The Design of Flat-Rate Pricing in the Ride-Hailing Market
The Question: Upfront Flat Fares or the Meter?
Ride-hailing platforms price trips in two common ways:
- Metered, pay per use: the meter runs in real time, so the final fare depends on the actual route and trip conditions.
- Upfront flat fares: an algorithm fixes the price at booking, regardless of the route or trip conditions.
The choice is really about who bears the risk of trip uncertainty. Under the meter, the rider absorbs a traffic jam or a detour. Under a flat fare, the driver does.
The industry has not reached a consensus. Uber has priced rides upfront for riders in the US since 2016. China uses a mix of both. In Hong Kong and Germany, Uber offers a metered taxi product alongside upfront-priced UberX.
We ask: how do riders and drivers respond to flat-fare versus metered pricing when trips are uncertain, and how should the flat-fare option be priced?
The Setting: A Natural Experiment in Singapore
On April 10, 2017, a leading taxi operator in Singapore introduced an origin–destination flat-fare option for app bookings, alongside its existing metered street-hail and metered app-booking options.
We use the operator’s administrative trip and booking records from December 2016 to March 2018. A regression-discontinuity-in-time design compares outcomes just before and just after the launch date.
Early Findings
1. Driver earnings fell
At the launch, drivers’ daily earnings dropped by S$7.9 (5.5%) per driver-day. Drivers completed fewer trips (−5.0%) and spent less time carrying passengers (−5.3%).
2. Drivers did not work less; they earned less per hour
Drivers did not stop working or cut their hours. Instead, the drop came within the shift: hourly earnings fell by 4.6%, and the idle gap between trips grew by about 1.9 minutes (+7.9%).
3. Flat-fare requests failed more often
Flat-fare booking requests were more likely to go unfulfilled than metered requests on the same route. This gap narrows as the flat-fare premium over the expected metered fare rises. Extrapolating, the excess failure rate disappears at roughly S$3 above the average flat-fare premium at the time.
Next Step: Pricing the Risk
We are building a two-sided dynamic structural model in which risk aversion is estimated rather than assumed:
- Drivers make forward-looking decisions about accepting, rejecting, and relocating, valuing each fare net of a penalty for trip-duration uncertainty.
- Riders choose among the flat fare, the metered taxi, and outside options such as public transit, facing risk over the final cost and waiting time.
With the model, we will find the optimal flat-fare premium: the markup over the expected metered fare that compensates drivers for the risk they take on and clears the market.
Takeaway
Upfront flat-fare pricing is a risk-allocation choice, not just a convenience feature. When the flat fare is set inappropriately, it silently taxes the side that bears the risk: here, driver earnings fell and flat-fare requests failed more often. The key design lever is the premium, that is, how much to mark up the flat fare over the expected metered fare.