Indonesia Telecoms Sector

Transcription

1 Asia Pacific/Indonesia Equity Research Telecommunication Services Research Analysts Colin McCallum, CA Contribution by Joel Ying, CFA, FRM Indonesia Telecoms Sector SECTOR REVIEW Predatory pricing II: This time it's data Figure 1: Cash flow yield of the Big 3 versus government bond yield curve 18.0% 16.0% 14.0% 12.0% 10.0% 8.0% 6.0% 4.0% 2.0% 0.0% -2.0% -4.0% 13E 14E 15E 16E 17E XL Indosat PT Telkom Government bond yield Data price points in Indonesia have declined by % since last July. We believe this has reached predatory pricing levels and that CDMA operators and GSM new entrants will never be able to generate ROIC over cost of capital. However, even at these revised prices, ARPU uplift from data is x for subscribers consuming bundled plans with 250 MB-4 GB of data/month; enough of a data-driven uplift to allow the existing big 3 GSM players to grow at 9.9% in FY13, and generate attractive ROIC medium term. PT Telkom offers double-digit FY13E earnings growth, a 6.0% free cash flow yield and a 4.1% dividend yield. We maintain our OUTPERFORM rating and Rp13,000 target price, in spite of a likely technical overhang (sale of 1.1% of shares in issue) which might crystallise in August Indosat s execution has improved, and despite a 15.2% downward earnings revision on higher interest costs, our core revenue and EBITDA forecasts are broadly unchanged, as is our Rp8,300 target price (48.2% potential upside and OUTPERFORM rating. XL Axiata offers 45.6% upside. XL s tariff moves in 4Q12 have stabilised market share as predatory price levels were reached, and XL started to gradually raise price points again in February We believe FY13 will prove to be the bottom for XL s earnings and 1Q13 earnings figure will be the bottom within FY13. Our DCF-based target price of Rp7,500 implies 45.6% upside: OUTPERFORM. DISCLOSURE APPENDIX CONTAINS 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 Client-Driven Solutions, Insights, and Access

2 Focus charts and tables Figure 2: Emerging market cellular operator tick-list PT Telkom Indosat XL Low smartphone penetration Yes Yes Yes Tiered (not unlimited) dataplans Yes Yes Yes No handset subsidies Yes Yes Yes Cash flows paid as dividends Yes Yes Yes M&A acquisition risk Medium Low Low Source: Credit Suisse research Figure 3: Total monthly fee previous (July 2012) Figure 4: Total monthly fee current (May 2013) Non-data users Low Low-Medium Medium Medium-high High volume Non-data users Low-Medium Medium-high Telkomsel Indosat XL Telkomsel Indosat XL Source: Company data Source: Company data Figure 5: Data price points have come down a long way, to what we would deem predatory price points Rp/MB Non-data users Low (250 MB) Low-Medium (700 MB) Medium (1 GB) Medium-high (2.5 GB) High volume (4 GB) Indonesia Previous n/a Indonesia Current n/a % Chg n/a -75.3% -72.2% -73.3% -68.3% -58.9% Source: Company data Figure 6: However, the ARPU uplift from data is still meaningful packages as a multiple of 1Q13 average ARPU (x) Non-data users Low (250 MB) Low-Medium (700 MB) Medium (1 GB) Medium-High (2.5 GB) High volume (4 GB) Telkomsel Indosat XL Source: Company data Figure 7: Indonesia telecoms sector comparative multiples Current Target Upside P/E (x) EV/EBITDA (x) FCF Yield (%) Div Yield (%) Price (Rp) Price (Rp) (%) FY13E FY14E FY13E FY14E FY13E FY14E FY13E FY14E PT Telkom 11,900 13, Indosat 5,600 8, XL 5,150 7, Bakrie n.m. n.m Asia ex Japan Integrated Asia ex Japan Wireless Indonesia Telecoms Sector 2

3 Predatory points reached on data At these price points, there is only room for 3 players Data price points in Indonesia have declined by % since last July. We believe this has reached predatory pricing levels and that CDMA operators and GSM new entrants will never be able to generate ROIC over cost of capital, particularly since bundled plans of voice, data and SMS mean that high voice and SMS interconnect rates still confer scale disadvantages on smaller players. Further investment from these players will likely be irrational. For the Big 3 players, on the other hand, even at these revised price points, inclusion of data within bundles is resulting in a x ARPU uplift for subscribers consuming bundles of 250 MB-4 GB of data/month. This level of data-driven ARPU uplift is enough to raise the revenue trajectory of the big 3 players, namely Telkomsel, Indosat and XL to 9.9% for FY13. It is also enough ARPU uplift to ensure that Indosat and XL, with 18.4% and 17.9% current revenue market share, can generate reasonable returns on capital in the medium term as their 3G network utilisation rises. Telkomsel, with 54.0% revenue market share, already generates 34.8%, and rising, ROIC. PT Telkom: Data growth outweighs share overhang Telkomsel is monetising the smartphone ARPU uplift effectively and looks set to deliver double-digit revenue growth (10.4%) in FY13. Scale advantages have also minimised the impact of the 3G investment phase, allowing relatively stable margins, double-digit FY13 earnings growth and FY13 ROIC of 34.8%. The growth trajectory of the fixed line division is also finally improving; we forecast 14.7% consolidated earnings growth in FY13. We expect 211 mn shares held in Treasury stock to be sold back to the market in August. While this also creates a small technical overhang (1.1% of shares in issue), a potential release of value from Telkom s tower assets provides an offsetting positive catalyst. Indosat: Strong core recovery, despite earnings drag Improved execution resulted in 16.5% YoY revenue growth for Indosat in 1Q13 and four consecutive quarters of revenue market share gains. With 900 MHz 3G now being rolled, boosting data revenues, we forecast that Indosat can grow FY13 cellular revenues at 11.1%, faster than the overall market. Furthermore, the 900 MHz strategy should allow Indosat to avoid the severity of the capex and opex shock suffered by XL in its GHz 3G rollout over the last two years. Accelerated depreciation and high interest costs, both as a result of legacy overspending in , are a material drag on headline earnings and we have cut our FY13E earnings figure by 15.2%. FY13 ROIC is also low at 3.8%, again a legacy issue. Crucially, headline earnings are on a sharp upward trajectory as accelerated depreciation rolls off, and ROIC is expected to reach 9.7% (close to WACC) by FY16. Indosat s 900 MHz 3G investment is worth making: OUTPERFORM. XL currently offers 45.6% upside XL s revenue market share fell by 1.6 pp to 17.9% in 1Q13, from 19.5% in 2Q12, triggering sharp share price underperformance since September The catalysts for XL s market share loss were improved execution at Indosat as well as some vulnerability to Hutch Indonesia s (in our view unsustainable) promotional activity. XL s profitability has also been badly squeezed by the impact of its accelerated GHz 3G rollout on capex and opex (in particular, infrastructure expenses). The impact was so severe we have cut our FY13 earnings forecast by 19.0%. On the other hand, we believe FY13 will prove to be the bottom for XL s earnings and that the 1Q13 earnings figure will be the bottom within FY13. XL started to raise price points again in February 2013, capex has already peaked and opex growth will also slow. As XL emerges from its 3G investment j-curve we expect ROIC to rise from 10.4% in FY13 to 14.8% in FY16: OUTPERFORM. Data price points just high enough for all of the big 3 to generate reasonable returns, but not high enough for sub-scale new entrants and CDMA players: We rate Bakrie Tel UNDERPERFORM We rate PT Telkom OUTPERFORM, with a target price of Rp13,000 We rate Indosat OUTPERFORM, with a target of Rp8,300 We rate XL OUTPERFORM, with a target of Rp7,500 Indonesia Telecoms Sector 3

4 Sector valuation Figure 8: Regional comparative multiples Closing Target Normalised P/E EV/EBITDA FCF yield (%) Div. yield (%) Code Curr. price Rating price 2013E 2014E 2013E 2014 E 2013E 2014E 2013E 2014E Integrated operators China Telecom 728.HK HK$ 4.12 O China Unicom 762.HK HK$ O Chunghwa 2412 TT NT$ U HTHK 215 HK HK$ 4.60 U KDDI 9433 JP 5,190 O 5, KT KS W 40,850 O 45, NTT 9432 JP 5,460 O 4, Softbank 9984 JP 5,820 O 6, PCCW 8 HK HK$ 3.94 O HKT Trust 6823 HK HK$ 8.02 N PLDT TEL PM P 3,210 O 3, SingTel ST SP S$ 4.00 O TNZ TEL NZ NZ$ 2.55 U CNU CNU NZ NZ$ 2.75 N (16.4) (12.0) SKB KS W 5,210 U 3, Telekom Malaysia T MK RM 5.46 U Telstra TLS AU A$ 5.07 N True Corp TRUE TB Bt 9.60 U 2.86 n.m. n.m (14.1) (3.8) - - NJA average integrated Mobile operators AIS ADVANC TB Bt O AXIATA AXIATA MK RM 6.91 O Bakrie BTEL IJ Rp U n.m. n.m (12.5) (11.1) - - Bharti BHARTI IN Rs O China Mobile 941 HK HK$ O DiGi DIGI MK RM 4.63 O XL EXCL IJ Rp 5,150 O 7, (1.4) FarEasTone 4904 TT NT$ N Globe GLO PM P 1,572 N 1, IDEA IDEA IN Rs O Indosat ISAT IJ Rp 5,600 O 8, LGU KS W 11,900 O 9, Maxis MAXIS MK RM 6.94 O M1 M1 SP S$ 3.38 O NTT DoCoMo 9437 JP 163,000 N 120, PT Telkom TLKM IJ Rp 11,900 O 13, Reliance RCOM IN Rs U SKT KS W 211,000 N 155, SmarTone 315 HK HK$ U StarHub STH SP S$ 4.47 N TAC DTAC TB Bt O Taiwan Mobile 3045 TT NT$ N NJA average mobile Asia average telecoms Note: (1) Rating: O = Outperform; N = Neutral; U = Underperform; R = Restricted, (2) The averages are based on market capitalisation, (3) The P/E for non-asian stocks are based on Credit Suisse adjusted EPS, (4) The financial years of KDDI, NTT, NTT DoCoMo, Softbank, Bharti, Reliance and SingTel are ended in March. For the sake of comparison, the FY05 of these companies in this matrix represents FY3/06 of their financial years and etc., (5) FCF yield = (EBITDA - interest exp. - tax - capex) / mkt cap Indonesia Telecoms Sector 4

5 Smartphone penetration 20 May 2013 At these price points there is only room for 3 players Smartphone growth despite subsidy discipline The trend of rapidly increasing penetration of smartphones in Indonesia and other Asian emerging markets has continued. Our current forecasts, which are built market-by-market, project 1.1 bn smartphone users in Asia ex-japan by 2015, up from 330 mn as at December While we expect the developed markets of Singapore, Hong Kong and Australia to reach penetration rates of % of population by that point, the key factor behind a tripling of total smartphone subscribers by 2015 will be a sharp increase in penetration rates of populous emerging markets. We forecast that smartphone penetration in China will reach 41.2% (from 11.8% as at FY12), India will reach 14.3% (from currently 4.0%), and Indonesia will reach 55.9% (from currently 15.3%). Thai smartphone penetration is expected to reach 67.4% by 2015, following the issuance of 3G spectrum, up from an estimated 17.2% as at FY12. As always, rising penetration is expected to result from a combination of rising nominal GDP, a decline in handset price points and a decline in data usage tariffs. We are very confident that our penetration forecast of 1.1 bn will be reached, if not exceeded, despite the current softening of regional (and global) nominal GDP growth rates, because the primary driver in this case is set to be a collapse in price points for smartphones (i.e., the second of the three drivers mentioned above). This in turn will be driven by a battle among four key operating systems the Apple ios, the Android system, RIM and Windows and the rapid commoditisation of smartphone production. Smartphone penetration rising rapidly helped by declining handset price points Figure 9: Smartphone penetration versus GDP Figure 10: Emerging markets will be the key growth driver 70.0% 60.0% 50.0% 40.0% 30.0% Malaysia Korea Taiwan Hong Kong Singapore Australia 80.0% 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% Indonesia Thailand 10.0% China Philippines 0.0% India - 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80, % 10.0% 0.0% Singapore Hong Kong Australia Korea Taiwan Thailand Indonesia China Philippines India GDP per capita Thus, 3G Smartphone penetration in the Indonesian market reached an estimated 15.3% of population in December 2012, almost doubling from the 8.4% level recorded in December We forecast the number of 3G smartphones will almost double again, to 65 mn, or 26.4% penetration, as at December The question is whether resulting data growth can be monetised Of course the key question for the investors assessing the Indonesian cellular services sector is whether or not rising smartphone penetration will drive higher revenues, higher EBITDA and higher cash flows for the operators. We have seen across the region, and, indeed, globally, that monetisation of rising data volumes from rising smartphone penetration is superior in markets in which the competitive and regulatory environment is benign enough to ensure (1) tiered, rather than unlimited, data plans, and (2) where operators do not subsidise smartphones. Data pricing structures are crucial to monetising this potential opportunity Indonesia Telecoms Sector 5

6 Figure 11: Emerging market cellular operator tick-list PT Telkom Indosat XL Low smartphone penetration Yes Yes Yes Tiered (not unlimited) dataplans Yes Yes Yes No handset subsidies Yes Yes Yes Cash flows paid as dividends Yes Yes Yes M&A acquisition risk Medium Low Low Source: Credit Suisse research From a top-down perspective, the Indonesian cellular sector does meet both criteria for successfully monetising data, while the low smartphone penetration level suggests high potential for data volume growth. In order to better understand the details of data pricing in Indonesia, we have updated the analysis first unveiled in our 24 July 2012 report Uncomplicating data pricing. In this report, we defined data access consumers in Indonesia, and across Asia, into six different standard categories: Non-data users (0 MB/month), Low data users (250 MB/month), Low-medium data users (700 MB/month), Medium data users (1 GB/month), Medium-high data users (2.5 GB/month) and High-volume data users, (4 GB/month). The key variable here of course is the quantity of data consumed. In reality, most consumers also purchase voice services. In order to capture the voice variable, but reduce the amount of distortion it might cause, we assumed that all of our customers wish to consume 200 minutes of outgoing voice calls per month. Figure 12: Requirement of six sub-group subscribers User type Non-data users Low data users Low-med data users Med data users Hi-med data users Very high data users Assumed data per month MB 700 MB 1 GB 2.5 GB 4 GB Assumed "outgoing" voice per month Source: Credit Suisse estimates We then looked into which data package each of our six standard consumers groups would choose under each operator under our coverage universe; of course we assumed that in each case the standard subscriber would be rational and would take up the cheapest package that contains the required amount of voice and data. This allowed us to build a picture of the implied ARPU level generated by each standard consumer in Indonesia and across the region. It also allowed us to calculate an implied tariff per MB across each standard data consuming group. Indonesia is meeting the monetisation criteria overall.. Figure 13: Total monthly fee previous (July 2012) Figure 14: Total monthly fee current (May 2013) Non-data users Low Low-Medium Medium Medium-high High volume Non-data users Low-Medium Medium-high Telkomsel Indosat XL Telkomsel Indosat XL Source: Company data Source: Company data Indonesia Telecoms Sector 6

7 In July 2012, the simple average implied ARPU of data users in Indonesia rose in a very steep upward curve from Rp31,583/month (US$3.2/month) for customers who did not use data services to Rp332,250/month (US$34.1/month) for those who consumed 4GB per month. Within this average there was also a clear differential between Telkomsel s price point (Rp398,750/month, or US$41.1/month, for a 4 GB user) and the other two players (XL and Indosat at Rp300,000/month and Rp298,000/month, respectively). Date price points have fallen sharply, but unlimited offers still avoided Over the past year data pricing in Indonesia has becoming increasingly complex. While the standard price point of Rp1/KB (US$0.103/MB) to Rp2/KB (US$0.203/MB) is still in force for pay as you go plans, the tiered data bundles have been significantly overhauled, and, with the inclusion of bonus data in bundles, the effective average data price point in the industry has declined materially. Figure 15: Total monthly fee comparison Rp/Month Non-data users Low (250 MB) Low-Medium (700 MB) Medium (1 GB) Medium-high (2.5 GB) High volume (4 GB) Indonesia previous 31, , , , , ,250 Indonesia current 30,000 62,378 75,711 75, , ,544 % Chg -5.0% -64.8% -63.5% -64.9% -62.9% -54.7% Source: Company data Thus, the average implied ARPU for subscribers using data has more than halved over the last year. Customers consuming 250 MB will now pay an average of Rp62,378, down 64.8% from the Rp177,083 which would have been paid last July. The average implied ARPU for high-volume subscribers using 4 GB of data has declined by 54.7%, from Rp332,250/month to Rp150,544/month. Reversing out the implied value of the 200 minutes of voice consumed by our standard consumers, the average price points for data have declined by 75.3% for low-volume (250 MB/month) packages, from Rp613/MB to Rp151/MB (S$0.02/MB). Figure 16: Effective "Data" price comparison Rp/MB Non-data users Low (250 MB) Low-Medium (700 MB) Medium (1 GB) Medium-high (2.5 GB) High volume (4 GB) Indonesia previous n/a Indonesia current n/a % Chg n/a -75.3% -72.2% -73.3% -68.3% -58.9% Notwithstanding the decline in the price per MB, for us the crucial factor is the absence of unlimited plans, which the operators have weeded out over the last year and a half. The requirement to pay a higher amount to consume larger volumes of data ensures that the economics of consumer behaviour match the cost structure of the operators. We view unlimited data plans, which still exist in some hyper-competitive markets such as Hong Kong, as extremely damaging for telecoms operators consumers are incentivised to consume more and more data, driving up capex, spectrum costs and opex for operator, yet operators receive no additional revenue from subscribers. Data prices comparable with other high-growth emerging Asian telco markets We also note that the even the pure data price points in Indonesia are comparable to those in other markets. In the Bt399, Bt599, Bt799 and Bt999 3G packages launched by AIS on 7 May, the data price ranges Bt /MB, or US$0.008-US$0.014/MB. In China, in 4Q12, China Mobile s data price was Rmb0.23/MB, or US$0.04/MB around four times higher, and Unicom s pricing was Rmb0.16/MB, or US$0.025 two and a half times higher. However, 74.2% of market leader China Mobile s data traffic is going over WiFi and China Mobile is charging only Rmb0.002/MB (US$0.0003/MB), or 100 x less (!) for WiFi access; Indonesian data pricing at US$ looks relatively high (and therefore better news for investors) in comparison...though price points have declined by 58.9%-75.3% Nevertheless, these are similar data price points to those found in Thailand Indonesia Telecoms Sector 7

8 leading to a material ARPU and revenue uplift Figure 17: Data pricing in stark contrast to Indonesia s low voice price average voice revenue per minute (RPM) Developed markets US$/min Developing markets US$/min Australia 0.13 China 0.03 Hong Kong 0.03 India 0.01 Korea 0.04 Indonesia 0.01 New Zealand 0.25 Malaysia 0.05 Singapore 0.12 Philippines 0.03 Taiwan 0.16 Thailand 0.02 Indeed, in the context of Indonesia s extremely low voice tariffs and resulting low blended average ARPU level, even at these revised price points the initial ARPU uplift as customers begin to consume data and further ARPU kickers as consumption levels rise, is extremely material. Thus, even low-volume data users consuming only 250 MB/month of data are required to pay 2.2x the average ARPU level generated by Telkomsel subscribers, 1.8x the average ARPU level generated by Indosat subscriber and 1.9x the average ARPU level generated by XL subscribers. This analysis gives credence to comments by management that at current data packages and average data consumption (less than 500 MB/month), ARPU levels are broadly doubling when customers purchase a smartphone. Figure 18: Standard data packages as a multiple of 1Q13 average ARPU (x) Non-data users Low (250 MB) Low-Medium (700 MB) Medium (1 GB) Medium-high (2.5 GB) High volume (4 GB) Telkomsel Indosat XL Smartphone penetration had reached only 15.3% as at December 2012, yet industry revenues for FY12 reached Rp102.4 tn. Assuming that all price points remained unchanged, and that data usage and the resulting ARPU generation remained at circa 2x, the increase in smartphone penetration from the current 15.3% to 100% by 2020 would lift industry revenues by a further 73.5%, to Rp177.7 tn. In reality, we would expect data price points to decline from current levels, but a strong elasticity response (i.e., higher data consumption per subscriber) could ensue, resulting in a similar multiple of ARPU expansion being achieved. Should this occur and should cannibalisation of traditional voice and SMS revenues continue to be minimised through use of bundled packages, our current forecast of Rp147.7 tn in revenues by FY20 could prove too conservative. In other words, our current implied assumption that telecoms expenditure would decline from FY12 s 1.5% of nominal GDP, to our projection of 1.2% of nominal GDP by FY20, could prove to be too bearish given the strength of unmet data demand in Indonesia, should the competitive environment remain broadly stable and a move to unlimited data pricing be avoided. Still a double-digit growth market into 1Q13 Thus rising smartphone penetration and 16.3% YoY data revenue growth contributed to double-digit (11.7%) YoY overall revenue growth into 1Q13 (though the data growth was also inflated by the introduction of SMS interconnect, and therefore a shift to reporting gross SMS revenue, starting from 2Q12). Indonesia Telecoms Sector 8

9 Figure 19: Indonesian cellular sector 1Q13 QoQ and YoY analysis (%) Cellular service voice data * 1Q13 data* Consolidated Consolidated revenue revenue revenue as % of total EBITDA net profit QoQ (%) YoY (%) QoQ (%) YoY (%) QoQ (%) YoY (%) revenue QoQ (%) YoY (%) QoQ (%) YoY (%) PT Telkom Indosat XL (10.7) Total industry *SMS, VAS and mobile data revenues Nevertheless, focusing on mobile Internet access revenues alone, data growth rates were also high. For example, XL s non-sms data revenues grew 15.9% YoY. In total, data revenues including SMS now contribute just under half (48.0%) of the total revenues generated by the big 3 cellular players. Double-digit YoY growth rate maintained into 1Q13 However, the 11.7% YoY growth rate represented a slowdown compared with the 14.7% and 16.0% YoY growth rates achieved in 3Q12 and 4Q12, respectively, and the 4.5% QoQ revenue decline shows that this year the structural positive of rising smartphone growth was not strong enough to outweigh the post-labaran seasonality. The main reason for this was weak revenue growth from the third player XL, which has grown at a slower pace than the industry for the past three quarters (3Q12-1Q13 inclusive). Thus, XL s share of total mobile industry revenues (including the CDMA players) has declined from 19.5% as at 2Q12 to 17.9% as at 1Q13. Figure 20: Net revenue and revenue market share Net revenue (Rp bn) 3Q11 4Q11 1Q12 2Q12 3Q12 4Q12 1Q13 Telkomsel 12,807 12,704 12,298 13,128 14,432 14,673 13,928 Indosat 4,360 4,164 4,080 4,447 5,124 4,839 4,751 XL 4,373 4,403 4,477 4,858 5,145 4,883 4,614 Hutch Others 1,731 1,762 1,645 1,724 1,783 1,701 1,717 Total 23,797 23,611 23,136 24,856 27,239 26,889 25,804 Revenue market share (%) Telkomsel 53.8% 53.8% 53.2% 52.8% 53.0% 54.6% 54.0% Indosat 18.3% 17.6% 17.6% 17.9% 18.8% 18.0% 18.4% XL 18.4% 18.6% 19.4% 19.5% 18.9% 18.2% 17.9% Hutch 2.2% 2.4% 2.7% 2.8% 2.8% 2.9% 3.1% Others 7.3% 7.5% 7.1% 6.9% 6.5% 6.3% 6.7% We put XL s weak revenue performance down to the revitalisation of Indosat in 2Q12 (due to a new management team, an improved network performance, and completion of a converged billing and campaign management system), as well as more aggressive competition from Hutch Indonesia (Not listed); XL s customer base proved more pricesensitive and more vulnerable to Hutch s activities, than that of either Telkom or Indosat. These dual pressures in turn resulted in XL lowering tariffs to regain subscriber momentum. For example, XL launched the 3x data quota for Hotrod 3G+ in September 2012, then followed up with bonus packages such as XLKU, which lowered both effective voice price points as well as data price points, on 30 November. These promotions were enough to stabilise XL s net addition share in 4Q12 despite XL s price points being on average twice that of Hutch Indonesia during the quarter. This suggested that the Predatory Price Point was reached, and there was also a clear pick-up in volumes of voice minutes, SMS and MB of data in 1Q13. The reason that these volume increases did not feed through into rising revenue into 1Q13 is the lag caused by lower price points in 4Q12 and in the first half of 1Q13. Indonesia Telecoms Sector 9

10 Enough growth for three players. High single-digits revenue growth possible for the Indonesian market in FY13 While the downward movement in price points from XL, and resulting YoY revenue growth of only 3.0% into 1Q13, acted as a drag on industry revenues in aggregate, both Telkomsel and Indosat were able to generate strong double-digit YoY revenue growth of 13.3% and 16.5%, respectively. Importantly, even XL s management felt confident enough to start raising price points in February 2013, and management stated that this was likely to continue to be done in the second quarter. As higher price points are applied to rising volumes, we expect strong revenue recovery into 2Q13 and beyond for XL. Thus, for FY13 as a whole, we forecast high single-digit revenue growth (9.6%), with market leader Telkomsel forecast to grow at 10.4% YoY, Indosat forecast to grow at 11.1% YoY, and XL, coming off a low trajectory in 4Q12-1Q13, expected to grow revenue at 7.2% YoY. Figure 21: Indonesian cellular revenue and revenue market share (Rp bn) 2009A 2010A 2011A 2012A 2013E 2014E 2015E 2016E Telkomsel 43,948 45,567 48,733 54,531 60,198 64,369 67,897 70,982 Indosat 13,929 16,027 16,751 18,489 20,533 22,081 23,550 24,743 XL 12,482 16,027 17,049 19,323 20,708 22,340 23,898 25,195 Hutch 820 1,339 2,007 2,883 3,517 4,221 4,643 5,014 Bakrie 2,984 3,055 2,863 2,626 2,522 2,706 2,853 2,989 Others 4,707 5,457 4,378 4,592 4,805 4,760 4,621 4,562 Total 78,869 87,473 91, , , , , ,486 Revenue growth Telkomsel 10.3% 3.7% 6.9% 11.9% 10.4% 6.9% 5.5% 4.5% Indosat -1.8% 15.1% 4.5% 10.4% 11.1% 7.5% 6.7% 5.1% XL 13.0% 28.4% 6.4% 13.3% 7.2% 7.9% 7.0% 5.4% Hutch 105.1% 63.4% 49.9% 43.6% 22.0% 20.0% 10.0% 8.0% Bakrie 20.0% 2.4% -6.3% -8.3% -4.0% 7.3% 5.4% 4.8% Others -3.0% 15.9% -19.8% 4.9% 4.6% -0.9% -2.9% -1.3% Total 8.3% 10.9% 4.9% 11.6% 9.6% 7.3% 5.8% 4.7% operational gearing also working in favour of the big 3 operators Telkomsel, with its scale advantages, continues to generate by far the highest returns on capital in the industry, at 31.9% in FY12. Telkomsel s large scale in revenue terms has also meant that the capex burden of rolling out blanket 3G coverage has been minimised. Even in FY12, the construction of 11,674 BTS in total, and 5,924 3G node B BTS s resulted in capex of Rp11.7 tn, a 21.5% capex-to-sales ratio. This had two implications. First, the knock-on impact of rising number of BTS on operational costs ( opex ) has been limited. Thus, Telkomsel s EBITDA margin declined by only 0.5 pp to 56.0% despite the 3G rollout. Second, a rising proportion of 2G assets have become fully depreciated, resulting in depreciation charges declining (and lowering the invested capital). Thus, Telkomsel achieved high operational gearing below the EBITDA line, with 22.5% YoY growth in net profit in FY12, and ROIC rose to 31.9% Looking into FY13, we expect the same dynamics to continue to apply. Thus, we forecast 10.4% revenue growth from Telekomsel as it continues to reap the benefit of rising smartphone penetration and competes on coverage and quality rather than price. Given the ongoing scale advantages and avoidance of subsidies we expect the EBITDA margin to decline by only 0.8 pp, resulting in 8.7% EBITDA growth in FY13. We expect depreciation charges to grow as more 3G assets are constructed, but for the charge to Telkomsel generates high return on capital and we expect returns to rise further Indonesia Telecoms Sector 10

11 grow more slowly than EBITDA, contributing to double-digit (11.1%) net profit growth. Returns on invested capital are forecast to rise from 31.9% in FY12 to 34.8% in FY13. We also expect to see core earnings growth for Indosat and XL, as utilisation rates on their 3G networks rise. Indosat has suffered from very low ROIC due to extremely inefficient investment in With the arrival of a new management team in 2009 capital expenditure almost halved from the Rp tn level of , but the high legacy capex, together with consistent revenue market share loss from 2Q07 though 1Q12, has left ROIC well below cost of capital. The impact of this continued to be felt in FY12, as it will in FY13 and FY14, as Indosat management recently amended the useful life of some legacy assets in order to clear out unproductive assets. This accounting policy change resulted in accelerated depreciation of Rp1.3 tn in FY12, lowering net profit and near-term ROIC, and this will continue with an Rp1.4 tn charge in FY13 and a Rp0.7 tn charge in FY14. On the other hand the commercial operations of Indosat have clearly improved since 2Q12, with the arrival of a new Chief Commercial Officer (CCO), improved network quality and completion of converged billing and campaign management system in May/June Thus, the cellular service revenue trajectory picked up from 4.5% growth in FY11 to 10.4% in FY12, and we expect this improved trajectory to continue with 11.1% cellular service revenue growth in FY13. Importantly, we believe this improved revenue trajectory can continue without an overly aggressive increase in capital intensity. While Indosat has noticeably lagged competitors Telkomsel and XL in the construction of 3G networks (as at December 2012 Indosat had only 4,596 3G Node B BTS but Telkomsel and XL had rolled out 15,433 and 13,142, respectively), Indosat was awaiting permission to roll out HSPA over its ample 900 MHz allocation. Use of 900 MHz HSDPA for blanket /large area coverage will be far cheaper than utilising MHz spectrum, allowing the cellular capex-to-sales ratio to remain below 40% through the rollout phase (in contrast to XL s 50.2% capex-to-sales ratio in FY12). Apart from lower capex, we expect a much less intense ramp-up in opex to accompany Indosat s 3G rollout, in contrast to the margin compression suffered by XL over the last 18 months. Again, this is due to far fewer BTS s being constructed (resulting in lower cell site rental expenses and power consumption). We therefore project that Indosat will report consolidated EBITDA growth of 10.2% in FY13 and 6.9% in FY14. ROIC is expected to improve to 5.8% in FY14 as the accelerated depreciation charges decline, and then to recover to 8.2% in FY15 and 9.7% in FY16. ROIC is expected to exceed the weighted cost of capital by FY17; we therefore view continued investment by Indosat, particularly using the 900 MHz HSPA strategy, as rational and likely to lead to reasonable returns. Indosat s ROIC is low for historical reasons but core results are improving and we expect rising ROIC FY14 onwards Indonesia Telecoms Sector 11

12 Figure 22: ROIC analysis of the Big 3 GSM operators (Rp bn) FY10 FY11 FY12 FY13E FY14E FY15E FY16E Telkomsel Revenue 45,567 48,733 54,531 60,198 64,369 67,897 70,982 Incremental revenue 1,619 3,166 5,798 5,667 4,171 3,528 3,085 Capex 8,197 8,500 11,700 10,655 10,943 10,864 11,357 Capex to sales (%) 18.0% 17.4% 21.5% 17.7% 17.0% 16.0% 16.0% ROIC (%) 24.2% 25.7% 31.9% 34.8% 37.0% 38.2% 39.5% Indosat Revenue 16,027 16,751 18,489 20,533 22,081 23,550 24,743 Incremental revenue 2, ,738 2,044 1,547 1,469 1,193 Capex 4,456 5,080 7,450 7,803 7,287 6,123 5,691 Capex to sales (%) 27.8% 30.3% 40.3% 38.0% 33.0% 26.0% 23.0% ROIC (%) 5.4% 5.0% 3.9% 3.8% 5.8% 8.2% 9.7% XL Revenue 16,819 18,004 20,267 21,652 23,284 24,842 26,139 Incremental revenue 3,737 1,185 2,263 1,385 1,632 1,558 1,297 Capex 4,848 6,522 10,176 8,620 6,520 5,465 4,966 Capex to sales (%) 28.8% 36.2% 50.2% 39.8% 28.0% 22.0% 19.0% ROIC (%) 17.3% 14.9% 13.1% 10.4% 10.9% 12.8% 14.8% With the rapid expansion in revenue market and scale, XL s ROIC rose from 6.9% in FY05 to reach a historical peak of 17.3% in FY10. While XL s revenue growth trajectory was slower than industry from 3Q12 through 1Q13, for FY12, its consolidated revenue growth was a respectable 12.5% YoY (albeit boosted the introduction of SMS interconnection). Instead, the decline in earnings and ROIC suffered in FY11 and FY12 was primarily related to costs. In particular, XL s decision to roll out 6,082 BTS in 2011 and 11,179 BTS in 2012, at a similar pace to the 6,066 and 11,674 BTS, respectively, rolled out by Telkomsel over the same periods, has a much bigger impact on XL s financial results, given that its revenue base is only one third of the size of Telkomsel s. Thus, XL s capex-to-sales ratio in FY12 was 50.2%, versus Telkomsel s 21.5%. Furthermore, the high capex had a knock-on impact on operating costs and EBITDA margin. For example, the total number of BTS (both 2G and 3G) rose by 39.5%, from 28,273 as at December 2011 to 39,452 as at December This in turn pulled up maintenance costs, cell site rental costs, power costs and so on, and so infrastructure expenses rose by 34.7% YoY, far faster than revenue. Gross EBITDA margin declined by 3.3 pp in FY11 to 49.4% and by further 3.6 pp in FY12 to 45.8%. EBITDA increased by only 0.7% in FY11 and 4.3% in FY12, and, after depreciation interest and forex, earnings declined by 2.1% in FY11 and 2.3% in FY12. ROIC declined by 2.4 pp in FY11 and a further 1.8 pp in FY12, to 13.1%. This trend continued into 1Q13, with YoY net cellular revenue growth slowing to only 5.6% YoY and gross EBITDA margin declining to only 40.1%; infrastructure expenses grew by further 8.0% QoQ and 12.3% YoY. With depreciation charges also rising, net profit declined by 44.6% QoQ and 52.7% YoY. Given this, we have lowered our FY13 revenue, EBITDA and net profit forecasts for XL by 2.5%, 6.8% and 19.0%, respectively. ROIC is expected to decline for the third consecutive year since the 3G rollout began, by a further 2.7 pp to 10.4% in FY13. On the other hand we believe FY13 will prove to be the bottom for XL s earnings, and we believe the 1Q13 earnings figure will be the bottom within FY13. Importantly, XL s management felt confident enough to start raising price points in February 2013, and management stated that this was likely to continue to be done in the second quarter. As higher price points are applied to rising volumes, we expect strong revenue recovery into 2Q13 and beyond for XL. The impact of the high number of BTS additions is also now clearly within the opex run-rate, and so the impact on EBITDA has already been felt. We XL s returns dipping due to an accelerated 3G rollout but XL is set to emerge from this investment j-curve Indonesia Telecoms Sector 12

13 note that the pace of BTS additions slowed to only 367 in 1Q13, versus 3,351 added in 4Q12, and we believe XL has clearly passed its peak rollout pace. The pace of opex growth should also therefore moderate. Big picture, we therefore continue to believe that while XL s EBITDA margin decline was steeper than anticipated in FY12 and 1Q13, this is a timing issue rather than a structural one; XL is facing a normal investment j-curve from which it will emerge. As the utilisation of the 3G network improves into 2Q13 and 2H13 and beyond, XL s EBITDA margin should rise, and as the capex-to-sale ratio declines ROIC should increase. Slightly higher price points being applied to the rising voice and data volumes will of course accelerate this process. We forecast a recovery in EBITDA margin to 42.8% for FY13 and 43.6% for FY14, up from the 40.1% achieved in 1Q13. ROIC is expected to rise to 10.9% in FY14 from a low of 10.4% in FY13 (though we note that even the FY13 ROIC is slightly above the WACC of 9.8%). We therefore believe XL is rational in pursuing its 3G rollout strategy and also its recent tactical data pricing strategy. As the figure below shows, the cash flow yields we project to be generated by all of the Big 3 GSM operators look attractive, particularly when compared with the current Indonesia government bond yield. PT Telkom s cash flow yield is attractive even in FY13, but for investors prepared to look two years ahead, the cash flow yields of both Indosat and XL are expected to exceed those of PT Telkom (and of course the government bond yield) by FY15, as 3G revenues rise and capital intensity moderates. It is also arguable that there can be higher confidence that the cash flows generated by Indosat will be paid out to equity holders in the form of dividends. XL is controlled by Axiata, while Indosat is controlled by Qatar Telecom. Given their existing portfolios of telco assets, any further international M&A activities are likely to be pursued by Axiata and Qatar Telecom, rather than XL and Indosat. Figure 23: Cash flow yield of the Big 3 versus government bond yield curve 18.0% 16.0% 14.0% 12.0% 10.0% 8.0% 6.0% 4.0% 2.0% 0.0% -2.0% -4.0% 13E 14E 15E 16E 17E XL Indosat PT Telkom Government bond yield...but not more than three In our August 2008 report Approaching predatory pricing we explained that its moves to slash voice points from 2Q07 through 3Q08 resulted in rapid scale expansion for XL, together with the destruction of the business models (and therefore deprivation of scale) of both CDMA players and also 1800 MHz-based GSM entrants. Both the CDMA players and the 1800 MHz-based GSM new entrants primarily used prices to attract subscribers and revenue growth, as they did not have comparable coverage or network quality (in the early stages of rollout, subscribers largely view coverage as the key Indonesia Telecoms Sector 13

14 differentiation point of network quality). The fixed wireless operators had the added disadvantage of limited functionality and handset ecosystem. Sure enough CDMA player Mobile 8 entered administration and eventually consolidated with the Sinar Mas CDMA operation to create SmartFren. PT Telkom s Flexi CDMA division has been starved of investment and revenue has collapsed by 60.0% between the 4Q07 peak and 1Q13. Bakrie Telecom, which had the strongest and most focused CDMA product and marketing team, has remained independent but as the figure below shows even its operation remains chronically sub-scale, with ROIC reaching -6.1% in FY12. Even if Bakrie Telecom s new business units, namely broadband wireless (dongle) access and the proposed launch of EVDO set top boxes with a keyboard that plug into TVs to give internet access, result in a return to revenue growth we forecast ROIC will remain well below the cost of capital. We therefore deem further capex/investment by shareholders to be irrational. Figure 24: Bakrie Telecom s ROIC analysis Rp bn FY10 FY11 FY12 FY13 FY14 FY15 Revenue 2,765 2,591 2,361 2,268 2,436 2,572 Incremental revenue 23 (174) (230) (93) Capex 1,964 1, Capex to sales (%) 71.0% 63.4% 10.1% 20.0% 20.0% 20.0% ROIC (%) 1.4% -1.7% -6.1% -5.6% -5.5% 2.5% Data alone cannot create a viable business model Similarly, at the current predatory price points for voice and SMS services the progress of the two GSM new entrants, Hutch Indonesia (Not listed) or Hutch and Axis (Not listed) since 2008 has been very limited. We put this down to several factors: (1) Inferior network rollout Inferior network coverage and high interconnect rates have created barriers to entry Initially due to regional rather than nationwide allocation of spectrum CDMA operators did not offer full nationwide coverage in Indonesia. Furthermore, regulatory requirements were such that services could only be offered within one local dial-code area; external use required roaming to a different phone number. While customers were prepared to accept this limited mobility and functionality when CDMA price points were one sixth of prevailing GSM price points, when GSM price points declined to CDMA levels the business model was irrevocably damaged. New GSM entrants, Hutch and Axis, had (and still have) the problem of 1800 MHz spectrum, which, due to its much shorter propagation range, raised the number of BTS that had to be built to match the coverage of the 900 Mz network of the Big 3 GSM players, namely Telkomsel, XL and Indosat. This accentuated the perceived lack of coverage which has persisted from 2005 to (2) A high interconnect rate and on-net/off-net tariff differential The high interconnect charge on GSM-to-GSM voice calls encouraged the creation of onnet communities. Indonesia s regulator, the BRTI, last cut the core local cellular-cellular interconnect rate by only 3.8%, from the Rp261/minute (US$0.029) which was effective from 1 January 2008 to 31 December 2010, to Rp251/minute (US$0.028) effective 1 January As a result of this relatively high interconnect rate, operators have chosen to compete primarily on the basis of on-net prices (since outgoing on-net calls cannot result in cash outflows, while off-net calls can). This has led to an extreme outcome whereby for example a two minute on-net call on Simpati PeDe at midnight will cost the user Rp206, while an off-net call to a cellular competitor will cost Rp3,000, or over 14x more! Clearly this confers a strong advantage to operators who already have scale. If a user selects a new entrant and has to primarily make off-net calls to friends, family and colleagues, that user will be subject to dramatically higher tariffs compared with a user who chooses a popular network and therefore predominantly makes on-net calls. Thus, Indonesia Telecoms Sector 14

15 new entrants off-net tariffs are effectively competing with incumbent s on-net tariffs, and this explains their complete lack of traction in the marketplace. (3) Introduction of interconnection on SMS Until June 2012 there was no such interconnect charge for SMS sent from one network operator to another. Thus, sub-scale new entrants such as Hutch and Axis were effectively encouraged to offer large bundles of SMS to their customer bases. Again, by definition, the sub-scale new entrants have much smaller subscriber bases than the big 3 GSM operators, so large volumes of SMS were sent off-net by those new entrants to the big 3 GSM players. The big 3 had to absorb this incoming traffic (and build the required capacity to facilitate it) but until June 2012 received no economic benefit for doing so. The introduction of a Rp23/SMS charge on 1 June 2012 turned the economics of this behaviour on its head; new entrants are now economically discouraged from sending large volumes of SMS to other networks and instead have to focus on aggressive promotional activity towards on-net SMS activity, using capacity built on their own. Rapid reductions in interconnect rates from current levels would conversely lower the scale barrier faced by new entrant GSM players. However, following a reduction in voice quality and increase in congestion during the price war period, BRTI has focused on regulations to promote voice quality and lower capex costs (e.g. tower sharing) rather than driving down price points. The BRTI correctly believes that current off-net pricing effectively subsidises on-net pricing, resulting in blended average rates which are already quite low and so we do not expect a change in policy towards lower interconnect rates. A data centric strategy has a very limited window of opportunity Some concerns have been raised about the prospect of both the CDMA players and also the GSM new entrants being more successful, and having a greater impact on the Big 3 operators in the data realm. In particular the lack of a domestic interconnect rate on a per minute or per SMS basis is thought to remove a critical disadvantage of the smaller players; on the other hand, we believe the following constraints will still apply: Bundled voice and data plans mean that interconnect rates are still a barrier (1) Inferior network rollout Lack of nationwide coverage is less of an issue for stationary internet users and we are not surprised to see CDMA operators such as Bakrie Telecom aggressively promoting use of dongles to fill their underutilised RF (i.e. radio) network resources where available. Lack of a nationwide backhaul network, constructed by each of the Big 3 prior to 2005, will also lead both the CDMA players and GSM new entrants to suffer data-driven network congestion issues earlier, even in areas where the RF network has been rolled out. Scale also counts in purchase of international bandwidth (the necessary interconnection for supply of overseas data access). This will likely remain a critical disadvantage for the smaller players until more content is cached locally. (2) Bundling strategies The key problem faced by the smaller players in the smartphone space, however, is that the disadvantages faced under voice and SMS continue to cast a shadow into the data era. Voice and SMS interconnect rates would indeed be irrelevant if 100% of subscribers had smartphones and only used VOIP and OTT (Over The Top) services such as WhatsApp and Viber. However, the reality is that only 15.3% of subscribers in Indonesia have 3G smartphones and so the vast majority of communication is still over voice and SMS both services are actually still growing in revenue terms. As we have seen, the reaction of the incumbents has been to launch bundled plans, offering attractive on-net voice and SMS tariffs with nationwide coverage, as well as data services. Thus, scale and the advantage of large on-net communities continue to be relevant and will remain so even beyond 2015, when we expect Indonesian smartphone penetration to reach only 55.9%. Indonesia Telecoms Sector 15

16 (3) Predatory pricing Part II The move by the Big 3 GSM operators to lower the effective price per MB by 59-75% to average Emerging Asian levels is arguably sub-optimal for their own returns on capital; it reduces the expected total industry revenue growth potential from data. However, by lowering price points in this way, we believe that the Big 3 have limited the room for GSM new entrants to undercut on pricing yet grow sufficient scale. We previously showed that with 54.0%, 18.4% and 17.9% revenue market share, Telkomsel, Indosat and XL are on track to generate FY13 ROIC of 34.8%, 3.8% (due to inefficient historic investment) and 10.4%, respectively. On the basis of our current forecasts for data revenue growth and required investment, we expect the respective ROIC to rise by 39.5%, 9.7% and 14.8%, respectively, by FY16. We believe that this prospect is therefore worthy of investment. Data prices are now too low. However, we believe GSM new entrants, like the CDMA operators, cannot at these price points reach scale sufficient to generate reasonable returns on capital. We estimate the return on capital generated by a GSM 1800 MHz and 3G player with 3.0% revenue market share and cumulative network investment of Rp22 tn (US$2.3 bn) to be -5.9%. Even in the unlikely event that this share of revenues could double simply through increased network utilisation (i.e., without further capital investment) ROIC would be negative, and a long way below cost of capital. The logical/rational move by these players is to stop investing. Any influence these smaller players have had on setting industry price points (i.e., by encouraging the larger players to cut data pricing) has, we believe, already played out with the arrival of predatory pricing on data. We are therefore not surprised to learn that XL, once again acting as price leader for the Big 3 operators, switched from lowering tariffs to raising tariffs in February for reasonable returns on capital ever to be generated by players with under 10% market share Figure 25: ROIC analysis of sub-scale GSM players Rp bn 3.0% share 6.0% share 9.0% share Revenue 3,500 7,000 10,500 EBITDA margin 20.0% 28.0% 38.0% EBITDA 700 1,960 3,990 Depreciation* 2,444 2,444 2,444 EBIT margin -49.8% -6.9% 14.7% EBIT (1,744) (484) 1,546 Tax rate 25.0% 25.0% 25.0% NOPLAT (1,308) (363) 1,159 Cumulative capex ,000 22,000 22,000 Payback years - revenue Payback years - EBITDA ROIC -5.9% -1.7% 5.3% *Assuming depreciation over nine years Source: Credit Suisse estimates Indonesia Telecoms Sector 16

17 PT Telkom: Data growth outweighs share overhang Strong data growth in cellular We have previously highlighted the attractive backdrop of the Indonesian cellular industry (15.3% 3G smartphone penetration, no unlimited data plans, no handset subsidies). Importantly, market leader PT Telkom (through cellular subsidiary Telkomsel) is keen to reap the benefit of rising smartphone penetration, by continuing to avoid unlimited plans and subsidies, and staying above tariff-based competition by focusing on coverage and quality (superior backhaul as well as radio capacity). The latest Dance like Agnes Simpati data add-on promotions such as Rp20,000 for 500 MB of data and Rp60,000 for 2 GB, are resulting in Telkomsel enjoying an uplift of around 2.0x the average ARPU of Rp37,783 when customers upgrade to a 3G smartphone and consume data (on average only 550 MB/month). Since Telkomsel already has sufficient scale, and has already built around 40% nationwide 3G coverage, this represents an attractive additional revenue kicker. We expect double-digit revenue growth from Telkomsel Telkomsel is also actively looking at moving beyond a dumb pipe, with M-payment strategies and content such as SingTel s newsloop. With strong data growth, primarily still from access/usage, but assisted by improving local content and applications, we continue to believe that double-digit (10.4%) revenue growth is possible for Telkomsel in FY13. We expect EBITDA margins to decline by only 0.8 pp in FY13 as revenues outstrip some fixed costs, softening the impact of a higher data mix, and thus we anticipate 8.7% EBITDA growth. Capex is forecast at Rp10.7 tn (18% capex to sales), broadly flat YoY, as the 3G BTS addition rate remains circa 6,000 per annum. With depreciation charges set to grow more slowly than EBITDA we expect a further year of double digit earnings growth (11.1%) from Telkomsel in FY13. to drop through to doubledigit earnings growth The 1Q13 figures showed that this trajectory is very much intact, with Telkomsel revenue, EBITDA and net profit rising 13.3%, 10.8% and 22.5%. If anything we believe there could be risk to the upside. Figure 26: Telkomsel P&L summary (Rp bn) E 2014E 2015E 2016E Revenue 43,948 45,567 48,733 54,531 60,198 64,369 67,897 70,982 Growth 10.3% 3.7% 6.9% 11.9% 10.4% 6.9% 5.5% 4.5% EBITDA 27,331 26,598 27,549 30,564 33,229 35,146 36,800 38,330 Growth 13.6% -2.7% 3.6% 10.9% 8.7% 5.8% 4.7% 4.2% EBIT 18,790 17,181 17,533 21,132 23,240 25,156 26,330 27,534 Growth 11.9% -8.6% 2.0% 20.5% 10.0% 8.2% 4.7% 4.6% Capex (12,660) (8,197) (8,500) (11,700) (10,655) (10,943) (10,864) (11,357) Earnings 13,160 12,362 12,824 15,715 17,457 18,930 19,850 20,795 Growth 15.2% -6.1% 3.7% 22.5% 11.1% 8.4% 4.9% 4.8% and also in fixed line The fixed line business grew revenue by 0.4% YoY in FY12, following three consecutive years of declines, as fixed line data revenue more than offset traditional voice declines. The revival of fixed line growth prospects is a key positive that looks set to continue. While plans to connect 15 mn homes to fibre by 2015, paid for entirely by copper replaced during the rollout, have been watered down somewhat (the rollout will take until 2016 and Telkom s fixed line capex to sales ratio is now expected to remain at around 25%) even the more gradual rollout of fibre should facilitate mid-single-digit fixed line revenue growth The fixed line business has returned to growth Indonesia Telecoms Sector 17

18 as the broadband subscriber base rises from only 2.7 mn as at March 2013 (only 29.3% of current fixed lines in service) and multimedia offerings expand. Our forecast of 5.5% YoY revenue growth in FY13 represents the highest fixed line revenue growth since 2007 (7.0% YoY), when the Flexi CDMA fixed wireless business was growing rapidly (29.5% YoY). However, the key difference in our minds is that the fixed line broadband access revenues will likely prove sustainable, while since 2008 Flexi revenues have more than halved (down 55.6%) The fibre rollout will also support 1 mn Wi- Fi hotspots. Telkom will initially sell Wi-Fi access to Telkomsel for hand-off in busy downtown areas, but may over time sell access to other cellular operators as well. Figure 27: Fixed line revenue and EBITDA projections (Rp bn) E 2014E 2015E 2016E Revenue 23,578 23,062 22,520 22,612 23,865 24,987 25,539 26,015 Growth -2.2% -2.2% -2.4% 0.4% 5.5% 4.7% 2.2% 1.9% EBITDA 9,229 10,505 9,009 9,193 10,047 10,394 10,471 10,536 Growth -12.6% 13.8% -14.2% 2.0% 9.3% 3.5% 0.7% 0.6% EBITDA margin 39.1% 45.5% 40.0% 40.7% 42.1% 41.6% 41.0% 40.5% Telkom expects cost savings to arise from the network upgrade, from personnel through to the release of space in buildings, which previously housed now-outdated switching equipment. These opex savings are of course longer term in nature and will only crystallise as the network modernisation project unfolds. However, PT Telkom has already spent Rp7.3 tn on an Early Retirement Programme running from 2002 through February This has brought the employee numbers down under 19,000, from 32,000 as at December We expect some operational gearing over personnel costs (the fixed line division s largest cost) to be apparent in FY13, for three reasons. First, the previous ERP programmes clearly create a lower base for ongoing employee cost increases. Second, while there will be another ERP in FY13, the budget is for Rp500 bn for employees, down from the Rp699 bn spend in FY12. This YoY decline will in itself generate EBITDA growth. Third, there has also been a pick-up in the rate of natural attrition (i.e., employees reaching retirement age and dropping off the payroll without an ERP. Around 520 employees, a similar number to the number taking an ERP will exit the payroll this year through natural retirement, further bringing down the run rate of the fixed line division s largest cost. Thus, we forecast a decline in the personnel cost to sales ratio, from 33.2% in 2012 to 32.2% in and ERP costs are flattening off This, together with an expectation of continued improvement in general and administrative costs (as achieved in 2012), leads us to expect a low single-digit (3.0%) increase in fixed line operating costs in FY13, versus the achievement of a 0.7% decline in fixed line operating costs in FY12 (on lower general and administrative costs and lower operating and maintenance costs). This, coupled with our forecast of pick-up in revenue growth, leads us to expect a second consecutive year of improving margins from the 40.7% achieved in FY12 to 42.1% in FY13. This, in turn, is expected to result in 9.3% YoY EBITDA growth in the fixed line division, the highest YoY growth since the 13.8% YoY increase in 2010 (when no ERP was conducted). Growth in consolidated earnings, rising cash flows On a consolidated basis we are forecasting 9.0% service revenue growth and, given broadly stable margins, 8.9% EBITDA growth. Indonesia Telecoms Sector 18

19 With PT Telkom s strong cash flow resulting in the net debt position reaching zero as at December 2012, and a net cash position building from FY13 onwards, net interest costs are expected to decline, resulting in high operational gearing below the EBITDA line. Thus, earnings are expected to rise 14.7% YoY into FY13, and our EPS forecast is 3.7% ahead of consensus. Double-digit earnings growth again looks possible in FY13... Figure 28: PT Telkom P&L summary Rp bn E 2014E 2015E 2016E Revenue 67,526 68,629 71,253 77,143 84,063 89,356 93,436 96,997 Growth 5.6% 1.6% 3.8% 8.3% 9.0% 6.3% 4.6% 3.8% EBITDA 36,560 37,103 36,558 39,757 43,276 45,540 47,271 48,866 Growth 5.6% 1.5% -1.5% 8.8% 8.9% 5.2% 3.8% 3.4% EBITDA margin 54.1% 54.1% 51.3% 51.5% 51.5% 51.0% 50.6% 50.4% Capex 17,900 12,651 14,600 17,300 16,621 16,690 15,461 14,349 Capex/Sales 26.5% 18.4% 20.5% 22.4% 19.8% 18.7% 16.5% 14.8% Earnings 11,332 11,537 10,965 12,850 14,740 15,801 17,251 18,126 Growth 6.7% 1.8% -5.0% 17.2% 14.7% 7.2% 9.2% 5.1% At PT Telkom s AGM on Friday 19 April shareholders approved a 65% payout ratio (Rp436/share) for FY12. The dividend payout matches the FY11 payout ratio and the same terminology was used, i.e., that the standard payout of 55% was maintained and that a 10% special dividend was added as a sweetener for investors. leading to a higher dividend payout As we expect the payout ratio to remain at 65% in FY13, our earnings forecasts also imply a 14.7% YoY increase in the dividend payable and a yield of 4.1% in FY13E. Since the free cash flow yield is higher (6.0%) there is room for higher dividend payments, and we forecast a 75% payout ratio for FY14. Figure 29: Indonesia telecoms sector comparative multiples Current Target Upside P/E (x) EV/EBITDA (x) FCF Yield (%) Div Yield (%) Price (Rp) Price (Rp) (%) FY13E FY14E FY13E FY14E FY13E FY14E FY13E FY14E PT Telkom 11,900 13, Indosat 5,600 8, XL 5,150 7, Bakrie n.m. n.m Asia ex Japan Integrated Asia ex Japan Wireless A stock split, Treasury Stock sale, and a tower sale? A five-for-one stock split was also approved at the PT Telkom AGM in April. Of course, this will have no impact on the valuation of the company, but simply result in the quoted share price being divided by five. In theory this could make retail participation easier, but only if individual retail investors are wishing to purchase a very small number of shares. We are unsure as yet on the exact date of the bonus issue. As shown in Figure 30 below, there are currently 1.0 bn shares held in Treasury Stock after four consecutive share buyback programmes. At the current share price levels the Treasury Stock has a total value of Rp12.0 tn, equivalent to 5.3% of the total market cap of PT Telkom, and, given an average entry price of Rp7,980/share, PT Telkom has arguably created value of Rp4.0 tn, or 1.7% of the market cap, by buying its own shares at a 32.9% discount to the current share price. Treasury Stock sale creates a technical overhang Indonesia Telecoms Sector 19

20 Figure 30: Share buyback programmes Program # shares Average Price (Rp/share) Period Share BuyBack I 211,290,500 8,657 Dec Jun 2007 Share BuyBack II 215,000,000 9,160 Jun Dec 2008 Share BuyBack III 64,284,000 7,238 Jun Dec 2009 Share BuyBack IV 520,355,960 7,309 May Nov 2012 Source: Company data The allowed time limit for holding the 211 mn shares purchased in Share BuyBack I across expires in August The three options available are cancellation of the shares, use of the shares to create a convertible bond, and sale of the shares back to the market, and the options for how to deal with this tranche, together with the stock held under Share BuyBacks II through IV, were discussed at the AGM. While it may seem illogical, PT Telkom management did not believe that politicians would ever sanction cancellation of shares purchased by Telkom and held as an asset. Thus, the PT Telkom board sought and received approval for the sale of the Treasury Stock. We therefore expect the 211 mn shares from Buyback I to be sold in the market in August This does create a near-term technical overhang amounting to 1.1% of the share capital in issue (Rp2.5 tn, or US$259 mn at current share price levels), but listing rules require the shares to be sold at a price above the 90-day average preceding the transaction. Use of the shares to create a convertible is unlikely on this particular occasion, though we expect this approach to be taken over the next two to three years for shares purchased in Share Buybacks II through IV. Separately, PT Telkom continues to look at strategic options regarding its tower assets (15,000 are held by 65%-owned subsidiary Telkomsel, 3,500 are held by 100%-owned subsidiary Mitratel, and a further 1,500 are owned directly by PT Telkom). Decisions regarding a potential listing may finally be made by August 2013, though no further information was given in Telkom s April 2013 AGM. This could, however, provide offsetting positive news flow around the time of the Treasury Stock placement. but not an overly material one in size Options for tower assets remain open Indonesia Telecoms Sector 20

21 Indosat: Strong core recovery continues, despite earnings drag Revenue has recovered through improved execution Indosat s cellular revenue declined 1.8% QoQ into 1Q13 on seasonality, but grew 16.5% YoY. The seasonal QoQ decline was more shallow than the 5.1% QoQ decline recorded by PT Telkom subsidiary and market leader Telkomsel, and the 5.5% QoQ decline recorded by XL, and so 1Q13 represented the fourth consecutive quarter in which Indosat has outgrown both XL and the industry. Revenue market share of the broader mobile industry (i.e., of total industry revenues including CDMA-based players such as Flexi and Bakrie Telecom) has therefore expanded from 17.6% as at 1Q12 to 18.4% as at 1Q13. We have previously attributed this structural improvement to four factors. First, there have been management changes; Indosat s execution in the marketplace was clearly reenergised with the 2Q12 arrival of the new Chief Commercial Officer Erik Meijer (previously Telkomsel sales and marketing, then Bakrie Telecom CCO). Second, some particular network issues suffered in 4Q11 and 1Q12 in Jabotabek were resolved with the upgrade of base station equipment, and network quality noticeably improved, making Indosat less exposed to price-based competition from some of the smaller (sub-3% revenue share) players in the marketplace. Third, completion of converged billing and campaign management system in May/June 2012 allowed much faster execution of marketing tactics and much smarter marketing campaigns on a cluster-by-cluster and indeed customer-by-customer basis (to drive up revenues without dropping core nationwide price points), in much the same way that XL and Telkomsel had done in 2011 and 1H12. Fourth, Indosat has weeded out unlimited data plans and is therefore monetising data appropriately. The fact that Erik Meijer is set to move from Indosat to Garuda after only one year in his post, following a request from the SOE minister, is clearly a blow to the first critical improvement factor mentioned above. Indosat believes it will not completely derail the company s momentum, though a replacement CCO will need to be found. Certainly we would acknowledge that engineering a significant turnaround in momentum, from declining market share to increasing market share, as initially achieved by Erik, was a more challenging task than maintaining the existing positive momentum. Importantly, the other infrastructure constructed (improved network, improved billing and campaign management) remains in place to support Erik s successor. Indosat has been gaining market share due to both new personnel and improved processes Unfortunately the Chief Commercial Officer is about to change again but the processes are intact and improving 3G rollout is expected to be capital efficient. Indosat has noticeably lagged competitors Telkomsel and XL in the construction of 3G networks. As at December 2012 Indosat had only 4,596 3G Node B BTS but Telkomsel and XL had rolled out 15,433 and 13,142, respectively, across more than 29 large cities in Indonesia. 3G delay worth waiting for, and not damaging to date Figure 31: 2G and 3G BTS per operator 2G BTS G BTS 2012 Additions G BTS G BTS 2012 Additions 2012 Telkomsel 33,114 38,864 5,750 9,509 15,433 5,924 Indosat 15,816 17,334 1,518 3,437 4,596 1,159 XL 23,363 26,310 2,947 4,910 13,142 8,232 Source: Company data Benchmarking against the 2G BTS coverage (and taking into account that GHz cell coverage is lower, but that population density in key cities is of course higher), we estimate that Telkomsel and XL s 3G networks cover 40-50% of Indonesia s 242 mnstrong population. Indonesia Telecoms Sector 21

22 Of course, with 3G smartphone penetration having reached only 15.3% as at December 2012, Telkomsel and XL have arguably built quite aggressively ahead of actual demand. We note that Indosat has been able to grow market share from 2Q through 4Q despite having much weaker overall 3G coverage, due to the fact that the 3G smartphone S-Curve is in a relatively early stage. One reason for Indosat s delayed rollout was management s decision to apply for use of its ample 900 Hz spectrum resources (it has 2x 10 MHz of spectrum, as shown in the figure below) for 3G HSPA rollout. After almost a year of lobbying, permission was granted through Decree Number 504/KEP/M.KOMINFO/08/2012 concerning Cellular Mobile Network Licenses, on 31 st August The decree allows Indosat to roll out any 3rd Generation Partnership Project ("3GPP") standard, including 3G HSPA, over 900 MHz spectrum and 1800 Mhz spectrum, as well as the MHz range. Figure 32: Spectrum allocation in Indonesia (MHz) Allocation in MHz 900 MHz 1.8 GHz 2.1 GHz Telkomsel 7.5*2 22.5*2 15*2 Indosat 10*2 20*2 10*2 XL 7.5*2 7.5*2 15*2 Source: Company data The flexibility granted by this decree will allow Indosat to use its existing spectrum resources to meet data demand in Indonesia. At the very least, the ability to use 900 MHz resources for HSPA meant that Indosat did not need to bid for either of the recently auctioned 5 MHz pairs of GHz spectrum, which were sold to Telkomsel and XL for Rp500 bn and an annual fee of Rp250 bn over ten years. In fact, the savings to Indosat will be larger than this, as rollout of blanket /large area coverage using 900 MHz HSDPA will be far cheaper than utilising MHz spectrum. The radio propagation is far superior (up to twice the distance), and far fewer Base Transceiver Systems (BTSs) may be required. Of course this advantage is only realisable in lower-density areas, or in providing seamless 3G coverage in travel corridors which connect high-density coverage areas; we fully expect Indosat to rely on its 10 MHz of MHz spectrum in key city centre areas, as it has done to date (and as competitors will continue to do). Furthermore, Indosat has chosen to construct multiband antennae which can cope with 900 MHz, 1800 MHz and 2100 MHz frequencies within one BTS. This reduces energy consumption, space used within the tower area, and due to the cabling system, which is being used, wind resistance and weight on the towers. Figure 33: The 900 MHz multiband antenna Figure 34: allows much wider coverage for lower cost Source: Indosat Source: Indosat Due to this superior design, and the fact that Indosat will need to build far fewer BTS to achieve the same coverage, we expect the following implications: Indonesia Telecoms Sector 22