Heterogeneous Effects of ICT across Multiple Economic Development in Chinese Cities: A Spatial Quantile Regression Model
Round 1
Reviewer 1 Report
The paper entitled
Heterogeneous effects of ICT across multiple economic development in
Chinese cities: a spatial quantile regression model
is well documented regarding the financial literature
and fine structured.
The Abstract provide sufficient information about
the scope of the paper.
Introduction provide enough evidence about
the literature regarding the topic of the paper.
Chapter 2. Methodology is clearly explained.
There may a problem with the equations - they look different in the document. For example
in the equation (8) the last term is not clear / visible.
All the equations should be rewritten , if the problem is visible to other, too.
(Do not take into consideration
if others can see normal equations, maybe it is my Acrobat that do not display them quite right.)
Chapter.3 Data is well explained.
Data sources are presented, and also the statistical description of the data.
We consider that some comparison with other studies data, if available, can be inserted.
Chapter 4. Results and Discussion shows extensive information about the results.
It is not clear,though, why the SAR results are presented, as the authors said in lines 321-322 that their findings could not be accurate:
"However, SAR model analyses influencing factors homogeneity effects on urban
economic development which may be not accurate".
We suggest to keep only the best results -QSAR and to move not so accurate results in Appendix, only for comparison.
There is necessary to extend to further explanations why the QSAR is a better methodology and provides better results.
In our opinion, there is also necessary to present a comparison of the results obtained with other studies.
For example, a phrase such as
"The results are in line / not in line with " can be inserted.
Chapter / section 5. Conclusions provides enough information about the study, results, limitations and future research.
Author Response
Response to Reviewer 1 Comments
Respected Reviewer,
Thank you very much for you reviewing and giving valuable suggestions. We have read your review report repeatedly, and tried our best to revise according to the report. Our responses are listed below. Please let us know if we could be of any further work.
Best Rewards,
Authors
List of Responses and Actions:
POINT 1: Chapter 2. Methodology is clearly explained. There may a problem with the equations - they look different in the document. For example in the equation (8) the last term is not clear visible. All the equations should be rewritten, if the problem is visible to other, too. (Do not take into consideration if others can see normal equations, maybe it is my Acrobat that do not display them quite right.)
Response 1: We have checked our paper and rewrote all the equations with Math Type to avoid invisible equation due to reading software difference. Thank you very much for you reminding us the possible mistake.
POINT 2: Chapter.3. Data is well explained. Data sources are presented, and also the statistical description of the data. We consider that some comparison with other studies data, if available, can be inserted.
Response 2:Thank you very much for you reminding comparison with other studies data. We found similar studies from Eaton and Eckstein (1997), Sharma(2003), and Black and Henderson(2003), their results are consistent with Zipf’s Law using data from Japan, India and the US. But our result in Fig. 1 shows some difference in the tail of the point range. Consequently, we further compare our result with the studies of Liu and Sun(2009). They compare the spatial distribution of innovation activities which is highly correlated to economic development in China and the US, and the spatial distribution of patent distribution at province level in China is similar to Fig. 1. The figure below is the rank-frequency distribution of invention patents of China in 2006 from paper by Liu and Sun(2009). Through the comparison, we will use QSAR model with reasonable confidence. Thank you for your valuable suggestion again and again.
The rank-frequency distribution of invention patents of China in 2006
POINT 3:Chapter 4. Results and Discussion shows extensive information about the results. It is not clear, though, why the SAR results are presented, as the authors said in lines 321-322 that their findings could not be accurate: "However, SAR model analyses influencing factors homogeneity effects on urban economic development which may be not accurate". We suggest to keep only the best results -QSAR and to move not so accurate results in Appendix, only for comparison. There is necessary to extend to further explanations why the QSAR is a better methodology and provides better results. In our opinion, there is also necessary to present a comparison of the results obtained with other studies. For example, a phrase such as "The results are in line / not in line with " can be inserted.
Response 3:Thank you for your valuable suggestion on results and discussion. It could be argued that, if we keep only the results of QSAR in main body, it would be hard to confirm the necessity of spatial lag term at different quantiles using statistic like Global Moran’s I because we cannot find similar statistic to test spatial independence at quantiles. Additionally, we confirm that SAR model is more suitable than SEM model in this research using Lagrange multiplier (LM) and robust Lagrange multiplier (Robust LM). If we move the results of SAR and SEM out of main body, we are afraid that the LM and Robust LM test results seem lack of evidence. Lastly, through homogeneity analysis compare results from SAR and SEM models, we try to confirm existence of ICT directive effect and spillover effect on urban economic growth which could be the basis to analyze heterogeneity.
Thank you very much for you reminding to present a comparison with other studies for explanations why the QSAR could be better. We find that more articles consider using SAR or SEM to analyze the same issue with ours, so we only present a comparison between results of SAR and other studies. According to your suggestion, we compare our results with Lin et al.(2017), Haini (2020) and Doong and Hu (2012). Our results are in line with theirs that we confirm ICT positive effect on economic, but ours are not in line with theirs that we could provide more details and statistical evidence using QSAR.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The paper investigates the heterogeneous economic effects of ICT between developed and less developed Chinese cities. To do this, it uses a quantile spatial autoregressive model. Two main results are obtained: first, it is found a significantly positive economic effect of ICT in local and neighbouring cities; second, ICT-related variables have a larger impact on less developed cities than in developed ones. These results suggest the existence of a double dividend from ICT, namely, increasing overall economic growth and triggering a process of convergence among cities.
A better understanding of the role of ICT in economic development is certainly needed. In this respect, the empirical results obtained in this paper can serve as a basis for theoretical research on the impact of ICT on total factor productivity and convergence among local economies within a country. This seems to me a relevant contribution.
My comments are listed below.
- In the abstract (on line 14), the authors refer to GDP per capital. However, in the analysis the dependent variable is GDP per worker.
- In the introduction (on lines 26 through 31), the authors give some statistical information regarding ICT penetration across countries and in China. However, no source is provided for this information. The authors should provide data sources.
- In the introduction (on lines 95 through 97), the authors argue that, unlike cross-country studies, “The case of cities can help us better understand the ICT-economy relationship since all cities share same education system, culture and national strategy. Therefore, any heterogeneity of ICT effect on economy across cities can be attributed to differences of urban economy.” However, it could be argued that, though the education system is the same, people’s education level could differ significantly across cities. Similarly, the existence of a national strategy does not mean that this strategy operates in the same way in every city; for instance, the national strategy would operate differently in agriculture-based local economies than in industry-based ones. Thus, even when analysing the impact of ICT across cities, some factors, such as education levels, sectoral structure, among others, may explain economic development. The authors should elaborate deeper their argument of leaving aside these factors. I worry about the effects of omitting relevant explanatory variables.
- In the introduction (on lines 132 through 134), the authors assert “As far as the authors know, only few articles consider both spatial dependence and heterogenous effects of ICT in one empirical research.” The authors should cite these few articles after this sentence. Also note that there is a typo in the sentence: it should say heterogeneous instead of “heterogenous”.
- On line 223, it is said that data come from China City Statistical Yearbook of 2018. Afterward, the authors refer to year 2017. This information should be provided at the beginning of the section.
- In subsection 4.2 (on line 313), it is said that “…will increase local city’s GDP per capital…” However, the dependent variable is GDP per worker.
- In Figure 2, set the same scale in y-axis (0.4 to 0.7, for instance) to allow an easier comparison of results in panels (a) and (b).
Author Response
Response to Reviewer 2 Comments
Respected Reviewer,
Thank you very much for you reviewing and giving valuable suggestions. We have read your review report repeatedly, and tried our best to revise according to the report. Our responses are listed below. Please let us know if we could be of any further work.
Best Rewards,
Authors
List of Responses and Actions:
POINT 1: In the abstract (on line 14), the authors refer to GDP per capital. However, in the analysis the dependent variable is GDP per worker.
Response 1: Thank you very much for reminding. Due to translation error, we used different nouns to express one concept. We have unified the name with GDP per capital.
POINT 2: In the introduction (on lines 26 through 31), the authors give some statistical information regarding ICT penetration across countries and in China. However, no source is provided for this information. The authors should provide data sources.
Response 2:Thanks for you reminding us to provide data sources. We have added source of Chinese data in the paper. And we also used newer data on global Internet development. The previous data reached only 2015.
POINT 3: In the introduction (on lines 95 through 97), the authors argue that, unlike cross-country studies, “The case of cities can help us better understand the ICT-economy relationship since all cities share same education system, culture and national strategy. Therefore, any heterogeneity of ICT effect on economy across cities can be attributed to differences of urban economy.” However, it could be argued that, though the education system is the same, people’s education level could differ significantly across cities. Similarly, the existence of a national strategy does not mean that this strategy operates in the same way in every city; for instance, the national strategy would operate differently in agriculture-based local economies than in industry-based ones. Thus, even when analysing the impact of ICT across cities, some factors, such as education levels, sectoral structure, among others, may explain economic development. The authors should elaborate deeper their argument of leaving aside these factors. I worry about the effects of omitting relevant explanatory variables.
Response 3: Thank you very much for your suggestion. We strongly agree your argument that educational levels and sectoral structure will influence economic growth of cities. The better method to solve the problem could be panel data QSAR, so the urban fixed effect in panel data model could allow us to control the time-invariant unobserved urban characteristic. And a set of time-varying urban controls (R&D intensity, education, industrial structure) associated to urban economic growth could be added to the panel data model as explanatory variables. We are trying our best to find suitable variables and more detailed data to support the empirical study. However, we cannot provide better empirical results due to the limitation of city level data. We have considered your suggestion in section 5 as the limitation of the paper. Thank you very much for you pointing out the lack of the paper, and we will take further study on the issue.
POINT 4: In the introduction (on lines 132 through 134), the authors assert “As far as the authors know, only few articles consider both spatial dependence and heterogenous effects of ICT in one empirical research.” The authors should cite these few articles after this sentence. Also note that there is a typo in the sentence: it should say heterogeneous instead of “heterogenous”.
Response 4:Thank you very much for you reminding to add articles. We have added some articles by Cardona et al.(2013), Corrado(2017) and Lin et al.(2017). They considered both spatial dependence and heterogenous effects of ICT without QSAR. Additionally, we have replaced “heterogenous” with heterogeneous. Thanks for you reminding the typo again.
POINT 5: On line 223, it is said that data come from China City Statistical Yearbook of 2018. Afterward, the authors refer to year 2017. This information should be provided at the beginning of the section.
Response 5: Thank you for you reminding. We have explained that the data refer to year 2017 on line 223. And we also have stated the source of the yearbook. The revised words are “The Chinese data of 275 cities in year 2017 are obtained from China City Statistical Yearbook of 2018. The yearbook is published by National Bureau of Statistics of China and supported by departments at provincial and county level.”
POINT 6: In subsection 4.2 (on line 313), it is said that “…will increase local city’s GDP per capital…” However, the dependent variable is GDP per worker.
Response 6:Thanks for you reminding. We have unified the name with GDP per capital in response on ARGUE 1. Thanks for reminding the typo again.
POINT 7: In Figure 2, set the same scale in y-axis (0.4 to 0.7, for instance) to allow an easier comparison of results in panels (a) and (b).
Response 7: Thank you for your suggestion about figure. We have revise Figure 2 as your suggestion for a easier comparison. The revised figures are as follow:
|
(a) |
(b) |
(a) lnICT (local ICT effect) (b) WlnICT (spatial spillovers of ICT)
Figure 2. Regression coefficients of ICT-related factors based on QSAR model.
Author Response File: Author Response.pdf
Reviewer 3 Report
This paper focuses on the heterogeneity of ICT effects on economic growth in cities with different level of economic development (most developed and less developed cities), focusing on the Chinese case. It applies a quantitative approach, based on spatial econometrics models.
This is a particularly good, and interesting, paper. The methodology is clearly and rigorously explained, and the findings are clear, coherent and relevant: they apparently contribute to increase the knowledge on this issue.
Moreover, I appreciated the clear and consistent structure of the paper. I mean, the sequence of the sections and the arguments presented in the paper.
Lastly, the literature is rich, very well presented, and the added value provided by the paper within the literature, I mean the innovative contribution provided by this work, is well explained, and highlighted.
I have only few comments and suggestions, to further improve the paper:
1) I would add a thematic map of China, representing the GDP per capita at the geographical level used in the analysis. A cartographic representation is always a useful tool, which helps readers to identify and interpreter the geographical question under examination. Consider, for example, that in the Introduction (Section 1) authors talk about the geographical gap between East and West of China (a map showing this pattern would help a lot).
2) I would spend more text to describe, exhaustively and precisely, the data used. I refer in particular to the beginning of Section 3, when authors mention the China City Statistical yearbook 2018. I wonder, for example:
- Can authors say something more about the source of the yearbook and the data? Which institution – public or private - produce this publication and these data?
- Are cities and regions the same spatial units in China? I mean, are they the same geographical level of administrative division / breakdown? Authors talk about cities and regions (see for example lines 223-226) as if they are the same, but I am not sure; they should be clearer about it.
3) I think that these results can be useful even in other countries, where there is a clear-cut geographical divide, like Italy and Germany. The North-South divide in Italy is a case of intra-national macro-regional gap very well-known in the international economic literature, which also concerns the diffusion and the use of ICTs. I think that, in the conclusions (Section 5), authors should reflect also on the potential of this analysis, which be applied also in other contexts and cases, like Italy. I suggest, in this respect, to take into consideration the main contributions on the North-South divide in Germany and in Italy (see for example: Bianchi et al 2019, Rivista economica del Mezzogiorno; Maseland 2014, Regional Studies; Musolino 2018, European Spatial Research & Policy; Felice 2015, MPRA Paper).
In conclusion, line 399: “larger” instead of “lager”. Please check all the text, as I have seen the same mistake in other parts.
Author Response
Response to Reviewer 3 Comments
Respected Reviewer,
Thank you very much for you reviewing and giving valuable suggestions. We have read your review report repeatedly, and tried our best to revise according to the report. Our responses are listed below. Please let us know if we could be of any further work.
Best Rewards,
Authors
List of Responses and Actions:
POINT 1. I would add a thematic map of China, representing the GDP per capita at the geographical level used in the analysis. A cartographic representation is always a useful tool, which helps readers to identify and interpreter the geographical question under examination. Consider, for example, that in the Introduction (Section 1) authors talk about the geographical gap between East and West of China (a map showing this pattern would help a lot).
Response 1: Thanks for your suggestion. We have drawn the cartograph of GDP per capital at the province level and added it in Section 1. The cartograph is as follows:
ARGUE 2. I would spend more text to describe, exhaustively and precisely, the data used. I refer in particular to the beginning of Section 3, when authors mention the China City Statistical yearbook 2018. I wonder, for example:
- Can authors say something more about the source of the yearbook and the data? Which institution – public or private - produce this publication and these data?
- Are cities and regions the same spatial units in China? I mean, are they the same geographical level of administrative division / breakdown? Authors talk about cities and regions (see for example lines 223-226) as if they are the same, but I am not sure; they should be clearer about it.
Response 2: Thank you for you reminding us to state data source. (1) The yearbook is produced by public institution. We have stated the source of the yearbook in paper that “The yearbook is published by National Bureau of Statistics of China and supported by departments at provincial and county level.” (2) Regions usually include a large number of cities in China. For example, Lin et al. (2017) divide three regions (Eastern, Central, and Western regions in China) including numbers of cities. We have replaced “region” with “city” to state clearly.
ARGUE 3. I think that these results can be useful even in other countries, where there is a clear-cut geographical divide, like Italy and Germany. The North-South divide in Italy is a case of intra-national macro-regional gap very well-known in the international economic literature, which also concerns the diffusion and the use of ICTs. I think that, in the conclusions (Section 5), authors should reflect also on the potential of this analysis, which be applied also in other contexts and cases, like Italy. I suggest, in this respect, to take into consideration the main contributions on the North-South divide in Germany and in Italy (see for example: Bianchi et al 2019, Rivista economica del Mezzogiorno; Maseland 2014, Regional Studies; Musolino 2018, European Spatial Research & Policy; Felice 2015, MPRA Paper).
Response 3: Thank you very much for your valuable suggestion. We are glad to adopt the suggestion and add a new paragraph to talk about possible cases, like Italy and Germany. We have read the literatures above carefully and considered how the diffusion of ICT can be used in Italy and Germany to narrow the regional divide in both nations. The added paragraph in section 5 is as follows:
The “double dividend” from ICT can be obtained by other countries where there is a clear-cut geographical divide. The spatial inequalities in digital development of households and individuals in Europe at regional level have been identified[49]. In the scenario, ICT penetration would also have dividend on narrow the regional gap. Musolino (2018)[50] concerns “perception gap” would contribute to regional development gap because Italian entrepreneurs have a stereotyped, much too negative, image of Southern Italy. Through bridge communication channel beyond barriers of geographical distance, ICT would help to tear down the “wall in the head” of entrepreneurs. Similarly, poor peripheries grow faster than richer ones throughout Germany[51] which could be contributed to ICT development strategy partly based on the empirical study of China.
Thanks very much for providing important and interesting literatures which have been quoted in our paper partly.
ARGUE 4. In conclusion, line 399: “larger” instead of “lager”. Please check all the text, as I have seen the same mistake in other parts.
Response 4: Thank you for you reminding the typo. We have checked all the text and revise the mistake.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The paper can be published in the revised form.
