The sharing economy: A futuristic taxi landscape (Part 3 – Pricing fares the same way as stocks)
Phuah Eng Chye (27 January 2018)
Price regulation is often the biggest casualty in an industry disrupted by information. This is reflected by the experiences in the financial services and airlines industries. However, unlike other industries where prices were decisively liberalised, this is not the case in the taxi industry where the regulation-deregulation debate seems caught in the perpetual swing of the regulatory pendulum.
In a well-known example in the 1970s, ten U.S. cities deregulated by removing limits on taxi licenses and controls over meter-rates and relaxing regulations such as extra vehicle inspections and driver testing. The results were mixed. “Although the supply of taxis expanded dramatically, fares often went up instead of down, and total cab usage often went down, which reduced incomes for companies and drivers. Long cab lines usually emerged at major sites like airports, frustrating drivers. Most re-regulated. Six of the ten also restored limits to the number of taxis.”
This example, often cited as a justification against deregulation, demonstrates that price regulation has its advantages. First, metered rates can address information asymmetry by providing a reference point for fair and pre-negotiated prices. Second, price regulation complements licensing restrictions to ensure orderly conditions and to protect income and value in the taxi industry. Indeed, the debate on fares is often intimately linked to income issues.
But the persistence of underserved areas and crowding in popular routes such as airports reflect the inadequacy of fare regulation in resolving demand and supply imbalances. In this context, there doesn’t appear to be a satisfactory theoretical framework for setting taxi fares to resolve these perennial problems.
In its review of the Tasmanian fare-setting mechanism, the Centre for International Economics noted the lack of clear objectives and processes for fare setting and changes in fares were irregular and mainly “in response to demands by industry”. It noted that “fares are currently increased using a taxi cost index, with no consideration of the level and structure of fares…The construction of the cost index has overweighted items that are included, because it excludes major cost components (labour and plate leases). This has led to a far higher weight on items such as fuel.”
Thus, price regulation is blunt and inflexible. It treats all distances alike whereas not all kms are equal. For example, some routes are more congested at certain times. In addition, price regulation can’t take into account variations arising from the likelihood of a backhaul customer ride or constant fluctuations in fuel costs. In particular, price can also be affected by personal factors such as the type of car, the individual’s perception on his income need, his time costs or because the pick-up or drop-off is on the way. Price regulation involves standardisation which implies cross-subsidies between profitable and unprofitable routes.
In the past, there was little alternative to fare regulation to address information asymmetry. But this is no longer true. Mobile connectivity and smart algorithms now makes it possible to offer dynamic pricing precision. While ridesharing platforms have arbitraged the taxi regulations to offer variable pricing, their dynamic pricing strategies either have competitive or profit objectives. One example is the use of subsidies to win market share at the expense of traditional companies.
The other is the controversy generated by surge pricing. Hubert Horan criticises Uber’s surge pricing in that it “simply raises fares (up to eight times normal levels) without prior warning. Given the short notice this does nothing to increase total taxi supply, but merely redistributes drivers to higher fare areas…Uber’s surge pricing does not increase efficiency; it simply prices taxis out of the reach of many current users, reducing both total taxi demand and overall economic welfare.” Surge pricing, at its worst, is viewed suspiciously as a form of price gouging – an attempt to maximise profits given supply shortages.
In this context, Hubert Horan believes there are severe limits to using on-demand pricing techniques to optimise driver assignments to overcome supply bottlenecks. Hubert Horan points out “research has long demonstrated that the timing of taxi demand is highly inelastic, (people want a cab at a very specific time) so variable fares will not change demand patterns, improve taxi utilization or increase total revenue. All forms of urban transport have similarly inelastic demand; the Long Island Rail Road has had peak/off-peak pricing for a hundred years but rush hour is still rush hour. No level of taxi discount will get anyone to shift their Saturday night plans to midday Tuesday.” Hence, “the instant gratification that on-demand services are supposed to provide make all these costs and challenges worse”.
The pricing algorithms adopted by the ridesharing platforms appears similar to the yield management techniques used by airlines but with two differences. First, the fares for the same flight are adjusted over a long period based on take-up and how demand elasticity changes as the departure date draws near. In taxis, immediacy and waiting time for a ride are critical.
Second, “markets” where one party has control over the information and is a dominant price setter are different from “markets” where prices are discovered through a quote-driven auction process. In the price-setting system, the demand-driven pricing algorithms of sharing platforms are based on strategies to achieve competitive goals such as market share or profit maximisation. But this may be at the expense of public goals.
In this context, technological advances now make it possible to implement the latter type of “market”. I have thus attempted to sketch a futuristic scenario where taxi fares are discovered the same way as stocks based on a willing buyer-willing seller negotiated process. The objective of using an auction process is to allow taxi drivers and customers to discover the fares to correct the imbalances between demand and supply and to improve the allocative efficiency of taxi resources. For example, the crowding of taxi drivers at airports suggests the fare is too high and should be lowered. Similarly, the fares at underserved neighbourhoods need to rise to attract supply.
But would participants have confidence if fares are constantly fluctuating by route and by time? Volatile fares lose the benefit of stable reference points which is needed to underpin confidence. In addition, higher fares cannot solve the problem of immediacy of supply. However, the uncertainties can be lessened if price discovery operates within a framework of long-term and repeated price patterns established by consensus between well-informed drivers and customers over a period of time.
While there are practical limits to solving immediacy, nonetheless prices can signal where persistent crowding and shortfalls occur and provide a mutual basis for more precise price adjustments either upwards or downwards. One interesting point is that prices would logically be higher in congested areas. The question is whether the benefits of the higher fares should accrue to the driver or to the public authorities in the form of a congestion charge?
Confidence can be reinforced in two ways. One is through information disclosures (of historical fares) to ensure a well-informed bargaining process (such as pre-and post-trade reporting in the case of markets). The other is through regulatory discipline. The regulators establish the behavioural norms and intervenes when conditions are disorderly and when they detect unfair, abusive, manipulative or predatory practices. Hence, prices play a role of signalling where there are supply excesses and shortfalls.
As a matter of comparison with the commercial platforms, the surge and yield management models operate without public obligations such as fairness or ensuring efficient allocation of taxi coverage. The shortcomings are exposed during when they attempt to subsidise fares to gain market share or raise fares to maximise profits. In my view, the best safeguards against monopolistic behaviour is to minimise entry barriers and to ensure that information is transparent and freely shared.
As part of this futuristic experiment, I propose the design of an app called “Taxifair” to demonstrate how a market-pricing system could work. The app would be an electronic form of haggling but with nudges to bring bargaining to a quick resolution.
Step 1: Passenger indicates a pick-up and a drop-off point and other requirements such as luggage, no of passengers or child seats.
Step 2: The app will suggest a default fair fare for a default route. The default fare is calculated based on past fares during the same time slot for a similar trip; adjusted for distance, time and fuel costs. The app could then prompt as to whether the passenger wished to identify cheaper or faster alternatives using different nearby drop-off and pick-up points. The passenger can then choose the default fare or to bargain by entering a lower fare (within an allowed price limit).
Step 3: The app will flash the request to the five closest taxis (by estimated pick-up time; the equivalent of order depth in markets). Driver will see the full details of the request (including whether passenger has chosen the default fare or to bargain) as well as his share of the fare, the time to pick-up point and to destination. In addition, the app will also flag if the passenger has a low reliability score (tends to frequently cancel requests, is late to the pick-up or a no-show). Driver can accept or offer a higher or lower fare.
Step 4: Passenger will see details of taxis that accepted; such as estimated time to pick-up, car model, driver reliability scores and the accepted or offered fare from the driver. Passenger can then select the driver to confirm the ride.
This is a simplified version of how fares can be determined based on a willing buyer-willing seller basis. There is substantial room for calibration and customisation. For example, the screen for passengers and drivers can be based on personalised filters (driver rankings, car, lowest fares) or pre-set preferences (routes, pooling). Fare adjustments can be made if there was an unexpected event – an accident causing a lengthy traffic jam and the cost can maybe be shared by the driver and the customer.
The app can also provide useful information to assist trip planning. There can a listing of last-done fares by route and by time for reference by passengers and drivers to assist price discovery. Passengers can save a list of favourite drivers for which he can contact to make trip pre-arrangements. Pooling transport can also be highlighted which will provide the comparative pick-up time and fares. Tourist surcharges can also be imposed to incentivise active coverage of popular locations. When passengers travel to underserved neighbourhoods, the app can indicate the expected time to order a taxi and the default fare for the return trip.
Overall, this analysis highlights the need to revisit price theory. At the moment, price theory is based on a paradigm of optimising welfare. This may be a practical and fair approach for a low-information environment where price signalling is a relatively primitive tool. But welfare is an ambiguous concept which assumes prices can be used to determine the prioritisation of the diverse objectives of private participants and to ignore public good objectives. This is also often confused with achieving economic efficiency.
These constraints don’t apply in a high-information environment where the speed, quantity and quality of information is substantially higher while the costs of relaying information are extremely low. Higher information transparency expands the dimensions of price signalling. in this context, different players would have their own private objectives in relation to their price strategies. The point is that price mechanisms (including the distribution of information) are not neutral especially when there is concentration or domination; and may favour select parties.
In this context, there are opportunities to design price mechanisms to incorporate public objectives such as economic efficiency, fairness, choice or opportunities. Towards this end, information transparency has an important role in ensuring confidence in the fairness of transactions, improving the quality of participation through (in the case of sharing) and ensuring the genuineness of activities.
Hara Associates (10 October 2015) “Taxi Economics – Old and New”. City of Ottawa Taxi and Limousine Regulations and Service Review. http://documents.ottawa.ca/sites/documents.ottawa.ca/files/documents/otlrsr_taxi_economics_en.pdf
Hubert Horan (Nov-Dec 2016) “Can Uber ever deliver?” Part One to Five. Nakedcapitalism.com
Phuah Eng Chye (13 January 2018) “The sharing economy: A futuristic taxi landscape (Part 1 – The future and the uncompromising economics of the past)”. Economicsofinformationsociety.com. http://economicsofinformationsociety.com/the-sharing-economy-a-futuristic-taxi-landscape-part-1-the-future-and-the-uncompromising-economics-of-the-past/
The Centre for International Economics (Draft report – January 2013) “Setting taxi fares in Tasmania”. Prepared for Office of the Tasmanian Economic Regulator. http://www.economicregulator.tas.gov.au/Documents/13541_CIE_Draft_Report_Taxi_Fare_Methodology_Inquiry_15_February_2013.PDF
 Hara Associates “Taxi economics – Old and new”.
 This is partly addressed through add-on time-based charges.
 Pre-trade transparency refers to the publication of current bid and offer prices and quantities. Post-trade transparency refers to the publication of details of successful transactions.