Completion Report

Project Number: 39176-043 Loan Number: 2831 August 2020

Azerbaijan: Road Network Development Program (Tranche 3)

This document is being disclosed to the public in accordance with ADB’s Access to Information Policy.

CURRENCY EQUIVALENTS

Currency unit – Azerbaijani manat (AZN)

At Appraisal At Project Completion 20 June 2011 28 September 2017 AZN1.00 = $1.27 $0.59 $1.00 = AZN0. 78 AZN1.70

ABBREVIATIONS

AAY – Azeravtoyol State Agency ADB – Asian Development Bank ADT – average daily traffic ARS – Azer Road Service BCP – border -crossing point DED – detailed engineering design DSC – design and supervision consultant EIA – environmental impact assessment EIRR – economic internal rate of return EMP – environmental management plan GAP – gender action plan GDP – gross domestic product IRI – international roughness index km – kilometer LAR P – land acquisition and resettlement plan MFF – multitranche financing facility MOT – Ministry of Transport PFR – periodic financing request PIU – project implementation unit PPMS – project performance management system PSC – project steering committee RNDP – Road Network Development Program RRP – report and recommendation of the President VOC – vehicle operating cost

NOTES

(i) The fiscal year (FY) of the Government of Azerbaijan and its agencies ends on 31 December. “FY” before a calendar year denotes the year in which the fiscal year ends, e.g., FY2018 ends on 31 December 2018.

(ii) In this report, “$” refers to United States dollars.

Vice-President Shixin Chen, Operations 1 Director General Werner Liepach, Central and West Asia Department (CWRD) Country Director Nariman Mannapbekov, Azerbaijan Resident Mission (AZRM), CWRD

Team leader Yagut Ertenliche, Senior Project Officer, AZRM, CWRD Team members Aziz Haydarov, Senior Portfolio Management Specialist, AZRM, CWRD Faraj Huseynbeyov, Senior Project Officer, AZRM, CWRD Afag Javadova, Project Analyst, AZRM, CWRD Jurgen Sluijter, Senior Transport Specialist, Transport and Communications Division (CWT C), CWRD Carmina Luna, Senior Project Officer, CWTC, CWRD

In preparing any country program or strategy, financing any project, or by making any designation of or reference to a particular territory or geographic area in this document, the Asian Development Bank does not intend to make any judgments as to the legal or other status of any territory or area.

CONTENTS Page BASIC DATA i I. PROJECT DESCRIPTION 1 II. DESIGN AND IMPLEMENTATION 1 A. Project Design and Formulation 2 B. Project Output 4 C. Project Costs and Financing 4 D. Disbursements 5 E. Project Schedule 5 F. Implementation Arrangements 6 G. Consultant Recruitment and Procurement 6 H. Gender Equity 7 I. Safeguards 7 J. Monitoring and Reporting 8 III. EVALUATION OF PERFORMANCE 10 A. Relevance 10 B. Effectiveness 10 C. Efficiency 10 D. Sustainability 11 E. Development Impact 11 F. Performance of the Borrower and the Executing Agency 12 G. Performance of the Asian Development Bank 12 H. Overall Assessment 12 IV. ISSUES, LESSONS, AND RECOMMENDATIONS 13 A. Issues and Lessons 13 B. Recommendations 14

APPENDIXES 1. Design and Monitoring Framework 16 2. Project Cost at Appraisal and Actual 17 3. Project Cost by Financier 18 4. Disbursement of ADB Loan and Grant Proceeds 20 5. Contract Awards of ADB Loan and Grant Proceeds 21 6. Chronology of Main Events 22 7. Status of Compliance with Loan Covenants 23 8. Environmental Safeguard Audit Summary 29 9. Gender Action Plan Implementation Matrix 32 10. Economic Revaluation 35

BASIC DATA

A. Loan Identification 1. Country Azerbaijan 2. Loan number and financing source 2831 ( ordinary capital resources ) 3. Project title Road Network Development Program (Tranche 3) 4. Borrower Republic of Azerbaijan 5. Executing agency State Agency for Azerbaijan Automobile Roads 6. Amount of loan $200.00 million 7. Financing modality Multitranche financing facility

B. Loan Data 1. Appraisal – Date started 20 June 2011 – Date completed 25 June 2011 2. Loan negotiations – Date started 13 October 2011 – Date completed 17 October 2011 3. Date of Board approval 14 December 2011 4. Date of loan agreement 27 July 2012 5. Date of loan effectiveness – In loan agreement 26 August 2012 – Actual 23 October 2012 – Number of extensions 6. Project completion date – Appraisal 30 June 2014 – Actual 28 September 2017 7. Loan closing date – In loan agreement 31 December 2014 – Actual 28 September 2017 – Number of extensions 2 8. Financial closing date – Actual 12 January 2018 9. Terms of loan – Interest rate LIBOR + 0.60% less credit of 0.20% – Maturity (number of years) 24 – Grace period (number of years) 4 10. Disbursements a. Dates Initial Disbursement Final Disbursement Time Interval 12 June 2013 12 December 2017 54.05 months Effective Date Actual Closing Date Time Interval 23 October 2012 12 January 2018 62.7 months

ii

b. Amount ($ million) Increased (Decreased) Cancelled Last Original during during Revised Amount Undisbursed Allocation Implementation Implementation Allocation Disbursed Balance Category (1) (2) (3) (4=1+2 –3) (5) (6=4 –5) Civil works 164.50 17.86 0.00 182.36 182.36 0.00 Consulting 1.00 3.21 0.00 4.21 4.21 0.00 services Contingencies 34.50 (21.08) 13.42 0.00 0.00 0.00

Total 200.00 13.42 186.57 186.57 0.00

B. Project Data

1. Project cost ($ million) Cost Appraisal Estimate Actual Foreign exchange cost 200.00 186.57 Local currency cost 50.00 33.27 Total 250.00 219. 84

2. Financing plan ($ million) Cost Appraisal Estimate Actual Implementation cost Borrower financed 38.00 33.27 ADB financed 200.00 186.57 Total implementation cost 238.00 219.8 4 Interest during construction costs Borrower financed 12.00 0.00 ADB financed Total interest during construction cost 12.00 0.00

3. Cost breakdown by project component ($ million) Component Appraisal Estimate Actual A. Base Cost 1. Civil works 164.50 182.36 2. Land acquisition 8.00 0.70 3. Consulting services 1.00 4.21 4. Taxes and duties 30.00 32.57 Subtotal (A) 203.50 219. 84 B. Contingencies 34.50 0.00 C. Financing Charges during Implementation 12.00 0.00 Total Project Cost (A+B+C) 250.00 219.8 4 Note: Numbers may not sum precisely because of rounding.

4. Project schedule Item Appraisal Estimate Actual Date of contracts with consultants Q4 2014 24 December 2014 Civil works contracts Construction of Masalli–Astara Motorway: Section B Date of award Q3 2012 15 March 2013 Completion of work Q4 2014 27 March 2017 Completion of tests and commissioning 27 September 2017

iii

Start of operations 18 September 2018 Construction of Masalli–Astara Motorway: Section C Date of award Q1 2013 18 October 2013 Completion of work Q4 2014 27 March 2017 Completion of tests and commissioning 20 October 2017 Start of operations 18 September 2018

5. Project performance report ratings Implementation Period Single Project Rating From 15 December 2012 to 7 April 2013 Actual problem From 12 April 2013 to 31 December 2013 On track From 1 January 2014 to 31 December 2014 On track From 1 January 2015 to 31 December 2015 On track From 1 January 2016 to 31 December 2016 On track From 1 January 2017 to 28 September 2017 On track

D. Data on Asian Development Bank Missions No. of No. of Person- Specialization Name of Mission Date Persons Days of Members a Consultation mission 1 13–15 May 2012 4 12 a,b,c,d Loan inception mission 25–29 Jan 2013 4 20 e,b,d,c Consultation mission 2 19–21 Apr 2013 2 6 b,c Loan review mission 1 5–10 Jul 2013 2 12 b,c Loan review mission 2 17–21 Sept 2013 3 15 b,c,d Loan review mission 3 4–6 Feb 2014 3 9 b,c,d Loan review mission 4 30 Aug–2 Sept 2014 4 20 b,c,e,a Midterm project review mission 30 Nov–9 Dec 2014 5 42 d, b,b,c,c Loan review mission 5 29–30 Jan 2015 4 8 b,d,c,c Loan review mission 6 13 May 2015 3 3 b,c,c Loan review mission 7 6–7 Oct 2015 2 4 b,c Loan review mission 8 22–27 Jun 2016 2 2 b,c Loan review mission 9 13–21 Oct 2016 4 1 b,c,d,d Loan review mission 10 1–9 Mar 2017 7 56 b,c,a,b,c,c,c, Loan review mission 11 28–30 Aug 2017 5 15 c,c,c,d,f Project completion mission July 1–2 2019 2 4 c,d a a= environment specialist, b = transport specialist, c = project officer, d = project analyst, e = social development specialist, f = portfolio management specialist.

I. PROJECT DESCRIPTION

1. After the collapse of the Soviet Union, most roads in Azerbaijan were in unsatisfactory condition because of inadequate funding for road maintenance and poor management of weight control on the roads. As a result, a large part of the road network needed reconstruction or rehabilitation. The poor road conditions resulted in high transport costs, long travel time, and danger to safety. Transport plays an important role in the national economy, and the country’s vehicle fleet was expected to rapidly increase over the medium term, which dictated road network expansion. To help the Government of Azerbaijan meet these challenges, on 28 September 2007, the Asian Development Bank (ADB) approved a multitranche financing facility (MFF) for the government’s Road Network Development Program (RNDP) for an aggregate amount of $500 million. On 11 December 2011, ADB approved a $200 million loan for RNDP Tranche 3.1

2. Complementing tranches 1 and 2 under the MFF, the project aimed to contribute to sustained economic and social development in Azerbaijan by improving the Masalli–Astara section of the Alat–Astara highway connecting Baku, the capital, to Azerbaijan’s southern border.2 The project outcome was the development of an efficient, safe, and sustainable transport corridor, linking the country domestically and internationally. The project output is described in para. 12.

II. DESIGN AND IMPLEMENTATION

3. The project’s scope was originally part of tranche 1, approved in 2007, which envisaged construction of three sections of the Masalli–Astara expressway: (i) section A, covering 0 kilometer (km)–22.1 km; (ii) section B, 22.1 km–45.1 km (Girdani–Penser); and (iii) section C, 45.1 km–61.4 km (Penser–Astara). Because of the significant price increases in the cost of construction materials during 2007–2008, the funding under tranche 1 turned out to be insufficient to finance the works on all three sections (the funding gap was about $160 million). After completion of the detailed engineering designs (DEDs) and the cost estimates for all three sections, sections B and C were removed from the scope of RNDP project 1 and eventually formed the scope of RNDP Tranche 3.3

4. An updated economic analysis was undertaken considering the revised costs of sections A, B, and C, as consecutive and integral parts of the same motorway. The analysis showed that the Masalli–Astara road remained economically viable with a revised economic internal rate of return of 12.9%. The sensitivity test to assess the impact from different benefits and costs showed that the project remained robust across the broad range of variations in the key parameters. The environmental assessment regulatory framework prepared for the RNDP was updated along with the environmental impact assessment (EIA) to reflect the requirements of ADB’s Safeguard Policy Statement (2009). The EIA concluded that the overall environmental impacts would be limited to the construction phase and could be minimized by adequate mitigation measures included in the environmental management plan (EMP). The land acquisition and resettlement framework was also updated to reflect the provisions of the Safeguard Policy Statement. The approval of the final land acquisition and resettlement plans (LARPs) by ADB was a condition for contract awards, and civil works would commence only after the LARPs were fully implemented and the necessary

1 ADB. 2007. Report and Recommendation of the President to the Board of Directors: Proposed Multitranche Financing Facility to the Republic of Azerbaijan for the Road Network Development Program . Manila. 2 ADB. 2020. Completion Report: Road Network Development Program–Project 1 in Azerbaijan. Manila (Loan 2354/2355[SF]-AZE); ADB. 2013. Completion Report: Road Network Development Program–Project 2 in Azerbaijan. Manila (Loan 2433-AZE) . 3 ADB. 2011. Loan 2354-AZE: Road Network Development Program–Project 1–Amendments to the Loan Agreement for Road Network Development Program.

2 monitoring reports submitted. At appraisal, the project was assessed to have no adverse impact on indigenous peoples.

5. At appraisal, the project was expected to help reduce poverty by creating an enabling environment and by increasing access to markets, jobs, and basic social services for the whole population, including the poor. 4 The project was expected to create about 400 person-years of jobs for road construction, of which about 200 person-years would be unskilled jobs and half of these would be for the poor. The project design discussed the need to monitor and minimize potential risks of HIV/AIDS and other sexually transmitted infections, as well as drug and human trafficking. The HIV/AIDS–associated risks were to be mitigated through an awareness and prevention program in the project area by an international nongovernment organization, and inclusion of provisions in the civil works contracts requiring contractors to take measures to protect construction workers from potential health risks.

6. The project design’s key assumptions were (i) continued political and macroeconomic stability in Azerbaijan, (ii) the government’s continued commitment to regional cooperation, (iii) effective implementation of the RNDP for 2006–2015, and (iv) timely provision of counterpart funding. The key risks to the project included (i) limited coordination among government agencies during implementation, (ii) rapid deterioration of the road network caused by weak axle-load controls and insufficient funding for maintenance, (iii) delay in implementation of LARPs for sections B and C, and (iv) delays because of contractors’ slow mobilization and importing equipment. The risk mitigation measures were based on the then-government’s transport institutional setup anchored around the Ministry of Transport (MOT), which was reorganized in January 2016, resulting in the reduced relevance of the at-appraisal risk mitigation measures.

A. Project Design and Formulation

7. Roads are the most important element of Azerbaijan’s in-land transport network. The share of passenger turnover by road increased from 76% in 1990 to 92% in 2007 and 98% in 2018. The share of freight turnover by road grew from 8% in 1990 to 47% in 2007 and 79% in 2018. 5 The Baku–Astara highway is a major strategic highway in Azerbaijan and an integral part of the North–South Corridor, the shortest transport route from Europe to the Persian Gulf.

8. Improvement of road infrastructure, therefore, was a government priority as reflected in the State Program on Poverty Reduction and Economic Development for 2008–2015 and the State Program on Socioeconomic Development of Regions 2009–2013. 6 Road infrastructure improvement was also a priority under ADB’s country and corporate strategies. 7 The government’s road network investment plan envisaged total financing of $3,500 million during 2006–2015, with the state budget as the main funding source supplemented by external financing.

4 Most villages were 0.5 km or more from the main roads, and most of the rural population, particularly the poor, had to walk to the main roads to access social services and buy necessities. 5 The State Statistics Committee, Republic of Azerbaijan. 6 Government of Azerbaijan. 2008. State Program on Poverty Reduction and Sustainable Development in the Republic of Azerbaijan for 2008–2015 . Baku; Government of Azerbaijan. 2008. State Program on Socioeconomic Development of Regions for 2009–2013 . Baku. 7 ADB. 2006. Country Strategy and Program Update: Azerbaijan . Manila; ADB. 2008. Strategy 2020: The Long-Term Strategic Framework of the Asian Development Bank, 2008–2020. Manila . Strategy 2020 indicated transport infrastructure development as a core area for ADB engagement.

3

9. The project was the outcome of the highly inaccurate preliminary design of RNDP project 1, prepared at the end of 2006, when the road alignment had not been finalized or sources of materials fully identified. 8 Increased government spending on infrastructure projects added inflationary pressure on key supplies, works, and services. Project implementation was delayed because of land acquisition and procurement of civil works contracts for sections B and C. The delays had a compound effect, resulting in significant price increases in the cost of construction materials during 2007–2008. During implementation, the design and supervision consultant (DSC) engaged under RNDP project 1 prepared the DED and updated the cost estimates. The estimates turned out to be far above those prepared at appraisal and the available financing highly insufficient to cover the project’s scope. The funding gap was about $160 million.

10. To solve the cost overrun, in March 2011, ADB considered additional financing for the project. However, as LARPs for sections B and C had not been finalized and the bidding documents were not ready, a change in scope was approved in June 2011 to remove these sections from RNDP project 1. After receiving the third periodic financing request from the government to finance sections B and C in August 2011, ADB again assessed the project for an additional financing modality. However, having processed the scope reduction with requisite amendment to the RNDP project 1 loan agreement and having informed the Board about the periodic financing request for RNDP Tranche 3, ADB opted to process the project as a stand- alone project of the RNDP. At completion, the rationale for processing the project as stand-alone turned out to be less relevant because section A construction, which was expected not to be delayed because of the low readiness of sections B and C, was eventually completed jointly with those sections in August 2017.

11. In September 2016, the government deferred construction of new cross-border facilities at the border with Iran, as the capacity of existing cross-border infrastructure was found to be sufficient to accommodate current and future cross-border passenger and freight traffic. Uncertainty was also growing on the readiness of infrastructure on Iran’s side. This necessitated a minor change in project, approved in July 2017, to exclude the last 2.4 km (km 59.0–km 61.4) from section C.9 About a year after project completion and given the agreement with Iran on the location of the cross-border point, the government allocated state budget funds to construct the 2.4 km link and the cross-border bridge as originally planned.10 The road works started in June 2019 and are scheduled for completion in December 2020. The cross-border bridge’s design is being finalized with the start of construction planned for late 2020. With the completion of this “last mile” connection, the project’s original scope will be realized.

8 ADB. 2020. Completion Report: Road Network Development Program, Project 1 in Azerbaijan . Manila . 9 Starting from the interchange at km 58.6, traffic would be diverted to the existing M3 highway up to the current border- crossing point. The embankment work for the 800-meter (m) section from km 59.0 to km 59.8 had been completed before the government’s decision on the border-crossing point. This 800 m section was expected to be utilized by the Azeravtoyol State Agency (AAY) to construct roadside facilities such as parking lots or warehouses. 10 Presidential Order on Additional Measures for the Construction of the Alat–Astara–Iranian Border Highway , 15 October 2018, no. 546.

4

B. Project Output

12. The project had one output: construction of sections B and C of a four-lane expressway between Masalli and Astara in the southern corridor, with access roads rehabilitated.11 Despite the almost 2.8-year delay and construction cost about 10% higher than planned, the project output targets were achieved. The details are in the design and monitoring framework (DMF) (Appendix 1).

13. The civil works in section B (km 22.15–km 45.10, Girdani–Penser) commenced on 18 May 2013 and, after three contract variations, were completed on 18 June 2017. The civil works in section C (km 45.1–km 61.4, Penser–Astara) commenced on 27 January 2014 and, after two contract variations, were completed on 15 July 2017. The final length of section C was until km 59 only because of the change in scope in July 2017 (para. 11).

14. In 2013, during the bidding process for both sections, ADB proposed installing high-speed fiber-optic telecommunication cables along the alignment of the entire Alat–Astara highway.12 Although the Azeravtoyol State Agency (AAY) initially supported the proposal, it eventually decided to proceed without the change in the scope of civil works, which would have required more time and cost to conduct due diligence, prepare engineering designs, and vary the construction contracts, which had already been awarded. In July 2018 (after the project road was completed), the feasibility report on introducing tolling on the Masalli–Astara expressway under project 1 of the Second Road Network Development Program 13 showed that cable infrastructure was needed along the road alignment and estimated that installing it would cost about 20% of the overall works on tolling infrastructure.14

15. The delivery of the project output was delayed because of (i) the slow pace of land acquisition; (ii) the need to move public utility infrastructure (e.g., power transmission poles) installed by other government agency on the already secured and cleared land for the project road; (iii) unforeseen construction site issues such as high water tables; and (iv) extreme weather conditions such as the subtropical rainy season from October 2013 to April 2014, constraining asphalt-laying works. All project sites had a high water table. To solve the issue, a subconsultant had to be engaged to carry out detailed geotechnical and hydrological surveys, which required more time and delayed project implementation further.

C. Project Costs and Financing

16. At appraisal, the total project cost was estimated at $250.0 million, of which the ADB loan was envisaged to finance $200.0 million (80% of the total project cost) and government counterpart funding estimated at $50.0 million. At completion, the project’s actual cost amounted to $219.85 million, of which ADB loan financing was $186.57 million (84% of the actual project cost) and counterpart funding totaled $33.27 million, which covered taxes and duties. A partial loan cancellation of $13.1 million caused by the excessive unused balance under the unallocated

11 Schedule 1 of the loan agreement for the project (i) does not provide for rehabilitation of access roads; and (ii) notes the component on project management support and consulting services for construction supervision, financial audit, social and environmental assessments, and institutional capacity development. 12 The utility duct would have allowed the laying of fiber-optic cables along the project road for multipurpose future use, such as tolling infrastructure requiring a high-speed data transmission channel. 13 ADB. 2007. Azerbaijan Second Road Network Development Program. Manila 14 Per the feasibility report, the cable and related works included provision and laying of optical fiber cable from the central server to all plazas, including high-density polyethylene pipes, warning tape, trenching works, and backfill. The government eventually decided not to introduce tolling on the Masalli–Astara expressway but to do so on the new M1 highway, expected to be completed in December 2020.

5

loan category was followed by the final cancellation of the undisbursed loan balance of $0.3 million upon the loan’s financial closing. The unallocated loan proceeds were reallocated once during implementation to works and consulting services. The reallocation to works was needed to fund the increased number of road works contracts, and reallocation to consulting services to finance the DSC previously financed under RNDP project 1. The government shouldered the loan financing charges.

17. Civil works contracts totaled $182.34 million at completion, about 11% higher than the $164.5 million estimated at appraisal because of contract variations and price adjustments for fuel and bitumen costs during implementation of the road works. Contracts were varied to meet the need for (i) more materials and earthworks such as geotextile and rockfill for slope protection (to rectify the high water tables noted in para. 15), (ii) additional drainage facilities and utility relocation works, (iii) more underpasses and animal crossings, and (iv) “New Jersey”–type concrete barriers to replace the median metal guardrails along the entire length of the road to improve road safety. Consulting services totaled $4.22 million, compared with the $1.00 million planned at appraisal, because the DSC contract was transferred from RNDP project 1 to the project. Details on project cost at appraisal and actual are in Appendix 2, those on project cost by financier in Appendix 3.

D. Disbursements

18. The first disbursement was made in June 2013 (against the expected Q4 2012), the last in December 2017 (against the expected Q4 2014). Loan disbursements totaled $186.57 million, compared with $200.0 million at appraisal. Because sections B and C were part of RNDP project 1 and were, therefore, at an advanced stage of procurement, advance procurement was used. The details of contract award and disbursements are in Appendixes 4 and 5.

19. All disbursements from the loan proceeds were made in accordance with ADB’s Loan Disbursement Handbook (2007, as amended from time to time). The direct payment procedure was used to disburse the loan proceeds to all project contractors and consultants. While the periodic financing request report provided for the statement of expenditures procedure to reimburse any eligible expenditure (such as individual short-term consultants) for any individual payment transaction of $100,000 equivalent and below, the procedure was not used under the project.15 The audited project financial statements did not identify major deficiencies or risks to financial management. Internal controls, fund management, and financial accounting of the project were generally satisfactory.

E. Project Schedule

20. The project loan was approved on 14 December 2011, the loan agreement was signed on 27 July 2012, and the loan became effective on 23 October 2012, with an original closing date of 31 December 2014. The contracts for sections B (15 March 2013) and C (18 October 2013) were signed 6 months later than originally envisaged.

21. Highly unfavorable weather conditions (heavy winter rains) led to five contract variations and implementation delays in road works by both contractors. The loan closing was subsequently extended by a total of 2 years and 9 months. Civil works were completed on 25 September 2017

15 Advance fund and the statement of expenditures procedures were used under RNDP project 1 to finance the project implementation unit (PIU) expenditures. The advance fund procedure, therefore, was not needed under the project.

6 and the loan was closed on 28 September 2017. The loan financial closing was on 12 January 2018, within the 4-month winding-up period. The project’s main events are in Appendix 6.

F. Implementation Arrangements

22. Initially, the MOT was the executing agency, and the Azer Road Service (ARS), an open joint-stock company under the MOT, was the implementing agency. Following the reorganization of the MOT in January 2016, the ARS assumed responsibility as executing agency from February 2016. The ARS’s name and legal status changed: (i) in 2016, the ARS was changed to the AAY; and (ii) in December 2017, the AAY was restructured as the State Agency of Azerbaijan Automobile Roads, a public legal entity. 16 The project steering committee (PSC) envisaged in the facility administration manual was not established. Instead, the government held meetings as needed, usually chaired by the deputy prime minister with the participation of representatives from agencies such as the MOT, AAY, Ministry of Finance, and Ministry of Economy. The PSC was not established because, following the reorganization of the MOT, the government concluded that the PSC would duplicate the existing coordination and reporting arrangements exercised by the Cabinet of Ministers with respect to externally financed projects. The empowered AAY also felt that the project might be further delayed because of the time the high-level PSC would require to convene and make decisions.

23. The project implementation unit (PIU) established in the AAY reported to the head of the AAY’s Investment Department; was headed by a project director; and included financial, environmental, social, and procurement specialists, as well as highway engineers. 17 Since January 2014, a program management consulting firm engaged under project 1 of the Second Road Network Development Program (footnote 14) assisted the PIU in project management, procurement, and capacity development. 18 The project-financed design and supervision consulting firm assisted the PIU in contract administration and on-site supervision, contractor quality control and assurance, safeguard compliance monitoring, and project monitoring and evaluation. 19

G. Consultant Recruitment and Procurement

24. No new consulting service recruitment was originally envisaged under the project, as the consultant under RNDP project 1 (the DSC) was expected to provide the necessary consulting services. To fund the eventual extension of the DSC’s services to cover the project’s duration, $1.0 million was included in the project..20 By June 2016, the funds for the DSC under RNDP project 1 were fully utilized, and from July 2016, the financing for the RNDP project 1 consultants was transferred to the project. Consulting services cost $4.2 million or $3.2 million more than planned because of the 2.8-year extension of the project. The DSC helped the AAY prepare monthly and quarterly progress reports, implemented monitoring and evaluation programs, and provided on-the-job training for contractors and the AAY staff. Performance of the DSC was satisfactory .

16 Government of Azerbaijan. 2016. Decree of the President of the Republic of Azerbaijan on the legal status of “Azeryolservis” OJSC . 9 March 2016. Baku; Government of Azerbaijan. 2017. Decree of the President of the Republic of Azerbaijan on the State Agency of Azerbaijan Automobile roads . 19 December. Baku. 17 The same PIU was also in charge of European Bank for Reconstruction and Development–financed road projects. 18 ADB. 2007. Report and Recommendation of the President to the Board of Directors: Proposed Multitranche Financing Facility Republic of Azerbaijan: Road Network Development Program. Manila. 19 The DSC was initially financed under RNDP project 1 and transferred to RNDP project 3 in July 2016. 20 Additional consulting services, if required, were to be procured in accordance with ADB’s Guidelines on the Use of Consultants by Asian Development Bank and its Borrowers (2010, as amended from time to time).

7

25. Civil works were procured in accordance with ADB’s Procurement Guidelines (2010, as amended from time to time). Following advance procurement action, two contractors were engaged though international competitive bidding for sections B and C. Procurement of all contracts followed ADB’s single-stage, one-envelope procedure without prequalification. Contractors maintained a generally good level of environmental management and orderly documentation. Most of the issues encountered during the construction phase were mainly on- site contamination, dust control, and solid and construction waste management, which were duly resolved following ADB recommendations. The performance of the contractors for sections B and C was overall satisfactory .

H. Gender Equity

26. The project is categorized as having some gender elements . At appraisal, the project was expected to have a positive impact on women. It was expected to promote women’s participation in monitoring project impacts on local communities, and the civil works contractors would be encouraged to employ women during construction. Direct employment of women in project activities was considered limited, but women were expected to benefit from improved access to basic social services and markets in the project area. Increased demand for goods and services was expected to benefit women involved in trading and services.

27. A gender action plan (GAP) was developed for the project (a single GAP covered RNDP projects 1 and 3). The draft GAP was submitted to ADB in October 2014 (18 months later than envisaged in the loan agreement). The GAP implementation matrix is in Appendix 9. The implemented gender activities included gender-sensitive road safety awareness training for communities along the project road, employment by the project of women from the project communities, employment of women as engineers by the construction supervision consultant, and employment of women for administrative support.21 Because of the nature of construction work as well as cultural factors, the gender composition of consultants and workers hired by the project was skewed toward men.

I. Safeguards

28. Environmental safeguards . The project followed the requirements of ADB’s Safeguard Policy Statement (2009) and was classified as environment category A. The project’s EIA concluded that the overall environmental impacts were limited to the construction phase and could be minimized by adequate mitigation measures included in the EMP. The project road’s alignment in section B slightly skirted a protected forest plot (lowland Hirkan Forest Reserve). Following the corrective action taken after the review mission in May 2012, the DSC deployed a biodiversity specialist before construction to do a detailed study and propose additional measures to mitigate any adverse impact on the forest’s ecosystem, as specified in the final environmental monitoring report. 22 The DSC submitted semiannual monitoring reports on the implementation of environmental safeguards.

29. Environmental management in section C during construction was considerably better than in section B. Throughout the construction of section C, the contractor maintained a high standard of environmental management and protection and delivered higher-quality reports. No major noncompliances occurred during construction. The noncompliances observed in sections B and

21 The contractor hired 2 women for administrative support and 57 for the auxiliary staff (canteen workers, utility workers, and office cleaners). 22 The final safeguard documents are accessible at https://www.adb.org/projects/documents/aze-39176-043-emr-0.

8

C were restricted to minor issues such as poor material storage, substandard waste storage arrangements, lax fire safety arrangements, and, at times, poor dust and traffic management. As the noncompliances were minor and short term, the contractors generally promptly dealt with emerging issues. On balance, the AAY, with the help of the DSC, ensured adequate environmental safeguard management in both sections.

30. In November 2019, ADB conducted an independent post-completion environmental safeguard audit for all projects in category A for environmental safeguards, concluding that the project had complied with the agreed environmental requirements. The audit summary is in Appendix 8.

31. Social safeguards . The project followed the Safeguard Policy Statement and the laws and regulations on land acquisition and resettlement in Azerbaijan in 2007, and was rated category A for involuntary resettlement and category C for indigenous people. 23 Because of the changes in land prices and valuation methods in Azerbaijan, in August 2011, the AAY submitted to ADB the revised (i) LARP for sections B and C, and (ii) land acquisition and resettlement framework. They outlined impacts, compensation policy, rehabilitation provisions, cost estimates, and the implementation schedule to compensate for the impacts related to the construction of section C.

32. Km 45+000 was considered the starting point of section C during initial LARP preparation. Because of the changes in separating sections B and C, however, the bidding documents defined km 41+500 as the end point of section B and start of section C. The key purpose of this amendment was to add four villages (Kokolos, Bala Shahagac, Kolatan, and part of Pensar) to section C and subtract the last 3 km from section C (Suparibag village, part of Archivan, and part of Gilekeran). 24 Because of additional land impacts, an addendum to the LARP was prepared in October 2013, providing details on the structures and households affected by the project. As per the agreed entitlement matrix, 1,288 affected persons were compensated, with no unresolved issues remaining.25 No social and cultural disruptions were caused to the affected households and communities. However, the following areas are recommended for more thorough examination in preparing future project designs and LARP documents: (i) more meticulous consultation at the project design stage and for the LARP, (ii) capacity building for municipalities and other local executive representatives to increase their understanding about ADB’s resettlement safeguard policy, and (iii) consultations with utility service companies.

J. Monitoring and Reporting

33. The project complied with the general covenants specified in the loan agreements, except that on audited project financial statements, which in some years were submitted up to 8 months late (partial compliance). The project did not comply with the special loan covenant on installing vehicle weighing stations on the project road mainly because of poor coordination between the construction activities managed by the PIU and the AAY’s road maintenance department. As a

23 Construction of sections B and C required about 172 hectares (ha) of privately owned lands and 88 ha of public land. Based on the detailed design, more than 4,500 people (1,200 households) were expected to be affected. Of these, about 600 (167 households) were expected to lose more than 10% of their agricultural land and about 79 (22 households) were assessed as requiring relocation. The project had no adverse impact on indigenous people. Ethnic Azerbaijanis account for most of the population in the project area. Other ethnic groups include the Talysh, who were not adversely affected by the project. 24 The last 3 km of the road alignment toward the border had to be agreed with Iran’s authorities. 25 ADB. 2011. Resettlement Plans for Section B and Section C ; ADB. 2013. Addendum to LARP .

9

result, installing weighing stations was not included in plans in 2011 and weighing stations were not built along the project road even though the installation plan provided for them in 2015. 26

34. The project partially complied with the special loan covenants on implementing the project in accordance with the provisions in the facility administration manual, and with the GAP. The facility administration manual provided for a PSC to oversee and coordinate the project. The PSC was never formally set up, and agencies coordinated with each other as needed through ad hoc meetings, resulting in less effective oversight and problem-solving. The AAY only partially implemented the GAP because of a lack of sex-disaggregated data, and because cultural restrictions prevented the intended involvement of women in construction works. No covenant was modified, suspended, or waived, as they remained conceptually relevant to the project design and important for sector policy dialogue with the government. On balance, the issues of compliance with the covenants did not significantly impact the project’s outcome and outputs. The status of compliance with loan covenants is in Appendix 7.

35. The project monitoring and reporting arrangements were overall satisfactory . The AAY provided ADB with the DSC’s monthly progress reports. The PIU established a project performance monitoring system and coordinated overall project implementation, including compliance with the loan covenants. The DSC worked closely with the PIU in supervising the contractors’ work, monitoring contractors’ conformity with environmental and social impact controls, supervising implementation by the contractor, and monitoring contractors’ conformity with traffic control and the road safety action plan during construction. The PIU environmental expert coordinated the EMP implementation and mitigation activities, resolved environmental issues, regularly monitored environmental quality, and reported the results to PIU management, which, if needed, sought ADB’s advice.

36. Separate financial accounts for the projects were established and financial accounts audited by an external auditor selected by the executing agency and acceptable to ADB. The AAY submitted the annual project financial statement audit reports to ADB. The audited project financial statements were submitted with significant delays of 6‒8 months during 2013–2016. The audit reports provided quantitative and qualitative assessments of the financial management systems adopted for the project, which were found satisfactory. The audit reports were found to be generally in order and acceptable to ADB and all had unqualified opinions.

37. The contractors, in consultation with the DSC, prepared grievance redressal mechanisms under their contracts and communicated such mechanisms to the affected communities during the public consultation meetings. Contractors maintained a grievance register, where all complaints and comments were logged, alongside details of the complainant, and how and when the issue was resolved. The DSC expedited timely and adequate remedies for the grievances by the contractors, in coordination with the AAY. In total, 25 grievances were recorded: 17 for section B and 8 for section C. No unresolved grievances remained.

26 The Decree of the Cabinet of Ministers No. 214s, 16 July 2015, provides for 32 weighbridge stations to be installed across the country. These included three stations to be installed on the project road at the final stage of road works, but no vehicle weighing station has yet been installed on the project road.

10

III. EVALUATION OF PERFORMANCE

A. Relevance

38. The project was highly relevant at appraisal and completion from socioeconomic and connectivity perspectives. It was also highly relevant for the government’s development plans and ADB’s country strategy. Using the additional financing modality might have been the preferred option as it would have preserved the natural integrity of the original economic and infrastructure investment scope of the project, without the seemingly artificial separation of the first section of the same road from the other two sections. Because of timing and readiness considerations, the project was processed as a stand-alone tranche, the second-best option that enabled delivery of the planned strategic infrastructure asset.

39. The DMF’s outcome indicators were related to development of the country-wide road network, and the outcome formulation to the southern road corridor. The outcome indicators followed the approach of RNDP project 1, which had country-wide road indicators, because these were (during preparation of RNDP project 1) the best available statistics. Although access road rehabilitation was part of the project output formulation in the DMF, no output indicator related to the access roads. As the MFF’s third project, it could adopt arrangements to enable use of the outcome and output indicators that would be better linked with it and enable a more accurate assessment of its performance during implementation and at completion. Although the DMF was revised during implementation to reflect the extended implementation timeline, the outcome and output indicators were not changed.

40. The project was vital to realize the Masalli–Astara expressway, which, in turn, is critical to shorten and modernize the Baku–Astara part of the country’s north–south transport corridor, which will anchor Azerbaijan as an efficient link between the southern and eastern European markets and the logistical and economic hubs in the Persian Gulf. Despite the issues with the modality and DMF indicators, and considering the significant connectivity gains, on balance the project is rated highly relevant.

B. Effectiveness

41. The project is rated effective because it achieved the outcomes and output targets, although later than planned. Because the government decided to open the entire Alat–Astara highway at once, the opening of the Masalli–Astara expressway took place in September 2018 or 1 year after project completion in September 2017, resulting in a 3.75-year delay. The main outcome of better connectivity and an efficient southern road corridor from Baku was largely achieved. The road users have started benefitting from reduced travel time and vehicle operating costs.27 Traffic volume on the road is increasing and will eventually expand economic activities in the project area. The international roughness index improved from 6.5 or more to 2.5 in 2017.

C. Efficiency

42. The project is rated inefficient because the economic internal rate of return (EIRR) calculated for the Masalli–Astara expressway was estimated at 12.8% at appraisal but

27 Completion of the four-lane Masalli–Astara expressway under RNDP projects 1 and 3 has reduced the average travel time from 90 to 40 minutes. The reduction of transport cost is estimated at 25%–30%.

11

reevaluated at 3.8% at completion.28 As the reevaluated EIRR is far below the ADB threshold of 12%, the project is considered economically unviable and indicates that the project was poor value for money. The reasons are (i) the actual cost of RNDP project 1 was about 60% higher than expected at appraisal in 2007; and (ii) traffic is significantly lower than expected, with average traffic on the Masalli–Astara expressway estimated at 4,800 vehicles/day, compared with 9,200 expected at appraisal. The economic downturn in 2015–2016 (following the oil price crunch in 2014) and traffic that was much lower than expected (mainly because of the economic downturn) were not foreseen at appraisal. The low efficiency, therefore, can also be attributed to exogenous factors beyond the project’s control. The details of the economic reevaluation are in Appendix 10.

D. Sustainability

43. Overall, the project is considered likely sustainable . Routine and periodic maintenance is carried out by the AAY through its regional units and financed from the State Road Maintenance Fund, established in December 2006. The government has been programming and allocating enough funds and resources annually for road maintenance based on established norms. 29 The overall spending for road maintenance has gone from $55 million in 2006 (about $9,700/km) to about $177 million in 2019 (about $11,500/km). 30 With assistance from the World Bank, seven new specialized regional road maintenance units were established under the AAY with the sole responsibility for highway maintenance; one is responsible exclusively for maintaining the new Alat–Astara highway. Next measures to strengthen the road maintenance framework might include introducing performance-based contractual relations between the regional units and the AAY to make utilization of the State Road Maintenance Fund more efficient.31

E. Development Impact

44. The project is rated satisfactory on development impact as the relevant impact target indicator—defined at the level of the country and not of the southern road corridor—was achieved. Trade turnover with neighboring countries reached $3.8 million in 2018. Freight traffic by road grew on account of improved road network and road conditions, reaching 16.857 million ton-km in 2018. Passenger transport by road also grew throughout the country and reached 25.276 million passenger-km in 2018. The project provides opportunities for further economic, industrial, and commercial development, increasing prospects for business and employment. With travel time reduced, most villagers can go to town centers more frequently, up from 5 to 17 times a year. The project’s nonphysical elements of project management also helped strengthen the

28 Economic reevaluation was done for the entire Masalli–Astara expressway because separate evaluation of the sections is impossible from economic and road transport integrity perspectives. For the same reason, the economic evaluation at appraisal was done for the entire Masalli–Astara expressway. 29 Presidential Decree of the Republic of Azerbaijan No. 1729, 12 October 2006. Establishment of the Automobile Road Specialized Budgetary Fund . Baku; The Cabinet of Ministers’ Order on the Road Fund Regulations, 25 January 2007, no.13; The Cabinet of Ministers’ Order on Road Maintenance Norms, 17 January 2006, no.12. The fund is sourced from revenues from the road tax, tax on imported cars, charges on technical inspection of vehicles, and simplified tax on transport service provision. The fund’s surpluses are rolled over to the next fiscal year. In 2019, the fund’s revenues totaled about $207 million and spending on road maintenance totaled $177 million, implying a surplus of $30 million. 30 Government of Azerbaijan, Ministry of Finance. Information on the 2019 State Budget Implementation. The per- kilometer road maintenance expenditure grew steadily from about $9,700 in 2008 to about $17,200 in 2015. Because of the economic crisis and sharp depreciation of the manat against the US dollar caused by the 2014 oil price shock, the per-kilometer maintenance expenditure dropped to about $11,400 in 2016 and $10,700 in 2017. It has since grown from about $11,000 in 2018 to $11,500 in 2019. 31 N. Valiyev. 2019. Modernizing Road Maintenance in Azerbaijan to Ensure Safer, More Efficient Travel . World Bank Blogs.

12 performance of the AAY as the country’s highway and regional road network developer and operator.

F. Performance of the Borrower and the Executing Agency

45. The government provided the required counterpart funds and the necessary support generally without delay. The executing agency closely coordinated and regularly monitored construction progress and the quality of physical works. The PIU, established by the executing agency and supported by the consultants, monitored project implementation on-site and prepared the required progress reports. The PSC was not established as envisaged in the facility administration manual, resulting in weakened interagency coordination and problem-solving (e.g., power transmission poles installed by another government agency on project land had to be removed at the cost of the project, no action was taken on the proposal to lay communication cables along the project road alignment, the GAP was weakly implemented, the weighing stations that were supposed to be installed on the project road were not).

46. The government and the AAY capably supported ADB missions during project preparation and implementation and at completion. The AAY ensured that project accounts and financial statements were audited by external auditors acceptable to ADB, although several audit reports were submitted late. The executing agency established the project performance management system. The government provided counterpart funding generally on time and in full. The AAY submitted the project completion report on time. On balance, the performance of the borrower and executing agency is rated satisfactory.

G. Performance of the Asian Development Bank

47. Project administration was undertaken by ADB headquarters until February 2014, when it was delegated to the ADB resident mission. ADB headquarters, however, continued providing guidance to the executing agency and resident mission on procurement and sector-specific technical issues. Most project review missions were conducted jointly by the resident mission and ADB headquarters staff members. Despite the issues with the GAP implementation, delayed submission of project audit reports, and the seeming lack of connection between the DMF’s outcome and output and their indicators, no financial management and results management experts joined the review missions. ADB’s approval of documents during processing and implementation was overall timely, and the payment of claims was prompt. In general, the government appreciated ADB’s timely advice on technical and contract administration matters. Overall, the performance of ADB is rated satisfactory .

H. Overall Assessment

48. The overall assessment of the project is less than successful, based on the ratings presented below.

13

Overall Ratings Criteria Rating Relevance Highly relevant Effectiveness Effective Efficiency Inefficient Sustainability Likely sustainable Overall Assessment Less than successful Development impact Satisfactory Borrower and executing agency Satisfactory Performance of Asian Development Bank Satisfactory Source: Asian Development Bank.

IV. ISSUES, LESSONS, AND RECOMMENDATIONS

A. Issues and Lessons

49. Inadequate engineering design . The DSC comprehensively revised the initial DED for sections B and C, which had been done during preparation of RNDP project 1 and originally covered sections A, B, and C. Despite the DSC’s revamp of the initial DED and award of the construction contracts, after the start of road works in sections B and C it became clear that additional geotechnical surveys had to be undertaken, including to test soils to improve the stability and safety of high embankments. Given this discovery, a specialized consulting firm was engaged through the works contract for section B to fully review the designs, undertake geotechnical investigations, conduct testing in high-risk project areas, and advise the contractors on necessary changes to the contracts. This all led to additional cost and implementation delays.32

50. Lack of interagency coordination and project oversight platform. The government- level PSC had not been established, constraining proactive problem-solving, promotion of the project’s soft agenda, and an integrated approach to infrastructure development. For example, the section B contractor had to do additional utility relocation works, including on power transmission poles installed in the project area by another government agency after the land had been acquired for the road works. The concerned agency claimed its actions were fully legitimate, and eventually the contractor had to relocate the poles, resulting in additional project cost and inefficient use of loan proceeds. The problems with the GAP implementation and the weighing station installation may also be attributed to the lack of a PSC as government platform, to which such issues could have been escalated for decision. Had the PSC been set up and operational, it might have facilitated a comprehensive review of ADB’s proposal on integrating the construction of a utility duct for the fiber-optic cable along the project road to modernize the telecommunication connection between central and southern Azerbaijan. Given the fragmented infrastructure planning, the PSC could have promoted integrated and more accountable project implementation.

51. Inadequate coordination between construction and asset management functions. Project investments and road maintenance functions in the AAY could have been more effectively synchronized, as seen from the lack of coordination between the PIU and the AAY road maintenance unit, which resulted in vehicle weighing stations to control axle overloading not being installed along the project road, while weighing stations along state-funded road projects were

32 This section discusses the issues, lessons learned, and recommendations specific to the project and complement those identified in the project completion report of RNDP project 1.

14 built.33 This situation could have been avoided if a workable coordination mechanism between the PIU and the road maintenance unit had been designed during project preparation and then financed during implementation. As the issue with the weighing stations had been encountered in RNDP project 1 and in previous road projects in Azerbaijan, incorporating a capacity-building component for the road maintenance unit in the project design might have been expedient.

52. Late submission of audited project financial statements. The audit firms’ performance was mixed. The two local audit firms, chosen using the least-cost selection method, provided poor reports that required corrections. Submissions of the audited project financial statements during 2009–2011 and 2013–2016 were significantly delayed, by up to 6–8 months. 34

53. Need to promote road sector sustainability . The transparency and accountability of road maintenance funding must be improved to extend the road network’s productive life. To this end, developing a comprehensive road maintenance program should be considered. It could cover (i) private sector participation (choice of the most suitable modalities and country areas); (ii) road asset management principles and funding (e.g., performance-based budget allocations to regional maintenance units, road asset inventory, and database); and (iii) enforcement of a vehicle axle load management system (e.g., establishing more vehicle weighing stations along the main roads and setting up an automated overloading control and monitoring system). 35

B. Recommendations

54. Raise the quality of DEDs. To ensure that DEDs are robust enough, engaging an independent engineer would be useful to (i) review the scope of the works for the DEDs; and (ii) validate the cost estimates and adequacy of the geotechnical, hydrological, and other surveys done by the DSC. This can be financed through ADB’s project readiness financing modality. It might be expedient to consider incorporating accountability provisions in the terms of reference of the DSC, such as not disbursing the retained payment or imposing a penalty within a certain time after DED completion if the DED needs to be significantly revised or additional surveys or assessments done because of the DSC’s lack of diligence.

55. Design and operationalize effective interagency project oversight and intra-agency coordination arrangements. Implementation of the project’s soft components such as road maintenance financing, axle load regulations, and gender equity requires an effective coordination platform for all concerned government stakeholders. It can mainstream innovations and help implementing agencies better align the use of loan funds with the rest of public spending. The adoption of an interagency oversight mechanism should be pursued in parallel to seeking the government’s approval to sign the loan agreement. Future projects should embed an operational arrangement to coordinate and synchronize project investment activities with the asset management function, including road maintenance. New projects could have a dedicated

33 Government of Azerbaijan. 2011. Decree of the Cabinet of Ministers of the Republic of Azerbaijan No, 142s . 3 June 2011. Baku; Government of Azerbaijan. 2011. Executive Order of the Ministry of Transport of the Republic of Azerbaijan No, 96/ü. 21 June 2011. Baku. The weighing stations were installed at km 49 (right side) and km 241 (left side) of the M2 highway (Baku–Alat–Georgia border), and at km 67 (left side) and km 84 (right side) of the M1 highway (Baku–Guba–Russian Federation border). 34 The quality of the project’s last audit report, covering transactions from the loan closing date on 28 September 2017 to the financial closing date of the Asian Development Fund loan on 30 April 2018, was adequate. This is because the auditor was selected based on a new arrangement adopted in March 2018; that is, auditors were selected from a list of eligible audit firms in Azerbaijan following the quality- and cost-based selection method (50:50). 35 Subject to the positive test results from the first tolling on M1 highway, application of tolling on other roads and their transfer to the private sector for operation and maintenance should be considered. Such expansion would free up resources and support the road sector’s long-term sustainability.

15

component on building capacity for road asset management, strengthening the governance framework, and expanding the scope of the State Road Maintenance Fund to public–private partnerships. ADB may consider providing attached technical assistance to help implement “soft” components on sector, institutional, or thematic (e.g., gender equality) covenants.

56. Further action or follow-up . Remedial measures should be employed to avoid delays in submission of audited financial statements in future projects. Some delays might be justified and will need to be consulted with ADB if revisions are proposed to the financial covenants. ADB should continue dialogue with the government and the AAY through technical assistance from ADB’s knowledge departments, focusing on vehicle axle load control, road maintenance funding and efficient contractual arrangements, road sector asset management, road safety and emission control, and gender equity. Regional technical assistance is ongoing from which the AAY can benefit.36

57. Timing of the project performance evaluation report. The project’s socioeconomic benefits are likely to better crystalize by the end of 2021, 1 year after the government completes and opens to traffic the M1 highway (the northern part of the north–south road corridor), now planned for the end of 2020. A project performance evaluation could be done in 2022–2023. By then, the project road will have operated for over 4 years, enabling a fair and comprehensive assessment.

36 ADB. 2019. Central Asia Regional Economic Cooperation: Knowledge Sharing and Services in Transport and Transport Facilitation (TA 9754-REG).

16 Appendix 1

DESIGN AND MONITORING FRAMEWORK

Design Summary Performance Indicators and Targets Project Achievements Impact By 2020 Greater economic development in Increase in trade with neighboring countries Achieved Azerbaijan and expanded to $2.7 billion (Baseline: $2.2 billion in 2006) Trade turnover with neighboring trade with neighboring countries reached $3.6 billion in countries 2018. Outcome By 2015 An efficient, safe, and Achieved sustainable southern 20% increase in freight traffic (Baseline: 206% increase to 15,532 million road corridor from Baku 7,536 million ton-km in 2005) ton-km in 2015 a

Achieved 5% increase in passenger traffic (Baseline: 218% increase to 23.825 billion 10,892 million passenger-km in 2005) passenger-km in 2015 a

Achieved Increased funding for road maintenance to Total funding was $185 million at least $150 million (Baseline: $55 million (AZN297 million) in 2017 in 2006)

Achieved 5% reduction in road traffic accidents 30% decrease to 2,220 road traffic (Baseline: 3,179 in 2005) accidents in 2015 a Outputs Partly achieved Sections B and C of four- Sections B (22.85 km) and C (16.4 km) of Sections B and C completed with lane expressway the Masalli–Astara expressway completed the delays, by September 2017 b between Masalli and by 2013 Astara on the southern Achieved corridor built, with access International roughness index (IRI) for IRI estimated at 2.0 after 2018 roads rehabilitated sections B and C of the expressway maintained at no higher than 5.0 within 5 years after project completion, compared with the average IRI value of 11.0 in 2006 km = kilometer. a Transport in Azerbaijan Statistical Yearbook 2018 . Baku. b Actual km-length output of the project is only 36.9 km, from km 22.1 to km 59 (see para 11). The remaining 2.4 km was financed out of the state budget one year after project completion. Source: Asian Development Bank and The State Statistics Committee of Azerbaijan.

Appendix 2 17

PROJECT COST AT APPRAISAL AND ACTUAL ($'000)

Appraisal Estimate Actual Foreign Local Foreign Local Item Exchange Currency Total Cost Exchange Currency Total Cost A. Base Costs 1. Infrastructure improvement component a. Subproject 1: Section B Road i. Civil works 92.00 0.00 92.00 97.94 0.00 97.94 ii. Land acquisition and resettlement 0.00 4.50 4.50 0.00 0.50 0.50 iii. Consulting services support 2.00 0.00 2.00 2.70 0.00 2.70 Subtotal 94.00 4.50 98.50 100. 64 0.50 101. 14 a. Subproject 2: Section C Road i. Civil works 72.50 0.00 72.50 84.42 0.00 84.42 ii. Land acquisition and resettlement 0.00 3.50 3.50 0.00 0.20 0.20 iii. Consulting services support 1.50 0.00 1.50 1.51 0.00 1.51 Subtotal 74 .00 3.5 0 77 .50 85.93 0. 20 86.13 Provision of taxes and duties 0.00 30.00 30.00 0.00 32.57 32.57 Total Base Costs 168.00 38.00 20 6.00 18 5.57 33.27 21 9.84 B. Contingencies 1. Physical and Price Contingencies 32.00 0.00 32.00 0.00 0.00 0.00 Subtotal (B) 32 .00 0.00 32 .00 0.00 0.00 0.00 Financing Charges during Construction 1. Interest and other charges 0.00 12.00 12.00 0.00 0.00 0.00 Subtotal (C) 0.00 12 .00 12 .00 0.00 0.00 0.00 Total (A+B+C) 20 0.00 50.00 25 0.00 18 6.57 33.27 219. 84 Source: Asian Development Bank estimates and Loan Financial Information System.

18 Appendix 3

PROJECT COST BY FINANCIER

Table A3.1: Project Cost at Appraisal by Financier

ADB Government Total Cost a % of Cost % of Cost Taxes and Amount Category Amount Category Amount Duties Item A. Investment Costs 1. Civil works 164.50 84.57% 30.00 15.43% 194.50 30.00 3. Land acquisition 0.00 00.00% 8.00 100.00% 8.00 0.00 4. Consultants 3.50 100.00% 0.00 100.00% 1.00 0.00 Subtotal (A) 168.00 81.32% 38.00 18.68% 203.50 30.00 Total Base Cost 168.00 81.32% 38.00 18.68% 203.50 30.00 B. Contingencies 1. Physical and Price 32.00 100.00% 0.00 00.00% 32.00 Subtotal (B) 32.00 100.00% 00.00 00.00% 32.00 C. Financial Charges during 00.00 0.00% 12.00 100.00% 12.00 0 Implementation

Total Project Cost (A+B+C) 200.00 80.00% 50.00 20.00% 250.00 30.00 % Total Project Cost 80.00% 20.00% 100% Note: Numbers may not sum precisely because of rounding. Sources: Asian Development Bank and Azeravtoyol State Agency estimates.

Appendix 3 19

Table A3.2: Project Cost at Completion by Financier

ADB Government Total Cost a % of Cost % of Cost Taxes and Amount Category Amount Category Amount Duties Item A. Investment Costs 1. Civil works 182.36 85.19% 31.70 14.80% 214.06 31.70 3. Land acquisition 0.00 00.00% 0.70 100.00% 0.70 0.00 4. Consultants 4.21 84.00% 0.88 16.00% 5.10 0.88 Subtotal (A) 186. 57 84.89% 33. 27 15.11% 219.8 4 32.5 8 Total Base Cost 186. 57 84.89% 33. 27 15.11% 219.8 4 32.5 8 B. Contingencies 1. Physical and Price 00.00 00.00% 00.00 00.00% 00.00 Subtotal (B) 00.00 00.00% 00.00 00.00% 00.00 C. Financial Charges during 00.00 0.00% 00.00 100.00% 00.00 0 Implementation

Total Project Cost (A+B+C) 186. 57 84.85% 33. 27 15.14% 219.8 4 32.5 8 % Total Project Cost 84.8 5% 15.1 4% 100% Note: Numbers may not sum precisely because of rounding. Sources: Asian Development Bank; Azeravtoyol State Agency estimates.

20 Appendix 4

DISBURSEMENT OF ADB LOAN AND GRANT PROCEEDS

Table A4.1: Annual and Cumulative Disbursement of ADB Loan Proceeds ($ million) Annual Disbursement Cumulative Disbursement Amount Amount Year ($ million) % of Total ($ million) % of Total 2012 00.00 00.00% 00.00 00.00% 2013 25.00 13.40% 25.00 13.40% 2014 32.20 17.26% 57.20 30.66% 2015 43.04 23.07% 100.23 53.72% 2016 48.04 25.75% 148.27 79.47% 2017 38.29 20.52% 186.57 100.00% 2018 0.00 0.00% Total 186.57 100.00% ADB = Asian Development Bank. Source: Asian Development Bank

Figure A4.1: Projection and Cumulative Disbursement of ADB Loan Proceeds ($ million)

Years 2012 2013 2014 2015 2016 2017 2018 Plan 0 8 48 116 180 Actual 0 25 57.2 100.24 148.24 186.57

Appendix 5 21

CONTRACT AWARDS OF ADB LOAN AND GRANT PROCEEDS

Table A5.1: Annual and Cumulative Contract Awards of ADB Loan Proceeds ($ million) Annual Contract Awards Cumulative Contract Awards Amount Amount Year ($ million) % of Total ($ million) % of Total 2012 00.00 0.00% 00.00 00.00% 2013 182.36 97.72% 182.36 97.72% 2014 00.00 0.00% 182.36 97.72% 2015 00.00 0.00% 182.36 97.72% 2016 4.22 2.28% 186.60 100% Total 186.6 100.0% ADB = Asian Development Bank. Source: Asian Development Bank.

Figure A5.1: Projection and Cumulative Contract Awards of ADB Loan Proceeds ($ million) Years 2012 2013 2014 2015 2016 2017 2018 Plan 0 80 165.5 165.5 180 Actual 0 182.36 182.36 182.36 186.57 186.57 186.57

22 Appendix 6

CHRONOLOGY OF MAIN EVENTS

Date Event

2006 9 November Concept paper 11–22 November ADB fact-finding mission

2011 20 July Project financing request received 30 August Management review meeting 14 December Loan approval

2012 27 July Loan agreement signed 23 October Loan effectiveness

2013 29 March Section B works contract signed 31 October Section C works contract signed

2014 30 November–9 December Midterm review mission

2016 21 July Design and supervision consultant contract transferred from tranche 1 to tranche 3 31 December First extension of loan closing date

2017 28–30 August Last review mission 28 September Second extension of loan closing date

2018 12 January Project completion date

2020 — Completion report approval — Completion report circulation

Source: Asian Development Bank.

Appendix 7 23

STATUS OF COMPLIANCE WITH LOAN COVENANTS

Reference in Loan Covenant Agreement Status of Compliance (a) The Borrower shall cause the Project to be carried out with Loan 2831 Complied. due diligence and efficiency and in conformity with sound Article IV The executing agency of the project was applicable technical, financial, business, and development Section 4.01 established and managed efficiently, practices. with engagement of financial, engineering, and other specialists. (b) In the carrying out of the Project and operation of the Project The Project was carried out in facilities, the Borrower shall perform, or cause to be performed, accordance with conditions set out in all obligations set forth in Schedule 5 of the loan agreement. Schedule 5 of the Loan Agreement. The Borrower shall make available, promptly as needed and on Loan 2831 Complied. terms and conditions acceptable to ADB the funds, facilities, Article IV The Ministry of Finance provided all services, land, and other resources, as required, in addition to Section 4.02 counterpart funds, staff, and other the proceeds of the Loan, for the carrying out of the Project and resources necessary to implement for the operation and maintenance of the Project facilities. projects as required under the loan agreement.

(a) Whenever applicable, in the carrying out of the Project, the Loan 2831 Complied. Borrower shall cause competent and qualified consultants and Article IV The Azeravtoyol State Agency (AAY) contractors, acceptable to ADB, to be employed to an extent Section 4.03 engaged qualified and competent and upon terms and conditions satisfactory to the Borrower and contractors and consultants ADB. acceptable to ADB in accordance with ADB guidelines and met all (b) The Borrower shall cause the Project to be carried out in requirements under the loan accordance with plans, design standards, specifications, work agreement. schedules and construction methods acceptable to the The AAY implemented the project in Borrower and ADB, as applicable. The Borrower shall furnish, accordance with the plan, design or cause to be furnished, to ADB, promptly after their specifications, and work schedule preparation, such plans, design standards, specifications and approved by the Bank. work schedules, and any material modifications subsequently made therein, in such detail as ADB shall reasonably request. The Borrower shall ensure that the activities of its departments Loan 2831 Complied. and agencies with respect to the carrying out of the Project and Article IV The AAY and its departments and operation of the Project facilities are conducted and Section 4.04 agencies adopted the latest coordinated in accordance with sound administrative policies international standards and sound and procedures. administrative policies and procedures. (a) The Borrower shall (i) maintain, or cause to be maintained, Loan 2831 Partially complied. separate accounts for the Project; (ii) have such accounts and Article IV The AAY maintained separate related financial statements audited annually, in accordance Section 4.05 accounts for the project and submitted with appropriate auditing standards consistently applied, by certified copies of audited accounts independent auditors whose qualifications, experience, and and financial statements to ADB. terms of reference are acceptable to ADB; (iii) furnish to ADB, Submission of audited project financial as soon as available but in any event not later than 6 months statements for 2014, 2015, and 2016 after the end of each related fiscal year, certified copies of such were significantly delayed by 6–8 audited accounts and financial statements and the report of the months. auditors relating thereto (including the auditors' management letter and opinion on the use of the Loan proceeds and compliance with the financial covenants of this Loan Agreement as well as on the use of the procedures for statement of expenditures), all in the English language; and (iv) furnish to ADB such other information concerning such accounts and financial statements and the audit thereof as ADB shall from time to time reasonably request.

(b) The Borrower shall enable ADB, upon ADB's request, to discuss the Borrower's financial statements for the Project and its financial affairs related to the Project from time to time with

24 Appendix 7

Reference in Loan Covenant Agreement Status of Compliance the auditors appointed by the Borrower pursuant to Section 4.05 (a). The Borrower shall enable ADB's representatives to inspect the Loan 2831 Complied. Project, the Goods and Works, and any relevant records and Article IV documents. Section 4.06 The Borrower shall ensure that any facilities relevant to the Loan 2831 Complied. Project are operated, maintained, and repaired in accordance Article IV with sound applicable technical, financial, business, Section 4.07 development, operational, and maintenance practices. The Borrower and the Project Executing Agency shall ensure Loan 2831 Partially complied. that the Project is implemented in accordance with the detailed Schedule 5 The project steering committee, which arrangements set forth in the FAM. Any subsequent change to Para 1 was envisaged in the facility the FAM shall become effective only after approval of such administration manual, was not change by the Borrower and ADB. In the event of any established. Instead, the government discrepancy between the FAM and this Loan Agreement, the held meetings as needed, usually provisions of this Loan Agreement shall prevail. chaired by the deputy prime minister with participation of representatives of agencies such as the Ministry of Transport (MOT), Azer Road Service (ARS) and/or AAY, Ministry of Finance, and Ministry of Economy. This led to deficiencies in interagency coordination, resulting in the weakened coordination and problem- solving. The Borrower shall ensure that the preparation, design, Loan 2831 Complied. construction, implementation, operation and decommissioning Schedule 5 The AAY implemented projects in of the Project and all Project facilities comply with (a) all Para 2 accordance with its applicable laws applicable laws and regulations of the Borrower relating to and regulations and ADB’s environment, health, and safety; (b) the Environmental Environment Policy. Regular reports Safeguards; (c) the EARF; and (d) all measures and were submitted to ADB. requirements set forth in the respective EIA and EMP, and any corrective or preventative actions set forth in a Safeguards Monitoring Report. The MOT shall, through ARS, ensure that the construction of Loan 2831 Complied. the Project road complies with technical specifications of the Schedule 5 The AAY adopted technical design, and that the construction supervision, quality control, Para 3 specifications of international and Project management are undertaken according to standards in implementation. internationally accepted standards and practices. The MOT shall ensure that sufficient funding and resources are Loan 2831 Complied. made available for implementation of the road maintenance Schedule 5 On 11 May 2016, the Cabinet of plan for 2008-2015. Para 4 Ministers approved a new organizational structure of the road maintenance administration. The road maintenance administration consists of 83 separate limited liability companies, all of them reporting to the AAY, including 7 in charge of all major highways, 10 for the greater Baku area, 10 for special purpose maintenance, 55 for all roads outside Baku except major highways, and 1 in charge of road bridges throughout the country.

Appendix 7 25

Reference in Loan Covenant Agreement Status of Compliance The MOT shall implement appropriate safety enforcement Loan 2831 Complied. measures in respect of the Project road and, through ARS, Schedule 5 The State Program of Azerbaijan monitor the accident rate and traffic volume after Para 5 Republic on Road Safety for 2012– commencement of the operation of the Project road. 2016 applies to all highways and city roads throughout the country. Road safety is enforced by Azerbaijan’s State Traffic Police (DYP) according the Chapter 4: Enforcement of Road Safety Rules (State Program of Azerbaijan Republic on Road Safety for 2012–2016). The DYP has appropriate facilities and traffic enforcement equipment to carry out roadside checks to ensure compliance with the road safety standards. The implementing agency submitted a project road safety plan to ADB and has undertaken final measures to comply with the plan. The MOT shall, through ARS, establish vehicle weighing Loan 2831 Not complied with. stations along the Project road to control the axle overloading Schedule 5 and formulate appropriate procedures and regulations to Para 6 As per the Decree of the Cabinet of enforce the axle load controls. Ministers of the Republic of Azerbaijan no. 214 S, 16 July 2015, 32 weighbridge stations were supposed to be installed in Azerbaijan, including three on the project road. This did not materialize as vehicle weighing stations were installed but not on the project road. The MOT shall, through the relevant agencies, enforce the Loan 2831 Complied. vehicle emission standards applicable on the territory of Schedule 5 Vehicle emission standards are Azerbaijan. Para 7 enforced by the DYP, who have appropriate equipment to carry out roadside checks to ensure compliance with vehicle emission standards legislation. The Borrower shall ensure that all land and all rights-of-way Loan 2831 Complied. required for the Project and all Project facilities are made Schedule 5 The land acquisition and resettlement available to the Works contractor in accordance with the Para 8 plan (LARP) was successfully schedule agreed under the related Works contract and all land implemented. In 2013, because of the acquisition and resettlement activities are implemented in additional land impacts, an addendum compliance with (a) all applicable laws and regulations of the to the LARP was prepared and fully Borrower relating to land acquisition and involuntary complied with. resettlement; (b) the Involuntary Resettlement Safeguards; (c) the LARF; and (d) all measures and requirements set forth in the respective LARP, and any corrective or preventative actions set forth in a Safeguards Monitoring Report Without limiting the application of the Involuntary Resettlement Loan 2831 Complied. Safeguards, the LARF, or the LARP, the Borrower shall ensure Schedule 5 The AAY attended to all related issues that no physical or economic displacement takes place in Para 9 in accordance with the agreed plan, connection with the Project until: and due diligence reports were (a) compensation and other entitlements have been provided prepared by an external expert in to affected people in accordance with the LARP; accordance with the LARP and (b) a comprehensive income and livelihood restoration program reviewed by ADB has been established in accordance with the LARP; and (c) a due diligence report prepared by an external expert in accordance with the LARP has been cleared by ADB.

26 Appendix 7

Reference in Loan Covenant Agreement Status of Compliance The Borrower shall make available necessary budgetary and Loan 2831 Complied. human resources to fully implement the EMP, and the LARP. Schedule 5 Necessary funding and human Para 10 resources were provided during implementation of the environmental management plan (EMP) and the LARP. The Borrower shall ensure that all bidding documents and Loan 2831 Complied. contracts for Works contain provisions that require contractors Schedule 5 EMP recommendations were to: Para 11 incorporated in the bidding documents (a) comply with the measures and requirements relevant to the by the design consultants and contractor set forth in the EIA, the EMP, and the LARP (to the compliance was closely monitored by extent they concern impacts on affected people during the supervision consultant. The works construction), and any corrective or preventative actions set out contracts include provision for the in a Safeguards Monitoring Report; contractors to prepare contracts (b) make available a budget for all such environmental and specific to the EMP. Monitoring reports social measures; were regularly submitted to ADB. (c) provide the Borrower with a written notice of any unanticipated environmental, resettlement or indigenous peoples risks or impacts that arise during construction, implementation or operation of the Project that were not considered in the EIA, the EMP, or the LARP; (d) adequately record the condition of roads, agricultural land, and other infrastructure prior to starting to transport materials and construction; and (e) fully reinstate pathways, other local infrastructure, and agricultural land to at least their pre-project condition upon the completion of construction. The Borrower shall do the following: Loan 2831 Complied. (a) submit semiannual Safeguards Monitoring Reports to ADB Schedule 5 Semiannual reports were submitted in and disclose relevant information from such reports to affected Para 12 a timely manner to ADB. ADB’s persons promptly upon submission; safeguard specialist visited project (b) if any unanticipated environmental and/or social risks and sites during the review mission in impacts arise during construction, implementation, or operation August 2014 to assess the of the Project that were not considered in the EIA, the EMP, or implementation process. the LARP, promptly inform ADB of the occurrence of such risks or impacts, with detailed description of the event and proposed corrective action plan; (c) prior to the award of the Works contract, engage qualified and experienced external expert or qualified NGO under a selection process and terms of reference acceptable to ADB, to verify information produced through the Project monitoring process, and facilitate the carrying out of any verification activities by such external experts; and (d) report any actual or potential breach of compliance with the measures and requirements set forth in the EMP or the LARP promptly after becoming aware of the breach. The Borrower shall ensure that no proceeds of the Loan are Loan 2831 Complied. used to finance any activity included in the list of prohibited Schedule 5 investment activities provided in Appendix 5 of the SPS. Para 13 The Borrower shall cause MOT and ARS to ensure that Loan 2831 Complied. contractors engaged under the Project (i) comply with all Schedule 5 Local men were employed by the applicable labor laws; (ii) use their best efforts to employ Para 14 contractors to work on the project site women and local people, including disadvantaged people, while female workers were mainly living in the vicinity of the Project; (iii) provide equal pay to men recruited for auxiliary support jobs at and women for work of equal type; (iv) provide and adequately the campsite office. equip first-aid, health and sanitation, and personal culturally Adequate sanitation and culturally appropriate hygiene facilities for male and female workers at appropriate personal hygiene facilities the Project site; (v) maximize female training and employment; for male and female workers were

Appendix 7 27

Reference in Loan Covenant Agreement Status of Compliance (vi) conduct an information and education campaign for provided at the project site and construction workers as part of the health and safety program contractor’s camp. The AAY strictly at campsites and adjacent communities during Project followed the child labor law. implementation; and (vii) abstain from child labor. Relevant contracts financed under the Project must include specific clauses on these undertakings. The Borrower shall cause MOT and ARS to ensure that within Loan 2831 Partially complied. 6 months of the Effective Date, the gender action plan is agreed Schedule 5 A gender action plan was developed in with ADB, that it is implemented in a timely manner over the Para 15 2014 for Loan 2354 and Loan 2831, Project period, and that adequate resources are allocated for which covered sections A, B, and C of this purpose. the Masalli–Astara highway. About 50% of the actions envisaged in the gender action plan have been implemented. Gender-sensitive road safety awareness training was conducted for all the communities in 21 villages along the road in sections A, B, and C during May 2015. More than 50% of the participants were female residents of these villages. The training sessions were conducted by the social safeguards specialist of the design and supervision consultant. Women and men were given equal opportunities during the construction phase. The Borrower shall ensure that the counterpart funding, land, Loan 2831 Complied. rights-of-way, facilities, and human resources required for the Schedule 5 The Ministry of Finance provided the Project, including cost overruns, are promptly provided for Para 16 necessary counterpart funds, staff, purposes of successful Project implementation. and other resources in a timely manner as required under the loan agreement The Borrower shall cause MOT and ARS to be adequately Loan 2831 Complied. staffed and equipped in a manner acceptable to ADB. Schedule 5 Para 17 The Borrower, MOT and ARS shall (i) comply with ADB’s Loan 2831 Complied. Anticorruption Policy (1998, as amended to date) and Schedule 5 acknowledge that ADB reserves the right to investigate directly, Para 18 or through its agents, any alleged corrupt, fraudulent, collusive or coercive practice relating to the Project; and (ii) cooperate with any such investigation and extend all necessary assistance for satisfactory completion of such investigation.

The Borrower, MOT and ARS shall cause the Project Executing Loan 2831 Complied. Agency and ARS to ensure that the anticorruption provisions Schedule 5 acceptable to ADB are included in all bidding documents and Para 19 contracts, including provisions specifying the right of ADB to audit and examine the records and accounts of the executing and implementing agencies and all contractors, suppliers, consultants, and other service providers as they relate to the Project. In the event that the Borrower proposes (i) any change in Loan 2831 Complied. ownership of the Project, (ii) any sale, transfer, or assignment Schedule 5 of interest or control in the Project, or (iii) a lease or contract or Para 20 otherwise material modification of MOT or ARS’s functions or authority over implementation and operation of the Project, the Borrower shall, at least 6 months prior to the implementation of such a proposal, consult with ADB and obtain its consent. The

28 Appendix 7

Reference in Loan Covenant Agreement Status of Compliance Borrower shall ensure that any such proposal is carried out in a legal and transparent manner.

The MOT shall prepare and submit to ADB quarterly progress Loan 2831 Complied. reports for the Project, which will include (i) a narrative Schedule 5 Reports were submitted on time. description of progress made during the reporting period; (ii) Para 21 changes in the implementation schedule; (iii) problems or difficulties encountered; and (iv) activities to be undertaken in the next reporting period. The MOT shall, through ARS, and with the assistance of the Loan 2831 Complied. supervision consultant engaged under Project 1 of the Facility, Schedule 5 A project performance management monitor and evaluate the Project impacts to ensure that the Para 22 system (PPMS) was established and Project facilities are managed effectively, and the Project the initial draft submitted to ADB in benefits maximized. The MOT shall, through ARS, collect the October 2014. The PPMS was data agreed with ADB at the inception, on Project completion, updated regularly and reports shared and 3 years after Project completion. with ADB. Within 3 months of the Effective Date, the MOT shall, through Loan 2831 Complied. ARS, establish a Project performance management system in Schedule 5 form and substance acceptable to ADB, in accordance with the Para 23 Project performance indicators agreed with ADB. The MOT shall, through ARS, undertake periodic Project Loan 2831 Complied. performance review in accordance with the Project Schedule 5 performance management system to evaluate the scope, Para 24 implementation arrangements, progress, and achievements of objectives of the Project and the Facility. ADB and the MOT shall meet regularly as necessary, and at Loan 2831 Complied. least annually to review the progress under the Project. In Schedule 5 Regular meetings were organized with addition, the Ministry of Economic Development, the Ministry of Para 25 stakeholders to discuss the progress Finance, MOT, ARS and ADB shall jointly undertake a midterm of the project and any changes to review of the Project in the year 2013, which will focus on (i) implementation arrangements or implementation progress; (ii) road subsector reform and remedial measures required to be performance; (iii) performance of consultants and contractors; undertaken to achieve the project (iv) status of compliance with the understandings in the FFA objectives. ADB and the AAY fielded and the covenants in the Loan Agreement; and (v) the need for regular project review missions to any midcourse changes in the scope or schedule of Project to review the project's implementation ensure full achievement of its impact. status, discuss pending issues, and agree on actions to be taken for improvement.

Appendix 8 29

ENVIRONMENTAL SAFEGUARD AUDIT SUMMARY

A. Introduction and Project Background

1. The Masalli to Astara (border with Iran) section of the M3 motorway was constructed as part of the $600 million Asian Development Bank (ADB) Multitranche Financing Facility (MFF) for the Road Network Development Program (RNDP) in Azerbaijan. This segment comprised two sections: (i) Masalli to Lenkaran, from kilometer (km) 142.9 to km 165.05 (Project 1); and (ii) Lenrakan to Astara, from km 165.05 to km 201.9 (Tranche 3). Both projects were completed on 28 September 2017. The financial closure for Project 1 was on 12 January 2018, and that of Tranche 3 on 30 April 2018.

2 During the project preparatory technical assistance in late 2007, an environmental impact assessment was prepared for the entire section under ADB’s category A. During implementation, Project 1 was to be implemented in accordance with Azerbaijan’s applicable laws and regulations and ADB’s Environment Policy (2002). Tranche 3 was to be implemented in accordance with Azerbaijan’s applicable laws and regulations and ADB's Safeguard Policy Statement (2009).

3 The executing agency for the project was the Azeravtoyol State Agency (AAY) (formerly the Azerbaijan Road Service). The construction supervision consultant (CSC) was Nippon Koei UK Ltd., supervising the projects with the following contractors:

Project Sections and Contractors

Tranche Project Contractor (i) Azerinşaatservice LLC (first contractor, Azerbaijan) Section A (0–22.15 km) (ii) KOLİN İnşaat, Turizm Sanayi ve Tranche 1 Ticaret A.Ş. (second contractor, Turkey) Lenkaran Bridge (iii) Ganja Korpu Tikinti-2 OJC (km 31+555 - 32+350) (Azerbaijan) Tranche 3 (iv) Özgün Construction Industry and Section B (22.15 - 45.00 km) Trading (Azerbaijan) Section C (45.00 – 61.4 km) (v) Polat LLC (Turkey)

4 One important feature in section B was part of the Hirkan Forest Reserve, which not long ago covered almost the whole Lenkaran Lowland. Owing to the reserve’s protected status, a biodiversity specialist was deployed by the CSC before construction to study the area and propose measures to mitigate any adverse impact to its ecosystem. The result of the specialized study was considered in the construction activities.

5. As part of project completion report’s preparation, an independent environmental audit was conducted to ensure that all environmental safeguards had been fully implemented and that there were no unresolved issues. A field mission to Azerbaijan on 10–15 November 2019 collaborated with the Azerbaijan Resident Mission (AZRM) and the project implementation unit (PIU). The field visit was done on 12 November 2019.

30 Appendix 8

B. Review of Project Environmental Documents

6. The AZRM provided the following documents for projects 1 and 3: the environmental impact assessment (EIA), biannual environmental monitoring reports, monthly progress reports, and other project-related documents. Safeguard specialists used them to become familiar with the issues the project faced during construction and to inform the environmental audit.

7. The EIA for the Masalli–Astara section was done in 2007 during the technical assistance, which produced the EIA with the corresponding environmental management plan (EMP). During the construction stage, a basic requirement for the contractors was to draft their site-specific environmental management plan following the requirements of the project EMP. Their EMPs were reviewed by the CSC in accordance with the stipulated project requirements.

8. The CSC’s biannual environmental monitoring reports for projects 1 and 3 provided information on implementing environmental safeguards for section A, Lenkaran Bridge, section B, and section C. Public complaints and grievances (mainly on local access impacts and dust) were recorded in grievance registers. Almost all project-related issues were resolved and considered closed before the project ended. The documents were detailed, and numerous nonconformance letters were sent to the contractors, showing that the CSC’s performance was satisfactory.

9. Based on the assessment of the environmental monitoring and management activities, the first contractor for section A (Azerinşaat) and the Lenkaran Bridge (Ganja Korpu) failed to fulfill the project’s environmental aspects. The second contractor for section A, Kolin, did far better than Azerinşaat and the site conditions improved. The final report revealed that the initial issues in section A were partly resolved at the end of the project. For the Lenkaran Bridge, the slopes underwent gradual natural revegetation and were no longer a concern at final inspection.

10. The implementation of the environmental safeguard requirements for sections B and C was satisfactory. At the final part of construction, the project road vicinity met the environmental requirements in a generally acceptable manner, and the documentary portions were satisfactory. The section C contractor was better than the other contractors at fulfilling environmental requirements.

C. Inspection of the Project Road

11. The field mission inspected the project road on the spot to determine the degree to which impacted areas had been restored as per project requirements. The mission visited the Masalli– Astara project road on 11 November 2019 with the assistance of the AYY and PIU. The mission inspected section A, interchange A1, section B, Lenkaran Bridge, section C, and the Astara interchange. The mission also inspected some sites accessible to the project road, such as the former contractors’ camps and materials sources.

12. The mission made the following general observations: (i) Vegetation, mainly grass and bushes, were growing along the roadsides, bridge embankments were fair to good. Most of the sites previously stripped for construction were covered with grass. (ii) The corridor was generally clean, with no remnants of construction material. (iii) Embankments were generally stabilized with grass cover against surficial erosion. (iv) Former camp sites were generally leveled, with few floorings remaining, possibly with the owner’s permission.

Appendix 8 31

(v) The inspected materials sites were still being operated by other entities, which might have accounted for the sites’ current condition.

D. Conclusions and Recommendations

13. The environmental audit generally evaluated the project’s compliance with prescribed environmental safeguards. Sufficient information was gathered and used to examine how the contractors implemented the required environmental measures and responded to local issues under the supervision of the CSC. The final environmental monitoring reports provided a good summary of all the contracts’ environmental aspects.

14. Based on the documentary review, the section C contractor (Polat) performed much better than the rest of the contractors. The first contractors in section A (Azerinşaat) and in Lenkaran Bridge (Ganja Korpu) had internal management issues and were deemed noncompliant. The replacement contractor in section A (Kolin) provided better environmental management and environmental conditions were acceptable in the end. The section B contractor (Özgün) and the section C contractor (Polat) maintained a generally good level of environmental management and documentation. Most of the issues encountered during construction were mainly on-site contamination, dust control, solid and construction waste management, and general housekeeping.

15. Based on the site visit, the impacted areas had recovered well, a manifestation of the satisfactory implementation of the project environmental management plan. No major adverse impacts were discovered, the environment was generally reinstated to acceptable levels, and all raised issued had been resolved.

16. Among the lessons learnt from the project are as follows: (i) Environmental compliance stems from a well-prepared site-specific EMP. It should be an indispensable requirement before any construction starts. (ii) Supportive management is necessary to have satisfactory level of environmental compliance. (iii) Some sanctions should be introduced in the contract as part of the enforcement mechanisms.

32 Appendix 9

GENDER ACTION PLAN IMPLEMENTATION MATRIX

Multitranche Financing Facility Road Network Development Program, Projects 1 and 3

Output 1: Enhanced road safety and improved travel condition Status

Provide gender-sensitive road At least 50% of participants involved in The training was held on 28 safety awareness training to all awareness training are female September 2016 for community members communities from Yukhari Nuvedi, Burjali, and Shiyakaran villages. A total of 21 residents participated, 12 (57%) of whom were female. Improve the road network with Road built to international safety Complied to make it safer than alternative standards routes Create roadside rest areas with Separate toilets for women in place and Not complied separate sanitary facilities for separate baby changing and/or feeding female travelers, including rooms in place breastfeeding and changing rooms for women travelling with babies Output 2: Improved livelihoods for residents of local villages

Give women equal opportunities Allocation of at least 20% of local Partially complied. to be employed during the construction workforce jobs to women, Involvement of women in the construction phase with equal 10% of whom are heads of households. construction industry had pay Equal pay for equal work at all grades, been limited (7.1% in 2018), where both genders are employed doing mainly because of the work of equal value remote location of the projects, the nature of the works, and limited availability of construction- skilled female professionals in remote locations. Urban projects and offices, including the Azeravtoyol State Agency office, have fair gender representation. Announce recruitment Culturally appropriate recruitment Complied opportunities and display announcements and notices to attract recruitment notices widely, female workers disseminated via local clearly targeted at women as media, such as newspapers and radio, or well as men displayed in local facilities frequented by women, before local hiring started

Provide basic facilities for Separate toilets for women Separate toilets were female and male construction provided for women and workers at construction sites men at the construction sites.

Appendix 9 33

Conduct training on financial At least 100 women participate in A total of 100 women—from management, entrepreneurship, trainings; at least 20 women draw up Lenkaran (40), Masalli (30), and business development business plans for micro or small Astara (10), and Lerik (20)— enterprises were trained. Training covered topics related to how to start and expand business; participants developed 38 business plans. The activities were implemented by the Sidus Group. During the project, out of 70 women who were not entrepreneurs, 38 (54%) started their own businesses. Of the 38 trainees, 12 received loans for $2,000 from the Molbulaq Non-Banking Credit Organization. Businesses the women engaged in were mostly small beauty salons, small hair salons, computer skills and language learning courses, cleaning services, babysitter services, and small catering services.

Of nine businesswomen with existing businesses, five (55%) expanded their business operations in agriculture farming, entertainment centers, and cafés. Three of five women each took out a loan of $5000 of from Molbulaq Non-Banking Credit Organization. Inform potential female Announcements Various methods were used employees about training to inform women about opportunities before recruitment. training opportunities. Disseminate and display notices Among them were about training opportunities in announcements to local local media and in local facilities nongovernment frequented by women organizations and local women associations. Output 3: Reduction in road construction–related health and safety threats to local females

Provide awareness training on At least one STI and drug and/or human Trainings are conducted on risk of sexually transmitted trafficking awareness-raising training continuous basis infections (STIs) and human event conducted for all project staff trafficking and how to avoid members and for each civil works contract package

34 Appendix 9

and/or reduce threats of Four STI (including HIV/AIDS) and drug At least one training was exposure to these risks and/or human trafficking awareness- held per month. On average, raising training events conducted for 8–10 females participated as residents of all project sites participated in per the attendance sheets by women per civil works contract from the project monthly package (minimum age to attend: 16 progress reports. years) Two women-only STI (including HIV/AIDS) At least one training was and drug and/or human trafficking held per month, with at least awareness-raising training events per civil 10–12 participants. works contract package conducted for women from project-affected villages

Output 4: Gender-related institutional capacity strengthening

Preparation and dissemination GAP manual known and implemented by A manual was written in the of the gender action plan (GAP) key managers local language. implementation manual The design and supervision One SSS recruited for 2 years for Complied consultant (DSC) to recruit intermittent time inputs during the Social Safeguards Specialist completion phase SSS) to support capacity in implementing and monitoring the GAP Ensure inclusion of gender Clause requiring gender sensitivity Complied. sensitivity in the terms of prepared and included as a nonnegotiable Staff members to be recruited reference for all new staff requirement in the template to prepare should have undergone members and consultants of the terms of reference for all new staff gender sensitivity training DSC and the Azer Road Service members and consultants and/or mentorship provided by the social development specialist in the contractors’ team. Ensure the use of gender- Gender disaggregation required by the Not complied sensitive indicators, sex- PPMS monitoring plan disaggregated data in the project performance management system (PPMS) monitoring and presentation of results

Appendix 10 35

ECONOMIC REVALUATION

Multitranche Financing Facility for the Azerbaijan Road Network Development Program, Tranches 1 and 3

E. Background and Scope

1. The principal output of the two tranches was about 59 kilometers (km) of limited-access four-lane road on a new alignment between a point south of Masalli via the city of Lankaran to Astara on the Iran border. 1 The project included the Lankaran Bridge and vehicle weighing and road maintenance equipment. An additional output was rehabilitation of 39.5 km of rural roads in the Ganja region. The four-lane road and the rural roads projects are entirely independent in economic terms and are, therefore, reevaluated separately.

2. The rationale for the Masalli–Astara road was to add road transport capacity to the international corridor between Baku and Iran; in 2010 the Astara border-crossing point (BCP) was reported to be Azerbaijan’s busiest international road border.2 The principal expected source of benefits was the saving in road user costs arising from diverting traffic from the old two-lane M3 road to the new alignment (also referred to as the M3). At the time of appraisal, the new road was assumed to open in 2010, but substantive construction started on various dates from 2011 to 2016 so the road was completed in September 2017 and formally opened in September 2018.

3. The rationale for the rural roads was to ensure continued and improved access from Ganja to the M2 east–west highway between Azerbaijan and Georgia. The roads were upgraded from categories III and IV to category II. They were constructed in 2011–2012.

4. At the time of the report and recommendation of the President (RRP) in 2007, buoyed by high oil prices, gross domestic product (GDP) was forecast to rise by 9.0% per year from 2008 to 2012, falling to 8.1% from 2013 to 2027. However, oil prices fell sharply in 2008–2009, rose from 2011 to 2014 but fell again in late 2014 and have since remained low. GDP growth was 5.5% per year from 2008 to 2012 and just 1.2% from 2013 to 2018. The drop in GDP rendered the 2007 traffic forecasts overoptimistic. At the time of the RRP the manat was trading at $1 = AZN0.85, but after floating in December 2015 it fell to its current parity of $1 = AZN1.7. This has had an adverse effect on the dollar value of journey time savings.

5. The city of Lankaran (2019 population of 229,000) is midway between the start and end points of the M3. It is the center of an important agricultural area known for tea and fruit. The coastal towns north and south of Lankaran attract domestic tourism.

F. Economic Analysis at Appraisal

6. The Masalli–Astara project was appraised at the time of the original RRP in 2007 and updated, in support of a periodic financing request PFR), in 2011. At the time of the original appraisal, based on 2006 traffic counts and 2006 origin and destination surveys, about 75% of traffic was assumed to be diverted to the new road. Normal corridor traffic was forecast to grow 8% annually over the evaluation period. Completion of the entire north–south road link from the

1 The intention had been to take the M3 to a new BCP, but the parties could not agree on its location and the final 2.8 km to the border was scrapped. 2 In tonnage terms. See T. Ziyadov. 2012. Azerbaijan as a Regional Hub in Central Eurasia . Baku: Azerbaijan Diplomatic Academy.

36 Appendix 10

Russian Federation to Iran would increase traffic by a further 10.5%. Project financial cost at 2007 prices (excluding taxes, duties, price contingencies, and financing costs) was $176 million (about $210 million at 2018 prices) and the economic internal rate of return (EIRR) 12.8%. Benefits included vehicle operating cost (VOC) and journey time savings, and safety benefits. The new road was expected to open in 2010.

7. The 2011 update used 2008 traffic and the same 2006 origin and destination survey results to estimate diversions. Traffic from 2006 to 2008 had not grown as expected at the time of the RRP and forecast traffic growth was reduced to 5% per year. Despite an increase in initial financial cost to $286 million (at 2010 prices, excluding taxes, duties, and price contingencies), the EIRR was little different at 12.9%. The 2011 evaluation estimated much higher journey time and crash cost savings than those at the time of the RRP. The rural roads were not subject to appraisal, while Lankaran Bridge was treated as an integral part of the Masalli–Astara project and not appraised separately.

G. Reevaluation of Masalli–Astara Road

1. Sectioning for Analysis

8. For economic analysis purposes the new M3 is divided into four sections from north to south, each connecting two interchanges (ICs). Section A (22.1 km) from IC1 (with the old M3 south of Masalli town) to IC2 (local road that connects Girdani to Lankaran), section B1 (7.9 km) from IC2 to IC3 (R48 junction), section B2 (5.3 km) from IC3 to IC4 (old M3 connecting to Lankaran south), and section C (23.8 km) from IC4 to IC5 (old M3 about 5 km from Astara town and 7 km from the BCP).

2. Demand Estimates

9. Traffic data exist for the old M3 from 2005 to just before the opening of the new M3 in 2018. In preparation for the reevaluation, 7-day 24-hour classified counts were undertaken at four sites along the new M3 from 26 September to 2 October 2019. The data make it possible to estimate historic traffic growth in the corridor and diversions from the old to the new M3. Table A10.1 summarizes average daily traffic (ADT) estimates on the old M3.

Table A10.1: Motorized Traffic Demand on the Old M3, 2005–2018

Year Location Small Buses and 3–4-Axle >4-Axle ADT Passenger Medium Trucks Trucks (vehicles/day) Vehicles Goods Vehicle s 2005 a Liman (Lankaran N) 4,756 2005 a N of Lankaran 4,292 2006 b Masalli–Lankaran 3,952 477 301 265 4,995 2006 b Lankaran–Astara 3,136 274 191 174 3,775 2008 c Masalli–Lankaran 4,870 422 247 266 5,805 2008 c Urga–Lankaran 3,936 755 579 478 5,748 2008 c Lankaran urban (N) 5,785 773 252 170 6,980 2008 c Lankaran urban (S) 7,847 715 342 234 9,138 2008 c Lankaran–Astara 6,030 591 371 262 7,254 2018 d Masalli–Lankaran 4,733 646 189 259 5,827 2018 d Lankaran–Astara 3,587 592 54 187 4,420

Appendix 10 37

ADT = average daily traffic. Note: Values shown have not been adjusted for seasonal effects. a World Bank. 2005. Azerbaijan Second Highway Project: Project Appraisal Document . Washington, DC. b Asian Development Bank (ADB). 2007. Proposed MFF for Republic of Azerbaijan Road Network Development Program . Manila. c ADB. 2011. Periodic Financing Request Report: Azerbaijan Road Network Development Program, Proposed Tranche 3. Manila. d Sheladia Associates for ADB and Azeravtoyol State Agency. 2018. Draft FSR for Installation of a Toll Collecting System along the M3 Alat–Astara Highway.

10. The salient features of Table 1 are the following: (i) Except in 2008, the ADT on the section south of Lankaran was about 75% of the traffic to the north. (ii) Growth since 2005 has been exceedingly low, typically 1%==2% per year (compared with GDP growth of 3%–9%). (iii) Small and medium-sized vehicles have been responsible for most of the recorded growth. (iv) The absolute volume of the largest trucks (>4 axles) has remained fairly constant at about 180–250 vehicles/day. The strongest historic traffic growth was from 2005 to 2008, when annual rates of 9%–20% were recorded (and GDP growth was 23%–25%).

11. Separately, truck traffic at the Iran–Azerbaijan BCP was put at 150 per day in 2005 3 and 180 per day in 2010, consistent with Table 1 data.

12. Table A10.2 shows the ADT estimates for the new M3 derived from the counts in September 2019. These values are used to generate with-project base-year traffic on the project road.

Table A10.2: Motorized Traffic Demand on the New M3, 2019 (vehicles/day)

Cars Pickups Minibuses Medium- 2-Axle 3–4-Axle Art Average Daily Location Section sized Trucks Trucks Trucks Traffic Bus es Km 142.9 A 4,119 151 319 177 17 205 274 5,262 Km 165.05 B1 2,730 69 146 80 11 174 226 3,436 Km 184.4 C 2,116 44 106 54 14 43 203 2,580 Weighted 3,002 90 195 106 14 133 234 3,774 Average km = kilometer. Source: Asian Development Bank estimates.

13. The drop in traffic in section B1 suggests that about 1,800 vehicles/day turn off the new M3, either east to Lankaran or to join the R48 westbound, which terminates at the small town of Lerik, 44 km west of Lankaran.

14. By comparing Table A10.1 without-project corridor traffic measured in 2018 on the old M3 with Table A10.2 traffic on the new M3, it is possible to estimate the percentage of corridor traffic that diverts to the new road (Table A10.3). The diverted percentage is about 80% when estimated using traffic weighted by distance,4 compared with about 76% at the time of the RRP and about

3 TERA International for ADB. 2006. TA4684: Preparing the Southern Road Corridor Improvement Project. 4 The with-project network is much longer than the without-project network, which distorts diversion estimates based on distance-weighted averages.

38 Appendix 10

60% at the time of the PFR. Weighted average diverted traffic in 2019 was 3,770 vehicles/day, 64% of the 5,900 vehicles/day forecast for 2019 at the time of the PFR (Table A10.5).

Table A10.3: Diverted and Corridor Traffic, 2018–2019

Average Vehicles per Corridor and/or Old M3 Traffic Section Day a Corridor without project IC1–LC LC–IC4 IC4–IC5 IC2–LC

Kilometers 26.6 9.8 24.5 6.2 ADT (vehicles/day)a 5,830 4,420 4,420 1,000 c 4,600 With-project post-diversion ADT 570 980 1,840 2,820 c 1,250 (vehicles/day)b Diverted Traffic (new M3) New M3 (with project) IC1–IC2 (A) IC2–IC3 (B1) IC3–IC4 (B2) IC4–IC5 (C)

Kilometers 22.1 7.9 5.3 23.8 ADT, vehicles/day 5,260 3,440 3,440 2,580 3,800 ADT = average daily traffic, IC = interchange, LC = Lankaran center. a Taken as old M3 traffic before 2019. b Traffic on the old M3 after diversion. c 1,000 vehicles/day local traffic added to this section both in the without-project and with-project post-diversion cases. Source: Asian Development Bank estimates.

15. Forecast annual GDP growth to 2024 is 2.4%–2.7%. In 2006–2008 passenger and small vehicle growth was about the same as GDP growth; for these vehicle classes, future growth is taken as 2.5% for 2019–2028, thereafter dropping to 2%. For medium-sized and heavy goods traffic, regional trade is an additional explanatory factor. Trade turnover with Iran, the Russian Federation, and Georgia in 2011–2018 fell to a minimum in 2015 but has since recovered, growing at 12% annually in 2016–2018 (in nominal dollar terms). Growth in trade with Iran, particularly relevant to demand for the project road, has grown even more strongly, at 56% in 2016–2018. Medium-sized and heavy goods and truck-trailer traffic are, therefore, forecast to grow at 3.8% per year (1.5 times GDP growth) to 2028, thereafter dropping to 3.5%.

Table A10.4: M3 Corridor Forecast Traffic Growth Rates

Period Cars Pickups Minibuses Medium- 2-Axle 3–4-Axle Art Trucks sized Bus es Trucks Trucks 2019–2028 a, b 2.5% (1.0) 2.5% (1.0) 2.5% (1.0) 2.5% (1.0) 3.8% (1.5) 3.8% (1.5) 3.8% (1.5) 2028–2042 b 2.0% (0.8) 2.0% (0.8) 2.0% (0.8) 2.0% (0.8) 3.5% (1.4) 3.5% (1.4) 3.5% (1.4) a Zero growth assumed before 2019. b Elasticities shown in parentheses. Source: Asian Development Bank estimates.

16. Table A10.5 compares revised forecasts on the project road with those made at the time of the PFR.

Table A10.5: Traffic Forecasts: Weighted Daily Average Traffic on the New M3 (vehicles/day)

Year 2019 2025 2030 2037 This reevaluation 3,774 4,411 4,979 5,790

Appendix 10 39

Year 2019 2025 2030 2037 Periodic financing 5,886 7,767 9,787 13,525 request, 2011 Source: Asian Development Bank estimates.

3. Economic Costs

17. Investment costs . Economic costs include (i) capital investment, which includes civil works, consulting services, and land acquisition and resettlement; and (ii) road maintenance. Costs related to taxes, duties, and financing charges during implementation have been excluded, as was the case at appraisal. Table A10.6 breaks down the actual investment costs of both projects. (Actual financial costs are incremental and exclude taxes, duties, and interest during construction.)

18. All predicted project costs and benefits are measured in 2018 economic prices expressed in US dollars. Traded goods are measured at world prices and nontraded inputs at domestic prices less indirect taxes multiplied by a standard conversion factor estimated at 0.98.5 A shadow wage rate factor (SWRF) of 0.96 was estimated and applied to unskilled labor used in road construction. A SWRF of 1.0 was applied to skilled and professional labor.

19. A residual value equivalent to 20% of the investment cost—estimated by applying the straight-line depreciation method to individual project items based on assumed lifespans—is included in the economic analysis.

Table A10.6: Actual Investment Costs of Tranches 1 and 3 Tranches Component Financial Cost a Economic Cost b (km) ($ million) $ million $ million/km

1 and 3 59 M3 sections A–Cc 333.7 341.3 5.79 22.1 M3 section A 151.3 155.4 7.0 36.5 M3 sections B and C 182.4 185.9 5.1 39.5 Rural roads 13.6 14.0 0.35 Construction 13.9 14.3 supervision

Land acquisition and 2.9 3.0 resettlement Total 364 373 km = kilometer. a Current prices excluding taxes and duties. b Constant 2018 economic prices. c Civil works, including Lankaran bridge and weigh stations. Source: Asian Development Bank estimates.

20. Maintenance costs . Identical maintenance regimes were assumed for both the project and without project cases of the M3, as shown in Table A10.7 below. Tranche 1 rural roads used the maintenance interventions shown below in the WP case, but in the WOP scenario no overlay was included.

5 Using the ADB simplified method based on merchandize imports of $9.200bn (CIF basis), exports of $21.03bn (FOB basis) and taxes on trade of $649m (averages in current $ for 2013-17 from Azeri national statistics and World Bank data) 6 An approximation based on the ratio of rural and urban incomes from all sources for 2016 (Azeri National Statistical Service: household income per head, 2016)

40 Appendix 10

Table A10.7: M3 Maintenance Interventions Intervention Unit Cost Works Effect Financial Economic Routine maintenance $4,720/km $3,850/km Pothole patching (annual) Elapsed time 6 months. 80% $16.5/m 2 $13.6/m 2 patched Crack sealing (annual) 80% of transverse thermal and 80% $5.3/m 2 $4.4/m 2 of wide cracks repaired Edge break repair (annual) $14.2/m 2 $11.7/m 2 90% repaired 40 mm overlay (IRI>10 and $29.5/m 2 $24.5/m 2 HDM-4 default interval of ≥ 8 years triggers) HDM = Highway Design and Maintenance, IRI = international roughness index, km = kilometer, m 2 = square meter. Source: Asian Development Bank estimates.

21. Economic benefits Using standard HDM-4 software the economic benefits considered at re-evaluation were (i) VOC savings and (ii) time savings. The data needed to make reliable estimates of crash cost savings were not available. Detailed without-project condition data were also not available. At appraisal in 2006 put international roughness index (IRI) at 9–11, with the southern section at the upper end of this range. For this reevaluation, initial (2014) without-project IRIs were assumed to be 12 on the southern section (IC4–IC5) and 11 elsewhere. The with- project IRI was set to 2.0 on opening. The 2018 toll road study 7 reported small passenger vehicle speeds of 62 km/h in 2018 over 200 km of the old M3 and estimated 110 km/h for the new four- lane road.

22. Assumed vehicle fleet characteristics are shown in Table A10.8. Values of passenger working time, crew and maintenance labor costs range from $3.5 to $1.9 per hour at 2018 shadow prices. (GDP per head in 2018 was $2.46/h). Non-working time is values at 30% of working time. Time values are increased by 1.7% per year from 2019 to reflect expected growth in GDP per head.

Table A10.8: Vehicle Fleet

Item Unit Cars Pickups Minibuses Medium- 2-Axle 3–4- Art sized Trucks Axle Trucks Bus es Trucks Axles No. 2 2 2 2 2 4 5 Km/year km 10,000 50,000 80,000 100,000 100,000 100,000 100,000 Service life Years 10 10 10 10 8 8 8 No. of passengers No. 3 2 13 35 1 0 0 Operating weight Tons 1.4 1.8 3.2 15 6 20 30 VEF Axles 0.02 0.02 0.14 1.41 0.91 3.5 3.8 Vehicle cost $ 23,000 19,000 18,000 43,000 21,000 90,000 100,000 Tire cost $ 125 80 80 150 170 240 240 Maintenance labor $/h 1.5 1.5 1.5 1.5 1.5 1.5 1.5 Crew $/h 0 0 2.5 2.5 2.5 2.5 2.5 Passenger working $/h 3.5 3.1 1.9 1.9 1.9 - - time

7 Sheladia Associates for ADB and Azeravtoyol State Agency. 2018. Draft FSR for Installation of a Toll Collecting System along the M3 Alat-Astara Highway.

Appendix 10 41

Item Unit Cars Pickups Minibuses Medium- 2-Axle 3–4- Art sized Trucks Axle Trucks Bus es Trucks Passenger non- $/h 1.1 0.9 0.6 0.6 0.6 - - working time Work-related trips % 75 85 75 50 50 - -

Source: Asian Development Bank estimates.

23. The shadow price of gasoline and diesel is estimated at $0.60 per liter (higher than the November 2019 retail price of about $0.45). This estimate is based on World Bank price forecasts for crude oil for 2019–2030.8

4. Results of the Economic Reevaluation of the M3

24. The results of the economic reevaluation are summarized in Table A10.9 and shown in more detail in Table A10.15. The economic indicators provided are EIRR, net present value, and benefit–cost ratio. The appraisal discount rate of 12% is applied.

Table A10.9: M3 Reevaluation Results

Costs as PVs ($ million) Benefits as PVs ($ million) NPV BCR EIRR CAPEX Mtce VOC Journey time ($ million) (ratio) (%) 245 1.0 51.6 23.7 -171 0.31 3.8 BCR = benefit–cost ratio, CAPEX = capital expenditure, EIRR = economic internal rate of return, Mtce = metric tons carbon equivalent, NPV = net present value, PFR = periodic financing report, PV = present value, VOC = vehicle operating cost. Source: Asian Development Bank estimates

25. The result indicates that the M3 project represented poor value for money, for two reasons:

(i) The actual cost was about 60% higher than expected at appraisal in 2007. (ii) Traffic has been significantly lower than expected: average traffic on the new M3 over 2019–2037 is now estimated at 4,800 vehicles/day, compared with 9,200 expected at the time of the PFR (Table A10.5). The predicted journey time saving for vehicles traveling the entire 59 km of the new M3 is about 30 minutes, and the average annual road user costs in the with-project case is 27% less than that in the without-project scenario.

26. Neither the cost overrun for section A nor the sharp fall in GDP and traffic growth could reasonably have been foreseen at the time of the RRP. An alternative evaluation, therefore, takes into account the financial consequences of the failure of the section A contractor and uses the traffic growth rates from the PFR (but applied to this reevaluation’s opening 2019 traffic). The former has the effect of reducing the total economic cost of civil works from $341.3 million (Table A10.6) to $329 million, while the latter increases average traffic levels after 2019 by about 37%.

8 World Bank Commodity Markets Outlook, Oct 2019

42 Appendix 10

Table A10.10: M3 Reevaluation Results—Alternative Scenario

Scenario Costs as PVs ($ million) Benefits as PVs ($ million) NPV BCR EIRR CAPEX Mtce VOC Journey time ($ million) (ratio) (%) PFR traffic 245 0.9 70.8 34.9 -141 0.43 6.4 cost 237 1.0 51.6 23.7 -163 0.32 4.0 Both 237 0.9 70.8 34.9 -133 0.44 6.6 BCR = benefit–cost ratio, CAPEX = capital expenditure, EIRR = economic internal rate of return, Mtce = metric tons carbon equivalent, NPV = net present value, PFR = periodic financing report, PV = present value, VOC = vehicle operating cost. Note: An alternative scenario reduces initial investment cost by 3.3% to allow for section A contractor failure and uses PFR traffic growth rates. Source: Asian Development Bank estimates.

27. Table A10.9 demonstrates significant sensitivity to traffic growth, bringing the EIRR from 3.8 to 6.4%.

D. Evaluation of Tranche 1 Rural Roads

28. Under tranche 1, 39.5 km of rural roads were upgraded. They were not subject to economic appraisal as part of the RRP in 2007 or as part of the PFR in 2011. No traffic or condition data are available. All the roads listed in Table A10.9 are minor roads in categories III and IV providing access to the M2 east–west highway between Azerbaijan and Georgia. Under tranche 1, they were upgraded to category II.

29. For evaluation purposes, existing roads were assumed to have single bituminous surface treatment (on a granular base) (SBST) wearing courses, last receiving treatment in 1990 and with an initial IRI of 10 in 2015. Initial traffic and fleet composition in 2011 were taken as the average of the 2019 counts on the local roads constructed under tranche 4, i.e., 964 vehicles/day. Traffic growth in 2011–2019 was assumed at zero. In the absence of local data, M3 growth rates from 2019 were taken from Table A10.4, with 10% generated traffic (passenger vehicles only) added.

Table A10.11: Tranche 1 Rural Roads Traffic, 2011 (vehicles/day)

Cars Pickups Minibuses Medium- 2-Axle 3–4-Axle Art Trucks Total sized Bus es Trucks Trucks 878 14 34 18 3 15 2 964 Note: Zero growth is assumed for 2011–2019. Source: Asian Development Bank estimates.

30. The upgrade comprises an asphaltic mix on a granular base (AMGB) pavement. Maintenance in both the with- and without-project cases followed Table A10.7.

Appendix 10 43

Table A10.12: Project 1 Rural Roads

Road Name Length (km) M2–Nemetabad (category IV) 4.2 M2–Kemerli (category III) 14.0 M2–Samedabad–Ahmadabad (category IV) 7.4 M2–Delimmemedi–Qushchular (category IV) 4.5 M2–Ashagi Salahli (category IV) 7.0 M2–Nadirkend (category III) 2.4 39.5

31. HDM-4 was used to evaluate the rural roads as a single project, with benefits confined to VOC and journey time savings. The results are summarized in Table A10.13 and shown in more detail in Table A10.16.

Table A10.13: Project 1 Rural Roads Reevaluation Results

Costs as Benefits as PVs ($ million) a NPV BCR EIRR PVs VOC Journey time ($ million) (ratio) (%) ($ million) 5.47 7.51 1.59 3.63 1.66 16.2 BCR = benefit–cost ratio, EIRR = economic internal rate of return, NPV = net present value, PV = present value, VOC =vehicle operating cost. a Including generated traffic benefits. Source: Asian Development Bank estimates.

32. While Table A10.12 demonstrates a sound case for the upgrades, the lack of data at any stage during project processing renders them purely indicative. Some validation of the results is provided by the results for tranche 4 local roads: the present values of benefits per km are similar: $0.23 m/km for tranche 1 and $0.21 m/km for tranche 4. (Both packages have similar traffic and road lengths, although tranche 4 has lower unit initial investment costs.)

E. Evaluation of Combined Tranches 1 and 3

33. Table A10.16 shows detailed evaluation results for the M3 and the tranche 1 rural roads. The combined summary, inevitably dominated by the results of the much larger M3 investment, is shown below. Table A10.14: Tranches 1 and 3 re-evaluation results

Costs as PVs ($m) Benefits as PVs ($million) NPV BCR EIRR Capex Mtce VOC Journey time ($m) (ratio) (%) 251 1.0 59.1 25.4 -167 0.34 4.2 BCR = benefit-cost ratio; EIRR = economic internal rate of return; NPV = net present value Source: Asian Development Bank estimates

34. Traffic growth emerged as a key variable in the re-evaluation of the M3. Table A10.15 shows sensitivity of the combined project to adoption of the M3 PFR traffic growth rates, and to a conventional range of costs and benefits.

44 Appendix 10

Table A10.15: Sensitivity analysis

Item Sub-project Tranches 1 and 3 M3 Rural roads EIRR% NPV $m EIRR% NPV $m EIRR% NPV $m Base case 3.8 -171 16.2 3.6 4.2 -167 M3 PFR traffic growth 6.4 -141 16.7 4.2 6.7 -137 Costs x 1.2 2.8 -221 14.5 2.5 3.2 -219 Benefits x 0.8 2.6 -187 14.1 1.3 2.9 -185 Switching values: Costs a -69.5 +67 -67 Benefits a +329 -60 +300 Source: Asian Development Bank estimates a The percentage by which costs or benefits need to change to give an EIRR of 12%

Appendix 10 45

Table A10.16: Tranches 1 and 3 Reevaluation Results (at 2018 world prices)

M3 (tranches 1 and 3) Tranche 1 rural roads Incremental cost Benefits Total Net Incremental cost Benefits Total Net Combined Investment Maintenance VOC Time benefits benefits Investment Maintenance VOC Time benefits benefits 2011 20.43 0 0 0 0.0 -20.4 6.99 -0.15 0.00 0.00 0.00 -6.99 -27.42 2012 44.93 0 0 0 0.0 -44.9 6.99 0.00 0.00 0.00 0.00 -6.99 -51.92 2013 42.85 0 0 0 0.0 -42.9 -5.92 -0.06 1.54 0.32 1.86 7.85 -35.00 2014 67.03 0 0 0 0.0 -67.0 0.00 0.00 0.82 0.15 0.97 0.97 -66.05 2015 56.19 0 0 0 0.0 -56.2 0.00 0.00 0.89 0.16 1.05 1.05 -55.14 2016 78.49 0 0 0 0.0 -78.5 0.00 0.00 0.96 0.17 1.13 1.13 -77.36 2017 48.64 0 0 0 0.0 -48.6 0.00 0.00 1.04 0.18 1.22 1.22 -47.43 2018 0.05 0 0 0 0.0 -0.1 0.00 -0.03 1.12 0.20 1.32 1.35 1.29 2019 0.00 0.23 8.52 3.31 11.8 11.6 0.00 -0.04 1.20 0.22 1.42 1.46 13.07 2020 0.00 0.23 9.77 3.87 13.6 13.4 0.00 -0.05 1.32 0.25 1.57 1.62 15.03 2021 0.00 0.23 11.15 4.51 15.7 15.4 -5.92 -0.01 1.45 0.29 1.74 7.68 23.11 2022 0.00 0.23 12.70 5.23 17.9 17.7 0.00 0.00 0.75 0.16 0.91 0.91 18.61 2023 0.00 0.23 14.42 6.05 20.5 20.2 0.00 0.00 0.83 0.17 1.00 1.00 21.25 2024 0.00 0.17 16.33 6.98 23.3 23.1 0.00 0.00 0.91 0.19 1.10 1.10 24.23 2025 0.00 0.16 18.40 8.02 26.4 26.3 0.00 0.00 0.99 0.21 1.20 1.20 27.46 2026 0.00 0.18 20.68 9.18 29.9 29.7 0.00 0.03 1.08 0.23 1.31 1.28 30.96 2027 0.00 0.23 11.86 5.35 17.2 17.0 0.00 0.04 1.18 0.26 1.44 1.40 18.38 2028 0.00 0.23 13.36 6.15 19.5 19.3 0.00 0.02 1.29 0.30 1.59 1.57 20.86 2029 0.00 0.28 14.89 7.02 21.9 21.6 -5.92 0.04 1.41 0.34 1.75 7.63 29.26 2030 0.00 0.37 16.56 7.98 24.5 24.2 0.00 0.05 0.57 0.21 0.78 0.73 24.90 2031 0.00 0.40 18.40 9.05 27.4 27.1 0.00 0.05 0.62 0.22 0.84 0.79 27.84 2032 0.00 0.39 20.42 10.25 30.7 30.3 0.00 0.05 0.66 0.23 0.89 0.84 31.12 2033 0.00 0.35 22.62 11.57 34.2 33.8 0.00 0.05 0.70 0.25 0.95 0.90 34.73 2034 0.00 0.35 24.86 12.94 37.8 37.5 0.00 0.05 0.74 0.27 1.01 0.96 38.41 2035 0.00 0.42 13.26 7.38 20.6 20.2 0.00 0.05 0.78 0.29 1.07 1.02 21.23 2036 0.00 0.42 14.83 8.33 23.2 22.7 0.00 0.05 0.83 0.32 1.14 1.09 23.83 2037 0.00 0.42 16.55 9.37 25.9 25.5 0.00 0.05 0.88 0.35 1.23 1.18 26.68 2038 0.00 0.42 18.44 10.51 28.9 28.5 0.00 0.02 0.94 0.39 1.33 1.31 29.83 2039 0.00 0.42 20.50 11.75 32.2 31.8 -5.92 0.04 1.01 0.44 1.44 7.33 39.15 2040 0.00 0.40 22.77 13.11 35.9 35.5 0.00 0.05 -0.08 0.23 0.15 0.10 35.57 2041 0.00 0.36 25.18 14.58 39.8 39.4 0.00 0.05 -0.10 0.23 0.13 0.08 39.48

46 Appendix 10

M3 (tranches 1 and 3) Tranche 1 rural roads Incremental cost Benefits Total Net Incremental cost Benefits Total Net Combined Investment Maintenance VOC Time benefits benefits Investment Maintenance VOC Time benefits benefits 2042 0.00 1.41 27.73 16.15 43.9 42.5 0.00 0.05 -0.11 0.23 0.12 0.07 42.55 2043 -71.72 0.42 13.36 8.34 21.7 93.0 -2.80 0.05 -0.12 0.22 0.10 2.84 95.84 PV at 12% 245 1.0 51.6 23.8 75.4 -171 5.51 -0.04 7.51 1.59 9.11 3.63 -167 NPV -171 NPV 3.63 -167 EIRR 3.8% EIRR 16.2% 4.2% BCR 0.31 BCR 1.66 0.34 Source: Asian Development Bank estimates

Appendix 10 47

LIST OF PARAMETER VALUES/ASSUMPTIONS

Price base year: 2018 Discount year: 2011 Currency of analysis: $ Construction start year: 2011 (M3 and tranche 1 rural roads) Construction end year: 2017 (M3 – opened 2018); 2013 (tranche 1 rural roads) First year of benefits: 2019 (M3); 2013 (tranche 1 rural roads) Appraisal period: to 2043 (25 years M3 operation plus implementation period) Numeraire used: World price numeraire Income elasticity of demand: 1.0 falling to 0.8 (passenger vehicles), 1.5 falling to 1.4 (goods vehicles) Value of time (in work, 2018): $3.5/hour (car passengers) - $1.9/hour (bus passengers) Value of time (non-work, 2018): $1.1/hour (car) - $0.6/hour (bus passengers) GDP growth assumption: 2019 2.5%, 2020 2.7% (ADB ADO 2019), 2024 2.4% (IMF WEO Oct 2019) Shadow price of labor: 0.9 (unskilled) Standard conversion factor: 0.98 Conversion factor applied to works, LAR and physical contingencies: 0.98 Conversion factor applied to supervision: 1.0 Conversion factor applied to taxes, duties, profits, transfers: 0.0