Ningxia Liupanshan Poverty Reduction Rural Road Development Project (RRP PRC ) ECONOMIC EVALUATION

Transcription

1 Ningxia Liupanshan Poverty Reduction Rural Road Development Project (RRP PRC ) A. Introduction ECONOMIC EVALUATION 1. The economic analysis covers both the Project as a whole and the individual components. For the Project as a whole, the analysis covers the overall economic rationale, the rationale for public intervention, the goals of the investment plan, and the general design of the plan and the overall economic benefit to be expected. Economic analysis was conducted in accordance with Asian Development Bank s (ADB) Guidelines for the Economic Analysis of Projects (1997) and Economic Analysis Retrospective 2005: Strengthening ADB s Aid Effectiveness. B. Project Cost 2. The project contains the upgrading of two types of roads: 7 rural trunk roads, consisting mostly of class IV township roads upgraded to Class III (exceptions are one road and an additional segment consisting of county roads upgraded to class II); and 21 rural feeder roads, consisting exclusively of village roads upgraded to class IV. 3. The rural trunk roads have a total length of kilometers (km) with almost 120,000 inhabitants residing in the catchment area. The upgrading of trunk roads makes up $ million 1, which is 86% of total road investment costs of the project. These roads connect to regional and national networks, allow for higher speeds and larger transport volumes. The benefits of trunk roads are mainly efficiency improvements on the network level. Benefits may be attributed to all transport users, of which 41% are poor. These benefits comprise reduced vehicle operating costs (VOC) and travel time. Safety benefits from reduced accidents have been left out due to a lack of data, although a significant improvement in road safety is foreseen. Road Name County Length (km) Table 1: Project Cost Trunk Road Upgrading Program Cost (M CNY) a No. of Total Beneficiaries No. of Poor Beneficiaries Road Type, Class Zhengqi Jiucai Sikouzi Haiyuan ,000 10,000 Township Road, Class III Shatang Gaodian Jingyuan ,000 1,000 Township Road, Class III Mengyuan Chunshucha Chengyang Yangping Pengyang ,000 12,700 Township Road, Class III Wangtuan Yuwang Tongxin b ,000 8,000 Township Road, Class II/III Jiangtai Xitan Pingfeng Xiji ,000 10,635 County Road, Class III Guating Guyuan Yuanzhou ,000 6,000 County/Township Road, Class II/III Wanzhang Sanying Yuanzhou ,000 5,000 Township Road, Class III Total , ,000 53,335 (44% poor) a Includes base cost, physical contingency (9%), and financial charges during implementation. b Length after realignment, original distance was 67.4 kilometers. 1 1 USD = CNY as of 15 June 2016.

2 2 4. The rural feeder roads have a total length of km and 104,000 inhabitants reside in the catchment area. The share of poor population is 58%. The upgrading of feeder roads makes up $34.07 million, which is 14% of total road investment costs of the project. Feeder roads are improved through paving of unsurfaced roads. These roads connect the villages to local markets and services and are currently characterized by very low speeds and small transport volumes. The benefits of these village roads are predominantly for the poor population, where upgrading brings major socio-economic benefits from improved and year-round accessibility to social services and markets. 5. Currently 8% of the feeder roads length is graveled and 43% are earth roads, the remaining 49% are sealed. These roads will be improved as follows: 49% will be newly constructed roads, 31% will be paved, 12% widened and 8% rehabilitated. Figure 1: Current Feeder Road Surface and Construction Type Road Name County Length (km) Table 2: Project Cost Feeder Road Upgrading Program Cost (million CNY) a No. of Total Beneficiaries No. of Poor Beneficiaries Road Type, Class Hongzhuang Dadian Yuanzhou ,000 5,600 Village Road, Class IV Shahexian Guhu Road Qiaowa Miaotai Yuanzhou ,000 3,600 Village Road, Class IV Ligou Xiaojiashengou Yuanzhou ,000 6,750 Village Road, Class IV Ke village Feng village Yuanzhou ,500 3,600 Village Road, Class IV Licha Dongjia village Yuanzhou , Village Road, Class IV Caichuan Yangjiayaoxian Yuanzhou , Village Road, Class IV Wangping Lizhang Xiji ,593 1,885 Village Road, Class IV Daying Yaerpo Xiji ,928 1,433 Village Road, Class IV Mawan Caonao Xiji ,006 1,230 Village Road, Class IV Bataijiao Quancha Xiji ,100 1,140 Village Road, Class IV Nanchuan Lujiagou Xiji ,300 1,104 Village Road, Class IV Tongfu village Shanghewan village Tongxin ,300 3,040 Village Road, Class IV

3 3 Road Name County Length (km) Cost (million CNY) a No. of Total Beneficiaries No. of Poor Beneficiaries Road Type, Class Majiajing Suocha Tongxin Village Road, Class IV Tongyu Road Lijiagangzi Tongxin ,000 8,900 Village Road, Class IV Liushubaozi Xiachen Road Chener Tongxin ,500 2,100 Village Road, Class IV Caomiaoxinwa Caochuan Pengyang ,700 5,700 Village Road, Class IV Xiaochadiaocha Qigeshan Pengyang ,800 2,800 Village Road, Class IV Xi'an Zhangwan Haiyuan ,000 2,000 Township Road, Class III Xiangtong Huitiaogou Haiyuan ,500 2,500 Village Road, Class IV Dongxia Digou Jingyuan ,000 1,500 Township Road, Class IV Zhangtian Jinglin Yangchuan Longde ,000 3,000 Township Road, Class III Total ,127 60,206 (58% poor) a Includes base cost, physical contingency (9%), and financial charges during implementation 6. The trunk and feeder road packages have been evaluated by comparing conditions over a 20-year evaluation period from the assumed year of opening for two cases: a reference case and a project case. The reference case represents future conditions in the absence of the project. Evaluations have been done by national consultants applying a Chinese model. To verify assumptions and reliability of the results re-appraisals have been done for all trunk and feeder roads. Hereto the Road Economic Decision Model (RED) has been used. The results from the FSR and RED-exercises appeared to be comparable. In this document the results from the re-evaluations are presented. Economic price conversion 7. Project costs and benefits are measured in 2016 economic prices expressed in domestic currency. Vehicle costs in the People s Republic of China (PRC) are close to international prices, the only adjustment made is to deduct taxes. 8. Initial investment costs were derived from financial cost estimated by design institutes by applying the following conversion factors and deductions: Adjustment for taxes and duties as follows: o Civil works: 11% o VAT on equipment: 17% o Land acquisition: 40% (average of taxes and duties on LAR costs of trunk roads) o Consultancy services: 6% Land acquisition costs represent the opportunity cost of lost agricultural production over the analysis period 2 Physical contingencies: o 9% for trunk roads o 5% for feeder roads Price contingencies and financial charges were excluded 2 In line with PRC directives the opportunity cost of lost production is estimated assuming 5% real annual growth in production (reflecting increasing productivity and conversion to crops of higher value). The stream of the projected value of lost production was discounted at 12%, adjusted by the standard conversion factor, and included as a firstyear implementation cost.

4 4 9. Next, the investment cost has been corrected (applying domestic price numeraire), based on the shadow exchange rate factor (SERF), the inverse of the standard conversion factor (SCF). Box 1 Standard Conversion Factor and Shadow Wage Rate Factor The standard conversion factor (SCF) can be calculated as follows: SCF = (M cif +X fob )/(M cif +T M -S M +X fob -T x +S x ) Where: M is imports at CIF prices, X is export at FOB prices, T is taxes and S is subsidies. For the PRC these values are: 3 X = $2.342 trillion (2014) M = $1.958 trillion (2014) T M = 4.74% (weighted average) T X = 0 The export subsidies program of the PRC has been suspended. 4 Export duties are gradually abolished too. 5 Import subsidies are assumed to be negligible. The above figures result in a SCF of (or a SERF of 1.022). To convert the wage component for unskilled labor, the shadow wage rate factor (SWRF) has been calculated. The daily minimum wage in the PRC ranges from CNY30 to CNY60. 6 Assuming a wage for unskilled labor in Ningxia province of CNY45, the average monthly wage is around CNY1,000. Income below CNY1,500 is taxed at 3%, 7 consequently the net wage factor for unskilled labor is Further, unemployment in Ningxia province was 4% in To calculate the shadow wage the following formula is adopted: SWRF = W*(1-t)*(1-u) Where: W= market wage, t is the income tax rate and u is the unemployment rate. Based on the above the SFRF applied to the unskilled labor component is 0.97 * 0.96 = For skilled labor the income tax component (10%) has been deducted. These conversion rates have been applied to the investment cost as presented in table 1 to obtain the economic value. This results in a (negligible) composite conversion factor of (see table 3). CIF = Cost insurance and freight, FOB = Free on board Export duties are only imposed on a few resource products and semi-manufactured goods. In 2013, China levied temporary tariffs on exports including coal, crude oil, chemical fertilizers, rare metals and iron alloy to conserve resources. Overall, the PRC is gradually abolishing export duties (

5 5 Construction Item Table 3: Economic Price Conversion (domestic price numeraire) Share within category (%) Conversion factor Share (%) Weighted Value Materials 45% Tradable materials Non-tradable materials Labor 20% Skilled labor Unskilled labor Equipment 35% Tradable Equipment Non-tradable Equipment Composite Conversion Factor Investment costs have been spread over 3 years in the proportions as shown in table 4. It presents the economic implementation costs, 9 inclusive of standard conversion factors, for the project components by year, exclusive of maintenance costs. The residual value of upgrade works was estimated at 20% of investment cost at the end of the evaluation period. 9 Excluding cost related to institutional development

6 6 Road Name Table 4: Economic Implementation Costs (CNY Million) Civil Works Equipment LAR Consultancy Services Taxes & duties Physical Contingencies Economic Cost Implementation Year Sub-total Trunk Roads 1, , % 44% 11% Zhengqi Jiucai Sikouzi % 40% Shatang (Huanghua County) Gaodian % Mengyuan Chunshucha Chengyang Yangping % 40% Wangtuan Yuwang % 50% 20% Jiangtai Xitan Pingfeng % 50% 20% Guating Guyuan % 40% Wanzhang Sanying % 40% Sub-total Feeder Roads % 0% 0% Hongzhuang Dadian Shahexian % Guhu Qiaowa Miaotai % Ligou Xiaojiashengou % Ke village Feng village % Licha Dongjia village % Caichuan Yangjiayaoxian % Wangping Lizhang % Daying Yaerpo % Mawan Caonao % Bataijiao Quancha % Nanchuan Lujiagou % Tongfu village Shanghewan village % Majiajing Suocha % Tongyu Lijiagangzi Liushubaozi % Xiachen Chener village % Caomiaoxinwa Caochuan % Xiaochadiaocha Qigeshan % Xi'an Zhangwan % Xiangtong Huitiaogou % Dongxia Digou % Zhangtian Jinglin Yangchuan % Total 1, , % 38% 9% Source: Asian Development Bank estimates.

7 7 11. Maintenance cost consists of routine and periodic maintenance. The annual operation and maintenance economic cost (routine) for all roads is assumed at CNY2,000/km with the project. Maintenance cost savings have been estimated for each forecast year by comparing the total maintenance expenses per kilometer in the with-project and without-project situations. Routine maintenance procedures such as drain clearance, grass cutting, minor shoulder and pavement repairs, and traffic sign repairs would be similar for all types of roads. Table 5 shows the unit costs of maintenance and operation that have been applied in the evaluation to produce estimates of annual maintenance costs. Type of maintenance Routine maintenance Periodic maintenance Type of road All roads Table 5: Road Maintenance Parameters Works Frequency Unit cost Drain clearance, grass cutting, minor shoulder, pavement, and traffic sign repairs, etc. Yearly CNY 2,000/km Trunk Resurfacing Every 10 years CNY 105,000/km Feeder Resurfacing Every 6 years CNY 42,000/km C. Traffic projections 12. The estimation of existing traffic volumes in AADT was conducted for each road by local design institutes based on historic traffic counts, and traffic analysis using connected roads when traffic data was not directly available for project roads. As no traffic counts could be conducted for the analysis, special attention has been given to the sensitivity analysis to lower traffic volumes. Table 6 shows the traffic estimation for all project roads per Table 7 shows the distribution of traffic per category derived from the FSR. Table 6: AADT on Project Roads (2017) Trunk Roads AADT (2017) Zhengqi Jiucai Sikouzi 1,806 Shatang (Huanghua County) Gaodian 1,175 Mengyuan Chunshucha Chengyang Yangping 1,046 Wangtuan Yuwang 3,611 Jiangtai Xitan Pingfeng 2,426 Guating Guyuan 4,070 Wanzhang Sanying 2,145 Feeder Roads Hongzhuang Dadian Shahexian 297 Guhu Road Qiaowa Miaotai 287 Ligou Xiaojiashengou 304 Ke village Feng village 145 Licha Dongjia village 202 Caichuan Yangjiayaoxian 147 Wangping Lizhang 151 Daying Yaerpo 153 Mawan Caonao 154

8 8 Feeder Roads Bataijiao Quancha 148 Nanchuan Lujiagou 150 Tongfu village Shanghewan village 350 Majiajing Suocha 350 Tongyu Road Lijiagangzi Liushubaozi 350 Xiachen Road Chener village 350 Caomiaoxinwa Caochuan 246 Xiaochadiaocha Qigeshan 246 Xi'an Zhangwan 2,165 Xiangtong Huitiaogou 244 Dongxia Digou 304 Zhangtian Jinglin Yangchuan 2,431 Sources: Local design institutes and Asian Development Bank estimates. Table 7: Distribution of Traffic by Road Type Trunk Roads Feeder Road Car Medium 35% 20% Bus Medium 20% 10% Truck Light 25% 20% Truck Medium 10% 0% Motorcycle/Tricycle 10% 50% Sources: Local design institutes and Asian Development Bank estimates. 13. Normal traffic growth estimates on trunk roads have been based on an analysis of economic growth projections from the FSR and ADB estimates. 14. An elasticity factor of 0.8 has been applied to the county annual growth rates to define the traffic growth rates (Table 8 and Table 9). Trunk Roads Table 8: County Economic Growth Projections County Annual County GDP growth rate Zhengqi Jiucai Sikouzi Haiyuan 15.1% 12.1% 9.4% 7.7% 6.3% Shatang (Huanghua County) Gaodian Jingyuan 13.4% 11.6% 9.4% 7.7% 7.5% Mengyuan Chunshucha Chengyang Yangping Pengyang 20.1% 11.8% 9.7% 7.9% 6.5% Wangtuan Yuwang Tongxin 15.6% 10.0% 7.8% 5.2% 3.7% Jiangtai Xitan Pingfeng Xiji 18.6% 8.9% 9.1% 7.4% 6.1% Guating Guyuan Yuanzhou 16.0% 11.9% 9.5% 7.7% 6.3% Wanzhang Sanying Yuanzhou 16.0% 11.9% 9.5% 7.7% 6.3% Source: Local design institutes.

9 9 Trunk Roads Table 9: Normal Traffic Growth Projections County Normal Traffic Growth Rates Zhengqi Jiucai Sikouzi Haiyuan 9.6% 7.5% 6.1% 5.0% Shatang (Huanghua County) Gaodian Jingyuan 9.3% 7.5% 6.1% 6.0% Mengyuan Chunshucha Chengyang Yangping Pengyang 9.5% 7.8% 6.3% 5.2% Wangtuan Yuwang Tongxin 8.0% 6.2% 4.1% 2.9% Jiangtai Xitan Pingfeng Xiji 7.1% 7.3% 6.0% 4.9% Guating Guyuan Yuanzhou 9.5% 7.6% 6.2% 5.0% Wanzhang Sanying Yuanzhou 9.5% 7.6% 6.2% 5.0% Source: Asian Development Bank estimates. 15. Given the lack of reliable information related to feeder roads, conservative traffic growth rates were estimated at 6% annual growth rate from 2017 to 2025 and 3% annual growth rate from 2026 to Generated traffic was estimated using a price elasticity of demand for transport of 0.5 for all categories of vehicles. The rule of half has been applied to calculate benefits from generated traffic. D. Project Benefits 17. The benefit streams of the roads upgrading comprise vehicle operating cost savings, travel time savings, and road safety benefits. Vehicle operating costs and project benefits have been calculated in the Roads Economic Decision Model (RED) by vehicle type on the basis of the economic costs (purchase, consumables, and maintenance) and typical vehicle characteristics. 18. Table 10 shows the five types of vehicle used in the economic evaluation, namely passenger cars, buses (medium), trucks (light), trucks (medium) and tricycles (the latter only on feeder roads). The different components of vehicle operating costs (VOC) by vehicle type in the table below are at constant mid 2016 financial prices and include valuations of driver time and maintenance labor using the average wage in Ningxia. Passenger time is excluded from the VOC. The components of VOC are valued at economic prices. The analysis is carried out using domestic prices. Cost Item Economic Unit Costs Table 10: Vehicle Characteristics (Trunk and Feeder Roads) Car Medium Bus Medium Truck Light Truck Medium Motorcycle / Tricycle New vehicle cost (CNY/vehicle) 101, , , ,950 6,482 Fuel cost (CNY/liter for MT, CNY/MJ for NMT) Lubricant cost (CNY/liter) New tire cost (CNY/tire) Maintenance labor cost (CNY/hour) Crew cost (CNY/hour) Kilometers driven per year (km) 12,000 25,000 40,000 40,000 8,000 Hours driven per year (hr) ,000 1,600 1, Service life (years)

10 10 Cost Item Car Medium Bus Medium Truck Light Truck Medium Motorcycle / Tricycle Percent of time for private use 40% 0% 0% 0% 50% Gross vehicle weight (tons) Source: FSR and Asian Development Bank estimates. 19. Passenger time savings have been calculated for work and leisure time. The value of leisure time was valued at 30% of work time. In the Social and Poverty Assessment, average GDP per capita in the project area was estimated at CNY12,778, i.e. CNY1,100 per month (2013 prices). It has therefore been assumed that car passengers earn on average CNY1,200 per month (110% of average revenue), while bus passengers earn CNY900 per month (75% of average revenue). Further, the average number of passengers per vehicle is estimated at 3 per car and 22 per bus, the percentage of business trips was set at 70% for car trips and 25% for bus trips. Value of time was therefore estimated as per Table 11 below. For goods in transit, no time saving has been considered. Vehicle Hourly wage (CNY) a Table 11: Car and bus passenger value of time % work time Value of working / non-working time Passenger time cost (CNY/pas-hr) No. of passengers Passenger time cost (CNY/veh-hr) Car % 100% Bus % 30% a Estimated on a base of 25 days worked 9 hours 20. Table 12 and 12 summarize road user costs and journey time savings for the trunk and feeder roads based on the assumptions of the existing network conditions, comparing the reference case with the project case. Upgrading the roads lead to improved driving conditions as expressed by a lower international roughness index. Due to the lack of reliable information, no accident and carbon emission cost savings were considered. Road Table 12: Road User VOC Savings VOC Base Scenario (CNY/veh.km) VOC Project Scenario (CNY/veh.km) IRI Car MGV IRI Car MGV Trunk Roads Feeder Roads Upgrade from Earth Road Upgrade from Gravel Road Upgrade from Paved Road IRI = international roughness index, MGV = medium goods vehicle, VOC = vehicle operating cost, VOT = value of time. Road Table 13: Vehicle Speed Function of Roughness VOC Base Scenario (km/hour) VOC Project Scenario (km/hour) IRI Car MGV IRI Car MGV Trunk Roads Feeder Roads Upgrade from Earth Road Upgrade from Gravel Road Upgrade from Paved Road IRI = international roughness index, MGV = medium goods vehicle, VOC = vehicle operating cost, VOT = value of time.

11 11 E. Results of the Economic Cost-Benefit Analysis 21. An economic analysis of each individual trunk and feeder road has been conducted over 20 years inclusive of project construction. The economic costs of the components include: (i) capital cost including land acquisition and resettlement, and environmental mitigation and monitoring costs at economic prices; and (ii) the costs of operation and maintenance (O&M) including the costs of replacement of depreciated equipment. Land acquisition and resettlement costs were estimated based on their opportunity costs, that is, the agricultural output foregone and the re-siting of displaced activities. Transfer payments, including taxes and interests, were excluded from the economic costs, while physical contingencies were included. 22. The economic viability of each component has been determined calculating the economic internal rate of return (EIRR) and the net present value (NPV) of the incremental cost and benefits. A project is considered viable if the calculated EIRR is higher than 12%, the social discount rate as prescribed by ADB. 10 Note that the projects generate significant unquantified socio-economic benefits (see section 0 below) and road safety benefits. Table 14 shows the evaluation results of the 7 trunk roads. Trunk Roads Table 14: Evaluation Results by Trunk Road (CNY million) PV Incremental Cost VOC Savings PV of Benefits Time Savings NPV (at 12%) EIRR (%) Zhengqi Jiucai Sikouzi % Shatang Gaodian (5.11) 10.2% Mengyuan Chunshucha Chengyang Yangping % Wangtuan Yuwang % Jiangtai Xitan Pingfeng (15.79) 11.3% Guating Guyuan % Wanzhang Sanying (1.10) 11.9% EIRR = economic internal rate of return, NPV = net present value, PV = present value, VOC = vehicle operating cost. Sources: Asian Development Bank estimates. 23. Table 14 shows the calculations for all trunk roads combined, resulting in an EIRR of 13.8%. Table 15: Economic Evaluation of All Trunk Roads (CNY million) Year Incremental Cost VOC Benefits VOT Benefits Net Benefits Based on ERCD presentation on updated guidelines for economic evaluation on 15 June 2016, Manila.

12 12 Year Incremental Cost VOC Benefits VOT Benefits Net Benefits NPV(at 12%) EIRR 13.80% 24. The evaluations of trunk roads show that the interventions are economically feasible with EIRR s ranging from 10.2% to 18.1%. The role of the feeder roads being to reach further into poor villages from the trunk network, the economic evaluation was also conducted at the county network level, including all trunk and feeder roads in each county (table 15), as well as for all trunk and feeder roads combined (Table 16). Table 16: Evaluation Results by County (Trunk and Feeder Road Combined) (CNY million) County PV Incremental Cost PV of Benefits VOC Savings Time Savings NPV (at 12%) EIRR (%) Haiyuan % Jingyuan (17.31) 7.8% Pengyang (2.42) 11.9% Tongxin % Xiji (45.39) 10.6% Yuanzhou % Longde (0.62) 11.5% Table 17: Economic Evaluation of All Roads (CNY million) Year Incremental Cost VOC Benefits VOT Benefits Net Benefits

13 13 Year Incremental Cost VOC Benefits VOT Benefits Net Benefits NPV(at 12%) EIRR 13.51% 25. At the county level, EIRR varies from 7.8% to 18.3%. The feeder roads generally show positive results since the upgrading from an unpaved condition is a significant improvement at relatively low intervention cost. Only Jingyuan county has an EIRR below the 12% threshold, since some feeder roads in this county show too little traffic to generate sufficient benefits from vehicle cost and time savings. However, upgrading these roads will have a significant socioeconomic impact on the communities linked to these roads. 26. The overall upgrading program of all trunk and feeder project roads shows an EIRR of 13.5%, which makes it significantly viable. Sensitivity Analysis 27. The project components were tested for responsiveness to changes in key variables applying (i) a 10% increase in cost, (ii) a 30% overestimation of traffic projections, (iii) a lack of O&M funding 11 and (iv) a 2 year implementation delay. Main risk related to climate change relates to more excessive precipitation that would impact the accessibility of certain roads and require additional funding for O&M. The theoretical scenario of climate change can therefore be modelled by the above sensitivity tests. Table 6 shows the results of the sensitivity analysis. Table 17 and 18 show the results for the trunk roads and counties. 11 Lack of funding for O&M was simulated by increasing VOC by 3% every year (cumulatively) between two major maintenance works and removing the residual value at the end of the study period.

14 14 Trunk Roads Roads Table 18: Sensitivity Analysis Road Program (trunk roads) (EIRR) +10% costs -30% traffic 2 year delay Lack of O&M Switching Value (to 12%) % of Cost % of Traffic Zhengqi Jiucai Sikouzi Road 12% 9% 13.9% 11.1% 107.3% 92.8% Shatang (Huanghua County) Gaodian Road 9% 6% 10.0% 8.2% 86.5% 115.2% Mengyuan Chunshucha Chengyang Yangping Road 11% 8% 13.2% 10.3% 101.0% 98.3% Wangtuan Yuwang Road 17% 13% 18.9% 16.2% 153.3% 64.8% Jiangtai Xitan Pingfeng Road 10% 7% 11.7% 9.4% 94.5% 106.1% Guating Guyuan Road 13% 10% 15.0% 12.0% 115.7% 86.0% Wanzhang Sanying Road 11% 8% 12.8% 10.0% 98.6% 100.7% All trunk Roads 13% 9% 14.6% 11.9% 115.0% 86.9% Trunk Roads Roads Table 19: Sensitivity Analysis Road Program (by county) (EIRR) +10% costs -30% traffic 2 year delay Lack of O&M Switching Value (to 12%) % of Cost % of Traffic Haiyuan 13% 10% 15.2% 12.6% 116.3% 86.0% Jingyuan 7% 4% 7.3% 6.1% 70.3% 143.0% Pengyang 11% 8% 12.7% 10.3% 99.0% 101.0% Tongxin 17% 13% 19.1% 16.5% 154.0% 64.2% Xiji 10% 7% 10.8% 8.9% 88.9% 112.4% Yuanzhou 12% 9% 13.9% 11.3% 108.1% 92.1% Longde 10% 7% 11.4% 11.2% 96.2% 103.7% All Project Roads 12% 9% 14.12% 11.73% 111.8% 88.8% 28. The economic viability of trunk roads and county road programs generally resists to an increase of project cost and a sharp decrease of benefits. To be noted that an implementation delay would increase the economic viability of the project due to the higher discount rate applied to capital cost. Only Shatang (Huanghua County) Gaodian Road (Jingyuan county) falls significantly under 12% in given scenarios. Risk Analysis 29. A Monte Carlo risk analysis has been conducted for the trunk and feeder roads program combined. Asymmetric triangular probabilities have been applied on total incremental cost (including maintenance cost) and total benefits by setting a minimum of 90% and maximum of 130% for the cost and a minimum of 70% and maximum of 110% for the benefits. Under these assumptions, the probability of a positive NPV for the project is 47.4%.

15 Table 19: Assumptions Triangular Probability Distributions Base case ENPV CNY million Variables Investment O&M Total benefits Base-case (present value) CNY million 1, ,321.9 Minimum % 90% 90% 70% Most likely % 100% 100% 100% Maximum % 130% 130% 110% Number of iterations # 10,000 Table 20: Results Monte Carlo Simulation Mean (expected ENPV) mcny Minimum mcny Median mcny Maximum mcny Standard deviation mcny Probability {ENPV>0} % 47.4% mcny = million yuan Figure 2: ENPV Probability Distribution

16 Figure 3: Cumulative Probability Distribution F. Socio-Economic Impacts 30. In order to describe possible impacts of rural roads a comprehensive literature research was undertaken, which is described in detail in chapter 4 of the Social and Poverty Assessment. In total 29 research documents 12 from 15 countries showed the following impacts of rural roads improvement: Impacts on agricultural production and marketing Transport induced local market development Impacts on wages, consumption and employment Impacts on poverty Impacts on transport speeds, costs and patterns Improved access to health services and education, and Impacts on gender 31. Most of these benefits have been already captured in the economic evaluation above through generated/induced traffic. However, the impact on wages, consumption and employment and access to health services and education have not been quantified, while representing a significant improvement, in real term value, of the affected population s wealth. 12 Asher 2015: India, Atsushi et al 2015: Brazil, Balisacan et al 2002: Philippines, Banerjee 2015: India, Chongvilaivan 2016: Timor-Leste, Cook 2005: Asia, Cook et al 2005: China, Dercon et al 2007: Ethiopia, Escobal 2003: Peru, Fan et al 1999: India, Fan et al 2000: China, Fan et al 2005: China, Ferf 2014: Kongo, Gertler et al 2014: Indonesia, Gibson et al 2002: Papua New Guinea, Glewwe et al 2002: Vietnam, Hettige 2006: Asia, Jalan et al 2002: China, South, Kandler/Bär 2004: Bangladesh, KfW 2013: Cambodia, Khandker 2006: Bangladesh, Kwon et al 2000: Indonesia, Mohapatra 2007: India, Qiaolun Ye 2006: China, Raballand 2009: Cameroon, Raychudhuri 2004: India, USAID 2006: Afghanistan, Van de Walle 2002: Vietnam, Van de Walle 2009: Vietnam, Warr et al 2006: Lao PDR.

17 For instance, the German Financial Cooperation with Cambodia (KfW 2013) conducted a major impact assessment study on their Rural Infrastructure Program (RIP) II in Bangladesh and observed an average increase of 197% on annual household income. A different study of road improvements in Bangladesh (Khandker, 2006) also showed a significant impact on men s agricultural wage (increases by 27%), fertilizer price (falls by about 5%) and aggregate crop indices (price indices increase by about 4%), while output indices rise significantly by about 30%). 33. The overall effect of road improvement on household per capita annual consumption was estimated at 11%. Given the size of the affected poor population and applying a linear increase of the average GDP per capita estimated in the Social and Poverty Assessment to reach 11% five years after the end of the construction works, this would represent a net present value of CNY 425 million over the project life in real value. 34. Fan et al. (2000) examined the factors which contributed to the exceptional growth and to the reduction of poverty in China during the past decades. Government spending on rural infrastructure (roads, electricity, and telecommunications) had substantial impact in reducing poverty and inequality as well, owing mainly to improved opportunities for nonfarm employment and increased rural wages. Among the three infrastructure variables considered, the impact of roads is particularly large. They conclude that that with every CNY 10,000 (about $1,200) spent on rural roads, eleven persons are lifted above the poverty line. In terms of impact on growth, for every yuan invested in roads, CNY 8.83 in rural GDP is produced. Roads yielded the largest return to rural nonfarm GDP, at CNY 6.71 for every yuan invested, 35 percent higher than the return to education investment. 35. Another frequent observation is the shift from agricultural self-employment to wageearning employment. Asher et al (2015) compiled large datasets from India s rural road construction program that has built paved roads to over 100,000 previously unconnected villages since it began in The authors find that rural roads increase economic well-being, as measured both by household earnings [ ] Road construction to previously unconnected villages lead to a 10 percentage point reduction in the share of households and workers in agriculture, with an equivalent increase in wage labor market participation. Other studies confirm the above findings on income poverty and employment. 36. Even though the observed impacts were very much dependent on the local economic and geographic conditions, a large consensus amongst the researchers may be observed regarding the positive effects of rural roads on income and poverty as well as on accessibility to social service and places of employment. Generally, there were two major impacts observed: Many studies confirmed that rural roads induce a market led local development which entails strong impacts on agricultural marketing and increasing incomes through farming. Other studies revealed that rural roads increased the revenues from non-farming activities. This implies a shift from subsistence agricultural to commercial agriculture or manufacturing, which is achieved though commuting or even resettlement to locations of employment. 37. Consequently, a number of authors argue that rural roads should be improved to the locations where poverty was most severe and through improved access to markets provide opportunities for subsistence farmers to integrate into the market economy and thus increase farm production, marketing and agricultural incomes.

18 Other authors argue that rural roads should be primarily improved to locations where economic opportunities are best and thus induce a dynamic process for commercial farming and manufacturing, which again creates places of employment. Only one ex-post study 13 on rural road impacts was conducted in the PRC which is representing the latter approach on economic development through job creation. 39. Widely it was perceived that rural roads have considerable positive impacts on poverty through the above effects. However, it has to be emphasized that roads are essential for development, but not sufficient to generate a process on its own. The ability of the poor and very poor to make significant economic use of a road depends on their asset base and the entitlements to resources and opportunities that they can command, as well as on the passage of time. Thus, it is not imperative that the very poor benefit from the road improvements. This problem may be overcome by selecting roads to villages with a large share of poor inhabitants. 40. Many authors claim that roads at the lower end of the network, such as upgrading of paths, tracks and feeder roads have a stronger effect than improving the main road network. This is especially the case, if in the past major road networks have already been improved. Expected Impacts in Ningxia 41. With specific focus on Ningxia, a field survey was conducted in September The details of this survey are presented in the Social and Poverty Assessment. The survey revealed that local inhabitants expect considerable impacts from the feeder road rehabilitation. Villages will receive new opportunities by receiving direct access to all-weather roads. This may imply that traders drive with larger vehicles directly into the village to purchase farm products, competition amongst traders strengthens, and more transport services may be offered. Most importantly, access is guaranteed all year around (if snow removal is well organized). Altogether this will have relatively larger impacts on poverty than upgrading of existing all-weather roads. Providing direct access to rural farmsteads and fields generates new opportunities for the farmers (i) to shift for higher value or heavy crops, (ii) to increase production, (iii) to market directly, or (iv) to traders. A direct impact on farm gate prices and thus income may be expected. 42. One of the strongest social impacts to be expected from improvements of feeder roads is the access to social services: health services may be reached more easily and ambulances can access the village. Children travel more conveniently and faster to school, banking services are more easily accessible and fire brigades are able to reach the village. 43. The improvement of trunk roads will have impacts on the overall efficiency of the road network and thereby reduce transport costs through shorter travel times, reduced vehicle operating costs and reduced environmental costs. This will have a certain impact on poverty as well, since farm gate prices will increase. However, the trunk road conditions in Ningxia are better than those of the feeder roads. Therefore, the economic impacts of upgrading trunk roads will relate more to faster travel times and a reduction of vehicle operating costs, while the socioeconomic impact of the upgrading of feeder roads, that are currently in a very bad condition, is more significant. 13 Qiaolun Ye (2006).