CSFB Mark Rose's outstanding High Yield analysis of WCII:
[an Email (*not* a PM) should elicit a copy of the full pdf with tabled data]
April 19, 1999 Industry: Telecommunications
WinStar Communications (WCII) Rating: Strong Buy
Dispelling a few myths about an undervalued credit.
· Myth 1. WinStar is going to run out of money before it achieves
positive EBITDA. WinStar is well funded. Two major events, Lu-cent
financing and a Williams capacity contract, occurred in late
1998 to insure the company's liquidity until it achieves positive op-erating
cash flow in 2001.
· Myth 2. WinStar's strategic decision to accelerate its network
rollout and expand its scope to international markets is a mis-take
that will hurt bond investors. Accelerating its network build-out
positions WinStar to meet its top priority in 1999, which is add-ing
on-net customers. In non-U.S. markets, telecommunications
competition, data services, and the Internet are just beginning to
emerge. One of the critical missing resources is local bandwidth, and
WinStar can address this opportunity quickly with its broadband
wireless technology. But, if WinStar is going to address the signifi-cant
international opportunity, it needs to gather spectrum and build
its networks now.
· Myth 3. Only 20% of all WinStar phone lines are carried on its
network. The reality is that 34% of all of WinStar's local phone
lines are carried on its network, and we expect the company to reach
an on-net penetration of 52% (of local lines) by the end of 1999. The
20% on-net statistic includes data only and long distance only lines
in addition to WinStar's local lines. Long distance only customers do
not have an on-net local connection. Data lines are not on-net, but
data services provided over leased lines (off-net) are still profitable
for WinStar.
· Myth 4. WinStar overpaid Williams Communications for long
haul dark fiber. The WinStar-Williams dark fiber contract was a
good deal for WinStar. WinStar did not pay $10,600 per fiber mile,
and we have broken-out the details to explain the economics. Win-Star
will migrate most of its long haul traffic to the Williams net-work
by the second half of 1999, which should eliminate variable
leasing expenses.
Liquidity concerns and an aggressive expansion of its business plan
(Myths 1 and 2) are the primary reasons that WinStar's debt has traded
at a discount to other CLECs. We think WinStar's liquidity is sufficient,
and the market should gain confidence in its expanded business
plan as the Company achieves solid revenue and on-net access line
growth. WinStar did not fully participate in the recent rally of
CLEC securities in the high yield market therefore we find good
relative value now in WinStar's capital structure! We also expect
WinStar's debt and preferred securities to outperform other CLEC
credits over the next three quarters as investors regain confidence
in the Company.
Valuation
Why does WinStar paper trade at a discount to other CLECs?
WinStar announced a number of strategic decisions in December 1998
at a time when the high yield market was recovering from one of its
worst performances on record. Emerging telecom paper performed the
poorest of all sectors in the high yield market down turn, so investor
sentiment was very cautious toward emerging telecom carriers. We
think that WinStar's announcement of an expanded business plan ac-companied
by greater EBITDA losses was mistakenly received by re-covering
high yield investors like salt on a wound.
Today, we find good relative value in WinStar's capital structure. We
think WinStar is comparable to other developmental CLECs such as
Hyperion and GST Telecommunications. Being conservative, we will
only compare WinStar to the weaker trading credit GST. The GST
10 ½% Senior Discount Notes trade around 11.25% on a yield to worst.
WinStar's 14% Senior Discount Notes trade around 14.60% on a yield
to worst. We think that this +335 bps spread in the yields is too wide.
WinStar may deserve to trade behind GST in the near term because of
its recently expanded EBITDA losses, but we think an appropriate
spread is +100-150 bps between comparable debt securities of GST and
WinStar. We therefore have a near term price target of 85, or 12 ¾%
YTW, for WinStar 14% Senior Discount Notes.
Based on our price target for WinStar's 14% Senior Discount Notes, we
have established the following targets for the rest of WinStar's capital
structure. We also think WinStar will outperform many CLEC credits
as investors regain confidence in the Company's ability to execute its
business plan.
Myth 1. WinStar will run out of money before reaching positive EBITDA.
On page 2, we display our WinStar EBITDA model (Table 1). We do
not expect the Company to need capital until mid-2001. In 1999, we
forecast WinStar to have EBITDA losses of $282 million, cash interest
expense of $56 million, and capital expenditures of $550 million. Win-Star
should also receive payments from Williams Communications, which purchased
for $400 million a 25-year IRU for 2% of WinStar's long term broadband
capacity on a hub by hub basis. WinStar should receive approximately $222
million in payments from Williams in 1999. Williams will make payments to
WinStar on a proportionate basis to the construction of hubs until WinStar's
total hub count reaches 270. (The hub count is detailed in Table 3.)
Despite the delivery of all hubs in the next three years, WinStar will
recognize revenues for this contract over the entire 25 years.
The losses, capital expenditures, interest payments and Williams' pay-ments
make-up the $654 million loss in free cash flow before financing
that we have estimated for WinStar in 1999. The Company begins the
year with $313 million in cash and has secured a vendor financing line
of $2 billion from Lucent, which we have factored into our model. We
expect WinStar to fund approximately $445 million of its capital ex-penditures
in 1999 with the vendor line. In addition to its vendor fi-nancing,
WinStar has raised $168 million from an equity offering in the
first quarter. We expect WinStar to exit 1999 with approximately $265
million in cash.
In the year 2000, we expect WinStar to consume approximately $625
million in free cash flow losses before any financing activity, and to tap
its vendor financing line for another $450 million during the year. This
will leave the company with $90 million in cash at the end of 2000, and
an operation with annualized revenues of $800+ million. We expect
WinStar to reach EBITDA breakeven by the second quarter of 2001. In
our model, we expect WinStar to be able to raise approximately $400
million in capital to continue the funding of its business. (Observe the
additional issues line in Table 1.)
Myth 2. WinStar's decision to accelerate and expand its business plan in
December will hurt bond investors. WinStar made two strategic deci-sions
in December 1998 after it had secured $2 billion of Lucent fi-nancing
and a $400 million Williams contract. WinStar decided to
accelerate and expand its network build-out into 45 U.S. and 6 interna-tional
cities by the end of 1999 and another 15 U.S. and 6 international
cities in 2000. The end goal for WinStar is to provide facilities-based
network services in 60 U.S. and 50 international cities connecting cus-tomers
over its broadband wireless and fiber networks.
Accelerating its network build-out positions WinStar to meet its top
priority in 1999, which is adding on-net customers. In 1998, WinStar
was aggressively selling services ahead of its network installations.
Customers were delivered local service over the incumbent local ex-change
carriers (ILECs) network on a resale basis. Resale lines are un-profitable
with a gross margin of approximately 20% or less. WinStar
learned in 1998, like several other CLECs, that the costs are prohibitive
to convert resale lines to on-net, so preselling markets ahead of network
deployment became too expensive of a strategy to continue.
At the end of 1998, WinStar made the decision to de-emphasize the ag-gressive
growth of resale lines and accelerate the construction of its
network, which creates more addressable buildings to market high-margin
on-net services. WinStar's New Millennium Sales program is
also a key initiative that the company implemented in November of
1998 to drive growth in on-net lines. The refocus to on-net growth has
an impact to WinStar's near-term financial forecast. WinStar is not able
to capitalize all of its expenses related to network construction, so the
acceleration drives growth in operating expenses as well as capital ex-penditures.
Our forecasted capital expenditures, cost of sales, and
SG&A expenses have been increased for 1999 and 2000 and are de-tailed
in our model on page 2. As WinStar's new pace of network con-struction
matures in the second half of 1999, we expect its revenue
growth to outpace spending growth, so EBITDA losses should begin to
decline.
WinStar's decision to expand internationally also contributes to the in-crease
in expenses in our model. There is a significant opportunity to
provide affordable bandwidth in non-U.S. markets. For example, in
European countries, the cost of a domestic E-1 (the European equivalent
of a T-1) ranges from 5-13 times the cost in the U.S.. Besides the op-portunity
to provide affordable bandwidth in under-served non-U.S.
markets, spectrum is starting to become available through auctions and
grants by governments in various countries. If WinStar is going to build
an international business using broadband wireless technology, now is
the opportune time to gather the necessary spectrum licenses.
WinStar has gained licenses in Tokyo, Osaka, Buenos Aires, Amster-dam,
Rotterdam, London, Birmingham, and Manchester. WinStar plans
to offer data and closed voice services in its targeted international mar-kets,
but it will not offer public voice services. By focusing on data and
private voice, WinStar avoids the legal issues in gaining public voice
licenses, expenses related to Class V voice switches, and the necessity
to interconnect with the incumbent carriers. We view WinStar's inter-national
strategy as targeted, opportunistic in delivering affordable
bandwidth, and streamlined to maximize speed to market and minimize
capital investment.
Myth 3. Only 20% of WinStar's phone lines are carried over its network.
This statement is actually incorrect. WinStar actually carries 34% of all
its local customer lines on its network, and above, we have a detailed
analysis of WinStar's access lines and revenue segments.
WinStar includes three types of “lines” in its reported access line count,
and, unlike some CLECs, it does not use a multiplier on its access line
trunks to PBXs. WinStar's reported access lines includes local lines, data
only lines, and long distance only lines. A customer with WinStar's local
service may have additional services, such as Internet access or long
distance, but the customer's access lines are counted as local lines and
only in POTs or DS-0 increments. For example, a WinStar customer
with five local lines and long distance service would register five lines in
WinStar's line count.
Local Service - Only local lines are in WinStar's on-net line count.
During its fourth quarter 1998 conference call, WinStar indicated that it
had 20% of all its reported access lines, 64,000 lines, on its network. If
the on-net lines are viewed in proportion to WinStar's local lines instead
of total lines, the ratio is 34%. By the end of 1999, we forecast approxi-mately
357,000 local lines with 184,000 or 52% of all local lines on
WinStar's network. Currently, only two CLECs that we follow, GST and
Hyperion, have achieved on-net line penetration above this percentage.
Data Services – Exiting 1998, we estimate that WinStar had 87,725 ac-cess
line equivalents (in DS-0 increments), 28% of its total lines, carry-ing
data only services. If a customer receives WinStar's data services in
combination with local dial tone, the access lines are counted with local
telecom customers. For 1999, we estimate that WinStar will grow its
data service access lines at a faster pace than local services and reach a
total of 187,900, 32% of the total, by the end of the year.
Most of WinStar's data only customers are connected to its backbone
network through resold leased lines provided by the incumbent local
exchange carriers (ILEC). Unlike resold voice lines, resold data lines are
profitable and have a gross margin of approximately 50%. WinStar's
Frame Relay and ATM backbone network is comprised of leased long
haul DS-3 capacity connecting 30 points of presence which are listed on
the GoodNet home page good.net.
We expect WinStar to raise its capacity level in the near term to OC-3
and transition its data backbone to the Williams network over the next
few quarters, which should reduce its operating costs related to leasing
long haul capacity. With its own fiber from Williams, WinStar will be
able to increase its backbone capacity with the growth of customer de-mand.
See details of the Williams contract below.
Long Distance Services – WinStar's strategy is to sell long distance
services in bundled packages with its local and data services. WinStar
has sold long distance separately to customers that are located in their
targeted buildings in order to establish an account relationship. Cur-rently,
long distance only customers account for 10-13% of total access
lines, but we expect this segment to trend downward as a percentage of
the total because the WinStar sales incentive program is focused on re-warding
sales of local, data, and bundled packages.
CLEC Revenue Analysis – We have carried our line analysis steps
further to back into estimates of WinStar's data and voice revenues. In
1998, WinStar made several acquisitions to establish its Frame Relay
and ATM data services. For the fourth quarter of 1998, we estimate that
WinStar had $18 million in data service revenues, which is an average of
$82 per data line per month. For 1999, we have conservatively backed
down our average revenue per line per month expectations to $69 and
estimate that WinStar will achieve $115 million in data revenues, 240%
growth over 1998. We think that the company is close to a $100 million
annualized revenue run-rate for data services today, which compares
well with other CLECs that have recently announced an increased focus
on data services.
In voice revenues (local and long distance only lines contribute to this
segment), we estimate that WinStar achieved $37.4 million in the fourth
quarter of 1998 with average revenue per line per month of $59. We ex-pect
the average revenue per line per month for 1999 to stay in the high
$50s range and WinStar to achieve $223 million in voice related service
revenues in 1999, 66% growth over 1998.
New Media Revenues and EBITDA losses – In table 3,we have also
presented our forecast for WinStar's New Media services. In New Me-dia,
WinStar continues to invest more on Internet content initiatives,
such as its Office.com found at Yahoo!'s portal site. New Media lost
$8.4 million in EBITDA in 1998, but we estimate that $5-6 million of
the New Media EBITDA loss was due to Internet content initiatives and
the remainder related to film distribution and other traditional media
services. The company is currently reviewing its alternatives for capi-talizing
and realizing its deserved valuation for New Media's businesses.
Myth 4. WinStar overpaid Williams Communications for long haul dark fi-ber.
The WinStar-Williams dark fiber contract was a good deal for
WinStar. Displayed above (Table 4) is our analysis of the WinStar-Williams
dark fiber contract. WinStar committed to pay Williams $644
million over a seven-year period in payments of $7.6 million per month.
Williams is committed to deliver to WinStar four fibers over 14,684
miles for 25 years when the network is completed by the end of 2000.
The contract also includes capacity leases of specified circuits from
Williams for at least 20 years, collocation spaces, equipment, and net-work
maintenance. In addition, Williams will carry all of WinStar's long
distance and data traffic over the specific routes where WinStar is pur-chasing
fiber in the interim until the network is completed.
WinStar currently spends approximately $7.8 million per month on long
distance and data capacity and access fees related to its long distance
services, so WinStar has favorably locked in its costs at $7.6 million per
month for these services, which is a level below its current lease rates.
Furthermore, we expect WinStar's customer base and volumes to con-tinue
to expand in the future.
We estimate the value of the interim services portion of the Williams
contract to be $160 million. We estimate the fiber, long term capacity
leases, network maintenance, equipment, and collocation spaces portion
of the contract to total $270 million. The company will receive assets
and services ahead of payments, so there is an embedded interest cost in
the contract. If $430 million is the present value of the Williams contract
today, and WinStar is paying $7.6 million on a monthly basis for the
next seven years, the implied interest rate is 12.3% and total interest paid
is $214 million.
In Table 4, we have also presented an analysis of the price for WinStar's
dark fiber, equipment, collocations, and maintenance services. If the pre-sent
value is $270 million, then WinStar is paying the equivalent of
$49.50 per fiber mile per month over the life of its IRU using a 12.3%
cost of capital. This is an attractive price considering that WinStar is receiving
much more than just dark fiber strands. For example, we
follow a long haul fiber carrier in the Northeast called NorthEast Optic
Networks, NEON, and it typically charges $65 per fiber mile per month
for its dark fiber and maintenance services in long term lease contracts.
Also, WinStar retains a right with Williams to increase its network by
two strands for an additional $51 million. This would lower its monthly
per fiber mile valuation to $39.26, which is even more attractive. |
