Improving Appraisal Methodology for Land Use Transport Measures to Reduce Risk of Social Exclusion

Round 1

Reviewer 1 Report

This is generally a well written paper.

A few minor points:

social capital normally measures the extent of our connections with other people. Of course trust is an important element of social capital but maybe it's better to use trust in this paper.

I would like to see some discussion about how this methodology can be used in the post covid world.

 

Author Response

Thank you for your comments/suggestions. We have dealt with these matters as follows:

Point 1. About trust. We have expanded the understanding of social capital at line 259 in the original version of the paper to respond to this point and included a footnote to further elaborate, as follows: 

• social capital: the benefit a person derives from social networks, trust and reciprocity within a community (or communities), with bridging and bonding social capital associated with social networks the main research focus here [69,70]. While social capital is a widely used, it remains a controversial concept, as there is no agreed or consistent definition or measurement. Yet, there are agreed concepts around networks, trust and reciprocity, and also about the difference between bonding and bridging forms of social capital. The authors have used networks because of the obvious connection between networks and mobility, the main research focus. In addition, it could be argued that trust and reciprocity can be developed more readily if people can meet and form networks.

[1] Of interest, trust was tried as an additional explanatory variable in the model in Table 2 but did not contribute significantly to risk of social exclusion. However, trust is highly correlated with bridging (p=.035) and bonding social capital (p<.001) networks, suggesting that building those networks will support development of trust

 

 

Point 2. About using this methodology in post COVID world

The following has been added from line 472 in the original paper.

By way of illustration, COVID 19 has encouraged greater working from home in many cities (REF). This greater local focus should be supportive of the growth of stronger communities, a core planning direction in cities such as London and Vancouver [10, 11]. Measures such as local trip making, social capital (especially bridging social capital), sense of community, wellbeing and neighbourhood disadvantage are all relevant (partial) indicators of strong communities. Initiatives to build stronger communities should be expected to see gains in some or all these performance indicators. Before and after studies can be used to assess how particular initiatives change one or other of these indicators at community or neighbourhood level (as applicable to the circumstances under consideration). The relevant unit benefit values derived herein, taking account of household income levels of affected people, can then be applied to the associated indicator changes, to produce monetary benefit measures, post implementation. Repeated before and after applications should then enable unit benefit values to be increasingly used in pre-implementation appraisals, as dose-response understanding of impacts on key variables accumulates. This broadens opportunities for demonstrating the benefit values associated with building stronger communities.

Reviewer 2 Report

The contribution of this paper is to extend previous research by monetising various factors related to social exclusion, highlighting the importance of understanding the impact of relevant factors and providing evidence to support the application of monetary values in the cost-benefit analysis appraisal framework for land-use and transport planning.

There are some minor questions or comments.

(1) In the literature review, would it be possible for the authors to tabulate all societal factors using a table?

(2) Throughout the text, the authors have extensively used the word 'pathway'; it is better to elaborate on this term in the first instance.

(3) In methodology, the author should have explained the indicators used in the current paper more clearly instead of a mix-up with comparisons with previous studies. And perhaps relocate lines 252-268 to the literature review section.

(4) In the discussion section, again, it would be better if a table could be provided which compares the current policy package of the case study and the proposed one based on your research.

Author Response

Thank you for your helpful comments/suggestions, which we have dealt with as follows. 

Point 1. There is no one list of who is socially excluded, as social exclusion varies between countries, time, and values. However, the most common group are people with a risk of social inclusion are those with a low income, are unemployed, have poor health, have a disability, frail aged, who have poor transport and housing, and have rural isolation. The United Nations expresses the view that social inclusion refers to people who have a lower opportunity to participate in society due to a lack of access to resources, many of which are itemised in the Sustainable Development Goals. They note that social inclusion processes require both addressing the drivers of exclusion, including certain policies, as well as discriminatory attitudes and behaviours, and actively bringing people in. The report notes that household surveys commonly omit some groups at high risk of exclusion, such as homeless persons. This research attempted to minimise this problem by purposefully accessing hard-to-reach people. We do not think that this issue can be handled in a table and also do not believe that it would add to the text if it was included.

 

United Nations Department of Economic and Social Affairs (2016) Leaving no one behind: the imperative of inclusive development: Report on the World Social Situation 2016, New York, https://www.un.org/esa/socdev/rwss/2016/full-report.pdf

Point 2. The text around line 105ff in the original version of the paper has been amended to clarify this use of the word ‘pathways’.

The evolution of that research has subsequently seen monetisation of potential benefits from reducing exclusion risk widened, to include contributions from bridging social capital, bonding social capital, sense of community, three conceptions of wellbeing, as well as neighbourhood disadvantage [24, 25]. These various ways of reducing risk of social exclusion are referred to herein as pathways to reduced risk of social exclusion.

Point 3. Lines 252-268 have been moved to the literature review section, from line 224 in the original version, with some minor re-wording of what follows to fit this change in.

Point 4. This was suggested for inclusion as a table but we prefer to extend the text by providing a couple of detailed examples of policy relevance, which would not work as well if summarized for a table. After what was line 466 in the original version of the paper, we have added:

The demonstrated benefit of trip making as a means of reducing exclusion risk poses questions for some current transport policies, such as the UK policy of providing public transport over much of the country by reliance on ‘competition in the market’, where service provision depends essentially on the financial viability of a service to its provider. The demonstrated high value of trip making derived herein suggests that the ‘competition in the market’ model can be expected to be regressive: services will be provided where users can pay, rather than where the societal benefits to at-risk people and wider society are likely to be substantial but fare revenues much less so. As a result, Stanley et al. (2022c) [8] argue for governmental financial support for transport initiatives intended to generate inclusion benefits. Such benefits are largely ignored under ‘competition in the market’, except to the extent some passengers may receive fare concessions and/or operators be provided with some financial assistance to support their operations (e.g., fuel tax rebates or subsidies).

Then continuing line 472, we have added:

By way of illustration, COVID 19 has encouraged greater working from home in many cities. This should be supportive of the growth of stronger communities, a core planning focus in cities such as London, Vancouver and Melbourne [10, 11]. Measures such as local trip making, social capital (especially bridging social capital), sense of community, wellbeing and neighbourhood disadvantage are all relevant (partial) indicators of strong communities. Initiatives to build stronger communities should be expected to see gains in some or all these performance indicators. Before and after studies can be used to assess how particular initiatives change one or other of these indicators at community or neighbourhood level (as applicable to the circumstances under consideration). The relevant unit benefit values derived herein, taking account of household income levels of affected people, can then be applied to the associated indicator changes, to produce monetary benefit measures, post implementation. Repeated before and after applications should then enable unit benefit values to be increasingly used in pre-implementation appraisals, as dose-response understanding of impacts on key variables accumulates. This broadens opportunities for demonstrating the benefit values associated with building stronger communities.

Reviewer 3 Report

1.Introduction: It is recommended to add content on the risks and negative impacts of social exclusion to enrich the research significance.

2. Lines 103, 140, 148, etc.: Two different citation formats appear multiple times in the same place in the text. Please modify them.

3. Line 247: Why is the wellbeing in this article only subjective? Why not include related research considerations such as effective wellbeing?

4. Line 258: Why did the definition of household income choose the 2008 Australian dollar price? It is now 2023, and there have been significant changes in social and economic development over the past 15 years. Is there a guarantee for the timeliness of research?

5.Line 289: What is the composition of N for all Melbourne samples reported in this article, N=765? What is the proportion of high-risk individuals with social exclusion in the sample?

6. Line 245: Reference citation appears [77], but there are only 64 references in total. Please verify if there are any missing references.

7. Line 406: Please explain how to achieve comparability between this study (2008 data) and the UK's 2019 data?

8. Results: The modeling results demonstrate that the monetary value of the influencing factors of social exclusion risk varies with income, and how does the monetization of social welfare under the role of public transportation reflect this? In addition, it is recommended to supplement the paths of social exclusion influenced by policies related to land use, transportation, and social transportation services.

9.What is the relationship between subjective well-being and social exclusion? Does the increase in the monetization value of subjective well-being mean reducing the risk of social exclusion?

10. It is recommended to supplement the conclusions and suggestions related to land use transport measures.

In Table 1 of row 316, it is recommended to add an explanation of the measurement method for Social exclusion risk factors.

12. It is recommended to unify the format of the tables in the text, including table name position, font, table lines, and formatting.

13.Can the scoring methods involved in lines 317-322 be included in the article?

14. Line 336 increases a person's subjective well-being (PWI) from the lowest level to the second lowest level, significantly reducing the risk of exclusion. However, further improvement in PWI has no significant impact on this risk. Can we discuss whether this is related to the setting of PWI scores? As mentioned in line 328 earlier, PWI is a subjective score.

15. There is an error in the layout of line 343. What is the content after "and"?

16.Can you provide a specific calculation process for the value of additional travel in lines 16.353-360?

17.Is the identification of factors influencing social exclusion risk in lines 512 to 514 a research conclusion, or is it only a part of the research content based on previous research?

Moderate editing of English language required.

Author Response

Thank you for your helpful comments/suggestions, to which we have responded as follows.

Point 1. We have added the following to the text to the original version of the paper at line 149.

Being socially excluded can involve many negative aspects for people, community and society. For example, for the individual, it can threaten the need for belonging, self-esteem, control, and meaningful existence are threatened, which may lead to negative affect, such as depression or anger (Eck et al. 2016). Having many people experiencing social exclusion in a community may risk social cohesion the building of community strengths and supports, and higher rates of crime. Society is likely to pay a higher cost the higher the numbers of people experiencing social exclusion, in terms of welfare and health costs. 

Eck, J., Schoel, C., & Greifeneder, R. (2016). Coping with or buffering against the negative impact of social exclusion on basic needs: A review of strategies. In P. Riva & J. Eck (Eds.), Social exclusion: Psychological approaches to understanding and reducing its impact (pp. 227-251). New York, NY: Springer.

Point 2. Done,

Point 3. We have added the following to the short summary of wellbeing concepts, around line 251 in the original version of the paper. 

• ... The focus herein on subjective wellbeing is because of the paper’s intention of exploring the idea of a subjective wellbeing threshold being required for inclusion. Also, subjective wellbeing is the most widely used wellbeing indicator in appraisal/impact assessment settings. Interested readers are referred to Stanley et al. (2021) [24] for monetary values of affective and eudaimonic wellbeing.

Point 4. Concerning use of A$2008 prices. We have expanded that line a little to clarify this point, as follows:

1. Household income – less than a threshold of $A500 gross per week (A$2008 prices, when the survey was undertaken), which was the highest level of welfare payment at that time.

Also, we have added the following immediately above Table 3, which contains the monetised values:

All these monetary values can be updated to current price levels for application.

Also relevant in this regard, the discussion points (1) about benchmarking subjective wellbeing values against UK and Australian values at lines 400-422 of the original version of the paper and (2) the apparent lack of influence of increased ICT use on trip making, at lines 486-505 in the original version, are both pertinent to making the case that the estimated values are still relevant.

Point 5 about survey respondent characteristics and high-risk people. An Appendix has been added to the paper which summarizes key respondent characteristics and compares them to Melbourne as a whole. Lines 306-309 in the original version of the paper talked about high-risk individuals.

Point 6. We have fixed up the references so all are now included.

Point 7 about benchmarking. We have re-worded the sentence on line 406 of the original version of the paper to respond to this point.

To benchmark the subjective wellbeing values derived in this research stream, they can be expressed as a proportion of household income and then compared to the UK proportions.

Point 8. The discussion section of the paper has been extended to provide several examples of how monetary values derived in this research program can be used in land use transport policy and planning. As an extension of line 360 in the original version of the paper we have added:

For example, if a new bus service is being planned, identifying prospective service users and their household incomes would enable application of relevant trip values in an evaluation of the case for providing the new service, within a cost-benefit framework.

Then before line 435 in the original version of the paper we have added the following paragraph, to summarise relevant examples, two of which are new additions to the paper. 

The high value of trips in contributing to reduced risk of social exclusion, particularly for those from lower income households, has a number of significant implications for transport policy. Three examples that follow illustrate this significance: for local public transport services, including establishing threshold service levels for inclusion; for setting priorities between measures that reduce trips versus those that reduce trip lengths and/or modal choices; and for 'competition in the market’ as a means of providing public transport services.

We have added the following example to the text, after line 456 in the original version of the paper:

Cities such as London and Vancouver are pursuing transport opportunity equity in their city planning largely by ensuring that there is a threshold public transport service level available across the city, supported by building strong communities [19]. The trip values derived in the current research stream can assist the development of such public transport service thresholds, since they support the process of quantifying associated inclusion benefits. Australian application, for example, has suggested minimum boarding rates of 6-7 per service hour are needed for a regional urban route bus service to generate inclusion benefits of a similar scale to service costs or 10-11 per service hour in Melbourne [15, 84].  

The other new example of application of the results for transport policy purposes has been added after line 466 of the original version of the paper.

The demonstrated benefit of trip making as a means of reducing exclusion risk poses questions for some current transport policies, such as the UK policy of providing public transport over much of the country by reliance on ‘competition in the market’, where service provision depends essentially on the financial viability of a service to its provider. The demonstrated high value of trip making derived herein suggests that the ‘competition in the market’ model can be expected to be regressive: services will be provided where users can pay, rather than where the societal benefits to at-risk people and wider society are likely to be substantial but fare revenues much less so. As a result, Stanley et al. (2022c) [8] argue for governmental financial support for transport initiatives intended to generate inclusion benefits. Such benefits are largely ignored under ‘competition in the market’, except to the extent some passengers may receive fare concessions and/or operators be provided with some financial assistance to support their operations (e.g., fuel tax rebates or subsidies).

Point 9 We have considerably revised lines 323-339 in the original version of the paper as follows, to clarify this point. 

Modelling undertaken by Stanley et al. (2021, 2022a) [24,25] used subjective wellbeing as a continuous variable. The authors suggested, however, that some minimum or threshold level of wellbeing may be needed for social inclusion but this idea was not tested. The current paper undertakes such a test, using the same database, as referred to above, with the resulting model then used to infer monetary values for changes in levels of various independent (pathway) variables. Subjective wellbeing scores (measured by the Personal Wellbeing Index PWI) [74] were split into 5 categories, with cut points at PWI scores of 6, 7, 8 and 9. Table 2 sets out the resulting ordinal logistic regression model. The coefficients attached to the respective independent variables are the natural logs of the odds ratios attached to those variables.

For the continuous variables, Table 2 shows that risk of social exclusion was significantly negatively correlated (p<0.05) with household income (per day squared) and daily trip numbers and positively correlated with age. For categorical variables, interpreting the impact of any variable depends on its significance level and on the size and sign of relevant coefficients as compared to the reference level for that variable. Thus, the model in Table 2 shows:

• a significant association between increased subjective wellbeing and reduced risk of social exclusion but this only applies for those with the lowest levels of subjective wellbeing (i.e., PWI1= PWI score below 6 out of 10), supporting the idea that a wellbeing threshold is important for inclusion. The 0.896 co-efficient attached to PWI1 infers that someone whose subjective wellbeing level is in category PWI1 is 2.45 times (95% CI = 1.29 to 4.65) more likely to be in a higher risk level for social exclusion than someone with the highest level of PWI (=PWI5, the reference level for this variable),. Alternative cut-points between PWI categories were tested during model development, the significance of this variable typically applying up to a score of around 6, so whole number cut-points have been used for ease of interpretation.

Point 10 about land use transport measures. The additions to the text in response to query 8 cover these matters in relation to land transport. For the land use side, we have added the following example as an extension from line 472 in the original version of the paper:

By way of illustration, COVID 19 has encouraged greater working from home in many cities. This should be supportive of the growth of stronger communities, a core planning focus in cities such as London and Vancouver [10, 11]. Measures such as local trip making, social capital (especially bridging social capital), sense of community, wellbeing and neighbourhood disadvantage are all relevant (partial) indicators of strong communities. Initiatives to build stronger communities should be expected to see gains in some or all these performance indicators. Before and after studies can be used to assess how particular initiatives change one or other of these indicators at community or neighbourhood level (as applicable to the circumstances under consideration). The relevant unit benefit values derived herein, taking account of household income levels of affected people, can then be applied to the associated indicator changes, to produce monetary benefit estimates, post implementation. Repeated before and after applications should then enable unit benefit values to be increasingly used in pre-implementation appraisals, as dose-response understanding of impacts on key variables accumulates. This broadens opportunities for demonstrating the benefit values associated with building stronger communities.  

Points11 (and 13). Concerning notes to Table 1 and further explanation of measurement methods. The table notes have been revised to meet these requests, the table notes now reading as follows:

a = index based on number of exclusion risk thresholds (out of 5) failed, these thresholds relating to household income, employment status, political activity, participation and social support, as explained in the literature review.

b = Index derived from four 6-point scales, relating to frequency of contacts (Never =1; Less than Once a Year = 2; More than Once a Year = 3; About Once a Month = 4; About Once a Week = 5; Most Days = 6) within close networks (members of your close family; members of your extended family; friends/intimates; neighbours). Possible scoring range = 4-24. For modelling, this variable was converted to a categorical variable by setting cut points between low/medium and medium/high ranges at scores of 17 and 20, to give approximately equal numbers of respondents in each category.

c = Index derived from two 6-point scales, relating to frequency of contact (same as in note ‘b’) with wider networks (work colleagues; people associated with groups in your community). Possible scoring range = 2-12. For modelling, this variable was converted to a categorical variable by setting cut points between low/medium and medium/high ranges at scores of 5 and 9, to give approximately equal numbers of respondents in each category.

d = Index derived from twelve 7-point scales (Sense of Community Scale) [71]. Possible scoring range is 12-84. For modelling, this variable was converted to a categorical variable by setting cut points between low/medium and medium/high ranges at scores of 54 and 63, to give approximately equal numbers of respondents in each category.

e = Index derived as average of eight 10-point scales (Personal Wellbeing Index) [74]. Treatment for modelling is discussed in the following paragraph.

Point 12. Done, thanks.

Point 13. Re scoring methods. This comment has been picked up in our response to comment 11.

Point 14. The revisions set out above in relation to comment 9 include our response to this query. 

Point 15 The 'and' is redundant and has been deleted. Thank you.

Point 16 about how values were calculated. The following text has been added in place of the sentence at lines 353-4 of the original version of the paper.

Differentiating the equation in Table 2 for risk of social exclusion (SOCEX) with respect to trips (to derive a marginal utility of trips = MUTRIPS) and household income per day squared (to derive a marginal utility of income = MUHINC, which includes multiplication by household income), produces an inferred value of an additional trip for someone at that particular household income level. For example, at mean sample household income of $226.33 per day, the following value of an additional trip results:

Marginal utility of trips (MUTRIPS) = -0.060.

Marginal utility of household income squared (MUHINC) = 2*-0.000009446*$226.33. =-0.00428

Marginal rate of substitution between trips and household income (the marginal value of a trip) = MUTRIPS/MUHINC (at sample average household income) = -0.060/-0.00428 = $A14.05 (rounded to nearest 5 cents).

Point 17 about the contribution of this paper. The sentence at lines 512 to 514 in the original version of the paper has been revised to clarify this point.

The paper has reinforced previous work by Stanley et al. (2022) [ 25], showing how monetary values can be derived for a number of factors found to influence risk of exclusion, these influencing factors (or pathways) being trips, bridging and bonding social capital, sense of community, subjective wellbeing and neighbourhood disadvantage. It has extended that prior research by showing that there are grounds to argue that a subjective wellbeing threshold needs to be achieved to reduce exclusion risk. The paper has also demonstrated how the monetary values of these various pathway influencers of risk of social exclusion are considerably higher for those from lower-income households, which argues for policies intended to reduce exclusion focussing closely on initiatives to benefit this group. Importantly, application of these pathway monetary benefit values …

 

Round 2

Reviewer 3 Report

Accept in present form

Minor editing of English language required