Negative Congestion Charging via Vehicle Occupancy 'Arbitrage'

A tradable approach for real world transport providers, municipalities and traffic light management using real time data and artificial intelligence

An Introduction July 2024

We call this Vehicle Occupancy Arbitrage because a market participant can 'buy low and sell high' by identifying trip O-D pairs where it would make sense to offer lower prices for more customers (2 - 3) per vehicle.

This is precisely the definition of an arbitrage opportunity, and it is very lucrative if it can be done consistently for specific O-D pairs in a locale.

The ultimate determinant of success is marketing and the rate determining step of uptake comes as a function of careful marketing and customer perception management.

In the first instance, once people can save 50% of daily costs on the very short trips, where sharing with others is no real burden if there is no deterioration in trip quality (especially during daylight hours), the shift will begin

Customers can be locked in with longer term contracts (100, 200 trip packages for example) and later, be able to sell a proportion of a package back if not needed (a fully fungible product)

The marketing begins with identifying customers on fairly regular routes and reducing their fares by 20%-80% immediately. Once captured, more options can be offered.

This becomes an effective dynamic on-demand bus route.

For larger capacity vehicles, the potential profits are even higher. This all depends on having enough liquidity available to sustain a viable model for 18 months.




A Tradable Relative Value Approach

• · Most cars are used only 1 hour per day and for 90% of trips undertaken carry only one occupant: the driver. Occupancy and default risk on operating costs are unambiguously linked as the risk of an operator not making enough money to survive is the risk of occupancy falling to zero .

Thus, the simple expedient of increasing average occupancy offers the chance to reduce costs for the customers while simultaneously boosting overall income for the operator. The total number of vehicles required for a certain person-flow rate is therefore inversely proportional to the occupancy rate. In order for the overall customer experience to be maintained wait times cannot be more than 2 15 minutes. As more customers shift to higher occupancy rides, the fewer vehicles on the road make for more predictable timings.

• · Both occupancy and wait time directly tradable as derivatives of cost and expected revenue in the marketplace - trip default swaps and occupancy puts for instance . The close relationship between occupancy level and price means the marketplace has a large arbitrage opportunity if we can search for highly fecund O-D pairs with high frequencies and relative low distances (<7km) or those longer trips that cost so much that even a 20% reduction in price is a significant incentive for customers to change behaviour
• · Our approach consists of searching the universe of O-D Pairs in order to identify tradable O-D pairs where either the frequency of trips make grouping highly likely or longer distances where the significant cost saving spurs behaviour change. Because people are averse to sharing without safeguards, a social network acts as a catalyst to this behaviour change.



Double Occupancy, Half Price: Trade Mechanics

Searching for cases where the total number of trips for a specific origin-destination (O-D pair) is above an arbitrary minimum and the total travel time is relatively low, prices can be reduced dramatically and revenues for the operator boosted dramatically.

• · Calculate the present value of income for every O-D pair from historic data and fixed cost per mile (£1.60/mile, £2.56 / km) over the duration of dataset . this approach is used to address any potential timing mismatches between customers (to make ridesharing happen we must solve the coincidence of wants), which most likely would wait up to 5
• minutes to get a ride at half the price anyway
• · Determine the potential extra revenue if a certain occupancy threshold is reached and the prices are reduced by a certain amount (Occupancy 2.0, Prices 0.5z)
• · Make an assumption about what % of customers will take up the shared rides and what the trip default rate would be when sharing is the only option (this gives us an average price per ride and allows us to calculate the effective hourly revenue for drivers)
• · Determine the amount of money (liquidity) needed to hedge this estimated trip default risk , i.e. TDS Liability.
• · Determine the cost of the liquidity based on expected default rate (only one-person riding) for an arbitrary time window.



Two decision variables :

Do we choose high cost, relatively high frequency O-D Pairs or ultra-high frequency, low-cost trips) which are over 90% of all the trips in the dataset)

What recovery value (liquidity cost / stuff up) equates to the loss for the cost of one traveller only in the vehicle at the reduced price?

The key point is a maximisation of revenue output coupled with a minimisation in price for customer and number of vehicles used per people flow.

Each O-D- pair has a naturally observed flow pattern.



Freq: 500




Neales Trip Analysis

High Wycombe; Bucks

comeins

We

mappins

National

European Electoral Regions




Migration Analysis and Summary

The data was collated for trips taken from January 2014 - December 2017. The total dataset comprised 1.4m items and covered trips from across the UK, emanating from the BUCKS headquarters.

Earnings were calculated simply by using either the direct geodesic or road distance and the acknowledged price per mile / kilometre.

Top 3 O-D Pairs:

HP13 6TQ - HP13 7YH (1814, 2.14km) HP13 6TQ - HP13 7LU (1526, 1.83km) HP13 6TQ - HP13 7HL (1397, 2.51km)

Top Origins: HP13 6TQ (165,283) HP13 6NN (60,391) HP11 2DQ (24,862)

Top Destinations:

HP13 6NN (101,179)

Highest Earning:

Most clustered trips O-D Pair:

Mean radius of top 95% of trips:

7km

Marketing Assumption 1

Flat fare of £3.20 for trips below 7km 50% Price Reduction if over 7km

Marketing Assumption 4

Flat fare of £3.20 for trips below 7km 50% Price Reduction if over 7km

Marketing Assumption 2

Flat fare of £3.20 for trips below 7km 50% Price Reduction if over 7km

Marketing Assumption 5

Flat fare of £2.50 for trips below 5km 50% Price Reduction if over 7km

Marketing Assumption 3

Flat fare of £3.20 for trips below 7km 80% Price Reduction if over 20km and more than 100 occurrences

Marketing Assumption 6

Flat fare of £2.00 for trips below 2.5km 40% Price Reduction if over 7km




Data Treatment

We have had to create certain procedures and patterns to analyse and model the data. For the departure times, we must look at the clustering, although historical patterns are of limited use since behaviours will change once we spend on marketing.

Signup Strategy

• · Identify the top 1,000 customers by
• 1) frequency of trip
• 2) total spend
• · Send an SMS message to those customers to identify if the number is still active and request consent for further marketing, via having them respond with an email address
• · Allow users to book by SMS in addition to the app
• · Collect all app searches and requests for service by a webhook to populate the analysis database
• · Run live Demand Mapping on the incoming trips to further preposition vehicles.



Negative Congestion Charge Example

If a council / municipal government levies a congestion charge during rush hours, then the economic calculation to use a single occupancy vehicle will change. Every vehicle must be charged, but exemptions are dealt with via reimbursement.

Such a scheme charges all for full use of the roadspacetime for all class of vehicle, regardless of fuel source and is the only fair and rational way to do so. All roadspacetime purchase data is anonymous and secured via Zero Knowledge Protocols (with salts and peppers).

Hypothecation:

Funds raised from the management of the Transit Exchange Platform and fees paid by the participating market-makers are used only for payments to citizens and maintenance of the road infrastructure.

Liquidity:

When a vehicle does not have a full load of passengers for its complement of seats, there is a need to allow the vehicle to depart with a less-than-optimal load factor. The term to describe this is market making where the empty seats are financed by an entity other than a customer. This entity is called a marketmaker and they exist in all exchanges.

Liquidity Provision Measure Calculation:

We can estimate how much 'float' is required for a fleet of a certain size, based on the expected default rate (not 100% of the vehicle filled) per hour.



Trip Default Insurance:

Trip insurance is something that the market makers must offer buyers, so that someone who turned down a single trip by themselves will have a guarantee they will be able to complete a trip within a certain time radius. Of course, higher availability of vehicles and fewer vehicles on the road at busy times due to the congestion charge will lead to lower default rates.

Similarly, vehicle operators need to be able to lock in certainty in what they will make, and they can do this in the same way as hotel booking sites allow hotels to lock in a certain minimum.

Using the merchant model, prevents market abuse.

Negative congestion charge payback

As its positive counterpart, this will be variable based on demand. All registered residents with cars living within a certain radius of the congestion zone will be eligible for payments. There will be other stipulations, like minimum miles/hours driven per year, residency status and vehicle excise duty paying status.

Present Value of Day Ahead Future Trips:

We can estimate what return an investment into the capital buffers will be for an investor, and target a 20 basis point daily return. The whole aim is to provide investors with an investment opportunity (like a money market fund with high liquidity) that is transparent and unable to be easily manipulated



Scenario 1.1

Occupancy >2.0, 30% price, medium distances, max wait time: 600 seconds

Scenario 1.2

Occupancy >2.0, 50% price, long distances, max wait time: 600 seconds

Scenario 1.3

Occupancy >2.4, 20% price, max wait time: 6000 seconds

Decision Variable 1:

Assume 50% occupancy, no more than 300 seconds variability from planned departure time

Decision Variable 2:

PV of the CDS premia equals the cost of a put




The Alchemy of Making Road Pricing into something Desirable

Q: At the core there is a booking system, or as you put it equalizing bank system. I do not get that completely in practice but will work further on my own assumptions.

A: There are several booking systems. TEXXI is both a Business-To-Government (B2G), Businessto-Business (B2B) and Business-to-Consumer (B2C) system. The very first iteration required us to do the B2C as well as the B2B and B2G simply because there were no other ridehailers in 2005 2008. The entire boom came about because I proved something thought impossible was indeed possible.

The topmost (B2G) layer booking system is for the use of the road spectrum. In the original realisation, roadspacectime "operators" bid with one another to buy road capacity in order to resell it to users. The fee goes to the Government which manages this as a "Commons". This is, incidentally, the model the UK virtual electricity market currently employs, where regional electricity companies (REC) bid against one another every morning for generated electrical energy capacity from the myriad power producers (GAS, COAL, NUCLEAR, HYDRO, SOLAR and STORAGE). The RECs then sell their allocations to customers (both large corporations and smaller companies and users) via a virtual market, permitting anyone anywhere to choose the cheapest price / best tariff for their particular circumstances even if technically their closest physical provider delivers the electricity. The price the user pays comes from the virtual market, using the free market to force competition and efficiency on to what would otherwise be moribund captive markets.

The B2B layer is where a large reseller sells capacity onto a smaller reseller, who then sells it to a particular consumer. A Coach or Taxi operator would buy capacity from a large reseller in order to




then operate their vehicles, whose utility and profitability is determined by selling seats to people who need transport. This is a direct corollary of an airline buying access to an airport, then selling seats on its aircraft.

The B2C layer is where "Joe/Jane the Citizen" interacts, buying a slot on the road in order to drive their private vehicle into an area (e.g. the Central Business District). Parking Prices are a version of this. This is a direct corollary of a person buying a takeoff slot for a private aircraft or a mooring slot for a boat on a waterway.




Contact Information

Occupancy Arbitrage & Trip Insurance Strategy

Eric Masaba +44 20-7993-2324 / TEXXI GLOBAL, TEXXI BUCKS (TRANSIT EXCHANGE XXI FRANCHISE 01)

Vehicle Operations

Abrar Hussain +44 1494 372 860 / NEALES TRANSPORT Azfar Rashid +92 333 5511235 / NEALES TRANSPORT

Marketing Operations MAMAER DIGITAL +44 20 7183 1663

Licence Sales TEXXI BUCKS +44 20 7183 1754

Analyst Certification

I, Eric Masaba, hereby certify that the views expressed in this research report accurately reflect my personal views about the subject.