~-d Transport& RoadResearchLaborato~

L~5r~~]

I Il:i:il Research Report 333

Pakistan road freight industry: Theproductivity and time use of commercialvehicles(by J L Hine

TRANSPORTAND ROAD RESEARCH LABORATORYDepadment of Transpofi

RESEARCH REPORT 333

PAKISTAN ROAD FREIGHT INDUSTRY: THE PRODUCTIVITY AND TIMEUSE OF COMMERCIAL VEHICLES

by J L Hine

Crown Copyright 1991. The work described in this report forms part of the programmecarried out for the OverseasDevelopment Administration, but the views expressed are not necessarily those of the Administration. Extractsfrom the text may be reproduced, except for commercial purposes, provided the source is acknowledged.

Overseas UnitTransport and Road Research LaboratoryCrowthorne, Berkshire, RG11 6AU1991

ISSN 0266-5247

Digest of Research Report 333 1991

PAKISTAN ROAD FREIGHT INDUSTRY THE PRODUCTIVITY AND

TIME USE OF COMMERCIAL VEHICLES

by

J L Hine

Introduction

A study of Pakistan’s freight transport industry was carried out by the Overseas Unit of the Transport and RoadResearch Laboratory in cooperation with the National Transport Research Centre, Islarnabad. As part of this studya ‘Vehicle Activity Survey’ was undertaken in which the different vehicle activities (i.e. loading, unloading,moving, resting) were recorded and timed as they occurred. The data were collected by both truck drivers andsurvey staff as they travelled with each vehicle on a continuous basis for periods lasting from five days to overfour weeks. hformation on distances travelled, costs incurred and freight tariffs charged was also recorded. In totaldata on 45 survey periods were collected covering 405 loaded and 327 empty trips.

The main purpose of this survey and the subsequent analysis was to investigate the following three aspects of timerelated costs:

a) The use of time savings following road investment.

b) The amount of time a vehicle spends in different activities (e.g. in traveling, loading, unloading, or at rest).

c) The relative importance of time and distance in the explanation of costs and tariffs.

Findings

The use of time savings

From a regression analysis relating trips made per day to mean trip working time elasticities between-0.84 to-1 were found. These imply that in Pakistan’s conditions, journey time savings following road investment

are likely to be fully translated into extra trips.

Time distribution of vehicle activities

Freight vehicles were in movement for 40 per cent of the time and including loading and unloading vehicles werein active use for over 12 hours per day.

The relative importance of time and distance

Vehicle operating costs depend on both time and distance but there is uncertainty as to the relative importance ofeach. Pakistan has a very competitive freight transport industry so tariffs probably reflect underlying costs.

A

dA Department of Transport<

A statistical analysis of the relationship between tariff, time and distance was undertaken but the high degree of

correlation between time and distance made it difficult to identify the separate explanatory power of eithervariable.

The work described in this Digest forms part of a programme of joint research between the Overseas Unit (HeadJ S Yerrell) of the Transport and Road Research Laboratory, UK and the National Transport Research Centre(Head M S Swati), Pakistan.

If this information is insufficient for your needs a copy of the full Research Report RR333, may be obtained, free

of charge (prepaid by the Overseas Development Administration) on written request to the Technical Informationand Library Services, Transport and Road Research Laboratory, Old Wokingham Road, Crowthorne, Berkshire.

CROWN COPYRIGHT. The views expressed in this Digest are not necessarily those of the Department ofTransport. Extracts from the text may be reproduced except for commercial purposes, provided the source isacknowledged.

CONTENTS

Abstract

1. Introduction

2. The valuation of commercial vehicletime savings

3. The road freight transport industryin Pakistan

4. The vehicle activity survey

4.1 Survey procedure

4.2 The pattern of vehicle use

5. The relationship between trip frequencyand trip duration

6. The relationship between revenue andjourney time and distance

7. Conclusions

8. Acknowledgements

9. References

Appendix A

Page

1

1

1

3

3

3

4

6

11

14

14

14

15

PAKISTAN ROAD FREIGHT INDUSTRY: THEPRODUCTIVITY AND TIME USE OF COMMERCIALVEHICLES

ABSTRACT

As part of a wider study of freight transport undertaken inPakistan during 1985-87 a survey of vehicle utilisationwas carried out. Different vehicle activities were timedand recorded on a continuous basis for periods lastingfrom five days to four weeks. Using these data it waspossible to investigate three aspects of time relatedcosts:-

i)

ii)

iii)

1

A series of elasticities relating trips to travel timeshowed that, in Pakistan’s conditions, time savingsfollowing road investment are likely to be fully used.

An analysis of vehicle time budgets found thatfreight vehicles were in active use (traveling,loading, and unloading) for more than 12 hours perday.

Both time and distance could each independentlyprovide a good explanation of tariffs. Althoughdifficult to identify their separate influences, the bestestimate suggests that time and distance accountfor 45 per cent and 55 per cent of tariffs respec-tively.

INTRODUCTION

The rapid growth in traffic volumes experienced by manydeveloping countries has, in recent years, led to anincreased interest in providing high capacity roads thatare primarily designed to reduce congestion and providejourney time savings. Despite this, the analysis of journeytime savings in developing countries has attractedcomparatively little research. The two main elements oftime-related costs are passenger time, and vehicle andcrew time. To improve our understanding of the latter forcommercial vehicles this Repoti presents and analysesdata collected from a survey carried out in Pakistanduring 1985-86.

The Repoti examines three aspects of time-related costs.These are:-

a) the use of time savings following road investment.Common-sense suggests that if road investmentbrings about journey time savings then, on average,commercial vehicles should be able to make use ofthe time saved by making extra trips. Howeverthere is disagreement among various investigations(Fleischer 1963, Dawson and Vass 1974, Thomas1983) about the extent to which this can beachieved.

b)

c)

the amount of time a vehicle spends in differentactivities (e.g. in traveling, loading, unloading, or atrest). This data is needed to help estimate thebenefits from changes in travel time for those costcomponents that are time dependent.

the relative importance of time and distance in theexplanation of costs and tariffs. Most componentsof vehicle operating costs can be separated intothose that are clearly time dependent (e.g. driver’swages) and those that are clearly dependent ondistance travelled (e.g. tyre wear). However forsome components, (such as depreciation ormaintenance), it is known that there is a depend-ency on both time and distance although there issome uncertainty as to the relative importance ofeach. An analysis of the extent to which tariffs canbe explained by time worked and distance travelledwould be useful in checking the validity of currentmodels of vehicle operating costs used in roadinvestment appraisal.

To collect information on the time utilisation of freightvehicles a “Vehicle Activity Survey” was undertaken. Datawere recorded by both truck drivers and survey staff whotravelled with each vehicle for periods lasting betweenfive days and four weeks throughout Pakistan. Thepurpose of the survey was to record the number of tripsmade and the time spent by the vehicle in differentactivities (namely traveling, loading, unloading, resting,and under repair) during each period. The data wererecorded 24 hours-per-day. In addition data were alsocollected on the distance travelled, the revenues earnedand the expenses incurred.

This Report forms part of a study of Pakistan’s freighttranspoti industry that was carried out under a pro-gramme of cooperative research between the OverseasUnit of the Transport and Road Research Laboratory andthe National Transport Research Centre (Islamabad). Arepoti providing a general description of the industry andpresenting more information on vehicle operating costsand freight tariffs has been published separately (Hineand Chilver, 1991).

2 THE VALUATION OFCOMMERCIAL VEHICLE TIMESAVINGS

In order to value the benefits of journey time savings forcommercial vehicles arising from road investmentassumptions are necessary about the extent to which thepotential time savings may be translated into productiveuse. Some authors, such as Dawson and Vass (1974),

1

assume constant working hours per year in their formula-tion of vehicle operating costs; this implies that all timesavings following a road investment will be fully used.

In vehicle operating cost formulations, which do notassume constant working hours, annual time-relatedcosts are divided by distance travelled per year. Thelatter figure may either be estimated by a formula derivedfrom vehicle speed or estimated by the user. Neitherprocedure is very satisfactory. For example both Winfrey(1969) and DeWeille(1966) have proposed a formulafrom which it can be calculated that a one per centdecline in travel time would bring about a one per centincrease in distance travelled. It is assumed that the timespent traveling per day will remain constant. This impliesthat, with a fall in journey times, the time spent workingper day (including traveling, loading and unloading) mustrise as more trips are made because of the increasedtime devoted to loading and unloading. The assumptionof constant traveling time is clearly unrealistic for com-mercial vehicles where loading and unloading time is animportant component of the working day. This approachwill nearly always tend to overestimate time savingsbenefits.

Other vehicle operating cost model formulations (such asTRRL’s “Road Transport Investment Model” or the WorldBank’s “Highway Design Model”) provide the user withthe option to predict utilisation following road investmentor to choose one of a variety of models to achieve thesame result. Again no empirical evidence is presented toshow how utilisation is likely to change following roadinvestment.

The assumption that travel time savings can be readilytranslated into extra trips has been questioned byFleischer (1962) and Thomas (1983). Their empiricalinvestigations suggest that there is an inflexibility invehicle operations relating to the constraints of drivers’hours and the scheduling of work which will prevent traveltime savings from being fully used, particularly in theshort-term.

It maybe thought that the more inflexibility there is invehicle operations then the greater the probability thattime savings will not be fully used following road invest-ment. To test this the author built a computer basedsimulation model in which the operations of a freightvehicle were constrained by a set of drivers’ hoursregulations and permitted loading and unloading times.The vehicle was assigned to undertake a randomsequence of trips of different lengths. The constraintswere shown to impinge on vehicle operations where itwas predicted that certain activities (i.e. loading, unload-ing and short trips) could not be completed within theworking day. When this happened the vehicle wasassumed to remain idle for the rest of the day and theactivity was postponed to the next day.

Using different input assumptions the results of the modelsuggested that while additional constraints on vehicleoperations could be expected to reduce efficiency therewere no a priori grounds to suggest that they wouldreduce the probability of using time savings. The modelprovided as many cases where total idle time fell (i.e.time savings were more than fully used) as cases where

total idle time rose (i.e. time savings were not fully used)following a reduction in trip times. The exact outcomewas shown to be dependent upon the particular triplength distribution and the set of constraints assumed inthe model.

There are several ways of trying to estimate how timesavings may be used following road investment. Oneapproach is to ask operators to estimate what theirresponse would be, given the predicted reduction injourney times. Even if it be assumed that each operatorcan forecast correctly his response the results are verydifficult to interpret because unless a lot of detail is knownthere is no simple way of comparing those who say theycan make productivity gains with those who say theycannot. It is possible, for example, that time savings will,on average, be fully used if only one operator in twentycan benefit.

Another approach is to look at changes in utilisationbefore and after an investment has been made. This wasdone by both Fleischer (1963) and by Thomas (1983).Fleischer looked at one company operating on the routebetween Grants Pass, Portland and Seattle on the WestCoast of the USA. There was a series of road improve-ments which gradually reduced journey times over manyyears. Fleischer found that an extra trip between Portlandand Seattle could only be made after journey times hadbeen reduced to a given amount. On the route betweenGrants Pass and Potiland he found that it was only whenthe firm could relocate its depot could the reductions injourney times be translated into an increase in thedistance travelled which occurred many years after thefirst road improvements, though it is not known howrepresentative or complete the case was. Without a totalsurvey of all operators using the route it is not knownwhether other operators were able to make use of thetime savings. Likewise it is possible that in other situa-tions with only one major route impressive productivitygains could occur with only very small time savings.

Thomas carried out an historical “before and after”analysis as well as a cross-sectional analysis to deter-mine how vehicle productivity might change followingroad investment. For the historical analysis, data wascollected on vehicle productivity before and after the newKuala Lumpur - Karak highway was opened in WestMalaysia which reduced average vehicle trip times from 4to 5 hours by about 45 minutes. By collecting data on theoperations of various types of commercial vehiclesThomas could not find any dramatic improvements invehicle productivity following the opening of the new roadsection. In fact for certain vehicle types the level of tripmaking per day fell in the Kuala Lumpur area; also duringthe period which was covered by the data collection largechanges were recorded in both the Malaysian commer-cial vehicle fleet numbers and in the levels of economicactivity. Both of these factors could have swamped anybeneficial effect of the road on vehicle productivity.

In the Malaysian cross-sectional analysis Thomasderived a number of elasticities for different categories ofcommercial vehicles between trips made per day andaverage travel time per trip. Most of the Malaysian datawas collected from a roadside interview survey in which

2

drivers were asked what trips had been made in theprevious 24 hours. The elasticities found ranged from-0.2 to -0.6; i.e. if travel times were reduced by one percent then trip making could have been expected to riseby between 0.2 and 0.6 per cent.

This analysis suggests that the models which predict aone per cent increase in distance travelled for one percent decline in travel time will clearly overestimate theeffects of journey time savings. However to estimate theextent to which vehicle working time savings are used itis necessary to include loading and unloading time withintrip times in the calculation of the elasticities. Loading andunloading times were not collected in the Malaysian studyand as a result the derived elasticities underestimate theextend to which time savings may be used followingjourney time savings.

It was against this background that it was decided todevelop further the cross-sectional analysis using vehicleactivity data from Pakistan.

3 THE ROAD FREIGHTTRANSPORT INDUSTRY INPAKISTAN

Currently in Pakistan there are about 45,000 trucks inoperation, of which about 95 per cent are privatelyowned. Road freight transpoti is organised on a freemarket basis and freight tariffs are determined competi-tively by supply and demand with little governmentintervention. The industry is dominated by a large numberof individual entrepreneurs operating a ‘hire and reward’service. Entry into the industry is cheap and easy; thereis a relatively lax licensing system and little enforcementof axle load limits or construction and use regulations.

During the study several different surveys were under-taken. General information on private road freighttransport was collected principally from the RoadsideInterview Survey of 3,500 truck drivers. Three quatiers ofthe trucks surveyed were two-axle Bedford trucks, 14 percent were two-axle Japanese trucks and the remainderwas divided between three-axle rigid vehicles andtractor-trailer combinations.

Less than one per cent were owned by a company for itsown account operations. Most trucks were found tooperate on the basis of picking up business where theycould and going from job to job as demand required. Itwas very common for drivers to work away from base forup to three weeks. The Roadside Interview Survey foundthat on average drivers of Bedford trucks returned tobase after 7 days and returned to their families after 17days.

The driver was found to be responsible for finding theload, for collecting revenue, and for repairing the truck.When he returned to base he was expected to accountfor the revenue earned and the expenditures incurred.About 80 per cent of drivers were employees, 17 per centowned their own vehicle and the remaining three per centowned a part share of the vehicle.

There is an extensive network of freight agents whoassist drivers to find a load. Virtually all agents werefound to be connected by telephone which appeared toplay an important pati in their business. Over 60 per centof loads were found using agents. In the vast majority ofcases vehicles could be found for a consignor within onehour.

Little specialisation in body types was recorded. About 85per cent of the Bedfords and 60 per cent of other truckswere high sided, while 8 per cent of Bedfords and 20 percent of other trucks were tankers.

Virtually all freight vehicles in Pakistan were found tohave sufficient space within the cab to seat four people.In addition there was usually a purpose built space on top

of the cab where a driver or his assistant could rest orsleep whilst the vehicle was in motion. Just over half thetrucks were found to have two drivers and one assistant,the remainder had just one driver and one assistant.

4

4.1

THE VEHICLE ACTIVITYSURVEY

SURVEY PROCEDURE

In Pakistan a high proportion of journeys last for morethan one day. In order to get an accurate picture of timingand duration of the different vehicle activities a “VehicleActivity Survey” was undertaken in which activities wererecorded and timed as they occurred. Data were col-lected from a sample of trucks on a continuous basis forperiods lasting from five days to over four weeks. Thedata were collected by both survey staff and by drivers.The survey staff travelled with their allotted truckscontinuously throughout the period, if necessay sleepingon board the truck as it travelled. Where drivers wereused to collect data they were paid an additional sum torecord their activities.

In total over 600 days of useful data were collected,about one fifth of this being recorded by drivers, the restby survey staff. This comprised 24 periods of data relatedto conventional two-axle Bedford trucks, seven periods toBedford tanker trucks and 14 periods to conventional twoand three axle Mercedes trucks. The latter trucks wereowned and driven by Afghan refugees. Basic datarelating to the trucks participating in the Vehicle ActivitySurvey are given in Table Al of the Appendix.

In view of the difficulty encountered in placing the surveystaff with vehicles many staff made use of their contactsto find vehicles with which they could work. Wherepossible, recorders sensitive to vehicle vibration werefitted to provide an independent check on the manualrecords of the survey staff.

Detailed records of the timing of all movements, restperiods, loadings, unloading, waiting periods, emergen-cies and repairs were collected on a continuous basis.Vehicle stops of less than 15 minutes were ignored.Cooperation was sought from the other drivers and

3

assistants traveling with the vehicle to help record thosedetails which the survey assistant or driver missed whilsthe was asleep. Additional data on the distance travelled,

costs incurred and revenue earned were also recorded.

4.2 THE PA~ERN OF VEHICLE USE

The Vehicle Activity Survey collected data on trucksengaged on a variety of patterns of operation andcovered a wide spectrum of trip lengths. The 45 surveyperiods covered 405 loaded trips and 327 empty trips.Distances were recorded for 92 per cent of the trips andfor these trips the mean trip distances, loaded and empty,were 347 kms and 150 kms respectively. Fifteen per centof loaded trips and 42 per cent of empty trips were under50 kms while eight per cent of the loaded trips and twoper cent of the empty trips were over 1000 kms.

To analyse the pattern of vehicle productivity and timeuse it was necessary, as far as possible, to allocate timespent in different activities into empty and loaded tripperiods. For each trip the times spent moving, loading,unloading, resting, or under repair were separatelytotalled. Because of the long journey distances one tripwas usually broken up into a series of movements andrest periods. On average each loaded trip and eachempty trip was composed of approximately 5 and 3separate movements respectively.

Because the vehicle was at least partially loaded duringloading and unloading these activities were defined to bepart of a loaded trip. In the analysis no distinction wasmade between resting time and waiting time. Waiting tounload was counted as part of a loaded trip while waitingto load after completing an empty movement wascounted as part of that empty trip. For a loaded trip tofinish the truck would have to be completely unloaded; soa sequence of multiple partial loadings or unloadingwould all be counted as part of one loaded trip. A loadingis necessary to finish an empty trip. So a series of emptymovements (made in any direction) was counted as partof one empty trip. Besides the loaded and empty tripsthere were also periods between two loaded trips whenthe vehicle did not move empty. In total 63 periods of thistype were recorded. Detailed activity data is shown inTable A2 of the Appendix. The mean periods spent ineach activity for the whole survey are shown in Table 1.

On average loaded trip periods lasted 21.8 hrs and emptytrip periods lasted 16.5 hrs. In total trucks were found tobe loaded 56 per cent of the time. Rest and waiting timeaccounted for 30 per cent of loaded trip time, but overall63 per cent of rest and waiting occurred while the truckwas empty. The average duration between the end ofloading and arrival at the destination was found to be14.9 hrs which gives an average loaded journey speed(including intermediate rest periods) of 23 kph. This is thesame overall journey speed found for loaded Bedfordtrucks in the Roadside Interview Survey (Hine andChilver, 1991) . Using data from the Vehicle ActivitySurvey an estimate of 109,000 kms was calculated forannual vehicle travel. This is very close to the meanestimate of 112,000 kms annual vehicle travel found forBedford trucks in the Roadside Interview Survey.

The distribution of total loaded and empty trip times foundby the survey is shown in Figure 1. Although 70 per centof loaded trips took less than 24 hrs, 13 per cent lastedlonger than 48 hrs.

A breakdown of time by activity is shown in Figure 2. Thisdemonstrates the high degree of time utilisation of theSUNeyed vehicles. Most vehicles worked round the clockwith activity interrupted only by short rest and waitingperiods. Vehicles were found to be moving 40 per cent ofthe time and loading or unloading a further 11 per cent ofthe time. Rest periods accounted for 46 per cent of thetime. In total trucks were in active use 51 per cent of thetime (i.e. over 12 hours per day).

The high level of working activity was maintainedthroughout the whole week. Apart from a slight reductionin loading and unloading, the level of activity was notsubstantially reduced on a Friday (the normal rest day)compared with other days of the week.

Figure 3 shows how vehicle use changed throughout theday. The most active movement times were between16.00 hrs and 02.00 hrs. Trucks were most likely to be atrest between 02.00 hrs and 08.00 hrs. Even at thequietest time of day, 06.00 hrs, 37 per cent of vehicleswere found to be working. The most active loading andunloading times were in the middle of the working daybetween 08.00 hrs and 18.00 hrs.

Figure 4 gives the distribution of the stati times of loadingand unloading. Although most loading and unloading took

TABLE 1

Mean Times Spent on Each Activity for Loaded and Empty Trip Periods

‘Moving Loading Unloading Resfing Repair TotalHrs Hrs Hrs Hrs Hrs Hrs

Loaded Trips 10.8 2.2 1.9 6.5 0.3 21.8Empty Trips 4.4 11.3 0.9 16.5Empty Periods BetweenLoaded Trips 7.1 0.6 7.7

Source: Vehicle Activity Survey

4

o

A-.

3m

Per cent of all vehicles

o ~m

&03 0

0 0 0

In o

0-.~

w oN

Per cent

G gG E

0-4.0.I I

4-8

8–12 -

12-16

16–20

20-24

24–28$-.:, 28-323m~ 32-36

38A0

4044

44+6

48-52

52-56

56-60

>60

—

Start times

Fig. 4 Loading and unloading stati times

place during the “normal working day” 13 per cent ofloading and 29 per cent of unloading started between20.00 hrs and 6.00 hrs. Unloading (which lasts less time)was more evenly distributed throughout the day thanloading. Because it is usually necessary to hire Iabour forloading and unloading the industry appeared to be quiteflexible in allowing for it to take place throughout thenight.

An analysis of the data showed that a rest or waitingperiod occurred in 57 per cent of the time periodsimmediately prior to a loading. On average this periodlasted 8.9 hrs. In other words, as expected, driverstended to take a break after completing a delivery andbefore collecting their next load. But many also took theopportunity for a break before unloading. In 40 per cent ofthe periods immediately prior to an unloading a rest orwaiting period occurred which lasted on average 6.1 hrs.

5 THE RELATIONSHIPBETWEEN TRIP FREQUENCYAND TRIP DURATION

In order to estimate the extent to which journey timesavings could be translated into extra trip making a seriesof elasticities was derived from the Vehicle ActivitySurvey data relating trip frequency to trip duration.

The survey data were collected over periods lasting up tofour weeks. During the survey some of the truck drivers

6

took their normal “day of~ recreational rest while otherdrivers did not. In order to estimate the extent to whichthis lack of uniformity may have biased the result anadditional analysis was undertaken using data adjustedfor long rest periods. A relationship had been foundbetween rest days per month and loaded journey timefrom data collected from the Roadside Interview Surveyand this was used to smooth the total amount of timedevoted to long rest periods in the “adjusted data se~referred to below. A further analysis was cartied outwhich excluded data from the tanker trucks which havetheir own unique pattern of operation. The following threedata sets were prepared:

i) The basic data set (45 cases)

ii) The adjusted data set (45 cases)

iii) The adjusted data set excluding tankers (38 cases)

Each data set was derived from the 45 survey periods ofthe Vehicle Activity Survey. For each case the number oftrips made per day, the moving and working time per tripand the mean working time per day was calculated.

One approach to estimating the utilisation of time savingsis to determine the relationship between total workingtime per day and average movement time per trip. If timesavings cannot be fully used then working time per day(i.e. all moving, loading and unloading time) would risewith mean movement time per trip. This analysis ispresented in Table 2.

.

TABLE 2

Regressions Relating Total Working Time Per Day To Movement Time Per Trip

Basic data set. N =45

i) Total working time per day = 11.0 + 0.11 (movement time per trip)(se = 0.04)

R2=0.13

Adjusted data set. N=45

ii) Total working time per day = 10.8 + 0.074 (movement time per trip)(se = 0.035)

R’= 0.09

A@usted data set excluding tankers. N=38

iii) Total working time per day = 11.5- 5.7x 10-5(movement time per trip)(se = 0.07)

R2=0

Although both of the first two regressions shown inTable 2 have positive slope coefficients and are signifi-cant at the 5 per cent level of probability they have verylow explanatory power. The last regression has no

significance at all. The analysis shows no consistentrelationship between total working time per day andmovement time per trip - hence there is little to suggestthat journey time savings cannot be fully used.

Elasticities derived between the number of trips made perday and movement time per trip (excluding loading andunloading time) cannot be used directly to estimate theextent to which time savings can be used followingjourney time savings. However they are useful forindicating how the total distance travelled per day mightchange with a change in travel time. Total distancetravelled may be calculated from the average journeydistance and the predicted number of trips made per day.

An elasticity of -1 between trips made and movementtime per trip is unlikely where loading and unloadingtimes are significant. An elasticity of this magnitude wouldimply that total working time per day (including moving,loading, and unloading time) will increase with travel timesavings.

The elasticity is the coef~cient “b” in the following regres-sion:-

Log (trips) = a + b x Log (journey time)

Table 3 gives the results of regressions relating trips perday to mean trip movement times, separately for “loadedtrips” and for “all trips”.

“All trips per day” refers to the mean number of loadedand empty trips made per day. Precise definitions ofloaded and empty trips are given in Section 4.2 above.

The regressions have high R’ values and are verysignificant. The elasticities lie within the range -0.70to -0.77, which are much higher than those found in theMalaysian study.

To estimate the extent to which time savings can be usedfollowing road investment elasticities using total workingtime were calculated. Regressions showing theseelasticities are given in Table 4. For each data set threeregressions are presented which relate the followingvariables:-

i) “all trips per day” to “mean total working time pertrip”

ii) “loaded trips per day” to “mean loaded working timeper loaded trip” ‘

iii) “loaded trips per day” to “mean total working timeper loaded trip”

The last provides an estimate (as far as is practical inPakistan conditions) of the response of “round tripmaking to total working time per “round trip”. In theanalysis total working time refers to all empty and loadedmoving time as well as loading and unloading time whileloaded working time is similar but excludes emptymovement time. ,.

Elasticities in Table 4 range from -0.84 to -1. All of theregressions are highly significant and have high R2values. Figure 5 gives a typical plot of the data used inRegression (xv). Overall, by introducing loading andunloading into the analysis, the elasticity values haverisen by 25 per cent. The mean value of the elasticities is-0.92; i.e. we could expect commercial vehicles to makeuse of over 90 per cent of the time savings following roadinvestment.

7

TABLE 3

Regressions Relating Trips Per Day to Moving Time Per Trip

Basic data set. N = 45

iv) Log (all trips per day) = 0.682-0.698 Log (movement time per trip)

(se 0.036)R,= Ogo

v) Log (loaded trips per day) = 0.518-0.709 Log (loaded movement time per trip)(se = 0.046)

R,= 085

Adjusted’data set. N =45

vi) Log (all trips per day) = 0.715-0.755 Log (movement time per trip)

(se = 0.032)R2 = 0.93

vii) Log (loaded trips per day) = 0.56-0.768 Log (loaded movement time per trip)

(se = 0.042)R2 = 0,88

Adjusted data set excluding tankers. N=38

viii) Log (all trips per day) = 0.724-0.765 Log (movement time per trip)(se = 0.043)

R’= 0.90

ix) Log (loaded trips per day) = 0.506-0.702 Log (loaded movement time per trip)(se = 0.048)

R2 = 086

a

TABLE 4

Regressions Relating Trips Per Day To Working Time Per Trip

Basic Data Set. N= 45

x) Log (all trips per day) = 0.909-0.838 Log (mean total worting time per trip)(se = 0.047)

R’= 0.88

xi) Log (loaded trips per day) = 0.873-0.913 Log (mean loaded working time per loaded trip)(se = 0.061)

R2 = 084

xii) Log (loaded trips per day) = 0.90-0.864 Log (mean total workng time per loaded trip)(se = 0.045)

R2 = Ogo

Adjusted Data Set. N =45

xiii) Log (all trips per day) = 0.968-0.913 Log (mean total working time per trip)(se = 0.041)

R2 = 0.92

xiv) Log(loaded trips per day) =0.951 -0.995 Log(mean loaded working time per loaded trip)(se = 0.054)

R2 = 0.89

xv) Log (loaded trips per day) = 0.969-0.932 Log (mean total worting time per loaded trip)(se=O.038)

R2 = 093

Adjusted data set excluding tankers. N=38

xvi) Log (all trips per day) = 1.00-0.95 Log (mean total working time per trip)(se = 0.054)

R2 = 089

xvii) Log(loaded ttips per day) = 0.88-0.925 Log(mean loaded working time per loaded trip)(se = 0.066)

R2 = 085

xviii) Log(loaded trips per day) = 1.03-0.98 Log( mean total working time per loaded trip)(se = 0,052)

R2 = O.gl

9

0.:

(

–0.4

–0.6

–0.8

\

.

.

●

✎

✎

✎✎

✎

●✼

✎ ✎

✎☛

✎

✎

✎

✎

✎

1 1 I 1 I 1 I I

\

I 10.6 0.9 1.2 1.5 1.8 2.1

Log working time per loaded trip

Fig. 5 Loaded trips per day against working time per trip

6. THE RELATIONSHIPBETWEEN REVENUE ANDJOURNEY TIME ANDDISTANCE

Road freight transport in Pakistan is very competitive andin this environment there is every reason to believe thatrevenues and tariffs reflect operating costs. To examinewhether it was possible to identify the proportions ofoperating costs that are dependent on time and distancean analysis was undertaken to determine the extent towhich tariffs could be explained by these variables. Thisinformation could be useful in checking the relativeimportance of time related costs within the more complexcost models used in road investment appraisal. Theanalysis was carried out using simple and multipleregression techniques with data collected from theVehicle Activity Survey.

Four sets of data from the Vehicle Activity Survey wereanalysed. Because the tariff rates for carrying petroleumproducts are set by oil companies directly on a distancebasis they were excluded from the analysis.

The sets of data were as follows :-

i) Bedfords - grouped data, 26 observations, datarelating to periods lasting from five days to fourweeks.

ii)

iii)

iv)

Bedfords - empty and loaded trip data, 176 obser-vations, data relating to trips including emptymovements.

Bedfords - loaded trip data, 201 observations, datarelating to loaded trips only.

Mercedes - loaded trip data, 77 observations, datarelating to loaded trips only (non tankers).

Simple regressions relating tariffs (Rs) to distance (kms)and tariffs to time (hrs) are shown in Tables 5 and 6.They show that both time and distance are good explana-tory variables of tariffs. The relationships are verysignificant and the R2 values are high. Distance is abetter explanatory variable for the grouped Bedford dataset and for the Mercedes data set whereas time is abetter explanatory variable for the two Bedford trip datasets. Moving time appears to be a better explanatoryvariable than working time.

Multiple regressions relating tariffs to working time anddistance are shown in Table 7. These provide a slightlybetter explanation than the simple regressions. The R*value is raised in three of the four data sets by betweentwo and three per cent compared with best alternativesimple regression. However in the last two regressionsthe second term has very little significance. As expected,there is clearly a high degree of correlation presentbetween time and distance.

TABLE 5

Regressions Relating Tariffs to Distance

Bedfords - grouped data set. N =26

xix) Tariff Revenue (for period)= 1499 + 2.50 (Empty & Loaded Distance)(se = 0.17)

R’= 0.90

Bedfords - empty and /oaded trip data set. N = 178

xx) Trip Tariff = 286 + 2.28 (Empty & Loaded Trip Distance)(se = 0.12)

R,= 0.67

Bedfords - loaded trip data set. N = 201

xxij Trip Tariff = 363 + 2.76 (Loaded Trip Distance)(se = 0.15)

R’= 0.62

Mercedes - loaded trip data set. N = 77

xxii) Trip Tariff = 165+ 6.14 (Loaded Trip Distance)(se = 0.39)

R2 = 0.76

11

TABLE 6

Regressions Relating Tariffs To Time

Bedfords - grouped data set. N = 26

xxiii) Tariff Revenue(for period) = -1634 + 86.2 (Empty & Loaded Work Time)(se = 7.13)

R’= 0.86

Bedfords - empty and /oaded trip data set. N = 178

xxiv) Trip Tariff = 246 + 73.7 (Empty & Loaded Moving Time)(se = 3.77)

R2 = 0.69

xxv) Trip Tariff = 93.6 + 66.9 (Empty & Loaded Work Time)(se = 3.47)

R2 = 0.68

Bedfords - loaded trip data set. N = 201

xxvi) Trip Tariff = 250 + 97.7 (Loaded Moving Time)(se = 3.9)

R2 = 0.76

xxvii) Trip Tariff = 34.6 + 87.6 (Loaded Work Time)(se = 3.6)

R2 = 075

Mercedes - loaded trip data set. N = 77

xxviii) Trip Tariff = 452 + 147 (Loaded Moving Time)(se = 12.2)

R2 = 066

xxix) Trip Tariff = 381 + 115 (Loaded Work Time)(se = 12.9)

R2 = 0.52

One way of testing the usefulness of the multiple regres-sion model when multicollinearity is present between theindependent variables is to compare the R2 value of theregression with the squared correlation coefficientbetween the independent variables. This is done in Table8 where it can be seen that for the two Bedford trip datasets the multiple regression R2 value is lower than thesquared correlation coefficient between the independentvariables. This suggests that multiple regression resultsfor these two data sets are not satisfactory. It appearsthat the multiple regression for the grouped Bedford dataset provides the best explanation for Bedford truckoperating costs.

The analysis is designed to reflect the relative importanceof time and distance for the freight transport market,however different vehicle types do operate in differentmarkets with different tariff levels. The age and size ofthe Mercedes trucks may account for the relatively lowproportion of tariffs explained by working time. Thesevehicles (which are imported from Afghanistan) are very

old and consequently have relatively low capital values,in addition, because they carry much more than theBedfords, Iabour costs account for a lower proportion oftheir total revenues and costs.

Table 8 provides an estimate of the percentage of themean tariff of each data set that is explained by time anddistance in the different multiple regressions. TheBedford grouped data set and the Mercedes data sethave a total percentage “explained” by time and distancethat is greater than 100 per cent. This is because theconstant term in the regression equation is negative.

Table 8 suggests that multiple regression analysis cannoteasily be used to explain tariffs. There is considerablevariation in the relative importance of time and distancefor the different data sets. Assuming that the groupedBedford data set provides the best explanation andadjusting for the constant term then working time anddistance account for 45 and 55 per cent of tariffs respec-tively. The grouped data regression suggests that

12

TABLE 7

Regressions Relating Tariffs to Time And Distance

Bedfords - grouped data set. N =26

xxx) Tariff Revenue (for period) = -350 + 1.61 (Empty & Loaded Distance)(se = 0.324)

+ 35.3 (Empty & Loaded Work Time)(se =11 .4)

R’= 0.93

Bedfords - empty and loaded trip data set. N = 178

xxxi) Trip Tariff= 141 + 37.8 (Empty & Loaded Work Time)(se = 8.06)

+ 1.1 (Empty & Loaded Distance)(se = 0.28)

R2 = 0.70

Bedfords - loaded trip data set N = 201

xxxii) Trip Tariff= 47.5 + 80.1 (Loaded Work Time)(se = 7.87)

+ 0.292 (Loaded Distance)(se = 0.27)

R2 = 075

Mercedes - loaded trip data set N =77

xxxiii) Trip Tariff= -3.08 + 5.43 (Loaded Distance)(se = 0.586)

+ 21.7 (Loaded Work Time)(se = 13.4)

R,= 078

TABLE 8

Proportion of Tariffs explained by Time and Distance

Per Cent Of Tariffs Multiple Squared corr.Explained By:- Regression coefficient

Constant Work R2 bemeenData Set Term Time Distance value Time & Distance

Bedfords -grouped data -2.8 46.3 56.5 0.93 0.86

Bedfords - empty& loaded trip 13.4 51.5 35.1 0.70 0.83

data

Bedfords - loadedtrip data 4.2 88.6 7.1* 0.75 0.79

Mercedes - loadedtrip data -0.1 16.5* 83.6 0.78 0.56

● Term not significant at 5 Y. probability.

13

revenue will increase by Rs 1.6 for each extra kilometretravelled and increase by Rs 35.3 for each extra hourworked.

An analysis of vehicle operating costs for Bedford trucksindicates that the key time dependent costs, namelycrew, capital costs, taxation and profit amount to aboutone third of total revenues. Although there is a margin ofuncertainty about how operating costs may be split upinto time and distance dependent elements, neverthelessthe tariff analysis probably overstates the importance oftime dependent costs.

7. CONCLUSIONS

A study of freight vehicle utilisation was carried out inPakistan. From the data elasticities relating trips madeper day to mean trip working time were calculated to lie inthe range -0.84 to -1 with a mean value of -0.92. Theanalysis suggests that in Pakistan’s conditions, journeytime savings following road investment are likely to befully translated into extra trips.

Freight vehicles were found to be used very intensively inPakistan. The sumey found that vehicle running ac-counted for 40 per cent of the time and that with loadingand unloading vehicles were in active use for over 12hours per day. Vehicles were used intensively at night;the most active running period was between 4.00 pm.and 2.00 am. Even at the quietest time of day, 06.00 hrs,37 per cent of vehicles were working. Although mostloading and unloading took place during the “normalworking day” 13 per cent of loading and 29 per cent of

unloading was started between 8.00 pm. and 6.00 am.

A statistical analysis of freight tariffs with trip times anddistances found that either time or distance could providea good explanation of tariffs. However the high degree ofcorrelation between time and distance made it difficult toidentify the separate explanatory power of either variable.The best analysis suggested that time and distance couldaccount for 45 per cent and 55 per cent of Bedford trucktariffs respectively.

8. ACKNOWLEDGEMENTS

The work described in this report forms pan of a pro-gramme of joint research between the Overseas Unit(Head: J S Yerrell) of the Transport and Road ResearchLaboratory, and the National Transport Research Centre(Head: M S Swati), Pakistan.

9. REFERENCES

De Weille J. (1966). Quantification of road user savings.World Bank Staff Occasional Papers No. 2. Baltimore.

Fleischer, G.A. (1963). Effect of Highway Improvementon Travel Time of Commercial Vehicles. HighwayResearch Record, Vol. 12.

Hine, J L and A S Chilver (1991 ). Pakistan Road FreightIndustry: An Ovewiew. Department of Transport TRRLReport RR 314. Transport and Road Research Labora-tory, Crowthorne.

Thomas, S (1983). The value of time savings in WestMalaysia: commercial vehicles. TRRL Report SR 792.Transport And Road Research Laboratory, Crowthorne.

Winfrey, R (1969). Economic analysis for highways.International Textbook Co. Scranton.

Dawson, R.F.F. and Vass P. (1974) Vehicle operatingcosts in 1973. Department of the Environment TRRLReport LR 661. Transport and Road Research Labora-tory, Crowthorne.

14

APPENDIX A:

TABLE Al

Vehicle Activity Suwey Data

Code Make Axles Truc~type Stati Survey Trips TripsNo. date days loaded empty

No. No. No.

123456789

101112131415161718192021222324252627282930313233343536373839404142434445

Total

BedfordBedfordMercedesBedfordBedfordBedfordMercedesBedfordBedfordMercedesMercedesMercedes

BedfordBedfordBedfordBedfordBedfordBedfordMercedesMercedesMercedesBedfordBedfordMercedesBedfordBedfordBedfordMercedesBedfordBedford.BedfordBedfordBedfordBedfordBedfordMercedesBedfordMercedesBedfordMercedesBedfordBedfordBedfordBedford

223222222333322222333332322332222222333232222

Simple truckSimple truckSimple truckSimple truckSimple truckSimple truckSimple truckSimple truckSimple truckSimple truckSimple truckSimple truckSimple truckSimple truckTankerSimple truckSimple truckSimple truckSimple truckSimple truckSimple truckSimple truckSimple truckTankerSimple truckSimple truckTankerSimple truckSimple truckSimple truckSimple truckTankerTankerSimple truckSimple truckSimple truckSimple truckSimple truckSimple truckTankerSimple truckSimple truckSimple truckSimple truckTanker

0310718503/07/8507110/850310718501/1 1/8517110/8521/1 1/8508/0518607109/8517/07/8523/08/8520/09/8508/1 0/8508/021861910218606/0418611/02/8622/03/8615/03/8611/02/861410318630/0318601/04/8616/04/8608109/8503/04/86o~i09ia516/04/8615/04/8608/03/8619/0318616/0518608/0518609/071a61810618606/05/8625109/8501lo51a602/09/85121061a61210218602110/85151091a6021031a626/05/86

137

1610161311

71912171819

605846677

1533

829161616.1017

59

197

128

156

151610161411

640

6464

2726

66

1078a849

10112013

665

14113

163

1855912

16682

21-23a

131512

2

405

43633

1256

106aa856a

121113665.

1462

174

1954

101,2

15591’

212478

1133

327

15

TABLE A2

Time Utilisation Data: Vehicle Activity Survey

CodeNo. Total Loaded Empty Loading Unloading Rest Repair Loaded Total Moving Moving Woting Loaded Worhngtime moving moving time time time time trips tips time time time

time timeworting time

/day /day /loaded /all all time /loadedttipa tips trips /loaded trips

tips

Hrs Hrs Hrs Hrs Hrs Hrs Hrs No, No, Hrs Hrs Hrs Hrs Hrs

123456789

1011121314151617181920212223242526272a2930313233343536373639404142434445

mean

Total

2a7,aI 5a.73a4.o210,0356.0303.0241.aI 5a,3421.93a6.53a4.a449,5447.0132.8449.3327,3422.1322.336a.1335,03a5.o364.0349.6775.5173.5690,5363.0360.2377.02ia.a392.0127,0la5.i430.a131.5304.2I 7a.o332.6120,0342.5354.5239a34a,5313.1252.5

t4657.0

91,575.497.6a8.5

107.0aa.oa9.344a

162.5127.3131.a174.0168.55a.a

137,6a2,0

ct4a.a‘ao.990.9

109.8106,576,071,2

332.346.8

1a6.7109,550.595.5a6.1

loa,631.261.a

101,532,574.44a.227.92a,7

1I 7.a133.537.763.0

142.071.a

4376,5

22.811,029.411.55.0

2a.o17.720.047.444.33a.a47.035.016.216.99.8

41 .aa.a

29.334,329.540.027.047,3Ii.a55.559.029.420.512.232.436a46.370.9ia.149,410.024.116.797,032,020.536.5

3.aal,5

1422.5

6,57.9

10.23,5

64.04a.o

6.79.a

20.a10.516.033.046.5

7.714.02a,o25.14a.a26,912.517.512.036.114.63.5

29,15.5

22.5Ia.o22.014.64,22.4

34.610.3la.7a.2

19,47,02.7

23.021.541,014.35.3

a56.o

6.35.67.04.5

49,053.511.6a.7

17.615.5I a.529,022,0

9.010.535.817,947.717.614,019.013,017,624.9

3.024.2

9,513.519.011.916.43.07.a

22.2a.9

14.65.0

16.17.3

1a,a1a.o22.329.010,27.3

158.34a.5

225.097.099.071,5

129.173.0

164.6165.3169.3153.5136.040.4

232.0169.2186.7120,6i9a,7159.5194.0216.0197.6347,4I oa,4349.7170.0244,3213.0

a6.5201.4

51.965.0

201.661.7

145.0Iol.a211,a

52.3100.0137.0126,3161.5136.7ai ,a

763,6 6759.9

2.510.114,75.0

32.014.05,42.19.1

25a10,513.039.0

0.73a,5

2.51.9

15.54,75.0

la,57,00.09.10.0

45.49.50.0

11.00.0

la,60,0I .a0.00.02.14.a

33,aa,o6.3

11.011.517,56.14.a

47a,5

0.500.61o,3a0.461.az2.060.600.910.570.430.500.430.430.72o.4a0.730,631.49o.a50,430.370.330.960.340.410.560.201,200.320.550.550.190.26o,a91.100.630.271.510.400.210.541,301.030.920.19

0.67

o,a31.060.75o,ao2,023.011.091.a21.14o.ao1.000.85o.a61.63o,ao1,321.312.311,70o.a60.750.661.920.530.691,150.462.470,640,991.16o,3a0,521.732.011,340,403.03o.ao0.491.022.101.791.15o.4a

1,21

15.3I a.a16.322.1

4.03.4

11.67.5

16.3I a.2I a.521 ,a21,114.715,3a.2

13.54.07.0

18.3I 7.a15,25.1

30.215.611.736.5

2.a19.117.212.131.230.9

6,35.49.3

24.11.3

14,339.316.72.94.2

11.a35.9

15,3

11.412.310.614,33.73.17.95.4

10.513.210.713.a12.78.3

10.35.1a.32.94.6

12.011.311,63.5

22.311,7

7.324.1

2.211.610.97.4

34.027.0

5.a4.67,3

19.41.2

11.330.711.02.a3,a9,7

30.7

11.2

12.714.312,015,47.55.79.76.9

12.415.212,a17.717.010,211.9a.6

10,16,06.3

14,214.414.15.4

24,613.09.0

26.23,1

15.314.7

9.137.629.6

7.46,39,2

23a2.1

14.933.7I 3.a4,96.5

11.433,2

13.6

17.422,319.224.1

a.17.3

14.910.520.121.920,a29.529.6la,918.014,617.4a,9

10.422,723,a20.2

8.933,a17.a15.041.5

4,a26.524.015,53a.336.0

9.9a.6

13.530.7

3,021.546.421.a

6.38.9

13.942,2

19.a

21.225,024.127.0

a.3a.4

17.913.924.82a,225a35.434.022.919.915.621,2

9.312.728.42a,a2a.210.93a, I21.71a.5al,2

6.430.626.419,175.159.114,311.a19.635.7

4,229a7a.725a

7,a11.314.2a3.o

26.3

16Printed in the United Kingdom for HMSO

DdK50120 11/91 C5 G2516 10170