The Small Vehicle of Large-Scale Change

How we can consciously embed sustainability impact within frugal but efficient mobility design

IPTs, by nature of carrying multiple passengers, are more efficient in terms of emissions per passenger kilometres compared to private cars[...] Additionally, if they are given more legal recognition and policy support in India, the IPTs would benefit from technological advancements and reductions in GHG emissions akin to the evolution seen in the cars/automobile sector.

https://www.dhirubhai.net/posts/dhanuraj_make-shared-mobility-part-of-sustainability-activity-7261566888671887360--lCc/

An article I came across recently about paratransit highlights a very important aspect about perceptions and how it can sometimes impede progress in unexpected ways. It so happens that it comes from D Dhanuraj, the Chairman of the Centre for Public Policy Research, with whom I recently had the pleasure of interacting at a national-level event. Perceptions can sometimes make moonshot solutions seem like the only viable ones, rather than getting one's hands dirty in upgrading the status quo in simpler ways.

When we think about sustainable mobility, the more "visible" solutions centre around vehicle electrification and direct reduction of emissions by introducing a new fleet of vehicles on the streets, such as e-autos, e-3W, e-vans, and more. While these solutions are obviously viable, actionable, and important, the slightly enhanced visibility of - and attraction to - a certain set of solutions often eclipses other, more frugal ones.

Frugality has a lot of latent energy. Let me show you how.

If we look at #intermediateparatransit (#IPT) or informal public transport from a technical lens, we find that it is essentially a large fleet of small-capacity vehicles. Intermediate paratransit comes in different flavours across the world: auto-rickshaw, vans, and shuttles in India, minibuses and pods in Europe and the US, and jitneys, matatus, tro-tros, and more in countries of the Global South. While each vehicle can carry fewer passengers than a bus, the compact footprint lends the fleet an innate flexibility that larger public transport operations do not possess.

Imagine that a fleet of lakhs of auto-rickshaws did not already exist in our cities. In this scenario, introducing new vehicles that created no tailpipe emissions would be a great self-starting solution to serve the ever-increasing urban mobility demand in a clean and sustainable way.

Unfortunately, that is not the case for our cities. They are currently overburdened by a huge fleet that is operated by independent operators, the auto-rickshaw drivers. The drivers are themselves self-starting entrepreneurs, and any systemic intervention we seek to design must empower them while also improving mobility operations.

This gives us the following high-level goals for our (mobility) system (in no particular order):

1. Reduce the carbon footprint of the act of providing last-mile or point-to-point mobility to commuters: (i) at the fleet level, and (ii) on a per capita basis
2. Improve the commercial opportunities for the operators (drivers or related stakeholders in the value chain) to help improve their economic standing
3. Improve access to traditional public transport for commuters
4. Reduce the dependence of citizens on private vehicles (especially cars) for local travel use cases

Side note: First-principles thinking can lead us to a different set of design goals that doesn't include the second one above. (Think centrally managed city-scale fleets.) But it would be a short-term moonshot, in that it will require at least one drastic change to the status quo. That would in turn likely bring with it an unsustainable and net-negative social impact. To reiterate, a moonshot solution is not a bad thing in itself (stay tuned for a follow-up soon), but the present discussion is to demonstrate that we can pursue other, more immediate and sustainable solutions too.

Mobility is a psychological and economic activity at its core. From this perspective, any system designed to move people around efficiently and sustainably must coexist with many other systems by default:

• Urban planning and infrastructure: With city planning broadly leaning towards density than sprawl (for minimizing energy intensity, lowering infra costs, and boosting economic efficiency, among other things), unbounded growth in low-occupancy vehicles on the roads is definitely not tenable. We must develop solutions for utilizing the passenger capacity of vehicles to a greater extent, especially for mass transit.

• Utility development and Smart Cities: Using sensing technologies, advanced control systems, and ML/AI layers for making mobility smarter is part of a natural progression. The question is how to formulate the right problem, towards solving which these tools will be used. Some problems include remote management of fleets, co-planning passenger operations with vehicle operations (such as EV charging), downtime management, minimize the parking footprint during fleet downtime, and so on.

• Economic and commercial systems: On one hand, labour and employment markets influence commuting patterns and access to job markets, and on the other, the commuting patterns enable micro-entrepreneurship at the fleet or vehicle level. For large-scale mobility systems, aligning the right economic incentives in a stable way across the mobility value chain (which includes the government) is crucial, but such stability is easier imagined than achieved. Of late, there are new avenues for attaining strong distribution of choice and distribution of opportunity - new datasets that were previously unavailable, interoperable digital networks, and real-time decision-making capabilities.

• Governance and policy systems: In recent years, we have seen the sometimes collaborative, often aggressive push-and-pull between innovative solutions and policy. Be it through the introduction of ride-hailing systems, autonomous vehicles, or unified mobility payments systems, the importance of a trilateral friendship between innovation, enterprise, and policy has been underscored many times over. Of course, that is what D Dhanuraj is emphasizing in the context of paratransit systems through the article quoted above.

• Energy and environmental systems: We have long quantified the nature of the mobility sector's energy draw and the consequent contribution to emissions. Any solutions we seek to implement to meet our 2°C climate change restraints in the next 4-5 decades have a direct correlation with the energy system. Besides global climate change, issues like urban air quality are prime challenges too, which are usually exacerbated by urban transport. The current example at hand, which is setting expectations of what kind of solutions the future might hold, is the dynamic shift anticipated in the energy system (gigawatts of upgrades to renewable electricity production and the grid) to accommodate EVs. While there is no undermining the unstoppable force of collective human effort, it is equally important to be implementing simpler and effective solutions while making for the dynamic shifts.

Phew. That was a bit much, eh? Anyways, let's come back to urban mobility and paratransit systems.

Today, most of the paratransit fleet runs on IC engines and is not electrified. But on the flip side, it serves more individuals per vehicle daily than the private vehicle fleet. That is to say, the per capita emissions of the fleet are relatively lower. (Some nuances exist, sure.)

Here is a quick estimate of why this is the case:

### ESTIMATING CO? EMISSIONS PER KILOMETER

CO??per?km = [Emission?factor?(kg?CO?/unit)] / [Fuel efficiency (km/unit)]

# CNG Three-Wheeler:
CO??per?km = [2.7 kg?CO?/kg] / [32.5 km/kg] ≈ 0.083 kg?CO?/km

# Petrol Car:
CO??per?km = [2.31 kg?CO?/liter] / [16 km/liter] ≈ 0.144 kg?CO?/km

# Diesel Car:
CO??per?km = [2.68 kg?CO?/liter] / [20 km/liter] ≈ 0.134 kg?CO?/km



### AVERAGE CO? EMISSIONS PER CAPITA PER KILOMETER

# CNG Three-Wheeler:
CO? per capita per km ≈ [0.083 kg CO?/km] / [3 passengers] ≈ 0.0276 kg CO?/pax/km

# Petrol Car:
CO? per capita per km ≈ [0.144 kg CO?/km] / [1.5 passengers] ≈ 0.096 kg CO?/pax/km

# Diesel Car:
CO? per capita per km ≈ [0.134 kg CO?/km] / [1.5 passengers] ≈ 0.089 kg CO?/pax/km



### SUMMARY

CNG three-wheelers can be* up to ~3.47 times and ~3.22 times less polluting than petrol cars and diesel cars respectively.

(*A more detailed model will help characterize emissions more accurately over the vehicle usage type and the vehicle lifecycle)        

Taking things up a notch, paratransit vehicles like CNG three-wheelers often operate in a mix of reserved and shared modes. The activity-based transit modelling paradigm, which absorbs the purpose of individual mobility needs, such as shopping, education, or work trips, informs us calmly that this is to be expected. To minimize the cost of providing mobility and balancing convenience and comfort, both modes will naturally try to coexist.

In the context of the Motor Vehicles Act, these two modes are related to the contract carriage and stage carriage modes respectively. Currently however, regulatory approval for shared rides happens to live in a grey area.

Well, with the optimistic thought of our current paratransit fleet being more sustainable than we previously thought, let us turn to our beloved ol' microeconomics, for a more nuanced understanding of what impact this has.

The demand curve (demand for mobility services):

In the case of a mobility 'service', which is not a 'product' of course, we can imagine that the demand curve represents the overall demand of rides at a particular price, rather than the willingness-to-pay for a set of rides by an individual consumer. (A consumer doesn't buy more than one unit of a mobility service at a time, so it helps to look at aggregate demand instead.)

From unit economics and from common sense, we can say that the demand for reserved rides broadly represents higher willingness-to-pay (WTP) and demand for shared rides broadly represents lower willingness-to-pay. From transportation modelling, and yet again, common sense, we know that the utility that a consumer derives from their chosen mode of transport is dependent not just on the price of the ride, but also the time that it takes to reach the destination using the chosen mode. (Going into more detail, we have may have several other parameters such as comfort, safety, waiting time, and so on, but we will stick to price and time for simplicity).

Therefore, the interesting balance which is up to us to determine ("us" = market + government) is the balance between solutions that the market provides to consumers for urban last-mile or point-to-point mobility, i.e., reserved rides and shared rides. In other words, how much of the aggregate demand for mobility services should be served by reserved and shared rides respectively.

### A (VERY) SIMPLIFIED MODEL

# A utility function is the mathematical representation of the economic utility derived from the use of a product or service at a certain price.

# Assuming the utility function to be quadratic: the linear coefficient represents the marginal utility of convenience and price from a particular number of people using the transport mode, and the quadratic coefficient represents diminishing returns from more people using the transport mode.

# Utility from reserved ride
U_r ( x_r ) = a_r * x_r  -  b_r * x_r^2

# Utility from shared ride
U_s ( x_s ) = a_s * x_s  -  b_s * x_s^2


# At a fundamental level, we seek to maximize the total utility across all consumers, i.e., all units of demand.

maximize [ U_r ( x_r ) + U_s ( x_s ) ]
subject to
x_r + x_s = D (total demand)
x_r ≥ 0
x_s ≥ 0


### EXAMPLE
U_r ( x_r ) = 20 * x_r  -  0.05 * x_r^2
U_s ( x_s ) = 30 * x_s  -  0.1 * x_s^2
D = 100

maximize [ 20 * x_r - 0.05 * x_r^2  +  30 * x_s - 0.1 * x_s^2 ]
subject to
x_r + x_s = 100
x_r ≥ 0
x_s ≥ 0


# Under these specific conditions in this limited example, the approximately optimal combination would be to serve 66.67% of the aggregate demand (of a city, for instance) via shared rides, and 33.33% via reserved rides, so as to maximize the utility.        

In the market, the two forces that are primarily responsible for determining the optimal combination are:

1. Available innovations, i.e., the types of mobility services with differentiated features and value propositions available to consumers, and
2. Policy interventions such as price caps or promotion of certain service types, i.e., the artificial pressure, whether positive or negative, on specific types of supply.

As we see in the market, there is no dearth of available innovations to improve utility in shared rides. However, policy is yet to catch up, and the silence around or prohibition of shared rides artificially reduces the legal supply of low-cost ride options, pushing shared rides to operate in informal arrangements.

Overall market demand for urban mobility:

As the urban population grows, the overall demand for mobility services of all types is increasing. This is akin to the demand curve slowly but surely shifting to the right, but this is a long-term movement. (A short-term shift, on the other hand, would be something like the odd-even rule for passenger cars, where the demand already exists, but is temporarily diverted to a different mobility mode, say, public transport.)

The cost of substitute products such as public transport are more affordable, but due to poor first- and last-mile connectivity, the accessibility of these services is low, leading to a net lower utility derived by commuters. On the other hand, in recent times, the prices of other substitutes like ride-hailing were artificially reduced by market players flush with cash in a bid to increase the portion of demand served by this mode and report higher utility from ride-hailing. Naturally, consumers started preferring cabs over auto-rickshaws in hordes. (Certainly, this was not a move for profitability as much as it was a move for cementing competitive positions, and this artificial effect can be expected to subside at some point.)

The adverse price impact and social impact of under-enforcement:

Regulation of paratransit is a state issue, but enforcement must happen at the city level where the paratransit fleet operates. Owing to a lack of data-driven processes and political reasons, the fleet size isn't curated to suit the needs of the city. In most cities, this has led to an over-supply of drivers and vehicles in the short-term. As demand is growing slowly as compared to the over-supplied fleet, this has in turn led to lower earnings per driver on average. Like I said before, mobility is a psychological and economic activity at its core. With earnings dipping, drivers are driven to sidestep the system and act out, hiking fares above legal limits, reducing service coverage, and undermining service quality. This leads to shifting the demand curve to the left, that is, lower demand due to lower trust in the system.

The changing nature of information asymmetry:

With upcoming technologies for mobility, such as real-time demand discovery, paratransit and mobility data, multimodal trip planning, and open networks, it seems like the sunset days of information asymmetry are almost here. An imbalance in information - about demand, supply, or market factors - exists between drivers and commuters, between the fleet and the city (transport department, planners, public transport operators, etc.), and between the commuter and the city. In a world of perfect and real-time information, it becomes theoretically possible to operate the full product mix - public transport, paratransit, ride-hailing, micro-mobility, and private vehicles - in an optimal manner, providing the required service to each consumer with precision. While this may not be possible in real life, this thought sets the tone of what each actor (driver, commuter, and city) can and should aspire to achieve as information asymmetry reduces.

This brings me, finally, to the central questions that beset us as we plan better urban mobility operations:

Why should we overlook the inherent advantages of a naturally-evolved system like IPT instead of deploying it for sustainability and economic impact?

Why should we not create systems that leverage the emergent properties of this complex system, and guide it towards optimal operation?

Let us revisit what we are dealing with, shall we?

We have a large fleet of small-capacity vehicles (auto-rickshaws, minibuses, vans, etc.) that has inherent flexibility, which today can be empowered with information balance, which is already replete with many innovations in service design, and is hopefully empowered by policy support in the near term.

I want to show you a glimpse of the future of paratransit and the extremely important role it will play for urban mobility:

• Greater availability of reliable and accessible rides: Deploying shared rides for standalone city-scale operations or for last-mile connectivity for public transport can increase the overall demand for mobility services that intrinsically have a lower carbon footprint. Moreover, with a data-driven approach, a city-scale fleet can be operated in an on-demand manner that makes rides more accessible for all citizens, while also increasing service quality and reliability.

• Higher driver earnings and a balanced fleet: With IPT modes and solutions better recognized under the policy regime, drivers stand to benefit from higher earnings via better demand visibility from smart apps and from higher commuter demand owing to new ride modes. With cutting-edge solutions on the horizon for providing real-time visibility of demand to drivers without the need for a platform model, reducing dead mileage and wasted time becomes easy. This way, cities can end up with more balanced fleets that operate street-hail at standard fare rates (platform-less model) as well as ride-hail (platform model or open network model).

• Determining the optimal combination of reserved and shared rides: The toy example we worked out above used dummy data. However, if granular data about paratransit operations is available, market players can begin to take better decisions in provisioning services, and long-term policies can be designed as a measured response, rather than by speculation.

• Fleet size and operational management: With better access to paratransit data and deeper integration with city IT systems, it will be possible to natively manage the city fleet, including determining the optimal fleet size, driver monitoring, commuter safety monitoring, over-the-air updates for applicable fares, reporting of sustainability metrics, managing parking spots for the entire paratransit fleet, and much more. Safety is one of the most important aspects that policy around shared mobility is concerned with. It is important to push for safety through innovation and action, rather than through restrictions and limitations.

• Better service modes and vehicles: To stay relevant and up-to-date with the innovation in service design, new service modes and vehicle types will come up, contributing greatly to service quality and consumer choice. The introduction of EVs into the fleet, requiring methods for managing the optimal operation with respect to vehicle downtime for charging, would be another avenue for innovation to shine. Vehicle autonomy is already a reality, and a path to implementation will soon be charted.

I imagine a future for our cities where point-to-point and shared mobility are central to the emergent flow of human capital, keeping cities clean, energy-efficient, and sustainable. To achieve this vision, the strong push for innovation that already exists should be coupled with perceptive support from policy. In the case of paratransit and shared mobility, it is literally "build, and they will come".

The frugal does have tremendous latent energy.

At Yuja we are working on technologies and products that help make this vision a reality.

??????