AT&T-TCI Deal, By the Numbers


Back on June 26 -- the Friday after AT&T Corp. and
Tele-Communications Inc. announced their $48 billion merger plans -- AT&T chairman C.
Michael Armstrong and chief financial officer Dan Somers, along with TCI president and
chief operating officer Leo J. Hindery Jr., organized a conference call to
enthusiastically extol the deal and to emphatically refute any naysayers.

The call, which began after the markets closed on the East Coast, ran for more than
two-and-one-half hours. Along the way, Somers and Hindery, in particular, laid out many of
the specific projections that the two companies assumed to help justify AT&T's paying
an estimated 12 times 1999 cash flow for TCI.

Since then, the companies have submitted detailed merger
plans to the government, and they are observing a fairly strict "quiet period,"
designed to avoid hyping the deal, until those plans are in final, public form, probably
later this month. But here is what some analysts are saying about some of those
projections today.

AT&T Consumer Services Co. Revenue Forecast

Lehman Bros. Inc. forecasts 7 percent and 10 percent revenue growth at AT&T Corp. in
1999 and 2000, respectively, with most gains from business services.

Here's how the consumer unit, including the competitive long-distance operation, shapes

<p> <strong id="d9e34-15-b">1999<br> Est.</strong> </p><p> <strong id="d9e40-19-b">Revenue</strong> </p><p> <strong id="d9e46-23-b">Difference<br> (%)</strong> </p><p> <strong id="d9e52-27-b">Difference<br> ($)</strong> </p>





Traditional Long-Distance




















<p> <strong id="d9e204-111-b">2000<br> Est.</strong> </p><p> <strong id="d9e209-114-b">Revenue</strong> </p><p> <strong id="d9e214-117-b">Difference<br> (%)</strong> </p><p> <strong id="d9e219-120-b">Difference<br> ($)</strong> </p>





Traditional Long-Distance




















Source: Lehman Bros. Inc.


A highlight of the June 26 call was Somers' prediction that
by adding the ability to offer video, local-calling and high-speed-data services, customer
churn in the residential long-distance business could be reduced by one-third.

Even with those other offerings, long-distance will remain
AT&T Consumer Services Co.'s core service, generating 57 percent of its revenue and 40
percent of its cash flow in 2002, so cutting that business line's churn is a huge deal. It
would add $200 million in annual revenue and $100 million in yearly cash flow to
AT&T's coffers.

In a way, that 33 percent churn reduction represents a new
conservatism at AT&T. In January 1997, when AT&T launched its "WorldNet"
Internet service, it targeted its own long-distance customers on the strength of some
eye-popping survey data.

Six months into a trial involving 400,000 AT&T
long-distance customers who were offered WorldNet, churn reduction was 30 percent to 40
percent when compared with a control group that wasn't offered WorldNet. And that was just
by adding dial-up Internet service, billed separately.

But marketing experts cautioned that "bundling,"
or "packaging" disparate services, is tricky. It requires some consumer
education, especially when it's a new service, like digital satellite TV or a cable modem.
Consumers understand cash on the table, or big discounts, more easily.

Bruce Leichtman, a former cable-marketing executive who is
now senior analyst, consumer media research and consulting at The Yankee Group in
Cambridge, Mass., said poll data showed that about two-thirds of consumers want bundled
telecommunications services.

But start penciling in specific services and watch the
percentages fall away. If the offer links local and long-distance calling plans, about 50
percent of consumers are interested. A combination of local, long-distance and cable
service appeals to about one-third of consumers. When Internet service gets tossed in on
top of local, long-distance and cable, the percentages fall "down to the teens,"
Leichtman said.

"It's not just as simple as putting local and
long-distance and cable together and saying that customers are not going to leave,"
Leichtman said. "You have to have some kind of way of incenting them." The
easiest way is to offer a cable discount, but that cuts into revenue.

One hopeful sign comes from TCI's digital-cable
deployments. According to Hindery, subscriber churn for that product is only about the
same level as what cable experiences when subscribers change residences.

"I couldn't be more pleased than I am with digital
cable," he told analysts in discussing third-quarter results Nov. 13. "We are
far ahead of plan both in raw numbers and stickiness. It's a very sticky, gluey product.
It'll complement well Mike Armstrong's digital plans for this company in '99."


Back in June, Somers and Hindery were optimistic about
making solid gains in the basic-cable business, too. Where AT&T is able to offer
telephony and data services to consumers in TCI's cable regions, cable penetration should
grow by 5 percent by 2002. That should be worth an incremental $500 million in revenue and
$200 million in cash flow.

AT&T also believes that it can add a full 10 percentage
points to its long-distance market share within TCI's footprint, worth $500 million in
revenue and $125 million in cash flow.

Leaving aside the long-distance-penetration estimate, the
cable number is a big stretch. SG Cowen Securities Corp. analyst Gary Farber said his
pre-merger estimate was for TCI's cable penetration to rise from 61.3 percent in 1998 to
62.7 percent in 2002, or 1.6 percentage points.

To get to 5 incremental percentage points, rising to 67.7
percent, would require compound annual subscriber growth of 3.5 percent -- "well in
excess of the rest of the industry, which we expect to grow by 1.5 percent during the same
time," Farber wrote in a report.

The strongest basic-subscriber growth in the industry this
year, at least among the big MSOs, looks to be the 2.8 percent (7,600 subscribers)
year-over-year gain reported by Cablevision Systems Corp. in the third quarter.

TCI's third-quarter basic-subscriber growth was actually
flat, although its subscriber count, at 12.5 million, was up by just under 2 percent from
the figure a year ago. Digital subscriptions are rising quickly, hitting 1 million total
and on pace to hit 800,000 in the systems that TCI will continue to own and operate after
its deconsolidation efforts close.

Two recent analyst reports, both bullish on AT&T-TCI,
shrugged off those AT&T forecasts. Lehman Bros. Inc.'s Larry Petrella projected 0.8
percent and 1 percent pro forma subscriber growth in 1999 and 2000, respectively.
PaineWebber Inc.'s Eric Strumingher looked for 1.4 percent subscriber growth in 1999, 2000
and 2001, dipping to 1.3 percent in 2002.


Back in their chatty days in June, TCI and AT&T
officials were eager to bash analysts who tossed around estimates of $10 billion to $20
billion just to clean up TCI's plant. "The wildest of cap-ex [capital-expenditure]
numbers are running around," an apparently amazed Armstrong declared.

For the record, TCI said its two-way upgrades would cost a
total of $1.8 billion, including more than $400 million that had already been spent at
that time. TCI planned to speed up the effort, in deference to AT&T, but that
shouldn't affect the total dollars spent, the company said.

On top of that, AT&T expected to spend about $1.3
billion per year to maintain the cable plant that it inherits.

In its third-quarter report, TCI claimed to be on schedule
with its rebuild plan, which aims to get 90 percent of its owned-and-operated systems'
plant in two-way condition before the end of 2000.

At the end of the third quarter, about 2.3 million homes
were passed by two-way plant, up from 1.3 million homes July 30 and pushing toward 3.5
million -- 20 percent of TCI's pro forma base of 17.5 million homes -- at the end of the

Two-way-activated homes could be marketed
Internet-protocol-phone service "tomorrow," Hindery said, if all other
technological questions were answered.

Cable analyst Fred Moran, of ING Baring Furman Selz LLC,
said nothing has turned up to alter those estimates of capital spending on TCI's plant.
The surprise is how far TCI has come on its own and how fast it has stepped on the cap-ex

"They're clearly upgrading much faster than we
expected, so far," Moran said. The other key number that AT&T is eyeing is TCI's
1.2 million digital subscribers. "I don't think that any of us on Wall Street were
confident that they could unroll it as fast as they have."

TCI's capital expenditures hit $525 million in the quarter,
including about $60 million for "partnerships" that have since closed.

Sanford C. Bernstein & Co. analyst Tod Jacobs said:
"The real issue with TCI is that there's a feeling [that its] plant isn't in great
shape, and that they can't execute on a technological plan."

Tony Werner, TCI's executive vice president of engineering,
helped to counter that perception with the rapid conversion of TCI's headends for digital
cable, but "the company has to keep proving itself" in progress toward getting
the plant ready for AT&T, Jacobs said.

The third-quarter snapshot provided some
"confirmation" that TCI is progressing at least as well as expected. Jacobs said
he hasn't had to change any of his estimates.


This is a big one, considered by many to be the key to the
deal. In June, TCI and AT&T used an estimate of $400 to $500 per home to activate
telephone service. Those incremental costs kick in after a customer signs up.

The facilities-based phone service will start with
circuit-switched service in 1999. Hindery said, "You'll see a large number of our
major metropolitan markets doing circuit-switched launches in 1999." TCI sources have
indicated that this effort, led by former cable CEO Marvin Jones, will target 10 markets.

AT&T's recent order of cable-telephony gear from Arris
Interactive, the joint venture between Nortel and Antec Corp., was based on an average
equipment cost of $450 per customer.

Digital-cable customers who have advanced boxes, such as
General Instrument Corp.'s DCT-5000 terminals, will already have much of the technology
needed for IP-phone service.

For those customers, the incremental cost to activate
IP-phone service is more like $250, according to Hindery. As a result, he said he was
"comfortable" using $350 as the average cost to wire a customer for IP-phone

That tracks with the IP-phone incremental-cost average of
$325 that Jacobs has been using for the per-home cost -- including headend switching and
routing gear, as well as customer equipment and installation.

Strumingher had a contrary view. He called IP telephony a
"smoke screen" that AT&T and TCI are using to give investors a reason to
believe in the combination.

In reality, Strumingher wrote, issues like writing
standards for IP routers in cable headends to communicate with telco switches could keep
IP telephony from being extensively deployed for "several years."

But the good news, from Strumingher's perspective, is that
AT&T doesn't really need to offer IP telephony. He thinks that the threat of IP
telephony, combined with pressure from courts and regulators, could persuade the regional
Bell operating companies to offer deep discounts to AT&T and others that want to
resell Bell capacity.