Future mobility in India: Implications for energy demand and CO2 emission

ARTICLE IN PRESS

0967-070X/$ - s

doi:10.1016/j.tr

�Tel.: +91 51

fax: +91512 25

E-mail addr1Growth rate

rather than sim

paper.

Transport Policy 13 (2006) 398–412

www.elsevier.com/locate/tranpol

Future mobility in India: Implications for energydemand and CO2 emission

Sanjay Kumar Singh�

Department of Humanities and Social Sciences, Indian Institute of Technology Kanpur, Kanpur 208016, Uttar Pradesh, India

Available online 23 May 2006

Abstract

The aim of this study is threefold. First, it aims to provide a reliable data set of land-based passenger traffic volumes in India from

1950–1951 to 2000–2001 for the five major motorized modes of transport—two-wheelers, cars, auto-rickshaws, buses, and railways.

Second, based on this data set, it aims to estimate the long-term trends in motorized traffic volume and modal split up to the year

2020–2021. Third, based on the projected values of aggregate traffic volume and modal split, this study aims to estimate the level and

growth of energy demand and CO2 emission from the passenger transport sector in India. It is found that the motorized traffic volume in

India will very nearly touch the mark of 13 000 billion passenger-kilometers in 2020–2021, out of which 91.7% will be provided by the

roads and the rest by railways. If there is no reduction in modal energy and CO2 intensities, energy demand is projected to increase from

1060.8 peta joules in 2000–2001 to 5584.4 peta joules in 2020–2021 and CO2 emission will increase from 19.80 to 93.25 million metric tons

of carbon equivalent during the same period.

r 2006 Elsevier Ltd. All rights reserved.

Keywords: Motorized mobility; Energy demand; CO2 emission

1. Introduction

Passenger mobility in India relies heavily on rail androad. Passenger travel by air and water is negligible incomparison to rail and road. On an average, an Indiantraveled 285 km in a year during 1950–1951, out of which185 km was by rail and 100 km by road. In a span of fivedecades their annual travel figure jumped to 3470—449 kmby rail and 3021 km by road. From 1950–1951 to2000–2001, passenger travel per capita (measured in termsof passenger-kilometers per capita; Pkm/cap) by road andrail taken together increased at the rate of 5.13% perannum against the per capita Gross Domestic Product(GDP) growth of around 2.23% per annum.1 It issurprising to know that between 1970–1971 and2000–2001, in light of an 88% population growth, absolute

ee front matter r 2006 Elsevier Ltd. All rights reserved.

anpol.2006.03.001

2 259 7501 (O); +91 512 259 8332 (R);

9 7510.

ess: [email protected].

is calculated as compound annual growth rate (CAGR)

ple annual growth rate. This is followed throughout this

motorized mobility (measured in terms of passenger-kilometers; Pkm) has risen by 888% in India. The growthof road-based passenger mobility (Pkm) during the last 50years has been around 9.17% per year against thecorresponding compound annual growth rate of 3.93%by rail. Consequently, the road share in passenger mobilityincreased from 35% in 1950–1951 to 87% in 2000–2001.Rapid increase in travel demand and increasing reliance

on road transport has serious implications for variousissues related to energy and environment. Already, trans-port sector is the major cause of air pollution in urbanareas. It contributes significantly to major environmentalchallenges both at local as well as global levels. Transportsector is a major consumer of energy particularlypetroleum fuels. Designing and implementation of effectivestrategies to reduce transport sector’s impact on environ-ment and energy demand require a reliable projection ofmotorized mobility for the forthcoming years. This paperlays the foundation for the same.The main aim of this study is to forecast the level of

traffic mobility, energy demand, and carbon dioxide (CO2)emission from the land-based passenger transportation in

www.elsevier.com/locate/tranpol

dx.doi.org/10.1016/j.tranpol.2006.03.001

mailto:[email protected]

ARTICLE IN PRESS

0

500

1000

1500

2000

2500

3000

3500

PKm

/cap

1950-51 1960-61Year

1970-71 1980-81 1990-91 2000-01

Auto-rickshawsCarsTrainsBuses Two-wheelers

Fig. 1. Mode-wise motorized mobility per capita in India.

3Limited time series data for traffic volume by mode of transport has

been compiled and published by the Centre for Monitoring Indian

Economy (CMIE), Mumbai, India and TERI Energy Data Directory &

Yearbook (TEDDY), New Delhi, India. For example, CMIE provides

estimates of road-based Pkm from 1950–1951 to 1988–1989 in 11 data

points. It does not present the complete series from 1950–1951 to

2000–2001 for which motor vehicle population data is readily available.

S.K. Singh / Transport Policy 13 (2006) 398–412 399

India up to the year 2020–2021. There are typically twoways to estimate future traffic mobility. The first approachis generally based on independent projections of trafficvolume per mode of transport over time. Typically, eachmodal projection is built on a different method, and thetotal traffic volume becomes simply an aggregate of theindependent estimates for the various modes. The secondapproach is based on projection of motorized mobility (theaggregate traffic volume of two-wheelers, cars,2 auto-rickshaws, buses, and trains) in a first step, and the relatedmodal split is computed afterwards. The second approachis a better one for developing long-term scenarios since ittakes into account the competition between modes(Schafer, 1998). Projections based on the first approachoffer some glimpses into the future, often in the form ofmultiple scenarios that bound a large range of possibilities,but offer little guidance about the most likely ones (Schaferand Victor, 2000). This paper follows the second approachto allow formulation of aggregate and long-term scenarios.The aim is to project total mobility and the share of eachmode far away in the future. This aggregate approach willbe useful when we analyze the transport systems’ impact onthe environment and energy demand. The forecastingmodel used in this paper is mainly built on only twoexplanatory variables: GDP and population. Annual datafrom 1950–1951 to 2000–2001 are used to forecast thefuture mobility in India. The statistical program LIMDEPVersion 8.0 is used for the required regression analysis.

Based on the projected values of aggregate traffic volumeand modal split, the paper estimates the level and growth ofenergy demand and CO2 emission from passenger trans-port sector in India using a scenario approach. Twoscenarios, ‘business as usual’ and ‘efficiency gain’, arediscussed in the paper. In the business as usual scenario,both energy and CO2 intensities of all transport modes areassumed to remain at 2000–2001 levels, whereas in theefficiency gain scenario intensities are assumed to reduce atthe rate of 1% per year from 2000–2001 onwards. It isfound that both energy demand and CO2 emission willincrease rapidly from 2000–2001 to 2020–2021 due to risein travel demand as well as greater reliance on automobiles.Even when we assume a reduction of 1% per year in energyand CO2 intensity of all modes of transport, energydemand and CO2 emission are projected to increase atthe rate of 7.5% and 7% per year, respectively between2000–2001 and 2020–2021.

The paper is organized as follow. Section 2 presents theaggregate as well as mode-wise traffic volume data from1950–1951 to 2000–2001. Section 3 deals with the model,model estimation, projection of per capita as well asabsolute mobility, and estimation of modal split changesfrom 2000–2001 to 2020–2021. Section 4 describes theenergy demand and CO2 emission scenarios. The paper’sfindings are summarized in Section 5.

2Cars include jeeps and taxis. This will be followed throughout the

paper.

2. Passenger mobility in India during the past five decades

The paper presents for the first time, annual time seriesdata of land-based traffic volume in terms of Pkm from1950–1951 to 2000–2001.3 These data account for the fivemajor modes of transport, namely cars, two-wheelers,auto-rickshaws, buses and railways. The data sources andestimation methods are summarized in the Appendix A.Passenger traffic volume in India increased from 102

billion passenger-kilometers (BPkm) in 1950–1951 to3536BPkm in 2000–2001 due to a 12.18-fold increase inannual distance traveled by the people (from 285 km in1950–1951 to 3470 km in 2000–2001), and a 2.84-fold risein population (from 359 million in 1950–1951 to 1019million in 2000–2001). Analysis of per capita mobility datashows that the average annual distance traveled by thepeople triples in every two decades (Fig. 1). Although alarge proportion of the mobility need is still catered to bythe buses, there is increasing reliance on automobiles inrecent years. For example, during 1990s per capita mobilityby two-wheelers, auto-rickshaws and cars increased by124%, 130% and 90%, respectively, against the corre-sponding increase of 60% for buses and a meager 27% forrailways. Thus, mobility share of private- and para-transitmodes4 increased from 16.2% in 1990–1991—21.2% in2000–2001, whereas share of both buses and railwaysdeclined during the same period (Fig. 2).Although railways had played a dominant role in

providing passenger mobility from the second half of the19th century to the early 1950s, the rail-system has beencontinuously loosing its ground from the late 1950s.

TEDDY provides similar series that seems to be based on CMIE

estimates.4Private- and para-transit modes include cars (including jeeps and taxis),

two-wheelers, and auto-rickshaws.

ARTICLE IN PRESS

12.9

0

10

20

30

40

50

60

70

80

90

100

Mob

ility

sha

re (%

)

1950-51 1960-61 1970-71

Year

Private-and para-transitmodes

Buses Railways

1990-911980-81 2000-01

65.1

45.333.0

26.316.6

21.216.2

10.0

9.8

7.9

46.8

6.6

28.4

57.263.7 67.2 65.9

Fig. 2. Share of different modes in providing passenger mobility in India.

0

500

1000

1500

2000

2500

3000

3500

4000

3500 5500 6500 7500 8500 10500 11500

PKm

/cap

GDP/cap4500 9500 12500

Fig. 3. Motorized mobility (car, two-wheeler, auto-rickshaw, bus, and

rail) per capita vs. GDP per capita in India between 1950–1951 and

2000–2001.

5Per capita GDP data, presented in Fig. 3, is taken from National

Accounts Statistics of India: 1950–1951 to 2002–2003 published by EPW

Research Foundation, Mumbai, India. GDP figures are at factor cost at

1993–1994 prices.

S.K. Singh / Transport Policy 13 (2006) 398–412400

Between 1950–1951 and 2000–2001, rail-based passengermobility increased at the rate of 3.93% per annum against thecorresponding annual growth rate of 9.17% for roads. Thiswas largely due to tremendous increase in the motor vehiclepopulation in the last three decades or so. Currently, motorvehicle population in India is growing at a rate of around10% per annum. In 1990–1991 there were about 21 millionvehicles in the country. After 10 years in 2000–2001, thisnumber increased by more than 2.5 fold to 55 million.

The vehicle population in India is growing faster in thecategory of two-wheelers and three-wheelers (auto-rick-shaws). During 1950–1951 to 2000–2001, the two-wheelerand three-wheeler population grew at an average annualrate of 15.6% and 14.9%, respectively. The growth inpopulation of car was relatively modest. In the last fivedecades, car population increased at the rate of 7.9% peryear. The corresponding figure for buses was 5.7%. If wecompare the growth rate in recent years, say from1995–1996 to 2000–2001, we find that both cars as wellas two-wheelers were growing at the rate of around 10.5%per year, and three-wheelers (auto-rickshaws) increased atthe rate of 11.3% per annum. During the same period buspopulation increased only at the rate of 4.5% per annum,whereas total motor vehicle population increased at therate of 10.2% per annum. As income of the peopleincreases demand for personalized modes of transport isexpected to increase more rapidly. The modal splitindicates that in 1970–1971, about 31% of the totalvehicles were two-wheelers, which increased to 70% in aspan of just three decades. As a result, the percentage shareof buses in total motor vehicle population declined from4.9% in 1970–1971 to 1.0% in 2000–2001.

3. Passenger mobility in India during the next two decades

3.1. The model

The demand for mobility depends on various socio-economic factors such as age distribution and household

composition, employment, educational level, supply ofpublic transport services, infrastructure availability, gov-ernment policy towards automobiles and transport, pricesof different transport services, fuel and vehicle prices,income of the people, etc. At the national level, therelationship between per capita mobility and factorsinfluencing the same can be written as

Pkm

cap¼ f ðX Þ, (1)

where Pkm=cap is passenger-kilometers per capita (repre-senting per capita mobility) and X is a vector of variablesdetermining the level of per capita mobility.Using Eq. (1), it is possible to estimate the future level of

mobility if data for each of the variables on the right-handside are available. However, time series data for many ofthese variables are not readily available for India. In thissituation, it is important to find out the key determinantsof mobility for which time series data are available. Schafer(1998), Dargay and Gately (1999), Schafer and Victor(2000), Preston (2001) and many others have shown thatthere is a close relationship between income and demandfor mobility. The strong relationship between income andmobility is found for both cross-country as well as time-series data. Fig. 3, which presents the relationship betweenper capita mobility and per capita GDP for India,reiterates the same.5 Therefore, in this paper, per capitaGDP will be used as the main explanatory variable toproject future passenger mobility in India. Assuming thattime captures the effects of the omitted variables, Eq. (1)can now be approximated by

Pkm

cap¼ f

GDP

cap; time

� �. (2)

In a practical forecasting problem, the statistical natureof the data-generating process is unknown and the

ARTICLE IN PRESS

7Although it is possible to estimate the saturation level, there is no

guarantee that the final estimate of the saturation level, a, is close to the

global optimum (Heij et al., 2004). Therefore, to estimate the models, it is

essential to get a reliable estimate of the saturation level of per capita

mobility.8Schafer (1998) shows that the time spent on travel per person per day

are virtually unchanged with respect to per capita income across the

countries. Although the reason for travel time budget stability is not very

clear, Marchetti (1994) argued that a travel time budget of around one

hour per capita per day reflects a basic human instinct. He argued that

perhaps security of the home and family, the most durable unit of human


forecaster’s task is to select a model that best approximatesthe ‘real-life’ data generating process. For a time series likethe per capita passenger mobility, it is conceivable that theseries converges to a maximum as income reaches a certainlevel. If we plot level of passenger mobility per capitaagainst GDP per capita, the graph is expected to look likesome sort of a S-shaped curve. Passenger mobility isexpected to increase slowly at the lowest income levels, andthen more rapidly as income rises, finally slow down assaturation is approached. There are a number of differentfunctional forms that can describe such a process; forexample, the logistic, Gompertz, logarithmic logistic, logreciprocal and cumulative normal functions (Dargay andGately, 1999).6 The logistic and Gompertz functions arethe two most widely used functional forms to describe aprocess represented by a S-shaped curve. This paper alsouses these two functions to model and forecast passengermobility in India.

The logistic model can be written as

Pkm

cap

� �t

¼a

1þ g expð�bðGDP=capÞt � lðtimeÞtÞþ et, (3)

where a is the saturation level and et is an error term atperiod t. All the parameters a, b, g and l are positive.

Similarly, the Gompertz model can be written as

Pkm

cap

� �t

¼ a exp �g exp �bGDP

cap

� �t

� lðtimeÞt

� �� þ et,

(4)

where a is the saturation level and et is an error term atperiod t. All the parameters a, b, g and l are positive.Parameters g, b and l define the shape or curvature of thefunction.

Models (3) and (4) need to be transformed into a linearform in order to estimate them using Ordinary LeastSquares (OLS) method. The logistic model (3) can betransformed in a linear form as follows:

lna

ðPkm=capÞt� 1

� �¼ b0 þ b1

GDP

cap

� �t

þ b2ðtimeÞt þ et,

(5)

where a is the saturation level and et is an error term atperiod t. Parameter b0 is positive and both b1 and b2 arenegative.

Similarly, the Gompertz model (4) can be transformed as

ln lna

ðPkm=capÞt

� �� ¼ b0 þ b1

GDP

cap

� �t

þ b2ðtimeÞt þ et,

(6)

6An overview of such functional forms is given in Meade and Islam

(1998); see also Meade and Islam (1995), Bewley and Fiebig (1988),

Franses (2002), and Mohamed and Bodger (2005). Application of such

functional forms to project the traffic mobility in India can be seen from

Ramanathan (1998), Ramanathan and Parikh (1999), and Singh (2000).

These studies in the Indian context have used time as the only explanatory

variable to project the future mobility in India.

where a is the saturation level and et is an error term atperiod t. Parameter b0 is positive and both b1 and b2 arenegative.Models (5) and (6) can be estimated by OLS provided we

know the saturation level a.7 It is possible to get a reliableestimate of the saturation level by making a reasonableassumption about the time spent on travel per person perday (i.e., travel time budget) and average speed of vehicles.For example, if all demands are met at an average speed of30 km/h and the travel time budget is fixed at 1.1 h/capita/day, the total annual distance travel would be around12 000 km/capita.8 So, if we assume that all demands aremet at a maximum possible (average) speed of 30 km/h,mobility per capita of 12 000 will be the saturation level.Considering the socio-economic characteristics of India(such as population density, rapid increase in telephonedensity, expected boom in information technology, greaterreliance on public transport, high fuel prices, etc.) and thelevel as well as growth in per capita mobility in thedeveloped world, 12 000 Pkm per capita appears to be theappropriate saturation level for India.According to the data published by the ECMT-Eurostat,

during the year 2000, land-based Pkm/cap in the EuropeanUnion (EU) countries were around 12 000. Although land-based mobility trends differ from country to country,mobility has been relatively stable since 1990 in the UnitedKingdom and Germany. The saturation level for these twocountries may be somewhere around 12 000 Pkm/cap.Analysis of the data published by the Bureau ofTransportation Statistics, USA shows that during the year2000, Pkm/cap by rail and road taken together in theUnited States of America was around 24 000. The US isalso approaching its saturation level since growth in percapita mobility is slowing down in recent years. Thedifference in the saturation level between the US and EU isprimarily due to the difference in the socio-economicfactors. Public transport is far more important forpassenger travel in the EU than in the US. The muchgreater car travel per capita in the US is largely due to farmore extensive motorway networks, much cheaper petrol,

organization, limits exposure to the risk of travel. Also, traveling is

naturally limited by other activities such as sleep, leisure, and work. Even

when time spent on any of these activities changes, there is evidence that

the travel time budget remains constant (Marchetti, 1994). Time-use and

travel surveys from numerous cities and countries throughout the world

suggest that travel time budget is approximately 1.1 h per person per day

(Schafer and Victor, 2000). It should be noted that the stability of average

travel time budget holds only for travel by all modes.

ARTICLE IN PRESS

Table 1

Parameter estimates of the logistic and Gompertz models (with t-statistic in parentheses)

Model Estimate

Saturation level, a ¼ 12 000

Logistic (5) b0 ¼ 4:29110ð127:9Þ, b1 ¼ �0:08260ð8:2Þ, b2 ¼ �0:04948ð34:3Þ; R2 ¼ 0:996; MAPE ¼ 3.93

Gompertz (6) b0 ¼ 1:71227ð101:1Þ, b1 ¼ �0:07661ð15:0Þ, b2 ¼ �0:01263ð17:3Þ; R2 ¼ 0:993; MAPE ¼ 6.08







Logistic (sat.: 16000)

Logistic (sat.: 12000)

Gompertz (sat.: 20000)



13000

12000

11000

10000

9000

8000

7000

6000

5000

4000

30002000-01 2005-06 2010-11 2015-16 2020-21

Year

PKm

/cap

Fig. 4. Assumptions and projections of land-based per capita mobility in

India.


motor vehicle, and roadway taxes and fees, and the longertrip distances both within and between cities. Sincetransport availability characteristics in India is more closeto the EU, we expect that the saturation level in Indiawould be around 12 000PKm/cap. This is close to what hasbeen estimated by Singh (2000). In his paper, Singh (2000)estimated that the saturation level for land-based passengermobility in India will be around 20 000BPkm. Assumingthat the Indian population will saturate at 1.65 billion,saturation level for mobility per capita would be around12000 km. However, both the logistic and Gompertzmodels will be estimated for three different saturationlevels, 12 000, 16 000 and 20 000 Pkm/cap to illustrate thesealternative paths of per capita passenger mobility in India.

3.2. Model estimation

The logistic model (5) and the Gompertz model (6) areestimated using the econometric software LIMDEP Ver-sion 8.0. Both the models have been estimated for threedifferent saturation levels: 12 000, 16 000 and 20 000 Pkm/cap. Annual data of passenger mobility per capita (Pkm/cap) and GDP/capita (Rs. in thousand at 1993–1994prices) from 1950–1951 to 2000–2001 are used for theestimation of the models. The variable time takes thefollowing values: 1 for 1950–1951, 2 for 1951–1952, 3 for1952–53,y, and so on.

Table 1 reports the estimation results. According to theR2 values, the models fit the data very well. We alsocompare the predicted values with the actual values ofPkm/cap over the sample period and found the same. Themean absolute percentage error (MAPE)9 presented inTable 1, is in the range of 3.86–3.93 for the logistic modelsand 5.77–6.08 for the Gompertz models. All the estimatedparameters have the expected signs and most are highlysignificant. To project the future per capita mobility up tothe year 2020–2021, we have to make reasonable assump-

9The MAPE is commonly used in quantitative forecasting methods

because it produces a measure of relative overall fit. The absolute values of

all the percentage errors are summed up and the average is computed.

tions about the per capita GDP growth rate. Between1993–1994 and 2003–2004 (the latest year for which percapita GDP figure is available), per capita GDP in Indiaincreased at the rate of around 4.3% per annum. Assumingthat GDP/capita will increase at the same rate up to2020–2021, passenger mobility per capita has beenprojected for the future. Fig. 4 presents the future percapita mobility up to 2020–2021 for different saturationlevels. Since according to MAPE the logistic model is abetter model than the Gompertz, and as discussed in theSection 3.1, saturation level is expected to be around12 000 Pkm/cap, the next stage of the discussion will bebased on the estimated logistic model with saturation levelat 12 000 Pkm /cap.10

3.3. Projection of per capita as well as absolute mobility up

to 2020–2021

On the basis of the estimated logistic model for12 000 Pkm/cap saturation level and assumptions concern-ing population and GDP, projections of per capita as wellas absolute mobility up to 2020–2021 are obtained. As

10One should note that the MAPE values for different saturation levels

under the logistic model are very close to each other, hence, it is difficult to

choose between them on the basis of their MAPE values.

ARTICLE IN PRESS

9770

0

15000

12000

9000

6000

3000

2005-06 2010-11

Year

2015-16 2020-21

PKm/cap

BPKm

55555045

6620

78178284

10419

12987

Fig. 5. Future mobility trends in India.


stated in the previous section, per capita GDP is assumedto grow at the rate of 4.3% per annum up to 2020–2021.Based on World Population Prospects: The 2004 RevisionPopulation Database published by the United NationsPopulation Division, population of India is assumed togrow at the rate of 1.56% per annum from 2000–2001 to2005–2006, 1.41% per annum from 2005–2006to 2010–2011, 1.27% per annum from 2010–2011 to2015–2016, and 1.11% per annum from 2015–2016 to2020–2021.

Fig. 5 presents the future mobility trends up to2020–2021. It shows that in 2020–2021, average Indianswill travel about thrice as many kilometers as they traveledin 2000–2001. Absolute passenger mobility in India at theend of 2020–2021 will virtually touch the mark of13 000BPkm. On an average, per capita mobility andabsolute traffic volume in India are expected to increase atthe rate of 5.31% and 6.72% per annum, respectively in thenext two decades. However, growth in both mobility percapita and absolute traffic volume are expected to be higherduring 2000–2011 than 2010–2021 (Table 2).

11This is somewhat similar to the findings of Schafer and Victor (2000)

for high population density countries such as India at per capita mobility

level of around 10 000 km.12In fact, latest (provisional) figure from the Ministry of Road

Transport and Highways, GOI, New Delhi reveals that the growth rate

of these vehicles is same from 2000–2001 to 2003–2004 as well.

3.4. Modal split changes

As shown in Fig. 3, increase in per capita GDP results inrise in per capita motorized mobility. Assuming that anaverage person spends some fixed time (approximately1.1 h per day) on travel, mean travel speed has to increasewith the increase in per capita mobility. Because differenttransport modes operate with different ranges of speed,increase in mobility changes the modal split towardsflexible and faster transport modes. Thus, as per capitamobility and per capita GDP increase, traffic share ofpublic transport modes such as buses and trains decreaseand share of private- and para-transit modes increase.Fig. 6 shows the relationship between mobility and share oflow speed public transport. The share of public transportmodes went down from 93.4% in 1950–1951—78.8% in2000–2001 in a linear fashion as per capita mobilityincreased from 285 km in 1950–1951—3470Km in

2000–2001. Similar relationship is found between share ofpublic transport modes and per capita GDP (Fig. 7). Bothprovide somewhat identical future values for the share ofpublic transport modes up to 2020–2021.It is estimated that the share of low-speed public

transport (buses and trains) in India will be around 52%during 2020–2021 (Fig. 8).11 Since the share of buses andtrains in the public transport modes are virtually un-changed from 1993 to 1994 onwards at around 84% and16%, respectively, we assume that the same pattern will befollowed up to the year 2020–2021. Based on thisassumption, the share of buses and trains in meeting thepassenger travel demand in future has been projected. It isestimated that 43.6% of traffic mobility in India in2020–2021 will be provided by the buses and 8.3% by thetrains (Fig. 8).Similarly, we projected the share of high-speed private-

and para-transit modes (cars, two-wheelers, and auto-rickshaws) up to 2020–2021. Fig. 9 presents the same alongwith the share of individual modes. It is estimated that thecombined share of these modes in India would be around48% during the year 2020–2021. Since, in this case also, theshare of cars, two-wheelers, and auto-rickshaws within theprivate- and para-transit modes are virtually unchangedfrom 1993 to 1994 onwards at 37%, 49%, and 14%,respectively, we assume that the same pattern will befollowed till 2020–2021. This assumption is made becausegrowth rate of car, two-wheeler, and auto-rickshawpopulation is more or less same from the last 7 years orso.12 This is not expected to change in near future sincemost of the car owner in India prefers to own a two-wheeler as well. Two-wheelers are perceived to be useful forshort distance travel and easy to drive even in congestedstreets and roads. Also, fuel cost for a two-wheeler in Indiais hardly 25% of that of a car. Therefore, as income of thepeople increases, the demand for car and two-wheeler islikely to increase more or less at the same rate. It is unlikelyto have crowding out effect of increased car ownership athigher level of income on demand for two-wheeler. Oneshould note that per capita vehicle ownership rate in Indiais still among the lowest in the world and it can easilysustain high growth path for these categories of vehicles forthe long period. Even if we assume that growth in carownership may be higher than the two-wheelers at higherlevel of income say from 2010–2011 to 2020–2021, theassumption that the mobility share of cars, two-wheelers,and auto-rickshaws within the private- and para-transitmodes will be unchanged is not very strong since theaverage occupancy of car is expected to decrease with theincrease in car ownership. Decrease in average occupancyof two-wheeler and auto-rickshaw is also expected at

ARTICLE IN PRESS

Table 2

Level as well as growth of land-based passenger mobility in India from 1950–1951 to 2020–2021

Per capita mobility

(Pkm/cap)

CAGR in per capita mobility

(since the previous period) (%)

Absolute mobility

(BPkm)

CAGR in absolute mobility

(since the previous period) (%)

1950–1951 285 — 102 —

1960–1961 395 3.3 171 5.3

1970–1971 661 5.3 358 7.6

1980–1981 1169 5.9 794 8.3

1990–1991 2125 6.2 1783 8.4

2000–2001 3470 5.0 3536 7.1

2010–2011 6620 6.7 7817 8.3

2020–2021 9770 4.0 12 987 5.2

50

60

70

80

90

100

0 1000 1500 2000 2500 3000 3500 4000

Tra

ffic

sha

re o

f pub

lic tr

ansp

ort

mod

es (%

)

Share =93.42 - 0.0043(PKm/capita); R2 = 0.98

PKm/capita

500

Fig. 6. Traffic share of public transport modes (buses and trains) at

different level of per capita mobility between 1950–1951 and 2000–2001.

50

60

70

80

90

100

0 2 8

Tra

ffic

sha

re o

f pub

lic tr

ansp

ort

mod

es (%

)

Share =100.81 - 1.903(GDP/capita); R2 =0.98

GDP/capita (Rs. in thousand at 1993-94 prices)

1412104 6

Fig. 7. Traffic share of public transport modes (buses and trains) at

different level of income between 1950–1951 and 2000–2001.

0

10

20

30

40

50

60

70

80

90

100

Shar

e (%

)

Aggregate share of public transport modes

Share of buses

Share of trains

2000-01 2005-06

Year

2010-11 2015-16 2020-21

10.511.612.9

65.972.2

60.765.5

55.0 58.3

49.051.9

43.6

8.3

78.8

9.3

Fig. 8. Share of public transport modes (buses and trains) during the next

two decades.

0

25

50

Shar

e (%

)

Aggregate share of private- and para-transit modesShare of two-wheelersShare of carsShare of auto-rickshaws

75

2000-01 2005-06

Year

2010-11 2015-16 2020-21

23.6

17.8

6.7

48.1

41.7

20.415.4

5.8

16.9

12.8

4.8

34.5

13.610.3

3.9

27.821.2

10.38.0

2.9

Fig. 9. Share of private- and para-transit modes (cars, two-wheelers, and

auto-rickshaws) during the next two decades.


higher level of ownership but at lesser rate than that of carwhen car demand exceeds demand for two-wheeler andauto-rickshaw.13 Therefore, assumption that mobility

13One should note that modal share of a particular category of vehicle

depends on three factors: no. of vehicles, annual utilization (in km), and

average occupancy. For example, in this case, modal share of

car ¼ (mobility provided by cars)/(mobility provided by private- and

para-transit vehicles) where mobility provided by car ¼ [car popula-

tion� annual utilization of car (in km)� avg. occupancy of car] and

mobility provided by private- and para-transit vehicles ¼ [{car popula-

tion� annual utilization of car (in km)� avg. occupancy of car}+{two-

wheeler population� annual utilization of two-wheeler (in km)� avg.

occupancy of two-wheeler}+{auto-rickshaw population� annual utiliza-

tion of auto-rickshaw (in km)� avg. occupancy of auto-rickshaw}]. It is

share of cars, two-wheelers, and auto-rickshaws withinthe private- and para-transit modes will be unchanged from2000–2001 to 2020–2021 appears to be plausible. Based onthis assumption, the share of individual high-speed modesin meeting the passenger travel demand in future has beenprojected (Fig. 9). It is estimated that, during the year

(footnote continued)

easy to see that even if car population growth is different than that of two-

wheeler and auto-rickshaw say from 2010 to 2011 onwards, modal share

of car may be stable since it depends on other factors as well.

ARTICLE IN PRESS

Table 3

Modal share and energy intensities in 2000–01 and projections for 2020–2021

Mode of transport 2000–2001 2020–2021 (business as usual scenario) 2020–2021 (efficiency gain scenario)

Modal share (%) Energy intensity

(MJ/Pkm)


(MJ/Pkm)


(MJ/Pkm)

Car 8.00 0.94 17.80 0.94 17.80 0.77

Two-wheeler 10.30 0.53 23.60 0.53 23.60 0.43

Auto-rickshaw 2.90 0.58 6.70 0.58 6.70 0.47

Bus 65.90 0.19 43.60 0.19 43.60 0.16

Rail (diesel) 7.74 0.24 1.66 0.24 1.66 0.20

Rail (electricity) 5.16 0.12 6.64 0.12 6.64 0.10

Rail (diesel and

electricity)

12.90 0.19 8.30 0.14 8.30 0.12

Total (weighted

average)

100.00 0.30 100.00 0.43 100.00 0.35


2020–2021, 17.8% of the land-based traffic mobility inIndia will be provided by the cars, 23.6% by the two-wheelers, and 6.7% by the auto-rickshaws. Analysis ofchanges in modal split reveals that, from 2000–2001 to2020–2021 the traffic share of buses and trains will declinewhereas the share of cars, two-wheelers and auto-rickshawswill increase substantially. The rapid growth in mobilityand more energy intensive modal split are expected to haveadverse impact on energy demand and CO2 emission.

142000–2001 energy intensities by mode (in mega joules per Pkm; MJ/

Pkm) have been estimated by the author. The vehicle mileage for car, two-

wheeler, auto-rickshaw, and bus has been assumed to be 12, 45, 35, and

4.5 km per litre of fuel (petrol/diesel), respectively. The average occupancy

has been assumed to be 3.18 for cars, 1.5 for two-wheeler, 1.76 for auto-

rickshaw, and 41.6 for bus. The energy equivalent value for 1 litre of fuel

(petrol/diesel) is taken as 35.85MJ. The estimates for energy intensities in

rail are taken from Ramanathan and Parikh (1999).

4. Energy demand and CO2 emission from passenger

transport sector

4.1. Energy demand

India’s energy consumption has grown rapidly over thelast two decades or so. Presently, it accounts for around3.3% of the world total. Coal (50.9%) and petroleum(34.4%) together account for more than 85% of India’senergy consumption, while natural gas (6.5%) and hydro-electricity (6.3%) account for much of the remainder. Thetransport sector in India is a major energy-consumingsector, particularly of petroleum products. About half ofthe total petroleum products consumption in the countrygoes into the transport sector in the form of HSD (high-speed diesel) and gasoline. India consumes close to 3% ofworld oil supply and imports around 65% of its require-ment. In 2000–2001, India consumed 102.5 million tonnesof petroleum products, out of which HSD alone wasaround 39 million tones (CMIE Energy database). Thetransport sector consumes more than 85% of the totalHSD consumed in the country.

In India, land-based passenger transport accounts for asignificant proportion of energy consumed in the transportsector. The aim of this sub-section is to estimate the presentas well as future energy demand in the passenger transportsector in the country. For this, we have to estimate themodal energy intensities of all major transport modes.

Table 3 summarizes the estimated 2000–2001 modal energyintensities and their projected values for 2020–2021.14

For 2020–2021, two scenarios are presented for energyintensities in the sector. First, in the business as usualscenario, the 2020–2021 energy intensities of all groundtransport modes are assumed to remain at 2000–2001levels. However, rail energy intensity will be differentbecause we assume that the electrified routes will cater tothe need of 80% of the rail passenger traffic in 2020–2021rather than the current level of 40%. Although, business asusual scenario may appear optimistic, historically fuelefficiency of automobiles in many countries includingOECD countries has not changed significantly during thepast 20 years or so. In addition, it is quite likely thataverage occupancy will decline at higher level of incomeand mobility. Therefore, even if automobile fuel efficiencyimproves, energy intensity of individual modes may remainunchanged.Second, in the efficiency gain scenario, the energy

intensities of all ground transport modes are assumed todecline at the rate of 1% per year up to 2020–2021 i.e.,around 18% reduction in energy intensity of all modes in aspan of 20 years. Efficiency gain scenario is also plausibleprovided government plays active role to achieve the same.Recently, as part of its economic reform programme, thegovernment has tried to change its role in the energy andenvironment sector. Dismantling price control, reductionin subsidy, opening of the energy sector to private andforeign investment, removing restriction on energy trade,and setting up the independent regulatory commission inenergy sector are examples of its changed role in energy

ARTICLE IN PRESS

15Sum of modal energy demand may not be equal to the total of the

sector since energy intensity figures have been rounded off to 2 decimal

places.16In the efficiency gain scenario also, sum of modal energy demand may

not be equal to the total of the sector since energy intensity figures have

been rounded off to 2 decimal places.


sector. The Ministry of Environment and Forests, Govern-ment of India, the nodal agency in the administrativestructure of the central government, for the planning,promotion, co-ordination and overseeing the implementa-tion of environmental and forestry programmes, has takennumber of measures to tackle the environmental problemsin the country. As far as environment consciousness inroad transport sector is concerned, government hastightened the vehicle emission standards, fitting of catalyticconverters in new petrol driven passenger cars aremandatory now, unleaded petrol has been made availablethroughout the country, and Bharat-II norms (akin toEuro-II norms) have been extended to the entire countrywith effect from April, 2005. In addition to these, CNGand LPG are permitted to be used as auto fuels. Alternativefuels like di-methyl ether, bio-diesel, hydrogen, electric andfuel cell vehicles, etc. are at various stages of experimenta-tion. Therefore, it is expected that the government will playan active role in promoting energy efficient technology andtake measures to improve fuel efficiency.

Measures such as improvement in traffic flow, removalof encroachment on roads, restraint on parking incongested areas, infrastructure improvement, adequateinspection and maintenance program for in-use vehicles,increasing the price of petrol and diesel, tax concession tofuel efficient vehicle, etc. have huge potential to improvefuel efficiency. In general, Indian roads face severe trafficcongestion problem. Growing traffic and limited roadspace have reduced peak-hour speeds to 5–10 km/h in thecentral areas of many major cities in the country (Singh,2005). Since fuel efficiency is optimum at much higherspeed, improvement in traffic flow, removal of encroach-ment on roads, restraint on parking in congested areas, androad infrastructure improvement will have potential toimprove fuel efficiency. Similarly, adequate inspection andmaintenance program for in-use vehicles, higher price forfuel, and tax concession for fuel-efficient vehicles cansignificantly improve fuel efficiency. One should note thathigher fuel prices induce people to prefer vehicles that arefuel efficient. Demand for fuel-efficient vehicles can beincreased by providing tax concessions for these vehicles.For example, today, the most efficient car available in theIndian market is about twice as fuel efficient as the leastfuel efficient car. Models such as Maruti 800, Maruti Alto,and Maruti Zen have an average petrol consumption ofabout 5–6 litres per 100 km whereas Ambassador, onceIndia’s most popular car, consumes 8–10 litres of petrol for100 km. Similarly, most efficient two-wheeler (such asHero-Honda with fuel efficiency of 1 litre per 80 km) istwice as fuel efficient as the least fuel efficient two-wheeler(such as Bajaj Scooter with fuel efficiency of 1 litre per40 km). Similar kinds of fuel efficiency differences exist inother categories of vehicles. Therefore, government policycan easily increase the fuel efficiency of the system throughpricing and taxation. Although it is beyond the scope ofthis study to estimate the efficiency gain potential ofdifferent policy measures, we assume that the government

action can reduce the energy intensities of all modes at therate of 1% per year up to 2020–2021. That is why; energydemand from the land-based passenger transportation inthe country is analyzed for two different possible modalenergy intensities in the future.

4.1.1. Business as usual (BAU) scenario

In this scenario, the average energy intensity in thepassenger transport sector in India is expected to increaseby around 43% in a span of 20 years from 0.30MJ/PKm in2000–2001—0.43MJ/PKm in 2020–2021. This is mainlybecause of increased dependence on automobiles to meetthe travel demand in future. Combined with the growth inaggregate transport demand and the projected modal splitchange, overall land-based passenger transport sectorenergy use in 2020–2021 is projected to be 5584 peta joules(Table 4).15 Energy demand in the sector is expected toincrease at the rate of more than 8% per year from2000–2001 to 2020–2021. Table 4 also presents the percapita energy consumption for transport in 2000–2001 and2020–2021. Energy consumption per person is projected torise by a factor of 4 from 1041MJ in 2000–2001 to 4201MJin 2020–2021.

4.1.2. Efficiency gain scenario

The BAU scenario shows huge increase in energyrequirements. Emphasis on efficiency improvements canreduce the energy demand significantly. Hence, in thisscenario, it is assumed that there will be continuous effortsto increase the energy efficiency. In particular, we assumethat energy intensity of all modes will decrease at the rateof 1% per year from 2000–2001 to 2020–2021. In thisscenario, energy requirement in 2020–2021 is estimated tobe around 4545 peta joules, around 1000 peta joules lessthan the BAU scenario.16 Similarly, per capita energyconsumption in 2020–2021 will be 3420MJ rather than4201MJ (Table 5). Even when there is a reduction inenergy intensity of all modes by 1% per year, passengertransport energy demand in India will increase at the rateof 7.5% per year from 2000–2001 to 2020–2021. This willhave huge implications for oil import bills of India. Oilimports, which meet around 65% of India’s currentdemands, will continue to grow due to rapid increase indemand from the transport sector.

4.2. CO2 emissions

The problem of climatic change is one of the mostserious consequences of the emission of large quantities ofCO2 and other greenhouse gases into the atmosphere.Transport in general and road transport in particular

ARTICLE IN PRESS

Table 5

Travel and passenger transport energy use in India: efficiency gain scenario

Mode of

transport

2000–2001 2020–2021 (efficiency gain scenario)

BPkm Energy

intensity (MJ/

Pkm)

Energy

demand (PJ)

Energy use

per person

(MJ)

BPkm Energy

intensity (MJ/

Pkm)

Energy

demand (PJ)

Energy use

per person

(MJ)

Car 283 0.94 266.02 261.06 2312 0.77 1780.00 1339.08

Two-wheeler 364 0.53 192.92 189.32 3065 0.43 1317.92 991.46

Auto-

rickshaw

102 0.58 59.16 58.06 870 0.47 408.96 307.66

Bus 2330 0.19 442.70 434.45 5662 0.16 905.97 681.56

Rail 457 0.19 86.83 85.21 1078 0.12 129.35 97.31

Total 3536 0.30 1060.80 1041.02 12 987 0.35 4545.45 3419.50

Table 4

Travel and passenger transport energy use in India: business as usual scenario

Mode of

transport

2000–2001 2020–2021 (business as usual scenario)

BPkm Energy

intensity (MJ/

Pkm)

Energy

demand (PJ)

Energy use

per person

(MJ)

BPkm Energy

intensity (MJ/

Pkm)

Energy

demand (PJ)

Energy use

per person

(MJ)

Car 283 0.94 266.02 261.06 2312 0.94 2172.98 1634.72

Two-wheeler 364 0.53 192.92 189.32 3065 0.53 1624.41 1222.03

Auto-

rickshaw

102 0.58 59.16 58.06 870 0.58 504.67 379.66

Bus 2330 0.19 442.70 434.45 5662 0.19 1075.84 809.35

Rail 457 0.19 86.83 85.21 1078 0.14 150.91 113.53

Total 3536 0.30 1060.80 1041.02 12 987 0.43 5584.41 4201.10

172000–2001 CO2 intensities by mode have been estimated by the author

based on the data provided by Ramanathan and Parikh (1999).18Sum of modal CO2 emission may not be equal to the total of the sector

since intensity figures have been rounded off to 2 decimal places.


constitutes a major share in the CO2 emissions. Vehiclesusing fossil fuels (diesel and gasoline) produce CO2

emissions in quantities that depend on the carbon presentin the fuel molecule. Globally, the transport sector nowcontributes 25% of all the CO2 emissions released into theatmosphere. Approximately 80% of those emissions arefrom road transport. Although, currently, India is one ofthe lowest per capita emitters of CO2, at 0.27metric tons ofcarbon equivalent, energy sector’s carbon intensity is high,and the country’s total CO2 emissions rank among theworld’s highest. In 2002, CO2 emission in India was around280 million metric tons of carbon equivalent which wasaround 4% of the world total (International EnergyAnnual 2002). Between 1980 and 2002, India’s carbonemission increased at an astonishing rate of 5.7% perannum against the world average of 1.26%.

Since land-based passenger transport sector accounts fora significant proportion of energy consumed in the country,its share in CO2 emission will be equally significant. Theaim of this sub-section is to estimate the present as well asfuture CO2 emission from the passenger transport sector inIndia. For this, we have to estimate the CO2 emissionintensities for all the major modes. Tables 6 and 7summarizes the estimated 2000–2001 CO2 intensities and

their projected values for 2020–2021 along with the level ofCO2 emission from the sector.17 In this case also, twoscenarios will be explained in line with the discussion in theprevious section.

4.2.1. BAU scenario

In the BAU scenario, the 2020–2021 CO2 intensities ofall ground transport modes except rail are assumed toremain at 2000–2001 levels. The rail CO2 intensity isdifferent because we assume that the electrified routes willcater to the need of 80% of the rail passenger traffic in2020–2021 rather than the current level of 40%. Assumingthat refined oil will continuously fuel the passengertransportation sector, CO2 emissions will grow more orless in proportion to the energy used by the sector. In theBAU scenario, CO2 emission is projected to increase from19.80 to 93.25 million metric tons of carbon equivalentin a span of 20 years between 2000–2001 and 2020–2021(Table 6).18 The projected average annual rate of growth in

ARTICLE IN PRESS

Table 6

Intensities and the level of CO2 emission from different modes of transport: business as usual scenario

Mode of transport 2000–2001 2020–2021 (business as usual scenario)

BPkm CO2 intensity

(grams of carbon

equivalent per

Pkm)

CO2 emission

(million metric

tons of carbon

equivalent)

BPkm CO2 intensity

(grams of carbon

equivalent per

Pkm)

CO2 emission

(million metric

tons of carbon

equivalent)

Private and para-transit modes

(car, two-wheeler, and auto-

rickshaw)

749 10.05 7.53 6247 10.05 62.78

Bus 2330 4.19 9.76 5662 4.19 23.72

Rail 457 5.50 2.51 1078 6.29 6.78

Total 3536 5.60 19.80 12 987 7.18 93.25

0

10

20

30

40

50

60

70

80

Kgs

of c

arbo

n eq

uiva

lent

Private- and para- transit modes

Buses Trains Total

57.41

70.17

19.43

4.185.12.46

14.6117.84

9.58

38.63

47.23

7.39

2000-01

2020-21 (Business as Usual Scenario)

2020-21 (Efficiency Gain Scenario)

Fig. 10. Per capita CO2 emission from different modes of transport

(kilograms of carbon equivalent).


CO2 emission, around 8%, is somewhat similar to thecorresponding growth in energy demand from 2000–2001to 2020–2021. Fig. 10 presents the per capita CO2 emissionfrom passenger transportation during the year 2000–2001and 2020–2021. In the BAU scenario, CO2 emission perperson is projected to increase at the rate of 6.63% per yearfrom 19.43 kg of carbon equivalent in 2000–2001 to70.17 kg of carbon equivalent in 2020–2021.

4.2.2. Efficiency gain scenario

In the efficiency gain scenario, the CO2 intensities of allmodes decline at the rate of 1% per year up to 2020–2021by assuming that this may happen due to the technologicalchange (fuel efficiency improvements and transportation fuelswith lower carbon content), improvement in transportationsystem management and/or some combination of these two.Emphasis on efficiency improvements is expected to reducethe CO2 emission significantly. In this scenario, the CO2

emission in 2020–2021 is estimated to be 76.36 million metrictons of carbon equivalent, around 18% less than the emissionin the BAU scenario (Table 7).19 Similarly, per capita CO2

19In the efficiency gain scenario also, sum of modal CO2 emission may

not be equal to the total of the sector since CO2 intensity figures have been

rounded off to 2 decimal places.

emission in 2020–2021 in this scenario is projected to be57.41 kg of carbon equivalent rather than 70.17 (Fig. 10).One should note that even if there is a reduction in CO2

intensity of all modes by 1% per year, the level of CO2

emission from passenger transportation in India willincrease at the rate of around 7% per year from2000–2001 to 2020–2021.

5. Concluding remarks

In this study, we projected the level of traffic mobility,energy demand, and CO2 emission from land-basedpassenger transportation in India up to 2020–2021. Thelevel of land-based passenger traffic in India increased atthe rate of 7.75% per year during last two decades from794BPkm in 1980–1981 to 3536BPkm in 2000–2001 and isexpected to increase at the rate of 6.72% per year duringthe next two decades. Land-based traffic volume in2020–2021 is projected to be nearly 13 000BPkm, out ofwhich, 91.7% will be provided by road and the rest by rail.An average Indian traveled 3470 km in 2000–2001, out ofwhich, 449 km was by rail and 3021 km by road. After twodecades, in 2020–2021, their annual travel figure isprojected to be 9770 km–811 km by rail and 8959 km byroad. Analysis of modal split reveals that the share ofpublic transport modes (buses and trains) in providingpassenger mobility in India will decline from 78.8% in2000–2001 to 51.9% in 2020–2021, whereas share ofprivate- and para-transit modes will increase from 21.2%to 48.1% during the same period. Among the private- andpara-transit modes, the share of two-wheelers is expectedto increase from 10.3% in 2000–2001 to 23.6% in2020–2021, while the corresponding increase for cars andauto-rickshaws will be from 8.0% to 17.8% and from 2.9%to 6.7%, respectively. The expected rapid increase inmobility and more energy intensive modal split will havehuge implications for energy demand and CO2 emissionfrom passenger transportation in India.Assuming that energy intensity of various transport

modes remains unchanged, the energy demand in thepassenger transport sector is expected to increase at the

ARTICLE IN PRESS

Table 7

Intensities and the level of CO2 emission from different modes of transport: efficiency gain scenario

Mode of transport 2000–2001 2020–2021 (efficiency gain scenario)

BPkm CO2 intensity

(grams of carbon

equivalent per

Pkm)

CO2 emission

(million metric

tons of carbon

equivalent)

BPkm CO2 intensity

(grams of carbon

equivalent per

Pkm)

CO2 emission

(million metric

tons of carbon

equivalent)

Private and para-transit modes

(car, two-wheeler, and auto-

rickshaw)

749 10.05 7.53 6247 8.22 51.35

Bus 2330 4.19 9.76 5662 3.43 19.42

Rail 457 5.50 2.51 1078 5.15 5.55

Total 3536 5.60 19.80 12 987 5.88 76.36


rate of more than 8% per year from 1060.8 peta joules in2000–2001 to 5584.4 peta joules in 2020–2021. Similarly,energy consumption per person is expected to rise by afactor of 4 from 1041MJ in 2000–2001 to 4201MJ in2020–2021. Even when we assume a reduction of 1% peryear in energy intensity of all modes, passenger transportenergy demand is projected to increase at the rate ofaround 7.5% per year.

Assuming that refined oil will continuously fuel thepassenger transportation sector, CO2 emission will growmore or less in the same proportion as energy demand. Inthe business as usual scenario (i.e., assuming that the CO2

intensity of all modes remains unchanged), CO2 emission isprojected to increase from 19.80 to 93.25 million metrictons of carbon equivalent in a span of two decades between2000–2001 and 2020–2021. In the efficiency gain scenario(i.e., assuming that the CO2 intensity of all modes declinesat the rate of 1% per year), the CO2 emission in 2020–2021is projected to be 76.36 million metric tons of carbonequivalent. One should note that even in efficiency gainscenario, the level of CO2 emission from passengertransportation in India is expected to increase at the rateof around 7% per year during the next two decades.

Apart from CO2, substantial amount of local pollutantslike carbon monoxide (CO), unburnt hydrocarbons (HC),nitrogen oxides (NOx), sulfur dioxide (SO2), lead (Pb), andsuspended particulate matters (SPM) are also emitted bythe passenger transport sector. The air pollution problemdue to vehicular emission in most of the metropolitan citiesin India is taking serious dimension and worsening people’squality of life (Singh, 2005). Pollutants from vehicularemission have various adverse health effects. One of themain pollutants SPM, particularly fine PM, has serioushealth effects, especially in the form of respiratory diseases.The ambient air pollution in terms of SPM in allmetropolitan cities in India exceeds the limit set by WorldHealth Organization (WHO).

India faces significant challenges in balancing itsincreased demand for energy with the need to protect itsenvironment from further damage. Population growth and

urbanization make the task all the more difficult. Rapidincrease in vehicle ownership will aggravate the alreadyexisting air pollution problem and urbanization willincrease the health risks from that pollution. In the absenceof coordinated government efforts, including stricterenforcement, air pollution is likely to continue to worsenin the coming years. India’s ability to safeguard itsenvironment will depend on its success in promotingpolicies that keep the economy growing while fulfillingthe energy demand in a sustainable manner.

Acknowledgements

This paper is part of a study on Passenger TransportMarket in India sponsored by the Indian Institute ofTechnology, Kanpur, India. I am thankful to the Directorand Dean (Research & Development) of the institute forproviding me with an initiation grant for the study. I wouldalso like to thank Prof. Surajit Sinha (Department ofHumanities and Social Sciences, IIT Kanpur, India) andtwo anonymous referees for their helpful comments andvaluable suggestions which considerably improved theexposition of this work.

Appendix A. Data descriptions

Although, annual time series data of the level of rail-based passenger mobility in India from 1950–1951 to2000–2001 is readily available (e.g., in Statistical Abstractof India published by the Central Statistical Organization,Ministry of Statistics and Programme Implementation,Government of India, New Delhi), similar series for road-based passenger traffic volume is not offered by any source.Therefore, it is decided to estimate the level of road-basedpassenger traffic from 1950–1951 to 2000–2001 on the basisof services provided by the different modes. The estimatesare based on the following passenger vehicles: (i) cars (ii)two-wheelers (iii) auto-rickshaws and (iv) buses. Carsinclude jeeps and taxis as well.

ARTICLE IN PRESS

Table A1

Motor vehicle population in thousand and its compound annual growth rate in percentage since previous period (in parentheses)

Year Cars Two-wheelers Auto-rickshaws Buses Total passenger vehicles Total motor vehicles

1950–1951 159.3 (�) 26.9 (�) 1.7 (�) 34.4 (�) 222.2 (�) 306.3 (�)

1955–1956 203.2 (5.0) 41.0 (8.8) 2.5 (8.8) 46.5 (6.2) 293.1 (5.7) 425.6 (6.8)

1960–1961 309.6 (8.8) 88.4 (16.6) 6.2 (19.9) 56.8 (4.1) 461.0 (9.5) 664.5 (9.3)

1965–1966 455.9 (8.0) 225.6 (20.6) 16.1 (20.8) 73.2 (5.2) 770.8 (10.8) 1099.1 (10.6)

1970–1971 682.0 (8.4) 576.0 (20.6) 36.7 (17.9) 91.4 (4.5) 1386.1 (12.5) 1865.0 (11.2)

1975–1976 779.0 (2.7) 1057.0 (12.9) 59.4 (10.1) 114.2 (4.6) 2009.6 (7.7) 2720.0 (7.8)

1980–1981 1160.0 (8.3) 2618.0 (19.9) 142.1 (19.0) 153.9 (6.2) 4074.0 (15.2) 5391.0 (14.7)

1985–1986 1780.0 (8.9) 6245.0 (19.0) 336.9 (18.9) 227.6 (8.1) 8589.5 (16.1) 10 577.0 (14.4)

1990–1991 2954.0 (10.7) 14 200.0 (17.9) 617.4 (12.9) 331.1 (7.8) 18 102.5 (16.1) 21 374.0 (15.1)

1995–1996 4204.0 (7.3) 23 252.0 (10.4) 1009.0 (10.3) 449.0 (6.3) 28 913.9 (9.8) 33 786.0 (9.6)

2000–2001 7058.0 (10.9) 38 556.0 (10.6) 1725.4 (11.3) 560.0 (4.5) 47 899.4 (10.6) 54 991.0 (10.2)

Table A2

The level of passenger mobility provided by different modes of road transport during selected years (in BPkm)

Year Cars Two-wheelers Auto-rickshaws Buses Road transport

1950–1951 6.38 0.25 0.10 28.99 35.72

1955–1956 8.14 0.39 0.15 50.52 59.19

1960–1961 12.40 0.84 0.37 80.15 93.76

1965–1966 18.27 2.13 0.95 123.29 144.63

1970–1971 27.33 5.44 2.16 204.72 239.65

1975–1976 31.21 9.99 4.68 306.88 352.76

1980–1981 46.48 24.74 8.38 505.81 585.40

1985–1986 71.32 59.02 19.86 757.48 907.68

1990–1991 118.36 134.19 36.40 1198.32 1487.27

1995–1996 168.45 219.73 59.49 1830.36 2278.03

2000–2001 282.80 364.35 101.73 2329.60 3078.49


Table A1 reports category-wise motor vehicle popula-tion in India for selected years between 1950–1951 and2000–2001. This is based on data given in the MotorTransport Statistics published by the Ministry of RoadTransport and Highways, Government of India, NewDelhi and Statistical Abstract of India published by theCentral Statistical Organization, Ministry of Statistics andProgramme Implementation, Government of India, NewDelhi. The traffic mobility provided by the differentcategories of vehicles has been computed after makingreasonable assumptions regarding their average annualutilization and average occupancy. These assumptions arebased on studies like National Transport Policy CommitteeReport (1980), Planning Commission, New Delhi; RoadDevelopment Plan 1981–2000 (1984), Indian Road Con-gress, New Delhi; Estimation of Road Transport Passengerand Freight Demand (1986), Study Report of Ministry ofSurface Transport, New Delhi; Report of Steering Groupon Transport Planning (1987), Planning Commission, NewDelhi; and Singh M. and Kadiyali L. R. (1990) writtenbook on Crisis in Road Transport published by theKonark Publishers Pvt. Ltd., New Delhi. Annual utiliza-tion of cars, two-wheelers, and auto-rickshaws are assumedto be 12 600, 6300, and 33 500 km, respectively. Average

occupancy of a car, two-wheeler, and auto-rickshaw areassumed to be 3.18, 1.5, and 1.76, respectively. Accord-ingly, the level of passenger mobility provided by thesemodes has been computed and presented in Table A2.Estimation of traffic mobility provided by the buses

requires data on bus population, average annual utiliza-tion, occupancy ratio (ratio of number of passengers to theseats offered), and average seating capacity. These data aretaken from various sources such as State TransportUndertakings: Profile and Performance (various issues)published by the Central Institute of Road Transport,Pune; TERI Energy Data Directory & Yearbook (variousissues) published by the TERI, New Delhi; and Singh M.and Kadiyali L. R. (1990) written book on Crisis in RoadTransport published by the Konark Publishers Pvt. Ltd.,New Delhi. Table A3 presents these data. Assuming thatthe average seating capacity is 52, the level of passengermobility provided by the buses has been computedand presented in both Tables A2 and A3. Rail-basedpassenger mobility data is readily available from1950 to 1951 onwards. Passenger mobility data for rail,road, and land (aggregate of rail and road) for selectedyears between 1950–1951 and 2000–2001 have beenreported in Table A4.

ARTICLE IN PRESS

Table A4

Trends in rail, road, and land-based passenger mobility in India

Year Rail pass.-km (in

billion)

CAGR in

percentage wrt

previous period

(rail)

Road pass.-km (in

billion)

CAGR in

percentage wrt

previous period

(road)

Land pass.-km (in

billion)

CAGR in

percentage wrt

previous period

(land)

1950–1951 66.52 — 35.72 — 102.24 —

1955–1956 62.90 �1.1 59.19 10.6 122.09 3.6

1960–1961 77.67 4.3 93.76 9.6 171.42 7.0

1965–1966 96.76 4.5 144.63 9.1 241.39 7.1

1970–1971 118.12 4.1 239.65 10.6 357.77 8.2

1975–1976 148.76 4.7 352.76 8.0 501.52 7.0

1980–1981 208.56 7.0 585.40 10.7 793.96 9.6

1985–1986 240.62 2.9 907.68 9.2 1148.30 7.7

1990–1991 295.64 4.2 1487.27 10.4 1782.91 9.2

1995–1996 342.00 3.0 2278.03 8.9 2620.03 8.0

2000–2001 457.02 6.0 3078.49 6.2 3535.51 6.2

Table A3

Growth of Indian bus industry; 1950–1951—2000–2001

Year Bus population Average annual utilization (kms) Occupancy ratio (percent) BPkm

1950–1951 34 411 36 000 45 28.988

1955–1956 46 461 41000 51 50.518

1960–1961 56 792 46000 59 80.149

1965–1966 73 175 54 000 60 123.285

1970–1971 91 406 59 000 73 204.717

1975–1976 114 193 68 000 76 306.878

1980–1981 153 909 79 000 80 505.807

1985–1986 227 608 80 000 80 757.479

1990–1991 331 100 87 000 80 1198.317

1995–1996 448 970 98 000 80 1830.361

2000–2001 560 000 100 000 80 2329.600


References

Bewley, R., Fiebig, D., 1988. Flexible logistic growth model with

applications in telecommunications. International Journal of Fore-

casting 4 (2), 177–192.

CMIE database on Energy and Infrastructure (various issues), Centre for

Monitoring Indian Economy (CMIE) Pvt. Ltd., Mumbai, India.

Dargay, J., Gately, D., 1999. Income’s effect on car and vehicle ownership,

worldwide: 1960–2015. Transportation Research Part A 33 (2),

101–138.

Franses, P.H., 2002. Testing for residual autocorrelation in growth curve

models. Technological Forecasting and Social Change 69 (2), 195–204.

Heij, C., et al., 2004. Econometric Methods with Applications in Business

and Economics. Oxford University Press, New York, p. 209.

International Energy Annual, 2002. Energy Information Administration.

Washington, DC, USA. Available on http://www.eia.doe.gov/emeu/

iea/carbon.html.

Marchetti, C., 1994. Anthropological invariants in travel behavior.

Technological Forecasting and Social Change 47 (1), 75–88.

Meade, N., Islam, T., 1995. Forecasting with growth curves: an empirical

comparison. International Journal of Forecasting 11 (2), 199–215.

Meade, N., Islam, T., 1998. Technological forecasting–model selection,

model stability, and combining models. Management Science 44 (8),

1115–1130.

Mohamed, Z., Bodger, P., 2005. A comparison of Logistic

and Harvey models for electricity consumption in New

Zealand. Technological Forecasting and Social Change 72 (8),

1030–1043.

National Accounts Statistics of India: 1950–1951—2002–2003. EPW

Research Foundation, Mumbai, India.

Preston, J., 2001. Integrating transport with socio-economic activity—a

research agenda for the new millennium. Journal of Transport

Geography 9 (1), 13–24.

Ramanathan, R., 1998. Development of Indian passenger transport.

Energy—The International Journal 23 (5), 429–430.

Ramanathan, R., Parikh, J.K., 1999. Transport sector in India: an

analysis in the context of sustainable development. Transport Policy 6

(1), 35–45.

Schafer, A., 1998. The global demand for motorized mobility. Transport

Research: Part A 32 (6), 455–477.

Schafer, A., Victor, D.G., 2000. The future mobility of the world

population. Transport Research: Part A 34 (3), 171–205.

http://www.eia.doe.gov/emeu/iea/carbon.html

http://www.eia.doe.gov/emeu/iea/carbon.html

ARTICLE IN PRESSS.K. Singh / Transport Policy 13 (2006) 398–412412

Singh, S.K., 2000. Estimating the level of rail- and road-based passenger

mobility in India. Indian Journal of Transport Management 24 (12),

771–781.

Singh, S.K., 2005. Review of urban transportation in India. Journal of

Public Transportation 8 (1), 79–97.

TERI Energy Data Directory & Yearbook (various issues), TERI, New

Delhi, India.

World Population Prospects: The 2004 Revision Population Database.

United Nations Population Division, United Nations. Available on

http://esa.un.org/unpp/index.asp?panel=3.

http://esa.un.org/unpp/index.asp?panel=3

Documents

Future mobility in India: Implications for energy demand and CO2 emission