80 Transportation Research Record 898

Vehicle Operating Costs Under Indian Road and

Traffic Coriditions
C.G. SWAMINATHAN AND LR. KADIYALI

A research project on road-user costs sponsored jointly by the Government of either widened initially to an intermediate width of
India and the World Bank has recently been completed in India. As part of 5.5 m or widened to the standard 7.0 m.
this project, an attempt was made to build relationships between vehicle operat· The number of motor vehicles in India is 3.7 mil-
in1tcost components and road, traffic, vehicle, and environmental factors. India lion, which register an annual growth of nearly 10
presents a wide variety of these conditions. Roads in India are improved in percent (2). The country produces about 60 000
stages in view of the paucity of funds. A major portion of the roads are single·
single-unit trucks annually that have a carrying
lane, bidirectional, and unpaved. Traffic is heterogeneous in character; it con-
capacity of 7.5 tonnes. The passenger cars are of
sists of fast-moving as well as animal-drawn vehicles. Climate and topography
change across the country. The work was carried out by collecting real-life
small capacity and are also produced in India;
data on cost of operation of about 1000 vehicles of different types. The re- annual production is about 40 000 vehicles.
sults prove that horizontal curvature, vertical profile, pavement roughness, and The traffic on Indian roads is heterogeneous in
pavement width are some of the important factors influencing vehicle operating character; it consists of a mixture of motorized
costs. These results are likely to be of great value in evaluating benefits that are vehicles of all types, sizes, and shapes (cars,
possible from highway improvements and in arriving at sound investment deci- jeeps, vans, three-wheeled light vehicles, buses,
sions. Since the conditions in India are typical of developing countries, the re- minibuses , trucks, agricultural tractors, semi-
sults can be of value to other developing countries. truck-trailers, truck-trailers, scooters, motor-
cycles, mopeds); cycles and cycle rickshaws; and
animal-drawn vehicles (bullock carts, camel carts,
As part of the development activity in Tndi.a, vast and horse carts) • There is no segregation of vehi-
investments are being m.:idc for the improvement of cles on the roads, with the result that traffic
roads. Since the resources are scarce and the de- operations are hazardous and slow. Figures 1-6
mands from the vario\]s sectors are heavy, economic illustrate the different types of vehicles on the
appraisal of schemes provides a good basis for in- Indian roads and the traffic conditions.
vestment decisions. A serious handicap in this di- The accident r ate in India is alarming. In 1977-
rection has been the lack of an adequate data base 1978, 21 300 persons were killed (2). The number
for road-user costs under typical Indian condi- killed per 10 000 vehicles is 64, which is very much
tions. In order to fill the gap in the process of higher than the rate of 3-7 in developed countries.
highway planning in India, the Central Road Research
Institute, New Delhi, has recently completed a re- ROAD USER COST STUDY IN INDIA
search project entitled Road User Cost Study, spon-
sored jointly by the Government of India and the Realizing the growing importance of roads and road
World Bank. As part of the project, research was transport in India and the serious data gap for ef-
undertaken on the relationships between the vehicle fective planning, the Government of India and the
operating-cost components and road characteristics, World Bank jointly sponsored an ambitious research
traffic factors, and vehicle characteristics. The project entitled Road User Cost Study. This proj-
research was carried out hy collecting real-life ect, which cost nearly $1 million (U.S.), was
data on the cost of operation for nearly 1000 vehi- started in 1978 and was completed in 1982 Cl-~).
cles of different types that operate in the dif- The Central Road Research Institute was the imple-
ferent parts of the country. The data were menting agency.
collected over a period of about 18 months. Simul- The total transportation cost is composed of
taneously, measurement of the road character is tics
on nearly 42 000 km of roads of the various types 1. Cost of initial construction of the facility,
waG completed with the aid of modern instruments. 2. Cost of periodic maintenance, and
From the data, relationships have been formulated 3. Road-user cost.
between each of the vehicle operating-cost com-
ponents and the roadway, traffic, and vehicle fac- ~he research project recently completed deals
tors that govern them. with the third component listed above. The Govern-
ment of India has plans to do further research on
ROAD AND TRAFFIC CONDITIONS IN INDIA the first two components listed above, so that a
highway design model can be developed with capabili-
Since India is predominantly an agrarian and rural- ties for evaluating a number of alternative schemes
oriented economy, road transport forms a vital sec- that have a range of road-design variables, mainte-
tor in planned development. The total length of nance strategies, and transport options.
roads in the country is l. 6 million km, of which The road-user cost itself has three important
only 40 percent are paved !ll· The major percentage component s :
of roads leading to the villages and small towns are
unpaved earth or gravel roads. The count r y has a l. Vehicle operating costs (VOC),
large percentage of single-lane bidirectional roads 2. Accident costs, and
(pavement 3-3.75 m wide). Even on the National 3. Highway travel-time costs.
Highways, which are the arteries of the nation, the
proportion of single-lane pavements is 35 percent We limit this paper to the consideration of voe in
(_!). Single-lane roads, which permit only one vehi- relation to roadway, traffic, vehicle, and environ-
cle to use the pavement at a time, result in slowing mental factors.
down of the traffic, hazardous maneu11ers, and voe is composed of the following elements: fuel,
greater wear and tear on vehicles. A system of lubricants, tires, spare parts, labor for mainte-
staged constructio n is gene r al l y fo l lowed i n the nance and repairs, depreciation, and fixed costs.
country, whereby selected single-lane roads are
Transportation Research Record 898 81

Figure 1. Single-lane road showing crossing between bus and bullock cart. Figure 4. Convoy of bullock carts on single-lane road obstructing motor vehicle.

Figure 2. Bullock cart on earthen road in rainy season.

Figure 5. Mixed traffic conditions on typical road passing through a small town.

Figure 6. Three-wheeled auto-rickshaw commonly found on Indian roads.

Figure 3. Agricultural tractor-trailer on unpaved road.
82 Transportation Research Record 898

These elements are influenced by a variety of fac- Vehicle ~

tors, the more important of which are listed below: Bus 640
Truck 232
1. Roadway factors Car 53
a. Pavement width Jeep 14
b. Surface type and roughness Total 939
c. Horizontal curvature
d, Vertical profile Table 1 gives the salient teatures of some or the
2. Vehicle factors typical makes of vehicles included in the survey.
a. Make Adequate geographic representation across the
b. Age country was ensured to capture the effect of varying
c. Horsepower and engine capacity topographical and roadway design features. Tables
d. Load carried 2-4 give the representation of the vehicles across
3. Traffic factors the factorial matrix. It will be seen that although
a. Speed some of the cells are unrepresented or underrepre-
b. Volume of traffic sented, the extremes of road conditions are covered
4. Environmental factors by the survey. When the vehicles were selected, it
a. Altitude was ensured that only such vehicles were included
b. Rainfall and climatic conditions that had a reasonable expectation of operating on
homogeneous routes during the survey period. The
The objective of the research on voe was to in- data collection on the voe components from the oper-
terrelate each of the voe components with the above- ators was done for a period ranging from 12 to 24
mentioned factors. In other words, the research months. This period was considered necessary since
project aimed at establishing an accurate statisti- the replacement of tires and major spare parts takes
cal relationship between each of the voe elements place over a long period of time. Carefully thought
and the above-mentioned factors, so that voe can be out advanced plans were devised, and the field in-
predicted for any given combination of roadway, ve- vestigators were specially trained to collect accu-
hicle, traffic, and environmental conditions in rate and reliable data, after they had gone through
India. the records maintained by the operators and had in-
tensive discussions with them on doubtful items.
METHODOLOGY FOR RESEARCH ON voe The information was collected on the various com-
ponents of voe such as fuel, lubricants, tires,
For determining fuel-consumption relationships, two spare parts, labor for maintenance and repairs, and
different approaches were used. The first and the fixed costs.
more accurate was to perform controlled experiments Whereas the spare-parts consumption was measured
on the common makes of vehicles specially procured in monetary terms, the rest of the consumption items
for the purpose. The second approach was to collect were measured in physical units. This renders it
data on fuel consumption of 939 vehicles of dif- possible to use the relationships in other countries
ferent types from a number of operators. The two with similar operating characteristics. In the cas_e
dif~erent approaches were then compared. of spare parts, inflation was accounted for by
For other vehicle operating-cost components such bringing all costs to a common reference year of
as tire wear, spare-parts consumption, maintenance 1978 with the use of price indices.
labor, depreciation, fixed costs, and lubricants, The number of tires replaced during the survey
the controlled-experiment approach becomes unmanage- period included new, retreaded, and used tires.
able and impractical. The more' expedient procedure They were brought to a common measure of equivalent
is to undertake a detailed survey of the cost of tires by applying suitable conversion factors deter-
operation in real life from a large number of opera- mined from a detailed survey of selected operators.
tors, and this approach, known as the user cost sur- It will also be necessary to determine the mone-
vey (UCS) has been tried successfully in Kenya (~) tary cost of tire replacement, which consists of
and in Brazil (1). new, retreaded, and used tires, all of which cost
differPnt amounts. A conversion factor of 0.727 has
USER COST SURVEY been determined from the data, which can be used
thus:
since the relationships are likely to be complex
between the voe components and the factors governing Cost of equivalent new tires= 0.727 x number of
them, the survey had to be carefully structured and equivalent tires as per UCS relationship x cost
adequately planned. The factorial matrix adopted of new tire.
for the survey consisted of the combinations of
factors given below: The data for the entire survey period were aver-
aged to yield the following dependent variables:
No. of
Factor Levels Description of Level 1. Fuel consumption (FC) in liters per 100 km;
Terrain 3 Level, rolling, or hilly 2. Engine oil consumption (EOL) in liters per
Horizontal 2 Low or high 1000 km;
curvature 3. Other oil consumption (OL) in liters per
Pavement width 4 Single-lane unpaved, 10 000 km;
and type single-lane paved, two- 4. Grease consumption (G) in kilograms per
lane paved, or dual- 10 000 km;
carriageway paved 5. Spare-parts consumption (SP) in paise per
kilometer ($1 U.S. = 9,00 rupees; 1 rupee • 100
Three types of vehicles were included: cars and paise);
jeeps, buses, and trucks. 6. Tire life (TL) in kilometers for equi1ralent
After an intensive search, 939 vehicles belonging new tire;
to 121 operators were selected for the survey. 7. Labor hours (LH) in hours per 1000 km;
These were grouped in the following way: 8. Labor cost (LC) in paise per kilometer;
Transpor tation Research Record 898 83

9. Speed of vehicles v8 , VT, Ve, VJ for vertical profile, roughness, and pavement width.
buses, trucks, cars, and jeeps, respectively , in The vehicles included in the survey were associ-
kilometers per houri and ated with nearly 42 000 km of routes spread over
10. Use of vehicles (OPDJ in kil~meters per day. different parts of the country and having a variety
of characteristics. The measurement of the charac-
MEASUREMENT OF ROUTE CHARACTERISTICS teristics by conventional surveying methods would
have been time consuming and was ruled out. A sys-
since voe data have to be related ultimately to the tem of route-characteristic measurement from an in-
route characteristics, it is necessary to evolve a strumented moving car was selected (8).
quick and reasonably accurate method of measuring The instrumentation consisted -of a distance-
the important route characteristics. A variety of measur ing device, a gyroscope for measuring hori-
roadway factors influence voe, but the following im- zo ntal curvature, a pendulum-type gradometer, and a
portant ones were considered: horizontal curvature, car-mounted roughness integrator.

Table 1. Salient features of some typical UCS

vehicles. Gross
Engine Unladen Laden
Vehicle British Capacity Weight Weight No. of
Category Model Horsepower (cc) (tonnes) (tonnes) Axles

Car Ambassador 50 1489 1.2 1.5 2

Jeep Mahindra 75 2350 0.6 1.2 2
Bus Tata 1210 112 4788 7.0 11.0 2
Bus Ashok Leyland Comet 110 6075 7.0 11.0 2
Truck Tata 1210 112 4788 6.0 13. 5 2
Truck Ashok Leyland Comet 110 6075 6.0 13 .5 2
Truck Tata 1516 145 5675 9.0 24 .0 3
Truck Ashok Leyland Taurus 125 6735 7.0 18.5 2

Table 2. Representation of buses across factorial matrix.

Plain Terrain" Rolling Terrain 3 Hilly Terrain'

Curvature (degrees/km)

Pavement Width and Type 0-50 >50 0-100 >100 0-200 >200 Total

Single lane unpaved (up to 5.0 m wide, roughness >8000 mm/km) 11 17 10 25 9 72

Single lane paved (up to 5 .0 m wide, roughness < 8000 mm/km) 36 67 9 115 21 248
Intermediate lane paved (up to 5.0-6.0 m wide, all roughnesses) 48 35 7 70 160
Double lane paved (>6.0 m wide, all roughnesses) 90 42 4 23 I 160
Total TBS 16T 30 233 3T 640

aTerrain defined es follows: plain= rise+ fall= 0-15 m/km; rolling= rise+ rail= 15·30 m/km; hilly = dse +fall..::: >30 m/km.

Table 3. Representation of trucks across factorial matrix.

Plain Terrain• Rolling Terrain" Hilly Terrain"

Curvature (degrees/km)

Pavement Width and Type 0-50 >50 0-100 >100 0-200 >200 Total

Single lane unpaved (up to 5.0 m wide, roughness >8000 mm/km) 6 10 14 30

Single lane paved (up to 5.0 m wide, roughness <8000 mm/km) 5 II 1 3 20
Intermediate lane paved (up to 5.0-6.0 m wide, all roughnesses) 5 6 3 10 24
Double lane paved (>6.0 m wide, all roughnesses) 93 28 9 28 158
Total 109 55 IT 39 IT ill
8Terrain deflned as follows: plaln =rise + fall= 0-J 5 m/km; rolling= rise+ fall= 1 5-30 m/km; hilly= rise+ foll= >JO m/km.

Table 4. Representation of cars and jeeps across factorial matrix.

Plain Terraina Rolling Terrain 3 Hilly Terrain"

Curvature (degrees/km)

Pavement Width and Type 0-50 >50 0-100 >100 0-200 > 200 Total

Single lane unpaved (up to 5.0 m wide, roughness > 8000 mm/km) 2 2
Single lane paved (up to 5.0 m wide , roughness <8000 mm/km) 11 1 2 14
Intermediate lane paved (up to 5.0-6.0 m wide, all roughnesses) 4 2 I 7
Double lane paved (>6.0 m wide , all roughnesses) 34 4 5 I 44
Total 34 19 8 6 67
8
Terrain defined as follows: plain =rise+ faH = 0-15 m/kmi rolling= rise + fall ~ 15-30 m/km; hilJy =rise +fall = >JO m/km.
84 Transportation Research Record 898

The distance-measuring device was developed by ADDITIONAL DEPENDENT VARIABLES

introducing a miniature Halda gear in the vehicle
i;;peedometer cable; thisl redt1('ea the 'lp1>1>t'l nf rnt;1- In addition to road-~eometry variables and rouqh-
tion in the ratio 20:1 in a subsidiary cable at ness, some additional variables were also used in
right angles to the speedometer cable. The revolu- the analysis: gross vehicle weight (GVW) in tonnes,
tions of the subsidiary cable are then converted power/weight ratio (PW) in kilowatts per tonne, and
into elastic pulses that activate an electromechani- number of tires on vehicle (N) •
cal counter. Some vehicles are operated on long-distance hauls
The gyroscope is a flight instrument used for and inherently involve long hours of use per day.
navigation of aircraft. One such instrument, oper- Others are very poorly used and deployed only for
ated on 110 V, was mounted in the instrumented car, local trips. In order to isolate these features,
and the car battery (12 V) supplied the needed power the following three 0-1 dummy variables were intro-
through a static inverter. The instrument records duced: ia) medium use (MU), 4-12 h/dayi ib) low use
the direction of movement at any time. By noting (LU), less than 4 h/day; and (c) high use (RU), more
the reading at the beginning and enu of a horizontal than 12 h/day.
curve and the corresponding distance measurements,
one can calculate the deviation angle, length of ANALYSIS OF voe DATA
curve, and radius.
The pendulum-type gradometer, designed and sup- Correlation Between Independent Variables
plied by the U.K. Transport and Road Research Labo-
ratory (TRRL), is a simple heavy aluminum alloy It was seen that the two important geometry vari-
pendulum that has a pointer needle at right angles. ables, curvature (CV) and rise and fall (RF), were
The needle makes contact at different points on a highly correlated. The use of both these variables
semicircular arc i the points correspond to the up- in the analysis was therefore likely to cause
ward or downward gradients. The readings so ob- trouble. Hence the analysis was done by using one
tained are recorded by an electromagnetic counter. of these variables at a time. The following rela-
Accurate measurement of roughness of the surface tionship was derived between the two variables:
was recognized to be the key to the derivation of
good voe relationships. A number of devices are CV = 1.4011 RF!n (1)
available for this purpose (9-ll). The subject of
roughness measurements, the n~ed for calibration of where R2 = 0.8436 (R is the correlation coeffi-
the instruments, and the equally pressing need for a cient). Pavement width and roughness were also
universally acceptable measuring scale are engaging similarly correlated. The two variables were tried
the attention of researchers all over the world. together and have been retained only if a satisfac-
The British towed fifth-wheel bump integrator is a tory t-value was obtained for the regression coeffi-
standard instrument for roughness measurements and cients. Otherwise, a transformation of the vari-
has been used successfully in the Kenya <il and the ables to the form RF/W was also successful.
Caribbean (12) studies by TRRL. Since sufficient
experience existed in the country with the use of Ana lysis
this instrument, it was selected as the basic
roughness-measuring device. A large roughness- For a satisfactory explanation of the effect of a
measurement program that uses the towed fifth wheel number of variables on a dependent variable, the
is slow and unmanageable and is fraught with the multiple linear regression technique is well known.
danger that the device frequently runs out of cali- A large number of equations were tried, and those
bration. Hence, a simpler device consisting of an that are presented here are the best.
axle-mounted integrator unit was used for the large- A feature of the Indian study was that a variety
scale measurements, and the towed fifth wheel was of vehicle operators--government, ?rivate, large
reserved only for setting the standards for calibra- fleet owners, medium fleet owners, and small fleet
tion. Calibration of the axle-mounted roughness owners--were covered by the ucs. It was thought
unit was done at frequent intervals on actual road likely that each category of operator had its own
surf aces of varying roughness values against the maintenance policies, which were reflected in voe.
standard towed fifth-wheel integrator at the same In order to isolate such effects, 0-1 dummy vari-
locations. By this means, conversion equations were ables were introduced as follows:
developed that enabled the roughness measurements
with the car to be converted into the standard Variable Symbol
towed-fifth-wheel values (13). These standarn val- Government operator G
ues are used in the analysiS: Private operator p
The UCS routes were conveniently split up into a Government undertaking with high intensity GH
system of nodes and links, which facilitated the of maintenance
transfer of data into computers for analysis. From Government undertaking with average inten- GA
the average route characteristics of each link, the sity of maintenance
average characteristics of routes on which the UCS Private undertaking with high intensity of PH
vehicle operated during the survey period were maintenance
worked out. These were later used as the dependent Private undertaking with average intensity PA
variables in the statistical analysis as shown below: of maintenance
Variable Unit of Measurement Symbol
In the equations given in the subsequent para-
Horizontal Average accumulated deviation CV
graphs, t-values are indicated in parentheses, *
curvature angle of the curves in de-
denotes significant at 5 percent level, and ** de-
grees per kilometer
notes significant at l percent level. The
Vertical Average rise and fall in RF
R2 -value is the square of the multiple regression
profile meters per kilometer
RG coefficient.
Roughness British towed-fifth-wheel
values in millimeters per
kilometer Fuel-Consumption Rela t i on s h i ps
Pavement Average pavement width in w The ucs data yielded directly the relationships
width meters given below:
Transportation Research Record 898 85

Car: EOL = 2.4816 + 0.0601RFu + 0.0004(RG*/W) R2 = 0.241 (7)

(4.77) (2.00)
FC = - 13.81 + 0.036 69RF + 0.000 342 3RG + (621.3*/V)
(-1.54) (1.17) (1.66) (2.60) Bus:
+ 0.005 005V2 .. R2 = 0.111 (2)
(3.06) OL =intercept + 0.004 066LK** + 0.002 889RF* + 0.000 821 7/RG**
(6.25) (2.13) (5 .32)
Jeep: - 0.3295W R2 = 0.400 (8)
(-1.38)
FC = -10.91+0.088 23RF + 0.000 591 8RG + (432.3•*/V)
(-2.20) (0.75) (0.57) (3 .46) Value of intercept varies from operator to operator;
+ 0.005 312v2 •• R2 = 0.919 (3) a typical value is 2.79.
(6.34)
Truck:
Bus:
OL =intercept+ 0.004 219LK* + 0.000 264 6RG
FC =intercept+ 0.001023LK**+0.091 77RF** + (469.7**/V) (2.56) (0.79)
(2.57) (5.18) (9 .21)
+ 0.2632GVW** R2 ~ 0.761 (9)
+ 0.006 243V2 •• + 0.5563GVW** R2 = 0.709 (4) (3.14)
(10.23) (6.52)
Value of intercept varies from operator to operator;
Value of intercept varies from operator to operator; a typical value is -0.5313.
for a typical operator it has a value of -5.85.
Bus:
Truck:
G =intercept+ 0.001481LK**+0.033 76RF**
FC = 7.170- 0.9202PW** + 0.1432RF0 - 0.3889W + (567.1 **/V) (4.27) (2 .38)
(1.68) (-5 .65) (4.68) (-1.82) (6 .24) -0.3634W** R2 = 0.489 (10)
+ 0.007 868V2 •• R2 = 0.526 (5) (-4.00)
(9.67)
Value of intercept varies from operator to operator;
A comparison of the fuel-consumption values pre- a typical value is 4.80,
dicted from Equations 2-5 with the actual consump-
tion on UCS vehicles was made. It was observed that Truck:
the actual fuel consumption was related to the ex-
perimental values as follows: G = -0.5784 + 0.003 85CV** + 0.000 546RG**
(-0.43) (2.78) (3 .77)
1. Jeeps: actual value 1.53 x experimental + O.l 790GVW* R2 = 0.262 (11 )
value, (2.48)
2. Cars: actual value 1.64 x experimental
value,
3. Buses: actual value 1.28 x experimental
value, and The equations for tire life are given below:
4. Trucks: actual value 1. 31 x experimental
value. Car:

The higher fuel consumption in actual practice is TL= 47 340- 101.8RF- 18.39(RG**/W) R2 = 0.558 (12)
due to the frequent deceleration and acceleration, (-0.88) (-7.08)
stop-and-go motion, travel at speeds other than
steady state, idle consumption at forced stops, Jeep:
etc., in short, the driver behavior. However, the
results demonstrate that a good deal of savings in TL= 48 270 - 3.8680RG" * R2 = 0.536 (13)
fuel consumption can be brought about by improving (-3.56)
the road and bettering the traffic-flow conditions.
Bus:
Lubricants
TL= intercept - 3.900LK** - 361RF** - l.227RG** + 91 l.3W*
The consumption of lubricants (engine oil, other (-3.36) (-7.48) (-12.30) (2.13)
oils, and grease) is a small portion of voe and is -1851VCR R2 =0.621 (14)
not easily amenable to accurate analysis in terms of (-1.34)
road and vehicle characteristics. All the same,
some interesting results for buses and trucks pre- Value of intercept varies from operator to operator;
sented below demonstrate how the road and vehicle a typical value is 35 900.
characteristics affect these items of cost.
Truck:
Bus:
TL= 27 471GA** + 33 682GH + 31 680PA** + 38 053PH*"
EOL =intercept+ 0.001777LK**+0.012 71RF + 0.000 671 3(RG**/W)
(6.87) (1.25) (5 .38) - 275.8RF** - l .238RG** + 2355W** - l 169(GVW/N)
2 (-4.42) (-2 .61) (3.04) (-0.99)
+ 0.1614(0HF)* R = 0.484 (6)
(2.01) - 3962VCR R2 = 0.680 (15)
(-1.84)
Value of intercept varies from operator to operator;
a typical value is 1.52. The equations, which have satisfactory R'-
values, indicate that tire life generally
Truck:
86 Transportation Research Record 898

1.Increases as road width increases, Car:

2.Decreases as roughness increases,
~. Decreases as rise and fall increases, LC ~ 0.~498SP R2 -0.647 (20)
4. Decreases as the load per tire increases,
5. Decreases as congestion builds up on the Jeep:
road, and
6. Is higher for private operators than for LC = 0.3677SP R 2 = 0.774 (21)
government operators.
Bus:
The relationships indicate how improving the
roads can bring about economy in tire consumption. LC= 0.4027SP R2 = 0.513 (22)

Spare Parts Truck:

The equations for spare-parts consumption are given LC = 0.3692SP Ri =0.837 (23)
below:
The above equations have good R 2 -values and can
Car: be reliably used, since the labor costs are directly
related to the spare-parts cost.
SP= intercept+ 0.0041RG R2 = 0.453 (16)
(1.71) Depreciation of Vehicles

Value of intercept varies from operator to operatori One of the important components of voe is deprecia-
a typical value is -11.10. tion. Unlike other components of voe, depreciation
is an item difficult to determine, especially in
Jeep: relation to roadway characteristics. Depreciation
and use of vehicles are interrelated. In order to
Log.SP= 1.069 + 0.5365 Log.LK** + 0.000 005 271 RG R2 = 0.725 (17) simplify the approach, the investigation was carried
(5.11) (0.08) out in two stages (14):

Bus: 1. Determination of use rates of vehicles in

kilometers per year as a function of road geometry
Log.SP= intercept+ 0.2900 Log.LK** + 0.000 466 6CV** and vehicle speeds and
(11.37) (2 .70) 2. Determination of depreciation rate per year
+ 0.000 073 lORG** + 0.1 lOOVCR + (1.624*/W) from a survey of prices of second-hand vehicles.
(3.65) (I.IS) (2.03)
+ 0.26070HF** R2 = 0.506 (18) By dividing the depreciation in a given year by
(9.91) the use in kilometers during the year, one arrives
at the depreciation cost component per kilometer.
Value of .inteccept varies from operator to operator r The data for determining the annual rate of de-
a typical value is -0.1947. preciation were collected from a large number of
second-hand vehicle dealers, financiers, brokers who
Truck: advance money for purchase of used vehicles, and
some selected vehicle operators in the ucs who sold
Log.SP= - 0.8979PA** - l.332SPH** + 0.0402GA** -0.5587GH** or scrapped their vehicle.
+ 0.2332 Log.LK** + 0.000 141 3RG** + (3.4830*/W) From the data on the initial purchase price, re-
(5.46) (5.35) (2.30) sale price (after adjustment for inflation), and age
+ 0.0531GVW** + 0.276 20HF** R2 = 0.641 (19) at resale, it was possible to develop the following
equations:
(4.22) (4.50)
The equations indicate that spare-parts consump- Passenger car:
tion generally
DV = 0.9223A R2 = 0.9693 (24)
1. Increases as roughness increases,
2. Increases as rise and fall or curvature in- Conunercial vehicle (bus and truck):
creases,
3. Decreases as pavement width increases, DV=0.8631A R2 = 0.921 (25)
4. Increases with the age of the vehicle,
5. Increases with the overhauls, where
6. Increases as the congestion mounts up,
7. Is higher for the government sector than for DV depreciated value,
the private sector, and price of a vehicle at age A years divided by
8. Is lower where the operator has a good main- price of a new vehicle, and
tenance policy and adequate maintenance facilities. A age (years).

The equations underscore the point that road im- Vehicle Speed
provements can bring about a good reduction in wear
and tear on spare parts. As roads are improved, the vehicles can travel at
increased speeds. Greater use per day and per year
Maintenance and Repairs Labor Cost results. Thus, the annual depreciation cost can be
spread over a larger kilometer age. Similarly, the
The best equations relate the maintenance and re- fixed costs and crew costs, which are dependent, can
pairs labor cost to the spare-parts cost, which can be spread over a large kilometerage.
be obtained from known roadway and operating char- The analysis of use was accomplished in two
acteristics. The equations are given below: stages: (a) establishing an equation relating vehi-
Transportation Research Record 898 87

cle speed and road characteristics and (b) building Thus the use in kilometers per year can be ob-
up the use equations by using the predicted speeds tained from the use rates in kilometers per day
from the first equation and other additional vari- times 365 times percentage use rates.
ables. The use in kilometers per annum provides the
The equations for vehicle speed are given below: basis for the calculation of depreciation cost
allocatable per kilometer when the yearly depreci-
Car: ation cost is known.

Ve= 55.88 - 0.4969RF** - 0.008 342(RG**/W) R2 = 0.709 (26) Fixed Costs

(41.33)(-7.14) (-5.33)
The fixed costs include crew costs and other fixed
Jeep: costs such as interest on capital, insurance, taxes,
registration fees, license charges, office expenses,
V1 = 52.08 - 0.005 21CV** R2 = 0.844 (27) and overhead.
(22.07)(-7 .66) The crew costs (in rupees) have been found as
follows (15) ($1 U.S. = 9.00 Rs):
Bus:
Cost/Bus/Day Cost/Truck/Day
Ve= intercept - 0.3188RF** - 0.000 795 ORG'"*
(-9 .80) (-4.17) Crew Categor y {Rs) (Rs)
Driver 64 40
+ 2.080W R2 = 0.611 (28) Conductor 48
(7 .12) Cleaner 24 16
136 65
Value of intercept varies from operator to operatori
a typical value is 31.12. The other fixed costs per year have been worked
out as given below (.!.&_) ($1 U.S. = 9.00 Rs):
Truck:
Fixed Costs/Year/Vehi-
VT= 44.78 - 0.3856RF** - 0.002 87(RG/W) R2 = 0.482 (29) cle (Rs)
(63.91) (-7.95) (-4.67) Type of Vehicle Including Excluding
012erator TyEe Taxes Taxes
The equations prove that vehicle speeds are Government Bus 78 142 47 065
closely related to geometry variables, pavement Private Bus 36 088 13 545
width, and roughness. Speeds decrease as roughness Private Truck 68 041 59 057
increases and increase as pavement width increases.
After obtaining the above speed equations in the The crew costs and fixed costs per kilometer can
first stage of regression, the second-stage regres- be worked out by dividing the above values by the
sion was carried out by using the predicted speeds corresponding use rates.
from these equations and some additional factors in The use in kilometers per annum provides the
order to arrive at utilization rates. The equations basis for the calculation of depreciation cost allo-
developed are given below: catable per kilometer if the yearly depreciation
cost is known.
Car:
CONCLUSIONS
UPD = -186.34LU** + 14.94MU*'" +6.4859VC)** (30)
(4.141) The results of the research amply prove the fol-
lowing:
Jeep:
l. Curvature and rise and fall of terrain influ-
UPD = l.7728V1 R2 = 0.035 (31) ence almost all voe components significantly. It
would thus be a wise policy to design the horizontal
Bus: and vertical profiles of roads after a careful con-
UPD = 30.35MU** + 206.75HU** -113.75LU** + 8.01 lV;* sideration of the voe with alternative design var-
(15.71) iables.
2. Pavement width significantly affects most of
+ 10.300HF** - 0.0609LK** R2 = 0.558 (32)
the voe components. 11.lthough paucity of funds in
(2.79) (-4.48)
India dictates the selection of pavement widths, the
economic loss caused by narrow roads can be enormous
•rruck: at certain levels of traffic volume. The ad hoc
norms for widening of roads by stages need to be
UPD = 68.12MU** + 221.35HU** - 85.11LU0 reviewed in light of voe relationships.
+ 5.1637v;* R2 = 0.492 (33) 3. Roughness has a dominant infJ.:uence on almost
(6.87) all of the voe components. The benefits that can be
derived for providing smooth surfaces can be large.
The above equations indicate clearly how the use The need for paving the unsurfaced roads becomes im-
of vehicles increases as speed increases, and speeds minent. Similarly, a careful policy of maintenance
are influenced largely by road geometry, pavement of the road surfaces at a satisfactory roughness
width, and roughness. level becomes self-evident.
In order to arrive at the use rates per year, the 4. Fuel shortages are a serious problem in
data from all the 939 vehicles were analyzed and the India. This research has shown clearly the ways and
percentage use rates were derived. These are given means of achieving fuel economy by improving roads
below: and controlling speeds.
Truck (%) Bus (%) Car and Jeep (%) 5. The voe relationships established can be used
~
Government 52.76 84.4 69.0 for economic appraisal of schemes in India and in
Private 78.80 86.0 86.2 other developing countries with similar road and
Overall 62.36 85.6 75.7 traffic conditions.
BB Transportation Research Record 898

6. When the cost of construction of the initial portes (GEIPOT). Research on the Interrela-
facilities and the cost of maintenance are con- tionships Between Costs of Highway Construc-
sidered with the road-user cost, a better under- tion, Maintenance, and Utilisation. Brasilia,
standing of the total ti: an s portation cos.t is possi- Rept. 1, Mid-Term Report, 1977.
ble. This will be the final a nd sat i sfactoi:y answer 8. L.R. Kadiyali and others. Measurement of Road
to all highway investment questions. The Indian Geometry and Roughness for the Road User Cost
Government's future i:esearch on highway design and Study. Journal of the Indian Roads Congress,
maintenance aspects will go a long way in this di- Vol. 42, No. 2, 1981.
rection. 9. T.D. Gillespie and others. Calibration of
Response-Type Road Roughness Measuring Sys-
The results pi:esented here are likely to be of tems. NCHRP, Rept. 228, 1980.
great value to highway planners in India and in 10. J .c. Wambold. State of the Art of Measurement
other developing countries that have similar condi- and Analysis of Road Roughness. TRB, Transpor-
tions. tation Research Record 836, 1981, pp. 21-29.
11. W.R. Hudson. Rough Roughness: Its Elements
REFERENCES and Measurement. TRB, Transportation Research
Record 836, 1981, pp. 1-7.
l , Basic Road Statistics, 1975-1979. Transport 12. H. Hide. Vehicle Operating Costs in the Carib-
Research Division, Ministry of Shipping and bean: Results of a Survey of Vehicle Opera-
Transport, New Delhi, India, 1980. tors. U.K. Transport and Road Research Labora-
2. Motor Transport Statistics, 1977-1978. Trans- tory, Crowthorne, Berkshire, England, TRRL
port Research Division, Ministry of Shipping Rept. 1031, 1982.
and Transport, New Delhi, India, 1980. 13. L.R. Kadiyali and E. Viswanathan. Roughness
3. C.G. Swaminathan and L.R. Kadiyali. Road User Measurements for the Road User Cost Study.
Cost Study: A Major Research Project in Trans- Central Road Research Institute, New Delhi,
portation in India. Presented at the Workshop India, RUCS Tech. Paper 80, Fifteenth Quarterly
on Transportation Research, School of Planning Report, 1981.
nnd Architecture, New Delhi, India, 1982. 14. L.R. Kadiyali and others. Depreciation of Ve-
4. Road User Cost Study: Draft Inception Report. hicles. Central Road Research Institute, New
Central Road Research Institute, New Delhi, Delhi, India, RUCS Tech. Paper 89, Seventeenth
India, 1977. Quarterly Report, 1982.
5. Road User Cost Study: Final Report, Vols. 15. T.V. Gopalaswami and others. Crew Cost of Ve-
1-8. Central Road Research Institute, New hicles. Central Road Research Institute, New
Delhi, India, 1982. Delhi, India, RUCS Tech. Paper 107, Eighteenth
6. H. Hide and others. The Kenya Road Transport Quarterly Report, 1982.
Cost Study: Research on Vehicle Operating 16. T.V. Gopalaswami and others. Fixed Cost of Ve-
Costs. U.K. Transport and Road Research Labo- hicles. Central Road Research Institute, New
ratory, Crowthorne, Berkshire, England, TRRL Delhi, India, RUCS Tech. Paper 108, Eighteenth
Rept. 672, 1975. Quarterly Report, 1982.
7. Empresa Brasileira de Planejamento de Trans-

Socioeconomic Evaluation and Upgrading of Rural

Roads in Agricultural Areas of Ecuador
JACOB GREENSTEIN AND HAIM BONJACK

A national socioeconomic methodology to evaluate rural roads was developed. The provinces of El Oro, Guayas, and Los Rios, lo-
This methodology presents the relationship among road accessibility, rainfall, cated along the Pacific Coast of Ecuador, are the
drainage conditions, engineering properties of the subgrade and pavement ma· biggest agricultural producers in the country.
terials, cost analysis, end agricultural benefits. A production-loss function was These three provinces cover 32 000 km 2 (about 11. 3
developed to determine the relationship between road surface conditions and percent of the total area of Ecuador). The annual
losses in quality and value to agricultural products. Approximately 6000 km
agricultural exports of this region in 1978 were
of rural roads along the Pacific Coast of Ecuador, 70 percent of which are dirt
roads that are not usable during the wet season, were evaluated by using this valued at $240 million. The population in 1979 was
methodology. A road inventory was conducted to evaluate surface conditions, 3 million, approximately 35 percent of which, or 1.3
soils and materials properties, drainage and structural facilities, geometric prop· million, is the rural population.
erties, and accessibility. The percentage of the cultivated area was determined There are approximately 6000 km, or about 1000
together with the type of crop and the influence area of each road. Population links, of rural roads in these three provinces.
density and the illiteracy rate within the area of influence were also determined. About 70 percent of these are dirt roads (Figures l
It was concluded that the socioeconomic evaluation of road improvement is and 2) that are not accessible during the wet sea-
best executed in two stages; the first, as a threshold analysis, is to determine son. The other 30 percent are constructed mainly of
the most economical alternative for each given traffic volume. The second local granular-cohesive materials that have low to
stage is a complete socioeconomic analysis with determination of the internal
medium plasticity. About 55 percent of these rural
rate of return, first-year benefit ratio, net present value, population density,
and illiteracy rate. Under a total budget of U.S. $34 million, 1300 km were roads carry less than 20 vehicles/day in both direc-
upgraded as a result of this study. tions; 75 and 90 percent carry less than 50 and 100
vehicles/day, respectively. Only 1.5-2.0 percent of