From: goldstein@delni.enet.dec.com
Subject: Re: Measured Service: What Does It Cost?
Date: 16 Oct 89 19:19:27 GMT
Organization: Digital Equipment Corporation, Littleton MA USA

In article <telecom-v09i0447m04@vector.dallas.tx.us>, comcon!roy@uunet.uu.net
(Roy M. Silvernail) writes:

> Do you have measured service? What are the actual rates? Do you have
> to juggle zones? Do you have a free-call area? If you were there for
> the beginning of measured service, what was the introduction like?
> (was there a public outcry? Was the public even consulted?)

Here in Mass., there are measured residence options but mostly it's
flat-rate.  Business is measured-only IF there are more than 160k
local lines, otherwise you can get either.  I.e., the boonies are
flat, but Boston is measured.

Measured local service, particularly for residence, is a truly awful
idea.  The usual justification is that it's "fair" that people who use
more should pay more.  But what is fair about monopoly rates that
don't correspond to costs?  Most local measured service plans don't
have any relationship to costs whatsover.

The classic study was done in Denver in the mid-1970s, where local
calls can go up to 53 miles.  The cost of the typical local call
turned out to be under a mill a minute.  Only the longest were around
2c/minute.  It's no doubt part of the Colorado PUC's public record,
but I don't have a reference.

New York State is fairly rigorous about cost-based rates.  New York
City, with its extremely high percentage of tandem switching, is all
measured.  Costs vary with time of day, and there are multiple
distance zones.  Residence can be timed or untimed.  Untimed is about
8c/call peak hour, timed about 7c plus a penny a minute after the
first five.  I don't have the details handy.  But in any case, NYC is
NOT typical of the rest of the country!

I once worked at a firm whose major business was intervening in telco
rate cases.  Measured local service was a common telco ploy to raise
rates.  The cost of measurement typically exceeded the cost of the
calls being measured!  Thus it was actually padding the rate base,
costing the ratepayers money, and not buying any actual benefit.  If
overpriced local calls cause people to talk less, then the actual
cost/minute of the network will go up.  That's terribly
counterproductive and makes poor public policy.  Typically 80% of
telco local cost is fixed, 20% usage-sensitive.  What usage sensitive
pricing plan was like that?  Usually it gets more than 50% of revenue
from usage.

Local calls, especially within a short distance (not the
Atlanta/Denver multi-office extended local areas) are incredibly
cheap, on a marginal usage cents per minute basis.  If the telco could
really justify the rate on the grounds of cost, it would be
economically valid and "fair".  But then it would be too cheap to
bother with.  Which means they really shouldn't bother, but they
always come back again and again...

     fred

From: goldstein@carafe.enet.dec.com (Fred R. Goldstein)
Subject: Re: Sneaky! Michigan Bell Pulls a Fast One on Everybody
Date: 26 Nov 91 20:27:56 GMT
Organization: Digital Equipment Corp., Littleton MA USA

In article <telecom11.960.4@eecs.nwu.edu>, deej@cbnewsf.cb.att.com
(david.g.lewis) writes ...

> I don't know if I'd go so far as to say "almost no relation", but I
> agree that price of local telephone service and cost of local
> telephone service are rather weekly related.  There is a reason for
> this, however; it's referred to by the regulatory agencies as "the
> public interest, need, and necessity."

> Usage-based pricing is part of a general trend towards cost-based
> pricing; It recovers costs based on the use of resources in way which
> is generally deemed by regulators to be fair and equitable, and the
> usage of those resources is relatively easy to measure.

This is the telco party line, but it's buncombe.  While it's true that
some tiny teeny amount of the telco's cost is related to local USAGE,
it probably costs them more to measure it than the usage itself costs.

What the telcos have is a monopoly.  They notice telephone usage per
line rising, and they know you can't get it from anyone else.  So they
want to raise its price.  The cost of hauling a LOCAL call usually
ranges from about a penny a minute (in the highest-cost places) down
to a small fraction of a mil per minute.  So the proposed rates are
literally ORDERS OF MAGNITUDE higher than the costs!

Is this FAIR?  Of course not.  It's heavy usage subsidizing light
usage.  Does your property tax or rent get set on how many times you
open the door of your house or apartment?  Do you rent your
refrigerator based on a price per each time you open the door?  (Hey,
dieters would benefit!)  OF COURSE NOT.  They aren't monopolies.  Cost
comes into the equation.  They can't get away charging too much for
what's almost free.

The FCC's "access" (CALC) monthly charges are an attempt to move
towards cost-based pricing, by charging a fixed price for fixed costs
(previously paid by LD usage).  Usage-sensitive local service is a
move away from cost-based pricing, as further proof of the extreme
monopoly power of the local telephone carriers.

If they want to charge based on COST, let them.  I'll pay a penny a
call and a mil per minute, peak hour, half-price off-peak.  Anywhere
in the Metro area.  They'd still be ahead of the game.


Fred R. Goldstein   goldstein@carafe.enet.dec.com 
or goldstein@delni.enet.dec.com   voice:  +1 508 486 7388 

Date: Sun, 16 Jan 1994 00:19:06 -0500
From: goldstein@carafe.tay2.dec.com
Subject: Re: Unmetered Local Service

Imagine, if you will, a new telephone rate structure announced by your
local Bell company, say, Ameritech.

Henceforth, your monthly residential rate will be $2 + $1 for each
letter of the alphabet removed from "A".  Thus Abolafia will pay $2,
Bernstein $3, Coletti $4, up to Zzzyandottie who will pay $27.  This
is perfectly fair, of course, because most people have names that are
before "M", and thus come out in the cheaper half.

Is this absurd?  It is only a bit less arbitrary than most measured
local service plans!  Telephone rates are set by regulators because in
a free market, price tends towards cost, but telephones are a monopoly
so there is no free market.  And in this monopoly world, the historical 
regulatory regime has NOT been cost, but "value of service".  And that's 
basically whatever your regulator says it is.

One approach, of course, is to price telephone service (80%+ of which
is a fixed cost, whether you never use it or never hang up) based upon
usage.  Illinois Bell and New York Telephone both do this a lot.  In
the case of NY City, the state regulators made them do a fancy cost of
service study, and they ended up with a complex three-rate-period
multi-band structure that bottoms out under a penny a minute for
short-haul nighttime calls.  (Residence can get untimed service.)
Note though that NY City has an unusually complex and costly infra-
structure, with so many COs, tandems, etc., to deal with.

The trouble with overcharging for usage is that it reduces the
economic efficiency of the network as a whole. Overprice (vis a vis
true cost) usage rates discourage usage, so the usage-insensitive
infrastructure (most of it) gets less usage than it should.  The
resource is wasted.

The FCC makes telcos produce true cost data when making up the rates
charged to long distnace carriers for delivering calls.  These rates
are then buggered by a formula that produces a subsidy (read: they're
marked up), and they're entirely usage-based.  But the mileage
component is only about 1/100 of a cent per mile per minute.  A call
anywhere in a LATA usually costs about the same.

Overcharging heavy local users or overcharging people named "Townson"
is equally wrong.


fred

Date: Wed, 19 Jan 1994 00:03:18 -0500
From: goldstein@carafe.tay2.dec.com
Subject: Value of Service Pricing

All of this discussion about flat-rate local service, etc., brings up
the issue of just how telephone company rates are set.  The traditional 
method is "Value of Service" pricing.  While it is largely viewed as
obsolete, especially at the federal level, it still plays a major role
in local phone rates.

The most notable example of this is in the way local rates are set in
a flat-rate state.  The monthly rate is based on a rate band, which is
based on the number of lines within your local calling scope.  So
Nowhere Falls, with 300 phones in its local calling area, pays
$8/month while Major City pays $15/month.  After all, it gets more
value for its local bill; Nowhere Falls customers pay lots more tolls.
But the actual cost of monthly service is higher in the boonies,
because the dominant expense is outside plant, and rural areas need
the longest average wires.

I saw a good illustration of this back in the '70s when I was working
on Telephone Rate Reports.  In Ohio, the highest rate band for (I
think) United Tel (Lima) was something like 50,000 phones.  For Ohio
Bell, though, that was a fairly low band.  So Lima customers paid
something like $16/month while a Bell customer in a similar town would
pay maybe $10.  Columbus customers would be in OBT's higher bands.
United thus had to subsidize its huge rural territory (its own cheaper
bands) with small-city customers, while OBT had big cities to do it.
Overall, Ohio Bell was much cheaper, but it wasn't because United was
inefficient.

Charges for touch-tone are, of course, another "value" element not
related to cost.  They add revenue to the system in lieu of collecting
full cost from everyone.  Rural customers need a subsidy (in order to
have universal service), but things get out of hand when the system
loses sight of cost.

Telephone service isn't like electric service.  A drop line to the
street doesn't cost the electric company much; their expense is mostly
in the generation, which is consumed by usage, not connections.  Water
used to be flat-rated in some of my area, but vastly higher costs
(mostly in sewage treatment) have led to major usage-based charges.
Heavy users generate more sewage, so it's fair.  (Our sewer bills are
on a usage basis, but use water meters as a surrogate.  It's a higher
per-CCF charge than water per se.)

So what costs do heavy phone users incur?  With older CO switches and
analog transmission, heavy local usage did incur some significant
expense, though never as much, on average, as the non-usage-sensitive
portion.  With today's costs, it's less so.  A modern CO costs around
$600/line with typical usage.  If average traffic were several times
higher (all modem freaks, etc.) then it would at most double, but more
likely increase by less than that.  Inter-CO transmission is also
cheaper, now that fiber optics are predominant; they have nearly
infiite bandwidth, though the multiplexors aren't cheap.

I have no trouble with telcos charging for usage at their true
incremental cost, marked up a for a reasonable (not double) profit.
But that's not what local measured service plans usually are.  Most
make usage cover several times its fair burden, thus creating a true
subsidy to light users from heavy users.

Value of Service pricing never claimed to be cost-justified; it was a
policy decision.  Local measured usage in most cases is really a Value
of Service plan disguised as cost-justified.  The real numbers just
don't add up.


Fred R. Goldstein  k1io  goldstein@carafe.tay2.dec.com
Opinions are mine alone; sharing requires permission

Newsgroups: comp.dcom.isdn,comp.dcom.telecom.tech
Subject: Re: Network Capacity and congestion WasRe: GSM/IS136 International CPN
From: fgoldstein@bbn.NO$LUNCHMEAT.com (Fred R. Goldstein)
Date: Wed, 03 Jun 1998 15:20:21 GMT

In article <35784f19.74202340@vnntp.ma.ultranet.com>, none@stopthespam.com
says...

bob>>> How about revamping the entire telco pricing structure. How about
>>> charging in relation to what it actually costs to provide the
>>> service ?
>>
>>Define "actually costs". :) As a rule the price of every call should cover
>>the *marginal* cost of that call, but the totality of calls also has to
>>pay for the *total* costs of running the network.  Typically the marginal
>>costs add up to a lot less than the total cost, so the marginal cost of
>>a call only gives the lower bound of what should be charged in a "rational"
>>tariff.
>
>Agreed. So there has to be some allocation. If there was any sense
>to the allocation, a business and residential customer would pay the
>same for the same service. In general, they don't. This is just
>the most flagrant example I can think of. I'm sure that if I/we had
>any insight to how costs are allocated to the various types of
>lines we'd be amazed.

It is a political/value judgement.  Resi pays less for local POTS, period.
There was once an attempt to cost-justify it but now it's a value judgement,
that resi should be subsidized by business.  I didn't say it's right, but
that's how the system in the US is set up.

>Or perhaps businesses that are in telemarketing should pay
>per *incoming call*, vs. outgoing. Ticket agencies, Support
>organizations, etc. These folks use a disproportionate
>amount of the network capacity (hey, we're back to the
>thread) as compared to what they pay the telco for.

The ENTIRE marginal cost of a call is nominally imputed, in most cases, to the
originator.  Telemarketers, ticket agencies, ISPs, etc., are therefore "demand
generators" but do not incur ANY costs for incoming calls, because by
definition the cost is assigned to the caller.  That's a ratemaking philosophy
decision but it's standard in the industry.

Some telcos are trying to "double dip" by demonstrating how costly ISP-bound
traffic is.  Trouble is, they're already dipping at the origin.  To impute
costs against the terminating side means they should de-impute costs against
the origin.  You think a Belld does that?  C'mon, what do you think they send
guys to "creative accounting" school for?  They figure the public and the
regulators are so innumerate that they could add 1+1=7 and get by.

The exception is local universal FLAT RATE service, where the cost of
local usage (typically 10-20% of the total cost) is entirely included in the
monthly rate.  Then, since nobody's paying by the call, both sides are paying
for their share of the total.  Note that ISP inbound circuits tend to cost a
lot more in such places!
--
Fred R. Goldstein   k1io    fgoldstein"at"bbn.com
GTE Internetworking - BBN Technologies, Cambridge MA USA  +1 617 873 3850
Opinions are mine alone; sharing requires permission.

Newsgroups: comp.dcom.isdn,comp.dcom.telecom.tech
Subject: Re: Network Capacity and congestion WasRe: GSM/IS136 International CPN
From: fgoldstein@bbn.NO$LUNCHMEAT.com (Fred R. Goldstein)
Date: Thu, 04 Jun 1998 19:54:44 GMT

In article <3576D3D8.4B61@attws.com>, jeffrey.rhodes@attws.com says...

>When are you ever going to admit that this is a subsidy, so that our
>grandmothers can be sure to call '911' Emergency service when they need
>to, but who otherwise perhaps only wait for incoming calls? Your
>precious Universal Service FLAT RATE is going away and is being replaced
>by bigger *bundles*.

My employer's ILEC affiliate uses flat rate pricing in lots of areas, I think
including your own, and makes tons of money doing it.  No subsidy.  It's called
"rate averaging".  Since the variable part of the cost is *de minimis*,
charging everyone the same means that everyone is paying *close to* their own
incurred cost.  In general, where measured tariffs are in effect (including the
Bell Titanic states and California), the usage component is priced at such a
high multiple over cost (usually 3-20 times) that the cost/price ratio is
FARTHER away from unity than in a flat-rate area.  They hate to admit this but
that's what the numbers show.  For instance, Bell Titanic charges 5.2c per
minute for a "zone 2" (8 mile) call, from say Cambridge to Lexington, but their
cost studies should reveal it to cost them less than a third of a penny a
minute.  Zero is much closer to 0.3 than 5.2 is to 0.3.

The rest of your post reveals such incredible ignorance, blended with
arrogance, that I'll return you to your Plonkee status.
--
Fred R. Goldstein   k1io    fgoldstein"at"bbn.com
GTE Internetworking - BBN Technologies, Cambridge MA USA  +1 617 873 3850
Opinions are mine alone; sharing requires permission.

Newsgroups: comp.dcom.isdn,comp.dcom.telecom.tech
Subject: Re: Network Capacity and congestion WasRe: GSM/IS136 International CPN
From: fgoldstein@bbn.NO$LUNCHMEAT.com (Fred R. Goldstein)
Date: Fri, 05 Jun 1998 15:51:15 GMT

In article <pk94sy0p1m8.fsf@se.bel.alcatel.be>, grayc@se.bel.alcatel.be says...

>Surely the *best* incentive would be a straight metered service?  The
>monthly bundle is really just a compromise to ease the withdrawal
>symptoms of customers addicted to flat-rate service ...

That's a common attitude amongst European PTTs, who are accustomed to raping
their ratepayers and who fear competition.  Straight metered service, as I've
pointed out, usually doesn't reflect cost.  Instead it provides a very strong
artificial disincentive to make efficient use of the most expensive part of the
network, the loop plant.

Of the cost of providing telephone service, the largest cost is in the local
loop.  And that's not going down very fast, since labor and copper prices
aren't.  Pair gain is getting cheaper (and wireless loop is improving), so the
high end of the cost curve (long loops) is coming down; it's the low end (plain
copper) that's pretty stagnant.  The actual traffic-sensitive cost of local
usage is extremely low; modern digital switches, with RAM-based networks, can
provide much higher traffic capacity, on the cheap, than the older analog ones.
Ditto for fiber optic plant, which can carry a lot more (thus cheaper per bit)
than before.

That's the phone industry's dirty little secret.  They (well, some of 'em; some
would rather brag to Wall Street about earnings growth!) caterwaul about how
much they're getting hurt by data traffic on the POTS network.

Real numbers:  Bell Titanic has admitted in public that their average cost of
local usage (BA-South, eg., metro Washington) is in the range of a quarter-cent
per minute.  Now the average (remember, flat rate is an averaged rate, not
free) resi line originates something like 10 hours/month.  Gee, that's
0.25*60*10=$1.50/month for all of the usage!  Now give the guy a modem and let
him on AOL 2 hours a day (above average, for sure) and you're up to $9/month
for the usage.

What would the Belgian PTT charge for that?  What would Bell Titanic like to
charge for that?  Does any PTT charge *cost* or even *cost plus* (not double or
more) for usage when service is measured?  Or do we see local calls at 4-6
cents a minute?  Strong disincentive to do *anything* with the POTS network, if
you ask me, but legal monopolies can gouge.

A truly cost-based metered scheme is possible, of course, but it would probably
cost more to meter than it's worth.  Charging one cent per three minutes is
cost justifiable, but telcos tend to get greedy.  Better to just average it all
than to charge too much.  Besides, subscribers WANT flat-rate service.  They
WANT to "overpay" for the convenience of not having to think about usage costs.
In a monopoly, you can ignore their wishes.  Once competition comes, you can't,
so if subscribers want flat rate, somebody will give it, or at least something
that looks a lot like it (see next paragraph) to them.

Re: Blocks of time.  The outliers are the "nailed" circuits, the ones up 700
hours/month, and I know of too many such users.  Nailed usage includes both
peak and off-peak (almost no cost) usage, but in any case these guys are a
major loss to the telcos in flat-rate areas.  A big block of time, say the
now-common (US West, BellSouth, PacBell) 200 hours/month, is essentially
unmetered to 90%+ of users, but nails the ones who nail circuits.  A
"semi-permanent circuit" rate, or AO/DI, are more sensible, cost-justifiable,
alternatives.  I've testified in rate cases in favor of a 200-hour cap (also
called "threshold pricing") as the norm for ISDN local usage.

Bell Titanic's policy is to measure ISDN and leave analog flat rated.  This
encourages heavy as well as nailed users to go analog, which is worse:  Analog
switch components are the ones that block most easily.  ISDN usage is generally
cheaper, though 64000 bps clear-channel is a bit costlier, in the USA, because
so much of the POTS network is not clear channel (hence only works at 56000).

>Analogue modems in general (and fax in particular) are a very inefficient
>way of using today's mostly-digital networks, so their continued popularity
>indicates some kind of distortion somewhere.

Once upon a time (late 1700s), a Scottish fellow named Adam Smith wrote a book
called "The Wealth of Nations" which became the basis of most modern
economics, including Marx and Keynes. It said that in a truly competitive
marketplace, cost and price naturally align.  "Value of service"
pricing violates this; it means that there is not a free market, so the
monopolist can use other pricing means.  The popularity of modems, however, is
not a distortion, except that ISDN is overpriced so its share is too low (in
the USA).

Data over POTS is economical (in the Smithian sense), because all of the
alternatives cost the carrier more than the $3-5/month that your typical modem
owner's local usage does.  Overpricing local usage in order to discourage
modems would be a DISoptimization, since it would force people off of a $3-5
solution onto a much costlier one, whatever that turned out to be.  Again,
nailed and very heavy users are the exception; they SHOULD be moved off.

--
Fred R. Goldstein   k1io    fgoldstein"at"bbn.com
GTE Internetworking - BBN Technologies, Cambridge MA USA  +1 617 873 3850
Opinions are mine alone; sharing requires permission.
Definitely not an Official GTE position.

Newsgroups: comp.dcom.isdn,comp.dcom.telecom.tech
Subject: Re: Network Capacity and congestion WasRe: GSM/IS136 International CPN
From: fgoldstein@bbn.NO$LUNCHMEAT.com (Fred R. Goldstein)
Date: Mon, 08 Jun 1998 14:14:40 GMT

In article <pk9d8ckns4r.fsf@se.bel.alcatel.be>, grayc@se.bel.alcatel.be says...

>> That's a common attitude amongst European PTTs, who are accustomed to raping
>> their ratepayers and who fear competition.
>
>Sorry, I don't see the connection.  This would not explain why GSM operators
>with no previous history as "PTT"s, operating in a competitive environment,
>meter calls.

GSM and other wireless operators have a very, very different set of numbers to
work with.  In the wireless world, the vast majority of the cost is traffic
sensitive:  There are no wires, and a given cell can support any number of
subscribers, but only a given number of calls at a time.  Compare this to
wireline, where most of the cost is in the wire, which costs the same whether
it's talking halfway around the world or sitting idle.

This BTW is a risk to the use of wireless local loop in places like the USA:
If there is a high marginal cost of usage, then the popular flat-rate package
will become harder to sustain.  This is not, however, fait accompli, as the WLL
business is still developing.

>Whilst it is true that cost and price will naturally align, allocating
>the cost of running a network among subscribers and their usage is not
>trivial.  If everyone pays only those costs which can be directly linked
>to their usage and theirs alone, the total cost of running the network
>is unlikely to be met.

Why do you suggest that?  Everyone pays their own cost, both direct and
indirect -- those costs which cannot be directly assigned are allocated.  There
is a huge body of telco accounting to cover this.  However, the allocations
still fall into TS and NTS baskets, and if the NTS basket is the bigger one,
then the NTS price (the monthly bill) should be the bigger one.  In other
words, heavy users (efficient use of wire) should not subsidize light users
(who are more likely to waste the wire).
--
Fred R. Goldstein   k1io    fgoldstein"at"bbn.com
GTE Internetworking - BBN Technologies, Cambridge MA USA  +1 617 873 3850
Opinions are mine alone; sharing requires permission.

From: floyd@tanana.polarnet.com (Floyd Davidson)
Newsgroups: comp.dcom.isdn,comp.dcom.telecom.tech
Subject: Re: Network Capacity and congestion WasRe: GSM/IS136 International CPN
Date: 9 Jun 1998 11:45:52 GMT

Chris Gray  <grayc@se.bel.alcatel.be> wrote:
>
>I would argue that the congestion-based approach leads naturally to a more
>efficient use of the network, whereas the "value-based" approach leads to
>distortions.  I say I *would* argue, because it's difficult to argue
>anything when nobody seems able to understand a word one is saying ...
>

A "congestion-based" approach leads to a customer base which is
adjusted to match the network design.  A technically high
quality network is designed and then pricing is set to allow
efficient operation.  It might not necessarily meet customer
needs, but it can be an example of quality implementation.

The problem with a congestion-base approach is the lack of
incentive to re-evaluate design philosophy or modify the
network to match customer needs as long as usage patterns can
be controlled by adjusting the price structure.  Changes in
customer use patterns can and will be discouraged to allow
existing equipment and design to continue functioning.

The "rapid" implementation of ISDN is a perfect example.
Attempts to reduce Internet access via the PSTN are another.

A "value-based" approach leads to a network which is designed to
match customer needs.  Rapid deployment of new technology to
meet perceived changes in customer needs necessarily puts
quality of network design in a secondary position.  Hence a more
capable network, of lower quality.  With a value-based approach
reliability gravitates to the minimum that can be tolerated.

Perhaps the World Wide Wait is an example...

It is an exceedingly complex problem and the solution is an
exceedingly complex mix of both methods.  There are always going
to be both winners and losers any time adjustments are made.
But fairly distributing the gains and losses is not nearly as
important as setting a policy that allows for a functional
network in the future.  Hence all of these concepts intended to
fairly distribute either network costs or network resources have
to be viewed in the perspective of successfully developing the
future of the network first, and the distribution second.

Personally I think adjusting pricing to reduce IP traffic on
the PSTN, as opposed to shifting the design philosophy of the
network to match customer demand, is abject stupidity.

  Floyd





--
Floyd L. Davidson                                floyd@ptialaska.net
Ukpeagvik (Barrow, Alaska)

Newsgroups: comp.dcom.isdn,comp.dcom.telecom.tech
Subject: Re: Network Capacity and congestion WasRe: GSM/IS136 International CPN
From: fgoldstein@bbn.NO$LUNCHMEAT.com (Fred R. Goldstein)
Date: Tue, 09 Jun 1998 21:16:05 GMT

In article <pk93edeorch.fsf@se.bel.alcatel.be>, grayc@se.bel.alcatel.be says...

>> Why do you suggest that?  Everyone pays their own cost, both direct and
>> indirect -- those costs which cannot be directly assigned are allocated.
>
>Isn't that what I said?  I've just re-read it and it still seems so to me.
>Yes, those costs which cannot be directly assigned are allocated.  There
>is more than one way to do this, and one approach is to look at the
>contribution each user is making to congestion in the network; users
>who make the network less useful to other users should pay more.  Another
>approach, advocated by Mr. Rhodes, is that users whose traffic us more
>valuable to themselves should pay more ("value-based" pricing).
>
>I would argue that the congestion-based approach leads naturally to a more
>efficient use of the network, whereas the "value-based" approach leads to
>distortions.  I say I *would* argue, because it's difficult to argue
>anything when nobody seems able to understand a word one is saying ...

The problem with that approach is that it's fixated upon congestion, which is a
traffic-sensitive matter, in a network whose non-traffic-sensitive costs are
far and away greater.  Economically, there is a cost of fixing congestion.
It's reasonable to have people who cause congestion pay it.  But the telcos are
caterwauling about how much they are spending to overcome Internet-related
congestion, and the numbers they pass around are *miniscule*.  Sure, a hundred
million bucks sounds like a lot, but when PacBell averages that over the years
they're talking about, it's a couple of cents PER MONTH per line. Not per
minute of use, but PER MONTH.

>This discussion was about network congestion; the local loop has little to
>do with that.

But it does -- the price of local service includes both loop-related and
use-related (NTS and TS), and congestion is solely a TS issue.  If the cost of
fixing congestion is small, relative to NTS (loops), then it's uneconomical to
try to reduce usage of the loops.

As someone else pointed out in e-mail, subscribers perceive *value* in having
flat-rate service, which is the (valued) right to make certain calls without
incurring a charge.  They perceive other *value* in the ability to receive
calls, and in other features and capabilities.  Flat rate service that costs a
little more per month that equivalent metered service is still a better value,
simply because subscribers value it.  Metering service therefore decreases the
value of the network, perhaps by more than the cost of fixing congestion.
--
Fred R. Goldstein   k1io    fgoldstein"at"bbn.com
GTE Internetworking - BBN Technologies, Cambridge MA USA  +1 617 873 3850
Opinions are mine alone; sharing requires permission.