Why is there dynamic surge pricing *for rides*?

“Surge Pricing Solves the Wild Goose Chase” is the title of the new paper by Juan Camillo Castillo and E. Glen Weyl, here is the abstract:

Why is dynamic pricing more prevalent in ride-hailing apps than movies and restaurants? Arnott (1996) observed that an over-burdened taxi dispatch system may be forced to send cars on a wild goose chase to pick up distant customers when few taxis are free. These chases occupy taxis and reduce earnings, effectively removing cars from the road and exacerbating the problem. While Arnott dismissed this outcome as a Pareto-dominated equilibrium, we show that when prices are too low relative to demand it is the unique equilibrium of a system that uses a first-dispatch protocol (as many ride-hailing services have committed to). This effect dominates more traditional price theoretic considerations and implies that welfare and profits fall dramatically as price falls below a certain threshold and then decline only gradually move in price above this point. A platform forced to charge uniform prices over time will therefore have to set very high prices to avoid catastrophic chases. Dynamic “surge pricing” can avoid these high prices while maintaining system functioning when demand is high. We show that pooling can complicate and exacerbate these problems.

Perhaps it is an analogy to suggest movie theaters might use more surge pricing if a low valuation buyer took up the seat for several showings of the movie rather than just one.


To some extent, we already have surge pricing in cinemas, since early shows and matinees are generally cheaper than evening shows.

Plus dvds, streaming, Ota movies,....

Interesting study. Weird analogy. If the business relationships between theaters and their many distributors weren't costly to change in a coordinated fashion we'd probably have more price discrimination than the matinee, senior, student and children discounts. A few more bad analogies below:

Perhaps if ridesharing companies made a significant portion of their revenue from selling concessions to captive audiences they would have less incentive to implement surge pricing to keep capacity below 100%.

Perhaps theaters would use surge pricing if a theater could create more seats in which they would have a revenue share when the prices for showings went up.

Why? Because the market dynamics are different. Surge pricing in ride sharing operations is a way of increasing supply. A restaurant or theatre can't increase supply by surge pricing. What they will do is decrease demand.

A more interesting question is whether a line up outside a restaurant or theatre increases or decreases demand.

As mentioned, these businesses do a certain amount of surge pricing. Different menus for lunch and dinner, service companies work to control the peaks and fill the slow times.

There is also exigencies of demand. Surges in demand for rides usually are driven by other events; weather, everyone wanting a ride home after a concert, etc. To be able to offer the service at all is valuable, so increasing supply is the value added of the service. If prices for a movie go up, or restaurants, they can be rescheduled or simply skipped.

The problem with taxi services is that the investment is in the permitting and buying of a monopoly. The value added is in pleasing a regulator or commission with a large fee, plus a few donations to politicians. Giving rides, being effective at meeting demand for the services is secondary.

A restaurant or theatre can’t increase supply by surge pricing. What they will do is decrease demand.

And worse than that, they would have to advertise heavily and instantly to get the reduced price recognized in the market to make the demand increased. This is a very silly comparison.

I agree with both of these comments.

A price increase by a movie theater won't create more seats.

Stelios tried to set up exactly this in the UK after easyjet. Easycinema or some such name? Think it still exists but it's not setting the world alight.

This reminds me of the Barro Grossman 1971 supply-side multiplier. An initial excess demand causes a reduction in supply, worsening the original fall in the volume of output. Households supply less labour because they are rationed in output markets so cannot spend any extra wages they earn. Another way to think of this is that people waste their time queuing, rather than working to produce goods. This is probably a part of what is happening in Venezuela now.

I was going to mention Barro as well, but not Barro-Grossman, instead the Barro-Romer 1987 AER article about why ski resorts do not use dynamic surge pricing. Congestion and imperfect information were key attributes in the model, which might be applicable to the market for taxis, pre Uber at any rate.

But maybe the model doesn't apply to taxis; Tyler has to be well aware of the Barro-Romer article because he co-authored a comment on it:

Yet another possibility: plain old inefficiency and irrationality. For decades economists wondered why sports teams did not charge higher prices for games that had higher demand. I think it is not a coincidence that teams finally started doing this around the same time that they started using more rigorous analytics a la "Moneyball".

Sadly, dynamic pricing and ticket re-sellers like StubHub have put the hurt on, or maybe even killed off, America's greatest capitalist entrepreneurs, ticket scalpers.

Tyler: maybe the movie version of surge pricing is making sure everyone leaves the theater when the movie is over. That way seats can't be occupied by the same person over and over again. It isn't a monetary change in price due to demand, but in some sense your overall price is higher if you get booted out than if you don't. And that would be more likely for a popular film.

That also goes some way towards explaining why movies ever kick people out (other than at closing time). To the theater, the marginal benefit of selling your particular seat is close to zero in most situations, but there's a fairly significant marginal cost to encouraging someone to leave. Why engage in that behavior at all, unless there's some other detail (like an implicit surge price) that we're missing?

Comments for this post are closed