1. Socio-Economic Impact Assessment (SEIA) Methodology for Urban Transport Projects Presentation at Hasselt University, Belgium 13 th May 2009 By: Anvita Arora, PhD CEO, Innovative Transport Solutions, Technology and Business Incubation Unit, Indian Institute of Technology, New Delhi, India Resident Representative, Interface for Cycling Expertise, The Netherlands

2. Urbanization in India  Relatively slow, yet one of the largest urban systems  30-50% slum dwellers, ‘unauthorized’ self constructed dwellings, close to work  Growth of informal sector often faster than formal sector Bicycle ownership 30-50 % Car ownership 3-13% Scooter/M-cycle 40-50%

4. Threat to sustainable scenario: Increasing car and MTW trips

5. Transport Modes of the Urban Poor

6. PatnaJaipur Hyderabad Lu cknow Rickshaw policies? Three wheelers paratransit? Two wheelers/three wheelers? Rickshaws,cycles peds?

7. Urban transport problems Poor rely on non-motorized transport but their facilitation is often ignored Small changes in public transport fare/service can significantly affect their mobility Restraints on informal transport sector limits affordable services to the poor Dominance of private motor vehicles marginalizes NMTs Women are badly served by transport system Poor are more vulnerable to injuries and pollution 7

8. National Urban Transport Policy ( NUTP ) As per the directives of the GOI- MOUD- UT – the various proposals for urban transport being prepared under JNNURM should comply with NUTP in order to be eligible for Central Govt. funding. The focus of NUTP is on the following strategies : 1. Equitable allocation of road space – with people as focus 2. Priority to the use of Public Transport 3. Integrated public transport systems 4. Priority to non motorised transport 5. Promote multilevel parking complexes 6. Create public awareness Delhi CDP priorities and projects have been identified based on above guidelines of NUTP.

9. Delhi City Development Plan Vision and Investment 1. Equitable allocation of road space – with people as focus 33% modal share of pedestrians – investment on pedestrian infrastructure 0.5% of total investment 2. Priority to the use of Public Transport 60% of vehicular trips by public transport – Capacity building of public transport – 3 projects LRT, Monorail, HCBS – investment 42% of total investment 3. Integrated public transport systems No investment 4. Priority to non motorised transport 0.8% of total investment 5. Promote multilevel parking complexes 2% of total investment 6. Create public awareness 0.2% of total investment

10. Where is the remaining 55% investment being made? Increasing Road Length – 32% Flyovers – 10% Road Widening – 8% Spl. Scheme for CP and old city – 5%

11. Investments in flyovers,road expansion and FOBs

12. Delhi Metro – First Phase (2005) 65 km, projected ridership 1.5 m/d, actual ridership 0.4 m/d, USD 7.1m loss/yr, 100% cost overruns Existing Rail Corridors, Delhi Metro Rail System 256 km by 2021, estimated cost USD~3500 million Final Phase (2021), 60% residents & 82% area not within walking distance!!! 460 km of arterial roads,`10000 buses carrying 6 m trips

13. World class metro Over crowded buses

14. 14 INTRODUCTION Transport is a critical link between economic and social development Transport is a derived demand – need based The benefits of improving transport infrastructure have traditionally been measured by performance criteria, like improved connectivity, travel time, speeds and fuel savings for the user.

15. 15 The problem The users are not a homogeneous group Some users may benefit, some may not, and some may not be affected at all Also the non-users may be impacted – an externality (+ve or –ve) Benefits and dis-benefits to users and non- users need to be understood and internalized by transport projects.

16. 16 Need of Study Transport investments advocate inclusion of social assessment in transport projects and prioritize poverty alleviation as an objective. Need to understand: Users as a disaggregated mass (differentiated by income, occupation, gender, age, ethnicity, etc.) The gap between access availability (transport infrastructure) and mobility issues (ability of different groups to utilize the infrastructure) and their correlation with poverty (especially with respect to livelihood opportunities). A need to develop a methodological framework or model for ensuring the inclusion of socio-economic issues of transport planning in policies and projects.

17. 17 The Context Delhi  Population of 13.8 million (Census, 2001).  Modal share - 62% of the vehicular trips (33% of all trips including walk) are made by bus with an average trip length of 10.7 Km (RITES, 1994).  Heavy investments in transport infrastructure, like grade separated junctions, road widening and the Delhi Metro Rail. The Delhi Metro is a representative case study of a capital- intensive urban transport project promising to accrue high benefits of accessibility and decongestion.

18. 18 Objectives & Research Focus Objectives: To understand the impact of Delhi Metro Rail on the accessibility patterns of the urban poor. To understand the impact of changed accessibility on mobility and the socio-economic status of the low- income households. To develop indices of accessibility, mobility and SEWB and to formulate an SEIA methodology. Research focus: To understand how accessibility and mobility affect the socio-economic well-being (SEWB) of the urban poor and how indices of accessibility and mobility can be integrated in SEIA methods.

19. November 200719 Hypothesis a) Introduction of the Metro rail system in Delhi has changed the accessibility for the urban poor. b) This change in accessibility has changed the mobility profile and the socio-economic well-being of the urban poor.

20. 20 Case Study – Target Group Urban poor affected by the Delhi Metro Rail Project Urban poor as the inhabitants of slums in the city  Urban Delhi poverty line at Rs 505.45 (USD 12.64) per capita per month, (Saxena, 2001)  For Delhi slums per capita income of less than Rs. 600 (15 USD) per month for 78% inhabitants (Anand, 2006) Two categories of low-income households selected:  those living in the vicinity (within 1 km) of the metro stations, and  those relocated due to the construction of the metro.

21. November 200721

22. 22 Methodology Household survey based data collected for target group. Dataset used to derive indicators of accessibility, mobility and SEWB. The indicators aggregated into indices of accessibility, mobility and SEWB by using the Principal Component Analysis (PCA) technique. The change in indicators and indices in the before and after metro scenarios used to assess the significance of the impact of the metro project on the urban poor. The correlation between accessibility, mobility and SEWB is modeled using linear regression to illustrate that the change in accessibility and mobility due to a transport project changes the SEWB of the community.

23. 23 Structure 1. Introduction 2. Socio-economic impact assessment (SEIA) – current practices 3. Transport and poverty 4. SEIA methodology for urban transport projects 5. Accessibility, mobility and socio-economic wellbeing 6. Case study – Delhi metro rail 7. Formulation of the socio-economic impact assessment (SEIA) model 8. Conclusions, contribution and scope for future work

24. 24 SEIA – CURRENT PRACTICES Social impacts – “the consequences to human populations of any public or private actions that alter the ways in which people live, work, play, relate to one another, organize to meet their needs and generally cope as members of society”. History SIA realized as important part of EIA since 1969 to 1980’s. Partially forced by project failures resulting from inadequate appraisal of projects on narrow economic and technical criteria (Rickson et al., 1990; Burdge, 1998). WHO has pointed out that the cost of submitting major proposals for social impact assessment was far less than the cost of correcting unforeseen negative impacts that occurred after implementation (Giroult, 1983, cited in Burdge 1990).

25. 25 The Indian Scenario: The Ministry of Environment and Forests, has a separate Environment Clearance manual for large construction projects (MoEF, 2006). However, the socio-economic aspects merit only a 3 point write-up in Annexure II. Questions to be answered: 7. Socio-Economic Aspects 7.1. Will the proposal result in any changes to the demographic structure of local population? Provide the details. 7.2. Give details of the existing social infrastructure around the proposed project. 7.3. Will the project cause adverse effects on local communities, disturbance to sacred sites or other cultural values? What are the safeguards proposed? These points highlight the inadequacy of inclusion of SIA in large infrastructure projects in India and re-iterate the need for comprehensive work on it.

26. 26 The methodologies reviewed in this section are: The funding agencies’ approach  The World Bank  Asian Development Bank The SCOPE framework The implementing agencies’ guidelines  The FDOT handbook The NGOs’ perspective  Queensland Families, Youth and Community Care, Australia Impact Assessment Methodologies

27. 27 The World Bank approach: larger policy framework, generic applicability, focus on institutional mechanisms. The ADB document: comprehensive but generic –not include the special problems of transportation projects. The SCOPE framework: formulation of a socio-economic framework of a community, emphasis on the need to quantify all parameters listed but no holistic assessment design. FDOT Guidelines: focus on land use impacts of transportation projects, communities influence the use of land and vice-versa and transportation projects influence both in a correlated manner. The Australian NGO approach: emphasizes on people and their need and reactions, concepts like community sensitivity indices and the vulnerable community groups. Discussion

28. 28 Conclusion The SEIA of a transportation project must answer the following: What is the impact area of the transport project (spatial and temporal)? Who is affected by the project?  What is their socio-economic structure?  What are their needs?  What are their demands?  What is their absorptive capacity? Which are the vulnerable groups?  What is the income differential in mobility and accessibility?  What is the gender differential in mobility and accessibility?  What is the socio-cultural differential in mobility and accessibility? What is the existing transport system used (formal/informal)? What are the potential adverse impacts?

29. 29 TRANSPORT AND POVERTY Defining Poverty “a multidimensional phenomenon, encompassing inability to satisfy basic needs, lack of control over resources, lack of education and skill, poor health, malnutrition, lack of shelter, poor access to water and sanitation, vulnerability to shocks, violence and crime, lack of political freedom and voice”. The World Bank (a,1999) “poverty must be seen as the deprivation of basic capabilities rather than merely as lowness of income” (Sen, 1999). Poverty impacts of transport interventions Complex because transport is an intermediate service – transport improvements reduce poverty not through increased consumption of transport per se but through improving the quality and security of access to work, markets, and services, and through release of scarce resources for consumption and production

30. 30 Issues Efficiency vs Equity: Good transport policy contributes to poverty reduction by enhancing efficiency and equity (Gannon, et al, 2001). Access and Livelihood needs of the urban poor: Urban transport interacts with employment issues for the poor in two main ways: indirectly by providing access to employment opportunities and directly through employment of low-income people in the transport sector

31. 31 Gender Bias: Women tend to have different travel needs deriving from the multiple tasks they must perform in their households and in their communities (Greico et al, 1997). Health Impacts of Transport: Pollution (air, water, noise) effects the urban poor particularly severely, since they are the least able to avoid or seek protection from them (UNDP 1998). Pedestrian and cyclist are most vulnerable to road accidents. The Shelter-transport-livelihood link: Access to affordable transport is one of the most important factors in determining livelihoods for the urban poor The rise of private vehicular traffic has decreased bus speeds and service levels drastically and made non- motorized transport dangerous and difficult. Travel for the poor has thus become slower and more difficult even as other economic and planning forces have caused many of them to be displaced from central informal settlements to more peripheral locations (Immers et al, 1993)

32. 32 Eviction and relocation The central concern of the process of eviction and relocation is the reduction in accessibility and mobility options of the urban poor, which directly affects their livelihood and thus social well being. Transportation aspects of eviction and resettlement People evicted because of transport projects Transport implications for evicted people (due to any project)

33. 33 ACCESSIBILITY, MOBILITY AND SOCIO- ECONOMIC WELLBEING Review definitions and discussions Define Accessibility, Mobility and SEWB for the study Postulate indicators and indications

34. 34 Accessibility is a description of the proximity of destinations of choice and the facilitation offered by the transport systems (including public transport and non-motorized modes) to reach them.

35. 35 Mobility is both the ability to travel to destinations of choice and the amount of movement necessary to do so.

36. 36 Socio-economic well-being is defined as the status of a household where the basic social and economic needs for survival are fulfilled and the household has the capacity to improve its quality of life.

37. 37 Notes on subscripts: A = access, E = egress, MLH = main line haul NMV = non motorized modes including walking, MV = motorized modes

38. 38 Notes on subscripts: TR = travel, HH = household NMV = non-motorized vehicle

39. 39 * Infrastructure rank score refers to the additive score of the types of services where the service which is formally provided and operational is given a value of 2, that which is self obtained has a value of 1, and that which is not available is given a value of 0

40. 40 Impact of Transport Project (Delhi Metro Rail) Change in Accessibility: Public Transport Accessibility (A PT ) The differences in indicators for both sets of Bus users and Metro Spatio-Travel Accessibility (A ST )  Direct impact – change in indicators of A ST of households in the vicinity.  Indirect impact – change in indicators of A ST of households relocated. Change in Mobility:  Direct impact – change in indicators of Household Mobility (M HH ) and Personal Mobility (M P )of households.  Indirect impact – change in indicators of M HH and M P of households relocated. Change in SEWB:  Direct impact – change in indicators of Social Well-being and Economic Well being of households in the vicinity.  Indirect impact – change in indicators of households relocated.

41. 41 CASE STUDY: Delhi Metro Rail Legend In vicinity Relocated Part map of Delhi showing Case Study Area of Metro Rail line and locations of household survey

42. 42 Bus users and Metro users

43. 43 Household Survey In vicinity of Metro line: No significant impact on their socio-economic and travel profile. Decrease in the availability of buses since several bus-routes were realigned by policy to improve metro ridership. Considering that only 8% of their trips are on bus and 77% by walk, 4% by cycle and 6% by rickshaw, it is unlikely that these trips will be replaced by metro trips.

44. Relocated due to the metro line: Significant change in their accessibility and travel profile and income. The increasing distance, time and cost of daily travel, along with reduced incomes has a negative impact on the households. The land-use accessibility has deteriorated as distance to education, health services and other urban services has increased for 52%, 63% and 52% of the households respectively. The transport accessibility has deteriorated even more as distance to bus stop has increased for 72% of the households and the bus frequency has seen an average decrease from 5 min to 63 min (almost 13 times) November 200744

45. 45 Formulation of SEIA Model The SEIA model is formulated in 3 steps  Step I: Estimating Indicators  Step II: Developing Indices  Step III: Formulating the Model DEVELOPMENT OF INDICATORS Illustrated values of indicators, their change and significance of that change due to the introduction of the metro ACCESSIBILITY (A) Vicinity: little change in distance to education and health services. Distance to urban services like vegetable markets, daily needs shops increased for 23.6% of the households. The bus service time- gap has decreased for 34% of households Relocated: all the indicators have changed for the majority of the households. Values higher showing deterioration of accessibility

46. 46 Change Category D education (diff)D health (diff) D services (diff)D busstop (diff)S bus (diff) Households in Vicinity of metro line Total Decrease 0.0%3.0%4.9%0.5%34.5% No change 98.0%93.1%71.4%80.3%65.0% Total Increase 2.0%3.9%23.6%19.2%0.5% Households relocated due to metro line Total Decrease 40.8%33.8%36.3%13.9%1.5% No change 7.5%3.5%11.9%14.4%0.0% Total Increase 51.7%62.7%51.7%71.6%98.5% No.IndicatorsSignificance of change for HH in metro vicinity Significance of change for HH relocated At 5% confidence level At 1% confidence level At 5% confidence level At 1% confidence level 1 D education Not significant 2 D health Not significant 3 D services Not significant Significant 4 D busstop Significant Not significant 5 S bus Significant Significance of change

47. 47 MOBILITY Household Mobility (M HH ) Vicinity: some change in the indicators of PCTR for work and other purposes but little change in the PCTR for education and the share of NMVs in the modes Relocated: all the indicators have changed for the majority of the households. For 49% households, the PCTR for work has increased and for 30% of the households it has decreased. For 71% of households, the PCTR for education does not change The PCTR for other purposes has increased and decreased equally. The share of NMVs in the mode used has decreased for 59% of the households. Change categoryPCTRwork (diff)PCTRedu(diff)PCTRothers(diff)Mnmv/Mall(diff) Households in Vicinity of metro line Total Decrease 9.4%3.9%13.8%5.4% No change 77.8%91.1%81.8%87.2% Total Increase 12.8%4.9%4.4%7.4% Households relocated due to metro line Total Decrease 29.9%10.4%35.3%58.7% No change 21.39%70.65%29.35%21.89% Total Increase 48.8%18.9%35.3%19.4%

48. 48 Personal Mobility (M P ) Vicinity: minimum change in the mobility indicators regarding travel for education (distance, time, cost). The distance, time to and cost of trips made for other purposes has changes a little. Relocated: mobility indicators for travel to work – distance, time and cost – have increased for 83%, 82% and 61% of the households respectively. The distance, time for education have but not the cost. Similarly for other purposes there is more change in distance and time than the cost of the trip. Change category Dwork (diff) Deduc ation (diff) Dother s (diff) Twork (diff) Teduc ation (diff) Tother s (diff) Cwork (diff) Ceduca tion (diff) Cother s (diff) Households in Vicinity of metro line Total Decrease 10.3%3.9%15.3%13.8%4.4%16.3%3.4%0.0%4.4% No change 72.9%90.6%72.4%69.5%88.7%71.9%91.1%100.0%93.6% Total Increase 16.7%5.4%12.3%16.7%6.9%11.8%5.4%0.0%2.0% Households relocated due to metro line Total Decrease 14.9%22.9%58.2%14.4%21.9%52.2%10.4%2.5%12.4% No change 2.5%43.3%9.0%3.5%42.8%8.0%28.4%93.5%65.2% Total Increase 82.6%33.8%32.8%82.1%35.3%39.8%61.2%4.0%22.4%

49. 49 Significance of change of mobility indicators No.IndicatorsSignificance of change for HH in metro vicinity Significance of change for HH relocated At 5% confidence level At 1% confidence level At 5% confidence level At 1% confidence level 1 PCTR work Not significant 2 PCTR edu Not significant 3 PCTR others Not significant 4 M nmv/Mall Not significant Significant 5 D work Not significant Significant 6 D education Not significant SignificantNot significant 7 D others SignificantNot significant 8 T work Not significant Significant 9 T education Not significant Significant 10 T others SignificantNot significant 11 C work Not significant Significant 12 C education Not significant 13 C others Not significant SignificantNot significant

50. 50 SOCIO-ECONOMIC WELL-BEING (SEWB) Vicinity: only two indicators IRS and Household income show change with the introduction of the metro. Relocated: all the indicators have changed for the majority of the households. The indicators most affected are female literacy (21% decrease), residency (100% decrease), Household income per person (66% decrease), Infrastructure rank score (33% decrease and 61% increase), and employment (8% decrease and 14% increase). Change category NGinschl/ Ngschage (diff) NAdults>=5/ Nadults (diff) IRS (diff) Yslum/ Ydelhi (diff)W/N (diff)I/N (diff) V/N (diff) Households in Vicinity of metro line Total Decrease0.0% 3.4%0.0% 9.9%0.0% No change55.67%100.00%78.3%100% 66.01%100% Total Increase0.0% 18.2%0.0% 24.1%0.0% NA44.33% Households relocated due to metro line Total Decrease20.9%3.5%32.8%100%8.0%65.7%5.0% No change41.79%82.09%5.97%0.00%78.11%19.4%94.53% Total Increase4.5%14.4%61.2%0.0%13.9%14.9%0.5% NA32.84%

51. 51 Significance of change of SEWB indicators No.IndicatorsSignificance of change for HH in metro vicinity Significance of change for HH relocated At 5% confidence level At 1% confidence level At 5% confidence level At 1% confidence level 1 NG inschl/ Ng schage Not significant Significant 2 N Adults>=5/ N adults Not significant Significant 3 IRS Significant 4 Y slum / Y delhi Not significant Significant 5 W/N Not significant 6 I/N Not significant Significant 7 V/N Not significant Significant

52. 52 DEVELOPMENT OF INDICES Principal components are calculated using PCA Different rotations are tried to maximize loading on the principal components (PC1, PC2,…PCn) so that they explain maximum percent of the total variance. Theoretically the ‘varimax’ rotation maximizes variance explained while increasing the large loading and decreasing the smaller loadings. The higher loadings in each PC are retained and the smaller loadings are discarded in a manner so that each PC clubs together similar/ correlated indicators in a logical manner. Each PC becomes a type of factor explaining the aggregate index and each PC is independent of the others. The loadings of the retained variables in each PC are taken as indicative weights for the indicators and calculated as a fraction of 1. The ‘variance explained’ are taken as relative weights for each PC to aggregate them as an index. The value of the index is calculated for each household.

53. 53 Accessibility A = E1(PC1) + E2(PC2) Where E1 and E2 are the eigenvalues And PC1 = d(Dbusstop) + e(Sbus) PC2 = a(Ded) + b(Dhealth) + c(Dser) Where a,b,….e are component loadings. The PC1 explains accessibility provided by the bus system and the PC2 explains the landuse accessibility. The PC1 and PC2 explain approximately 55% of the total variance. The aggregated index reads as follows for the 4 data sets: In Vicinity-before metro A = 0.49(D ed ) + 0.57(D health ) + 0.62(D ser ) + 0.63(D busstop ) + 0.62(S bus ) ……… 1-a In Vicinity-after metro A = 1.07(D ed ) + 0.17(D health ) + 0.35(D ser ) + 0.52(D busstop ) + 0.52(S bus ) ……… 1-b Relocated-before metro A = 0.91(D ed ) + 0.27(D health ) + 0.49(D ser ) + 0.54(D busstop ) + 0.52(S bus ) ……… 1-c Relocated - after metro A = 0.34(D ed ) + 0.39(D health ) + 0.53(D ser ) + 0.72(D busstop ) + 0.50(S bus ) ……… 1-d

54. 54 Mobility M = E1 (PC1) + E2 (PC2) + E3 (PC3) + E4 (PC4) Where E1,E2, E3 and E4 are the eigenvalues AndPC1 = b(PCTReducation) + e(Ded) + h(Ted) + k(Ced) PC2 = c(PCTRothers) + f(Dothers) + i(Tothers) + l(Cothers) PC3 = a(PCTRwork) + d(Dwork) + g(Twork) + j(Cwork) PC4 = Mnmv/Mall Where a, b, ….l are component loadings. The PC1 explains the trip for education, PC2 explains the trip for other purposes like social, health, religious and PC3 explains the trip to work and PC4 explains only a single indicator of use of non-motorized modes. The PC1, PC2, PC3 AND PC4 explain approximately 65% of the total variance. The weight ages of the PCs imply that the trip for education and other reasons like buying daily need supplies would have a higher impact on the mobility index than the work trips, though the difference is not significant. Since M hh indicators are seen as desirable mobility and M p as undesirable mobility they are ascribed opposing signs in the index.

55. 55 In Vicinity-before metro M = [0.53(PCTR work ) + 0.79(PCTR education ) + 0.55(PCTR others ) + 1.68(M nmv /M all )] – [0.65(D work ) + 0.85(D education ) + 0.74(D others ) + 0.62(T work ) + 0.85(T education ) + 0.75(T others ) + 0.25(C work ) + 0.17(C education ) + 0.63(C others )] ……..…. 2-a In Vicinity-after metro M = [0.53(PCTR work ) + 0.78(PCTR education ) + 0.63(PCTR others ) + 1.39(M nmv /M all )] – [0.64(D work ) + 0.85(D education ) + 0.65(D others ) + 0.62(T work ) + 0.85(T education ) + 0.69(T others ) + 0.25(C work ) + 0.18(C education ) + 0.38(C others )] ……..…. 2-b Relocated-before metro M = [0.67(PCTR work ) + 0.75(PCTR education ) + 0.55(PCTR others ) + 1.58 (M nmv /M all )] – [0.74(D work ) + 0.80(D education ) + 0.61(D others ) + 0.73(T work ) + 0.80(T education ) + 0.70(T others ) + 0.53(C work ) + 0.22(C education ) + 0.31(C others )] ……..…. 2-c Relocated-after metro M = [0.73(PCTR work ) + 0.54(PCTR education ) + 0.28(PCTR others ) + 1.23 (M nmv /M all )] – [0.83(D work ) + 0.84(D education ) + 0.89(D others ) + 0.78(T work ) + 0.80(T education ) + 0.86(T others ) + 0.84(C work ) + 0.78(C education ) + 0.86(C others )] ……..…. 2-d

56. 56 SEWB SEWB = E1 (PC1) + E2 (PC2) + E3 (PC3) Where, E1, E2 and E3 are the eigenvalues AndPC1 = e(W/N) + f(I/N) + g(V/N) PC2 = c(IRS) + d(Yslum/Ydelhi) PC3 = a(NGinschl/ NGschage) + b(Nadults>=5/ Nadults) Where a, b, …. g are component loadings PC1 explains economic well-being, PC2 explains condition of physical infrastructure and PC3 explains social well-being. Together, the three PCs explain 60% of the variance. The aggregated index reads as follows for the 4 data sets: In Vicinity-before metro SEWB = 0.61(NGinschl/ NGschage) + 0.42(Nadults>=5/ Nadults) + 0.83(IRS) + 0.61(Yslum/Ydelhi) + 0.66(W/N) + 0.65(I/N) + 0.14(V/N) ……… 3-a In Vicinity-after metro SEWB = 0.57(NGinschl/ NGschage) + 0.46(Nadults>=5/ Nadults) + 0.71(IRS) + 0.62(Yslum/Ydelhi) + 0.63(W/N) + 0.63(I/N) + 0.19(V/N) ……… 3-b Relocated-before metro SEWB = 0.68(NGinschl/ NGschage) + 0.68(Nadults>=5/ Nadults) + 0.93(IRS) + 0.14(Yslum/Ydelhi) + 0.62(W/N) + 0.62(I/N) + 0.22(V/N) ……… 3-c Relocated - after metro SEWB = 0.68(NGinschl/ NGschage) + 0.66(Nadults>=5/ Nadults) + 0.60(IRS) + 0.65(Yslum/Ydelhi) + 0.72(W/N) + 0.67(I/N) + 0.06(V/N) ……… 3-d

57. 57 Significance of change in the Indices No.IndicesSignificance of change for HH in metro vicinity Significance of change for HH relocated At 5% confidence level At 1% confidence level At 5% confidence level At 1% confidence level 1 Accessibility Significant 2 Mobility Not significant Significant 3 SEWB Not significant Significant

58. 58 THE SEIA MODEL Correlation between Accessibility, Mobility and SEWB modeled in two ways 1. Correlation between the indices 2. Correlation of dependent index with independent indicators Correlation between indices Methods for linear correlation: 1. parametric: Pearson correlation (Continuous data) 2. non-parametric: Spearman correlation (Rank order data assumed) Data SetA & MM & SA & S ParametricNonparaParametricNonparaParametricNonpara In Vicinity- b4 metro -0.0010.0040.1760.1800.0350.084 In Vicinity- aft metro 0.1280.1080.1120.0890.2770.280 In Vicinity- change -0.157-0.2020.0140.114-0.170-0.177 Relocated- b4 metro -0.0340.0550.1690.1340.0570.140 Relocated- aft metro 0.001-0.049-0.039-0.090-0.065-0.125 Relocated- change 0.026-0.027-0.219-0.2290.0160.045 TOTAL -0.223-0.3350.1220.1150.0200.034

59. 59 Linear regression of dependent index with independent indicators This has been tried for the following equations (for all 4 data sets, and all repeated for each set) Index of mobility and indicators of accessibility M = a + b(AI i ) +c(AI j )+…+x(AI n ) ……………..4 Index of SEWB and indicators of mobility SEWB = a + b(MI i ) +c(MI j )+…+x(MI n ) ……………..5 Index of SEWB and indicators of accessibility SEWB = a + b(AI i ) +c(AI j )+…+x(AI n ) ……………..6 Index of SEWB and indicators of both accessibility and mobility SEWB = a + [b(AI i ) +c(AI j )+…+x(AI n )] + [b(MI i ) +c(MI j )+…+x(MI n )]..7

60. 60 Summary of Results of Linear regression No.Model usedData setR 2 valueP value for F-test 1 Equation 4 In Vicinity- before metro0.0220.49 2In Vicinity- after metro0.0200.55 3Relocated- before metro0.0250.43 4Relocated- after metro0.0510.07 5TOTAL0.1030.00 6 Equation 5 In Vicinity- before metro0.2830.00 7In Vicinity- after metro0.2570.00 8Relocated- before metro0.2000.00 9Relocated- after metro0.2830.00 10TOTAL0.2020.00 11 Equation 6 In Vicinity- before metro0.1570.00 12In Vicinity- after metro0.1300.00 13Relocated- before metro0.0110.83 14Relocated- after metro0.0120.81 15TOTAL0.0370.00 16 Equation 7 In Vicinity- before metro0.3610.00 17In Vicinity- after metro0.3310.00 18Relocated- before metro0.2310.00 19Relocated- after metro0.2950.00 20TOTAL0.2340.00

61. 61 Interpretation of Results Equation 4: no significant correlation between the index of mobility and the indicators of A,  A does not affect M significantly. Equation 5: there is a significant correlation between the index of SEWB and the indicators of M,  M affects SEWB significantly. Equation 6: there is a significant correlation between the index of SEWB and the indicators of A for the households residing in the vicinity but the correlation is not significant for the households relocated Equation 7: there is a significant correlation between the index of SEWB and the combined indicators of A and M,  A and M affect SEWB significantly. Comparing the R 2 values of all the models, the best results are given by Equation 7, implying that the SEWB is explained best when the affects/contributions of indicators of both A and M are considered. However, it is observed that the R 2 values change for the households after the introduction of the metro. For the households located in the vicinity, the affects if A and M on SEWB become less significant after the metro and for the households relocated, they become more significant.

62. 62 Significance of Coefficients (Eqn 7) IndicatorDescription In Vicinity-b4 metro In Vicinity-aft metro Relocated-b4 metro Relocated-aft metro CoeffP(2Tail)CoeffP(2Tail)CoeffP(2Tail)CoeffP(2Tail) CONST435.20.006308.10.019318.20.013515.50 A1 SD education -81.30.041-43.80.123-2.60.812-10.70.736 A2 SD health -15.70.353-23.00.153-27.30.059-11.00.484 A3 SD services -69.90-17.60.477-1.10.958-4.60.238 A4 SD bus-stop 65.60.11830.90.037295.90.0885.30.704 A5 S bus -0.10.9291.00.0994.10.51-0.20.57 M1 PCTR work 102.5089.70126.40105.60 M2 PCTR education 45.30.15154.00.06853.50.344-1.40.966 M3 PCTR others 31.90.22445.80.05456.20.00431.00.042 M4 M NMV /M all 59.30.67525.00.831-37.90.746-280.30 M5 D work -4.70.013-2.70.063-1.80.4260.30.581 M6 D education 2.50.8144.00.704-16.20.3234.30.567 M7 D others -1.50.721-2.30.62-3.30.4546.40.005 M8 T work 0.00.9090.00.88-0.80.0380.00.844 M9 T education -0.60.29-0.60.2740.10.912-0.30.479 M10 T others -0.40.371-0.40.443-0.80.28-0.90.038 M11 C work -0.90.558-2.00.135-1.60.364-3.00.012 M12 C education 1.10.91.50.8617.70.631-9.10.485 M13 C others 1.20.5942.10.6193.30.384-6.00.045 Note: The indicator coefficients with P value significant at 90% confidence levels have been highlighted as the coefficients are significant can be included in the models.

63. 63 Interpretation of Results Comparative study of the coefficients shows that: Different coefficients contribute to the model significantly for different data sets. The number of significant coefficients increases after the introduction of the metro in the households both living in the vicinity and relocated due to the metro. The PCTR for work is the only indicator that is significantly consistent across the board. The cost of travel has no significance in explaining SEWB if relocation not there but it becomes significant when they are relocated. A study of the coefficients of the combined dataset to get an overview of whether the coefficients are +ve or –ve shows that approximately 90% of the significant indicators and 72%of all indicators are correlated to the SEWB index in accordance with the empirically observed behavior (expected indications)

64. 64 Final Equations The final equations derived from the application of Equation 7 using significant indicators are illustrated below: SEWB Vb4 = 435.2 - 81.3(SD education ) - 69.9(SD services ) + 102.5(PCTR work ) - 4.7(D work ) ………… 8-a SEWB Vaft = 308.1 + 30.9 SD bus-stop ) + 1.0(S bus ) + 89.7(PCTR work ) + 54.0(PCTR education ) + 45.8(PCTR others ) - 2.7(D work ) ………… 8-b SEWB Rb4 = 318.2 - 27.3(SD health ) - 295.9(SD bus-stop ) + 126.4(PCTR work ) + 56.2(PCTR others ) - 0.8(T work ) ………… 8-c SEWB Raft = 515.5 + 105.6(PCTR work ) + 31.0(PCTR others ) -280.3(M NMV /M all ) + 6.4 (D others ) - 0.9(T others ) - 3.0(C work ) - 6.0 (C others ) ………… 8-d

65. 65 Interpretation of results- The PCTR for work most important positive determinant of SEWB. This implies the trips to work made by a household ensure the SEWB, The distance to work is consistently a negative indicator for households implying that increase in distance to work will negatively affect SEWB. The introduction of the metro changes the indicators which affect SEWB. Also, more numbers of indicators have a significant impact on SEWB after the introduction of the metro. This implies that the introduction of a new transport system restructures the determinants of SEWB, making the households more vulnerable by increasing the number of significant indicators. HH in Vicinity: Since bus routes and services have been affected by the introduction of the metro, they become significant indicators affecting SEWB. This implies that the introduction of a new transport system makes the existing transport system important in determining SEWB.

66. 66 HH Relocated: Travel for purposes other than work and education is affected by the relocation. While the distance for these trips contributes positively to SEWB, the time and cost of these trips contributes negatively to it. The commuting cost had no significant correlation with SEWB before relocation, after relocation it has a significant negative impact on SEWB of the households. Ratio of NMV to all modes used has become a significant indicator after relocation. The high negative value of this indicator implies that the reduction in this ratio (implying reduction in use of NMV in the household) has a severe negative impact on the SEWB of the households. Since the process of relocation has increased distances to destinations of choice for the household, beyond comfortable NMV distances, this indicator implies that the modal shift from NMV to motorized modes has had a negative impact on the SEWB of the relocated households.

67. 67 8. Conclusions Impact of Metro on the poor household in its vicinity No significant impact on the SEWB and Mobility While the landuse accessibility remains unchanged too, the transport accessibility has changed as distance to the bus stops has increased for 19% of the households and bus services have become non-existent for 33% of the households. Impact of Metro on the poor households relocated There is significant impact on Accessibility, Mobility and SEWB The land-use accessibility has deteriorated as distance to education, health services and other urban services has increased for 52%, 63% and 52% of the households respectively. The transport accessibility has deteriorated even more as distance to bus stop has increased for 72% of the households and the bus frequency has seen an average decrease from 5 min to 63 min (almost 13 times)

68. 68 The mobility of the households have increased significantly. The PCTR for work has increased for 49% of the households and decreased for 30%, implying change in number of trips made for work in the households. The share of NMVs amongst the mode used has decreased for 59% of the households. The mobility indicators for travel to work – distance, time and cost – have increased for 83%, 82% and 61% of the households respectively The SEWB indicators most affected are female literacy (21% decrease), residency (100% decrease), Household income per person (66% decrease), Infrastructure rank score (33% decrease and 61% increase), and employment (8% decrease and 14% increase). The indicators of adult literacy and vehicle ownership show least change with 82% and 94% respectively in the no change category. The results imply that relocation due the metro has had a significant negative impact on the SEWB of the poor households.

69. 69 Correlation of SEWB to Accessibility and Mobility SEWB is affected by indicators of both accessibility and mobility  SEWB is negatively correlated to spatial distance to education, health and other urban services  It is positively correlated to PCTR for work, education and other purposes  It is negatively correlated to travel distance, time and cost The significance of indicators changes with change in situation like the new metro line or relocation due to it  PCTR for work is positively correlated with SEWB and has the highest coefficient in all datasets, indicating the mobility for work is important in ensuring their SEWB, whatever be their situation  Cost of travel has no significance in explaining SEWB of the urban poor but it becomes significant when they are relocated and now have to pay heavily for the travel

70. 70 In conclusion… This study illustrates that the accessibility and mobility and hence the socio-economic well-being of the urban poor is affected by its introduction in the urban transport system. While they may not be expected beneficiaries of the project, the dis- benefits accrued to them due to the project need to be assessed, and hence mitigation measures planned when proposing the project. Hence, it is important to conduct Socio Economic Impact Assessment (SEIA) studies for a new project over disaggregated groups, specifically including impacts on the most vulnerable group – the urban poor.

71. 71 Policy recommendations The definition of the impacted population for a transport project should include not only the expected users but the non-users affected by it too. The accessibility and mobility needs of the urban poor need to be studied and the urban poor should be seen as captives of the systems they are using. Introduction of any policy or project that changes their status has to be carefully monitored for impacts. The cost-benefit analysis of a transport project should include the dis-benefits to non-user groups and the costs of compensation/mitigation measures inbuilt as part of project cost. Only then should a project be declared feasible. The Government should constitute a statutory body responsible for the SEIA of all infrastructure projects before they are given approval for implementation. This is in keeping with the social welfare function of the Government. All funding mechanisms for transport projects should have inbuilt monitoring and evaluation protocols with stringent SEIA guidelines.

72. 72 Contribution of research This dissertation tries to understand how the SEWB of the urban poor is impacted by large transport projects. The impact on the accessibility and mobility of the non-users of the new system is as important as the impact on the expected users and needs to be internalized by transport projects. The dissertation proves that the relocation of the poor is one of the most severe negative impacts of a transport projects and needs to be taken in account in impact assessment studies. The dissertation has redefined the concept of mobility into its positive and negative aspects. It has formulated indicators of accessibility, mobility and SEWB and aggregated them into indices. It has modeled how SEWB is affected by accessibility and mobility and, in doing so, has formulated a generic methodology of SEIA which is applicable in understanding the impact of large urban transport projects like expressways, flyovers etc on the urban poor.. Different intervention scenarios can be compared for their impacts and mitigation measures planned accordingly. This would lead to internalizing the external cost of the impact of transport projects on the urban poor.

73. 73 Scope for future work Literature review has shown that even amongst the urban poor, the women are poorer that the men, suffering from poverty of money, time and resources. Assessing the gendered impacts of transport projects would give additional depth to the process of SEIA. The WHO has declared road accidents as the number one disease in the world. The health impacts of transport need to be included more comprehensively in the SEIA method after a necessary review of the literature on the same. The qualitative data about socio-economic conditions and the opinions and choices of people are another aspect of SEIA which requires further research. Different techniques like stated preference models can be used to include qualitative data. The benchmarking of the various parameters/indicators needs to be carried out to identify acceptable level of adverse impacts of transport projects. The impacts on accessibility, mobility and SEWB need to ascribed value in terms of money and resources to formulate compensation packages where necessary. This study should further lead to mitigation measures and alternative recommendations to minimize adverse impacts of transport projects on the urban poor.

74. November 200774 THANK YOU