October 11, 2003


The Economist's Tom Standage writes about the telecom industry:

Economist.com: ...Telecoms is an infrastructure-intensive business, and because infrastructure takes a long time to build, telecoms firms have to make bets on the level and nature of future demand. As it turned out, the bets made during the technology bubble of the late 1990s were spectacularly wide of the mark.

In particular, the mania that accompanied the rise of the internet persuaded many investors that demand for data-network backbone capacity, to pipe internet traffic across continents, was about to explode. Since 1997, internet traffic has roughly doubled every year. Much of the industry, however, was convinced that traffic was doubling every 100 days. This widely quoted statistic, which originated at WorldCom, became an essential ingredient of business plans and conference presentations during the bubble. It assumed unimpeachable status when it appeared in a report published by America's Department of Commerce in April 1998.... European firms gambled that the expected surge in demand for fixed-communications capacity would be accompanied by a similar leap in demand for mobile capacity, and paid euro109 billion ($125 billion) for licences to operate “third-generation” (3G) mobile networks. They, too, took on a huge pile of debts.

But the expected deluge of data never came. In the four years from the beginning of 1998, says Andrew Odlyzko, a telecoms guru at the University of Minnesota, the amount of fibre in the ground increased fivefold. Meanwhile, advances in the technology of feeding signals into fibres at one end and extracting them at the other increased the transmission capacity of each strand of fibre 100-fold, so total transmission capacity increased 500-fold. But over the same period demand for transmission capacity merely quadrupled, a rise that could easily be accommodated by existing networks.... Mr Odlyzko estimates that around $150 billion was spent building unnecessary telecoms networks in America and another $50 billion in other parts of the world.

When it became clear that the predicted explosion of demand was not going to happen, operators frantically cut their prices, hoping to fill their empty pipes. Equipment-makers' sales collapsed and their share prices tumbled. Nortel, a once high-flying maker of telecoms gear, saw its market capitalisation fall from around $400 billion in the summer of 2000 to just $3 billion two years later....

The post-bubble opportunities seem to lie in exploiting three main trends. The most visible growth area is the continuing rise of mobile phones, which have overtaken fixed-line phones to become the most widespread communications devices on earth. Their number is expected to rise from 1.3 billion today to 2 billion by 2007, and they are being increasingly used to do much more than make phone calls, providing new opportunities for wireless operators and equipment makers.

The second trend is the growth of high-speed or “broadband” internet access, which is booming in many parts of the world.... A third promising area is in the corporate-telecoms market. As large firms look for ways to cut costs and move operations overseas, many are adopting new internet-based technologies that can interconnect regional offices cheaply and securely and allow voice and data to flow over the same network...

Posted by DeLong at October 11, 2003 08:23 AM | TrackBack


I wish I had a nickel for every business plan I have ever seen (1996-2000 but none since then) based on "the convergence" of the Internet -- the theory being that all voice, data, and consumer video would be carried over the same packet based network. Ha.

Having said that packet switching for telco is growing quite nicely, thank you, but traditional cell-based switching still seems to be %95+ of the market. It is hard to get real numbers because prior to the bubble burst we had such a huge overabundance of inventory at all levels (didn't Cisco burn about $2B worth rather than dump it in the market?) that has only recently burned off to managable levels. In March of 2000 we were facing 55-week lead times on many components, then one year later we're buried in the stuff.

Funny thing is that everyone what was happening when it was happening -- that we were in a bubble. Yeah there were some idiots (Dow 36,000 anyone?) who thought the growth would go on forever but by and large a lot of us were biting our nails and trying to prepare for the bang -- then could do little about it when the bang did happen.

Ever go to an investor with a business plan that forecast 50% of your last year revenues? When everyone else (including your own customers) were prediciting double or triple digit growth? PER QUARTER? Let me tell you it can't be done. They would just throw you out and get someone in there who would "make it happen" (by which they would mean tell them what they wanted to hear.)

Well we survived, and we're back to growth again. Can't say it's been fun.

Posted by: Alan on October 11, 2003 02:03 PM

As a number of people savvier than I have pointed out, it wasn't just mania, it was also fraud that helped fuel the overbuilding of the telecom infrastructure. The competitors of Worldcom, Global Crossing, Enron Broadband et al felt both compelled and enticed to overbuild -- compelled because market pr requirements forced them to keep up with others' public capacity growth statements, and enticed, because other market players appeared to be making money hand over fist while expanding their networks and capacity. If, for example, AT&T had come out and said they didn't believe Worldcom's numbers because those numbers didn't reflect anyone else's experience, they and their stock price would have been laughed off the street. It is so much easier to fuel a bubble when you can just make your sales and expense numbers up.

In addition, even if everyone had been acting rationally, you would expect substantial overcapacity right now because a) no sane telecom engineer digs holes in the ground with the expectation of digging more holes in the same place less than 10 years (and preferably 20-40) hence and b) in a world driven by increasing returns, aka the fevered pursuit of monopoly rent, any single player expecting to "win" would want a network capable of handling all the traffic expected.

Oh, and that goldarn moore's law thing in conjunction with 20 years of previously-underfunded bright ideas didn't help either.

Posted by: paul on October 11, 2003 02:11 PM

Of course the same geniuses that were predicting this data explosion were also doing everything they could to hold it back. Not content with slow rollout of DSL (and every legal tactic along with several non-legal to prevent non-telco firms from getting into DSL and taking up the slack), they insisted on a wide variety of measures that discouraged home users from starting up their own servers (where server is a grandiose term for things like your own web site or blog); measures like charging for a static IP address or charging for multiple IP addresses (which forced people to use NAT boxes), or simply putting in the terms of service that people were not allowed to operate servers.
All these petty restrictions have prevented home users from being able to experiment with a wide variety of services that would be using more bandwidth --- eg widespread use of VNC or equivalents (so that the tech savvy can help their non-tech friends out easily), or kitting your home machine out with a microphone and speaker to establish a simple cheap home-security system.

Posted by: Maynard Handley on October 11, 2003 02:25 PM

I loved this article as it reminds us that expected future sales is one of the key elements of any valuation exercise. And anyone remotely familiar with valuing an enterprise should realize that high growth must be a short-run phenomena even for a new industry. Eventually, the new sector will find a finite portion of the world economy's budget. But then I remember reading an LA Times discussion where the reporter was told by the CEO of Cisco Systems that his DCF models justified his company's $550 billion valuation because of high expected growth. I wonder if any of these forecasters are now helping Art Laffer and his colleagues forecast U.S. economic growth and its implications for Federal tax revenues?

Posted by: Harold McClure on October 11, 2003 04:33 PM

An excerpt from an analyst addressing employees of AT&T, in the year 2040:

“I want to thank you for inviting me here today and tell you how honored I am to be in the room with the people who invented the phone, Unix, and so many revolutionary technologies and ideas.

First, a little history lesson. I think we all know that AT&T precipitated the new world. I see some wise old ‘heads in the audience who must remember the days when the Mighty Bell was singing Wall Street songs of love about how you were going to bring high-speed internet access to every vertebrate mammal on the crust of the earth; and in exchange, Wall Street wrote you love notes on the backs of bonds worth untold billions to finance the network rollout that would bring us all to convergence. Your stock went up; the stock of all the companies who sold you hardware went up; probably the stock of the people who sold you chairs, paperclips and ham sandwiches went up.

Then, a decade later, you were stuck telling Wall Street that long-distance telephone really was a good business, and would they forget about all that Internet stuff? Wall Street took your credit cards away, and all the companies you brought pretty toys from dried up and rolled away like tumbleweeds in a ghost town, and when they went down the entire market went pop like the big bubble that it was.

You were broke, I was broke, pretty soon the government was so broke they sold Wyoming to the Indonesians. We’re ruined. Ruined!

And it’s All. Your. Fault.

The most amazing thing happened: the audience laughed. She was so surprised she stopped talking, and the audience, seeing her look of surprise, laughed even harder.

She decided to push it. “And you didn’t even buy us a refrigerator box to live in. Thanks, Ma Bell.” She heard a thud as someone in the third row laughed so hard that when he doubled over his head hit the chair in front of him.

“Okay, so now there is no global telecommunications infrastructure in a traditional sense. No more wires, no more phone poles except in neighborhoods where it’s considered classy and retro, and the whole thing is kept up by the same guy who’s a model railroad buff and leads a Boy Scouts troop. Yet, you were able to contact me two days ago to fill in for some slacker who will remain unnamed, and I don’t think there were any ponies or smoke signals involved.

Space: it’s the final frontier, and millions of Americans can reach it from their backyard. Every person in this room is talking to hundreds of thousands of non-hardened personal sattelites put there by people whose previous high point of technical self-sufficiency was lighting the backyard barbecue.

And now you, AT&T, need to find a reason to live. A reason to get up in the morning. You were too big to die. But there’s only one reason for you to live: empire. When you built this building it was Rome. It was the Coliseum. You had dot-com companies kill each other in gladiator combat in your front yard so you’d have something to watch during lunch. Washington’s regulators fanned you with palm fronds when they weren’t busy getting you a cold frosty one. You had an empire, and you lost it. But you can have an empire again, and I’m going to tell you how.”

Posted by: Lisa Williams on October 11, 2003 10:18 PM

Hey! One (if not "the") reason the data streams to fill broadband channels have not arrived on schedule is "special interests" who captured the HDTV problem. Congress insisted that (1) there be only a single standard -- backwardly compatible with good ol' "I Love Lucy" vintage NTSC -- and (b) that the political process (rather than market, artistic, technical or -- theoretically -- religious processes) would define what that single standard should be.

Had the HDTV industry followed the historical pattern of home video (Beta, Laserdisc, VHS, etc) or PC (Commodore, Apple, Kaypro, etc) or
cell phone (TDMS, PCM, etc etc) there would have been a whole bunch of excess data to fill the cable, DSL, fiber, etc etc channels.

A great deal of it would be crap.

But eventually a Bill Gates-like winner would shake out and take over.


I suppose some think this would be WORSE than the five year delay we had while Congress dithered? That the simultaneous collapse of
all the tele-comm companies from "bad timing" was BETTER than market competition with actual
consumer-driven preferences?

Posted by: Pouncer on October 12, 2003 03:06 PM

While I have no strong feelings either way about the political machinations of HDTV in the US, in what possible sense is it backward-compatible with NTSC? The claim makes no sense at any technical level. The HDTV standard is digital, not analog. The signal modulation is different. Neither NTSC camera, nor editing, nor viewing equipment can be used for HDTV. Heck, even the freaking picture aspect ratio is different.

Look, it's fine to view the world through your ideological prism, but let's at least try to get the basic facts of the story correct.

Posted by: Maynard Handley on October 13, 2003 12:58 AM
Post a comment