Although none of the 37 contracts under this project experienced any implementation delays, the performance of the contractors could have been better with respect to the quality of work and the technical specifications and construction standards. Specific issues were (i) design-related, affecting certain circuit-breakers, concrete poles for distribution lines, the size of pre-cast foundations, cross-arms, and system earthing; and (ii) construction-related, such as the stringing of transmission and distribution lines, installation of self-supporting insulated wire distribution lines, compaction of ground, installation of pre-cast foundations for steel lattice towers, anchor bolts for substation equipment, and grouting. These issues were either partially resolved or recommended for better technical specifications in future tranches of the MFF. Employing national and international best practices will be key to improving the quality and technology in future tranches. Encouraging contractors to improvise where possible and introduce new technology standards, incorporating lessons learned and issues identified in the technical specifications for the initial tranche, and prompting the project management units to be more proactive in using the project management consultant to develop the capacity of the operational staff on new technology and standards will also be helpful.
This MFF’s tranches 2 and 3 were supposed to be approved in November 2017 and March 2018, respectively. However, the restrictive approach taken by Azerbaijan to public external borrowing, which was absent at appraisal time, led to the deferment of these tranches. It might have been possible to push through with these tranches if they were processed close to the start of tranche 1, with the locations and routes identified upfront and thoroughly analyzed for potential construction and safeguard impacts. Overlapping processing and implementation of the first and subsequent tranches would have maintained the implementation momentum and maximize the benefits from the MFF.
Although the actual cost for the civil works sections increased by about 50%, the project still ended with unused loan proceeds of $9.65 million (Loan 2205) and SDR1.40 million (Loan 2206). These were sufficient to finance the local roads, which were estimated to cost $3.8 million. A thorough fund disbursement analysis should have been conducted during project implementation. The PCR recognized the need for greater due diligence during project design and formulation, which includes carrying out adequate cost estimations and setting aside enough resources to cover contingencies. The IEM also highlighted this lesson for future road projects.
Some indicators for impact and outcome are not readily available nor closely linked to the project. Since data are generally not readily available in Azerbaijan, sources for evaluation indicators should be identified in more detail at appraisal. Documents such as those prepared at appraisal and project completion should include sufficient data for subsequent verification and follow-up. Specific omissions were traffic data, unit benefits, and unit costs that were needed to fully comprehend the economic analysis.
At appraisal, two sections were intended to be financed by ADB. The Yevlakh-Ganja section involves much higher traffic than the Gazakh–Georgian border section. Thus, the Yevlakh-Ganja section brought greater benefits and greater socioeconomic impact to the project area. If ADB had financed the Yevlakh-Ganja section instead of the Gazakh–Georgian border section, the project might have been rated efficient as the recalculated EIRR would have been higher than the threshold of economic viability. Thus, selection of the ADB-financed project section should have been made more carefully.
For instance, the detailed design, based on a precise analysis (e.g., technical inputs, standards specifications, and cost estimates, among other things) helps minimize potential cost overrun that often result to implementation delays. Moreover, design details could benefit from reliable traffic forecasts (i.e., use of reliable income elasticities of demand) and adequate surveys, including geological hazards. This could help in developing appropriate design and in preventing major scope and design changes during implementation. On the whole, identification of any potential design issues, along with appropriate risk mitigation strategies during the design phase, helps enhance readiness and quality of entry.
For example, projects 1 and 3 of the whole program required the rehabilitation of 120 km access roads to sections A, B, and C. These were key project components since most of the economic benefits were to be derived from local roads traffic. However, projects 1 and 3 failed to rehabilitate the secondary roads due to lack of firm financial sources. This was a contributing factor to the lower than expected traffic volume and consequently, to a low EIRR.
Specifically, institutional capacity assessment of executing and implementing agencies at the onset is vital in planning for appropriate measures to support the agencies during the program. Technical assistance for capacity building helps complement such assessments by filling in gaps in capacity to enhance their financial, procurement, and technical capacity skills. These could include familiarization with ADB and government procedures and requirements, strategic planning, among other things, which are critical in implementing MFF tranches.
Overly optimistic forecasts of traffic volumes can stem from using outdated or arbitrary parameters and unrealistic assumptions. For instance, elasticity values of road transport demand may have changed over time. Also, the use of national GDP as a proxy for economic activities at the regional or local level could result in unrealistic traffic forecasts. Both of these could lead to an overly designed project. Accurate traffic forecast and reliable growth estimates contributes to establishing a good framework for road engineering design, including more reliable cost estimates. All these require critical assessment of traffic demand projection, especially during project appraisal phase.
While the actual cost for the project civil works sections increased by 50%, the project still ended with unutilized loan proceeds of $9.65 million (Loan 2205) and $1.40 million (Loan 2206 [Special Funds]). These were more than sufficient to continue implementing the local roads component estimated to cost $3.8 million, of which $3.0 million was to be funded from the loan. Even with a 50% increase in cost, the local roads civil works could still have been implemented.
Azerbaijan has significantly improved its road infrastructure quality. This is primarily attributable to the newly constructed high-quality motorways. Under the government’s annual state investment programs and the 5-year state programs on regional development, sizable funds are envisaged for the construction and rehabilitation of local roads. To increase the benefits for local communities and promote in-land connectivity, the planning and implementation of local road works should be better synchronized with those on motorways.