Morgan Stanley report on Telstra & Structural Separation

M O R G A N  S T A N L E Y  R E S E A R C H A S I A / P A C I F I C

Morgan Stanley Australia Limited+  Sachin Gupta, CFA

S.Gupta@morganstanley.com +61 3 9256 8942

Morgan Stanley Asia Limited+  Navin Killa

Navin.Killa@morganstanley.com +852 2848 5422

September 21, 2007

Stock Rating  Telstra Corporation

Underweight

Overcoming FTTN

ICnaduutsitoryu sV iew  Uncertainty  Key Ratios and Statistics

Reuters: TLS.AX Bloomberg: TLS AU

Australia Telecommunications

Conclusion: Telstra, as an integrated telco, continues   Price target  A$3.94 to face various operational, transformation and  Shr price, close (Sep 20, 2007)  A$4.40 regulatory uncertainties in the near term. The debate on  Mkt cap, curr (mn)  A$54,750 fibre-to-the-node investment might also not be resolved  52-Week Range  A$4.97-3.52 any time soon due to disagreements on access prices.  Sh out, basic, curr (mn)  12,443.1 EV, curr (mn)  A$68,540

• The economics of FTTN are reasonably weak for  Net debt/cap (08e) (%)  53.6 Telstra, as (a) there is limited incremental ARPU; and  ROE (08e) (%)  29.0

(b) it will cannibalize its existing fixed-line revenues.  Shrs out, basic, per-end (08e) (mn)  12,443 S'hldr eqty (08e) (mn)  A$12,243

Based purely on incremental return analysis, we

RNOA (08e) (%)  16.8 estimate the FTTN investment is unlikely to generate

a positive return in the first three years of investment.  Fiscal Year (Jun)  2007 2008e 2009e 2010e

• Alternative carriers also cannot make this investment  ModelWare EPS (A$)*  0.30 0.29 0.28 0.29 without Telstra's co-operation, as issues like sub-loop  EPS, basic, rpt'd (A$)  0.26 0.27 0.27 0.28 unbundling, co-location and backhaul are complex,  Revenue, net (A$ mn)  23,950 24,434 24,864 25,127 costly, and require access to Telstra's networks.  ModelWare net inc (A$ mn)  3,716 3,573 3,447 3,612 P/E  15.4 15.3 15.9 15.2

• Another emerging risk is if ALP wins the next election  P/BV  4.6 4.5 4.5 4.5 and implements an open-access' broadband  EV/EBITDA  6.8 6.7 6.7 6.6 framework, we estimate ~A$600-900mn in revenues  D* =ivPylelda s(e%s)e e explanation of Morgan Stanle6y. 1ModelWa6r.e4later in t6h.is4note.  6.4

are at risk.  e = Morgan Stanley Research estimates

A possible way of overcoming this investment,

operational, and regulatory uncertainty, could be to

structurally separate the Networks and Services

business, in our view. Similar to our TCNZ analysis in

May this year – we estimate this could result in ~20%

upside to our base-case valuation, or ~10% upside to

the current share price. There is no political will at this

point to review this scenario; however, a change of

government at the next election could be a catalyst to  Mcoomrgpaann iSetsacnolevye rdeodeisn aitnsdr esseeeakrsc htor edpoobrtuss. iAnessas rwesituhl t, revisit this debate. A vertical split of an integrated carrier  investors should be aware that the firm may have a

like Telstra would be unprecedented and a complex  conflict of interest that could affect the objectivity of this exercise. A number of different structures are possible,  report. Investors should consider this report as only a with varying valuation outcomes.  single factor in making their investment decision.

Customers of Morgan Stanley in the U.S. can receive We have made no changes to our earnings  independent, third-party research on the company

covered in this report, at no cost to them, where such forecasts, and maintain our Underweight call. In  research is available. Customers can access this

view of the current structure, we do not believe the stock  independent research at

is cheap at 15.9x F2009e earnings. The risks in the near  www.morganstanley.com/equityresearch or can call term remain high – especially with rising competitive  1-800-624-2063 to request a copy of this research.

pressures and possible delays in transformation  For analyst certification and other important benefits.  disclosures, refer to the Disclosure Section.

+= Analysts employed by non-U.S. affiliates are not registered pursuant to NASD/NYSE rules.

Telstra Financial Summary

Exhibit 1

Telstra Financial Summary (Year End June 30)

Income Statement 2006 2007 2008E 2009E

Normalised Growth (%)

Sales  2.4% 4.3% 2.3% 1.8% EBITDA  -3.1% -0.3% 0.9% 0.6% Operating profit  -7.0% -2.6% -2.2% -1.8% NPAT  -11.1% -5.5% -4.1% -3.5%

Profitability (%)

EBITDA  45.1% 43.1% 42.6% 42.1% Operating profit  29.0% 27.1% 25.9% 25.0% NPAT 17.4% 15.7% 14.8% 14.0% ROE  29.8% 29.3% 28.5% 27.8% ROA  18.5% 17.3% 16.3% 15.7% ROIC 18.0% 17.2% 16.3% 15.8%

Stability (%)

Net debt to equity  99.2% 107.6% 115.5% 121.1% Net debt/(net debt+equity) 49.8% 51.8% 53.6% 54.8% Current ratio   0.62 0.57 0.58 0.57 Interest coverage (X)   11.0 9.4 8.8 8.4

Operating Statistics 2006 2007 2008E 2009E Fixed Lines (m) 9.9 9.8 9.5 9.2 Blended Fixed ARPU ($/mth) 62.3 60.8 59.4 58.3 Aust Mobile Subs (m) 8.43 8.88 9.27 9.56 Blended Mobile ARPU ($/mth) 44.5 45.4 45.0 45.0 No. Employees 44,452 43,411 40,387 38,161

E = Morgan Stanley Research estimates

Source: Company data, Morgan Stanley Research

Balance Sheet 2006 2007 2008E 2009E Cash/deposits 689 823 930 898 Accounts receivable 3,721 3,891 3,982 4,053 Inventory 224 332 332 332 Other current assets 265 307 307 307 Current assets 4,899 5,353 5,550 5,589

Fixed assets 23,592 24,607 25,045 25,346 Goodwill 2,073 2,126 2,126 2,126 Investment & other assets 5,660 5,789 6,280 6,434 Non-current assets 31,325 32,522 33,451 33,906 Total assets 36,224 37,875 39,001 39,495

Accounts payable 3,570 4,207 4,419 4,555 Short-term borrowings 1,982 2,743 2,743 2,743 Other current liabilities 2,347 2,484 2,484 2,484 Current liabilities 7,899 9,434 9,646 9,782

Long-term borrowings 11,442 11,619 12,619 13,119 Other LT liabilities 4,049 4,242 4,242 4,242 Non-current liabilities 15,491 15,861 16,861 17,361 Total Liabilities 23,390 25,295 26,507 27,143

Net Assets 12,834 12,580 12,494 12,352

Paid in capital 5,569 5,611 5,611 5,611 Reserves (160) (258) (258) (258) Retained earnings 7,177 6,976 6,890 6,748 Minorities 246 251 251 251 Total shareholders' equity 12,832 12,580 12,494 12,352

Cash Flow 2006 2007 2008E 2009E EBITDA 9,551 9,858 10,054 10,220 Tax paid  (1,882) (1,618) (1,456) (1,432) Interest paid  (933) (1,087) (1,171) (1,237) Other 926 367 121 65 Operating Cash Flow 7,662 7,520 7,549 7,616

Capital expenditure  (4,255) (5,652) (4,858) (4,561) Investments  (48) (330) (100) (103) Divestments 255 305 0 0 Free Cash Flow  3,614 1,843 2,591 2,952

Dividends paid  (4,970) (3,479) (3,484) (3,484) Debt increase/(reduction) 487 1,778 1,000 500 Equity Issued 0 0 0 0 Net Cash Flow  (869) 142 107 (32)

Investment Case

Summary & Conclusions

Telstra, as an integrated telco, continues to face various operational, transformation and regulatory uncertainties over the medium term. FTTN is a key issue that might not be resolved any time soon due to disagreements on access prices. Telstra has suggested a price of A$59/mth for a 512k connection, the G9 has proposed prices ranging from A$25-45/mth and the government is suggesting a price of A$35-60/mth for rural broadband (for OPEL JV).

We do not believe Telstra is likely to make this investment, unless the access prices are high enough to partially compensate for existing wholesale revenues

and generate a reasonable return on the new investment. The economics of FTTN are reasonably weak, as (a) there is limited incremental ARPU; and (b) it will cannibalize its existing high margin fixed-line revenues.

• Based purely on incremental return analysis, we estimate the FTTN investment is unlikely to generate a positive return in the first three years of investment. Even then, the returns generated in the medium-term of 6-7% are below Telstra's WACC of 9.1%.

• On a standalone basis, we estimate that to generate a return of 14-16% on a A$4 billion FTTN investment, the average access price would need to be A$37-39/mth, including a A$10/mth charge from node-to-home. However, we estimate Telstra currently generates an average ARPU of A$84/mth from a wholesale DSL customer (voice and data), therefore a price of A$37-39 would result in a loss of A$45-47/mth ARPU.

• Alternative carriers (G9) also cannot make this investment without Telstra's co-operation, as issues like sub-loop unbundling, co-location and backhaul prices and terms are complex and costly, and require access to Telstra's networks.

• Another emerging risk is if ALP wins the next election and implements an open-access' broadband framework, we believe Telstra could lose up to A$600-900mn in annual revenues.

Our European team recently downgraded BT Group's (BT.L, 308.25p) rating to Underweight, with one of the key reasons being FTTN uncertainty – see Fibre Risk Awakens: EPS Support Fade: Underweight, for more details (20 Sept 2007). ULL migration is also accelerating – 90k in recent weeks, up from 50-60k previously. Telstra faces the same risks.

One possible way of overcoming this investment and operational uncertainty for Telstra, and to reduce the risk of further regulatory intervention, could be to structurally separate the Networks and Services business, in our view.

Similar to the analysis we undertook for TCNZ in May this year (see "Unlocking Value through Structural Separation", dated May 21, 2007) – we estimate this could result in around 20% upside to our base case Telstra valuation of A$3.94, or 10% upside to the current share price.

A vertical split of an integrated carrier like Telstra is unprecedented and a hugely complex exercise. A number of different structures are possible, with varying valuation outcomes. In our simple analysis, we have assumed that the hypothetical separation results in the creation of two separate entities – one focused on maintaining the access network businesses (Networks), and the other on selling services to consumers (Services). We have valued these businesses on an EV/EBITDA basis, along with Telstra's other businesses like mobiles, Sensis and Foxtel. This is summarized in Exhibit 2 below.

The company, the regulators and the government have not made any official comments in relation to this. However, we think that a change of government at the next election could be a catalyst for this debate to start. Telstra may itself review a separation scenario as a strategy to unlock/preserve value. We will monitor any such debate closely – and have not assumed this as our base case.

One of the main debates on Telstra is if the company can meet its 2010 targets for which the market remains divided. Telstra is holding another strategy day on November 1, where network transformation is likely to be a key focus. All these debates will become irrelevant and the long-term outlook will change if: (a) there is a FTTN decision beforehand; and/or; (b) if there is further intervention from the regulator/government, which could result in a structural separation of the company.

We have made no changes to our earnings forecasts, and maintain our Underweight call. In view of the current company structure, we do not believe the stock is cheap at 15.9x F2009e earnings. The risks in the near term remain high – especially with rising competitive pressures and possible delays in transformation benefits.

Exhibit 2

Telstra Hypothetical Break-Up Value

Source: Company data, Morgan Stanley Research. E = Morgan Stanley Research estimates  

Why Is FTTN Return Dilutive?

FTTN is an incremental investment, for which Telstra needs to generate an incremental return to get a positive return. It is difficult to see what additional ARPUs could be generated via FTTN, which is currently not possible on copper. It is not obvious that customers would pay significantly more for broadband speeds over above those delivered by ADSL2+ services. In fact, services like IP-TV would only mitigate some of the revenue decline, and given the excess capacity generated by fibre networks, it could even accelerate the rate of decline in broadband prices.

In our view, the two main issues for FTTN are: (a) wholesale access prices; and (b) access terms. Both these issues are unlikely to be resolved any time soon. The problem is that the access price, which Telstra needs and wants to charge on FTTN, is not acceptable to the regulators. On the other hand, the competitive carriers (like the G9) cannot make this investment happen without Telstra's co-operation. We make the following points:

1. Based purely on incremental return analysis on FTTN, we estimate this investment is unlikely to generate a positive return in the first three years of investment.
2. On a standalone basis, we estimate that to generate a return of 14-16% on an A$4bn FTTN investment, the average access price would need to be A$37-39/mth, including an A$10/mth charge from node-to-home. However, we estimate Telstra currently generates an average ARPU of A$84/mth from a wholesale DSL customer (voice and data), therefore a price of A$37-39 would result in a loss of A$45-47/mth ARPU.
3. The G9 cannot make this investment happen without Telstra's co-operation. Issues like sub-loop unbundling, co-location and backhaul are complex and costly.
4. Cost savings and market share outcomes alone can be viable reasons for this investment. We estimate if Telstra generates A$200-300mn in annual cost savings, increases its annual retail broadband share by 3%, and slows down the rate of PSTN decline to 2-3% per annum, the breakeven price for FTTN is A$50-60/mth.

We review each of these points in more detail below.

1. FTTN Economics

The economics of Telstra's fibre-to-the-node (FTTN) as an incremental investment are relatively weak. We estimate this investment is unlikely to generate a positive return in the first three years of investment. Even then, the returns generated in the medium-term of 6-7% remain below Telstra's WACC of 9.1%. See Exhibit 4 for more details.

It is important to note that this analysis is based purely on incremental returns on FTTN – and not considering other benefits like market share loss outcomes or cost savings from FTTN.

Key assumptions to note are:

• We assume capex per sub or A$1,000. Optus recently estimated that access capex per sub could be as high as A$2,500 for FTTH.

• Telstra increases its retail share to 60% from ~47% now.

• Incremental ARPU per retail sub of A$20/mth. We have not assumed any incremental ARPU per wholesale sub

– as FTTN access charges will be a replacement of ULL charges, and are likely to be lower than the current wholesale ARPU of A$84/mth. However, if we do assume an additional A$10/mth in incremental wholesale ARPU, the FTTN investment becomes positive in year two.

Exhibit 3

FTTN Build Economics

Source: Optus Presentation – Australian Telecom Summit 2007

Exhibit 4

FTTN "Incremental" Return Analysis

2008E  2009E  2010E 2011E 2012E 2013E  2014E 2015E

DSL Subscriber Base  4.1  4.7  5.3 5.8 6.2 6.6  6.9 7.3

% Metro  80  80  80 80 80 80  80 80

% Rural  20  20  20 20 20 20  20 20

Potential FTTN Subs  0.8 1.8 3.0 5.2  5.5 5.9

% FTTN  20 40 60 100  100 100

Australian Households  8.5  8.7  8.9 9.1 9.2 9.4  9.6 9.8

Homes Passed  1.74  2.66 3.62 4.62 4.71  4.80 4.90

% Homes with FTTN  20  30 40 50 50  50 50

Capex/Sub  1,000  1,000 1,000 1,000 1,000  1,000 1,000 Total Capex  1,741  923 959 996 92  94 96 Cumulative Capex  2,663 3,622 4,618 4,711  4,805 4,901

Subscriber Mix

% Wholesale  40  40 40 40 40  40 40

% Retail  60  60 60 60 60  60 60

Wholesale FTTN Subs  -  0.3 0.7 1.2 2.1  2.2 2.3 Retail FTTN Subs  -  0.5 1.1 1.8 3.1  3.3 3.5 Total FTTN Subs  -  0.8 1.8 3.0 5.2  5.5 5.9

Incremental ARPUs

Wholesale  0 0 0 0  0 0 Retail  20 20 20 20  20 20

Total revenues  122 194 349 593  777 821 EBITDA  73 116 209 356  466 493

% Margin  60 60 60 60  60 60

D&A  133 181 231 236  240 245 EBIT  (60) (65) (22) 120  226 248

Cumulative Investment  2,663 3,622 4,618 4,711  4,805 4,901 Cumulative Depreciation  133 314 545 781  1,021 1,266 Net Investment  2,530 3,308 4,073 3,930  3,784 3,635

Incremental Return (%)  -2 -2 -1 3  6 7

Source: Company data, Morgan Stanley Research, E = Morgan Stanley Estimates

2. FTTN Price – Wholesale DSL vs. ULL vs. FTTN  Exhibit 5 On a standalone basis (irrespective of who makes this

G9 Proposed Price s

investment), to generate a return of 14-16% (post-tax) on a

A$4bn investment, we estimate an average access price of

A$27-29/mth would be needed, and including a further

A$10/mth node-to-home charge, the total average wholesale

price would be A$37-39/mth.

We have used 5mn broadband subscribers in our calculations

– which we estimate is the size of the total metro broadband market over the medium term.

This is similar to the prices proposed by G9 in their Special Access Undertaking in May 2007. Therefore, it appears that the G9's calculations are also based on a standalone  investment.  

90

$84

80

However, this investment cannot be considered in  70

isolation for Telstra. Telstra has legacy revenues to protect

60 $50-60 which will largely be replaced if FTTN investment is made.

50

We estimate Telstra currently generates an average ARPU of  40

A$84/mth from a wholesale DSL customer (voice and data),  30

therefore a price of A$37-39 would result in a loss of  20

A$45-47/mth ARPU. Or, Telstra currently generates around  10 $14.5

A$2bn in wholesale revenues – therefore a price of

A$37-39/mth would imply a loss of around 50% of these  0 Wholesale DSL ULL Wholesale FTTN revenues. Moreover, Telstra would fail to generate a return on

this investment.  Source: Company data, Morgan Stanley Research

Without FTTN, ULL is a major risk for Telstra. As shown in Exhibit 6, moving from wholesale DSL to ULL, Telstra could potentially lose A$70 ARPU per subscriber per month. With FTTN, we estimate Telstra would need to charge an average access price of A$50-60/mth for it to be value neutral – and this could be a much better outcome than ULL.

Exhibit 7

FTTN: RoIC and Monthly Charge for "Standalone" Investment

Source: Company data, Morgan Stanley Research

3. Issues for Alternative Carriers  4. Cost Saving and Higher Market Share Potential

We believe it is difficult for alternative carriers (other than  Although the economics of FTTN are relatively weak (based Telstra) to make this investment happen without Telstra's  on incremental revenues), the potential for lower operating involvement and co-operation.  costs and the need to defend market share, can be two

possible reasons for this investment.

As shown in Exhibit 7, the average access price of A$37-39/mth is based on a 5mn total subscriber base. If we were to exclude Telstra's market share of ~50%, then the average price rises to ~A$75/mth – which would result in higher retail prices – not the desired outcome.

The G9 cannot make this investment without accessing Telstra's infrastructure. There are several other issues including sub-loop unbundling, co-location and backhaul – for which prices and access terms are important.

In a recent study on sub-loop unbundling in the Netherlands by Analysys Consulting for the Dutch market, it estimates that a business case for an access seeker with similar economic viability to that of continuing use of ULL for 60% of population would require both:

• a market share > 55% of all broadband lines (including cable); and

• a significant increase in ARPUs (incremental ARPU of 10/mth).

Therefore, the economics under sub-loop unbundling are difficult from an access seeker's perspective. The market size for an access seeker is much more limited under an FTTN model versus ULL. For example, under a ULL model, the potential market for an access seeker is the entire customer base around an exchange – whereas under an FTTN model, the potential market is reduced to 1-2 streets.

Exhibit 8

Sub Loop Unbundling Structure

Source: Analysys, January 2007

We estimate if Telstra generates A$200-300mn in annual cost savings, increases its annual retail broadband share by 3%, and slows down the rate of PSTN decline to 1-2% per annum, the breakeven price for FTTN is A$50-60/mth. As highlighted in Exhibit 11, we estimate:

• To maintain the current wholesale revenue of A$2bn, the average access price would need to be A$40-45/mth. This includes wholesale FTTN revenues, node-to-home charge of A$10/mth, some existing DSL lines, and ~A$200mn in additional revenues as a result of higher retail broadband share and slowdown in PSTN decline.

• If Telstra generates A$200-300mn in cost savings, it implies that for an average price of A$27-30/mth, this investment can be earnings breakeven. This price, although it would be earnings breakeven, means that this incremental investment has not generated an incremental return.

• Assuming a ROIC of 14% on a new investment is reasonable, and implies an average price of A$27/mth – as per Exhibit 7.

• Therefore, for FTTN to be value breakeven, we estimate an average access price of between A$50-60/mth is required.

The question remains if this price will be acceptable to Telstra or the regulators. The average breakeven price could be lower if the cost saving potential is higher than our forecasts, or if Telstra wins more market share.

As highlighted in Exhibits 9-10, different carriers have varying estimates of cost savings from FTTN/FTTP. Verizon estimates FTTP can reduce network costs by 50-60%, while AT&T estimates 38% lower costs from FTTN. Most of these cost savings are driven by lower plant maintenance, installation and customer care costs, as DSLAMs are moved closer to the customer.

Exhibit 9  Exhibit 10

FTTP Cost Savings  FTTN Cost Savings

Source: Verizon  Source: AT&T

Exhibit 11

Revenue Impact from FTTN

Wholesale FTTN  Addition Node

FTTN Price  Lines  Charge Wholesale DSL Lines FTTN Benefit Total Revenue Net Impact

2.2  @ A$10/mth 0.5 2009e

$25.0  660  264 367 197 1,488 (401) $27.5  726  264 367 197 1,554 (335) $30.0  792  264 367 197 1,620 (269) $32.5  858  264 367 197 1,686 (203) $35.0  924  264 367 197 1,752 (137) $37.5  990  264 367 197 1,818 (71)

Source: Company data, Morgan Stanley Research

Overcoming FTTN Uncertainty A Possible Break-up?

As highlighted in the previous section, we do not believe FTTN investment will happen in the current structure, which may be more feasible under a structurally separated Networks business.

We acknowledge there is no political will at this point to review this scenario – however a change in government (or even if the current government is re-elected), could be a catalyst to revisit this debate. Telstra may itself review a separation scenario as a strategy to unlock/preserve value. We will monitor any such debate closely – and have not assumed this as our base case.

Globally, structural separation is one of the emerging debates, as incumbent carriers are finding it difficult to justify incremental investment in networks without destroying overall returns, and as a possible way of unlocking some value.

• The key argument in favor of it is that it removes behavioral obstacles to competition, resulting in more competition, investment and innovation.

• The key argument against it is that it could possibly result in reduced infrastructure based competition, and requires high implementation costs.

A separation of Telstra was considered back in 2003; however, it was short-lived, due to the expected implementation costs, and was considered to destroy shareholder value. The Minister at the time decided to implement accounting separation' instead, and relied on the Trade Practices Act (Part XIC and XIB) to address competition issues in the country. There were also concerns at the time that the incumbent telecoms are so integrated that it is impractical to separate Telstra's network and services businesses.

Most of the submissions lodged by various industry participants at the time conceptually agreed with the merits of a separation, but recommended against it due to the practical challenges associated with implementing such reforms.

Telstra at the time estimated total separation costs of up to A$2bn with on-going costs of A$80mn per annum. This was based on previous estimates made by Verizon in the US. Evidence since then has been that operational costs are much lower – BT's total operational separation costs in 2006 were 70mn (A$175mn), and TCNZ estimates its operational

separation costs could range around NZ$200-500mn. There is not likely to be too much difference between operational and structural separation costs. Also, there is an argument that with the transformation of networks to IP platforms, separation should be relatively easier and cheaper.

What was Telstra's View in 2003?

Telstra in its submission in 2003 (Parliamentary Inquiry into Structural Separation, Submission No 59), strongly recommended against structural separation and concluded, "It would lead to a reduction in national efficiency, an increase in telecommunications costs and to higher prices for consumers. Breaking up the Telstra business and network would eliminate the economic efficiency benefits that come from operational integration. Creating an artificial boundary line between what Telstra could and could not do would inevitably be arbitrary and the long-term effects damaging.

Its four conclusions were:

1. Any break-up of the network would be arbitrary and impose significant structural rigidities, which would hamper innovation and technological improvements.
2. Structural separation would impose significant costs on Australian consumers.
3. Structural separation will reduce the operating efficiencies that are currently used to help fund uneconomic services, particularly in rural and remote Australia.
4. Structural separation will send strong negative signals to investors, especially international investors, as such a dramatic policy intervention increases sovereign risk and runs against the trend of regulation elsewhere around the world.

Of the above, we agree with (3). As shown in the previous section, in the current structure it is perhaps even more difficult for Telstra to make investments that would address technological improvements, unless the access prices are very high – which is not likely.

FTTN is one reason  1.  ADSL2+ won't improve speeds for all - Analysis from Telstra, its competitors, regulators and the government have  the Broadband Stakeholder Group suggests that

not been able to reach a decision on FTTN, and the issue is  ADSL2+ with speeds of 10Mbps and above is only likely not likely to be resolved any time soon, in our view. At the  to be available to around 30% of households and that same time, operational risks for Telstra appear to be rising  currently true 8Mb speeds are only available to 20% of from ULL.  households.

FTTN economics do not stack up for Telstra unless the access prices are very high, which the ACCC/coalition/ALP is not going to allow. It is an incremental investment with limited scope for incremental returns. In the current structure, (a) it is difficult to see this investment happening any time soon; and (b) it is difficult to see it being value accretive for Telstra. One of the main reasons why Telstra would want to make this investment is to reduce the risk of ULL, we believe.

At the last result, there were early signs of the ULL threat becoming more transparent. Telstra reported a 100% increase in ULL SIOs to 239k, which only added A$8mn in additional ULL revenues. This reflected lower access prices during the year and significant revenue dilution from moving from wholesale to ULL. Domestic wholesale access lines dropped 180k – of which ULL represented 2/3rd of the decline. With FTTN outcome delayed until mid-next year, and ULL prices coming down further, we expect the rate of wholesale access line decline to accelerate, especially now with Optus also focusing on expanding its on-net subscribers.

As stated earlier, FTTN is even more difficult for the other carriers (like G9). Issues like sub-loop unbundling, co-location at the node, backhaul, etc. are complex, costly, and require Telstra's co-operation. More importantly, for access seekers, the market size around a node is much smaller than around an exchange via ULL, therefore pricing and access terms are far more important under FTTN.

Our European team recently downgraded BT Group's rating to Underweight, with one of the key reasons being FTTN uncertainty, which will negatively impact cash flows (see "Fibre Risk Awakens: EPS Supports Fade: Underweight", dated September 20, 2007, for more details).

The ULL migration rate in the UK is also accelerating – 90k in recent weeks, up from 50-60k previously. Telstra faces the same risks.

Although the team believes that an FTTN rollout could be years away for BT, the key driver is likely to be rising competition issues. The three reasons highlighted are:

2. Competitive forces are at work: Mobile broadband, via increased sales of datacards, HSPA network upgrades, and mobile network refarming, will continue to force fixed networks to maintain a bandwidth advantage' over the longer term.
3. UK behind on fibre despite being ahead on broadband penetration. An increasing number of European countries have already begun fibre investment.

Once again, the economics are reasonably weak, and it is not clear if the regulator will provide some incentives. Recent comments from Ed Richards (new head of Ofcom) at the US congressional sub-committee on fibre investment in April 2007 suggest regulatory incentivisation in this regard is not necessary:

" We have looked at going much further and introducing policies of regulatory forbearance. Some incumbents in Europe, though not BT, have called for regulatory holidays' for NGN investments essentially the removal of all pro-competition rules. We do not agree that this is necessary to secure NGN investments, and we think the price of such a policy in a UK context would be extremely high. We would be sacrificing competition in return for an investment that BT can and will make in any event."

Exhibit 12

BT's ULL Weekly Net Additions

Source: Company data, Morgan Stanley Research

ALP's Open-Access Is Another  Exhibit 13

As we have seen in regional/rural areas, government  Telstra ULL vs. Wholesale SIOs subsidies are one way of encouraging these investments

2.5

1.5 1.2 0.9 0.6 0.3 -

while keeping a lid on access prices. Perhaps government

subsidies are also what are required in metro areas to see this

2.0

investment happen. Or, as the ALP is proposing, an "open

1.5

access" public/private structure for FTTN investment. This

makes sense for consumers, competitors, regulators, but not

1.0

for Telstra, in our view.

0.5 We estimate Telstra could lose up to A$600-900mn in

revenues in an open-access' broadband framework.  -

05 06 07 E E E E

20 20 20 011 012 013 014

• There are currently 2mn wholesale subscribers  2008E 2009E 2010E 2 2 2 2 2015E generating total ARPU (voice + data) of A$84/mth, of  Wholesale Lines (M) ULL Lines (M)

which we estimate 50% could be at risk (based on current  ESo=urMceo:r gCaonmSptaannyledy aRtae,sMeaorrcgha ne sStitmanalteeys Research

retail shares) A$1bn revenue at risk;

• Telstra generates A$6.3bn in PSTN and retail broadband  

revenues, which could come under pricing pressure, and  EPxohitbeit n14t ial Revenue Impact with ALP's Open-Access we estimate these could be impacted by 5-10% loss of

A$300-600mn in revenues;

• This is offset by a A$10/mth node-to-home charge for wholesale subs benefit of A$120mn;

In a public-private partnership structure (as proposed), assuming the government will have approximately 50% share, and for the rest of the market, Telstra maintains its 50% retail broadband share (at A$40/mth average FTTN charge) benefit of A$600mn;

• Net impact: A$600-900mn revenue loss.

Telstra can and has the capacity to make this investment on its own without any subsidies; however, the problem is that the pricing it needs and wants to charge to get an economical return would result in higher consumer prices – not the outcome desired by the regulators. This is a dilemma for Telstra, as without this investment, ULL risks become more real.

Source: Company data, Morgan Stanley Research

In our view, a structural break-up could be the answer. This would allow the company to separate its Network business into a different entity that could charge regulated access prices to all carriers, including Telstra. This is similar to the regulated access prices that BT's Openreach is allowed to charge.

Even the Current Government Could Re-Think  In our view, the advantages of a break-up for Telstra could The relationship between Telstra and the current government  be:

is not at its best. The government wants to see further

network investments at a reasonable price, and both parties  The Network business could achieve more certainty on have different views on what is reasonable. The government  regulations and investments, and could be rated as a has commissioned an Expert Taskforce to review the FTTN  regulated utility given its stable cash flows;

proposals, and has explicitly stated that the Taskforce is to

have particular regard to:  Similarly, the Service business could be perceived as a

growth business – with a better growth profile than Telstra

1. the government's strong commitment to robust  as a whole entity;

competition and the long-term interests of end-users;

• Separation could result in more efficient capital allocation

2. the commitment that all people in Australia have access  based on the levels of risks; and

to quality telecommunications services at affordable

prices; and  It could remove regulations in the Services business.

3. the need for investors to earn returns on their  The disadvantages, however, could be:

investment commensurate with the cost of their

investment and their risks.  Separation could see more retail competition but reduce infrastructure-based competition – there are inherent

The government has also asked for consideration of access  synergy benefits between networks and service

price and non-price terms; potential for compensation for  provision, which encourage access seekers to make stranded DSLAM investments; sub-loop unbundling price;  more investments.

rural subsidies; and risks of duplicate networks. These are

extensive and complicated issues, which will require  It will be a costly and time-consuming exercise. BT significant input from all participants.  estimates Netco separation costs of 70mn (A$175mn),

and TCNZ estimates its operational separation could

In the recent OPEL decision, the Minister (Senator Helen  range from NZ$200-500mn. A separated network would Coonan) stated that retail prices in regional/ rural areas will  require its own management, location, systems, etc. range from A$35 to A$60/mth for a 12Mb connection. This

compares to Telstra's proposal of 512k connection for  It could result in a loss of efficiency and may result in A$59/mth of FTTN. Therefore, it will be difficult to envisage an  higher prices.

access pricing decision that will satisfy all parties.

• There is a risk that the rate of return may not be sufficient

Also in the subsequent release, the Minister stated that  to encourage further investment.

"OPEL is a structurally separated 'wholesale only'

company, that will operate and maintain the network, and that  We acknowledge that a break-up of an integrated telco like will sell services on a transparent and equivalent basis to  Telstra is a very complicated exercise, and the valuation parent entities, Elders and Optus, and to any other broadband  outcome will depend on the specifics of the model. There are provider in the market."  a number of issues that need to be considered – such as the

transfer pricing regime, regulated returns, asset allocations, Perhaps, it is possible that the government may consider  USO (Universal Service Obligation) obligations, and whether alternative measures like a structural separation in metro  the Networks and Services businesses can compete in each areas too.  other's markets.

A break-up of Telstra appears to be gaining industry

support. At a recent industry conference (Australian Information Industry Association) in Sydney recently, a number of smaller carriers (Primus, Macquarie) and the ACCC suggested that there is a need for industry reform, and ALP (if it wins the next election) could be best placed to do it.

What are the different Structural Separation Options?  We believe any of the (a), (b), or (c) is possible – but in our The Allen Consulting Group in its study in December 2006,  analysis, we have only considered (a). Option (d) appears "Structural Separation of Telstra – why is it needed, and what  less likely given the existing mobile competition.

can be done", proposed the following options (Exhibit 15):

One of the key issues in such a scenario is likely to be what

1. Vertically splitting the wholesale and retail fixed-line  assets are included in the Network separation. Based on the elements, namely the copper access network, from the  proposals submitted by TCNZ and BT, they offered to include retail fixed-line business;  all "fixed-line local access bottleneck assets", which will enable it to deliver equivalence between access seekers and
2. Selling the HFC network to create infrastructure-based  promote competition in downstream services. These were competition;  cable, local access and regional backhaul fibre, civil infrastructure and buildings, cabinets and transport
3. Divesting Foxtel ownership;  electronics, connection points such as MDFs, and du-port facilities. It did not include any assets that service providers
4. Separating Telstra's mobile business from the rest of its  can build themselves – such as DSLAMs and mobile assets. businesses.

Exhibit 15

Structural Separation Options

Source: The Allen Consulting Group, December 2006, "Structural Separation of Telstra – why is it needed, and what can be done"

Emerging International Trends  And last but not least, investor confidence has not been BT in the UK has established Openreach,while Eircom in  dampened. The creation of Openreach and its own Ireland and TCNZ in New Zealand are reviewing the  separate reports provides a clearer picture of the financial possibility of structural break-up, along with a number of other  performance of different parts of the business. The European carriers. The incumbent telcos are finding it difficult  increased transparency is likely to lead to BT having

to justify new investments like Fibre, as incremental returns on  greater analyst coverage and greater access to capital these are limited, and these investments are margin dilutive to  funding in the financial markets. Helping to create a

the overall business. Another reason is to reduce regulatory  climate of confidence for sustainable infrastructure intervention.  competition, investment and innovation, BT has shown a

relatively strong share performance compared with many There are limited case-studies of a full vertical separation.  of its European peers since it announced its undertaking Most of the separation that has occurred (AT&T, etc.) is more  to functionally separate."

along the infrastructure lines.

In addition, Ed Richards, CEO, OFCOM, made the following comments in the same newsletter:

Feedback from BT

BT's Head of Global Interconnection/Regulation, Grant Forsyth, made the following comments in an industry newsletter (La Lettre, March/April 2007 edition):

"The benefits of functional separation:

• BT, Ofcom, competitors and consumers all benefit from the undertakings. First of all, BT benefits from retaining the efficiencies of a vertically-integrated operator and removing the uncertainty of future harsh regulatory remedial actions, thereby allowing it the ability to invest and innovate with greater freedom.

• Without this assurance, BT would undoubtedly have been more reluctant to invest in its 21C next generation network.

• Ofcom benefits through having a clearer regulatory focus on the incumbent telecommunications operator, which is now subject to strict oversight of its compliance with nondiscrimination principles.

• Furthermore, competitors can have greater confidence in the industry through a level playing field which will result in increased investment and innovation leading to greater choice and lower prices to the benefit of all consumers.

How is the policy working in practice? So far, very well. Openreach went from a theory to a practical reality in six months. Its creation has prompted a new wave of investment in the UK telecoms market which in turn has triggered a major price war in the broadband market.

• Importantly, there have been big benefits for BT itself – we have been able to deregulate retail markets and BT's share price has risen partly because of confidence that there is a new stability in the relationship with the regulator.

• Ironically, some European incumbents who were initially very hostile to functional separation are now seriously examining it for this reason.

How relevant is this UK experiment to other regulators? We certainly don't believe that all regulators would need to follow the UK approach to achieve effective competition – this depends on national market circumstances. The degree of Functional Separation required in different national markets would also differ. But we do believe that all regulators should have the powers to impose functional separation under the EU Framework even if only as a power of last resort.

Hypothetical Break-up Value

A vertical split of an integrated carrier like Telstra would be  (a) Network Value

unprecedented and a hugely complex exercise. A number of  Our base-case valuation of the network business of A$23.4bn different structures are possible, with varying valuation  is based on the following assumptions:

outcomes.

Similar to the analysis we undertook for TCNZ in May this year (Unlocking Value through Structural Separation) – we estimate under a hypothetical break-up scenario that Telstra shares could be worth A$4.80 per share with a range of A$4.06 to A$5.53 per share.

Replacement value per line of A$2,000; Regulated pre-tax WACC of 10%; 20-year depreciation schedule; EV/EBITDA of 8.5x.

The key variables in this valuation are:

1. What is the Network business worth?
2. What is the Service business worth once the Network business is separated?
3. What is the value of the rest of the businesses?

This is summarized in Exhibit 17 below, and we discuss each of these in more detail below.

We again stress that the valuation outcome will depend on the structure chosen for separation (vertical or horizontal), which assets are included, and other commercial arrangements. These can substantially influence the magnitude of shareholder value created/destroyed under a separation model.

For the purpose of this hypothetical exercise, we have assumed a high-level break-up of the Network (local loop or copper network) and Service businesses. We have not assumed any divestments or a change in the capital structure.

The value of Network is perhaps the most complicated part of this valuation exercise. These are depreciated assets, where the book value is not reflective of the current asset value or the replacement value. For example, total PPE as at June 2007 was A$24.6bn. However, this includes Telstra's mobile assets, copper, cable, other infrastructure, etc.

Therefore, to value Telstra's fixed-line copper networks, we

1. estimate the replacement value per line; (b) apply a regulated rate of return to determine its annual EBITDA contribution; and (c) then apply a utility type multiple to work out the market value of these networks.

Exhibit 16

Telstra's PPE (June 2007) A$mn

Land & Site, 22 Plant, Equipment and  Buildings, 467 Motor Vehicles, 680

Communication Assets, 23,428

Source: Company data, Morgan Stanley Research

Exhibit 17

Telstra Hypothetical Break-Up Value

Pre Network Separation  Post Network Separation

Revenues  EBITDA Margin EBITDA Revenues  EBITDA Margin EBITDA

Source: Company data, Morgan Stanley Research. e = Morgan Stanley Research estimates

What is the Replacement Value?  Exhibit 19

We have used the "replacement value" concept to determine  FTTN versus FTTH Estimates (US$) the Regulated Asset Base (RAB) – using BT's multiples as a  

starting point. For BT, the RAB is 11bn, which is based on  current cost accounting. Based on the total WLR, ULL and  retail lines of ~25mn, the cost is approximately 450 per line,  or A$1,200 per line. However, given Australia's land-size, this  could potentially range from A$1,000 to A$2,500 per line.  

Based on these estimates, we value Telstra's existing copper  network at between A$9.2bn and A$22.9bn, with a base  case valuation of A$18.4bn using an average price of  A$2,000 per line. An important point to note is if the whole  existing network were to be replaced, the technology of choice  would now be fibre-to-the-home (FTTH), not copper, for which  Source: Morgan Stanley Research; Global Telecom Outlook Day 2006 the replacement costs could be as high as A$40-50bn. This is  

based on our estimates that the cost of FTTH rollout is around  Exhibit 20

A$4,000-5,000 per line.  Replace Value at Different Price s (A$bn)

25,000

It is important to note that the RAB is different to the market

value of these assets. In utility businesses, RAB is the capital  20,000

value of the assets used by regulators in setting prices or price

limits for utility companies. Utilities tend to trade at a

15,000

substantial premium to its RAB (20-70%) due to their stable

predictable returns, and higher gearing potential. To work out

10,000

the market value of Networks, we need to determine what the

regulated WACC is.

5,000

Regulated WACC?

The objective for the regulator is to set the regulated return'

-

1,000 1,250 1,500 1,750 2,000 2,250 2,500

which would imply that the enterprise value of the asset is

close to or equal to the RAB. If the regulated return is below

Source: Company data, Morgan Stanley Research

Telstra's average cost of capital, then it would not be able to   attract additional investment. On the other hand, if the return

Exhibit 21

Pre-tax WACC of Utilities

is too high, then prices will be too high, which will result in  

9.60% 9.40% 9.20% 9.00% 8.80% 8.60% 8.40% 8.20% 8.00%

lower usage.  

In Exhibit 21, we highlight regulated returns for various utilities  in Australia and New Zealand. The average pre-tax WACC is  9%. For telecommunications, it could be argued that the  return should be higher to account for various technology  risks.  

In BT's case, the regulator has agreed to a 10.1% pre tax  WACC on Openreach – which is 270bps higher than BT's  WACC of 7.4%.  

AGL CitiPower TXU United Energy Multinet Powercor Envestra

Using a similar rate of 10% pre-tax WACC for Telstra's  Source: Company data, Morgan Stanley Research Network business, and a 20-year depreciation schedule,

we estimate the regulated return or annual EBITDA would

be A$2.76bn.

Market Value of Networks?  Exhibit 22

Applying an 8.5x EV/EBITDA multiple, we estimate the  Market Value of Networks

market value of Networks to be A$23.4bn. This represents  Access Lines  9.2 million a ~30% premium to Regulated Asset Base (RAB). The  Replacement Cost  2,000 question is if 8.5x is the right multiple.  Replacement Value  18,366

Our UK team values Openreach at 6.8x EV/EBITDA. If we  Pre-tax WACC (%)  10 were to use the same multiple for Telstra's Network – our  EBIT  1,837 hypothetical base case valuation reduces to A$4.40.  Dep (5%)  804 EBITDA  2,640

There is no guarantee that the Network business can meet its regulated returns. For example, in the recently published regulatory accounts, Openreach achieved a pre-tax return of 9.7% in 2007, which has implications for the premium that can be applied to RAB. Our European team recently reduced their premium to RAB for Openreach to 0% (~20-30% previously) mainly due to competition concerns (migration to mobile datacards)

For Telstra, we make the following points:

• Regulated WACC for Telstra could be higher given the higher risk-free rate in Australia;

• Telstra's current WACC of 9.1% is higher than BT's 7.4%;

• Relative to BT, competition is arguably more benign, and Telstra is more integrated to manage competitive risks.

Given limited comparables in the telecommunications sector, we also use valuation multiples for Australian utilities. Based on IBES data, the average EV/EBITDA for Australian utilities is 12.3x, with a range of 7-16x. Telstra's copper business has utility type characteristics – irreplaceable utility-like local loop infrastructure, stable predictable returns, and potential for high gearing.

Using a range of replacement value of A$1,000 to A$2,500 per line, and applying an EV/EBITDA multiple of 7.5-10.5x, we derive a hypothetical break-up valuation range of A$4.37 to A$5.51 per share (this assumes our current base-case values for all other businesses).

At 7.5x  20,662 At 8.5x  23,417 At 9.5x  26,172

Source: Company data, Morgan Stanley Research

Exhibit 23

Australian Utility 2008e EV/EBITDA

16.00 14.00 12.00 10.00 8.00 6.00

s ta et v

lin APA AGL sN

 Wind A DUET Viridis  Au

BB Hasting Transfield  BB Infra Envestra BB Power SP  Energy De

Challenger Infra

Source: I/B/E/S estimates

Exhibit 24

BT's Regulatory Reporting Summary

Source: BT Plc.

Exhibit 25

Hypothetical Valuation Scenarios at Different Replacement Price s and EV/EBITDA Multiples

EV/EBITDA Multiple

7.5  8.0 8.5 9.0 9.5  10.0 10.5

At A$1,000 replacement value  $4.37  $4.42 $4.48 $4.53 $4.59  $4.64 $4.70 At A$1,500 replacement value  $4.47  $4.55 $4.64 $4.72 $4.80  $4.89 $4.97 At A$2,000 replacement value $4.58  $4.69 $4.80 $4.91 $5.02  $5.13 $5.24 At A$2,500 replacement value  $4.68  $4.82 $4.96 $5.10 $5.23  $5.37 $5.51

Source: Morgan Stanley Research

Exhibit 26

Hypothetical Valuation Scenarios of Telstra's Network