COST-BENEFIT ANALYSIS FOR INVESTMENT DECISIONS, CHAPTER 17: APPRAISAL OF UPGRADING A GRAVEL ROAD Glenn P. Jenkins Queen’s University, Kingston, Canada and Eastern Mediterranean University, North Cyprus Chun-Yan Kuo Queen’s University, Kingston, Canada Arnold C. Harberger University of California, Los Angeles, USA Development Discussion Paper: 2011-17 ABSTRACT The purpose of this chapter is to illustrate how a proposed investment in upgrading a gravel road to a tarred surface should be evaluated. The project is located in the Limpopo Province of South Africa. It involves upgrading of two existing, mainly gravel, roads into a tar surface road connecting Sekhukhune and Capricorn districts. The whole route has several sections starting from Flag Boshielo to Mafefe, Sekororo, Ga Seleka and finally to Mmatladi. It has been estimated that more than 98% of the sections of gravel road are considered in either poor or fair condition. The main users of the existing gravel road are mini-buses and private vehicles transporting people from local areas to Lebwakhomo and other towns. The predominant economic activity in the region is small-scale agriculture, carried out on a number of irrigation schemes. An lesson from this case is the importance of evaluating segments of a road separately if the traffic on the segments or the cost of upgrading are significantly different across segments. To be Published as: Jenkins G. P, C. Y. K Kuo and A.C. Harberger, “Appraisal Of Upgrading A Gravel Road” Chapter 17.Cost-Benefit Analysis for Investment Decisions. (2011 Manuscript) JEL code(s): H43 Keywords: Road upgrade, maintenance cost savings, road segmentation, vehicle operating costs, time savings,
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
COST-BENEFIT ANALYSIS FOR INVESTMENT DECISIONS,
CHAPTER 17:
APPRAISAL OF UPGRADING A GRAVEL ROAD
Glenn P. Jenkins Queen’s University, Kingston, Canada
and Eastern Mediterranean University, North Cyprus
Chun-Yan Kuo Queen’s University, Kingston, Canada
Arnold C. Harberger
University of California, Los Angeles, USA Development Discussion Paper: 2011-17 ABSTRACT The purpose of this chapter is to illustrate how a proposed investment in upgrading a gravel road to a tarred surface should be evaluated. The project is located in the Limpopo Province of South Africa. It involves upgrading of two existing, mainly gravel, roads into a tar surface road connecting Sekhukhune and Capricorn districts. The whole route has several sections starting from Flag Boshielo to Mafefe, Sekororo, Ga Seleka and finally to Mmatladi. It has been estimated that more than 98% of the sections of gravel road are considered in either poor or fair condition. The main users of the existing gravel road are mini-buses and private vehicles transporting people from local areas to Lebwakhomo and other towns. The predominant economic activity in the region is small-scale agriculture, carried out on a number of irrigation schemes. An lesson from this case is the importance of evaluating segments of a road separately if the traffic on the segments or the cost of upgrading are significantly different across segments. To be Published as: Jenkins G. P, C. Y. K Kuo and A.C. Harberger, “Appraisal Of Upgrading A Gravel Road” Chapter 17.Cost-Benefit Analysis for Investment Decisions. (2011 Manuscript) JEL code(s): H43 Keywords: Road upgrade, maintenance cost savings, road segmentation, vehicle operating costs, time savings,
CHAPTER 17:
1
CHAPER 17
APPRAISAL OF UPGRADING A GRAVEL ROAD
17.1 Introduction
The purpose of this chapter is to illustrate how a proposed investment in upgrading a gravel road
to a tarred surface should be evaluated. The project is located in the Limpopo Province of South
Africa. It involves upgrading of two existing, mainly gravel, roads into a tar surface road
connecting Sekhukhune and Capricorn districts. The whole route has several sections starting
from Flag Boshielo to Mafefe, Sekororo, Ga Seleka and finally to Mmatladi. According to the
Roads Agency Limpopo (RAL), the proposed road consists of sections D4100, D4250, D4190,
D4050 and D1583, with the exception on D4100 where a section of 25 km was already tarred. It
has been estimated thatmore than 98% of the sections of gravel road are considered in either poor
or fair condition.1
The main users of the existing gravel road are mini-buses and private vehicles transporting
people from local areas to Lebwakhomo and other towns. The predominant economic activity in
the region is small-scale agriculture, carried out on a number of irrigation schemes. At present,
no specific tourist sites are operational in the area, but it is expected that the Lekgalameetse
Nature Reserve may become a tourist attraction in the near future.
The project is expected to serve some 35,000 people living in the immediate vicinity of the route,
and provide a convenient access to the existing and future developments in agriculture, tourism
and mining sectors.
The section of the proposed road consisting of segments D4100, D4250, and D4190 is about
81km long and it is part of the Spatial Development Rational, Golden Horse-shoe and
theDilokong sub-corridor.It is expected to support the Provincial Economic Development
1ARCUS GIBB Ltd., “Limpopo Integrated Infrastructure Development Plan: Phase II – Benefit Cost Analysis of Selected Projects”, Final Report, Appendix I: Flag Boshielo to Mafefe to Sekororo and Ga Seleka to Mmatladi, (March 2004).
CHAPTER 17:
2
Strategy.At the present time, this road serves a number of communities including 20 villages that
are located directly on its wayincluding the town of Lebwakhomo, and the communities around
the Flag Boshielo Dam. Upgrading this link will ensure a convenient accessfor the regional
population to the Lebwakhomo and Groothoek Hospitals, Jane Furse and Lebwakhomo Police
Stations, and possible future sites of agriculture and tourist projects.
The other component of the proposed road improvement, consisting of segments D4050 and
D1583, is about 75km long. This section already serves more than 28 villages located directly on
the route, the town of Lebwakhomo, and the Lekgalameetse Nature Reserve. The improved road
will facilitate an easier access to the hospitals in Lebwakhomo, Groothoek and Sekororo, as well
as to police stations in Jane Furse and Lebwakhomo. Once improved, this road will provide a
direct link to Tzaneen and Phalaborwa, making it convenient forvehiclesto travel across the
Province.
17.2 Project Costs
The project was proposed to take three years to construct, starting in 2005 and ending in 2007.
For sections D4100, D4250 and D4190 that pass through a relatively flat terrain and comprising
about 81 km, an average construction cost of R 1.301 million per km was estimated. For sections
D4050 and D1583 that are located in mountainous area and stretching for about 75 km, the
estimated costs of upgrading are higher, averaging R 1.459 million per km.
It is typical to include some provision for linking roads, which will connect the upgraded road
with other roads and projects en-route. About 15 km of linking roads were estimated as a part of
this road improvement project. An average construction cost of these linking roads is expected to
be R 0.700 million per km. Ten small river-crossings are also included in the project; their
estimated costis R 0.040 millionper km.2
In addition to the physical construction costs, professional fees are levied at 12% of the total
construction expenditures. A provision for contingencies also accounts for additional 10% of the
2 Ibid, p. 3-2.
CHAPTER 17:
3
total construction costs. In South Africa, the VAT at 14% rate is imposed on the total
construction costs, exclusive of professional fees and contingencies.
In terms of timing, Sections D4100, D4250 and D4190, located in Sekhukhune district, have a
higher traffic volume and will be upgraded in the first phase, starting at beginning of year 2005.
The second phase of construction will upgrade Sections D4250 and D4190, located in Capricorn
district in 2006. The last phase will upgrade one-third of 75 km of D4050 and D1583 located in
mountainous area in 2006 and the remaining 50 km in 2007.3 It is assumed that the costs of
linking roads, river crossings, professional fees, contingencies and VAT will be evenly spread
over three years.
The total tax-inclusive investment cost of the project over three years is expected to be R 307
million in 2005 prices. The detailed cost breakdown of total investment is presented by road
section and time schedule in Table 17.1.Road sections D4100 and D4250 will be upgraded first
in 2005 and 2006 and sections D4190, D4050/D1583 will follow in 2006 and 2007.
3 In 2005, 34 km of D4100 and 27 km of D4250 will be built while 7 km of D4250, 20 km of D4190, and 25 km of D4050/D1583 will be built in 2006. Finally, 50 km of D4050/D1583 will be constructed in 2007.
CHAPTER 17:
4
Table 17.1: Breakdown of Project Investment Costs (millions of Rand in 2005 Prices)
The roads of this project are owned and operated by the RAL. There is no toll imposed on road
users now, nor will it be tolled after the roads are upgraded from gravel to tarred surface. As
such, no financial revenues are expected from road users. Therefore,nofinancial evaluation will
be carried out in this project. The financial outlays by the RALwill simply follow the time path
of project expenditures.The objective of this chapter is to examine whether this investment
promises to increase the economic welfare to residents of South African society as a whole.
To evaluate the economic impact of upgrading a gravel road, one has to measurehow its effects
that differ from what one would likely have observed in the absence of the project. This
incremental impact analysis entails developing two alternative scenarios: “with” and “without”
the proposed road improvement. The “without” scenario, which assumes the absence of the
project, does not contemplate anymajor rehabilitation or capital outlaysthat will be spent on the
existing gravel roads. It does, however, assume that regular normal maintenance and
rehabilitationoperations will continue on these roads,so that the incremental impact of the
proposed project will not be overstated when compared to the “without” project scenario.
CHAPTER 17:
5
The capital expenditures of a tarred surface are typically justified by its lower annual
maintenance costs as compared to a gravel surface.However, several other types of benefit must
be accounted for when conducting the evaluation from the economicpoint of view. They should
include reduction in vehicle operation costs for road users due to the improved road surface, time
savings for road users due to the increasedaverage speed of vehicles, and a possible reduction in
the costs of accidents and other fiscal externalities.
Once the road is upgraded, road users will commence to travel on the tarred road. Since the total
construction phase of this project will take three years and each section of the road will take
approximately six months to upgrade, some improved sections may serve longer than others if
the project is terminated at the same time. For the purpose of this evaluation, the project is
assumed to last at least 20 years until 2027 and no salvage value remaining.
In measuring the economic benefits of transportation projects, one must distinguish between
those who would use the existing road even in the absence of the improvement, and those whose
travel would be newly induced as a consequence of the improvement. The benefits to the first
group are measured in reduction of vehicle operating costs and time costsbetween traveling on
the gravel and the tarred road. The benefits to the second group are measured by one half of such
savings in vehicle operating costs and time costs (see Chapter 16).
To ensure a consistent transformation from all the financial costs into the economic costs used in
the economic evaluation, a number of adjustments are made to convert these financial values into
their corresponding economic values. To do this, Commodity Specific Conversion factors for
several key project input variables are estimated, based on the methodology outlined in Chapters
10 and 11.
After all the annual benefits and costs are estimated for the “with” and “without” project
scenarios, the incremental net benefits are discounted over the project life by the economic
opportunity cost of capital for South Africa to see if the net present value is greater than zero.
CHAPTER 17:
6
In what follows, we will first examine individually thetraffic forecasts“with” and “without” the
project, plus the savings in maintenance costs, the vehicle operating costs,and time costs for each
type of vehicle, and then assess the project in terms of its economic feasibility, its impact on the
stakeholders affected by the project and finally, the risk inherent with this project.
17.4 Maintenance Costs
The upgraded road is expected to require substantially less maintenance care in terms of costs
and repair frequency as compared to the existing gravel surface. In the case of the “without”
project scenario, maintenance activities will include all regular and periodic maintenance
expenditures and the rehabilitation costs of the existing road, in order for it to be held within the
maintenance standards of theRoad Agency. Table 17.2 presents the engineering estimates of
maintenance costs of tarred (“with” project) and gravel (‘without” project)roads per kilometer by
type and frequency of maintenance activity for 2004. These estimates are then adjusted to year
2005, based on the annual inflation rate of 6.5%.
Table 17.2: Road Maintenance Costs for Tarred and Gravel Road
(millions of Rand per km)
Amount Road Surface
Type of Activity
Frequency 2004 2005
Tarred(With Project)
Routine Intermediate
Periodic
Annual Every 3 Years
Every 10 Years
0.030 0.150 0.500
0.032 0.160 0.533
Gravel (Without Project)
Blading Wearing Course Heavy Gravel
Annual Every 2 Years Every 5 Years
0.035 0.200 0.350
0.037 0.213 0.373
Sources: ARCUS GIBB Ltd., “Limpopo Integrated Infrastructure Development Plan: Phase II – Benefit Cost Analysis of Selected Projects”, Final Report, Appendix I: Flag Boshielo to Mafefe to Sekororo and Ga Seleka to Mmatladi, (March 2004).
As previously mentioned, the construction of the project starts in 2005 in certain sections of the
road and ends in 2027 for the purpose of this evaluation.
Given the estimates of the above maintenance costs per kilometer and the length of upgrading of
various road sections, the annual financial maintenance costs are estimated and presented in
Table 17.3 for “with” and “without” project scenarios over the life of the project. One can then
CHAPTER 17:
7
estimate annual savings in financial maintenance expenditures after roads are upgraded from
gravel to tarred surface.
Table 17.3: Estimates of Annual Financial Maintenance Costs
(millions of Rand in 2005 Prices)
Tarred Road (With Project) Gravel Road (Without Project) Routine Intermediate Periodic Total Routine Intermediate Periodic Total
The projected demand for traffic is the most important element in the economic analysis of a
road project. The traffic forecast model used in the present analysis is based on astudy completed
for the Road Agency, and most of the parameters and assumptions of its model are kept
unchanged. The model is built around six groups of road users, differentiated by vehicle type and
purpose of journey: heavy goods vehicles (HGV), light goods vehicles (LGV), agriculture
transport, tourists, passenger cars, and mini buses. For practical purposes, we combined LGV
with agriculture transport. Thus, our traffic projections are carried out for five types of traffic.
The projected demand for traffic must be forecasted over the life of the project for each of the
five vehicle categories under both the “with” and “without” project scenarios.
CHAPTER 17:
8
17.5.1 Traffic Level Without the Project
Given the generally poor road conditions indicated earlier,there is low traffic volume on the
existing gravel road. The main users of the road are mini-buses and private vehicles transporting
people from local areas to Lebwakhomo and other towns. The predominant economic activity in
the region is small-scale agriculture, carried out on a number of irrigation schemes. No specific
tourist sites are operational in the area, but it is expected that the Lekgalameetse Nature Reserve
will become a tourist attraction in the near future. The improved road will also provide a direct
access for tourists from Flag Boshielo area to Tzaneen and Phalaborwa. Economic activity in the
region is being stimulated by gradual development of mining resources.
In 2003, the total annual average daily traffic (AADT) was about 285 vehicles in D4100, 380 in
4250, 380 in 4190, and 60 in D4050/D1583. The proportions of total traffic on the first three
roads were 72% for mini-buses, 26% for passenger cars, and 2% for heavy goods vehicles. On
the remaining road sections D4050/D1583, the total traffic wassplit equally between mini-buses
andprivate passenger cars.
Passenger Cars
The initial levels of passenger car traffic in 2003 for sections D4100, D4250, D4190, D4050,
D1583 were calculated from the total AADT counts multiplied bythe estimated proportions of
the traffic type. The average volumes of passenger car traffic on these segments were found to be
75, 100, 100, 30 and 30 vehicles per day in 2003.4For later years the volume of passenger traffic
on each segment is assumed to grow by 4.0% over the life of the project until 2027.5
Tourists
4 For instance, the share of passenger traffic on D4100 segment is 26%, and total AADT is 285. Then, the number of passenger cars is 75. 5 Over the past ten years since 2001, the annual GDP growth rate in South Africa was about 3.4%.
CHAPTER 17:
9
Tourist trips are expected to follow sections D4050 and D1583, starting in 2005 with AADT of
6. For the following years, the traffic is expected to rise annually by 4%.
Mini-Buses
The annual increase in mini-bus traffic is linked to the growth of passenger traffic, and the
volume on all road segments rises by 4.0% per year. The initial AADT counts for sections
D4100, D4250, D4190, D4050 and D1583 were estimated at 204, 272, 272, 30 and 30,
respectively. The assumed 4.0% growth rate of passenger and mini-bus traffic is considered a
conservative estimate of traffic volume.
Agriculture and Light Goods Vehicles
A number of small irrigation schemes are located within reach of the D4100 section. Most of
these schemes are expected to become operational in the next four years as the Department of
Agriculture completes rehabilitation and transfer of the affected properties to their farmowners.
The improvement of the road will bring about reduced costs of transport. Agriculture and LGV
traffic is expected to start in 2005 with AADT of 4.0. The future growth rate of agriculture and
LGV on D4100 section is assumed to be 5.0%. On sections D4190 and 4250, the movement of
LGV and agriculture transport will start in year 2005 with AADT of 4.0 and then gradually reach
AADT of 6.0 in year 2007, thereafter a constant growth rate of 5.0% is assumed. No LGV and
agriculture traffic is expected on sections D4050 and D1583.
Heavy Goods Vehicles
Most of HGV traffic on the proposed road is expected to originate from the irrigation schemes,
which will be operational in the next few years. Agricultural produce grown on the farms will be
transported to bigger towns of Polokwane, Lebwakhomo, and possibly Burgersfort. For the
D4100 segment of the proposed road, the HGV traffic will most likely consist of agriculture
transportation, plus a very few mining vehicles. In the absence of firm plans for mining
development, it is difficult to predict what will be the additional mining HGV traffic volumes.
CHAPTER 17:
10
For this particular segment, the traffic volume is expected to gradually grow from AADT of 6.0
in 2003 to AADT of 15.0 in 2010, after which an annual growth rate of 5.0% is assumed. The
possible construction of the Flag Boshielo dam wall would bring added traffic to this segment for
two years, but no firm decision has been taken in this regard, this traffic is not included in the
forecast.
On road sectionsD4250 and D4190, the 2003 count of 8.0 AADT is modeled to rise by a rate of
5.0% annually throughout the entire period of 2003-2027. No regular HGV traffic is expected on
sections D4050 and D1583.
The traffic projections for the “without” project scenario over the life of the project are presented
in Table 17.4.
Table 17.4: Projected Traffic by Road Section and Vehicle Type for “Without” Project Scenario
Since the conditions of the existing road vary significantly among road sections, vehicle
operating costswill also differ by section as well as by vehicle type.
Vehicle operating costs (VOC) include consumption of gasoline and oil, the wear-and-tear on
tires, and the repair expenditures on vehicles. Their estimates for this project are based on the
Roads Economic Decision (RED) model developed by the World Bank, and modified for South
Africa by CSIR Transportek in 2003.6 Estimates are made by type of vehicle and by terrain,
depending upon the degree of roughness of road, measured according to the International
Roughness Index. The VOC is expressed as a function of the degree of road roughness (see
Appendix 17A).
Since the original model’s output was expressed in 2003 prices, VOCs for “with” and “without”
project scenarios for different road sections were estimated by varying the degree of road
roughness in the prices of 2003. The estimates were then adjusted by an annual inflation rate of
6.5% over the next two years to 2005 prices. For example, on section D4100 the VOC for private
passenger cars was originally estimated from the RED model at R 3.261 per vehicle km in 2003
prices. This estimate applied to flat terrain in the absence of road improvement, and to a degree
of road roughness measured at 10.0 (see Appendix 17A). The cost was then adjusted for inflation
to become R 3.699 expressed in 2005 prices. After the road is upgraded from gravel to tarred
surface, the road roughness is improved from the index 10.00 to index 2.0. The resulting VOC,
re-estimated from the RED model and adjusted for inflation, was R 2.500 per vehicle km in 2005
prices.
The same procedure was used to estimate the average VOC per vehicle km for other vehicle
types and other road sections. Estimates of the average VOC for each vehicle type traveling on
6Archondo-Callao, R., “Roads Economic Decision Model (RED) for Economic Evaluation of Low Volume Roads: Software Users Guide”, Version 2.0, 3/15/01, the World Bank, Washington, D.C. The model was customized for South African conditions by CSIR Transportek at the request of the Development Bank of Southern Africa (DBSA). Available at <http:www.dbsa.org>.
CHAPTER 17:
15
each road section before and after the road improvement are presented in Table 17.7.These data
provide the basis for our later estimates of total project-induced annual savings in vehicle
operating costs.
Table 17.7: Vehicle Operating Costs for “With” and “Without” Project Scenarios
In addition to the vehicle operating costs, time cost of occupants travelling on the road can also
be an important factor in the economic evaluation of the road improvement project. The time
cost of the travellers can be determined by the speed of the vehicle and the time value of
travellers. The former will be influenced by the condition of the road and the volume of the
traffic while the latter is related to the wages and salaries of the driver and other occupants in the
vehicle.
Since this project is located in a low traffic volume region(see Section 17.5), vehicle speed is
unlikely to be affected by the volume of the traffic. Rather, the average vehicle speed is
determined by the roughness of the road. As a consequence, the average vehicle speed measured
in this project is based on the RED model developed by the World Bank, taking into
consideration of the specific conditions in the Limpopoprovince of South Africa. Our speed
estimates are for different vehicletypes and various terrain and road conditions. As with the
VOCs, speed is measured as a function of the degree of roughness.7 The estimating equations
7 Passenger cars and tourist traffic correspond to “car” class, mini-buses are linked to “light bus” class, LGV and agriculture transport are presumed to be in “light truck” class, while HGV corresponds to “heavy truck” class. The
CHAPTER 17:
16
used for each vehicle type were originally estimated for year 2003, and are shown in Appendix
17B.
The results are presented in Table 17.8. For example, without the road improvement the average
vehicle speed for passenger cars traveling on the gravel road D4100 is 68.8km per hour, using an
international roughness index of 10.0. With the road upgraded to a tarred surface, the roughness
index is reduced to 2.0 and the vehicles can reach average speeds of 86.6 km per hour. This
increase in vehicle speed, together with thevalue of time per hour allows us to estimate the value
of project-induced time savings per vehicle km on D4100.The methodology forestimating the
value of the vehicle-km for each vehicle type will be explained below.
Table 17.8: Average Speeds of Vehicles for“With” and “Without” Project Scenarios
The next step is to estimate the average occupancy of each vehicle type and the time value per
hour for its occupants. In regard to the average vehicle occupancy, a road user survey is used to
obtain a reliable estimate. The time value of passengers ismeasured by wage rates for skilled and
unskilled labor. For valuation of time saving for tourists, additional information is needed
whether a particular tourist group is from overseas or domestic, and their respective average
wage rates must be known.
speeds are estimated on a flat terrain with roughness index of 2.0 for tarred road, and on a flat terrain with the index of 10.0 for gravel road.
CHAPTER 17:
17
The information regarding vehicle occupancy and labor wage rates was obtained from thestudy
by ARCUS GIBB.8 For passenger cars, an average rate of 1.2 skilled occupants is used. For
tourist trips, it is assumed that there are, on the average, 1.5 tourists per vehicle and also on
average 0.5 tourist guide per vehicle, thus making a total of 2.0 occupants per vehicle. For mini-
buses, 10.0 unskilled commuters comprise an average travel group. Note that for HGV, LGV and
agriculture vehicles, the driver’s salary is a direct cost of transportation and has already been
accounted for as part of vehicle operating costs.
The wage rate for unskilled labor is taken as R 7.15 per hour, and the rate for skilled labor, it is
R18.25 per hour.9 For tourists, who are likely to belong to the skilled category, only one half of
their wage rate, R 9.13 per hour, is taken as value of time. For LGV, HGV and agriculture traffic,
the value of time saving is dependent on the content and value of their cargo, and the respective
value of delivery delays. Because of the wide diversity of agriculture, mining, and other goods
that could be potentially traveling on the proposed road, they should be estimated separately,
insofar as possible. The total time saving for each vehicle type can then be estimated.
For passenger cars, the value of time saving per vehicle-km is equal to the value of time per
vehicle-km on the gravel road minus the value of time per vehicle-km on the tarred road. For
example, on section D4100, the value of time per vehicle-km for a single occupant of a
passenger car with a wage rate of R 18.25 per hour and traveling at the speed of 68.8 km per
hour, is R 0.2654 per vehicle-km. With the same value of time, but traveling at a speed of 86.6
km per hour the time cost is R 0.2107 per vehicle-km. The estimated value of time saving for a
passenger car is then about R 0.0656 per vehicle-km with 1.2 occupants.In a similar fashion, the
value of time saving for passenger cars on sections D4250/D4190 and D4050/D1583 is estimated
as R 0.0656 and R 0.0437 per vehicle-km, respectively.
8 ARCUS GIBB, “Limpopo Integrated Infrastructure Development Plan: Phase II – Benefit Cost Analysis of Selected Projects”, Final Report, Appendix I: Flag Boshielo to Mafefe to Sekororo and Ga Seleka to Mmatladi, March 2004, from p. 4-2 to p. 4-3. 9 The study by ARCUS GIBB places values of R 6.71 and R 17.14 in 2004 on unskilled and skilled hourly wages, respectively. An inflation adjustment of 6.5% was applied to obtain the 2005 wage rates, resulting in R 7.15 and R 18.25.
CHAPTER 17:
18
For tourist trips on D4050/D1583 section, the estimated time saving per vehicle-km is R 0.0455.
This is derived from the time saving for the average 1.5 tourists and 0.5 skilled occupants of a
typical vehicle.10 No substantial volume of normal tourist traffic is expected on other sections of
the upgraded road.
Min-Buses Traffic: The same method is applied to measure the value of time saving for mini-
buses. For section D4100, the resulting estimate is R 0.3102 per vehicle-km.11 The value of time
savings formini-bus traffic on sections D4250/D4190 and D4050/D1583 is estimated as R
0.3102 and R 0.2133 per vehicle-km, respectively.
In the case of freight for LGV and agriculture traffic, a different approach is needed to estimate
the value of time saving. The improved road will allow LGV and agriculture transport to move at
higher speed, which means a faster turnover of the vehicle fleet and more productive use of the
vehicles. In the long run the owners of cargo will need fewer vehicles, thus resulting in savings
of capital costs. Suppose a new truck costs R 1.30 million, its average utilization factor is 70%,
and the real rates of depreciation and financial return on investment are 15.4% and 10.0% per
annum, respectively.12 On section D4100 alone, the value of time for LGV/agriculture traffic per
vehicle-hour will beR 53.95.13 If the time saving per a 34-km trip due to speed increase is 0.181
hour, the value of capital savings can be estimated at R9.760 per vehicle-trip on this section.14
In addition to capital savings, there is also saving of driver’s wages that will add up to R 1.294
per 34-km vehicle-trip.15 Thus, the combined value of time saving for heavy traffic is R 11.054
10 The value of time saving for tourists is equal to R 0.0273 per vehicle-km (= [(R 9.13 / 60.7) – (R 9.13 / 69.1)] * 1.5 occupants). For skilled occupants of a tourist vehicle, the estimated time saving amounts to R 0.0182 per vehicle-km (= [(R 18.25 / 60.7) – (R 18.25 / 69.1)] * 0.5 occupants). The summation of the value of time for both kinds of occupants gives us a figure of R 0.0455 per vehicle-km on section D4050/D1583. 11 Estimated as ([R 7.15 / 60.4] - [R 7.15 / 81.9])* 10.0 occupants = R 0.3102 per vehicle-km on section D4100. 12 The average annual cost structure of truck transportation was obtained from the Vehicle Cost Schedule, published by the Road and Freight Association (October 2001). 13 The value of time for LGV vehicle is estimated as: R 1,300,000 * (15.4% + 10.0%) / (365 * 24 * 70%) = R 53.95 per vehicle-hour. Alternatively, one can use annual rental charges for LGV vehicle divided by the number of hours the vehicle is actually transporting merchandise. 14 The amount of time saved per trip is equal to 0.181 hour per vehicle trip (= 34 km / 53.9 km/hr - 34 km / 75.5 km/hr. The value of capital savings can then be estimated as R 9.760 per vehicle trip (= 53.95 R/hour * 0.181 hour/vehicle-trip) on section D4100. 15 The value of driver’s wage savings is estimated as: R 7.15 hour * 0.181 hour/vehicle-trip = R 1.294 per 34-km vehicle-trip.
CHAPTER 17:
19
per vehicle-trip on section D4100. Using the same approach, the value of time saving is
estimated for sections D4250/D4190 and D4050/D1583 as R 15.281 for a length of 47 km and R
16.459 for 75 km per vehicle-trip, on the corresponding section.
17.8 Economic Appraisal
The economic appraisal of a project is concerned with the effect that the project has on the entire
society and inquires whether the project increases the economic welfare of society as a whole. It
looks at the present value of all the incremental annual economic benefits and costs generated
throughout the project life, including savings in vehicle operating costs, time costs of travellers,
maintenance costs, and other costs such as accidents and other externalities. The present value of
annual benefits minus costs over the project’s lifetime is then compared with the capital
expenditures incurred on upgrading the road.
The annual savings in maintenance, VOC and time costs will be quantified in the next section.
As regards the impacts of an improved road project onaccidents, itcould be important because of
changes in the number of accidents and damages in monetary value on property and human
bodies.In general, they should be properly assessed “with” and “without” project scenarios. This
component, however, may not be significant in this project due to low volume traffic and it is
therefore not included in this study.
Other externalities such as various taxes and subsidies involved in key project inputs are
captured in Commodity Specific Conversion Factors (CSCF). Three key CSCFs are identified in
this project. They are infrastructure construction and maintenance costs, truck transportation, and
passenger care transportation; and their corresponding CSCFs are estimated at 0.876, 0.850, and
0.922, respectively, based on methodology outlined in Chapter 11 and the empirical estimation
carried out elsewhere.16 These CSCFs allow us to convert all financial costs of the project inputs
into the corresponding economic costs in order to construct the economic resource flow
statement of the project.
16 Taken from Cambridge Resources International, “Integrated Investment Appraisal: Concepts and Practice”, Appendix G: Commodity-Specific Conversion Factors for Non-Tradable Goods and Services in South Africa, (2004).
CHAPTER 17:
20
17.8.1 Annual Savings in Maintenance Costs, VOC and Time Costs
This section summarizes total savings in maintenance costs, vehicle operating costs and time
costs generated by upgrading the gravel road to the tarred road.
Maintenance Costs
Annual savings in financial maintenance costs have been estimated (costs with the project minus
what they would have been in its absence)and are presented in Table 17.3 by road section. These
costs are multiplied by the conversion factor for maintenance costs at 0.876 to generate annual
savings of economic resource costs. Details for each year over the life of the project are shown in
Table 17.9 by road section and by frequency of maintenance.
A positive result means that some cost savings will be generated, while negatives imply a net
increase in resource costs. In this case, each type of maintenance activity will generate savings in
economic resource costs. The estimated present value of these savings (using the economic cost
of capital for South Africa at 11.0% as the discount rate17)due to road improvement amounts to R
137.8 million in 2005 prices. About 57.1% of the total savings stems from reduced costs of
intermediate maintenance. Savings in periodic maintenance account for 35.7% of the total, and
savings on routine maintenance account for the remaining 7.2%.
17Kuo, C.Y., Jenkins, G.P., and Mphahlele, M.B., “The Economic Opportunity Cost of Capital in South Africa”, South African Journal of Economics, Vol. 71:3, (September 2003), pp. 525-543.
CHAPTER 17:
21
Table 17.9: Savings in Economic Maintenance Costs (Millions of Rand in 2005 Prices)
The results of the risk analysis were simulated for 10,000 runs. The simulation results shown in
Figure 17.1 indicatesthat the expected value of the project’s economic NPV is R 135.7 million,
which is higher than the deterministic value of R 105.2million. Figure 17.1presents therange of
possible project outcomes that the economic NPV can take and the likelihood of the occurrence
of these values. It ranges from the minimum gain of R 0.7 million to the maximum gain of R
359.5 million. There is zero probability that the economic NPV of the project may become
negativeunder all possible circumstances defined earlier in the risk analysis.
Figure 17.1: Results of Risk Analysis on Economic NPV
CHAPTER 17:
38
The expected resource costof this project to the Road Agency is about R 161.0 million. This is
created by the heavy initial capital cost that is partially offset by savings in maintenance costs
arising from the improvement of various road sections.Figure 17.2 shows that under any
circumstances, the Agency will not have a positive PV of net benefits from this project, because
savings in road maintenance costs are unlikely to overweigh the initial investment outlays.
Figure 17.2: Results of Risk Analysis on Road Agency NPV
Each of the five classes of road users stands to gain from the improvement of this road. The
expected values of their respective benefits accruing are all greater than the values in the
deterministic case. At the same time, their respective probability of getting negative benefits is
virtually non-existent.This simply means that all road users will surely gain as a result of this
project. This is illustrated below for private car passengers and mini buses in Figure 17.3 and
17.4, respectively.
The risk analysis indicates that the expected benefits received by the private car passengers is R
59.5 million, which is larger than the deterministic case of R 48.8 million. Theirnet benefit is
subject to significant variability, however, with a standard deviation of R 11.4 million. The gain
ranges from the minimum R 37.4 million to the maximum R 106.0 million. There is zero
probability of a negative NPV for this group.
CHAPTER 17:
39
Figure 17.3: Results of Risk Analysis for Car Passengers
By the same token, the mini bus usersare expectedto gain a mean value of R 215.5 million. This
resultsmainly from the VOC savings and time savings of the existing (“without project”) road
users. This gain is subject to a standard deviation at R 41.4 million. Nevertheless, the probability
of getting a negative benefit for this group is nil, as shown in Figure 17.4.
Figure 17.4: Results of Risk Analysis forMini Buses
CHAPTER 17:
40
It should be noted that the net benefit of upgrading Section D4050/D1583 is negative, with a
mean value of R 39.2million, under all possible circumstances defined within the risk analysis
presented earlier. There is zero probability a positive benefit resulting from improving this road
section as of 2005. On the other hand, the other two sections, D4100 and D4250/D4190, have
positive net benefits with the means of R57.1 million and R117.8 million, respectively, and zero
probability of getting a negative benefit.
17.11 Concluding Remarks
Thischapter has followed the integrated investment appraisal methodology to evaluate the
upgrading from gravel to tarred surface of several road segments in the Limpopo Province of
South Africa. The project is to be carried out by the Road Agency of Limpopo, and no tolls are
to be charged. The purpose of this study is to inquire whether the resources investedin each
segment promise to be outweighed by its economic benefits.
Typically, benefits generated by road improvement include the reduction (perhaps increase) in
resource costs on maintenance by the Road Agency, reduction in vehicle operating costs forroad
users due to improved road surface, time saving for road users due to an increase in average
speed of vehicles, and possible reduction in the costs of accidents. This chapter has illustrated
how to evaluate all but the last of these components.
For the Road Agency, this project suggests that savings in maintenance costs for R 137.8 million
are not sufficient to cover the initial capital expenditures of R 277.2 million required by the
improvement. However, the improvement of the road would generate a substantial benefit in
terms of the savings of time and operating costs accruing to different classes of road users. As a
result, the net economic benefit for the project is R 135.7 million.
The stakeholder analysis is also carried out for this project. The main beneficiaries of this
investment will be the owners and users of mini-buses for R 215.5 million and of passenger cars
for R 59.5 Million. Other road users, tourists, owners of lightand heavy freight vehicles will also
CHAPTER 17:
41
share the benefits of the improved road by R 2.4 million, R 5.3 million, and R 14.0 million,
respectively.
It is important to point out that the proposed road consists of three sections for which an
individual economic assessment was carried out in this study. It appears that while the overall
economic NPV of the project is expected to be approximatelyR 135.7 million, section
D4050/D1583 has an expected negative NPV of R 39.2million when evaluated on its own. This
section should therefore be excluded from the investment package, an act which would raise
overall net benefits of the project to R 174.9 million.
CHAPTER 17:
42
Appendix 17A:Estimation of Vehicle Operating Costs
The vehicle operating cost (VOC) estimates are based on the road economic decision model
(RED) developed by the World Bank and modified for South African realities by CSIR
Transportek in 2003.18. The costs are essentially dependent upon terrain, roughness of the road,
and type of vehicle. For each category, VOC is expressed as a function of roughness in the form
of cubic polynomials, differing by category of road and type of vehicle. The general form is:
VOC = a0 + a1* R + a2* R2 + a3* R3
Where R stands for the degree of road roughness, which is standardized and expressed in terms
of the international roughness index. The results are shown in Table 17A.1. Costs are obviously
expected to incur higher for traveling on a mountainous road than on a level road and on a rough
road than on a smooth one.
18 Archondo-Callao, R., “Roads Economic Decision Model (RED) for Economic Evaluation of Low Volume
Roads: Software Users Guide”, Version 2.0, 3/15/01, The World Bank, Washington, D.C., USA. The model is customized for South African conditions by CSIR Transportek at the request of the Development Bank of Southern Africa (DBSA). Available at <http:www.dbsa.org>. The vehicle operating costs could be also computed using the functions calibrated by AFRICON Pietersburg for both light and heavy vehicles. “The Long Term Consequences of Various Budget Levels and Flood Damage Assessment on the Northern Province Road Network”, Vol. 1, Prepared by AFRICON Pietersburg, Prepared for Roads Agency Limpopo, October 2000.
CHAPTER 17:
43
Table 17A.1: Estimates of Average VOC by Terrain, Type of Vehicle, and Roughness (Rand per Vehicle km in 2003 Prices)
Coefficients of the VOC Function VOC (Rand per Vehicle km) Terrain Vehicle
Type a0 a1 a2 a3 R at 2.0 R at 10.0 R at 20.0 Car 2.198 -0.028 0.017 -0.00031 2.204 3.261 5.790 Utility 2.197 -0.016 0.017 -0.00031 2.229 3.388 6.048 Light Bus 2.600 0.027 0.012 0.00010 2.702 4.164 8.722 Medium Bus 3.047 0.027 0.012 0.00011 3.148 4.611 9.227 Heavy Bus 3.788 0.026 0.013 0.00010 3.894 5.463 10.360 Light Truck 2.954 0.162 0.012 -0.00022 3.324 5.554 9.218 Medium Truck 3.397 0.181 0.013 -0.00023 3.809 6.251 10.259
Heavy Truck 4.840 0.223 0.012 -0.00021 5.334 8.083 12.489
Flat &
Paved
Artic. Truck 6.856 0.287 0.023 -0.00047 7.518 11.552 18.021 Car 2.495 -0.028 0.015 -0.00027 2.497 3.458 5.843 Utility 2.533 -0.015 0.015 -0.00027 2.562 3.631 6.145 Light Bus 2.943 0.025 0.011 0.00014 3.036 4.391 8.811 Medium Bus 3.583 0.038 0.009 0.00018 3.698 5.065 9.462 Heavy Bus 4.534 0.045 0.010 0.00017 4.665 6.154 10.799 Light Truck 3.669 0.161 0.010 -0.00016 4.028 6.092 9.512 Medium Truck 4.260 0.200 0.009 -0.00015 4.696 7.042 10.774
Heavy Truck 6.672 0.244 0.009 -0.00014 7.195 9.835 13.861
Since the original model’s output was expressed in prices of 2003 level, the final results were
inflated into the prices for year 2005 by using an annual inflation rate of 6.5% for each of the
years 2003 and 2004. Table 17A.2 presents the resulting VOC estimates in 2005 prices by type
of vehicle, terrain and road roughness.
Table 17A.2: Estimates of Average VOC by Terrain, Type of Vehicle, and Roughness
(Rand per Vehicle km in 2005 Prices)
CHAPTER 17:
44
VOC (Rand per Vehicle km) Terrain Vehicle Type
R at 2.0 R at10.0 R at 20.0 Car 2.500 3.699 6.567 Utility 2.528 3.843 6.860 Light Bus 3.065 4.723 9.892 Medium Bus 3.571 5.230 10.465 Heavy Bus 4.416 6.196 11.751 Light Truck 3.770 6.299 10.455 Medium Truck 4.320 7.090 11.636 Heavy Truck 6.050 9.168 14.166
Flat &
Paved
Artic. Truck 8.527 13.103 20.440 Car 2.720 3.837 6.598 Utility 2.735 3.983 6.906 Light Bus 3.267 4.837 9.913 Medium Bus 3.786 5.366 10.496 Heavy Bus 4.624 6.336 11.786 Light Truck 3.997 6.425 10.487 Medium Truck 4.505 7.210 11.670 Heavy Truck 6.211 9.274 14.200
Mountainous
& Paved
Artic. Truck 9.459 13.420 20.459
CHAPTER 17:
45
Appendix 17B:Estimation of Average Vehicle Speeds
Estimates of the speeds are also based on the Roads Economic Decision Model (RED) developed
by the World Bank and modified for South Africa. The average vehicle speed is a function of
several factors, and the models were developed to estimate speeds for different types of vehicles
under various terrain and road conditions.
For each category, the average vehicle speed (S), expressed in the number of kilometers per
hour, is calculated below as a function of the degree of road roughness:
S = b0 + b1* R + b2* R2 + b3* R3
The coefficients of the equations were estimated for each type of vehicle by terrain and the
international road roughness and are shown in Table 17B.1. One would expect the vehicle speed
is faster when road is smoother for the same type of vehicle. Similarly, given the same degree of
road roughness, on average vehicles move faster in flat road as compared to the case in
mountainous road.
Table 17B.1: Estimates of Average Vehicle Speeds by Terrain, Type of Vehicle, and Roughness
(Km per Hour)
Coefficients of the Speed Function Speed (km per hour) Terrain Vehicle Type b0 b1 b2 b3 R at 2.0 R at 10.0 R at 20.0
Car 87.310930 0.169627 -0.275873 0.007344 86.61 68.76 39.11 Utility 80.447861 -0.396431 -0.205073 0.005667 78.88 61.64 35.82 Light Bus 84.601774 -0.941928 -0.210988 0.006361 81.92 60.44 32.25 Medium Bus 74.745730 -0.124356 -0.220741 0.006043 73.66 57.47 32.31 Heavy Bus 74.578226 -0.158145 -0.216589 0.005939 73.44 57.28 32.29 Light Truck 79.405651 -1.691452 -0.131311 0.004518 75.53 53.88 29.20 Medium Truck 74.316529 -1.822427 -0.104362 0.003841 70.28 49.50 26.85 Heavy Truck 61.903260 -0.930893 -0.108541 0.003330 59.63 45.07 26.51
Flat &
Paved
Artic. Truck 85.917625 -5.437581 0.099941 0.000352 75.45 41.89 19.96 Car 67.707011 1.136416 -0.238360 0.005474 69.07 60.71 38.88 Utility 60.813997 0.637159 -0.178888 0.004143 61.41 53.44 35.15 Light Bus 63.274717 0.715527 -0.222646 0.005474 63.86 53.64 32.32 Medium Bus 53.016936 0.683988 -0.160811 0.003647 53.77 47.42 31.55 Heavy Bus 51.129871 0.615620 -0.145875 0.003264 51.80 45.96 31.21 Light Truck 53.549798 0.218623 -0.144733 0.003569 53.44 44.83 28.58 Medium Truck 49.905646 -0.038640 -0.117567 0.002995 49.38 40.76 26.07
Heavy Truck 36.174877 0.023484 -0.062253 0.001431 35.98 31.62 23.19