Price – Splitting the Pie – Access Charges

Understanding the debate over Net Neutrality means understanding how money flows between the many telecommunications providers who may handle a single packet. Smart readers of my post on the cost of pricing complexity like Aswath Rao and Rajesh Raut have commented that “bill and keep” – the usual model for connection between Internet backbone providers – is not the only possible model and may not be the best one.

This subject is unfortunately complicated.  But, if you read on, I promise you’ll understand both why it is nonsense say that Google is using at&t’s pipes “for free” and why at&t CEO Ed Whitacre and other traditional telco execs think this is exactly what is happening.

When you access Internet content or make a phone call – particularly an international call – many companies and many networks are involved in getting photons back and forth between you and the web site you’re accessing or between you and the person you’re calling.  Not unreasonably, the companies want to make money from their networks.

The revenue sharing models used on the Internet evolved  on the telephone network long before there was an Internet.  The four basic models are:

1. access charges
2. net settlement
3. bill and keep
4. facility rental including line sharing

To add further complexity, there are both access and backbone networks although most access networks include some backbone. Access networks connect to individual subscribers – residential or business.  In the US your local landline is part of an access network operated most likely by your local RBOC (regional bell operating company) but possibly operated by one of the thousands of ILECs (independent local exchange carriers) which serve mainly rural areas.  The towers of your mobile provider are another access network.  And whatever physical connection you have to the Internet is a third unless it’s DSL in which case it shares the physical copper which makes the RBOC or ILEC access network.  Told you this was complicated.

The court ordered breakup of AT&T in the 1980s separated the former monopoly into a host of regional access companies (the RBOCs) and backbone carrier AT&T.  MCI in its early days was a competitor of the backbone carrier.  One of the purposes of the breakup was to allow competition at the backbone level (“long distance” as we called it when distance still mattered).  When AT&T controlled both access and backbone, the reasoning went, it was impossible for a competitive backbone carrier to get any business.  Good thinking.

Formerly, we’d paid a bill to a single phone company – almost always Ma Bell.  Now we each had to choose a long distance (backbone) company as well.  If you didn’t choose, you got assigned to a long distance company just to keep things fair.  Initially, the RBOCs were not allowed to sell long distance and AT&T was not allowed to provide local residential service (access).

From the beginning, the RBOC networks provided both access and backbone since local and regional calls stayed totally on their network.  But, if Bell Atlantic was your RBOC and MCI your long distance carrier and you were calling someone in California (Pacific Bell), your call traversed all three networks.  MCI billed you for these long distance calls and paid “access charges” out of the proceeds to both Bell Atlantic who you used as an access network and Pacific Bell, the access network of the person you called.

Initially, when domestic long distance was still very expensive, the pennies per minute paid for access were a very small part of the cost of a call.  But competition at the backbone level worked.  Prices of long distance calling began to fall.  Access charges didn’t fall, however, since you still had no way to access your long distance carrier except through Bell Atlantic and AT&T had no way to get the call to the person in California except through Pac Bell.

There was no competition in access.  Sound familiar?

Although they were thought of as dowdy compared to the backbone carriers, these were actually great days for the RBOCs with their local monopolies.  The more the price of long distance went down, the greater the number of long distance calls and the greater the access revenue.  Almost every penny of access revenue went right to the bottom line of the RBOCs.  They had no marketing cost for the long distance calls that AT&T and MCI were advertising and discounting heavily.  They were already charging us a fixed price per month to rent us our access lines but now they were receiving ever greater access charge revenues from us indirectly whenever we used our lines for long distance calls.

Politically, access charges became very important.  RBOC rates were (and still are to some extent) regulated at the State level.  The monthly cost for a local line, which generally includes unlimited or a healthy amount of local calls, is very visible and politically sensitive.  It was hard for the RBOCs to get increases approved.  But the access charges are invisible; they’re paid by the long distance provider and are part of your long distance bill but never broken out.  So access charge revenue per line went up (stable monopoly prices, more volume) while the cost of a basic line rose more slowly. 

To some extent, long distance subsidizes local service through access charges.  The RBOCs effectively lobbied with this argument and the social corollary that poor people benefit from the subsidy since richer people make more long distance calls than poor people.  (However, calls made on business lines didn’t and don’t pay access charges by the minute.)

I was at AT&T when the dime a minute long distance plan was offered.  The cost of carrying the call over AT&T’s backbone was estimated at about one cent a minute.  Typically three cents per minute went to pay access charges.  Customer care and billing cost abut four cents per minute – remember, the RBOCs had neither customer care or billing expenses for these long distance minutes.  The remaining two cents had to cover marketing and AT&T’s bloated corporate overhead.  Don’t feel sorry for AT&T.  Most customers were not and are not on the most efficient calling plan.  It’s still easy to pay more than a dime a minute (or less) for domestic long distance.

OK.  Enough for today’s lesson.  Just remember the telco execs were brought up in system in which they both rented you an access line AND got to charge you again (this time indirectly) when you used it or when someone used your access line to call you.  It’s not surprising they’d like to charge twice for Internet access.

Next lesson is net settlement and a special case called reciprocal compensation. Also how the Telecommunications Reform Act of 1996 didn’t break the RBOCs local monopolies but how the Internet almost did.

BTW, I’m going into so much detail because the issues in the Net Neutrality debate in the US are important.  How they’re dealt with will affect both the quality of our online life and our national competitiveness.

The first post in this series on price is about the cost of complexity.

The second post is about FON, the WiFi sharing scheme, as a great experiment in pricing.

The third post is about “freeloaders”.

The fourth post is on unseating an incumbent with disruptive pricing.

Comments

[Lawrence]

Excellent article about access charges for broadband communications. T1 internet service broadband communications is the future of internet technology at http://www.1-satellite-tv-facts.com/T1-Internet-Service.html

[DG Lewis]

Parkite, Tom:

Actually, there are a number of reasons intrastate LD is more expensive than interstate LD. One is the inertia of the competition reason Tom mentions - the MFJ reserved intraLATA toll to the RBOCs, so there was no competition, and many intrastate calls are intraLATA.

I say "inertia", because intraLATA toll was opened up to competition in 1999, so theoretically intraLATA intrastate prices would have dropped with competition. However, as Parkite guessed, intrastate access charges are regulated by the state PUCs, which have historically been more favorable to the local carriers than the interexchange carriers. So intraLATA intrastate is more costly to someone competing with the local carrier, holding the costs artificially higher.

InterLATA intrastate is more expensive for the same reason - higher access charges due to the regulatory regime.

The fascinating thing will be to see what happens to the regulatory positions of the remaining RBOCs now that all them own interLATA carriers (SBC/BellSouth-AT&T, Verizon-MCI, USWest-Qwest). By my guess, SBC/ATT is now a net payer of access charges rather than a net recipient; Verizon is probably still a net recipient, but it's close. How long this fact will take to percolate through the organizations to the places where regulatory policy is set, and what their reactions will be, will be interesting to watch.

[Tom Evslin]

VCMC, Parktite, DG Lewis:

Thanks for the encouragement. Was afraid I might be getting a little too geeky even for the FOC crowd.

Parktite:

The consent degree didn't allow AT&T and MCI (RIP) to compete in IntraLata (in practice, Intrastate) so the rates there are set by the RBOCs without competition. There are no formal access charges on this traffic since it is onnet to the RBOCs.

However, if you buy unlimited US calling from Vonage or one of their competitors, it does include intrastate LD. Skype calls from a PC in the US to any phone in the US or Canada are currently free.

DG:

You're right and I should have been more precise. Modal access (eg. a dedicated T1) doesn't pay per minute access charges but a single business line from a dentist or startup.com does.

[vcmc]

Maybe this means I'm a geek, but I find this series fascinating - not to mention incredibly relevant to the Net Neutrality debate. Thanks for the insight.

[Parkite]

Great summary post..........perhaps you can answer this.....why are intrastate long distance calls so much more expensive than interstate calls? Still 2 access charges. The only difference i see is that in the case of an intrastate call both access charges are made by the same RBOC (or ILEC). Is this a case of the state PUC granting the RBOC the ability to charge a higher rate for access charges when the call is intrastate?

[DG Lewis]

Tom,

Good overview of access charges. When I first heard Ed Whitacre's comment, I had exactly the same thought: "Of course he wants both ends to pay for the internet traffic -- that's the access charge domain he grew up on."

A couple of nits.

One, you write, "calls made on business lines didn’t and don’t pay access charges by the minute." That's only true for those business lines that were/are directly connected to the LD company's network. What most people think of as "business lines" - the phone line at the local real estate office, for example - are plain old phone lines, and LD calls to and from those lines do pay per-minute access charges.

Even the directly-connected business lines almost always use access provider facilities, and the access provider gets monthly revenue from the LD provider for them, with the same advantages as per-minute access revenue (invisible to the end customer, no marketing expense, monopoly pricing).

If I recall correctly, RBOC access revenues (collectively) run about 55% switched/45% dedicated ("special access"), although the growth in data traffic is pushing the ratio more towards dedicated - when a customer puts in a T1 or Fractional T1 to tie a location to their corporate WAN, it's a special access circuit from the access provider.

Two, you write, "the RBOCs had neither customer care or billing expenses for these long distance minutes." The RBOCs did (and do) have significant organizations for carrier access ordering, provisioning, and maintenance, and these organizations are often held to a higher performance standard than the equivalent residential or business care organizations - AT&T and MCI were far more willing and able to go after the RBOCs for failure to perform that you or I are. They're not "end-user" facing care organizations, but they are there and they cost money. The RBOCs and ILECs also have to perform access billing for all those minutes. Again, the AT&Ts and MCIs were far more able to audit the bills - and complain to the state PUCs and FCC if they found errors - so the billing had to be accurate, and that's not hardly free.

While those comments apply to the pre-merger world, I'd find it surprising if they changed significantly. The LD part of AT&T is still paying somewhere north of $6B/year in access charges (though that'll drop by a billion or so when the BellSouth merger closes), and the local part of AT&T is still collecting somewhere on the order of $4B.