2. Network and responsibilities - hti.amhti.am/edfiles/files/voroshumner/AARCroadsfinancingstrategy.pdf · AARC final draft road maintenance strategy June 2015 2 (c) length under

AARC final draft road maintenance strategy June 2015 1

ADB TA8710 ROAD MAINTENANCE FINANCING STRATEGY FINAL DRAFT

1. Introduction The Government of Armenia is currently preparing a road financing strategy to meet the needs of road users and the communities in a cost effective and efficient way, while maximizing the value of its investment in road maintenance. This final draft was prepared in June 2015 by TA consultants on assignment to ADB. Its aim is to assist in the development of the strategy. It evaluates alternative funding scenarios for the interstate and republican networks. Local roads are also included, but at a lower level of detail commensurate with the lack of data.

The term maintenance covers routine maintenance (treated as a recurrent item and including drain clearing, removal of snow, pothole patching and shoulder repairs) and capital maintenance (including seals, overlays and reconstruction, often referred to collectively in Armenia as rehabilitation).

2. Network and responsibilities

2.1 Network lengths The total road network is (at April 2015) given as 7,530km, as set out below. This base network is often referred to as the common use network. If urban roads (other than those in Yerevan) are included, the total reaches around 10,800km (see for instance table 3 in the World Bank’s 2011 report: Improving the Sustainability of Road Management and Financing in Armenia). However, the 3,000km of minor urban roads do not attract direct central government support for maintenance activities.

The base network of 7,530km has not changed for several years. The local network is currently under review, however, and its length is expected to fall by 130km. Parts of the interstate and republican networks carry negligible or no traffic, in most cases because they end at closed borders.

Table 1: Base national road network

Designation Length in km

Ownership AADT, veh/day,

2014 MOTC Marzpeteran Yerevan

mayoralty

Interstate (M)a 1,759 1,714 - 44 2,900b

Republican (H) 1,966 1,955 - 11 1,980b

Local (T) 3,806 - 3,806c - N/A

Total 7,530 3,669 3,806 55

Source: MOTC (rounded to nearest km), TA consultants Notes: (a) M, H and T are initial letters of road numbers. Local (T) roads also carry an initial numeric

regional identifier (b) weighted average of 2014 motorized traffic on ARD’s HDM4 networks


(c) length under review (see text)

2.2 Designations and standards The designations interstate, republican etc are functional designations. There are also four technical classes, ranging from class I (four lanes with a lane width of 3.6m and an hourly capacity of 2,800 PCU per lane) to class IV (two lanes of 2.7m, hourly capacity up to 500 PCU/lane). There are no technical standards for low volume (e.g. single lane) roads, an omission which should be rectified and which appears as a recommendation in the action plan shown at annex B.

Of the interstate network of 1,759km, approximately 160km have four lanes. Net of the final expected total length of approximately 500km of the North-South Road Corridor Investment Program, the two lane equivalent length of the interstate network is 1,364km.

Around 70 percent of the base network is paved. This leaves approximately 2,000km of mostly local roads that are unpaved (source: table A3.1 of TA8249). Most of these are roads that were formerly paved but whose pavements have completely deteriorated.

With the exception of the first three tranches of the North-South Road, which have a concrete pavement, all paved roads have an AC pavement. (See also section 8.2).

3. Network condition Officially, road condition is reported according to road maintenance decree no 1419-N (2010). Table 1 in the decree uses potholes (as percent of the pavement area) and shoulder and roadside drain conditions (percentage not complying with norms) as criteria for assessing a road section as “good”, “fair” or “poor”. For interstate roads, the relevant ranges of pothole percentages are <0.3, 0.3-2.5 and >2.5 for good, fair and poor respectively. For republican roads the ranges are <0.5, 0.5-3.0 and >3.0 and for local roads <0.8, 0.8-4.0 and >4.0. There are similar ranges for shoulder and roadside drain condition. Once a road section is assessed as “bad” it is removed from routine maintenance and considered for capital maintenance. At the time of reporting just under 500km of interstate and republican roads are in this category. This procedure makes sense contractually (visual assessment can be used to measure contractor performance) but the omission of roughness is a weakness. It is also the case that the central ranges of pothole percentages are indicative of a wide range of conditions.

Potholes and shoulder and roadside drain condition are, pending implementation of a pavement management system (PMS), assessed visually. For the time being the more reliable indicator is IRI, being measured annually on at least most of the interstate and some of the republican networks.

The tables below show estimated IRIs for the interstate and republican networks from 2010 to 2014. The comparison is distorted because the networks used to estimate average IRIs (taken from the Armenian Roads Directorate’s HDM4 databases) are not identical for each year shown. Of the two networks, the interstate network is much better represented by the HDM4 database; values for the republican network are based on a 10-20 percent sample. It shows, with caveats concerning data quality and its ability to represent the entire network, that interstate network average IRI has risen by 0.2-0.3 per year. While the local road network is included in the strategy, there are no systematically held pavement quality data. For the purposes of estimating maintenance costs it is


assumed that paved local roads have the same pavement quality characteristics as the republican network.

Table 2: Network IRIs

Network, year <4 (good) 4-6 (fair) 6-9 (poor) >9 (bad) Weighted average

Interstate

Interstate, 2010 19% 74% 6% 1% 4.8

Interstate, 2013 12% 61% 22% 5% 5.5

Interstate, 2014a 13% 54% 30% 3% 5.6

Republicanb

Republican, 2010 19% 15% 49% 17% 7.0

Republican, 2014 29% 38% 22% 11% 5.7

Source: ARD (2013), TA consultants using ARD HDM4 network databases (2010 and 2014) Notes: (a) derived from HDM4 network database, which omits some road sections earmarked for

rehabilitation and may present a misleadingly optimistic picture (b) based on 10-20 percent sample of network

4. Targets In principle, investing in road maintenance is worthwhile up to the point where total transport cost (road agency plus road user costs) are minimized; an approach of this sort is followed in the 2011 World Bank paper. A similar economic criterion is to invest up to the point where the gains in the form of reduced road user costs are outweighed by the incremental costs needed to secure the gains. It is a useful approach when assessing total financing needs, but for operational use it is more convenient to specify a target network condition, either as a network average or as a distribution.

Thus, the Armenia Development Strategy 2014-25 (ADS) sets target road conditions for 2014, 2017, 2021 and 2025. By 2025 the percentages of roads that should be in fair condition are 100, 60 and 45 for interstate, republican and local roads respectively. There are no condition definitions in the ADS, but in IRI terms fair condition for an asphalt road would typically be 3.5-5.5m/km (see World Bank 2011).

Annual condition targets are also set in the MTEF (Medium Term Expenditure Framework) by MoTC’s financial department, expressed for each network as both IRI and as a percentage in poor condition (using decree 1419 criteria). For 2015 IRI targets are 6.29 and 7.5 for interstate and republican roads; no targets are given for local roads. The interstate and republican targets suggest an acceptance of a decline from 2014 performance (see table 2) and acceptance that ADS targets are no longer achievable. The MoTC continues to have a target that 60 percent of the network should be in a “maintainable” condition (approximately equivalent to an IRI<6).

In strategic terms there is a need to replace the targets with values that are achievable with the resources expected to be available; such an item is included in the action plan at annex B.


5. Historic and planned future expenditure Table 3 shows historic actual expenditure from 2009 to 2013, and planned expenditure for 2014-17 (the actual budget for 2014 is not yet approved).

In current terms, support for road maintenance has fallen from AMD49bn in 2009 to AMD19bn in 2013, an annual fall of approximately 20 percent at constant 2015 prices, or 14 percent if IFI-supported rehabilitation is omitted. Historically, IFI funding of road rehabilitation has made a major contribution, but this has also declined, to the point where all maintenance activities represented just 0.4 percent of GDP in 2013. As acknowledged in the ADS, the fall in funding corresponded to the financial crisis (Armenia’s GDP fell by 14.1 percent in 2009 (source: World Bank)). At 2015 prices, average annual maintenance expenditure over the period 2009-13 was AMD33bn, but a more representative figure for 2012-15 is approximately AMD20bn, or US$42m. The MOTC interstate and republican networks absorb most of the budget: for 2015 only approximately AMD1bn is allocated to the rehabilitation of local roads. Routine maintenance budgets are subject to a 5 percent annual increase, but this is in nominal terms and generally represents little or no increase in real terms.

Average costs per km of rehabilitated road over the period 2009 to 2013 have averaged approximately AMD100m (US$200,000) at 2015 prices (calculated as the weighted average of capital maintenance over the km of rehabilitated road in table 3 and including VAT), a value that is consistent with the costs of medium and heavy maintenance interventions used in this study (table 5). Routine maintenance averaged over the lengths shown in table 3 amounts to AMD1.9m/km (US$4,000), comparable to the routine maintenance unit cost adopted here.

There is no “correct” level of road maintenance expenditure as a proportion of GDP, but 0.4 percent is well below the usually quoted range of 1 to 2 percent range of expenditures found in those countries around the world whose infrastructure is already well developed and have GDP growth rates of 2 to 3 percent (source: Africa's Transport Infrastructure: Mainstreaming Maintenance and Management, Gwilliam et al for World Bank, 2011). This is consistent with the ADS, which indicated that an average of around 1.2 percent of GDP would be spent from 2015 on the roads sub-sector.


Table 3: Roads sector maintenance expenditure, executed and planned, 2009-18

Item Source 2009 2010 2011 2012 2013 2014a 2015a 2016a 2017a 2018a

GDP, current AMDbn 3,142 3,460 3,778 4,001 4,273 4,526 4,800e

Exchange rate, AMD per US$f 364 374 373 402 410 415 476f

Inputs: expenditure in AMDbnh

MoTC capital maintenance GoA 12.0 7.4 5.4 7.2 3.9 13.2 4.0 4.0 4.0 4.0

MoTC routine maintenance GoA 5.8 6.2 6.1 6.3 6.2 5.7 6.0 6.2 6.5 6.9

Marzes routine maintenance GoA 0.8 0.8 0.8 0.8 0.8 0.8 0.8

North-South Highway ADB 0.5 7.0 8.0 29.6 19.9 7.4 16.7 32.0

EBRD Bagratshen border (M6) EBRD 0.5 1.5 1.8 2.2

LRNIP Rehabilitation program World Bank 30.8 10.7 10.4 6.5 8.4 2.6 3.5 3.3 5.9 2.3

Total roads All 48.5 24.3 22.4 27.0 26.5 51.1 34.7 30.7 43.5 47.4

Of which maintenance, AMDbn All 48.5 24.3 21.9 20.0 18.5 21.5 14.8 23.3 26.9 13.2

As % of GDP 1.6% 0.7% 0.6% 0.5% 0.4% 0.5% 0.7% N/A N/A N/A

Km rehabilitated 411 350 236 187 134 N/A N/A N/A N/A N/A

Km subject to routine maintenance 3,246 3,246 3,159 3,184 3,298 N/A N/A N/A N/A N/A

Sources: Draft 2015-25 financing strategy for maintenance of common use roads, MoTC, Dec (2009-13), MOTC (2015) and MTEF Notes: (a) planned budget according to MoTC finance department

(e) estimate for 2015 is projected from 2014 by assuming 4 percent inflation and 1 percent GDP growth (f) annual average exchange rates to 2014 from EIU Armenia Country Report, Jan 2015. 2015 rate is Bank of Armenia rate on 12 April 2015 (g) taken from the 2015-17 MTEF entry for the LRNIP (and subject to confirmation) (h) all expenditure values include VAT at the standard rate of 20%


6. Estimates of future resources

6.1 Introduction Road maintenance levels of investment are in practice constrained by both GDP and the national budget. The two are of course linked. The possibilities for raising further transport-related sources of finance are not considered in this document.

6.2 ADS forecasts The ADS stated that, starting from 2015, base scenario contributions to the transport sector were to be around 1.4-1.5 percent of GDP, of which 85 percent was to be aimed at the road network. This can be interpreted as an allocation of 1.2 percent of GDP to the roads sub-sector. The ADS does not identify how much sub-sector expenditure should be spent on road maintenance, although it is clear from the ADS text that maintenance was seen as a priority. At 2015’s estimated GDP, adoption of the ADS allocation would mean AMD58bn for the sub-sector.

The ADS was predicated on a GDP growth rate of 5.5 to 6.5 percent, a forecast that was consistent with other contemporary estimates in March 2014. The current (2015-17) MTEF, adopted in July 2014, assumes similar growth rates.

6.2 Changes since ADS More recent GDP growth forecasts paint a very different picture from those adopted in the ADS. In January 2015 the World Bank was forecasting 3.3, 3.7 and 4.1 percent for 2015, 2016 and 2017 respectively. The EIU (Economist Intelligence Unit) was making similar forecasts at the same time: 2.7 and 3.4 percent for 2015 and 2016. Since then, reduced remittance flows combined with muted external demand led to revised forecasts and in late March ADB reduced its forecast growth for 2015 and 2016 to 1.6 and 2.3 percent (source: Asian Development Outlook 2015), while the World Bank’s Armenia country manager forecast 0.8 percent for 2015 at a news conference in Yerevan on 31 March.

6.3 Budget considerations The planned national budgets in 2012 and 2013 were 26.2 and 25.4 percent of GDP respectively. Roads maintenance expenditure (excluding development expenditure but including rehabilitation and LRNIP disbursements) in each of those years amounted to approximately AMD20 billion, or 1.9 and 1.7 percent of the national budget respectively.

Given the competing claims on the national budget, which can be expected to rise during a period of low economic growth, ADS allocations to the transport sector proposed in the ADS for 2015 onwards now appear unrealistically high.

Despite the competing claims, total allocations to the road sector have held up quite well: expressed at 2015 prices the planned expenditure for 2015-18 in table 3 represents approximately 0.7 percent of estimated 2015 GDP. About 55 percent of this is to be invested in the IFI-supported North-South Road Corridor, a capacity enhancement project. Approximately 0.3 percent of GDP (AMD13bn) remains annually for maintenance, including the World Bank supported LRNIP.

This is well below the level of approximately AMD20bn (approximately 0.4 percent of GDP) allocated up to 2014. Even AMD20bn is completely inadequate if the network is not to deteriorate to a point


from which it will be difficult (i.e. increasingly costly) to recover (as will be shown later in this paper). A level even lower than this cannot be regarded as a part of a credible bid for resources from the state budget.

6.4 Business as usual financing constraints Table 4 shows a range of business as usual (BAU) scenarios, i.e. scenarios that are consistent with allocations up to 2014. It assumes that the national budget continues to represent 25 percent of GDP (estimated at AMD4,800bn for 2015) and that road maintenance allocations lie between 1.7 and 1.9 percent of the budget. The amounts shown for road maintenance are simple averages over the period to 2025.

Table 4: BAU scenarios

GDP growth rate: 0% 2% 3.5%

AMDbn/yr for road maintenance inc VAT 20-23 23-25 24-27

Source: TA consultants’ estimates

7. Previous estimates of road financing requirements There have been three previous estimates of road financing requirements in recent years:

• Improving the Sustainability of Road Management and Financing in Armenia, World Bank, 2011

• ADB TA8249, draft final report, Sep 2014 • Consulting Services for Roads Financing Study, LRNIP, second technical report, May 2015

The World Bank’s 2011 paper used HDM-41 to estimate total transport costs (TTCs, i.e. the sum of road agency and road user costs) for four scenarios, ranging from minimizing TTCs over a 20 year evaluation period to maintaining the current network condition. Their conclusion was that an “eliminate the rehabilitation backlog in 10 years” scenario was the best compromise between a very costly unconstrained TTC minimization strategy and a low cost “maintain current condition” scenario.

Even the eliminate backlog scenario was costed at US$98m per year over the first five years, followed by US$74m per year over the following 15 years (i.e. US$860m in total over 10 years, an annual average of US$86m). Averaged over their assumed network of 5,740km (which at the time represented the length of the designated interstate and republican networks) the average annual expenditure was US$15,000/km over the first ten years. The five year level of expenditure implied an annual funding gap of U$30m when compared with 2011 planned expenditure of US$68m. Under this funding scenario it was expected that 55 percent of interstate and republican roads would be in good or fair condition in 2015 (say an IRI<6), rising to 65 percent in 2020. In the event 2011 expenditure was around US$10m less than the authors assumed and the 2011 funding gap would have been around US$40m. Note that this gap only applies to the assessed 5,740km interstate and republican network. 1 HDM4 – Highway Development and Management 4 – is a standard software tool for evaluating road investments. It can be used for appraising discrete projects as well as for planning investments in entire networks.


TA8249 estimated annual required expenditure of US$237m to US$287m for a network of 7,400km, the range corresponding to alternative assumed pavement lives. Target pavement quality seems to have been somewhat higher than in the World Bank’s assessment – 2025 IRIs of 3.3 and 4.5 for interstate and republican roads respectively. Expressed as expenditure per km per year the TA8249 estimates amount to US$32,000-39,000, approximately double the World Bank’s unit expenditure estimates.

The LRNIP Roads Financing Study (2015) also used HDM4 to estimate annual maintenance costs for four scenarios:

1. Unconstrained: US$112m (AMD53bn) 2. Average IRI within 1.5 of now: US$75m (AMD36bn) 3. Continue spending as now: US$30m (AMD14bn) 4. Routine maintenance only: US$14m (AMD6bn)

(Values are annual averages over a twenty year period and are taken from tables 1 and 2 of the May 2015 second technical report).

8. Revised estimates of road financing requirements

8.1 Approach For this revised financing strategy, fresh estimates of road financing are made using HDM4 (v2.08) in its strategy mode. HDM4 requires considerable amounts of input data; the principal assumptions are set out in annex A. There are acknowledged weaknesses in the quality of the base inventories, especially for the republican and local networks, and it must therefore be accepted that current estimates should be revised following implementation of the RAMS that is currently under procurement.

Investment within HDM4 is very “lumpy” as a result of the fairly coarse aggregation of sections by traffic and roughness classes, described below, hence forecast pavement quality displays more variability from year to year than it would in reality. The link between budget constraint and actual expenditure is also a loose one as a result: HDM4 picks (lumpy) interventions that have EIRRs over 12 percent within the budget constraint. If there are insufficient worthwhile interventions, HDM4’s predicted expenditure will fall well below the constrained budget. This is the case here.

8.2 Interstate and republican networks The networks analyzed are the 2014 interstate and republican roads networks shown in table 1, with adjustments made for expected progress on the North-South road corridor project during the forecast period and for rehabilitation of 92km of the M-6 (an ADB co-financed project), also expected to take place during the forecast period. Recorded total traffic and roughness (the main drivers of prioritization) are believed to be of reasonable quality for the interstate and some of the republican network.

Each of the interstate and republican networks is grouped by traffic band (T1 to T7) and by roughness band (R1 to R4), as shown in the table below. Each group (e.g. T1R1) is treated as a section in HDM4. In the analysis, roughness is also linked to pothole frequency, cracking and year of most recent rehabilitation. Details of these assumptions, traffic growth rates etc can be found in annex A.


Table 5: Networks by traffic and roughness class

Traffic band R1

IRI<4

R2

4<IRI<6

R3

6<IRI<9

R4

IRI>9

Interstate % of km % of km % of km % of km

T1 (AADT<500 veh/day) T1R1 6.0% T1R2 1.9% T1R3 0.0% T1R4 0.0%

T2 (500<AADT<1,250) T2R1 0.2% T2R2 10.2% T2R3 4.9% T2R4 2.0%

T3 (1,250<AADT<2,000) T3R1 0.0% T3R2 11.1% T3R3 4.6% T3R4 0.2%

T4 (2,000<AADT<3,000) T4R1 0.0% T4R2 7.5% T4R3 4.7% T4R4 0.2%

T5 (3,000<AADT<4,500) T5R1 3.6% T5R2 13.4% T5R3 5.7% T5R4 0.3%

T6 (4,500<AADT<6,000) T6R1 2.2% T6R2 5.5% T6R3 7.4% T6R4 0.0%

T7 (AADT>6,000) T7R1 1.4% T7R2 4.2% T6R3 2.8% T7R4 0.0%

Republican

T1 (AADT<500 veh/day) T1R1 1.5% T1R2 0.0% T1R3 7.1% T1R4 4.3%

T2 (500<AADT<1,250) T2R1 5.1% T2R2 18.4% T2R3 6.4% T2R4 1.0%

T3 (1,250<AADT<2,000) T3R1 10.4% T3R2 10.4% T3R3 4.3% T3R4 0.0%

T4 (2,000<AADT<3,000) T4R1 8.5% T4R2 2.3% T4R3 0.0% T4R4 0.0%

T5 (3,000<AADT<4,500) T5R1 1.2% T5R2 4.2% T5R3 0.0% T5R4 0.0%

T6 (4,500<AADT<6,000) T6R1 2.6% T6R2 1.3% T6R3 3.8% T6R4 5.6%

T7 (AADT>6,000) T7R1 0.0% T7R2 1.5% T6R3 0.0% T7R4 0.0%

8.3 Local network There are no systematically held traffic or condition data for local roads. The assumption made here for the paved part of the local network is that the roughness distribution is the same as, but that the traffic is half that of the republican network (i.e. T1 now represents 0-250 veh/day etc).

8.4 Interventions Table 5 shows the interventions used in HDM4. Useful benchmarks are the unit cost of the LRNIP road rehabilitation works, which averaged US$260,000/km excluding VAT (AMD106m) for 73.3km of road section contracts awarded in mid-2013 (source: World Bank LRNIP procurement plan, 15 Oct 2014) and an average of US$290,000/km (including VAT) estimated by ARD for the rehabilitation of 300km of roads in 2015. Expected useful lives of each intervention are 6-7, 10-12 and 20 years for light, medium and heavy capital maintenance interventions respectively. The HDM4 construction defects parameter, CDB, was set to 0.1 (instead of the default of zero) in order to replicate expected lives.

Table 6: Maintenance interventions

Intervention US$ per km

incl VAT US$/m2 AMDm/km AMD000/m2 Components

Routine maintenance 3,400 1.6 Pothole patching, crack sealing, shoulder repair,


Intervention US$ per km

incl VAT US$/m2 AMDm/km AMD000/m2 Components

ditch cleaning and winter maintenance

Light capital maintenance 140,000 20 67 9.5 DBST with average 3cm levelling, crack sealing & pothole patching

Medium capital maintenance

210,000 30 100 14

Milling and/or average 3cm levelling, 40-50mm AC overlay & pothole repair

Heavy capital maintenance 300,000 43 140 20

Reconstruction including minor earthworks, sub-base & 40mm AC pavement

Source: ARD rates and ongoing work on the North-South Highway Note: all costs shown are at financial prices and include VAT. For conversion to economic prices, VAT is

deducted and a conversion factor of 0.96 applied (see annex for details)

Routine maintenance is an annual activity. Capital maintenance interventions are triggered during runs of HDM4 as described below. HDM4 predicts a trade-off between routine and capital maintenance: higher investment in capital maintenance reduces on-carriageway routine maintenance (principally pothole patching).

8.5 High expenditure maintenance strategies

8.5.1 Interstate and republican networks The strategies for the interstate and republican networks represented by tables 6 and 7 respectively are simple ones, commensurate with the quality of the data available, and would be capable of considerable refinement by introducing compound criteria if a better quality database were available.

Both strategies are “backlog-heavy”, that is, reconstruction is the preferred option for roads carrying at least 500 veh/day and in a poor condition, always provided the intervention is good value for money (i.e. has a net present value (NPV) of at least zero). In both cases the strategies result in approximately 40 percent of the network being reconstructed over the 2015-25 period.

Table 7 shows the strategy applied to the interstate network. 60 percent of the network carries traffic in bands 4 and above and it is interventions in bands T4-T7 that determine outcomes.

Table 7: High expenditure strategy – interstate network

Traffic band R1 R2 R3 R4

T1 Light when IRI

reaches 5 Light when IRI

reaches 5 Medium when IRI


reaches 6

T2 Light when IRI Light when IRI Heavy when IRI Heavy when IRI



reaches 5 reaches 5 reaches 10 reaches 10

T3 Light when IRI


reaches 6 Heavy when IRI


reaches 10

T4 Light when IRI




reaches 10

T5 Medium when IRI




reaches 10

T6 Medium when IRI




reaches 10

T7 Medium when IRI




reaches 10

Source: TA consultants

The strategy applied to the republican network is slightly different (table 8). In this case 70 percent of the network carries traffic in bands 1-3. Without invoking heavier interventions for roads in these bands, even a backlog heavy approach will not achieve significant network improvement. Table 8: High expenditure strategy – republican network


T1 Light when IRI

reaches 5 Light when IRI



reaches 10

T2 Light when IRI




reaches 10

T3 Medium when IRI




reaches 10

T4 Medium when IRI




reaches 10

T5 Medium when IRI




reaches 10

T6 Medium when IRI




reaches 10

T7 Medium when IRI




reaches 10

Source: TA consultants

8.5.2 Local road network For practical analytical purposes the local network is split into paved and unpaved. Neither traffic nor pavement data are available for either. Table 9 shows the interventions adopted for the paved local road network. Essentially the 40mm overlay is now the heaviest form of capital maintenance and replaces the reconstruction option in tables 7 and 8. This is adopted because otherwise estimates of total financing needs would be dominated by the network about which least is known. Restricting the heaviest intervention to an


overlay limits the financing needs of this part of the network, but it also puts a severe limit on the scope for pavement quality improvements.

Table 9: Overlay-based strategy – paved local road network

Traffic banda R1 R2 R3 R4

T1 Light when IRI

reaches 5.5 Light when IRI

reaches 5.5 Medium when IRI

reaches IRI=8 Medium when IRI

reaches IRI=8

T2 Light when IRI

reaches 5.5 Light when IRI



reaches IRI=8

T3 Light when IRI




reaches IRI=8

T4 Medium when IRI




reaches IRI=8

T5 Medium when IRI




reaches IRI=8

Source: TA consultants Note: (a) each band represents half the traffic of its namesake in table 7

For the unpaved part of the local network it is assumed that it is converted to full gravel status over the ten year period at a cost of US$50,000/km (AMD24m/km). Re-gravelling, at the same unit cost, is assumed to take place every five years, together with annual maintenance at US$1,700/km (AMD0.8m/km). The average annual cost of this approach would be approximately US$3,000/km (AMD1.5m/km). A good gravel road should have an IRI of around 6.5. However, as this approach to dealing with these highly degraded roads is speculative at present, estimates of its network IRI are not made.

8.5.3 Results of the high expenditure strategies The table below summarizes the results of running HDM4 in its unconstrained mode, using the strategies outlined above. At AMD45bn per year this is twice the lower end of expected BAU resources in table 4 and represents just over 0.9 percent of GDP; it can be dismissed as unaffordable in the present economic climate. The weighted average expenditure over the paved network is US$15,000/km/year (including VAT), the same numerically as the World Bank’s 2011. The total of AMD45bn is 15 percent lower than the LRNIP estimate.

Adopting this strategy would ensure that 60 percent of the network is maintainable by 2025.

Table 10: High expenditure results

Network Expenditure per year, 2015-2025a

Expenditure per km per year

Network average IRI

2017-20

Network average IRI in 2025

Network average IRI, 2015-25

Interstate US$31m (AMD14bn) US$19,000 5.8 5.3 5.4

Republican US$32m (AMD15bn) US$16,500 6.2 5.5 6.1




Network average IRI

2017-20



Local - paved US$16m (AMD8bn) US$9,700 6.0 6.1 6.4

Local - unpaved US$16m (AMD7.5bn) US$8,000 - - -

All US$95m (AMD45bn) US$13,000 6.0 5.6 6.0

Source: TA consultants’ HDM4 runs Notes: (a) all values include VAT

8.6 Business as usual (BAU) strategies BAU strategies are modeled by taking the high expenditure approaches described above and applying a similar budget constraint to each of the high expenditure strategies.

In the case of the interstate and republican networks this means reducing the proportion of the network reconstructed from 37 to 15 percent. For the local paved network the constraint means reducing capital maintenance interventions by 50 percent. For the local unpaved network the rate of conversion to a full graveled status is halved and subsequent re-gravelling omitted.

Network conditions will continue to deteriorate under this scenario, with an annual rate of IRI deterioration of approximately 0.3 (a comparatively low value that arises from the assumed low traffic growth rates). Much of the paved network will be unmaintainable by about 2027.

Table 11: BAU results


Expenditure per km per yeara

Network average IRI

2017-20



Interstate US$15m (AMD6.9bn) US$9,100 6.2 7.5 7.5

Republican US$15m (AMD7.3bn) US$7,800 6.8 8.0 7.2

Local - paved US$9.3m (AMD4.4bn) US$5,600 6.2 7.8 6.7

Local - unpaved US$5.8m (AMD2.7bn) US$3,000 - - -

All US$45m (AMD21bn) US$6,200 6.4 7.8 7.1

Source: TA consultants’ HDM4 runs Note: (a) all expenditure values include VAT

8.7 Proposed strategy Both the high expenditure and BAU scenarios produce unacceptable results: the former should generate good long term results but is unaffordable while the latter is affordable but the network continues to deteriorate.

The returns on investing in road maintenance are, other things being equal, higher on roads that carry higher traffic volumes. If resources are tightly constrained, as they are in Armenia, it may make sense to concentrate capital maintenance on interstate and republican roads. (This would be better


approached by first defining a core network, as now included in the action plan shown in annex B). Accordingly, the proposed strategy comprises:

• allocate bare minimum routine maintenance funds to local roads at a rate of US$2,000/km/year. This implies AMD3.5bn per year, approximately three times the current allocation

• allocate the minimum needed to the interstate and republican networks to minimize further deterioration and achieve stable, if not ideal, network conditions

Results are shown below. Implied expenditure is AMD24bn per year, with ratios of 40:45:15 for the three networks. The total represents a 20 percent increase on the level of maintenance expenditure prior to 2014. Expressed as a ten year average, it is within the range of BAU estimates in table 4. It represents 0.5 percent of 2015’s expected GDP. Under this strategy, an average of 120km per year would be subject to capital maintenance (overlays and rehabilitation) over the ten year period (well within historic capabilities).

This scenario is less backlog-heavy than the others examined and expenditure over the first five years differs little from the ten-year average.

Table 12: Proposed strategy: results



Network average IRI

2017-20



Interstate US$20m (AMD9.4bn) US$12,000 6.2 6.3 6.3

Republican US$24m (AMD11bn) US$12,000 6.2 6.4 6.3

Local US$7.4m (AMD3.5bn) US$2,000 - - -

All US$51m (AMD24bn) US$6,500 6.2b 6.3b 6.3b

Source: TA consultants’ HDM4 runs Note: (a) including VAT

(b) interstate and republican networks only

Figure 1 compares network condition for the three funding scenarios. Following the high expenditure scenario will eventually lead to improved network conditions, while constraining expenditure to around AMD21bn (table 10) results in a steadily deteriorating network. The proposed scenario manages to maintain network condition to within approximately 1.0 of today’s average IRI, but only on the interstate and republican networks.


Figure 1 Entire network: condition for alternative funding scenarios

9. Deferring capital maintenance It is frequently asserted that timely capital maintenance is good value for money. To quote the 2011 World Bank report: “for every dollar of deferred maintenance there is an associated cost of four dollars to road users” (this is a general result dating from 1988 and is not necessarily representative of Armenian conditions). Essentially the argument is this: deferring maintenance saves money, at least in the short term, but this is outweighed by the costs imposed on road users.

A limited investigation into the costs and benefits of deferring capital maintenance on the interstate network was undertaken. In the deferred scenario, only routine maintenance is undertaken until 2021; thereafter the strategy in table 6 is followed. When compared with the unconstrained scenario the deferred scenario costs US$108m less over the ten year period. In terms of network condition in 2025, the deferred scenario is rather worse – see the figure below.


Figure 2 Interstate network: deferred capital maintenance

2

3

4

5

6

7

8

9Av

erag

e IR

I in

m/k

m

Base case - nodeferralDeferredmaintenance

The penalty for delaying capital maintenance is increased road user costs (and, ultimately, higher freight rates and passenger fares). The table below sets out the costs and benefits over the 2015-25 period. In this instance the PV of road user cost benefits is approximately four times the incremental agency costs of undertaking unconstrained capital maintenance, lending some support to the World Bank’s thesis. Thus, timely capital maintenance is good value for money with national economic benefits that considerably exceed its costs.

Table 13: Costs and benefits of deferring capital maintenance

Change arising from timely maintenance

PV at 12%

US$m 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025

Increase in agency cost 92 1.3 27.9 57.7 27.8 12.1 67.8 -60.5 54.6 -15.8 -47.8 -17.5

Drop in road user cost 365 36.9 38.0 45.1 56.9 61.7 75.4 117.8 67.3 102.4 73.7 65.5

Net benefit of timely maintenance 273 35.6 10.2 -12.6 29.1 49.6 7.6 178.3 12.7 118.2 121.5 83.0

Benefit-cost ratio 4.0

Note: (a) routine maintenance is not included in the table

10. Summary of results This final draft strategy paper has assessed three levels of road maintenance funding. Funding at or close to BAU will result in a continuing deterioration in road network condition, with increased road user costs, ultimately passed to users in the form of higher freight rates and passenger fares. This deterioration is already evident from the decline in interstate network condition from 2010 to 2014 seen in table 2.

If all economically worthwhile maintenance investments are taken, network condition would slowly improve in the medium and long term. At least 60 percent of the paved network would be in a


maintainable condition by 2025. However, implied expenditure would be twice the level of maintenance expenditure to 2014 (and over three times proposed 2015-18 expenditure) and a little less than one percent of GDP. It would be similar to ADS target levels. This is considered unaffordable under current economic conditions.

The proposed strategy recommends concentrating capital maintenance on interstate and republican roads on the grounds that, other things being equal, the returns are higher on roads carrying higher traffic. Under this strategy, the condition of interstate and republican roads would remain within approximately 1.0 of today’s average IRI. The implied expenditure expressed as a ten year average is AMD24bn per year. This represents a 20 percent increase on levels of expenditure prior to 2014, but is within the upper end of the BAU estimates in table 4. It represents 0.5 percent of 2015’s expected GDP. This scenario is less backlog-heavy than the others examined and expenditure over the first five years differs little from the ten-year average.

Although this level of maintenance budget allocation is unlikely in the short term, the MoTC hopes that, as commitments to development projects such as the North-South road project reduce, maintenance funding will slowly increase to levels recommended here. Together with the implementation of proposals to improve the cost-effectiveness of maintenance works, set out in annex B, it is hoped that road conditions will slowly improve.

An assessment of the costs and benefits of timely capital maintenance, when compared with deferring it, shows that not deferring it is a sound investment, with the benefits of reduced road user costs approximately four times the additional costs of the maintenance.

11. Measures being taken and proposed directions Provided funds are allocated efficiently, road maintenance is a sound investment. Provided it is properly implemented, use of HDM4 to guide prioritization will ensure that the basic criterion, namely that a capital maintenance project’s net present value (NPV) is at least zero, is met.

The current scope for technical prioritization is limited and is confined to the shaded cells in figure 3. Apart from increased allocations to road maintenance in accordance with the proposed strategy, the way forward is to increase the role of prioritization. Steps are being taken to acquire state of the art pavement survey equipment that will enable ARD to implement a full road asset management system (RAMS) in time for the 2017 budget (see the action plan appended to this strategy).

It is proposed that MCA (multi-criteria analysis) will, from 2016, become part of prioritization. It will reflect the criteria set out in decree no 54 of 25 Dec 2014 (Maintenance and Development Prioritization decree), shown in table 11 below. The decree contains a number of qualifications. First, EIRRs will continue to be the principal criterion to select long lists of candidate roads to IFIs, following which they will be screened using the population and regional condition indicators described below, but with each carrying an equal weight of 0.5. Second, the decree assigns top population and condition scores to interstate roads, meaning that in practice EIRR is the only prioritization criterion when comparing one interstate road with another. Third, the state may override decree 54 priorities for a wide range of reasons and, fourth, ARD may propose a revision to the scope of MCA.


It is readily accepted that there are situations where an NPV criterion is unrealistic, for example when screening rural roads for which there are no traffic data and only visual pavement quality assessments are available. Under these circumstances, population within say 2km of each side of a road, provided the data are available at little cost, could be a useful aid to prioritization.

However, following implementation of the RAMS, it would be straightforward to use properly acquired pavement information in conjunction with moving observer traffic counts as the basis for prioritization.

Table 14: Decree no 54 multi-criteria analysis

Criterion Indicator Score Weight

Improvement of road communication (population): number of beneficiariesa

>5,000 3

0.3 1,000-5,000 2

<1,000 1

Economic efficiency: EIRRb

>12% 3

0.4 10-12% 2

<10% 1

Proportional regional development: (% of roads in region rated as “bad”c)

>40% 3

0.3 25-40% 2

<25% 1

Source: decree 54 Notes: (a) population in settlements connected by the proposed road (b) HDM4 output. EIRR is the only criterion applied to interstate roads (see text)

(c) applies to all roads (regardless of classification) in a region. “Bad” is established using the criteria of decree no 1419-N (based on potholes, shoulder and drain condition)



Figure 3 Capital and routine maintenance contracting cycle

Rehabilitation & other capital maintenance

Summer & winter maintenance of maintainable roads

Key:= some technical prioritization takes place

MOF budget

MTA

Loca

l roa

ds

5 ye

ar u

nit r

ate

cont

ract

s w

ith

mar

zes

Annu

al

spec

ifica

tion

of

km

Annu

al

agre

emen

ts w

ith

mar

zes

Annu

al

mai

nten

ance

w

orks

ARD

list o

f reh

ab

proj

ects

MOF budget

Road

s cou

ncil

appr

oval

Tran

spor

t PIU

for

cont

ract

ing

Reha

b et

c w

orks

Inte

rsta

te &

re

publ

ican

&

capi

tal w

orks

on

loca

l roa

ds

5 ye

ar u

nit r

ate

cont

ract

s w

ith

MO

TC

Annu

al

spec

ifica

tion

of

km

Annu

al

agre

emen

t with

M

OTC

ARD

inte

rsta

te &

par

tial r

epub

lican

H

DM n

etw

ork

data

base

s

Oth

er re

publ

ican

& lo

cal r

oads

: ARD

&

mar

z en

gine

ers'

ass

essm

ents

(no

traf

fic d

ata)

Non

-tec

hnic

al

prio

rity

list

HDM

hig

h EI

RR li

st

MOF budget


12. Budgetary issues Although the financing alternatives discussed above relate to the entire road network, the fragmented nature of budgeting for roads remains an issue. Budgeting for road maintenance is divided between (i) MOTC, (ii) marzpeteran (for local roads), (iii) communities and (iv) Yerevan. Some segmentation of budgeting is inevitable and can serve equity interests, but a common approach to prioritization would improve value for money for the limited resources available. The implementation of a RAMS will make this a possibility.

13. Outcome evaluation system A system of setting targets and reporting on outcomes is embedded in the MTEF. Many of the targets represent outputs (km maintained, bridges repaired etc) rather than outcomes. Outcome target KPIs are:

• average IRIs for the interstate, republican and local roads (but no local roads IRI included in the 2015 MTEF)

• proportions of networks in poor condition (as defined in decree 1419) for each of the three networks

The achievement of the past three years’ target KPIs is reported by the MoTC and included in the MTEF document. In the action plan it is suggested that the targets be revised once this strategy is updated (itself dependent on implementation of the planned RAMS).


ANNEX A HDM4 INPUTS

1 Pricing issues All predicted project costs and benefits are measured in 2014-15 economic prices. Traded goods are measured at world prices and non-traded inputs at domestic prices multiplied by an aggregate conversion factor (CF). Taxes and duties are excluded. All prices in the HDM4 analysis are expressed in US$. A standard conversion factor of 0.97 is applied to non-tradable items and labor. A shadow wage rate factor (SWRF) of 0.7 is used to estimate the economic price of unskilled labor, estimated to account for 5 percent of the construction cost. This leads to an aggregate CF of 0.96. (Estimates of the SCF, SWRF etc are taken from appendix 12 of the RRP for the ADB North–South Road Corridor Investment Program, project no 42145, Sep 2009).

At the time of reporting (April 2015) the official exchange rate was AMD473 = US$1.

2 Traffic growth and fleet composition

Previous work No traffic growth assumptions were made as part of TA8249. The road sub-sector analysis that was annexed to the North-South RRP does however include a set of traffic growth rates, themselves taken from earlier MCC work on forecast GDPs per head and income elasticities of 1.1 to 1.2. Their assumptions for the period 2015-25 are shown below. The World Bank’s 2011 paper assumed 5 percent traffic growth over their entire 20 year evaluation period.

Table A.1 Traffic growth assumptions for 2015-25, North-South Road Corridor Investment Program

Item 2015-25

GDP per head:

Low 4.0%

Central 5.0%

High 6.0%

Traffic growth:

Low 4.4%

Central 6.0%

High 7.2%

ARD 4.0%

ADB rural roads PPTA 6.0%

Source: RRP for ADB North-South Road Corridor

Time series of traffic counts Unusually, fairly long records of traffic volumes are available from nine automatic traffic count (ATC) stations on several trunk roads in Armenia from 2005 to 2014. They are summarized overleaf. As the data originate from ATCs they should not require adjustment for diurnal, weekly or seasonal effects.


The data serve to illustrate the enormous variability in Armenian traffic conditions, making it difficult to draw general conclusions. However:

• in general, traffic growth since stabilization following the low point of the economy in 2008-9 has been much lower than growth since 2005 (approximately 3 percent versus 13 percent), but conclusions may be biased by the apparent high growth rates between 2005 and 2006

• recent growth on the M2 from Yerevan to Iran has been close to or below zero for all vehicle classes (site 3)

• routes to Georgia have shown relatively strong growth, although this growth has conspicuously not applied to truck-trailers (sites 1, 4, 5 and 9)

• light vehicles at sites closest to Yerevan have tended to show strong growth (sites 1, 2, 6 and 8)

Looking only at traffic well beyond Yerevan’s urban influence in order to assess country-wide growth (sites 3, 4, 5, 7 and 9), annual growth of all traffic since 2010 has averaged 2.8 percent, somewhat lower than average annual GDP growth over the same period (4.0 percent). Light vehicle traffic grew at 8 percent annually over this period, while goods vehicle traffic declined sharply at around -20 percent.

The above suggests an income elasticity of around 0.7. This, though, would be uncharacteristically low and an overall income elasticity of approximately one is taken, with a higher elasticity for light vehicles (cars, SUVs, light goods and buses), say 1.1, and a lower elasticity for goods vehicles, including truck-trailers, say 0.8.

Forecast traffic growth rates GDP growth forecasts for Armenia have fallen sharply over the three months since January 2015. In January 2015 the World Bank was forecasting 3.3, 3.7 and 4.1 percent for 2015-17. The EIU was making similar forecasts at the same time: 2.7 and 3.4 percent for 2015-16.

By late March, however, ADB had reduced its forecast growth rates to 1.6 and 2.3 percent for 2015-16 (source: Asian Development Outlook 2015), and the World Bank’s Armenia country manager forecast 0.8 percent for 2015 at a news conference in Yerevan on 31 March.

For the purposes of forecasting traffic growth a GDP growth rate of 2 percent is adopted and for the HDM4 projections the following are adopted:

• 2.2 percent for small vehicles • 1.6 percent for goods vehicles (including truck-trailers)

These growth rates are lower than those used by other studies, as shown in the table the start of this section, but are justified by the weaker economic outlook.

Fleet composition At this strategic level it is sufficient to use standard fleet compositions. Two are used:

• for roads with initial traffic up to 3,000 veh/day, a fleet composition derived from the most recent set of LRNIP roads is used (400km with average traffic of 1,400 veh/day)


• for roads carrying initial traffic greater than 3,000 veh/day, the average fleet composition for 2009-14 for ATC stations 3, 4, 5, 7 and 9 is taken, then proportionately subdivided according to the LRNIP composition (e.g. the ATC percentage for goods vehicles is allocated between light, medium and heavy goods vehicles using the same split as in the LRNIP fleet)

The two fleet compositions are shown below.

Table A.2 Vehicle fleet compositions

Traffic level Car LGV Minibus Med bus Large bus MGV HGV T-T

≤3,000 veh/day 84.50% 5.30% 4.80% 0.60% 0.10% 3.10% 1.30% 0.30%

>3,000 veh/day 76.60% 8.20% 4.40% 0.60% 0.10% 4.90% 2.00% 3.20%


Table A.3 Interstate traffic data, 2005-14, veh/day, sites 1-4

Year GDP 1: M1 20+700 Yerevan-Gyumri-

Georgia 2: M2 9+380 Yerevan-Meghri-Iran 3: M2 200+320 Yerevan-Meghri-Iran 4: M3 135+160 Margara-Vanadzor-Georgia

100 Light veh Goods T-T Total Light veh Goods T-T Total Light veh Goods T-T Total Light veh Goods T-T Total

2005 13.9% 114 8,316 1,351 71 9,738 17,032 4,963 630 22,625 560 185 15 760 569 282 10 861 2006 13.2% 129 8,813 1,428 82 10,323 17,586 5,389 624 23,599 726 238 38 1,002 487 877 53 1,417 2007 13.7% 147 6,195 3,441 317 9,953 18,352 5,673 526 24,551 No data 903 416 62 1,381 2008 6.9% 157 6985 4306 329 11,620 13,111 7,057 797 20,965 No data 1,362 714 402 2,478 2009 -14.2% 134 8,649 3,454 281 12,384 15,552 7,216 638 23,406 3,129 779 226 4,134 1,562 270 82 1,914 2010 2.2% 137 9,435 2,089 183 11,707 22,451 6,401 523 29,375 3,128 1,397 417 4,942 2,189 656 87 2,932 2011 4.7% 144 14,196 1,364 189 15,749 26,484 5,198 593 32,275 3,041 1,324 328 4,693 1,775 323 22 2,120 2012 7.2% 154 13,863 2,555 360 16,778 26,546 4,674 466 31,686 2,310 645 127 3,082 2,093 251 35 2,379 2013 3.5% 160 14,842 1,712 143 16,697 29,505 3,713 286 33,504 2,506 302 98 2,906 2,166 160 14 2,340 2014 2.6% 164 No data 30,299 2,676 188 33,163 3,677 304 90 4,071 2,287 135 6 2,428

AAGR 2005-14 % 9.6% 18.1% 42.1% 7.0% 9.5% -7.8% -13.2% 5.5% - - - - 18.8% 20.7% 87.7% 17.6% AAGR 2010-14 4.0% 18.4% 6.5% 11.2% 13.5% 8.0% -19.3% -20.2% 3.2% 7.1% -27.3% -28.4% -1.2% 2.0% -31.2% -33% -3.3%

Sources: ADB N-S RRP (2005-7) and ARD (2008-14) Notes: (a) AAGR = average annual growth rate (b) GDP growth rates from World Bank sources


Table A.4 Interstate traffic data, 2005-13, veh/day, sites 5-9

Year

5: M3 125+830 Margara-Vanadzor-Georgia

6: M4 10+320 Yerevan-Sevan-Ijevan-Azerbaijan

7: M4 80+300 Yerevan-Sevan-Ijevan-Azerbaijan

8: M5 9+400 Yerevan-Armavir-Ijevan-Turkey

9: M6 59+050 Vanadzor-Alaverdi-Georgia

Light veh Goods T-T Total Light veh Goods T-T Total Light veh Goods T-T Total Light veh Goods T-T Total Light

veh Goods T-T Total

2005 1,900 290 20 2,210 15,110 4,288 86 19,484 1,338 104 13 1,455 17,572 1,425 49 19,046 532 141 48 721

2006 2,251 985 112 3,348 17,933 4,833 111 22,877 2,126 144 31 2,301 16,946 2,829 40 19,815 702 571 92 1,365

2007 2,372 1,578 245 4,195 17,605 6,853 928 25,386 825 103 33 961 18,863 3,477 862 23,202 670 504 123 1,297

2008 2,868 1,721 210 4,799 15,017 5,912 576 21,505 438 271 74 783 14,798 5,356 1,313 21,467 1,174 681 261 246

2009 4,246 1,394 191 5,831 15,894 5,561 393 21,848 1,480 403 125 2,008 21,626 603 60 22,289 1,444 590 173 2,207

2010 4,326 966 137 5,429 22,803 6,450 528 29,781 1,949 316 33 2,298 22,944 2,186 46 25,176 2,063 250 164 2,477

2011 4,348 956 120 5,424 25,740 5,256 695 31,691 1,687 255 33 1,975 22,957 2,875 49 25,881 1,802 348 174 2,324

2012 4,400 876 56 5,332 25,707 4,722 553 30,982 3,691 690 121 4,502 24,198 3,589 59 27,846 1,492 143 41 1,676

2013 4,653 458 88 5,199 28,305 2,990 203 31,498 3,369 503 94 3,966 26,019 1,833 46 27,898 2,074 176 79 2,329

2014 5,453 285 82 5,820 30,688 1,692 114 32,494 3,593 341 46 3,980 27,187 856 21 28,064 3,159 271 94 3,524

2005-14 13.3% 18.3% 56.9% 12.6% 9.2% -6.4% 70.4% 6.7% 36.0% 32.5% 50.9% 28.3% 6.2% 33.7% 213% 4.6% 25.6% 35.0% 26.7& 24.5%

2010-14 6.2% -23.7% -3.9% 1.9% 7.8% -27.2% -24.0% 2.3% 25.8% 23.0% 48.3% 25.6% 4.4% -11.5% -12.4% 2.8% 15.4% 14.3% 10.3% 14.1%

Sources: ADB N-S RRP (2005-7) and ARD (2008-14)


2 Vehicle characteristics and cost, labor costs and values of time Vehicle characteristics and the costs of vehicles and labor, and the values of working and non-working time are shown in the following tables.

Vehicle costs exclude taxes and are the local Armenian costs of new vehicles. They are considered representative of world prices of the types of vehicle used in Armenia and no further adjustment is made.

Representative wages range from AMD3-5,000 per day for a laborer, through AMD100,000/month for a truck driver to AMD300,000 for a chief engineer. Social costs add around 24-26 percent. Approximate labor costs per hour are:

• crew and maintenance labor, large vehicles and SUVs etc: AMD700 (US$1.50) per hour • crew and maintenance labor, small vehicles: AMD450 (US$1.00) per hour

In the absence of stated preference results, values of working time are based on local wages:

• for car and SUV passengers: AMD1,500 (US$3.00) per hour • for marshrutka and bus passengers: AMD500-1,000 (US$1-2) per hour

Non-working time is valued at 0.25 times that of working time.


Table A.5 Vehicle characteristics

Vehicle class Large car Minibus (marshrutka)

Medium bus Bus Light goods

Medium goods Heavy goods 4x2 tractor + tandem trailer

HDM-4 base type Car large Minibus Bus medium Bus heavy Light truck Medium truck Heavy truck Articulated truck

PCUs 1.0 1.3 1.5 2.0 1.3 1.5 2.0 2.5

Fuel type Petrol/CNG CNG CNG CNG Diesel Diesel Diesel Diesel

No of tyres 4 4 6 6 6 6 10 14

Tyre type Radial Radial Radial Bias Radial Bias Bias Bias

No of axles 2 2 2 3 2 2 3 4

Km/year 10,000 80,000 100,000 100,000 100,000 100,000 100,000 100,000

Hours driven/year 200 2,000 2,000 2,000 2,600 2,600 2,600 2,600

Working time, h/yeara 2,000 2,200 2,200 2,200 3,300 3,300 3,300 3,300

Service life (yr) 15 12 10 10 8 8 8 8

Av no of passengers 3 12 25 40 1 1 0 0

% time private use 100 10 0 0 20 0 0 0

Work related passenger trips, % 50 20 20 10 0 0 0 0

Gross vehicle weight, t 1.4 3.20 7.5 15.0 3.00 13.00 20.00 30.00 Source: manufacturers’ and consultant’s estimates Notes: (a) working time = driving time + time spent loading, unloading and waiting for work. Working time has to be specified for HDM-4. Large values used for cars and motor-cycles in order to avoid spurious

savings in time related VOCs


Table A.6 Vehicle, tire and labor costs, and values of time

Unit Large car/SUV

Marshrutka Medium bus Bus LGV MGV HGV Articulated truck

Typical model: Camry/Land-Cruiser

Gazel NeoLaz Laz Lemberg Uaz pickup Uaz 3303 Kamaz 8 tnn

Vehicle cost USD 23,000 18,000 30,000 43,000 19,000 21,000 40,000 90,000

Replacement tire USD 125 80 170 330 80 170 330 330

Maintenance labor USD/h 1.50 1.00 1.20 1.50 1.00 1.20 1.50 1.50

Crew USD/h 0 1.20 1.20 1.20 0 1.20 1.20 2.40

Annual overhead USD 900 1,100 2,300 10,000 1,200 2,700 6,500 6,500

Passenger working time USD/h 3.00 1.50 1.00 1.00 1.00 1.00 0 0

Passenger non-working timea USD/h 0.75 0.38 0.25 0.25 0.25 0.25 0 0

Source: local dealers, recent construction contracts, consultants’ estimates Note: (a) 25 percent of value of working time


3 Fuel costs There is no refinery in Armenia. Russia is the dominant supplier of refined products and CNG although Iran supplies a small proportion of CNG. Gasoline is subject to customs fees, excise duty and VAT at the standard rate of 20 percent. Diesel is subject to customs fees, excise duty and a small environmental tax, but not to VAT. CNG appears to be free of customs and excise duties but is subject to VAT.

Gasoline retails at AMD450-500, depending on octane rating. If VAT and customs and excise duties are deducted, the price net of tax is AMD350-390. The lower end of this price range corresponds to the border parity price for gasoline refined from crude priced at US$55/bbl, i.e. around US$0.73/liter. Thus the price net of taxes and duties is a reasonable approximation of the economic price.

Diesel retails at AMD450-470, the higher price being for Euro (ultra-low sulfur) diesel. As diesel does not carry VAT, it is evident that there are large margins in the Armenian supply chain. For HDM4 it is reasonable to assume an economic price for diesel of US$0.70/liter, i.e. slightly lower than gasoline as it has slightly lower refining costs.

Per equivalent liter of gasoline, CNG’s retail price is approximately a third of that of gasoline, meaning that a representative economic price would be approximately US$0.25/liter. Many small vehicles, perhaps as many as 70 percent, are able to use CNG. For the economic evaluation it is assumed that CNG represents 40 percent of gasoline and CNG consumption, giving an effective economic price for gasoline of US$0.54/liter.

4 Road conditions A set of the most salient road conditions is associated with each of the four roughness bands used in the HDM4 strategy analysis, drawn from the ARD HDM4 databases.

Table A.7 Road conditions

R1 R2 R3 R4

Interstate

IRI roughness, m/km 3.6 5.0 7.5 11.9

All structural pavement cracks, % 4% 5% 7% 10%

Potholes/km 10 50 80 200

Last rehabilitation 2012 2010 2008 2001

Republican

IRI roughness, m/km 3.3 5.0 7.5 11.9

All structural pavement cracks, % 4% 8% 11% 15%

Potholes/km 10 50 80 200

Last rehabilitation 2012 2011 2008 2002


ANNEX B ACTION PLAN AND TIMETABLE

No Action Aim Timing Responsible Source of finance

Notes

1

Continue IRI survey with existing equipment (principally interstate roads)

Have working RAMS used for prioritization of all periodic maintenance and rehabilitation works for 2017 budget year

Annual

Ministry of Transport and Communication, “Armenian road directorate” SNCO

State budget of RA

Limited scope (1,500km/yr)

2 Continue video recording to cover entire network By end May 2015


State budget of RA

Stop once RAMS implemented

3 Procure pavement and road survey equipment and RAMS software

By end October 2015


External source

Under procurement

4 Survey equipment and software installation and training

Nov-Dec 2015 Ministry of Transport and Communication,

No financing needs

5 Prepare pavement condition survey schedule Jan-Feb 2015

Ministry of Transport and Communication,

No financing needs

Scope will depend on staffing; to meet minister’s order of 8 Apr 2014 should be entire network annually



Notes

6 Undertake pavement condition survey Annually from Spring 2016


No financing needs

7

Install additional automatic traffic count (ATC) stations, say 10 on interstate and 20 on republican network

Budget-dependent


No financing needs

Use existing interstate ATC technology

8 Revise target pavement quality standards in light of available funding

Realistic targets For 2017 budget year Ministry of Transport and Communication,

No financing needs

Dependent on implementation of reliable RAMS

9 Review decree 54 criteria Screen candidate roads with sparse traffic data

Before 2017 budget year Ministry of Transport and Communication,

No financing needs

Consider revising criteria and replace decree with use of RAMS and moving observer traffic counts

10 Define core highway network

More cost-effective use of limited funds Before 2017 budget year


No financing needs

With improved traffic data and strategic priorities, identify priority road sections for capital maintenance

11 Consider adoption of low volume road standards

More cost-effective use of limited funds 2015-20


No financing needs

Current standards do not include a single lane road, suitable for village access

12 Review maintenance interventions; consider more use of slurry seals



No financing needs



Notes

and surface dressing

13 Assess options for unpaved roads



No financing needs

2. Network and responsibilities - hti.amhti.am/edfiles/files/voroshumner/AARCroadsfinancingstrategy.pdf · AARC final draft road maintenance strategy June 2015 2 (c) length under

Documents