Unlocking Cities: The Impact of Ridesharing Across India
Unlocking Cities: The Impact of Ridesharing Across India
Commissioned by
Vincent Chin, Mariam Jaafar, Suresh Subudhi, Nikita Shelomentsev, Duong Do and
Irfan Prawiradinata
April 2018
ABOUT THIS REPORT
Over the last eight years, there has been a rise in the number of ridesharing firms
providing services in India. These include Uber and Ola. Ridesharing has intro-
duced a new mode of transport to commuters. Despite being in a nascent stage,
ridesharing has begun to influence the transport landscape. It is evident that
ridesharing has the potential to be a vital part of the solution to the region’s
transportation needs.
Uber has commissioned the Boston Consulting Group to assess the potential
benefits that greater adoption of ridesharing may bring to Asian cities. The findings
in this report were reached through research utilising publicly available transport
data, interviews with transport experts, and primary research involving commuters
in each city. The cities covered in this report are Delhi, Mumbai, Bangalore and
Kolkata. For comparison purposes, we have illustrated key indicators for Singapore,
Kuala Lumpur, Jakarta, Surabaya, Bangkok, Hong Kong, Taipei, Ho Chi Minh City,
Hanoi and Manila.
2 Unlocking Cities
In Brief
•• Growth in population and wealth have led to an explosion in transport demand
in India—an increase of 8x since 1980.
•• This explosive growth has put significant strain on Indian transport infrastructure.
Major Indian cities are constantly ranked among the world’s most congested.
•• Ridesharing can play a key role in ensuring higher efficiency in the use of existing
assets such as private vehicles. Ridesharing is a new point-to-point transport
model characterised by seven features: (1) Flexible supply base, (2) Smart dis-
patching, (3) Dynamic pricing, (4) Customer network effect, (5) Dynamic routing,
(6) Demand pooling and (7) Feedback collection and management system.
ǟǟ Supplementing incomes
The level of congestion is significantly higher in Indian cities than comparable cit-
ies around Asia, averaging 149%. This is partly attributable to India’s large popula-
tion and high population density, as well as an under-developed public transporta-
tion network (especially rail-based transport). Additionally, the use of private car is
relatively high in Indian cities. It is the most common mode of private transporta-
tion across the four cities surveyed (modality share ranges from 22%-45%).
High levels of private car usage have led to significant inefficiency. Ridesharing of-
Rideshare vehicles fers one way to improve the utilisation of these vehicles. Our study found that ride-
were able to achieve a share vehicles average 1.95x higher utilisation per year (measured as people-kilo-
high level of efficiency metres per vehicle per year). Furthermore, on an average, 80% of the commuters
in comparison to we surveyed expressed some willingness to forgo purchasing a car if rideshare could
private cars meet their desired level of service. In this way, rideshare adoption can potentially
help to reduce private car ownership across these cities.
Ideally, rideshare vehicles are able to achieve a high level of efficiency due to sever-
al advantages inherent in their technologies and operating models:
•• Dynamic pricing to incentivise drivers onto the road during peak demand
4 Unlocking Cities
ber of vehicles required to meet travel demand and bring several benefits in the
four cities covered.
Specifically, in Delhi and Mumbai, the key benefits would include (1) Providing al-
ternatives to car ownership, (2) Accelerating public transport adoption, and (3) Sup-
plementing incomes. In Bangalore and Kolkata, the key benefits would include (1)
Providing alternatives to car ownership, (2) Optimising the timing of infrastructure
investment, and (3) Supplementing incomes.
In this optimal scenario, a reduction of 33%-68% in private cars and congestion re-
duction by 17%-31% can be achieved across these cities. Consequently, this could
significantly help these cities improve their amenity by saving approximately 760 to
22,000 acres of unnecessary parking space in each city.
Index
800
600
400
200
1985
2003
1980
1981
1982
1983
1984
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
Sources: World Bank; OECD; National Center for Sustainable Transportation; BCG analysis.
In this report, we focus on the current state of transportation and congestion in four of
Congestion is the largest cities in India, namely Delhi, Mumbai, Bangalore, and Kolkata, where total
estimated to cost the congestion costs were estimated to be as much as USD 22 billion per year. We then esti-
four cities USD 22 mate the potential positive impact of ridesharing in each city. Indian cities share similar
billion per year challenges in terms of high congestion and under-developed public transport networks
unable to meet demand. However, each city’s transport network has unique character-
istics. The cities covered in this study may be further divided into two categories:
•• Delhi and Mumbai: These two cities are mega urban centres with relatively
more developed modern public transport systems. According to our interviews,
the Delhi metro is frequently considered to be the best quality transportation
network in the country, while Mumbai’s suburban rail, new metro line, and Bus
Rapid Transit (BRT) are known to be the busiest nationwide. Nevertheless,
congestion levels remain high due to a large number of private vehicles and the
low quality of road infrastructure. With the burgeoning population and the
growing prosperity of Delhi and Mumbai, the reliance on cars is expected to
increase, adding more pressure to road networks.
•• Bangalore and Kolkata: Relatively smaller than Delhi and Mumbai in terms of
population, these two cities have less modern and more road-based public trans-
port networks. While investments are being made to build more modern rail lines,
the current lack of public transport capacity, coupled with huge private vehicle
ownership numbers (both cars and motorbikes), has caused intense congestion.
Going forward, each of these cities are expected to face diverse challenges that
would require different approaches to achieve sustainable solutions. However, each
city will need to reduce their reliance on private vehicles and find more efficient
point-to-point transport alternatives.
•• Overall, travel by public transport accounts for 19% and 54% of kilometres
travelled in Delhi and Mumbai, respectively. The need to further improve
overall quality, network reach and ease of access from feeder transport modes
have been recognised as key challenge areas. Governments of both cities have
indicated that maintaining control over vehicle growth and encouraging the use
of public transport are their key objectives going forward.
•• For Kolkata and Bangalore, congestion levels are relatively higher than in
other cities, despite their smaller populations. This is driven by the limitation of
their older public transport networks which are primarily road-based, along
with a significant growth in private vehicles. Looking forward, a combination of
infrastructure improvement, addition of more modern mass transit as well as
efficient alternatives to vehicle ownership are likely to help curb congestion.
6 Unlocking Cities
EXHIBIT 2 | Peak Hour Congestion (% Additional Time to Travel in Peak Hours)
200
171
162
150
135 132 134
129
112
105
100
79
65 68 70
63 67%
57
50
0
Delhi Bangalore Singapore Surabaya Taipei Bangkok Manila
Sources: TomTom traffic index; Google API; Uber; Government statistics; BCG analysis.
Note: Asia average taken from average of East Asian cities based on TomTom traffic index.
10 9.60
5.92
6
4.8 0
1.97
2
0
Delhi Mumbai Bangalore Kolkata
Congestion
129% 135% 162% 171%
level
Sources: Centre for Urban Economic Studies, Department of Economics, University of Calcutta, Kolkata, India.
Dynamic pricing. Demand and pricing signals allow rideshare to efficiently mobi-
lize drivers to where they are most needed. Areas where customer demand signifi-
cantly exceeds driver supply will be shown to have higher pricing multiples (also
known as dynamic price or price surge). The higher the demand for drivers, the
8 Unlocking Cities
higher the price multiplier. Through the app, participating drivers can check re-
al-time market conditions (demand in the areas, pricing multiples, etc.). While there
have been debates on the use of dynamic pricing, this mechanism does provide
drivers with both information and incentive to move to high-demand areas and pro-
vide customers greater accessibility and choice.
Customer network. Rideshare allows drivers to access a large and expanding net-
work of passengers, thus decreasing the time drivers spend waiting for a booking.
This also increases the efficiency of supply-demand matching and facilitates pool-
ing services. As more customers join these platforms, the revenue pool for drivers
also increases, attracting new drivers to join. The flow-on effect of a growing net-
work is essential to realize the full potential of rideshare.
One factor appreciated by customers using rideshare, and which encourages loyalty,
is the added convenience of electronic payment, stored profiles, and digital records.
Frequent bonuses are given to customers who introduce new riders or drivers to
these platforms. In addition, car-hire platforms worldwide have introduced incen-
tives such as discounts and points redeemable for prizes to encourage repeat book-
ings (for example, the GrabRewards programs in Southeast Asia, Uber and SPG Pre-
ferred Guest Program, the Lyft and Delta Frequent Flyer partnership in the US,
etc.). Together with a simplified registration process, these features have encour-
aged rideshare growth.
Dynamic routing. Once driver and rider are matched, drivers are provided with re-
al-time updates on traffic conditions and the best route to the pickup location and
destination. The dynamic routing feature keeps drivers informed of the best routes
so they don’t have to rely on their personal experience and knowledge of the roads.
This feature not only maximizes driver benefits and helps to avoid congestion, but
also ensures that customers are adequately informed about their trips. Knowing a
driver’s location, contact details, and estimated arrival time, as well as the approxi-
mate cost of their trip, provides customers with an added sense of security, especial-
ly if they are new to the area.
Demand pooling. With sufficient levels of adoption, rideshare can improve vehicle
efficiency further through demand-pooling or trip-sharing. Drivers may have better
income opportunities because they are able to pick up multiple passengers on a sin-
gle trip. Pooling also lowers fares for customers, who share the cost of the trip. Ride-
share quickly maps available routes for multiple bookings, compare existing book-
ings to identify potential synergy, and direct drivers to pick up multiple passengers
on similar routes. Perhaps most importantly, demand- pooling also reduces the
number of cars required to meet transportation demand. It can help to alleviate
traffic congestion and reduce environmental impact.
This means that the reduction in private car ownership would also support
‘liveability’ in these cities. With reduced car ownership, land previously used for
parking spaces, could be re-allocated to enhance living conditions, such provid-
ing additional housing and social infrastructure. Fewer private cars usage would
also help to significantly reduce CO2 emissions.
In the case of Kolkata and Bangalore, the cities stand to benefit the most from
ridesharing by:
Respondents (%)
120
92 91
90 86 86 88
83 83 84
79 79 81
Ø 77
60 57
51
36
30
0
Delhi Bangalore Hong Taipei Jakarta Kuala Manila
Kong Lumpur
Mumbai Kolkata Singapore Hanoi Ho Chi Surabaya Bangkok
Minh
EXHIBIT 5 | Willingness for a Planned Car Buyer to Forgo Purchase, Provided Rideshare Meets
Desired Levels of Availability, Price, Timeliness
Respondents (%)
88 89 90
90 85 85 87
81 81 81 82 82 82 Ø 84
79 79
9 18
23 28 37 42 42 Highly willing
30 40
39 45
60 38 41
52
73
68
30 58
51 54 51
47 47 47 Somewhat willing
41 42 40
38
30
0
Delhi Bangalore Singapore Kuala Taipei Surabaya Hanoi
Lumpur
Mumbai Kolkata Bangkok Hong Jakarta Manila Ho Chi
Kong Minh
Under this scenario, ridesharing substantially reduces the number of vehicles re-
quired due to the higher people-kilometres ratio provided by each rideshare vehicle
as compared to private cars. Congestion would also decline significantly as a result
of the reduction in the number of cars trips. Furthermore, significant space, total-
ling as much as 33,000 acres, could be re-purposed from vehicle parking.
EXHIBIT 6 | Efficiency Gap Between Rideshare Vehicles vs. #1 Common Modes5 of Private Vehicles
(Car or Motorbike)
3.3x
100,000
1.6x
1.4x 1.5x
80,000
3.4x
1.7x 2.7x
60,000 3.2x
1.3x
1.8x 1.8x
2.0x 1.9x
40,000 1.7x
20,000
0
Bangalore Mumbai Jakarta Taipei Manila Hong Surabaya
Kong
Kolkata Delhi Kuala Bangkok Singapore Hanoi Ho Chi
Lumpur Minh
(%)
80
71 73
68 70
66
63 63
60
60 56 57 55
53
46 46 46
40 42
39 39
40 35
33 31
22 24
20 15
11
8 8
0
Delhi Bangalore Singapore Manila Jakarta Bangkok Hanoi
EXHIBIT 8 | Road Congestion During Peak Hours Before vs. After Rideshare (2017)
(%)
200
-19%
-17%
150 -27%
-31%
100
50
0
Delhi Mumbai Bangalore Kolkata
10
8 2.98
6
1.01
4 1.30
6.62
4.91
2 3.50 0.37
1.60
0
Delhi Mumbai Bangalore Kolkata
Congestion
31% 27% 17% 19%
Reduction
Source: Centre for Urban Economic Studies, Department of Economics, University of Calcutta, Kolkata, India; BCG analysis.
EXHIBIT 10 | Estimated Space That Can be Saved by Adopting Rideshare Assuming Rideshare
Substitutes for Private Cars
ACRES SAVED WITH RIDESHARE LANDMARK EQUIVALENT
Like traditional taxis, auto-rickshaw drivers could stand to benefit from working with
rideshare. This benefit is illustrated through the collaboration between Uber and Kaa-
li-Peelis taxis (refer to section “Utilisation of taxi” below for more details). However,
due to the price differences, consumers are unlikely to switch from rickshaws to ride-
share. Thus, we have excluded auto-rickshaws from our quantitative analysis.
Motorbikes: In the cities studied, our model estimates that the share of motorbikes
ranges from the lowest in Mumbai (12%) to the highest in Bangalore (32%), as mea-
sured by proportion of the total KM travelled. While these levels are lower com-
pared to those of private cars, they still comprise a considerable proportion of each
city’s transportation demand. This makes motorbikes a significant contributor to
congestion.
As an extension to the base scenario above, where rideshare is a substitute for pri-
vate cars, we also examine the effects it may have on motorbike population. While
replacing motorbike usage with rideshare adoption is less feasible given their cost
differences, our calculation shows that even if only ~1/3 of the current motorbikes
could be substituted with rideshare in addition to the base scenario above, it could
reduce between 22% to 38% of the total vehicles in each of the cities studied.
In a recent Shared Mobility Survey across six key cities in the US, rideshare users
reported the lowest car ownership levels. On average, 30% of shared mobility
users, including rideshare, reported shedding their vehicles.10
A recent study conducted across six US cities, namely Chicago, Los Angeles, Nash-
ville, San Francisco, Seattle and Washington DC, showed that only ~5% of ride-
EXHIBIT 11 | Estimated Public Transport Demand in Relation to Public Transport Capacity in 2022
% KM travelled by public transport
100
8 18 3
80
65 63
61 58 62
60
52 52
46
39 40
40 36 35
30 29
26 23 26 24
19 18 17
20 13
9
2
0
Delhi Bangalore Singapore Taipei Bangkok Manila Surabaya
This particular benefit was demonstrated in Jakarta where, in 1992, the government
introduced a policy that required vehicles to carry at least three occupants when
travelling on main routes during peak hours (3-in-1 policy). This policy, however,
lifted in 2016 due, at least in part, to concerns regarding the informal passen-
ger-for-hire (i.e. ‘jockey’) economy that had emerged as a consequence.12
A recent study by researchers at Harvard and MIT universities, found that following
the repeal of this policy, morning and evening congestion on the newly-liberalised
routes leaped by a staggering 46% and 87% respectively. Moreover, not only did
congestion leap on those central Jakarta roads where car-pooling was previously
mandated, it also increased during non-peak hours on routes and in areas that
were never subject to the pooling rule in the first place. In the hour following
the evening peak, for example 19:00-20:00, the repeal of this policy coincided with
a roughly 50% increase in delays.13
A common challenge in cities is matching the transport supply with commuter de-
mand to ensure sufficient supply during peak hours, and reducing supply during
off-peak hours, to minimise kilometres travelled without passengers (‘unproductive
miles’). Research in San Francisco indicated that rideshare vehicle ‘unproductive
miles’ is approximately half of taxis (as a percentage of total miles).15 Respon-
siveness to flexible demand allows ridesharing vehicles to be potentially more effi- 43% of shared-mobili-
cient in meeting demand, without adding to congestion, at times of lower demand. ty users reported an
increase in their use
Benefit 3: Accelerating public transport adoption of public transport in
A study published by the National Academy of Sciences, which covered several ma- several major US
jor US cities, found that 43% of shared-mobility users reported an increase in their cities
use of public transport, while only 28% of individuals reported reduced use of pub-
lic transport.16 Where public transport use has increased, the study suggests that
ridesharing is used to complement public transport, and can support a “car light”
lifestyle. This positive effect, however, may be limited due to economic conditions
and timing of travels. For example, a recent shared-mobility study found no clear
correlation between peak-hour use of rideshare and long-term increase in public
transport usage.17
Nevertheless, transit authorities are recognising the potential for ridesharing to act
as a feeder mechanism to public transport. In Portland, Oregon, for example, a local
transit authority (TriMet) has integrated rideshare booking capabilities into its pub-
lic transit app, as a means to enhance intermodal efficiency to public transport hubs.
Despite the obvious potential benefits of ridesharing, concerns have emerged about
the interaction between ridesharing and other transport modes such as taxi opera-
tors and public transport players. BCG has therefore explored these concerns and
suggested potential solutions. From our assessment, we found that a net positive
outcome can be realised for all stakeholders—ridesharing is not and need not be a
“zero sum” game. For Indian cities to achieve net positive benefits, several condi-
tions must be achieved:
This risk can be mitigated further by rideshare platforms and governments working to-
gether to establish programs that make ridesharing services an appealing complement
to public transport. For example, governments can work with ridesharing platforms to
provide commuters with live inter-modal travel data and to establish discounts or pool-
ing schemes for feeder transport to arterial public transport infrastructure.
Utilisation of taxis
The rise of rideshare has been perceived to reduce taxi ridership in some cities.
However, Ministers in both Singapore and Malaysia have suggested that rideshare
has served as a positive complement to taxis, particularly in peak hours.23, 24, 25
“In the morning peak hours where we have an inadequate supply of full-time taxi drivers,
many of the commuters’ interests are served because there are supplementary drivers that
Governments can also play a role in ensuring that taxi companies improve their com-
petitive position, while offering commuters better outcomes. For example, taxis
should be able to access the same technologies available to ridesharing vehicles. Both
taxis and private vehicles can form part of the flexible supply base necessary to real-
ise the congestion benefits outlined above. In particular, governments should ensure
that taxis can use apps to connect with passengers, while ensuring that taxis can avail
themselves of supply-demand matching mechanisms including dynamic pricing.
Partnerships between rideshare platforms and taxi companies can also benefit taxi
drivers. Recent examples of partnerships between rideshare platforms and taxi
companies include UberTAXI in Taiwan, UberFLASH in Malaysia and Grab’s part-
nerships with multiple Singaporean and Vietnamese taxi companies. These part-
nerships promise to benefit taxi drivers by offering them access to technology
which may allow more responsive matching of supply to demand, thereby increas-
ing vehicle utilisation and ridership. They may also benefit drivers by offering them
access to large networks of potential passengers.
In India, Uber collaborated with Kaali Peelis taxis in Mumbai to pilot usage of the
rideshare app among Kaali Peelis drivers. The sandbox experiment yielded strongly
positive results as drivers using the app were able to record 30% higher efficiency
(in terms of total KMs travelled per hour) compared to the control group, resulting
in 44% income uplift per hour.24
The demand for transport will continue to grow across Asian cities, leading to op-
portunities for incumbent transport models to evolve and for new transport models
to enter—ultimately leading to better transport outcomes for commuters.
Within the ridesharing industry, dynamic pricing has been the underlying mechanism
through which driver supply can respond flexibly to variable passenger demand. It also
promotes accessible pricing, since passengers do not need to cross-subsidise vehicles for
their idle time. Nonetheless, as ridesharing continues to evolve globally, its use of dy-
namic pricing, specifically the level of surge imposed, has met with debates.
Hall, Kendrik and Nosko (2015), showed that dynamic pricing helped rideshare, spe-
cifically Uber, in New York City, to effectively match demand-supply while main-
taining required service levels. Comparing similar peak events, when surge prices
worked as intended (e.g. after a sold-out concert in Madison Square Garden), there
was a rapid increase in active drivers within the applicable areas, matching that of
demand (~2x increase). This kept customers’ waiting time at normal level, averag-
ing 2.6 minutes, and 100% ride completion rate.
On the other hand, this efficiency was lacking when surge prices failed (e.g. New
During 26-minute Year Eve 2014-2015 during 26-minute surge pricing outage). The completion rate of
surge pricing outage rides dropped drastically to ~10% at the lowest point (compared to 100% in earlier
the completion rate of example). Waiting time also increased to 6 minutes at peak, more than 2 times the
rides dropped drasti- normal level.
cally to ~10% at the
lowest point Similar patterns were observed in Indian cities when state governments enforced
ceilings for surge prices. For example, in Bangalore, as a result of regulatory reform,
Uber imposed a ceiling on its surge prices starting from the night of May 30th, 2016.
While average demand levels did not differ significantly before and after this surge
ceiling introduction, the average estimated arrival time (i.e. customers’ waiting time
to be picked up) increased for both UberPool and UberGo services (two of its most
popular services by volume of rides) in the subsequent 10 day period. The increase
in waiting time was lower for UberGo, potentially due to its higher volume. Howev-
er, waiting time during high-demand hours (i.e. top 3 hours each day with the most
number of booking requests) increased significantly, by 12% and 15% for UberGo
and UberPool, respectively.
The impact on the level of unmet demand (i.e. level of booking requests unful-
filled) was even more acute. Average level of unmet booking requests during
high-demand hours increased by more than 3 times, in the 10 days after surge ceil-
Castillo, Knoepfl and Wey (2017) referred to this situation (longer waiting time and
high unfulfilled requests) as the “Wild-Goose-Chase” (WGC). Due to a lack of drivers
willing to provide services, commuters were matched with those who were far away,
resulting in long waiting times and high cancellation rates. This inefficiency is unfa-
vourable for both drivers (waste of time and income) and commuters (long waiting
time, high cancellation rate and lower chance of being able to book a ride). Using
simulations based on Uber data in Manhattan in 2016-2017, their study showed that
the usage of surge pricing is the best method to maximise total welfare.
The same researchers, explored alternatives for rationing demand and supply, all of
which resulted in undesirable effects on the whole transportation system. For exam-
ple, one way to reduce peak demand was randomly denying riders or denying those
with long waiting time. This was inefficient compared to a pure price mechanism
where drivers match with riders who have the highest willingness to pay. To increase
supply, expanding the dispatch radius was considered, though it would lead to high-
er waiting time, and the possibility of riders being denied a trip during peak hours.
The above evidence, though limited, seems to suggest that dynamic pricing is neces-
sary to enable the full benefits of ridesharing. Restrictions such as surge ceilings ap-
pear to immediately lower these benefits to below optimal for both commuters and
drivers. Hence, while there have been various debates and regulation changes sur-
EXHIBIT 12 | Customers’ Waiting Time Before and After Surge Ceiling Was Introduced
0.00 0.0
[Before] [After] [Before] [After] [Before] [After] [Before] [After]
Mean ETA Mean ETA Mean ETA Mean ETA Mean ETA Mean ETA Mean ETA Mean ETA
(high- (high- (high- (high-
demand) demand) demand) demand)
Indexed level of booking request un-fulfilled Indexed Level of booking request un-fulfilled
during high-demand hours during high-demand hours
5 5
4 4
3 3
2 2
1 1
0 0
Avg. increase 3.1x 3.4x
In the case of ridesharing, this review and revision process would be most effective
if the government and the industry were to closely collaborate. While state govern-
ment could be more open to reviewing and adjusting current rules, industry players
could supply them with the necessary data enabling effective decision making.
•• Price: Across the cities surveyed, approximately 82% of non-rideshare users cite
prices as a key deterrent. The majority of respondents indicate that prices must
be at least 25% lower than their current preferred mode of transport to make
them consider pooling.
•• Travel time: Similarly, 69% of the respondents indicated longer commute times
as being a deterrent.
Ridesharing has the potential to positively impact the transport environment across
India. Substantial growth in adoption is necessary in all cities, if benefits are to be
realised on a sustained basis. A combination of improved service offerings from
ridesharing platforms as well as support from state and regional regulators would
be required.
Notes
1. World Bank Database, retrieved January 15th, 2018
2. Based on indexed population and GDP per capita (constant) growth from 1980-2016 for Asian
countries among top 100 GDPs in world
3. Delhi traffic chaos costs Rs 60,000 crore annually, Times of India, 2017
4. World Bank Global Infrastructure Outlook database, retrieved January 20th, 2018
5. Defined by modality share by passenger-km
6. Delhi government statistic, retrieved January 15th, 2018
7. Le Vine & Polak, 2017
8. Hampshire, Simek, Fabusuyi, Di, & Chen, 2017.
9. During this period, other rideshare platforms continued to operate within Austin. The portion of
former Uber and Lyft customers who migrated to these alternative platforms is roughly the same as
the portion that migrated to cars (roughly 40%).
10. Feigon and Murphy, 2018
11. Feigon and Murphy, 2018
12. Anggun Wijaya, 2016; Tempo.co, 2016
13. Hanna, Kreindler, & Olken, 2017
14. Lyft Blog,2015; Uber Data, 2017
15. San Francisco County Transportation Authority and Northeastern University, 2017
16. Shared mobility and the transformation of public transit, Feigon, Murphy, 2016. Shared mobility
defined as public transit, bike sharing, car sharing, ridesharing, and similar modes
17. Feigon and Murphy, 2018
18. San Francisco County Transportaton Authority, 2017; Schaller, 2017
19. Hall and Krueger, 2016
20. Farrell and Greig, 2016a.
Analysis conducted
1. Qualitative research:
The qualitative research conducted as part of this report takes two primary
forms:
2. Quantitative research:
The quantitative research conducted as part of this report was used to model
the potential benefits provided by ridesharing and pooling under different
adoption scenarios
Survey methodology
In January 2018, we surveyed commuters across Delhi, Mumbai, Bangalore and Kol-
kata, covering approximately 300 commuters per city. The commuters surveyed
used a wide range of transportation.
Our results suggest that, rideshare represents an average of 10% of transport used
in the four Indian cities surveyed. These results support our quantitative assess-
ment that rideshare adoption is in its nascent stages.
25
20 19
15
13 13
11 12
11 11
10
10 9 Mean: 10
8 8
7
5 5
3
0
Delhi Bangalore Hong Bangkok Jakarta Singapore Kuala
Kong Lumpur
Mumbai Kolkata Taipei Hanoi Ho Chi Surabaya Manila
Minh
APPENDIX EXHIBIT 2 | Reasons Cited for not Adopting Rideshare in Comparison to Respondent
Preferred Mode of Transport
% of respondents
N = 440
100
82
80
70 69
37 59
60
26 25
23
40
29
46 44 44 10 Highly agree
20 37
19 Somewhat agree
0
High prices Safety concern Longer time taken Lack of availability/ Lack of awareness
difficulties in
booking
In the majority of cities surveyed, two-thirds of current non-users state that they
would be willing to adopt pooling if price, availability or speed worked out to be
better than that of their most preferred mode of transportation.
BCG surveyed the likelihood of commuters to purchase a car in the next 5 years.
Overall, commuters in all four cities indicated high intentions to purchase, averag-
ing at 87%. This relatively high probability can be attributed to the cities’ increasing
wealth and the fact that cars are considered to be a symbol for success.
Key Finding 5: Majority of drivers are willing to consider driving for rideshare to
supplement income
Overall, across the cities studied, 69% of respondents either somewhat or highly
agreed with the statement that they would be willing to use their own cars to work
as rideshare drivers. Enthusiasm for driving was highest in Kolkata at 77%, whereas
the level of willingness ranges from 65-69% in the other three cities.
2. Quantitative analysis
The focus of the quantitative analysis is to assess the impact of ridesharing on road
congestion under different scenarios of rideshare adoption. We define congestion as
the percentage of time difference in travelling during peak and non-peak hours
compared to the time it would take to travel the same distance at posted speed lim-
its. We have assessed peak hours at 7-9AM and 6-8pm.
% of respondents
N = 440
80
69 70
68
60
40
20
0
Up to 25% cheaper Slightly more easily available Slightly faster
APPENDIX EXHIBIT 4 | Percentage of Respondents Who Plan to Buy a Car Within the
Next Five Years
Respondents (%)
120
92 91
86 86 88
90 83 84
81 83
79 79
Ø 77
60 57
51
36
30
0
Delhi Bangalore Hong Taipei Jakarta Kuala Manila
Kong Lumpur
Mumbai Kolkata Singapore Hanoi Ho Chi Surabaya Bangkok
Minh
Source: BCG survey.
Respondents (%)
88 89 90
90 85 85 87
81 81 81 82 82 82 Ø 84
79 79
9 18
23 28 37 42 42 Highly willing
30 40
39 45
60 38 41
52
73
68
30 58
51 54 51
47 47 47 Somewhat willing
41 42 40
38
30
0
Delhi Bangalore Singapore Kuala Taipei Surabaya Hanoi
Lumpur
Mumbai Kolkata Bangkok Hong Jakarta Manila Ho Chi
Kong Minh
Respondents (%)
88 89 90
90 85 87
85
81 82 82 82
77 Ø 81
9 18
69
65 66 23 28 37 42 42 Highly willing
30 40
45
60
24 37
26 25
73
68
30 58
51 54 51
45 47 47 47 Somewhat willing
40 40 40 40
0
Delhi Bangalore Singapore Kuala Taipei Surabaya Hanoi
Lumpur
Mumbai Kolkata Bangkok Hong Jakarta Manila Ho Chi
Kong Minh
3. Traffic volume on road during the defined periods (peak, non-peak hours)
Road congestion in peak hours among the Indian cities studied averages at 149%,
significantly higher than those of South East Asian and East Asian cities. This
means that, on average, commuters take 1.5 times longer to travel a given distance
in peak hours compared to travel time during non-peak hours.
f
1 2
Actual drive-speed Post speed limit
f
3 4
Traffic volume on road Road capacity
5 6
Passenger car
# vehicles by type
equivalent conversion
7 8 9
Total people-KM Average occupancy by
Annual KM per vehicle
demand per vehicle type vehicle type
10 11
Total people-KM Modality share per
demand of the city transport mode
0
• % of additional travel time on average in peak, non-peak • Tom Tom Traffic Index
Road congestion
hours, when compared to driving at post speed limit • Government statistics
2
• Government data
Post speed limit • Post speed limits on highways, urban roads per city
• Press search
3 • Academic studies on
Traffic volume • Total traffic measured in passenger car equivalent units Transportation Engineering
on road on the road in peak, non-peak hours • Government statistics
• UBER data
5
• Number of vehicles by type: private cars, buses, taxi, • Government statistics
# vehicles by type
motorcycles, ridesharing cars and etc. • UBER data
7 Total people-KM
• Total distance travelled by the population using each of • Government statistics
demand per vehicle
the modes of transport • Survey
type
9 • Government statistics
Average occupancy
• Average number of people in a vehicle per trip • Survey
by vehicle type
• Expert interviews
200
171
162
150
135 132 134
129
112
105
100
79
65 68 70
63 67%
57
50
0
Delhi Bangalore Singapore Surabaya Taipei Bangkok Manila
Sources: TomTom traffic index; Google API; Uber; Government statistics; BCG analysis.
Note: Asia average taken from average of East Asian cities based on TomTom traffic index.
100 5 3 4 3 1 3 3 0 1 2 2 1 1 2 2 1
5 5 8 1 3 1 7 9 3
2
6
8
5 15 4 20
12 26
80 26
26 17 23
32 28
26 53 60
60 19 71
22
30 87
54
40 45 80 55
38
54 58 20
47 30 52
20 36 15
19 18 23
13 17
8 9 10
0 2
Delhi Bangalore Hong Taipei Bangkok Jakarta Surabaya
Kong
Mumbai Kolkata Singapore Kuala Ho Chi Manila Hanoi
Lumpur Minh
The ability for ridesharing vehicles to provide greater transportation benefits de-
pends on the difference in people-kilometres each rideshare vehicle provides in
comparison to other modes of transport. The figure below compares the estimated
people-kilometres ratio provided by ridesharing vehicles against private cars in
each city. Based on our estimate, ridesharing is between 1.4x-3.3x more efficient
than using a privately owned car.
In our study, we quantified the number of vehicles that could be taken off the road
in a scenario where the most widely owned private vehicles were substituted for
ridesharing. For example, in a market where private cars provide the second highest
form of modality and ridesharing provides the fifth highest form of modality, we as-
sessed how many vehicles could be saved if ridesharing became the second highest
form of modality.
APPENDIX EXHIBIT 11 | Average Annual People-kilometers Travelled Per Vehicle Type With Pooling
3.3x
100,000
1.6x
1.7x
1.2x
80,000 3.4x
1.7x 2.7x
60,000 3.2x
1.3x
1.8x 1.8x
2.0x 1.9x
1.7x
40,000
20,000
0
Bangalore Mumbai Jakarta Taipei Manila Hong Surabaya
Kong
Kolkata Delhi Kuala Bangkok Singapore Hanoi Ho Chi
Lumpur Minh
The increase in travel demand in the city can potentially be met by a greater adop-
tion of public transportation. Nevertheless, we estimate that the required increase
in rail network infrastructure may be greater than the capacity which will be creat-
ed by 2022 in Mumbai and Bangalore.
APPENDIX EXHIBIT 12 | Percentage of Private Vehicles (Car and Motorcycle) and Total Vehicles
Reduced with Rideshare With Pooling
(%)
80
73
71 70
68
66
63 63
60
60 56 57 55
53
46 46 46
40 42
39 39
40 35
33 31
22 24
20 15
11
8 8
0
Delhi Bangalore Singapore Manila Jakarta Bangkok Hanoi
200
-19%
-17%
150 -27%
-31%
100
50
0
Delhi Mumbai Bangalore Kolkata
10
8 2.98
6
1.01
4 1.30
6.62
4.91
2 3.50 0.37
1.60
0
Delhi Mumbai Bangalore Kolkata
Congestion
31% 27% 17% 19%
Reduction
Source: Centre for Urban Economic Studies, Department of Economics, University of Calcutta, Kolkata, India; BCG analysis.
100
8 18 3
80
65 63
61 58 62
60
52 52
46
39 40
40 36 35
30 29
26 23 26 24
19 18 17
20 13
9
2
0
Delhi Bangalore Singapore Taipei Bangkok Manila Surabaya
Mumbai Kolkata Hong Kuala Ho Chi Jakarta
Kong Lumpur Minh
APPENDIX EXHIBIT 16 | Estimated Space That Can be Saved by Adopting Rideshare Assuming
Rideshare Substitutes for Private Cars
ACRES SAVED WITH RIDESHARE LANDMARK EQUIVALENT
We estimate that the parking area that could be saved from a reduction in private
vehicles as a result of rideshare (in each city studied), ranges from 759 to 22,369
acres.
Mariam Jaafar is a Partner in BCG’s Singapore office and a member of Singapore’s Committee
on the Future Economy. She is also on the board of GovTech, the agency responsible for implemen-
tation of Singapore’s Smart Nation agenda. She has worked closely with multiple public sector cli-
ents across Asia Pacific on topics related to the digital economy, giving her a unique understanding
of the policy perspectives of the Singapore government. She can be contacted via e-mail at Jaafar.
[email protected].
Suresh Subudhi is a Partner in BCG’s Mumbai office of The Boston Consulting Group. He is the
Global Leader of Infrastructure sector and part of the Industrial Goods and Public Sector practice
leadership team. You may contact him via e-mail at [email protected].
Nikita Shelomentsev is a Project Leader based in Kuala Lumpur. He has over ten years of experi-
ence with consumer and automotive businesses in Europe and Asia Pacific. He has also supported
governments on strategic and change management topics such as innovation, economics and so-
cial development. He can be contacted via e-mail at [email protected].
Duong Do is a Consultant in BCG’s Ho Chi Minh City office and Irfan Prawiradinata is an Asso-
ciate in the Jakarta office. They can be contacted via e-mail at [email protected] and Prawiradi-
[email protected].
Acknowledgements
The authors would like to thank Panagiota Papakosta and the broader BCG GAMMA team for their
traffic congestion analysis, Jamshed Daruwalla, Jasmin Pithawala, Kim Friedman and Pradeep Hire
for design and production assistance and the BCG knowledge teams and Visual Services
department.