Bidding to Catch a Ride
Most of us are familiar with Uber and Lyft. When needed, you open your app, request a ride, and arrive at the destination you desire. The price of the ride: varies. Some days at certain times the price is very cheap, other times it is twice or thrice what you would pay, commonly known as surge pricing. You, as a user of Uber or Lyft, have no control over the pricing, depending on demand the pricing is calculated automatically. However, Didi Dache, a Chinese ride sharing program, handled price surging in a different way (before being ceased by the government).
On Didi Dache, instead of prices being automatically adjusted based on demand, riders can bid tips in fixed increments on top of the base price of the ride. This method of implementing auction theory and users bidding for rides was intended to create to an optimal matching of suppliers and demanders.
Payoffs for the Driver
As seen in this table, the driver can either accept the passenger with the current bid, or wait for a higher bid but risk the chance of not getting anymore passengers.
Payoffs for the Rider
As seen in this table, the rider can either increase the bid to their own limit or not give any and risk the chance of getting no driver, or increase their bid to the maximum bid and pay more than they wanted to.
Both these tables tell us that there is a decision to be made, both on the driver and the rider side, that can greatly effect their payoffs. This process, controlled by both the drivers and the riders, was designed to have the market clear out in an organic way. The more one wanted to pay for a ride, the more likely they would get a ride. Although this model may seen like it would be efficient, as it allows the riders and drivers to control the price, there were several weaknesses. The main one being that this model incentives collusion. The drivers can all come together and say that they will only accept bids higher than a certain amount, creating an false price surge and demand. On the other hand, the riders can all come together and say that they are not going to bid over a certain amount or at all. Either the collusion of drivers or riders would cause this process to not work as designed. And because this is the real world, rider and drivers are going to collude, and the process will fail to work as it is supposed to. So in theory Didi Dache’s model was good, in actuality it did not work out well.
Link: https://medium.com/@jensenloke/alternative-auction-model-in-ride-sharing-platforms-4d2a0e951cb2 (Images of tables taken from the website)