Integrating travelers’ perceived factors in modeling the ceiling price of toll road a case study of bot projects in vietnam

VIETNAM NATIONAL UNIVERSITY, HANOI VIETNAM JAPAN UNIVERSITY DO VIET HUNG INTEGRATING TRAVELERS’ PERCEIVED FACTORS IN MODELLING THE CEILING PRICE OF TOLL ROAD A CASE STUDY OF B.O.T PROJECTS IN VIET NAM MASTER’S THESIS Hanoi, 2019 VIETNAM NATIONAL UNIVERSITY, HANOI VIETNAM JAPAN UNIVERSITY DO VIET HUNG INTEGRATING TRAVELERS’ PERCEIVED FACTORS IN MODELLING THE CEILING PRICE OF TOLL ROAD A CASE STUDY OF B.O.T PROJECTS IN VIET NAM MAJOR: INFRASTRUCTURE ENGINEERING CODE: PILOT RESEARCH SUPERVISOR: ANNEX LIST OF FORMS FOR MANAGEMENT Dr NGUYEN HOANG TUNG Hanoi, 2019 ACKNOWLEDGMENT During the period of studying to accomplish the topic of master’s thesis, there is not only self effort but also the enthusiastic guidance of teachers, as well as the encouragement of my family and friends First of all, I would like to express my gratitude sincerely to my supervisor, Lecturer Dr Nguyễn Hoàng Tùng who instructed directly me to conduct this research by making particular guidance I also would like to give special thanks to whole teachers at Vietnam Japan University, who have conveyed me the precious knowledge for the duration of my learning here Especially Professor Dr.Sci Nguyễn Đình Đức and Dr Phan Lê Bình who always have stimulated me to try my best and have inspired me a lot of motivations for studying as well as working in the life Besides that I am grateful to Professor Hironori Kato, the Co-Director of Master's Program in Infrastructure Engineering, for supporting me actively when I had a threemonth Internship Program in Japan Finally, I would like to thank my family and co-workers, who always are beside me during the implementing this master's thesis Although there have been lots of my attempts in the research process, due to limited ability and experience of self, this master's thesis still exist some unavoidable shortcomings So I am looking forward to hear sincere feedbacks from teachers and colleagues in order to supplement and complete more this research in the future Hà Nội, Jun 2019 Đỗ Việt Hùng i ABSTRACT The Build-Operate-Transfer (BOT) scheme is considered as an attractive means by the Vietnamese government to develop new infrastructure projects They were started to widely carried out from 2011 and it is predicted to continuously increase in the future The toll price of BOT road given by investor is set under the ceiling price that is regulated in Circular of Ministry of Transport It can be seen that most of BOT toll road projects have applied charge that are nearly equal to the given ceiling prices Nevertheless, the ceiling price at this Circular did not consider the scale of road, as well as did not take into account of travelers’ willingness to pay to toll price There has not ever researches of the topic to previous studies Therefore, this study focused on integrating “willingness to pay of travelers” to model of toll ceiling price in BOT road project by using a bi-level programming model to formulate for a simple two-road network to optimize toll price and propose ceiling pricing for BOT projects Some important findings of ceiling price include: (1) If government does not restrain toll price on BOT road, that leads to a negative social welfare; and (2) there must exist price which is maximum threshold of toll price for BOT road in order that social welfare becomes positive; and (3) ceiling price is set to be equal to minimum of “willingness to pay” and the threshold of toll price above ii TABLE OF CONTENTS ACKNOWLEDGMENT i ABSTRACT ii TABLE OF CONTENTS iii LIST OF FIGURES v LIST OF TABLES vi LIST OF ABBREVIATIONS vii CHAPTER INTRODUCTION CHAPTER LITERATURE REVIEW CHAPTER METHODOLOGY 3.1 The model setting 3.2 Solution procedure 11 3.2.1 Step 1: Determining route choice of travelers 11 3.2.2 Step 2: The operators determine the toll prices to maximize their own benefit respectively 13 3.2.2.1 BOT road 13 3.2.2.2 Existing road 14 3.2.3 Step 3: Evaluation social welfare 15 3.2.3.1 Case 1: All two roads get maximum their own benefit at the same time 15 3.2.3.2 Case 2: Only BOT road get maximum his own benefit under condition that it is positive social welfare 16 3.2.4 Step 4: Integrating the factor named "willingness to pay of travelers" to guarantee users' benefit 17 3.2.5 Step 5: Determining ceiling price of toll road under condition that optimal benefit of investor and the government, which taking into account "willingness to pay" of travelers 18 3.2.5.1 Case 1: If p1’ is less than WTP (i.e p1’ ≤ WTP) 19 3.2.5.2 Case 2: If p1’ is greater than WTP (i.e p1’ ≥ WTP) 19 iii 3.2.5.2.1 Recommendation 1: Downscale capacity of BOT road 19 3.2.5.2.2 Recommendation 2: Keep capacity of BOT road 20 CHAPTER CASE STUDY 22 4.1 Introduction for case study 22 4.2 Willingness to pay of travelers choosing this road to travel 23 4.3 Determining toll ceiling price of BOT road 25 4.3.1 Case 1: Toll ceiling price for vehicle has under 12 seats 25 4.3.1.1.1 Recommendation 1: Downscale capacity of BOT road 26 4.3.1.1.2 Recommendation 2: Keep capacity of BOT road 27 4.3.2 Case 2: Toll ceiling price for vehicle has under 30 seats 29 4.3.2.1.1 Recommendation 1: Downscale capacity of BOT road 30 4.3.2.1.2 Recommendation 2: Keep capacity of BOT road 31 4.4 Estimation WTP as regression function 32 4.4.1 Case 1: Estimation WTP function of travelers have car under 12 seats 33 4.4.2 Case 2: Estimation WTP function of travelers have vehicle under 30 seats 35 CHAPTER CONCLUSION 39 5.1 Conclusion 39 5.2 Limitation 39 REFERENCES 40 APPENDIX 41 iv LIST OF FIGURES Page Fig 1.1 Relationship between government, investor, travelers Fig 1.2 Number of BOT projects in Vietnam incl estimation Fig 1.3 Rank of Vietnam in the world about service and quality of road transport infrastructure Fig 1.4 Comparison of benefit between BOT road vs non-BOT road Fig 1.5 Toll price in BOT scheme in Vietnam Fig 1.6 GDP per capita and Toll price of some countries Fig 3.1 A network model 10 Fig 4.1 PV-CG Expressway and NH.1A 22 v LIST OF TABLES Page Table 1.1 Ceiling Prices in National Highway Table 1.2 Ceiling Prices in Expressway Table 4.1 Major parameters of two roads 23 Table 4.2 Characteristic of interviewees 24 Table 4.3 Descriptive statistic about WTP of travelers for toll price 25 Table 4.4 Regression Coefficient of WTP function 33 Table 4.5 Regression Coefficient of WTP function 35 Table 4.6 Regression Coefficient of WTP function 36 Table 4.7 Regression Coefficient of WTP function 37 vi LIST OF ABBREVIATIONS BOT Build – Operate - Transfer PPPs Public Private Partnerships PV-CG Expressway Pháp Vân – Cầu Giẽ Expressway NH.1A National Highway No.1A WTP Willingness to pay vii CHAPTER INTRODUCTION The more rapidly the economy grows, the more greater demand for infrastructure are In most countries, infrastructure was built from the state budget However, there are many items which need be spent by governmental budget and private capital is a good fund to complement these shortages Nowadays, more and more governments have encouraged private investor which take part in public investment projects Therefore, in order to support public infrastructure, Public Private Partnerships (PPPs) have become a major scheme (Hodge and Greve, 2007) and reduce public sector budget shortages (Kwak et al., 2009) PPPs are widely used to supply many infrastructure projects in the world Infrastructures invested by PPPs make economic efficiency increasing (Zhang, 2005) and facilitates the overall development of social infrastructure (Li et al., 2016b) In Vietnam, according to Decree No 63/2018/ND-CP: “PPP” is an investment form which is carried out on the basis of a contract between the State and an investor, in order to construction, renovation, operation, business, management of infrastructure works, provision of public services Build-Operate-Transfer (BOT) model is a form of PPP which has extensive applications in infrastructure projects The BOT scheme is gaining popularity and booming in public infrastructure around the world (Tan, 2012) It is adopted as an innovative way to sponsor for infrastructure construction in both developing and developed countries (Subprasom, 2004) In recent years, BOT arrangements have contributed to accelerating economic growth and improve quality service delivery and operation efficiency (Akintola et al., 2003) In Vietnam, the Build-Operate-Transfer (BOT) scheme is considered as an attractive means by the Vietnamese government to develop new infrastructure projects According to Decree No.63/2018/ND-CP, parties involving in this contract project include: the government, private investor and travelers Q = -2,4313 P + 147 751 The function of the flow volume that is responses of users with another toll price was determined in a series of bid questions, and toll price has equation: Q = -4,7888 P + 234914 (In detail, see at Appendix) The equilibrium point is determined by Eq.(25a) and (25b): Equilibrium price: pe = 36 973 (VND/turn/car) Equilibrium volume: Qe = 57 857 (PCU/24h) So, the ceiling price of toll road is set to be equal to pe can be expressed as follows: Ceiling price = pe = 36 973 (VND/turn/car) We conducted to examine the validation of Qe by comparing it with practical value, Qr is determined from direct questions about "willingness to pay" of toll price Percentage of flow volume who accepted equilibrium price was determined as a formula, is that: ΔQe = Qe 57857 100% = 100% = 39,63% Q1 146000 Meanwhile, percentage of flow volume who accepted equilibrium price was determined by the fact, is that: ΔQr = 40,71% (detail of calculation, see in Appendix) Error of value calculated by formula and by the fact is: Δ= |ΔQe -ΔQ | r ΔQe 100% = 40,71-39,63 100% = 2,72% 39,63 That means, the formula of toll ceiling price is reasonable 28 4.3.2 Case 2: Toll ceiling price for vehicle has under 30 seats By the standard namely TCVN 4054-2005, Highway – specifications for design, the vehicle has under 30 seats is equivalent to times PCU Likewise, by the Chapter 4, first order needs to determine is the price p1’, that is the maximum value threshold of toll price for BOT road so that social welfare become positive: p ≤ = + ( +2 ) + +2 (20) Where: y1 is capacity of BOT road, y1 = 158 400 (PCU/24h) = 79 200 (vehicle/24h); y2 is capacity of NH.1A, y2 = 52 800 (PCU/24h) = 26 400 (vehicle/24h), Q0 is total flow volume, Q0 = 197 000 (PCU/24h) = 98 500 (vehicle/24h) M is cost of maintenance and operation, M = 194 (millions/24h), t0 is free flow travel time, t0 = 0.25 (hour) λ is parameter that transfer time to fare, λ = WTP / Δt = 37 450 / 0.5 = 74 900 (VND /hour/vehicle) Substitute value of parameters above into Eq.(20), we obtain: p1’ = 56 423 (VNĐ/turn/vehicle) Secondly the WTP of travelers is determined in table 5.3, we have: WTP = 37 450 (VNĐ/turn/vehicle) Comparison values of p1’ with WTP, it can be seen that WTP < p1’, since there are two recommendation as mentioned in Chapter 29 4.3.2.1.1 Recommendation 1: Downscale capacity of BOT road The toll ceiling price is set to be equal to WTP: Ceiling price = WTP = 37 450 (VND/turn/vehicle) So, the flow volume of BOT road can be calculated as follows: Q11 = y1 y2 Q0 + (p2 − WTP) y1 +y y2 λ.t0 (24) Where: y1 is capacity of BOT road, y1 = 79 200 (vehicle/24h); y2 is capacity of NH.1A, y2 = 26 400 (vehicle/24h), Q0 is total flow volume, Q0 = 98 500 (vehicle/24h) M is cost of maintenance and operation, M = 194 (millions/24h), t0 is free flow travel time, t0 = 0.25 (hour) λ is parameter that transfer time to fare, λ = 74 900 (VND /hour/vehicle) p2 is toll price of NH.1a, p2 = Substitute value of parameters above into Eq.(24), we obtain: Q11 = 34 275 (vehicle/24h) Corresponding to number of lanes is that, n11 = Q11 n y1 Where: n1 is number of lanes of BOT road, n1 = (lanes) Thus, n11 = (lanes) 30 4.3.2.1.2 Recommendation 2: Keep capacity of BOT road When the investor still keeps the capacity of BOT road, it is necessary to trade-off benefit of travelers and investor So, there must exist an equilibrium value belonging to range from WTP to p1’ By the Eq.(6a), we have the function of flow volume and toll price, can be expressed: Q = -1,0575 P + 73 880 The function of the flow volume that is responses of users with another toll price was determined in a series of bid questions, and toll price has equation: Q = -2,19 P + 121667 (In detail, see at Appendix) The equilibrium point is determined by Eq.(25a) and (25b): Equilibrium price: pe = 42 197 (VND/turn/vehicle) Equilibrium volume: Qe = 29 255 (VND/turn/vehicle) So, the ceiling price of toll road is set to be equal to pe can be expressed as follows: Ceiling price = pe = 42 197 (VND/turn/vehicle) We conducted to examine the validation of Qe by comparing it with practical value, Qr is determined from direct questions about "willingness to pay" of toll price Percentage of flow volume who accepted equilibrium price was determined as a formula, is that: ΔQe = Qe 29 255 100% = 100% = 40,07% Q1 (146 000/2) Meanwhile, percentage of flow volume who accepted equilibrium price was determined by the fact, is that: ΔQr = 43,39% (detail of calculation, see in Appendix) 31 Error of value calculated by formula and by the fact is: Δ= |ΔQe -ΔQ | r ΔQe 100% = 43,39-40,07 100% = 8,29% 40,07 That means, the formula of toll ceiling price is reasonable 4.4 Estimation WTP as regression function Using the Regression tool in Excel software in order to analyze independent variables affecting WTP of travelers for toll price, including: Age, Gender, Education level, Marital status, Career Status, Income, Travel Frequency, Travel time acceptance level, Comfortableness level, Safe feeling level, and Charge pricing acceptance level The regression function is described as follows: WTP = C + a1.Age + a2.Gen + a3.Edu + a4.Mar + a5.Ca + a6.Inc + a7.Freq + a8.Time + a9.Comf + a10.Safe + a11.Toll Where: C : Intercept coefficient of regression model; Age : Age of interviewee; Gen : Gender of interviewee; Edu : Education level of interviewee; Mar : Marital status of interviewee; Ca : Job status of interviewee; Inc : Income of interviewee; Freq : Travel frequency on BOT road of interviewee; Time : Travel time acceptance level when travelling on BOT road of interviewee; Comf : Comfortableness level when travelling on BOT road of interviewee; 32 Safe : Safe feeling level when travelling on BOT road of interviewee; Toll : Charge pricing acceptance level when travelling on BOT road of interviewee 4.4.1 Case 1: Estimation WTP function of travelers have car under 12 seats First, we consider objects who has car under 12 seats So, outcome of running regression model are showed in table as below: Table 4.4 Regression Coefficient of WTP function of travelers who have car under 12 seats Variable Intercept Symbol Coefficients P-value - 16.280 6.544E-05 Gender Gen 0.963 0.4533 Education level Edu 0.642 0.2348 Marital status Mar -1.266 0.2614 Career status Ca 0.267 0.6261 Age Age 0.006 0.9067 Income Inc 0.200 7.007E-08 Travel Frequency Freq -0.030 0.0039 Travel time acceptance level Time -0.542 0.6220 Comfortableness level Comf 0.709 0.4641 Safe feeling level Safe -0.574 0.2493 Charge pricing acceptance level Toll 3.089 2.972E-23 33 So, regression function is described as below: WTP = 16,28 + 0,96.Gen + 0,642.Edu – 1,266.Mar + 0,267.Ca + 0,006.Age + 0,2.Inc - 0,03.Freq – 0,542.Time + 0,709.Comf – 0,574.Safe + 3,089.Toll The analytical results show that the multiple correlation coefficient (Multiple R) is approximately 0,746 at the same time, F is equal to 21,45 with Significance F is equal 7,74E-28, much smaller than 0.05 That explains that the selected linear regression model is very suitable R-Square = 0,557 means that the independent variables in the model have explained about 55,7% of the variability of Y (Willingness to pay) The remaining 44,3% is due to random factors and other factors not included in the model It can be seen that variables including: Gender, Education level, Marital status, Career status, Age, Travel time acceptance level, Comfortableness level, Safe feeling level, have P-value is greater than 0,05; that means these variables are not significant statistic in confidence level of 95% Variables including: Income, Travel frequency, Charge pricing acceptance level, have P-value is less than 0,05; that means these variables are significant statistic in confidence level of 95% Remove those variables which are not significant statistic in confidence level of 95% out of regression model At this time, the regression model of WTP has also independent variables including: Income, Travel frequency, Charge pricing acceptance level The regression function is described as: WTP = C + a1.Inc + a2.Freq + a3.Toll So running again this model, and outcome is showed in table below And regression function is described as follows: WTP = 17,449 + 0,214 Inc - 0,028 Freq + 2,952.Toll 34 Table 4.5 Regression Coefficient of WTP function of travelers who have car under 12 seats Variable Symbol Coefficients P-value Intercept - 17.449 9.19E-35 Income Inc 0.214 5.611E-08 Travel Frequency Freq -0.028 0.0026 Charge pricing acceptance level Toll 2.952 1.75E-28 Look at the result of regression function, it can be seen that income and charge pricing acceptance level are proportional to WTP, and travel frequency is inversely proportional to WTP In detail, when Income increases by million VND per month, WTP of traveler will increase by 214 VND/turn/car When Travel Frequency increases by traveling turn/year, WTP of traveler will decrease by 28 VND/turn/car And when charge pricing acceptance level increase by level, WTP of traveler will increase by 952 VND/turn/car 4.4.2 Case 2: Estimation WTP function of travelers have vehicle under 30 seats Second, we consider objects who has car under 30 seats Seeing the table of Descriptive statistic above, it can be seen that variables including Gender, Education level, Marital status, Career status has statistic value of 100% Since, it is necessary to remove these variables out of the model At this time, the regression model of WTP has also independent variables including: Age, Income, Travel frequency, Travel time acceptance level, Comfortableness level, Safe feeling level and Charge pricing acceptance level Thus, regression function is described as follows: WTP = C + a1.Age + a2.Inc + a3.Freq + a4.Time + a5.Comf + a6.Safe + a7.Toll 35 So, outcome of running regression model are showed in table as below: Table 4.6 Regression Coefficient of WTP function of travelers who have car under 30 seats Variable Symbol Intercept Coefficients P-value - 14.417 3.23E-05 Age Age 0.078 0.092 Income Inc 0.316 5.47E-06 Travel Frequency Freq -0.002 0.038 Travel time acceptance level Time -1.674 0.048 Comfortableness level Comf 3.125 0.0002 Safe feeling level Safe -0.492 0.259 Charge pricing acceptance level Toll 1.943 1.58E-07 Thus, regression function is described as below: WTP = 14,417C + 0,078.Age + 0,316.Inc -0,002.Freq -1,674.Time + 0,3125.Comf 0,492.Safe + 1,943.Toll The analytical results show that the multiple correlation coefficient (Multiple R) is approximately 0,830 at the same time, F is equal to 29,16 with Significance F is equal 8,47E-21, much smaller than 0.05 That explains that the selected linear regression model is very suitable R-Square = 0,689 means that the independent variables in the model have explained about 68,9% of the variability of Y (Willingness to pay) The remaining 31,1% is due to random factors and other factors not included in the model 36 It can be seen that variables including: Age and Safe feeling level, have Pvalue is greater than 0,05; that means these variables are not significant statistic in confidence level of 95% Variables including: Income, Travel frequency, Travel time acceptance level, Comfortableness level and Charge pricing acceptance level, have P-value is less than 0,05; that means these variables are significant statistic in confidence level of 95% Remove those variables which are not significant statistic in confidence level of 95% out of regression model At this time, the regression model of WTP has also independent variables including: Income, Travel frequency, Travel time acceptance level, Comfortableness level and Charge pricing acceptance level The regression function is described as: WTP = C + a1.Inc + a2.Freq + a3.Time + a4.Comf + a5.Toll So running again this model, and outcome is showed in table Table 4.7 Regression Coefficient of WTP function of travelers who have car under 30 seats Variable Symbol Coefficients P-value Intercept - 17.928 9.18E-11 Income Inc 0.320 4.52E-06 Travel Frequency Freq -0.003 0.005 Travel time acceptance level Time -1.792 0.034 Comfortableness level Comf 2.928 0.0004 Toll 1.830 4.36E-07 Charge pricing acceptance level 37 And regression function is described as follows: WTP = 17,928 + 0,32.Inc -0,003.Freq -1,792.Time + 2,928.Comf + 1,83.Toll Look at the result of regression function, it can be seen that Income, Comfortableness level and Charge pricing acceptance level are proportional to WTP, and Travel frequency and Travel time acceptance level are inversely proportional to WTP In detail, when Income increases by million VND per month, WTP of traveler will increase by 320 VND/turn/car When Travel Frequency increases by traveling turn/year, WTP of traveler will decrease by VND/turn/car When Travel time acceptance level increases by level, WTP of traveler will decrease by 1792 VND/turn/car When Comfortableness increases by level, WTP of traveler will increase by 2982 VND/turn/car When Charge pricing acceptance level increase by level, WTP of traveler will increase by 1830 VND/turn/car 38 CHAPTER CONCLUSION 5.1 Conclusion This study solved the problem mentioned in Chapter When taking into account “Willingness to pay” to toll charge, benefit of travelers was determined Some important finding of ceiling price are summarized as follows: (1) If government does not restrain toll price on BOT road, that leads to a negative social welfare (2) There must exist the price (p1’) which is maximum threshold of toll price for BOT road in order that social welfare become positive (3) Ceiling price is set to be equal to minimum of WTP and the price (p1’) If p1’ is less than WTP, Ceiling price is set up to p1’ If not, there are some recommendation including: (3-1) Down to scale capacity of road in order to price p1’ decreases to be equal to WTP, and set up ceiling price to be equal to WTP (3-2) Keeping the capacity of road and find out equilibrium price (pe) that make flow volume calculated by formula and one calculated by travelers’ response are equal, and set up ceiling price to be equal to pe 5.2 Limitation This study was considering a network has only two road In fact, a road infrastructure system include lot of roads and they interact each other about economy and society Thus, in practical flow volume on each road can change by positive or negative trend Therefore, the flow volume can not apply this formula mentioned in this study In addition, the integrating travelers’ perceived factor to formula is weak 39 REFERENCES Akintola A et al, 2003 Public-Private Partnerships: Managing Risks and Opportunities Blackwell Publishing Company, Iowa, USA Carpintero S., Gomez-Iban˜ez J A., 2011 Mexico’s private toll road program reconsidered Doi:10.1016/j.tranpol.2011.05.005 Chen A., and Subprasom K., 2007 Analysis of regulation and policy of private toll roads in a build-operate-transfer scheme under demand uncertainty Transportation Research Part A 41 (2007) 537–558 Government of India report (GoI), 2009 Report of the Committee of Secretaries on Review of Toll Policy for National Highways (www.infrastructure.gov.in) Guo X., and Yang H,, 2009 Analysis of a Build–Operate–Transfer Scheme for Road Franchising International Journal of Sustainable Transportation, 3:312–338, 2009 Hodge, G.A and Greve, C., 2007 Public–private partnerships: an international performance review Public Adm Rev 67 (3), 545–558 Kwak, Y.H et al, 2009 Towards a comprehensive understanding of public private partnerships for infrastructure development Calif Manage Rev 51 (2), 51–78 Li, Y., Wang, X., Wang, Y., 2016b Using bargaining game theory for risk allocation of public-private partnership projects: insights from different alternating offer sequences of participants J Constr Eng Manag 143 (3), 04016102 Mills G., 1999 Welfare and profit divergence for a tolled link in a road network Journal of Transport Economics and Policy, 1995, 29: 137–146 Subprasom, K., 2004 Multi-Party and Multi-Objective Network Design Analysis for the build-operate-transfer Scheme A Dissertation for the Degree of Doctor of Philosophy in Civil Engineering, Utah State University 40 APPENDIX Percentage of Flow Volume 40% 34% 6% 0% 35,000 30,000 25,000 20,000 15,000 1% 50,000 5% 14% 45,000 11% 48,882 price p1' 35,000 Current toll 31% 40,000 35% 30% 25% 20% 15% 10% 5% 0% Toll price (VND/trip) Fig 0.1 Relationship between WTP and percentage acceptance travelers who has car under 12 seats Percentage of Flow Volume 40% 35% 56,423 price p1' 29% 25% 19% 10% 5% 0% 2% 40,000 35,000 30,000 25,000 20,000 0% 1% 0% 60,000 45,000 Current toll 15% 50,000 17% 45,000 20% 55,000 32% 30% Toll price (VND/trip) Fig 0.2 Relationship between WTP and percentage acceptance travelers who has car under 30 seats 41 Flow Volume (PCU) 80,000 74,460 70,000 y = -2.4313x + 147751 62,654 60,000 50,498 50,000 40,000 35,040 30,000 38,341 26,184 20,000 65,000 1,460 50,000 45,000 40,000 35,000 30,000 - 60,000 14,600 y = -4.7888x + 234914 55,000 10,000 Toll price (VND/trip) Fig 0.3 Flow volume function for travelers has car under 12 seats Flow volume is determined by result in a series of bid questions Flow Volume (vehicle) 30,000 26,292 25,000 23,360 21,005 20,000 y = -1.0575x + 73880 15,717 15,000 11,680 10,000 y = -2.19x + 121667 5,000 65,000 55,000 50,000 45,000 40,000 60,000 1,460 - Toll price (VND/trip) Fig 0.4 Flow volume function for travelers having vehicle under 30 seats Flow volume is determined by result in a series of bid questions 42 .. .VIETNAM NATIONAL UNIVERSITY, HANOI VIETNAM JAPAN UNIVERSITY DO VIET HUNG INTEGRATING TRAVELERS’ PERCEIVED FACTORS IN MODELLING THE CEILING PRICE OF TOLL ROAD A CASE STUDY OF B.O.T PROJECTS IN. .. 1.3 Rank of Vietnam in the world about service and quality of road transport infrastructure When choosing to travel on BOT roads, although length of road between BOT road and non BOT road are... the independent variables in the model have explained about 68,9% of the variability of Y (Willingness to pay) The remaining 31,1% is due to random factors and other factors not included in the