印尼水泥业：对水泥供需前景不担心.docx

Global ResearchIndonesia CementUBS Evidence Lab inside: Indo Infra Part 1No worries on cement demand-supply outlook?UBS Evidence Lab satellite imaging & demand model indicate higher utilisation We believe our EBITDA margin improvement thesis remains intact, despite lower-than- expected volume growth in 2019. We expect 5mt (5% of 2019 capacity) incremental domestic supply until 2022, as recent geospatial data indicates two cement plants in Java (Grobogan and Singa Merah) to come online. We expect domestic volume to recover to +2.3% YoY in 2020 (2019: -0.2% YoY) and +5% YoY in 2021, driven by improving credit cycle and stable inflation, which would translate to rising sector utilisation and healthy price competition. We believe the market is pricing in an EBITDA margin 130bp below our estimate for 2020E, and we see more upside for Semen Indonesia.Limited supply pressure from ongoing new plant constructionContinuing our initial reports in August 2018 assessing plant construction through satellite imaging, we use data from UBS Evidence Lab's geospatial team to monitor eight locations (six are new). We now expect total Indonesia domestic capacity to reach 105mtpa by 2022 (previously: 102mtpa) after including Hongshi/lmasco's plant in East Java. However, we think two new cement plants would have limited distribution scope given their locations and lack of port access. Meanwhile, given their lack of plant expansion, we believe Conch and Siam Cement are focusing on profitability rather than market share.Domestic cement demand could gradually improve from Q220Our demand forecast is based on a proprietary Indonesia cement consumption model (autoregressive order: ARI MAX) with three input variables unchanged from the last iteration (Indonesia crude price, inflation and M2 money supply). Our base scenario indicates demand could recover in Q419, cool off in Q120, then rise to >5% level by Q420. We attribute this to lower oil price assumption (high base in H119) and softer economic activity in 9M19 (M2 money supply is a two-quarter leading indicator).Reiterate our preference for Semen IndonesiaWe trim our price targets for Semen Indonesia (SMGR) and Indocement (INTP) by 4-8% due to more conservative 2020 volume forecasts. For INTP, our2020E EPS is similarto consensus; however, for SMGR our2020E EPS is 27% above consensus as we believe market is undervaluing synergy with recently acquired Solusi Bangun Indonesia (SMCB).9 December 2019EquitiesIndonesiaBuilding MaterialsBobSetiadiAnalyst +62-21-2554 7032Joshua Tanja, CFAAnalyst+62-21-2554 7030Figure 1: Valuations summaryCompany(Rp)Rating(US$m)2020E2020E202 IE2020E2020E2020EIm YTD12mSemen Indonesia12,425Buy5,26017.29.58.51224.412.7-1.88.04.6Indocement20,700Neutral5,43837.220.317.41972.28.40.2 12.23.2PriceMarket capPE (x) EV/EBITDA (x) EV/tonne (US$) FCFyield(%) ROE (%) Performance (%)Forecast changesOldNewOldNewOldNewOldNewOldNewSemen IndonesiaBuyBuy15,70015,000338340730721876853IndocementNeutralNeutral23,90022,000488444600552742662Above data as of 6 December 2019. Source: Reuters, UBS estimatesThis report has been prepared by PT UBS Sekuritas Indonesia. ANALYST CERTIFICATION AND REQUIRED DISCLOSURES BEGIN ON PAGE 23. UBS 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.Impact to our forecastsWe have added capacity from Semen Singa Merah (Hongshi/lmasco) into our calculations and now expect 5mt additional industry capacity in the next three years (1.6% CAGR in 2019-22E), with total industry capacity to reach 106mt by 2022E (2019E: 101mt). We think this would lift the industry utilisation rate to 71% by 2022E (2019E: 69%). We expect this to support ASP growth in the industry, and forecast ASP to rise at a 1.8% CAGR in 2019-22E. Competition-wise, we expect SMGR and INTP to own 71% of total industry capacity in 2022 (2019: 74%).Figure 20: Cumulative capacity (mt) vs utilisation (%)山山Reo 山 LNNO 山RLO 山。CXI寸0 山篦0 山后0 山 0CXI3 山6一.寸0 600 6LCXJ06 二O 830 800 8R0 8 二0 二占Figure 21: Cement capacity breakdown by company - SMGR and INTP to own 71% of capacity in 2022ESource: Indonesia Cement Association, UBS estimatesCumulative capacity Utilization (RHS)Source: Indonesia Cement Association, UBS estimatesFigure 22: Three-year growth for industry ASP, supply and demand90.90.49%12.5%(uss-tonne).oo 6o.8 7o o50.o.oS 64 3l.b%6.1%6.0%2%2.6%.7%/u /w /fe 9 9 9 2 2 64208642024 1111 - -HDXO AA01-0404-0707-1010-1313-1616-19E19-22 ECAGRCAGRCAGRCAGRCAGRCAGRCAGRASP(RHS) ' » Capacity DemandSource: Indonesia Cement Association, Company data, UBS estimatesUBS ResearchUBS ResearchIndonesia Cementreturn 个return 个PIVOTAL QUESTIONSQ: Will domestic cement volume growth continue to accelerate?UBS VIEWYes, but at a more gradual tempo. Based on our Indonesia cement consumption model, we expect domestic cement volume growth to improve to 2.3% YoYin 2020E (2019E: -0.2% YoY), starting in Q220. We expect this to further improve to 5.0% YoY in 2021E on the back of increased economic activity, steady oil prices, and stable inflation.EVIDENCEThe UBS Evidence Lab Data Science team examined 18 potential variables for predicting Indonesia's domestic cement demand growth and determined the three with the greatest predictive power: 1) Consumer Price Index; 2) oil price; and 3) M2 money supply. We fed historical data and UBS forecast variables into an ARIMAX model, which in turn indicated that, on a four-quarter moving average basis, Indonesia's domestic cement demand will bottom in Q220 and gradually improve to 5% YoY in Q421.WHAT'S PRICED IN?SMGR and INTP management expect 2020 industry volume growth of 2-4% YoY. We believe the market is expecting similar numbers for 2020.We expect a gradual recovery in domestic cement demandAfter starting the year with high optimism (SMGR and INTP were guiding 2019 industry volume growth target of 4% YoY; UBS-e: 2.1% YoY), Indonesia domestic cement growth contracted 1.1% YoY in 10M19. We attribute this to the presidential election and weak global growth weighing on Indonesia's economic activity, and we now forecast 2019E domestic volume growth of -0.2% YoY. However, we expect domestic cement demand to gradually improve to 2.3% YoY in 2020E and 5% in 2021E, driven by improving credit growth (lower interest rates and healthier company balance sheets), plus stable inflation level and a slight pullback in oil prices.Figure 23: Indonesia 4Q moving average for domestic cement growth12.0%10.0%8.0%6.0%4.0%2.0%0.0%-2.0%10.0%-4.0% -UJTzbcf WIZmcr 山 IZZO WTZTO 山。Z1/O wozmo 山。zzo 山。ZTO L1J6L 寸cr 6T8 6X0 6IIO 8百 8m noTzu mo 9WUJTzbcf WIZmcr 山 IZZO WTZTO 山。Z1/O wozmo 山。zzo 山。ZTO L1J6L 寸cr 6T8 6X0 6IIO 8百 8m noTzu mo 9WSource: Thomson Reuters, CEIC, UBS Evidence Lab Data Science teamRevisiting Indonesia cement demand model. To forecast domestic demand growth, we referred to work done by the UBS Evidence Lab Data Science team to develop a cement consumption model. This iteration of our Indonesia cemect consumption model uses the same three input variables (Indonesia crude price, inflation and M2 money supply) but we change the model estimate into four- quarter moving average YoY changes (previously quarterly YoY changes). We also update the modelling method from linear model (LM) to autoregressive integrated moving average with explanatory variable (ARIMAX) to resolve the issue of residual plots (differences between predicted and actual results). This resulted in better R2 of 89% (previously: 40%) and directional accuracy 76% (previously: 43%).Predicted YoY Actual YoYFigure 24: Indonesia cement demand growth - predictedFigure 25: Model coefficients for the Indonesia cementvs actualconsumption model25% 2.0%>0> uozdEnsuoo一仁 0131.5%1.0%0.5%0.0%-0.5%CPIOil PriceM2 from Bank Indonesia 2 Qtr Lead-1.0%Source: Thomson Reuters, CEIC, UBS Evidence Lab Data Science teamSource: Thomson Reuters, CEIC, UBS Evidence Lab Data Science teamSource: Thomson Reuters, CEIC, UBS Evidence Lab Data ScienceteamKey assumptions and scenarios. Of the three input variables, oil price remains the most indicative of Indonesia cement demand. For our base case (2020E: 2.3% YoY; 2021E: 5.0% YoY), we expect oil price to remain relatively stable in the next two years, and Indonesia inflation to be at 3-3.4% in the same period. An indicator of economic activity of Indonesian households and corporate spending, we forecast M2 money supply growth to recover from 7% in 2019E to 9.7% by 2021E. This would stem from an easing interest rate cycle and better balance sheet position, which would ultimately stimulate credit growth, in our view.Figure 27: Indonesia crude price 4Q moving average (US$/bbl)llZCMCOE 寸寸 Sin 99rz 8866OOL 5i- 5i- nr-iCJ CM 04lOOlCOlCOlCOlCOlCOlCOlCOlC*>lCOlOOOOOCOO0OOOO00OOOOOOBaseUpside DownsideLLNC|COCOp"xtgin994Z886600L V- T- 5- v- V- T- m- T- T V V- 51 T- t T- v CM CM C'J lCOlCOlCOlCOlCOt- CQlCOt-COt- COlCQl 0OOOOOOOOOOO0OOOOOOOOBaseU osideDownsideFigure 26: Indonesia quarterly inflation YoY (%)scFigure 28: Indonesia M2 money supply growth 4Q MA YoYBaseUpside DownsideLCXECSource: Thomson Reuters, CEIC, Bank Indonesia, UBS Evidence Lab Data Science teamFor our domestic cement demand upside scenario (2020E: 4.5% YoY; 2021E: 7.1% YoY), we assume the government retains current fuel and electricity prices, as well as lower food imports. Meanwhile, we assume the completion of a trade deal between the US and China would create higher economic activity across the globe thus increasing oil prices. Lastly, we assume Indonesia government would be able to attract foreign investment while at the same time stimulate the housing market through favourable fiscal and macro policies.In our downside scenario (2020E: -2.5% YoY; 2021E: -2.5% YoY), we assume that, to reduce the budget deficit, the government returns fuel and electricity pricing formula to market pricing. Meanwhile, we assume trade-war tensions reescalate and the US share of oil supply grows faster than expected, resulting in pressure on oil prices. We expect these events could hurt the investment appetite of Indonesia households and businesses, resulting in slower credit growth.-5%*448866 OOllCXICxICOCO 寸寸 LOLO994N886 寸。O O O O O v-i- i- r- i ii iiczbz寸z寸z寸z寸z寸e寸z寸z寸川寸可寸2寸e 000OOO00O00O00OO0OOO00O00O山RNO山尽0山寻。山 6LWDBase Downside UpsideSource: Thomson Reuters, CEIC, UBS Evidence Lab Data Science team, UBS estimatesFigure 30: Key assumptions in our domestic cement growth forecastsOil price (US$)CPI (%)M2 money supply (%)YoY cement growthOldNewOldNewOldNewOldNewBase201751.23.89.97.6%201867.53.26.94.7%20I9E61.161.73.23.17.37.00.1%-0.2%2020E63.560.34.03.28.89.05.1%2.3%202 IE63.73.19.75.0%Upside20I9E64.361.73.03.17.86.71.4%-0.2%2020E68.569.43.43.010.89.47.3%4.5%202 IE76.02.811.07.1%Downside20I9E59.561.73.33.16.76.7-0.4%-0.2%2020E57.051.34.94.46.06.51.8%-2.5%202 IE43.74.96.1-2.5%Source: Thomson Reuters, CEIC, Company data, UBS estimatesOn a quarterly YoY basis, we expect relatively flattish cement demand in the first three quarters of 2020E, before demand begins to increase in Q420E. This is mostly driven by a delay in the impact of improving economic activity (ie, M2 money supply growth), which we expect would kick-in from Q419.山RLO 山ON寸。 山 ozeo 山寻。 盘3 山 6LW3 6L8 6 Ro 6二0 830 8L8 8 Ro 8LL0 二寸。 Z.LCO0 二No Z.LLO 9LWD 9LCO0 9 Ro 9 二OActualOldNewSource: Company data, Indonesia Cement Association, UBS estimatesIn the quarter a year after the quarter in which the 2009/2014 presidential elections were held, domestic cement demand improved by 3.5%/3.8% YoY. In 2019, the presidential election was held in the same quarter as the Lebaran holiday, which further impacted volume growth. Taking the high base impact of H119 into account, we expect oil price to decline by 7% YoY in Q220E. As a result, we expect domestic cement volume to grow 1.2% YoY in Q220E (Q219:- 3.8% YoY).Figure 32: Quarterly YoY domestic cement growth in presidential election years2009 Election2014 election2019 electionElection Year1 year afterElection Year1 year afterElection Year1 year afterQI-5.8%17.7%3.6%0.5%0.0%0.0%*Q2-8.0%6.0%4.2%-1.1%-3.8%1.2%*Q3-0.8%3.0%2.2%5.7%-0.9%0.8%*Q417.7%1.0%3.0%9.6%2.8%*5.5%*Note: *1) Q419-Q420 numbers based on UBS forecasts; 2) Period highlighted orange/red indicate quarters when the presidential elections were held/cabinet were announced.Source: Company data, Indonesia Cement Association, UBS estimatesUBS Researchreturn 个Figure 33: The market is implying 130 bps lower EBITDA margin than our estimate24.4%2020E2021E UBS EBITDA margin Implied EBITDA margin by marketSource: Reuters, UBS estimates. Data as of 6 December 2019Reverse-engineering EBITDA margin from the current market priceWe believe our forecasts differ with the market and sell-side consensus at the EBITDA margin level rather than top-line. To calculate how the market is pricing in EBITDA margin, we reverse engineered the current market price. Our methodology is as follows: We calculate sector weighted average EV/EBITDA multiple using our DCF models for SMGR and INTP to come up with our base case of what we think is a fair valuation for the sector. Using oursectorweighted average EV/EBITDA multiples, wethen calculate thecurrent EV adjusted with the time value of money to determine the sector EBITDA implied by the market. Then we estimate implied EBITDA margin by dividing our implied sector EBITDA with our sector revenue forecast.We believe the current pricing indicates margin improvement for next year, but we believe the market is underestimating the impact of cost efficiency program as well as rising utilisation for both SMGR and INTP. The market is currently implying EBITDA margin of 22.1% in 2020 and 23.1% in 2021 (2019 UBSe: 19.1%), or 130 bps below our forecasts.Indonesia CementWHATS PRICED IN?Figure 34: Implied 2020E EBITDA margin from current market price: SMGR is 3.2% below our estimate and INTP is 1.4% above our estimate.24.6