On Tue, 11 Apr 2006 17:52:29 +0000, Jose wrote:

Why should I pay to keep some remote airstrip open if you won't pay to
keep my less-remote airstrip open? People who live far out there
shouldn't depend on me for support. Right?

Who is "you"?

Airports, like any kind of network node, gain/lose value by network
effects. The more airports exist, the more possible destinations for any
trip and therefore the more value there is in any one airport.

So airports need to be treated, economically, like a network. One
doesn't just price out an individual node as an individual node has little
benefit. But an incremental node does have benefit (the amount of which
is determined by a function on the number of nodes that already exist).

Unfortunately, the US has some very bad examples of this. For example,
consider any state road that leads, at the border, to another state's
road. If one state were to close their road, the value of the other
state's road would drop (albeit not to zero). Yet there's no cross
funding mechanism available.

Of course, this is also because that type of analysis is difficult to
accomplish and impossible to prove.

The Interstate road system avoids this problem by adopting a single-payer
model. That's part of the basis for FAA investments in airports, and it
is economically sound.

In other words, everyone should be paying to maintain the entire airport
network. I may never use (for example) LGA. But that it exists has value
for me (even if it's just to keep the crowds down a little at EWR {8^).
Similarly, I might never fly into 47N. But that it exists as a possible
destination makes my home airport that much more valuable.

ATC functions are, I think, different. My use of that service is often
forced upon me because of the requirements imposed to satisfy another
user. That is, I need to chat with TRACON to get home because "home" is
in class B and class B exists to satisfy the carrier crowd.

With airports, every new airport provides at least a little value for
everyone. With ATC, value is actually mostly accrued only to a small set
of users despite its impact upon many.

Yet, somehow, I feel like a single model could be built to cover both
cases.

I cannot believe that nobody has really studied this.

- Andrew