Thursday, September 5, 2013

Can airlines survive a classless society? Maybe not. Can they survive without first class? Yes.

As previously established, this blog is firmly anti-price discrimination.  So it was with interest that I read this post from Rory Sutherland, defending price discrimination in airfares as a necessary evil, and encouraging that other services emulate their practices.

Sutherland's basic point is this: by dividing airplanes into first-class and coach-class seating, airlines are able to attract wealthy, high-paying passengers who subsidize most of the cost of air travel.  As a result, airlines can sell tickets to the rest of us plebes at improbably affordable prices.  (Although he doesn't use the term, what he's describing is textbook price discrimination: selling more expensive tickets to people who are willing to pay more.)  He goes on to suggest that a similar two-tiered system could be replicated elsewhere; that we'd be able to provide more subways, school buses, and other niceties if society was a little less sensitive to class divisions in our services:
I sometimes suggest that we would similarly benefit from having different classes of travel on the London Underground. If the first two carriages in each train cost three times as much as the others but offered free Wi-Fi, and were furnished not with basic seats but with the sumptuousness of an Edwardian-era New Orleans brothel, you could afford to run more trains. Almost everyone finds this idea repellent. 
But I’d like to issue a challenge to any libertarians, economists, ethicists or software gurus reading this column. How do you get people with wildly differing willingness or ability to pay to fund some common good other than through redistributive taxation?
He's right in part and wrong in part.  Building off Sutherland's point, Felix Salmon does better: he points out that price discrimination exists both between first- and coach-class customers, and among them.  Two people sitting in the same might have paid wildly different amounts for their ticket.

To the extent that price discrimination is occasionally necessary, Felix gives a far better example.  Maintaining certain for-profit services might require a degree of "redistributive taxation" in the form of price discrimination, but that price discrimination can't be used to justify class distinctions in services.  Class distinctions simply wouldn't accomplish what Sutherland thinks they would.

This is the part where I use strained hypotheticals to show you what I mean.  (What can I say? Law school habits die hard.)

So imagine that you own an airline that flies three routes every day. It's expensive to provide those routes and you'll only make back your costs if you fill at least a third of them with high-paying customers.  Fortunately for you, high-paying customers like the flexibility your airline provides (three routes! Every day!), so each day all the rich people in town choose one of the routes and fly it.  You don't know in advance which one they'll pick, so you have to maintain all three.  And after they buy their tickets, you may as well sell tickets on the other routes to low-paying customers--you can’t get rid of the routes without losing your core clientele, and there’s no reason to not make an easy buck or two on the excess capacity.

In that situation, it can accurately be said that high-paying passengers are subsidizing the low-paying passengers.  Selling tickets to the rich is the airlines' raison d'être.  The airline was built for them; the rest of us get to come along for the ride, which is ultimately good for us.

You might point out that this is highly unrealistic.  In real life, maybe you'd expect the profit-generating customers to split more-or-less evenly between the three routes.  But if that's the case, it's virtually guaranteed that a large portion of the capacity along each route will be used to fly loss-generating customers.  In this instance, you're better off just reducing capacity overall; i.e., buy smaller planes.

One way to think about this is to realize that in order for a rational airline owner to spend money providing any given seat, there must be some chance that a profit-generating customer will fill it.

Another way of looking at it is to realize that if the value of a service for profit-generating customers is its capacity, and if that required capacity is in excess of what profit-generating customers can consume, then you'd expect de facto price discrimination that benefits society at large.  By selling tickets at continually lower prices until all seats are filled, the airline ensures it is generating the maximum possible profit from its available resources.  Some of the seats might be sold at a loss--the airline would save money by not producing the seat and therefore not selling the ticket--but the required capacity isn't created, no one, rich or poor, would buy a ticket.

It's not just transportation.  Salmon's post talks at length about price discrimination in the newspaper industry, which is a pretty good example of another service with similar characteristics.  A lot of a newspaper's credibility and currency are derived from the scale and scope of its readership; this is a big part of why, for instance, the New York Times is better-regarded paper than the Toledo Blade.  But the number of profitable subscribers to the Times is far smaller than the number of readers required to maintain its current global reach.  So it makes sense to institute a system in which wealthy customers are encouraged to pay full price while other customers can avail themselves of bargains or visit the site for free or sneak past the paywall.

So far, so good.  But here's the problem with Sutherland's argument: this logic only works when the consumers of a particular service are undifferentiated.  When you know that a service is mostly attracting profit-generating or loss-generating customers, then the reasoning breaks down.

Think again about the three air routes above.  Now, let's imagine that our airline enhances one of the routes with luxury service, free drinks and food, spacious seating arrangements, and the rest.  In doing so, it hopes to consistently attract the rich passengers to that route.  That's all well and good--it's certainly the airline's prerogative to do this if it wants.  But since it now knows its profitable customers will end up flying the one route and not the two others, why even maintain the second two?  They're losing money!  It's better off scrapping them.

Alternatively, the airline could create two classes of seat on every plane.  One provides excellent service to the rich, and one provides minimal service for everyone else.  But again, if you know which part of your service will attract profitable customers, why provide the rest of your service to everyone else?  Once again, you ought to just buy smaller planes.

Or think about newspapers again.  If papers know that both classes of consumer are attracted to the whole product, then price discrimination sounds like a good idea.  But what happens if they discover that rich subscribers are mainly reading the real estate section, while poor moochers are mainly reading the funnies?  Why spend money producing the funnies?

In short, it's unlikely that a private firm would provide a service that its knows will be consumed at a loss.  This makes intuitive sense, it's worth noting.  It also shows the defect in Sutherland's reasoning.  In all of his examples, services are split into two tiers: a fancy service for the rich and a crappy service for the poor.  But as a mechanism for cross-subsidization, tiered service doesn't work--the service provider is always better off just providing the upper tier.

One might counter that Sutherland didn't mean that low-paying consumers are truly loss-generating. But if that's the case, it's hard to see what the benefit of price discrimination is to the rest of us.  If none of the tickets are in any respect losing money, the service provider can still profitably provide the same level of service with standardized (and for some people, lower) prices.  It's true that expensive tickets might help subsidize cheap ones while holding profits constant, but I'm dubious that this forms a great case for the social necessity of  price discrimination.

Of course, all this raises the question of why some services are tiered.  And there are couple of answers to that.  First, there is still an element of price discrimination here--of the consumer-gouging kind.  Even if it's not strictly necessary to keep your firm afloat, price discrimination is still a good way to make a few extra dollars on the back of hapless consumers.  Second, I'm sure there's some profit in creating services tailored to certain needs.  Some people just want first-class seating; others want cramped, cheap seats; airlines cater to both in various degrees.  And within those categories, Sutherland's argument works just fine.  Airlines need a certain capacity of each seating class, and fill any excess with bargain tickets sold to low-paying customers.  The problem is that, while possibly necessary to sustain the industry, this practice is also necessarily invisible to consumers.

You might wonder why I care so much about this relatively esoteric point.  There are a few reasons.  First, as I've said before, price discrimination is actively harmful to consumers, so we need to be careful about rationalizing its widespread adoption.  Thinking through the problem in detail helps us see when certain forms of price discrimination are and are not necessary; I think we can safely say, for instance, that airlines can exist just fine without dividing their passengers into castes.  Besides, class divisions in society seem generally pretty corrosive, so arguments in favor of them ought to be tested.

But more broadly, I just don't believe policymakers think enough about the mechanics of consumer spending.  Companies sink tons of research and money into creating novel ways to squeeze a few dollars more out of each consumer purchase, and no one bats an eye.  People accept as an obvious truth that price discrimination and similar transactional innovations can change behavior on the producer side of the supply and demand curve.  But people rarely consider that these things have an equal and opposite effects on consumers.  Depressed consumer purchasing incentives and reduced consumer surplus are every bit as real and worthy of discussion as sales tactics--so let's discuss them.

No comments:

Post a Comment