Indonesian Telecoms Sector

Transcription

1 Asia Pacific Equity Research Telecommunication Services (Telecommunication Services ID (Asia)/Telecommunication Services TH (Asia)/Telecommunication Services CN Research Analysts Colin McCallum, CA Indonesian Telecoms Sector COMMENT Investing for data monetisation Figure 1: Monthly fee by operator (USD) higher ARPU for higher data use DISCLOSURE APPENDIX AT THE BACK OF THIS REPORT CONTAINS IMPORTANT DISCLOSURES, ANALYST CERTIFICATIONS, AND THE STATUS OF NON-US ANALYSTS. US Disclosure: Credit Suisse does and seeks to do business with companies covered in its research reports. As a result, investors should be aware that the Firm may have a conflict of interest that could affect the objectivity of this report. Investors should consider this report as only a single factor in making their investment decision. CREDIT SUISSE SECURITIES RESEARCH & ANALYTICS BEYOND INFORMATION MB 700MB 1000MB 2500MB 4000MB Telkomsel Indosat XL Cellular dynamics now improving. The cellular market is less attractive than the fixed broadband space in Indonesia, with higher penetration and a larger number of players. Furthermore, growth has been extremely lopsided, with only the market leader, Telkomsel, growing market share at the expense of XL, Indosat and smaller operators. However, after a consolidation phase, price points for voice have stabilised and data price points are now following. How far does mobile data pricing need to increase? Link Net and Telkomsel already generate attractive returns on invested capital (ROIC) of 23.3% and 35.5%, respectively. If Indosat and XL increased service revenue by circa 20 21%, they could increase their ROIC to double digits, covering their costs of capital. We show that this can be achieved before 2020 even on current data pricing in Indonesia, since: (1) the operators already enjoy a % ARPU uplift (albeit from a low base) as customers begin to consume 700MB of data, and (2) smartphone penetration is set to more than double by December 2017 (to reach 79.2%). Stock calls. We retain Link Net, with its strong balance sheet and focus on attractive fixed broadband as a top Indonesian telco pick. XL also looks interesting. Although we revise down our FY15 earnings by a further 78.9%, largely on forex, our DCF-based target price declines just 5.5% to Rp5,150, leaving 99.6% potential upside. We cut Indosat's FY15 earnings by a further 9.8% on forex, but maintain our OUTPERFORM rating given 24.3% upside to our target price. While there is only single digit upside to our (unchanged) Rp3,100 target price for PT Telkom, its strong operating performance and lack of US$-denominated debt make a viable 'large cap' investment choice. Client-Driven Solutions, Insights, and Access 9 5

2 Focus charts and tables Figure 2: Price per MB in USD Indonesian operators Figure 3: Indonesia versus Thailand and China (USD) Low Low-Medium Medium Medium-high High volume Telkomsel Indosat XL Low Low-Medium Medium Medium-high High volume China (4G w/o subsidy) Indonesia Thailand Note: Price per MB. Figure 4: BTS by operator as at 1Q15 suggests data growth will be shared in a three-player market '000s 2G BTS 3G/4G BTS Total BTS Proportion Telkomsel 46,996 43,556 90, % Indosat 22,212 18,544 40, % XL 36,331 16,431 52, % Hutch Indonesia* 14,000 14,000 28, % SmartFren 2,000 4,000 6, % Total BTS 121,539 96, , % Source: Company data for Telkomsel, Indosat, XL and SmartFren, *Credit Suisse estimates for Hutch Indonesia Figure 5: The foreign currency debt profile has been a key driver of Indonesian telco share price performance Foreign Net debt in currency foreign Gross foreign currency debt in 1Q15 (in foreign currency thousand) currency in 1Q15 (in Rp bn) Sensitivity on FY15 post-tax profit (IDR +/- 10% against forex) % net profit deduction if IDR depreciate s by 10% Per share value erosion (gain) due to the Rupiah depreciation YTD (in Rp / share) Value impact as % of our DCF valuation Share price change YTD(%) Consensus EPS change YTD(%) Link Net USD 12, /-Rp15.8bn 2.0% 5 0.1% 15.3% -8.5% PT Telkom USD 200,168* 2,617* -/+Rp111.5bn -0.6% % 2.8% -1.1% JPY 7,681, /-Rp80.5bn 0.4% Indosat USD 890,200 11,638 +/-Rp1,103bn 365.2% % 8.5% -56.2% XL Axiata USD 1,561,000 20,409 +/-Rp1,934bn 289.9% % -41.2% -82.2% * PT Telkom has Rp1,177 bn net financial assets in US dollars. Figure 6: However, DCF and multiples suggest this is priced in for XL and Indosat Close Target Upside P/E (x) EV/EBITDA (x) FCF yield (%) Div yield (%) Rp price price (%) 15E 16E 15E 16E 15E 16E 15E 16E Link Net 5,575 8, % % 4.2% 0.0% 1.4% XL 2, % % 9.3% 0.4% 4.6% Indosat 4, % % 6.4% 0.3% 2.1% PT Telkom 2, % % 5.7% 3.3% 3.6% TBIG 7,500 5, % % 1.1% 1.5% 3.2% TOWR 3,990 3, % % 2.8% 1.4% 1.7% NJA integrated % 5.4% 4.1% 4.3% NJA Mobile % 7.1% 3.1% 3.2% Indonesian Telecoms Sector 2

3 Investing for data monetisation Fixed healthier than cellular, but cellular is improving With only 6.1% penetration of households as at December 2014, with only two players enjoying any scale, and with strong barriers to entry, the fixed broadband market in Indonesia offers both high growth and high returns on capital. The cellular market is less attractive, with higher penetration and a larger number of players. Furthermore, growth has been extremely lopsided, with only market leader Telkomsel growing market share at the expense of XL, Indosat and smaller operators. However, after a consolidation phase, price points for voice have stabilised and data price points are now following. With only 6,000 base stations (BTS), we do not believe that a consolidated SmartFren operation has the scale to disrupt the improving competitive dynamics in the cellular sector. How far does mobile data pricing need to increase? Link Net and Telkomsel already generate attractive returns on invested capital (ROIC) of 23.3% and 35.5%, respectively. If Indosat and XL increased service revenue by circa 20 21%, they could increase their ROIC to double digits, covering their costs of capital. We show that this can be achieved before 2020 even on current data pricing in Indonesia, since (1) the operators already enjoy a % ARPU uplift (albeit from a low base) as customers begin to consume 700MB of data, and (2) smartphone penetration is set to more than double by December 2017 (reaching 79.2% versus 31.3% as at December 2014). Upcoming interconnect changes might benefit Indosat and XL's EBITDA to the tune of up to %. Given that the specific problems suffered by both XL (lack of spectrum, followed by integration of Axis) and Indosat (delays in approval for 900MHz 3G, followed by delays in rollout) have now been largely resolved, we are prepared to forecast that XL and Indosat's cellular revenue growth will more closely match industry growth rates from 2016 onwards at 6.9% YoY versus Telkomsel's 7.4%. Balance sheet issues big, but not insurmountable While operational progress has been one factor behind share price performance year-todate in 2015, balance sheet position, and in particular exposure to USD-denominated debt, has been a significant driver, in an environment where the Indonesian rupiah has devalued by 8.9%. PT Telkom and Link Net, both of which have net cash positions and minimal foreign currency exposure, have enjoyed share price appreciation. More surprisingly, Indosat's share price has also appreciated YTD (though it declined in 2014). However, we note that even without hedging the impact on (DCF) value from the Rp1.0 tn increase in Indosat's debt burden caused by the 8.9% adverse currency movement YTD represents only Rp186/share (just 3.3% of our DCF-based target price, and just 13.4% of Indosat's projected capex expenditure for a single year). While XL is more exposed, even on an unhedged basis the impact on (DCF) value from the Rp1.7 tn increase in XL's debt burden caused by the rupiah's 8.9% fall YTD represents only Rp208/share. This represents just 4.0% of our DCF-based target price a relatively small impact when compared with XL's 41.2% YTD share price collapse! Top picks Link Net and XL We retain Link Net, with its strong balance sheet and focus on the attractive fixed broadband as a top Indonesian telco pick. XL also looks interesting. Although we revise down FY15 earnings by a further 78.9%, largely on forex, our DCF-based target price declines just 5.5% to Rp5,150, leaving 99.6% potential upside. We cut Indosat's FY15 earnings by a further 9.8% on forex, but maintain our OUTPERFORM rating given 26.6% upside to our (unchanged) target price. While there is only single digit upside to our (unchanged) Rp3,100 target price for PT Telkom, its strong operating performance and lack of USD-denominated debt make a viable 'large cap' investment choice. While the fixed broadband market is more attractive, competitive dynamics in cellular are improving ARPU uplift from smartphones suggests that XL and Indosat can enjoy higher ROIC USD-denominated debt exposure is very painful for earnings, less so for DCF valuation Top picks: Link Net and XL Indonesian Telecoms Sector 3

4 Sector valuation Figure 7: Comparative multiples Close Target Mkt cap Normalised P/E (x) EV/EBITDA (x) FCF yield (%) Div yield (%) Ticker Ccy price Rating price (US$ bn) 15E 16E 15E 16E 15E 16E 15E 16E Integrated operators China Telecom 728 HK HK$ 4.33 O % 2.4% 2.9% 3.3% China Unicom 762 HK HK$ O % 3.6% 2.1% 2.5% Chunghwa 2412 TT NT$ 98 U % 5.6% 4.6% 5.0% HTHK 215 HK HK$ 3.20 O % 5.8% 3.7% 4.5% KDDI 9433 JP 3,336 N 3, % 17.3% 5.4% 5.7% KT KS W 29,900 O 37, % 18.1% 2.7% 3.0% NTT 9432 JP 4,980 O 5, % 15.2% 4.0% 4.2% Softbank 9984 JP 7,608 O 8, % 22.4% 0.5% 0.5% PCCW 8 HK HK$ 4.59 N % 8.4% 5.7% 6.6% HKT Trust 6823 HK HK$ 9.70 N % 6.9% 5.9% 6.9% PLDT TEL PM P 2,702 U 2, % 4.8% 6.1% 6.3% SingTel ST SP S$ 4.06 N % 7.2% 4.4% 4.7% SPK SPK NZ NZ$ 2.91 N % 8.1% 6.2% 6.5% CNU CNU NZ NZ$ 2.77 N % -16.2% 0.0% 5.1% TM T MK RM 6.41 N % 7.4% 3.6% 4.0% Telstra TLS AU A$ 6.33 U % 6.2% 4.8% 5.0% Jasmine JAS TB Bt 5.10 O % -2.4% 36.1% 3.1% True Corp TRUE TB Bt U % -0.3% 0.0% 0.2% Link Net LINK IJ Rp 5,575 O 8, % 4.2% 0.0% 1.4% HKBN 1310 HK HK$ 8.49 N % 5.5% 4.3% 5.5% Asia avg integrated % 10.2% 3.5% 3.7% NJA integrated % 5.4% 4.1% 4.3% Mobile operators AIS ADVANC TB Bt 242 N % 6.2% 5.7% 6.3% AXIATA AXIATA MK RM 5.96 O % 4.6% 4.0% 4.4% Bakrie BTEL IJ Rp 50 U n.m. n.m % -30.1% 0.0% 0.0% China Mobile 941 HK HK$ O % 6.2% 2.8% 2.8% DiGi DiGi MK RM 5.07 O % 4.7% 5.1% 5.2% XL EXCL IJ Rp 2,580 O 5, % 9.3% 0.4% 4.6% FarEasTone 4904 TT NT$ U % 6.5% 5.1% 5.1% Globe GLO PM P 2,652 U 1, % 3.0% 3.3% 3.7% IDEA IDEA IN INR U % 0.6% 0.4% 0.4% Indosat ISAT IJ Rp 4,345 O 5, % 6.4% 0.3% 2.1% LG Uplus KS W 12,000 O 12, % 9.6% 2.3% 2.9% Maxis MAXIS MK RM 6.38 O % 4.1% 4.1% 4.1% M1 M1 SP S$ 3.05 N % 7.3% 6.6% 7.2% NTT DoCoMo 9437 JP 2,800 N 2, % 4.0% 2.5% 2.9% PT Telkom TLKM IJ Rp 2,965 O 3, % 5.7% 3.3% 3.6% Reliance RCOM IN INR U % 19.8% 0.4% 0.3% SKT KS W 246,000 O 325, % 40.9% 4.1% 4.3% SmarTone 315 HK HK$ O % 5.5% 3.5% 4.1% StarHub STH SP S$ 3.83 U % 6.2% 5.7% 5.7% TAC DTAC TB Bt 65 O % 8.7% 6.3% 7.7% Taiwan Mobile 3045 TT NT$ 103 U % 5.6% 5.4% 5.7% Asia average % 6.6% 3.0% 3.2% mobile NJA mobile % 7.1% 3.1% 3.2% Asia average % 8.1% 3.2% 3.4% telecoms NJA telecoms % 6.5% 3.4% 3.6% Indonesian Telecoms Sector 4

5 Fixed healthier than cellular, but cellular is improving Fixed line broadband enjoying preferable dynamics As set out in our 7 January 2015 report, Fixed line: Back to the future!, a confluence of factors including content development and e-commerce, fixed line/pay TV convergence, fixed line/cellular convergence and declining equipment price points is resulting in a fixed line renaissance in emerging Asian markets such as Indonesia. The scope for fixed broadband growth is particularly strong in Indonesia because penetration is currently still so low. Very little fixed line infrastructure was built in Indonesia following the financial crisis leaving penetration as at December 2014 at only 6.1% of households. Indeed, if we define broadband access as services capable of providing speeds over 3Mbps, penetration of households was only 0.8% of households as at December 2014; amongst PT Telkom's subscriber base of 3.6 mn, only 116,0000 enjoyed speeds of 3Mbps and above last December, though just in the last six months this has increased to 427,000. Fixed broadband penetration is rising rapidly from a very low base Figure 8: 2014 broadband penetration (of households) 140.0% 132.7% 120.0% 107.5% 98.5% 100.0% 80.0% 75.0% 60.0% 44.8% 40.0% 30.2% 32.4% 20.0% 6.1% 6.1% 9.1% 0.0% Figure 9: 2014 YoY increase in broadband penetration 5.0% 4.2% 4.0% 3.3% 3.0% 2.1% 2.0% 1.5% 1.1% 1.0% 0.8% 0.5% -0.4% 0.0% 0.2% 0.0% -1.0% Source: Credit Suisse estimates Source: Credit Suisse estimates Importantly, the growth in penetration is currently being shared between only two large players, namely the aforementioned PT Telkom, which has a total of mn fixed lines in service gradually being upgraded to fibre, and Link Net, which had mn homes passed with Hybrid Fibre Coax (HFC) cable as at 30 June These first movers on broadband are set to enjoy a meaningful barrier to entry versus potential new entrants, in our view, for two reasons. (1) First, while licences for fixed broadband rollout are not restricted or expensive in Indonesia, in contrast to cellular licences which rely on scarce radio spectrum resources, rollout requires technical expertise, as well as permits from local authorities. Access to permits area by area can take up to seven months, involves more than one layer of bureaucracy, and acts as a severe constraint on network rollout schedules. (2) Second, rollout requires access to capital. As set out below, the decline in equipment price points and growth in income levels are such that rolling out fixed broadband services generates an attractive return on capital (circa 24.3% in Link Net's case). However, the crucial difference faced by any potential 'me-too' operator following Link Net into an area would be the likely number of customers that those operators could hope to achieve per node. If the penetration rate of homes passed drops to 15%, the return on capital drops to only 5.1% given the very heavy fixed costs. A two-player market can generate high returns on capital Indonesian Telecoms Sector 5

6 Figure 10: ROIC analysis HFC rollout in Indonesia 30% home passed penetration 30% penetration residential only 15% penetration residential only Rp '000s US$ Rp '000s US$ Rp '000s US$ Monthly ARPU Annual revenue 5, , , EBITDA margin 55.8% 55.8% 55.8% 55.8% 50.0% 50.0% EBITDA 3, , , Depreciation (years) Depreciation (823) (70) (823) (70) (1,516) (130) EBIT 2, , Tax rate (%) 25.0% 25.0% 25.0% 25.0% 25.0% 25.0% Tax (600) (51) (474) (40) (230) (20) NOPLAT 1, , Capex 7, , ,640 1,166 ROIC 24.3% 19.2% 5.1% Payback revenue (years) Payback EBITDA (years) * ARPU defined by Credit Suisse as FY14E total service revenue divided by projected average broadband subscriber numbers Source: Credit Suisse estimates We conclude that there is certainly room for one next generation network in high income residential areas in Indonesia, but not two or more. Shareholders in new entrants such as MNC Kabel (Not listed) and MyRepublic (Not listed) would therefore likely be better served by a strategy of rolling out to the 93.9% of households that are not yet connected to high-speed broadband, rather than following Link Net or PT Telkom's Indihome service down the same streets. In any case, neither MNC Kabel nor MyRepublic appears to be progressing quickly. We understand that as at June 2015 MNC Kabel (Not listed) has around 65,000 homes passed, and only 10,000 customers. MNC Kabel's management was quoted in the local press (Bisnes Indonesia) in October 2014 as targeting '800,000 homes passed by June 2015'. In March of this year, this target shifted down and back to '600,000 by December 2015', and as of now the target seems to have drifted down towards '500,000 by December 2015'. The aforementioned difficulties and delays in obtaining permit approvals are one of the key reasons behind the downward revisions by competitors, and we do not expect these structural issues common to all fixed line rollouts in Indonesia to be resolved any time soon. MyRepublic has yet to give any rollout targets for 2015 or Even PT Telkom's fiberisation continues to be relatively slow, with an estimated 500,000 fibre homes passed at present; some distance from stated targets of 3.0 mn IndiHomes by December 2015 (and some distance from Link Net's mn homes passed). However, in terms of total subscriber numbers (including around 2.27 mn subscribers with 'broadband' speeds of only kbps), PT Telkom remains the largest player in Indonesia, thanks to its nationwide footprint and legacy copper network. Link Net is already the second-largest operator, with 392,000 broadband subscribers as at December 2014, representing 10.2% market share by subscriber. We expect that this market share level will be broadly maintained in spite of an acceleration from PT Telkom and the gradual ramp-up of new players. However, given Link Net's 'cherry picking' focus on high income areas (to ensure high penetration of homes passed and therefore high ROIC), we expect its ARPU differential (Rp274,000/month for broadband services alone, versus Rp128,000 at PT Telkom) to be maintained. Thus, PT Telkom's broadband revenue market share is expected to decline by circa 0.5% p.a, while Link Net's broadband revenue share is expect to rise from 19.0% by circa 0.4% p.a, despite the entry of MNC Kabel and MyRepublic. Room for net entrants only in other geographical areas Indonesian Telecoms Sector 6

7 Figure 11: Broadband subscriber forecasts 2011 to 2018 '000s CAGR PT Telkom 1,789 2,341 3,013 3,400 4,650 5,900 6,900 7, % LinkNet % Others % Total 1,982 2,631 3,346 3,842 5,308 6,773 7,970 9, % Households 60,204 60,994 61,783 62,524 63,273 65,399 66,182 66, % Penetration 3.3% 4.3% 5.4% 6.1% 8.4% 10.4% 12.0% 13.7% 25.1% Market share PT Telkom 90.3% 89.0% 90.1% 88.5% 87.6% 87.1% 86.6% 86.1% -1.9% LinkNet 9.7% 11.0% 9.9% 10.2% 9.9% 9.8% 9.8% 9.9% -0.4% Others 0.0% 0.0% 0.0% 1.3% 2.4% 3.1% 3.6% 4.0% 2.3% Figure 12: Broadband revenue 2011 to 2018 Rp bn CAGR PT Telkom 4,000 4,200 4,433 4,918 6,667 8,563 9,870 10, % LinkNet ,197 1,572 2,047 2,472 2, % Others % Market 4,354 4,949 5,386 6,295 8,475 10,918 12,713 14, % Market share PT Telkom 91.9% 84.9% 82.3% 78.1% 78.7% 78.4% 77.6% 76.7% -0.5% LinkNet 8.1% 15.1% 17.7% 19.0% 18.6% 18.8% 19.4% 20.3% 0.4% Others 0.0% 0.0% 0.0% 2.9% 2.8% 2.8% 2.9% 3.0% 0.1% Our forecasts, therefore, project that penetration will rise from 6.1% of households in 2014 to 13.7% by 2018, with the number of broadband customers more than doubling over the period. Given the rollout constraints, we believe that the competitive environment can remain relatively benign, resulting in an expectation of stable average revenue per user as speeds increase. Indeed, Link Net recently succeeded in driving up overall ARPU levels for its bundled broadband and Pay TV products throughout March 2015 (on 7th, 15th, 23rd and 31st, depending on each customer's billing cycle). For example, its D'Lite package price was increased by 5.9% from Rp339,000 to Rp359,000, and this was tolerated by customers given additional four cable TV channels. Link Net's Elite package price was increased by 12.8%, with customers' speeds being raised from 10Mbps to 12Mbps, and an additional six cable TV channels added. With low penetration, high growth, and a benign competitive environment, we forecast that fixed broadband market revenues will grow at a 26.4% three-year compound annual growth rate (CAGR) across The cellular market has been growing, but lopsided Compared with fixed broadband, cellular penetration and smartphone penetration are already relatively high at 133.9% and 31.3% respectively as at December This is largely due to PT Telkom's decision to focus on allocating resources to its cellular division, Telkomsel, following the financial crisis, together with a flood of capital into the sector across the period. The shift from a de facto two-player market to a tenplayer market led to an aggressive price war in as new and re-capitalised players attempted to build scale, and 2G penetration rose sharply, catching up with regional peers. Since 2008, however, the Indonesian cellular sector has become somewhat more stable, entering into a consolidation phase. While there are still seven players in the market, as Figure 13 below shows, the 'Big 3' operators now comprise 93.5% revenue market share, Indonesian Telecoms Sector 7

8 up from 90.3% as recently as 4Q12. This has occurred due to (1) what we would call de facto consolidation, such as the loss of subscriber traction and termination of investment in the Bakrie Telecom and Telkomflexi fixed wireless CDMA offerings and (2) actual consolidation through the acquisition of Axis (Not listed) by XL and the merger of Mobile 8 with Smart. Figure 13: Revenue market share 4Q12 1Q13 2Q13 3Q13 4Q13 1Q14 2Q14 3Q14 4Q14 1Q15 Telkomsel 54.3% 53.7% 54.2% 55.3% 56.0% 56.2% 56.8% 57.9% 58.5% 58.8% Indosat 17.9% 18.3% 18.0% 17.5% 17.1% 17.1% 16.7% 16.7% 17.0% 16.8% XL 18.1% 17.8% 18.1% 18.1% 17.6% 18.7% 19.9% 18.9% 18.1% 17.9% Others 9.7% 10.3% 9.7% 9.2% 9.3% 8.1% 6.6% 6.5% 6.4% 6.4% Total 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% The resulting improvement in competitive intensity has facilitated an improvement in revenue generation across the cellular market, with growth improving from a low of 2.5% YoY in 3Q13 (albeit affected by a roll-off in growth after the introduction of SMS interconnection in June 2012) to 7.1% in 1Q15. Figure 14: YoY revenue growth trajectory (%) 4Q12 1Q13 2Q13 3Q13 4Q13 1Q14 2Q14 3Q14 4Q14 1Q15 Telkomsel 15.5% 13.3% 10.8% 7.6% 9.3% 9.8% 10.3% 10.0% 11.3% 12.1% Indosat 16.2% 16.5% 8.4% -4.2% 1.2% -2.1% -2.2% 0.3% 6.0% 5.4% XL 10.9% 3.0% -0.3% -1.5% 3.1% 10.3% 16.2% 10.2% 9.7% 2.8% Others 3.3% 7.0% -0.1% -4.2% 1.9% -17.7% -28.4% -25.8% -27.0% -14.5% Total 13.5% 11.2% 7.1% 2.5% 6.0% 4.9% 5.4% 5.0% 6.5% 7.1% However, we observe that the improvement in revenue growth has been tilted towards industry leader Telkomsel, which has clearly out-executed XL and Indosat. Telkomsel has enjoyed two engines of growth. Taking Telkomsel's stronger-than-expected growth of 12.1% YoY in 1Q15 as an illustration, the fact that voice revenue grew by 7.9% YoY and even SMS revenue grew by 5.1% YoY, was a critical driver to double-digit revenue growth being achieved overall. Telkomsel has successfully taken advantage of stabilising competition (and the recent, aforementioned, XL-Axis consolidation) to gradually raise voice and SMS price points where possible. This has been pursued on a regional basis; Telkomsel is now run through 206 pricing 'clusters' in order to maximise opportunities from regional differentials in competitive intensity. However, data is already proving to be an even stronger growth driver. Telkomsel's 3G smartphone subscriber base increased 54.2% YoY to 48.1 mn in 2Q15 representing 33.4% of the total subscriber base, and we will show later in this report that as customers upgrade to 3G smartphones, their ARPU levels are rising by at least 40%. Thus, Telkomsel s average subscriber base increased by 5.7% YoY into 2Q15, and blended average ARPU was still able to rise by 7.6% YoY into 2Q15 to drive 13.8% YoY revenue growth; a further acceleration versus the already very strong 1Q15 revenue result. Why have XL and Indosat not kept up? However, it is only the market leader, Telkomsel, which has been achieving double-digit organic growth; XL's YoY revenue growth rate in 1Q14 4Q14 was flattered by the (very expensive) acquisition of Axis, which was completed in March We observe that, taken in aggregate, Indosat and XL together have endured organic revenue growth rates in low-single digits. Indonesian Telecoms Sector 8

9 In the case of Indosat, revenue declines through late 2013 and 1H14 were in large part due to the lack of a working data network. This, in turn, led to an overreliance on voice and particularly SMS services during 2012 and early 2013, which subsequently became a drag. The lack of a competitive data offering in turn was due to Indosat's deliberate strategy of waiting for approvals to build 3G on 900MHz, so that its basic coverage rollout could be conducted in a less capital-intensive and more cost-efficient manner. Indosat finally received approval in September However, the implementation of the 900MHz rollout, by upgrading its network to multiband antennae and the refarming of 900MHz spectrum from 2G for allocation to 3G services, took much longer than expected. Full completion of Phase 1 coverage construction and refarming in all core cities inside and outside, actually took two years, and was only delivered in 3Q14, a critical nine months later than what had been planned initially. While the presence of a strong commercial director, Erik Meijer, helped Indosat continue to grow in 1H13 despite late delivery on the network, his departure, together with the inability to address consumers' data demands, resulted in a loss of revenue momentum which became particularly evident in 2H13 and 1H14. It is in our view no coincidence that XL's revenue growth trajectory seems to have mirrored that of Indosat. Specifically, XL, which had invested heavily in 3G in grew faster than the market at Indosat's expense during the 2H13 and 1H14 period, when Indosat did not have a strong data product in place and commensurately 'missed out' on industry growth. XL's growth was further boosted in-organically with the acquisition of Axis for US$865 mn, which was completed on 19 March Axis had circa 3.0% market share by revenue versus XL's 18.7% market share at the time. However, we suspect that Axis' subscriber base was of relatively low quality, and some of the subscribers churned off within six months of the acquisition. This was possibly aggravated by the XL-Axis network integration. While, in contrast to Indosat, XL's integration of the Axis network was executed in line with management's detailed plans (and within nine months), there was inevitably some disruption as the Axis spectrum resources were refarmed and XL's base transceiver systems (BTS) recalibrated in 2H14. There was also, unavoidably, a shift in XL's management focus from marketing to network integration. That the heavy integration work was done in 2H14, at just the point when Indosat's 900MHz 3G network was finally in place resulted in a clear deterioration in XL's revenue growth trajectory and a clear improvement in Indosat's. This has continued into 1H15. Overall, we observe that with the largest player, Telkomsel, executing well and growing market share, it has been mathematically impossible for both Indosat and XL to also grow market share. The two smaller players have therefore alternated with growth spurts and consolidation phases, as their spectrum and network constraints were exposed and then addressed. However, the cellular environment is improving As mentioned, Telkomsel has taken advantage of stabilising competition to gradually raise voice and SMS price points where possible. With total voice volumes still rising (due to increasing numbers of subscribers), this has driven mid-single digit voice revenue growth. Importantly, as they emerge from their network upgrade programmes, both XL and Indosat are now following Telkomsel upwards on voice pricing, as shown in Figure 15 below. While Telkomsel is still able to charge a premium given its superior network coverage and service quality, the premium is no longer expanding, and is in fact narrowing slightly as XL and Indosat catch up with Telkomsel's strategy following industry consolidation. With less pressure from smaller players such as TelkomFlexi, Bakrie Telecom and Axis, Indosat and XL are looking to improve voice monetisation. Even Hutch Indonesia (Not listed) has been less aggressive on pricing over the last six months, as the patience of the Hong Kong head office for negative EBITDA and negative cash flows finally appears to be wearing thin ten years after licence acquisition. XL has been particularly aggressive in raising price points, or, more specifically, in reducing discounts offered to new subscribers within starter packs. As set out in our 17 Indosat's growth disappointed in 2013 and 2014 and XL disappointed in 2H14 and 1H15 Voice price points have stabilised and are rising in some geographies Indonesian Telecoms Sector 9

10 March 2015 report, Short term pain for long term gain?, XL is no longer wishing to chase after the small minority of customers (possibly 10% of subscribers) who are so price sensitive that they are prepared to change phone numbers each month, and churn from one operator to another subscribing to discounted starter packs. Regrettably, over the past decade, operators have tacitly encouraged this behaviour, by keeping starter pack pricing (including discounts) lower than recharge pricing. This has created additional costs for operators in terms of SIM cards, promotional material, idle capacity (held for customers who have churned) and foregone revenue (since starter pack pricing is cheaper). XL is now keen to end this dynamic. XL has therefore reduced the supply of starter packs, which has raised the effective retail price point of starter packs in some areas, and the discounts offered on starter packs have largely been removed. Figure 15: Voice revenue per minute (Rp) Q12 1Q13 2Q13 3Q13 4Q13 1Q14 2Q14 3Q14 4Q14 1Q15 Telkomsel Indosat XL Importantly, Telkomsel has now turned its attention to driving up data pricing and, again, there is some evidence that XL is already following. Having successfully raised voice tariffs over the last 18 months, Telkomsel is now attempting to better monetise the acceleration in smartphone penetration (now rising more than 1.0 pp per month, more than double of 2013's run rate). Telkomsel's data bundle price points were therefore raised 15%, effective 24 November Telkomsel has now turned its attention to driving up data pricing The chart below shows the effective average data price per MB achieved by the 'Big 3' operators. While the chart shows that Telkomsel's average data price per MB still actually declined QoQ into 1Q15 and 2Q15, from Rp71/MB in 4Q14 to Rp37/MB in 2Q15, this is because as customers shift from expensive 'pay as you go' pricing to the tiered data packages on offer, the effective data price declines. Similarly, as customers commit to higher volume plans, there are implied volume discounts (a lower price per MB). With data volume growing rapidly, this driver offset Telkomsel's 15% price rise across its packages. However, had Telkomsel not raised price points, the yield per MB would have fallen even faster. Overall, Telkomsel's objective in data pricing is to slow the rate of decline, so that the 31.0% YoY growth rate in data revenue delivered in 2Q15 will move closer to the 131.1% YoY increase in data volume delivered over the period. Indonesian Telecoms Sector 10

11 Figure 16: Data price per MB (Rp) Q13 1Q14 2Q14 3Q14 4Q14 1Q15 2Q15 Telkomsel Indosat XL XL seems to be following Telkomsel's lead on data pricing again, with the effective price per MB declining only 2.3% QoQ from Rp40/MB in 4Q14 to Rp39/MB in 1Q15. XL's belief is that, following the completion of the Axis integration, it now has comparable network, spectrum resources, and, therefore, data service quality, to that delivered by Telkomsel, at least within XL's core coverage areas. XL's stated intention within the next 12 months is therefore to use the stabilising competitive environment to narrow the gap in data pricing versus Telkomsel. The overall goal of XL's management is in fact identical to that of Telkomsel's management team to convert more of the rapid growth in data volume (+87.9% YoY for XL into 1Q15) into growth in revenue. Will Indosat derail this improvement? It is clear from the shape of Indosat's 1Q15 results namely a large (3.3 mn) increase in the number of subscribers, and a 33.4% QoQ decline in the implied data price from Rp57/MB to Rp38/MB that Indosat is not yet following Telkomsel and XL on a strategy of improved data monetisation. While, as we saw, Indosat's revenue trajectory began to improve in 2H14 on the launch of the 900MHz 3G service, we suspect that management is keen to opportunistically make back more of the ground lost during 2013 and 1H14. Tactically, this has been assisted in 1Q15 by scooping up some of XL's low-end subscribers, who have turned away from XL following the removal of aggressive starter pack discounts. The fact that Indosat's Chief Marketing Officer, Pak Joy, had previously been XL's chief marketing officer, has meant that Indosat has been aware of XL's strategy, and is well-positioned to design offers and marketing plans that could take advantage of it. Indosat's management has suggested that shifting price points upwards with Telkomsel and XL is desirable, and our current view is that Indosat will follow in 2H15. In fact some of Indosat's more aggressive data offers have already been withdrawn. Thus, we believe that Indosat has simply been keen to 'steal a few bases' before following upwards on price, rather than to re-instigate a tariff war. The risk to this thesis is really two-fold. First, there is a chance that Indosat's management starts to believe that the recent subscriber growth can lead directly to improved profitability, and therefore maintains the current, more aggressive pricing tactics rather than following Indosat's data price declines have yet to slow.but we still expect Indosat to follow Telkomsel and XL Indonesian Telecoms Sector 11

12 Telkomsel and XL upwards. Second, there is a chance that XL could lose so much momentum that its management changes direction back towards discounting starter packs and lowering price points to attempt to recoup short-term subscriber growth. At present, we see very little likelihood of a reversal of strategy from XL, and so the greatest risk to the improvement in competitive dynamics is that Indosat gets 'too greedy' and fails to follow Telkomsel and XL upwards on pricing. Indosat's management has continually expressed a preference for higher price points, and its profitability is so low that, as we shall see, more stable data price points are vital to achieving a more reasonable return on capital. Our base case is therefore that Indosat will indeed follow Telkomsel and XL upwards, rather than risk a reversal in the direction of industry price points. Will SmartFren derail this improvement? In contrast, the restructured SmartFren might be expected to be disruptive. SmartFren was formed by the merger of Mobile 8 with Smart Telecom, owned by the Sinar Mas Group (Not listed). The two operators originally used CDMA, and entered financial difficulties after suffering from a lack of critical mass with what proved to be a transitional technology. As such, SmartFren arguably represents industry consolidation (a 'positive' for industry structure), rather than a 'new' entrant (which is in most cases negative for competitive dynamics). However, following a spectrum swap, SmartFren has emerged with 30MHz of unpaired 2.3GHz spectrum, on which it is launching TD LTE services. This is leading the Sinar Mas group, which is ultimately a well-capitalised business group, to attempt to resuscitate SmartFren as a cellular service. SmartFren has an interesting spectrum allocation Figure 17: Spectrum allocations by operator 850 MHz 900MHz 1.8GHz 2.1GHz 2.3GHz Telkomsel 5*2 7.5*2 22.5*2 15*2 Indosat 2.5*2 10*2 20*2 10*2 XL 7.5*2 22.5*2 15*2 Hutch Indonesia 10*2 10*2 SmartFren 10*2* 30*1 * SmartFren holds 5MHz but is also using 5MHz from Bakrie Telecom. In general, we tend to see four negative impacts from the entry of well-capitalised, infrastructure-based players into a cellular market. Firstly, spectrum costs are likely to rise in those markets in which spectrum is auctioned by the government though this is not the case on this occasion. Second, when the new entrant launches its starting point of zero revenue market share, combined with heavy fixed costs (depreciation, amortisation and interest charges) following spectrum purchase and network rollout, encourage it to attempt to gain revenue share. This is often done through: (1) tariff cuts and (2) aggressive promotional activity, such as advertising, dealer commissions and possibly handset subsidies. This, in turn, will affect both the revenue (through market share loss and potentially lower market price points) and EBITDA margins (through increased 'defensive' expenditure) of incumbent operators. Finally, if industry price points are dragged lower, stimulation of traffic volumes through lower prices is likely to lead to higher capex across all players, including incumbent. Incumbents may also choose to spend more on capex to preserve quality differentials versus (invariably cheaper) new entrants as long as possible. In summary, incumbent's spectrum costs could rise, revenue could fall, margins could fall and capex could rise. All of these impacts are DCF-negative, and will also feed through into earnings over time. To be clear, this set of impacts does not pre-suppose that the 'new entrant', in this case SmartFren, will actually generate positive returns on invested capital (ROIC) in its own cellular business. In fact, given the heavy investment required, we doubt that it will. Furthermore, in the case of the Indonesian market, we suspect that the impact of SmartFren's tariffs and promotional activity may not have a discernible impact on the 'Big 3' operators (Telkomsel, Indosat and XL). The reason is scale. As set out below, but lacks network resources Indonesian Telecoms Sector 12

13 SmartFren's current network is so small that it is unlikely to be able to influence nationwide price points. At present SmartFren only has 6,000 BTS, and all of those were originally constructed for 850MHz CDMA services. They are currently in the midst of being upgraded to FDD-LTE functionality. Were SmartFren to catch up with Indosat and XL on BTS rollout, in order to attempt to become a more serious nationwide competitive threat, this will clearly involve a great deal of time and (wasted) capital. As we will see later in this report, XL and Indosat currently generate returns on capital of only 3.6% and 4.1%, respectively, in spite of commanding revenue market share of 17.9% and 16.8%, respectively. At present, SmartFren's revenue market share is just 2.6%. It has Rp5.5 tn in net debt and EBITDA is barely in positive territory, resulting in a net debt to EBITDA ratio of 19.9x, and this is before the actual 4G launch costs kick in. Figure 18: BTS by operator as at 1Q15 000s 2G BTS 3G/4G BTS Total BTS Proportion Telkomsel 46,996 43,556 90, % Indosat 22,212 18,544 40, % XL 36,331 16,431 52, % Hutch Indonesia* 14,000 14,000 28, % SmartFren 2,000 4,000 6, % Total BTS 121,539 96, , % Source: Company data for Telkomsel, Indosat, XL and SmartFren, * Credit Suisse estimates for Hutch Indonesia Furthermore, the fact that there are already other small, sub-scale, players in the Indonesian market, with similar business models, suggests that they could bear the brunt of any success by a revitalised SmartFren in gaining initial traction in city areas. In particular, as shown in Figure 18 above, Hutch Indonesia looks vulnerable given its own price-based strategy, as well as its lack of available spectrum for the LTE rollout (since its 1800MHz spectrum is currently used to support voice services for its existing 2G subscriber base). In any case, since we estimate that over 20.0% revenue market share at current price points would be required for SmartFren to be able to generate reasonable returns on capital on a nationwide network, and since we believe that such an outcome is impossible given the relative strength of spectrum resources and network assets of the 'Big 3' operators, we suspect that after a few years of value destruction, the Sinar Mas Group will likely further consolidate SmartFren, for example by selling it for its spectrum resources. Indonesian Telecoms Sector 13

14 How far does mobile data pricing need to increase? Data is the growth driver monetisation is crucial Link Net is currently generating an attractive return due to its focus on the high growth/high return fixed broadband market, and in particular its 'cherry picking' strategy; Link Net does not roll out fresh nodes unless it can identify a minimum of 100 target homes (with the correct income demographics), per node. Link Net's FY14 return on invested capital (ROIC) was 23.3%. Link Net's FY14 ROIC was 23.3%... Cellular operators do not have the luxury of 'cherry picking' coverage to this level of granularity. In order to attract subscribers, cellular players must be able to demonstrate broad, seamless coverage, if not on a nationwide basis then at the very least across the whole of Java Island and the key cities outside of Java. Given that this overall coverage capex requirement, together with associated operating costs, is common to all of the operators, the relative scale or revenue market share is the primary determinant of returns on invested capital. Thus we observe that Telkomsel, with 58.5% service revenue market share as at 4Q14, generated an FY14 EBITDA margin of 56.2%, enjoyed an asset turn of 1.2x, and delivered a return on invested capital (ROIC) of 35.5%. and Telkomsel generated ROIC of 35.5% In contrast, Indosat generated an ROIC of just 4.1% and XL generated an ROIC of 3.6%. We consider these returns to be below their costs of capital; we estimate their respective weighted average costs of capital (WACC) to be 11.2% and 10.7%, respectively. The crux of the problem is that, due to the aforementioned coverage capex requirements, Indosat and XL's invested capital, at Rp36.9 tn and Rp35.2 tn, respectively, represents 68.1% and 65.0%, respectively, of Telkomsel's invested capital. Yet at Rp24.1 tn and 23.6 tn, respectively, Indosat and XL's service revenue represents only 36.4% and 35.6% of Telkomsel's FY14 service revenue of Rp66.3 tn. Given network operating costs, Indosat and XL's EBITDA margins are also much lower than Telkomsel's, at 41.8% and 36.6%, respectively. Indosat generated an ROIC of just 4.1% and XL generated an ROIC of 3.6%. Figure 19: FY14 ROIC analysis Rp bn Link Net Telkomsel Indosat XL Subscribers (000s) ,585 63,200 59,623 Service revenue 2,136 66,252 24,085 23,569 EBITDA 1,231 37,258 10,059 8,623 EBITDA margin 57.6% 56.2% 41.8% 36.6% Depreciation to sales (%) 18.3% 17.6% 33.3% 29.5% EBIT ,586 2,039 1,666 EBIT margin 39.3% 38.6% 8.5% 7.1% Tax rate (%) 26.1% 25.0% 25.0% 25.0% NOPLAT ,190 1,529 1,249 Average invested capital 2,658 54,107 36,852 35,181 Asset turn ROIC 23.3% 35.5% 4.1% 3.6% The most obvious way for Indosat and XL to drive up their returns is to generate more revenue. Specifically, if Indosat and XL increased service revenue by circa 20 21%, respectively, while keeping operating cost increases limited to circa 10%, then both companies would generate returns in line with cost of capital, and therefore cease to destroy value. Indonesian Telecoms Sector 14

15 Figure 20: ROIC analysis how much more revenue do Indosat and XL need to generate? Rp bn Indosat XL Actual service revenue 24,085 23,569 Required uplift (x) Required uplift (Rp bn) 4,867 4,848 Required service revenue 28,951 28,417 Operating costs (15,428) (16,440) EBITDA 13,523 11,977 EBITDA margin 46.7% 42.1% Depreciation charge (8,020) (6,958) Depreciation to sales (%) -27.7% -24.5% EBIT 5,503 5,019 EBIT margin 19.0% 17.7% Tax rate (%) 25.0% 25.0% NOPLAT 4,127 3,764 Average invested capital 36,852 35,181 Asset turn ROIC 11.2% 10.7% We note that the revenue uplift required by each player represents around 4.2% of FY14 industry revenues. We find it extremely unlikely that Indosat and XL will be able to achieve a combined 8.4% industry revenue share capture from Telkomsel, which has in fact been gaining, rather than losing market share given its super network coverage and quality. Thus, for Indosat and XL, the most credible route to achieving a return in line with cost of capital is for industry revenues to grow (and, if possible, for them to grow slightly faster than the industry as a whole). As a reminder, the industry is growing, having expanded by 7.1% YoY into 1Q15, though as mentioned Indosat and XL only grew by 5.4% and 2.8%, respectively, while Telkomsel achieved 12.1% YoY revenue growth. Figure 21: YoY revenue growth trajectory (%) 4Q12 1Q13 2Q13 3Q13 4Q13 1Q14 2Q14 3Q14 4Q14 1Q15 Telkomsel 15.5% 13.3% 10.8% 7.6% 9.3% 9.8% 10.3% 10.0% 11.3% 12.1% Indosat 16.2% 16.5% 8.4% -4.2% 1.2% -2.1% -2.2% 0.3% 6.0% 5.4% XL 10.9% 3.0% -0.3% -1.5% 3.1% 10.3% 16.2% 10.2% 9.7% 2.8% Others 3.3% 7.0% -0.1% -4.2% 1.9% -17.7% -28.4% -25.8% -27.0% -14.5% Total 13.5% 11.2% 7.1% 2.5% 6.0% 4.9% 5.4% 5.0% 6.5% 7.1% We see two potential paths to higher revenue for Indosat and XL, namely, (1) rising data volumes with more stable data price points or (2) rising data price points. We expect rising data volumes to drive revenue Fortunately for all of the 'Big 3' operators, rising data volumes are, in our view, extremely likely in Indonesia. As set out below, smartphone penetration was still only at 31.3% in December However, the structural driver of rapidly declining smartphone price points, as a result of commoditisation of chipsets, operating systems, batteries, screens and processors, is extremely powerful. Thus, the cheapest TD LTE (4G) smartphone in China now costs US$48. Thus, price points have fallen 76.8% since we started tracking TD LTE prices in May 2014 and have fallen 40.0% in the last four months alone. Smartphone price points continue to decline This dynamic is important for Indonesia, since smartphone price points have hit a 'sweet spot' for engaging middle income consumers. Thus, the pace of smartphone penetration has increased from less than 1.0 pp per month in 2014, to currently circa 1.3 pp per month, and we expect smartphone penetration in Indonesia to reach 79.2% by December Indonesian Telecoms Sector 15

16 Smartphone penetration 11 August 2015 Figure 22: Smartphone penetration 2014 vs 2017 Figure 23: Smartphone penetration versus GDP/capita 100.0% 90.0% 80.0% 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 0.0% % 90.0% 80.0% 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% Thailand Taiwan Malaysia China Indonesia Philippines India Korea Hong Kong Singapore Australia 0.0% - 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 GDP per capita Source: Credit Suisse estimates Source: Credit Suisse estimates While the amount of data consumed per smartphone customer is not clearly or consistently disclosed by operators, we estimate that, on an average, Indonesian smartphone customers use circa MB/month. This is higher than the average usage of only 349MB/month in China, but still some way short of Thailand's GB/month and developed market data use of circa GB per month. We have found that as smartphone screen sizes increase, processors get faster, and network download speeds increase, smartphone users tend to consume increasing amounts of video and other multimedia content and this is what drives data use from a few hundred MB per month into GB. As the popularity of YouTube and similar content increases in Indonesia, there should be a strong structural driver towards higher data use per subscriber. Figure 24: Data use per 3G subscriber (1Q15) 2,000 1,800 1,600 1,400 1,200 1, Telkomsel Indosat XL MB/month Figure 25: Indonesia in a regional context China Indonesia Thailand Singapore Hong Kong Taiwan MB/month Source: Credit Suisse estimates Given this, there should be scope for rising revenue as both the number of smartphone subscribers rises, and the amount of data use per subscriber increases, so long as data is not made available to customers in the form of 'unlimited' data plans. Data price points need to at least stabilise, or potentially increase The key question is therefore whether or not the Indonesian operators are going to price data appropriately to drive service revenues from what is likely to be rising data volumes. So far, evidence suggests that data pricing in Indonesia is sub-optimal. We have set out our standard analysis of data packages below (which assumes that rational customers find the cheapest way to consume 200 minutes of voice and 250MB, 700MB, 1GB, 2.5GB and 4GB of data). The good news is that the operators are not offering unlimited data plans; the pricing curve slopes up and to the right, as subscribers must pay more to use more. Indonesian Telecoms Sector 16