connection between our research and global development agendas, such as the Sustainable Development Goals (SDGs) and inclusive development, with a particular emphasis on the relevance and applicability of our findings to other developing countries.

 

Question 6. Heterogeneity analysis: The heterogeneity analysis provides valuable insight into the different impacts of digitization. Extending these results could make them more effective. For example: Discuss targeted interventions that could maximize benefits for small farmers or regions with aging populations. Elaborate on the equity implications of your findings, particularly in terms of reducing the digital divide in rural areas.

Author's Response: Thank you for the reviewer’s valuable comments on the heterogeneity analysis. The author fully agrees with and accepts the suggestions provided. The digital divide refers to the gap between different social groups in terms of access to and use of modern information technologies, particularly focusing on the differences in digital literacy and skills. In response to this, we have included a more detailed and targeted discussion in the research's discussion and conclusion sections, approaching it from the perspective of equity.

 

Question 7. Sustainability and environmental impact: While the study touches on sustainability, consider more detailed research on how the digitization of agriculture contributes to long-term environmental benefits, such as reducing pollution and improving resource efficiency. Linking these results to global efforts in sustainable agriculture could strengthen the impact of the research.

Author's Response: Thank you for your valuable suggestions. Your recommendation to strengthen the analysis of environmental benefits is indeed crucial, and it is an indispensable aspect of agricultural digitization research. In light of your comments, we have made the following additions and improvements to the manuscript:

(1) Discussion Section: We have emphasized the environmental benefits of agricultural digitization in the discussion section.

(2) Conclusion and Policy Recommendations: In the conclusion and policy recommendations section, we have proposed specific measures to achieve long-term environmental benefits through agricultural digitization.

Question 8. Future directions of research: Conclude the paper by identifying areas for future research, such as investigating the cost- effectiveness of digitization initiatives or studying their long-term impact on rural livelihoods and environmental sustainability. This could provide a guide for further research in the field.

Author's Response: Thank you for your valuable suggestion. Your comment regarding the need to clearly outline future research directions is highly appreciated. In response, we have added a section on research prospects in the discussion, building on the analysis of the limitations of this study. This enhances the overall structure and completeness of the paper.

Please refer to lines 561-568 for detailed modifications.

 

 

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

The theme of this paper is cuncurrent and the methods used for analysis are robust. Authors deserve commendation for undertaking this reserach. However, this revewer has some questions on the detailed issues of the manuscript. The reply from the authors will contribute to better understanding of the paper.

what are the 12 types of the agricultural techinologies mentioned in line 298-, explaining them will give clues to understand the context.

in Tble 1. if any of one aspect of dizitization of agriculture mentioned therein is fulfilled, no matter how crude they may be, vlue of 1 is assinged or there are some more comprehensive criteria. If it is the  first case, then this reviewer assumes there could or should be more counts with assigned value of 1. However the mean value is only 0.293. explain.

in the same table the health status need tobe explained. what makes the individuals to be considered healthy.

in Fig,1, the captions in the left sid eof the graph illigible making it possible to follow the explanations int he context. the quality of the graph and expalantions should be improved.

this reviewer could not understand the statement of line 387, "driven by a single component....... comprehensivr impact" which is seemingly to strong a statement.

land operation scle mentioned in line 416-; small sclae being 100 acre and so on. this reviewer wonders if 100 acre is a small scale in China, in many countries it is not, explain for confirmation.

the explanation of household human capital and old age need to be clarified cross checking with these 2 variables. This will be needed to substantiate what authors say, "digital technology can compensate for skill deficiencies......, they may be attributed  to low human capital often engage in rudimentary.....and so on" and the explanations in the section of elderly agr, workers. Again, this gets related with previous comment, are we to consider any mechanical tools with some digital aspect, no matter how crude they may be, to be considerd as "digitizatised" ? and in such case will not every technology be said to be digitized? then low mean value in table 1 mentioned above is a perplex.

in table 6, the constant value for large scale operation is not significant, explain.

 

Author Response

Revision Notes (III)

 

Question 1. What are the 12 types of the agricultural technologies mentioned in line 298, explaining them will give clues to understand the context.

Author's Response: Thank you for pointing this out. The author fully agrees with and accepts the reviewer's suggestion. Therefore, we have listed 12 specific types of technologies in the text as footnotes. These technologies include: soil testing and fertilization, crop cultivation management, pest and disease control, mechanized production, energy-efficient and high-efficiency facility agriculture, water-saving irrigation, animal health and hygiene farming, comprehensive utilization of crop straw, clean and renewable energy for agricultural use, network information technology services, improved seed varieties, and disaster prevention and reduction. Please refer to lines 302-307 for detailed modifications

 

Question 2. In Table 1. if any of one aspect of digitization of agriculture mentioned there in is fulfilled, no matter how crude they may be, value of 1 is assigned or there are some more comprehensive criteria. If it is the first case, then this reviewer assumes there could or should be more counts with assigned value of 1. However, the mean value is only 0.293. explain.

Author's Response: Thank you for pointing this out. Regarding the issue of the low mean value, we would like to address it as follows: (1) Real-World Context: In reality, agricultural digitalization projects in China are still in the process of being rolled out, and not all areas or sectors of agricultural production have been digitalized. The cost of digitalization investments is high, and the adoption rate remains low for the years covered by the data. Even today, full implementation of agricultural digitalization across China has not yet been achieved. (2) Scope of Digitalization in This Study: This paper focuses on agricultural digitalization related to land productivity. The survey covers six types of agricultural digitalization, of which this study focuses on four related to crop production. The other two are related to fisheries and livestock digitalization. Since the dependent variable is based on crop yield (which reflects land-based agricultural production), digitalization in fisheries and livestock sectors is not included in the analysis. As a result, the mean value for agricultural digitalization in this study is lower compared to the original dataset. However, this does not affect our research, as the sample is sufficiently well-distributed into treatment and control groups.

Additionally, in response to the reviewer’s concern about the potential bias in the construction of the core explanatory variable, we have added a robustness check in the paper. Instead of measuring whether a village has implemented agricultural digitalization projects, we now measure it by the number of digitalization projects implemented. For example, if a village has implemented four types of agricultural digitalization projects, the variable for agricultural digitalization (dig) is assigned a value of 4, and so on. The estimated coefficient is 0.197, which remains significantly positive at the 5% level. This approach further strengthens the robustness of our analysis and provides additional evidence for the validity of our conclusions. Please refer to lines 381-387 of the revised manuscript for specific estimation results and modifications

 

Question 3. In the same table the health status needs to be explained. what makes the individuals to be considered healthy.

Author's Response: Thank you for pointing this out. We agree with your comment. The CLES survey assessed the health status of respondents, categorizing them into five levels: loss of labor capacity, poor, fair, good, and excellent. In this paper, we measure the health status of household members by the proportion of individuals with "good" and "excellent" health. Please refer to Table 1 for specific definitions

 

Question 4. In Fig 1, the captions in the left side of the graph illegible making it possible to follow the explanations in the context. the quality of the graph and explanations should be improved.

Author's Response: We thank the reviewer for this suggestion. We have revised the corresponding figures and improved their quality to ensure that the content within the figures is clear and legible.

 

Question 5. The reviewer could not understand the statement of line 387, "driven by a single component....... comprehensive impact" which is seemingly to strong a statement.

Author's Response: Thank you for pointing this out. We agree with your view. As a result, we have removed the overly assertive phrases ("not driven by a single component" and "comprehensive impact") and replaced them with more cautious wording ("may play a role"). Additionally, we have emphasized that the value of this decomposition analysis lies in offering a more detailed perspective for understanding. Please refer to lines 379-381 for detailed modifications

 

Question 6. Land operation scale mentioned in line 416-; small scale being 100 acre and so on. this reviewer wonders if 100 acre is a small scale in China, in many countries it is not, explain for confirmation.

Author's Response: Thank you for pointing this out. We agree with your comment. Therefore, we have made revisions and separately analyzed the cases of land holdings under 50 acres, between 50 and 100 acres, and above 150 acres, instead of using 100 acres as the dividing unit. The estimation results show that, with this revised approach, the larger the land operation scale, the greater the impact of agricultural digitization on land productivity, and the effect is more significant. This revision alleviates concerns about potential estimation bias caused by the classification of household land operation scale. Please refer to lines 451-455, as well as columns 1-3 in Table 7 for detailed modifications

 

Question 7. The explanation of household human capital and old age need to be clarified cross checking with these 2 variables. This will be needed to substantiate what authors say, "digital technology can compensate for skill deficiencies......, they may be attributed to low human capital often engage in rudimentary.....and so on" and the explanations in the section of elderly age, workers.

Author's Response: Thank you for the reviewer’s valuable suggestion. The recommendation to cross-check the relationship between household human capital and aging variables to better support the paper's conclusions is highly important. In response, we employed a stratified analysis approach to verify the robustness of the aging (human capital) heterogeneity effect within different human capital (age) groups, providing specific statistical evidence. The estimation results are shown in Appendix Table 1.

The results indicate that for households with lower human capital, the impact of agricultural digitization on improving land productivity is significantly higher, regardless of the degree of aging, compared to households with higher human capital. Similarly, for households with higher levels of aging, the effect of agricultural digitization on land productivity improvement is notably higher, regardless of their human capital status, compared to households with lower aging levels. This test enhances the rigor and completeness of the heterogeneity analysis and provides stronger empirical support for the core argument of the paper.

Please refer to lines 487-493 of the revised manuscript for specific additional content.

Appendix Table 1.  Heterogeneity Analysis – Cross-Validation of Household Human Capital and Aging Variables

Variable	High Human Capital		Low Human Capital
	Aging Level (Low)	Aging Level (High)	Aging Level (Low)	Aging Level (High)
	landpd	landpd	landpd	landpd
dig	0.043*	0.171*	0.261*	0.315**
	(0.024)	(0.088)	(0.153)	(0.143)
Constant	6.332***	6.519***	7.033***	4.924***
	(0.427)	(0.549)	(0.392)	(0.819)
Fix county	Yes	Yes	Yes	Yes
Fix year	Yes	Yes	Yes	Yes
Control variables	Yes	Yes	Yes	Yes
Observations	530	439	675	377
R-squared	0.131	0.163	0.114	0.175

 

Question 8. Again, this gets related with previous comment, are we to consider any mechanical tools with some digital aspect, no matter how crude they may be, to be considered as “digitized”? and in such case will not every technology be said to be digitized? then low mean value in table 1 mentioned above is a perplex.

Author's Response: Thank you for your valuable suggestions. In this paper, not every technology, or every mechanical tool with digital functions, is considered as part of agricultural digitalization. The control variables include a variable for "use of agricultural machinery" (mech), with a mean value of 0.993, which is much higher than the mean value of agricultural digitalization (0.293). In the CLES survey, the questions regarding "whether uses agricultural machinery" and "whether implemented the following agricultural digitalization projects" are two separate questions. The latter includes options such as digitalization in crop production, seed industry, agricultural machinery equipment, and agricultural reclamation. Therefore, in this paper, we do not assume that a village has achieved agricultural digitalization simply because a household uses machinery with certain digital functions. Regarding the concern about the low mean value of agricultural digitalization in Table 1, this has already been addressed in our response to the second question, and we will not repeat it here.

 

Question 9. In table 6, the constant value for large scale operation is not significant, explain.

Author's Response: Thank you for pointing this out. There are various reasons why the constant term may be statistically insignificant. In econometric regressions, it is common for the constant term to be insignificant. Although the significance of the constant term is not typically a core focus of regression analysis, its statistical result and economic meaning still require proper interpretation. Below are the potential reasons for an insignificant constant term and its possible economic implications: (1) Inclusion of Sufficient Control Variables: If the model includes all key explanatory variables, these variables may explain most of the variation in the dependent variable, leaving little unexplained variation, which can cause the constant term to be insignificant. (2) Data Distribution Characteristics: If the mean of the data is close to zero or if the explanatory variables are highly correlated with the changes in the dependent variable, the constant term may become insignificant. (3)Small Sample Size: A small sample size can lead to instability in estimators. In such cases, the standard error of the constant term may be large, resulting in reduced statistical significance.

In the heterogeneity analysis conducted earlier in this paper, the sample size for households with land holdings of more than 200 acres was relatively small. This small sample size likely increased the standard error of the constant term, reducing its statistical significance.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The author has answered all my questions. no more questions.

Author Response

Thank you for recognizing our work. Your suggestions have been immensely helpful in improving our manuscript. Once again, we sincerely appreciate your valuable feedback and guidance!

Reviewer 3 Report

Comments and Suggestions for Authors

Most of the comments were well addressed. However, couldnt find the Appendix Table 1. anywhere in the manuscript. It must be provided.

Authors ahve supplied PSM analysis as an additional writeup to support the main analysis, which can be said to be an academic excercise. However, what caliper and why was it used in the caliper matching is not clear. Further, "After matching, the differences in sample characteristics are significantly reduced, and both the common support and balance assumptions are satisfied." these results generated during the analysis must be shown to the readers, in graph or table in the manuscript main contexr or appendix, for better understanding.

Author Response

Question 1. Comments and Most of the comments were well addressed. However, couldn’t find the Appendix Table 1 Suggestions for Authors anywhere in the manuscript. lt must be provided.

Author's Response. In response to your comments, we have relocated Appendix Table 1 into the main text (now Table 8) and added descriptions to facilitate reviewers' and readers' understanding. The relevant modifications can be found in lines 495–503 of the revised manuscript.

Question 2. Authors have supplied PSM analysis as an additional writeup to support the main analysis, which can be said to be an academic exercise. However, what caliper and why was it used in the caliper matching is not clear. Further, "After matching, the differences in sample characteristics are significantly reduced, and both the common support and balance assumptions are satisfied. These results generated during the analysis must be shown to the readers, in graph or table in the manuscript main context or appendix. for better understanding

Author's Response. (1) Methodology for caliper matching and rationale: Thank you for raising this point. Following general guidelines, we set the caliper value at 0.25 times the standard deviation of the propensity score (0.03) to ensure covariate balance between the treatment and control groups after matching, while minimizing the loss of sample size. We have also included information on the two other matching methods employed. Relevant revisions are provided in lines 393–398 of the manuscript. (2) Visualization of post-matching sample characteristics: We agree with your suggestion to present post-matching sample characteristics visually. In Appendix 1, we have added a table displaying standardized bias statistics for covariates after matching, allowing readers to better understand the matching effectiveness and its implications for the study results. Details can be found in lines 637–638 of the revised manuscript.

Author Response File: Author Response.docx