India Equity Research Reports, IPO and Stock News Visit http://indiaer.blogspot.com/ for complete details �� ��
Bharti Airtel (BHARTI)
Telecom
Not so fast. We take a closer look at the risks associated with ascribing a positive NPV
to Bharti’s Africa investment. Even as the combination of top-down Africa market
growth potential and confidence in Bharti’s execution is appealing, there are potential
pitfalls to a positive NPV case, including—(1) anchoring our forecasts to the transaction
price, (2) overestimating market growth and OPM expansion potential, and (3)
underestimating competition in Africa.
A lot of things need to go right for value creation from the acquisition
At 12X TTM EV/EBITDA on proportionate ownership, Bharti paid a substantial premium to the 5-
6X trading multiple of Africa and other emerging market telcos. Strategic premium for control is a
valid argument, but the same should drive meaningful improvement in operations to justify the
price paid and create value on top of the same. This is indeed possible, but there are challenges—
we examine them closely in this report.
Bharti’s African strategy hinges around successful adoption of its minutes factory model in the
acquired African operations. The model is about driving usage elasticity by cutting prices
(increasing affordability) and earning returns by being the lowest cost producer of minutes. Bharti
has done it successfully in India, as reflected in its industry-leading revenue/EBITDA market share
and return ratios. We are strong believers in Bharti’s execution capabilities. However, the challenge
is elsewhere; other key variables need to go right. We see three key variables, all of which need to
go right for the acquisition to turn value neutral or accretive –
Market growth – projected high GDP growth in the ‘continent of the century’ and low wireless
penetration justify potential; but one needs to watch out for affordability factor.
Market share gains – key swing markets (Nigeria, Kenya, Tanzania, Uganda and Ghana) have
formidable competitors in the likes of MTN, Vodacom, Millicom, Safaricom, etc.
EBITDA margin expansion – hinges on price elasticity being meaningfully revenue accretive, and
on Bharti’s ability to gain market share and establish/maintain cost leadership.
Even as pricing environment in the Indian wireless market remains benign for the time being, the
structural over-capacity issue with the sector remains a reality which could seek correction in many
ways, most of which are likely to be negative for the sector. In addition, as we discuss in this
report, Street’s optimism on the Africa business runs risk of disappointment, in our view. Valuation
at 7.5X FY2012E EV/EBITDA is not comforting, either. REDUCE.
A closer look at the three key variables – market growth, market share gains,
and EBITDA margin expansion
Market growth – watch out for the affordability factor
Purported ‘Continent of the century’, high projected GDP growth rate, low headline wireless
penetration (15-50% in most of Bharti’s operating countries) and even lower real
penetration (adjusted for multi-SIM usage) – what’s not to like about the market potential
Bharti has gained access to through the acquisition? The price paid, for one; but we are
repeating ourselves and intend to keep the acquisition price out of the operational
challenges discussion; nonetheless, before we move on, one final reference to the price paid
– the challenge is not about absolute market growth, absolute market share expansion for
Bharti or EBITDA margin expansion; it is about all of these being enough to justify the price
paid and create value in addition. Coming back to the market growth challenge, here are
two issues to watch out for –
Affordability – we plot ARPU as a percentage of GDP per capita (both nominal) on one
axis and wireless penetration on the other axis for several markets (see Exhibit 1 below).
Median ARPU as a percentage of GDP per capita for our analysis universe is about 2.1%
with a clear trend of declining ARPU as a percentage of GDP per capita as penetration
rises. Essentially, on an average, individuals spend roughly 3-4% of their income on
wireless telephony and the average % spend per user declines as wireless penetration
increases (see Exhibit 2) – pretty straight forward.
This is where it gets interesting – ARPU as GDP per capita for the 15 countries
where Bharti acquired Zain’s assets is in the 4% (Gabon, 114% penetration) to
54% (DRC, 14% penetration) range. This is clearly above the trend line and, more
importantly, way above the levels other emerging markets were at similar penetration
levels. Now, near-absence of wireline telephony can explain part of this reality as wireless
spend captures a bulk of a household’s telephony spend in these markets. We also note
that the multi-SIM impact argument is futile as it impacts values on both the axes (lower
real wireless penetration but higher ARPU as % of GDP per capita).
Implications – penetration potential is huge but could come at the cost of significant
erosion in ARPU. A couple of other things bear attention – (1) income distribution
(inequality) matters; increasing penetration to the next level would likely involve sharp
deterioration in marginal ARPU, and (2) prices can not be set in isolation; costs involved in
setting up the network and running the business (and wireless telephony is a high fixedcost,
high capex business) play a key role. The cost structure in most African markets is
high (high equipment costs, low power availability, shortage of skilled labor) and this
could serve as a floor to pricing and in turn determine the ceiling on penetration at
different points in time.
Elasticity is risky business – a blanket (for existing as well as new customers) cut in
pricing always results in reducing revenues from the existing base of minutes carried on
the network. Revenue accretion takes place if the pricing cut results in the minutes
carried increasing enough to compensate this loss. An increase in minutes can happen in
two ways – (1) increased usage from existing subs, and (2) subs base expansion through
market share gains and new customer acquisition (true elasticity is when new customer
acquisition is higher than would have happened even without the price cut). With
revenue accretion and little incremental operating costs (typical on networks with low
capacity utilization), absolute EBITDA accretion takes place – the essence of price elasticity.
However, and this is the risky aspect of elasticity – what happens if the expected elasticity
(growth in minutes on price cuts) does not pan out? Customers do not increase usage on
expected lines, new subscriber additions disappoint on numbers or quality, or
competition matches prices killing market share gains? Revenues stagnate or even decline.
The wireless revenue trajectory in the Indian market over the past eight quarters is a good
example. Hyper-competition drove massive price cuts, subs additions zoomed. However,
minutes growth from new subs and higher usage from existing subs clearly have not
compensated for the pricing decline. Result – a 100% growth in subs base, 6% growth
in revenues and likely an absolute EBITDA decline for the industry in the past eight
quarters (see Exhibit 3).
We do admit that the last two years have been a period of hyper competition in the
Indian market and hence not the best argument for lack of elasticity in the market.
Nonetheless, we have seen declining elasticity in the Indian market since the beginning of
FY2008, after 5 years of meaningful usage elasticity in the market as tariffs kept dropping.
Essentially, affordability plays a key role in determining the existence and extent of
elasticity in the market. Exhibits 4 and 5 trace the historical RPM, MOU, and ARPU trends
for Bharti in the Indian market.
Low MOU in Africa (Bharti reported an MOU of 112 in its African ops versus 454 in India
for the Sep 2010 quarter) is a valid argument in favor of elasticity. But, at the end of the
day, it is about affordability-led penetration and that equation is clearly not as straight
forward – it is tricky and challenging at best, in our view, as discussed in the previous
paragraphs.
Market share gains – its no red carpet out there
Bharti is ranked #1 in 5 of its 15 markets, #2 in 6, #3 or lower in the rest 4 and has a
different set of competitors in each market. A common theme across most markets – Zain
had lost subs market share consistently for a few quarters running prior to the asset sale to
Bharti. Exhibit 6: gives Bharti’s subs market share and share rank trends in these markets for
the past nine quarters while Exhibit 7: depicts the competition Bharti faces in the 15 African
markets. Blame a lack of focus and poor execution from Zain and you have a solid case for a
turnaround and improvement in market share now that the asset is in more capable hands.
We agree but highlight two risks–
Competition in key markets is formidable and these markets are as important to the key
competitor(s) as they are to Bharti.
Case in point #1 – Nigeria, the consensus ‘swing’/’make or break’/’key’ market for
Bharti Africa. Nigeria contributed 34% to Bharti Africa’s revenues, 29% to EBITDA, and
35% to subs, based on last reported country-wise numbers (4QCY09). It is also the 2nd
largest wireless market in Africa, has modest wireless penetration (55% headline), and
the market leader (MTN) has significant lead in terms of rev/subs market share and enjoys
substantially better EBITDA margin (59% versus Zain’s 30% in CY2009). Essentially, it
satisfies all parameters to be a key focus market for Bharti Africa turnaround.
Problem/ risk? MTN is unlikely to be a pushover – Nigeria is an extremely important
market for MTN (29% of its revenues, 42% of EBITDA for 1HCY10), MTN enjoys
advantages of network/capex leadership (7,000+ cell sites versus Bharti’s ~4,000) and
strong on-net community, and has a strong balance sheet. MTN reported a modest 0.1X
net debt/ EBITDA at the group level (MTN Nigeria has higher leverage, we believe) at end-
June 2010. Moreover, MTN’s scale matches Bharti and hence Bharti may not enjoy as
much scale-led equipment cost benefit over an MTN as it would over others.
Bharti also faces MTN as a competitor in four other markets – Congo B (Bharti a close #2
to MTN on subs market share), Ghana (Bharti #4, MTN the leader with a 53% share),
Uganda (Bharti a distant #2 to MTN), and Zambia (MTN a distant #2 to Bharti). That
makes it five markets where MTN is either #1 or #2 on subs market share and will serve
as a formidable competitor to Bharti. We also note that it has not been easy to dislodge
well-entrenched market leaders from their perch, especially if they have a strong balance
sheet; one does not need to go farther than Bharti in India to illustrate this. Hence, even
as the jury is still out on the (revenue) market share battle between Bharti and MTN,
Street’s rush to declare Bharti a winner is a tad premature, in our view. Exhibit 8 depicts
MTN’s condensed financial statements
Case in point #2 – Kenya; a low current market share, high potential market. Safaricom,
with nearly 80% subs market share, leads the 5-player market comfortably. More
importantly, Safaricom operates only in the Kenyan market and hence the importance of
the Kenyan market to Safaricom can not be overstated. In addition, Safaricom enjoys
strong competitive advantages in the form of network coverage, on-net community, and
the hugely successful M-PESA mobile money transfer service. We also note that Vodafone
has a 40% stake in Safaricom.
Bharti has attempted market share gains through a price war and seen some initial
success, per our channel checks; however two issues there – (1) competitors, including
Safaricom, have responded and hence, it would be too early to call it a one-way street for
Bharti, and (2) subs and minutes market share gains need to translate into improvement
in revenues and EBITDA and results on this front are yet to be seen. The second point
brings us to the next risk on the market share gains aspect.
Good versus bad market share gains. Growing subs and minutes market share through
massive price discounting is a risky strategy, in our view. It is probably necessary for a new
player trying to gain share from large incumbents, but it is critical to differentiate
between good market share and bad market share gain.
India again serves as a good example. This is what has happened in the Indian market in
the recent phase of hyper competition and price discounting – the new entrants have
gained reasonable subs market share, modest minutes market share, but unflattering
revenue market share; the incumbents have held on to (or lost very little of) their revenue
market share despite losing subs market share; market revenues have remained stagnant,
leading to stagnant revenues/ EBITDA for the leaders, and increasing revenues/operating
losses for the new players. Essentially, we emphasize that one needs to avoid the pitfall of
getting excited about subs market share gains without appreciating the revenue/EBITDA
impact of the same.
EBITDA margin expansion – will happen, but management target (40% by FY2013E)
looks challenging
EBITDA margin expansion for Bharti Africa is possibly the easiest value-creation driver to bet
on, at least directionally. There is a lot of low hanging fruit out there for a player like Bharti
in the form of network capex/opex efficiencies, IT/BPO outsourcing, and sales & marketing
expense leverage. The issue is not the direction but the extent and more importantly, timing.
This would be determined by two factors–
Price elasticity should be meaningfully revenue accretive – we have discussed this in
greater detail earlier in the note. The essence is this – more minutes on the network
increases operating costs (variable costs move in tandem while fixed costs go up in steps
as capacity/coverage expands) and hence, margin expansion comes from revenue
accretion. We are only suggesting here that the extent and sustenance of elasticity in the
African market is an unknown and one needs to build in a safety cushion on the extent of
elasticity led revenue growth and margin expansion.
Bharti needs to soon establish and maintain cost leadership in respective markets.
This is again a critical aspect of the minutes factory model – becoming the lowest cost
producer of minutes is a must. Bharti has demonstrated its ability to manage costs
effectively in the Indian market and needs to replicate the same in Africa. As mentioned
earlier, we are believers in Bharti’s execution capabilities.
A word of caution here – note the role of scale in cost efficiencies. Most cost reduction
initiatives can be replicated in the medium-to-long term – for example, Vodafone and
Idea have adopted Bharti’s innovative network/IT outsourcing model successfully in India
– in fact, Idea’s cost per minute is 75% of what Bharti’s cost per minute was at similar
traffic volumes. The only meaningful cost advantage Bharti now enjoys versus these
players in India is its larger scale of operations. Unfortunately, it is a mixed bag for Bharti
in Africa as far as scale advantage in individual markets is concerned. Exhibit 9 compares
the cost structure of Bharti’s Africa wireless business with that of MTN and Bharti South
Asia wireless (KIE estimates).
Bottom line – too early to build a positive NPV case for the acquisition
Lest we are misunderstood, we clarify that we appreciate the strategic rationale of the
acquisition. Bharti’s India business has turned FCF positive and it is only prudent that Bharti
looked to enter another growth market. Africa clearly fits the bill both as a growth market
and an ideal turf for Bharti to replicate its low-cost minutes factory model. We only caution
against hurrying to build a positive NPV case for the acquisition based on the bullish topdown
view on the African market and confidence in Bharti’s execution.
The price paid for the acquisition should serve as an important input to the whole equation
and we caution against anchoring forecasts to the acquisition price. In addition, a positive
NPV case builds in aggressive assumptions on each of three variables discussed in this note
(market growth, market share gains, and EBITDA margin expansion) and runs a risk of
disappointment, in our view. We refrain from justifying a positive NPV on a long-term DCF
— Bharti’s strategy in individual markets continues to evolve, competitor- and market
response to the same remain to be seen, and the first signs of margin expansion are still at
least a couple of quarters away. We base our valuation of Bharti Africa (negative Rs29 per
share of Bharti) on FY2012E EBITDA – we ascribe a 7.0X EV/EBITDA multiple to
proportionate EBITDA and adjust Africa net debt for minority interest. Exhibit 10 gives our
end-March 2012E SOTP valuation for Bharti. Exhibit 11 gives the comparable valuations of
wireless stocks across emerging markets.
Other risks/issues worth a mention
In addition to the potential risk of overestimating operational turnaround potential or
underestimating the required time to do so, there are a few other issues worth a mention –
Acquisition would expose Bharti to significant currency risks. Exposure to multiple
currencies will also increase the risk profile of Bharti’s earnings, in our view. Exhibit 12
depicts the currencies of various countries forming a part of the proposed acquisition.
High political uncertainty in the geography leads to high volatility in these currencies.
Zain’s reported US$ performance in these countries in CY2009 versus CY2008 bears
testimony to the substantial potential currency risk arising out of this acquisition – Zain
reported a yoy revenue decline of 12% in these countries in CY2009, with substantial
negative impact of currency depreciation; in constant currency, Zain would have reported
a revenue growth of 6% yoy, as per our computation.
Unhedged floating-rate acquisition debt keeps Bharti exposed to interest rate risk.
Bharti raised the acquisition debt (US$8.3 bn thus far, another US$700 mn remains
payable to Zain) at LIBOR (1-month, in our view) + 195 bps. Commenting on interest rate
cycle turns is outside our domain – however, we would urge investors to be cognizant of
the risk. A fixed-for-floating swap typically costs ~200-250 bps, per our channel checks.
Exhibits 13-14 depict historical LIBOR trends; LIBOR is near its all-time low.
Non-fungibility of Africa cash flows demands a holding company discount, in our
view. Each of Bharti’s 15 operating companies acquired through the Zain Africa
acquisition is a separate legal entity, paying taxes locally. Acquisition debt is held in a
Netherlands SPV. Access to cash flows from different opcos at the SPV or the parent level
would involve repatriation and attract distribution taxes, in our view. Moreover, cash
flows are not fungible across opcos. This would be especially true in opcos where Bharti
does not hold a 100% stake – this would cover 9 of the 15 total markets. Of course,
Bharti would look to raise its stake to 100% wherever possible (regulations in some
countries mandate a local partner), in our view. Hence, for now, one could argue for a
holdco discount on the PV of opco cash flows, although we do not ascribe any such
discount in our SOTP.
Lastly, given that Bharti does not have 100% ownership in all the opcos, the source of
EBITDA growth for Bharti Africa is also important. To illustrate the point, ownershipweighted
EBITDA for Bharti in Africa was 80.3% of the reported Zain Africa EBITDA in
CY2009. With Nigeria (65.7% ownership), the consensus swing market, likely to
contribute a majority of expected EBITDA growth, the proportion of ownership-weighted
EBITDA attribution is likely to come down in the coming years. Exhibit 12 depicts Bharti’s
ownership in various opcos.
Our forecasts for Bharti Africa
Exhibit 15 gives our key financial and operational forecasts for Bharti Africa. While our
forecasts build in Bharti missing its stated revenue/EBITDA target of US$5 bn/US$2 bn by
FY2013E, we do not see our estimates as conservative. We build in fairly reasonable 15%
revenue CAGR over FY2011-13E and an EBITDA margin expansion of 10.4% pts from
2QFY11 levels by FY2013E. Putting it slightly differently, we are building in incremental
revenues and EBITDA of US$1.2 bn and ~US$750 mn for FY2013E over annualized 2QFY11
comparables (incremental EBITDA margin of 63%).
No comments:
Post a Comment