The Impact of Urban Public Transport on Residential Transaction Prices: A Case Study of Poznań, Poland

Round 1

Reviewer 1 Report

This paper deals with an interesting topic, i.e. the effects of public transport policies in housing prices. I believe this topic is of good interest, since collective public transport is often seen as a "social" solution, which should increase the life of the poorer; however, an efficient collective public transport raise even more the interest of the richer, moreover in the case of big cities. 

Based on this premise, I have a few comments to improve the structure and contents of the paper:

• I believe the authors should divide the introduction from the literature review into two different sections.
• The introduction should better highlight the research question of the paper and the innovation provided in the used method. The authors, indeed, highlight that the paper use a consolidated approach, but the innovation with respect to this approach is not clear
• Literature review could include more papers focusing on the relationship between transport and social inclusions, e.g.:
  • Lucas, K. (2012). Transport and social exclusion: Where are we now?. Transport policy, 20, 105-113.
  • Giuffrida, N., Inturri, G., Caprì, S., Spica, S., & Ignaccolo, M. (2017). The impact of a bus rapid transit line on spatial accessibility and transport equity: The case of Catania. In Transport Infrastructure and Systems (pp. 753-758). CRC Press.
  • Currie, G., Richardson, T., Smyth, P., Vella-Brodrick, D., Hine, J., Lucas, K., ... & Stanley, J. (2010). Investigating links between transport disadvantage, social exclusion and well-being in Melbourne–Updated results. Research in transportation economics, 29(1), 287-295.
• The authors should include, at least literature review and in the discussion section, some reflections on the social consequences of the relation between housing prices and public transport improvement. Do you think the method could be used by public administrations? How should the transport decision-makers behave to improve the quality of life of the poorer?

Author Response

Thank you very much for your comments. I am submitting a revised paper, I have highlighted in red in the document the corrections that address your comments.

Author Response File: Author Response.pdf

Reviewer 2 Report

Dear Author(s):

Thank you for the opportunity to review the submitted article “The impact of urban public transport on residential transaction prices: a case study of Poznan, Poland” (IJGI-1492892). I have read the article with interest and have attached my comments and notes below. While I find the application to be of interest and well presented, I feel there are empirical improvements that can be included to strengthen to work – particularly as the literature and methodologies related to hedonic modelling of urban amenities is already quite expansive and making use of sophisticated spatial methods and quasi-experimental designs.

There are a number of missing elements from this work related to the empirical specification that I would like to see. There is little discussion regarding the choice of model and why weighted least squares methods were used, missing diagnostics from the estimates (concerning things like multicollinearity between distance variables especially), and lack of description for the variables used and their respective sources. Further, considering the explicit spatial nature of the data and the study context, I am surprised that no spatial effects (through the dependent variable or error term) have been explored. There are likely to be omitted variables here and controlling for spatial autocorrelation could impact final estimates. At the very minimum including spatial fixed effects (location dummy variables) may capture some of these dynamics.

I am also surprised at the insignificant result on housing prices related to the proximity to bus stops. Given the relatively small sample size of the data, and vagueness over the measure of distance used, it is difficult to be confident in these estimates. For example, it is not known what measure of distance (Euclidean or network) is used, or whether other potentially more meaningful measures were modelling. A dummy variable for having a bus stop nearby; a count of all stops within a given walking radius; controlling for the fact that different stops may have different number of routes available and thus be more accessible. Without ruling out or trying these different robustness measures, there can be little confidence in saying that bus stops have no impact on prices.

Some sections of the paper would also benefit from some reworking to help the flow – namely the second section on Materials and Methods. Here, I would suggest to more explicitly define the two – first speak to the data and study region (materials) and then to the methodology being used. In the data and study region section, it would be useful to know more about the context of the area and population, the prevalence of private vs. public transport use, demographics, etc. This could help inform on issues of external validity of the results.

Further, while obviously not the point of this work, it would be worthwhile to speak to the state of the residential real estate market during 2020. Given that the transactions are all occurring in 2020 – a time where lockdowns, moving, and all sorts of dynamics were at play, it would be useful to know what was going on in the local housing market. In some countries, trading had to stop, in others transactions could continue. Some areas had particular strong demand pressures as people tried to move into new dwellings once working from home became more engrained. These dynamics all could potentially have important impacts on how residents are using public transportation – especially in COVID-19 scenarios when people were adverse to these modes of transport. There could be particular effects coming from transactions in 2020 and decreasing attractiveness of public transport at a result to should be spoken to in the paper.

On a similar note, it would be good to know whether the number of transactions in 2020 is representative, or comparable to times when the real estate market was ‘normal’. Are the 2,561 transactions observed during 2020 low compared to other years; or structurally occurring in specific areas or markets? Making sure that the data is representative of a normal scenario, and not a market with particular specific pressures (caused by COVID and lockdowns), is important for validating the estimated results.

Overall I find the study to be well presented, however the empirical work could be improved. This is especially considering the wide range of existing hedonic estimates on public transport in the literature which take into account spatial effects and interactions – often making use of a quasi-experimental design to discern impacts and effects. Tightening up this section to include a better description of the variables used and the chosen model – along with the appropriate diagnostics for model fit (and importantly in this case multicollinearity between distance variables) would benefit this work.

I wish you best of luck with any revisions of this manuscript, and a happy holiday season.

Best regards.

Author Response

Thank you very much for your comments. I am submitting a revised paper, I have highlighted in red in the document the corrections that address your comments.

Author Response File: Author Response.pdf

Reviewer 3 Report

“The impact of urban public transport on residential transaction 2 prices: a case study of Poznań, Poland”

Thank you for allowing me to read and comment on your article. I think your manuscript needs extensive revision. The main drawbacks of the paper are the following.

1. Introduction

• The content of this section contains general information, so some related literature reviews are needed.
• Please refer to the submission guide, such as a spacing guide between the last character and citation of a sentence.

2. Materials and Methods

• According to Table 1, there are many variables that you found from the present studies, but the criteria for selecting only those variables adopted to the research model is unclear. In this aspect, some more detailed explanations and criteria are needed.
• Regarding as the reduced scope of variables, there should be evidence and related data such as surveyed data needed. For example, the variable, “Distance to schools”, focuses on the accessibility only to the primary schools and kindergartens. This paper already mentioned that accessibility to schools is very important to young children because older students attending secondary schools can go to school by public transportation. However, this study focuses on the general housing prices, so it should contain all other kinds of schools as long as you employ the variable of “Distance to schools”.

3. Results

• The variable, “Age of a building” must affect the strong impact on the housing prices, but it is not applied in the research model of this paper. A more detailed explanation and evidence should be depicted if that variable is not needed in this model.
• (Table 4) Please explain the criteria why you divided the variables, “Number of rooms” and “Floor” into three.
• (Table 5) 5 independent dummy variables of 13 ones are included in the model(OLS). The more independent dummy variables, the better R² Moreover, the result mostly depends on the P-values of the dummy variables because the P-values of the variables such as bus stop and school are 0.725 and 0.249 respectively. These values don’t have explanatory power, but the other variables such as Shopping centers do.
• (Table 5) This study focuses on the impact of the accessibility to public transportation on housing prices, but there are only two public transportation-related independent variables: “Distance to tram stop” and “Distance to bus stop”. Moreover, the P-value of tram stop is less than 0.001 for both OLS and WLS, and that of the bus stop is 0.725 for OLS and 0.718 for WLS. In this case, it’s difficult to determine the impact of those variables without detailed information. For example, the VIFs for those variables should be checked for multicollinearity because the P-value of the tram is too small.

Author Response

Thank you very much for your comments. I am submitting a revised paper, I have highlighted in red in the document the corrections that address your comments.

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Please see the attachment. 

Comments for author File: Comments.pdf

Author Response

Thank you very much for your comments. I am submitting a revised paper, I have highlighted in red in the document the corrections that address your comments. Please see the attachment.

Author Response File: Author Response.pdf