中國(guó)教育行業(yè)行業(yè)報(bào)告：什 么比學(xué)習(xí)更好

1、Table of Contents HYPERLINK l _bookmark0 ExecutiveSummary3 HYPERLINK l _bookmark1 Why Choose If You CanHaveBoth?3 HYPERLINK l _bookmark3 Whats So Great AboutThis Business?5 HYPERLINK l _bookmark4 Secularly growing demand, with bigconsolidationpotential5 HYPERLINK l _bookmark5 Highly stable,highlyvis

2、ible8 HYPERLINK l _bookmark6 Cash is beautiful, and theyhaveplenty9 HYPERLINK l _bookmark8 Is Online the Futureof Education?14 HYPERLINK l _bookmark9 How big is onlinemarkettoday?14 HYPERLINK l _bookmark10 Before we move on15 HYPERLINK l _bookmark11 Online vs. offline: How do theystackup?19 HYPERLIN

3、K l _bookmark13 Wherestheopportunity?23 HYPERLINK l _bookmark14 Who will be thewinners?24 HYPERLINK l _bookmark15 Food for thought: Life-time value ofastudent28 HYPERLINK l _bookmark16 EDUversusTAL32 HYPERLINK l _bookmark17 ValuationAnalysis38 HYPERLINK l _bookmark18 InvestmentRisks45 HYPERLINK l _b

4、ookmark19 Appendix: Whos Who inChina AST?47Companies54 HYPERLINK l _TOC_250002 NewOrientalEducation55TALEducationGroup65 HYPERLINK l _TOC_250001 76 HYPERLINK l _TOC_250000 Koolearn93Executive SummaryWhy Choose If You Can Have Both?We see a lot to like about the after-school tutoring industry in Chin

5、a HYPERLINK l _bookmark2 1. Demand is non-cyclical, growth shows secular trends, earnings are stable & visible, and the cash-flow profile is stellar, in our view. In particular, we believe the Big Two (EDU & TAL) are best-positioned to succeed now and into the future, with a long runway of market sh

6、are opportunity, backed by well thought-out strategy and best-in-class executions. All this might be well-known, but it still makes a very compelling story however we slice it (see P5-13 for our birds eye analysis of industry dynamics).Granted, these stocks arent cheap, but we are more sanguine, tak

7、ing the view that high potential stocks are rarely cheap. We believe investors should be willing to pay up to ride the compounding engine from this secular growth, which in turn will shave off multiples fast and generate multi-year returns. We believe current prices will have seemed a great entry po

8、int in hindsight a view that is supported by hindsight today on their share price performances over a year, two, or three years ago. Detailed valuation analysis (e.g., a reverse DCF, The Rule of 40, or stub value; P38-44) suggests that valuation concerns are overdone, and that these stocks are surpr

9、isingly affordable when viewed in the light of visibility and quality ofearnings.More often than not, investors also ask, “Then which camp are you in? EDU or TAL?” We respond: “Why not have both if we can?” We think this is particularly true at this juncture, as the two companies seemingly diverged

10、on strategic direction, trying to seize opportunity via different paths: EDU fortifying its offline and focusing on profit, while TAL has taken a leap into online and pursuing growth (See P32-37 for our thoughts on EDU vs. TAL).Our in-depth analysis on online vs. offline ASTs (such as unit-economics

11、, serviceable addressable market, product efficacy, etc.; P14-31) suggests online will surely enjoy disproportionate growth, but offline can take the lions shares of China AST for longer than many think. We identify two key winners EDU for offline and TAL for online who can build a strong moat in re

12、spective areas, in turn growing secularly for many years.What are the risks? Besides the usual (e.g., policy risk, risks on corporate/listing structure, FX, etc.), any meaningful slowdown in growth or decay in earnings visibility could drive fast de-rating of valuations. That said, neither seems lik

13、ely at this point, especially for the quality names like EDU/TAL, in our view.We do have stock preferences given investment horizon (Dec-2020 price targets), though it not necessarily reflect our long-term on the companies. Our pecking-order is EDU (OW) TAL (OW) (N) Koolearn (see individual company

14、pages for detailed discussion on eachstock).1 Our discussion of industry throughout this report refers to the K-12 after-school tutoring, i.e., non-formal private companies that provide teaching of core subjects (e.g., English, math, Chinese, etc.) for Kindergarten Grade 12 students. This doesnt inc

15、lude overseas consulting and other test-prep businesses (such as TOEFL, GMAT, CET etc.), which account for about 32%/11% of EDU/TALs revenues.New Oriental (EDU US): OW, with a PT of $143 (+26% potential upside). EDU is our top-pick, which we see as in the early innings of a substantial profit upcycl

16、e, for which the stock looks affordable. We comfortably forecast EDUs OP to triple in three years; key driver will be margin, as a higher mix of matured centers, rising utilization, and alleviating offline competition will push up OPM to 17% by FY22E (vs. 12% in FY19) on JPMe. Some investors seem co

17、ncerned about its (very) long term as EDU has seemingly given up on online for now. Not us we say EDU can and will build a strong moat in offline AST, will have the lions share of AST and grow secularly for longer than manythink.TAL Education (TAL US): OW, with a PT of $47 (+18% potential upside). W

18、e do think TAL offers the most definable long-term outlook, fueled by super- normal growth in online (which we estimate to print 80% CAGR and account for 40% of revenue in three years), and its best-in-class execution and proven track record give us lots of comfort. That said, we, as a sell-side, ca

19、nt help but think about near-term earnings and valuations. Given transitory pressure from online, we expect TALs margin can start to recover only from FY23E, until which time its profit momentum will inevitably trail EDUs. This, plus its valuation premium over EDU (50% on headline P/E, or 20%+ on st

20、ub P/E ex- online), lead us to place it as our 2nd preferred stock fornow.GSX (GSX US): N, with a PT of $16 (+13% potential upside). GSX is probably the most debatable yet interesting name in the sector, we think. On one hand, the numbers thus far look amazing, growing top-line at 500%+ while printi

21、ng the highest OPM in online education thanks to its star-teacher strategy and viral marketing, enabling better scalability of teachers and higher ROI on marketing. On the other hand, however, the sustainability of strategy is highly debatable, given severe concentration risk (nearly half of sales c

22、ome from Top 10 teachers) and yet-to-be-proven ability to cultivate another troop of star teachers. Were not saying its impossible, but it seems a challenging task. Given limited history, we dont have confidence to make a call and rate the stock as Neutral better safe than sorry, we thought. For sho

23、rt-term investors, we see GSX as a good trading stock into near-term results, where we expect sustained and above- guidancemomentum.Koolearn (1797 HK): UW, with a PT of HK$11.5 (-25% potential downside). Simply put, we struggle to find a path it to make good profits in the foreseeable future. We app

24、reciate niche-market potential of its unique small- group product (DFUB) that can cater to relatively affluent households in (very) low-tier cities, with superior quality than other available options; but the product is hardly scalable and penetration is likely slow, as per our analysis, leaving que

25、stions on its profitability. Meanwhile, its large-group offering is unlikely to gain strong momentum given late-comer disadvantage and lack of competitive edge. All-in, we expect Koolearn can turn profitable only by FY2023E, with less- than-stellar top-line growth for online pure-play; thus we rate

26、the stockUW.Table 2: China AST: Valuation comparison (prices as of 16 October 2019)CompanyJPMPT Rating (LC)Mkt cap 3m (US$ b) (US$Stock perf (%) P/E (x)*PEG (x)EV/EBITDA (x)*EV/Sales (x)EPS growth*3-mo YTD* 1-yearCY20E CY21E CY22E19-21ECY20E CY21ECY20E CY21ECY20E CY21E CY22EEDU US New OrientalOW 143

27、.0 113.718.0125 18% 107% 77%TAL US OW47.040.023.7103 10% 49% 65%GSXUS GSXN16.014.13.315.5 40% 34%n.a.1797 HK KoolearnUW11.515.01.83.0 46% 47%n.a.27.142.044.0n.m.20.0 15.41.030.4 22.01.529.8 19.61.2 68.3n.m.17.728.830.412.020.718.31133.14.036%39%159%36% 30%38% 39%48% 52%Note:Profitsarenon-GAAPbased,e

28、xcludingshare-basedcomps.GSXandKoolearnsYTDshareperformancesarecalculatedfromlistingprices. Source: Company data, J.P. Morgan estimatesWhats So Great About This Business?Secularly growing demand, with big consolidation potentialFigure 1: China AST: K12 market size (based on revenue)50845637041529332

29、8207233261228203255318285354393466426Rmb B5084563704152933282072332612282032553182853543934664266005004003002001000201120122013201420152016201720182019EOfflineOnlineY/Y growth(RHS)14%12%10%8%6%4%2%0%Source: Frost & Sullivan, J.P. Morgan estimatesChinas K-12 AST market is huge and is only going to ge

30、t bigger, in our view. Structural tailwinds that helped the industry to double in size since 2013 (+12% CAGR to RMB508B in 2019E, according to Frost & Sullivan) remain intact, namely rising income, cultural obsession to get into top schools, and strong propensity to spend money on education. Simply

31、put, demand for AST will remain intact due to deeply-rooted cultural aspects, as long as the competitive and test-based nature of Chinese education system remains in place.Frankly, the industry analysis may not even seem necessary, as the market has been growing at an almost boringly steady +12-13%

32、p.a. with no big fluctuation (except 2018, which was impacted by one-off regulatory changes). This was led by gradually rising penetration and prices (Figure 2-3), while the number of K-12 students remained steady; we dont see why or how the trend will change materially, considering the penetration

33、is still well below some other Asian countries (such as Korea or Japan, where penetrations are north of 50%; Figure 4). Structural, steady, and non-cyclical growth should remain intact for many years, in our view.Figure 2: K-12 AST revenue growth breakdown (offline only)+12%+12%+12%+12%3%3%3%3%+11%2

34、%+11%+9%3%+8%3%3%9%9%9%9%9%8%+12%+12%+12%+12%3%3%3%3%+11%2%+11%+9%3%+8%3%3%9%9%9%9%9%8%5%6%12%10%8%6%4%2%20122013201420152016201720180%2012201320142015201620172018Figure 3: K-12 AST penetration in China (offline only)26.0%26.3%26.4%26.6%26.8%25.5%24.1%595051535456576026.0%26.3%26.4%26.6%26.8%25.5%24

35、.1%59505153545657585654525048462019E442019E201220132014201520162017201828.0%27.0%26.0%25.0%24.0%23.0%22.0%OfflineASPY/YStudent enrollmentY/YStudents enrollments(mnpeople)Penetration (%,RHS)Source: Frost & Sullivan, J.P.MorganestimatesSource: NBS, Frost & SullivanFigure 4: K12 AST penetration (based

36、onstudentenrollment)Figure 5: China Education CompulsoryEducation58%37%13%22%58%37%13%22%KindergartenPrimaryMiddle(Junior)Middle (High)Source: Frost & Sullivan, J.P. Morgan estimates. Note: China penetration rate is based on studentenrollmentdatain2018;JapanpenetrationisbasedonNIERresearch2015;Korea

37、 penetration is based on KOSIS research2015.Source: TAL, J.P. MorganThis should already sound attractive for many investors. But the real beauty of the industry, in our view, lies in its huge potential for consolidation.According to various industry figures, there are around half a million AST insti

38、tutions across China, over 90% of which generate less than RMB10m in annual revenue, and the Big Two companies have less than 6% market shares (or 67% even including No 3). The fragmentation is even more severe in lower-tier cities, where the Big Players have faced two stumbling blocks lower income

39、levels (= affordability issue) and difficulties of sourcing quality teachers locally (= access to key resources).Figure 6: China consumer: concentration ratio of the top 3 (2018) 63.7%55.8%51.5%18.6%14.5%9.5%6.3%60%60%15%70% 63.7%55.8%51.5%18.6%14.5%9.5%6.3%60%60%15%Figure 7: China internet: concent

40、ration ratio of the top 3 (2018)Home applianceOnline videoE-commerceOTAHome applianceOnline videoE-commerceOTAOffline ASTOffline AST90%90%90%85%BrewerySportswearLiquorSeasoningSupermarketOnline adsOnline ASTSource: Company data, NBS, Euromonitor, Frost & Sullivan, J.P. Morgan estimates. Note: Liquor

41、 is based on 2017 dataBrewerySportswearLiquorSeasoningSupermarketOnline adsOnline ASTOnline gamingMobile paymentSource: Company data, Frost & Sullivan, J.P. Morgan estimates. Online AST includes Yuanfudao, which reportedly generated RMB1.5B revenues in 2018.Online gamingMobile paymentNonetheless, th

42、e industry consolidation has been in the making (Figure 8), as these issues have been gradually addressed with improving income at low-tier cities and technology development (such as via online or dual-teacher model; See P15-18 for details).And the catalyst came.Since last August, the government has

43、 stepped up oversight on AST with tougher regulations (Tables 22-23, P46), to raise the quality of AST institutions. Local-level implementation has not yet kicked-in for many provinces; but when new rules are strictly enforced, the sheer costs of compliance will substantially raise the “barriers to

44、survival”, likely forcing many smaller players out of the market, in our view.Moreover, Beijings initiative to reform Gaokao (college entrance exam) into a nationally standardized/uniform exam will be a tailwind for the national players, as itd remove the core competitive edge of smaller local playe

45、rs (i.e., more localized curriculum catering to nearby schools). We thus see industry consolidation as an inevitable and accelerating trend, and the Big Two should garner a big chunk of this.All of these suggest tremendous runway for the Big Two to grow structurally by expanding shares in a secularl

46、y growing industry. We forecast the Big Twos combined market shares to double to 12% by 2022, in turn implying +35% revenue CAGR for the two (Figure 9).Figure 8: K-12 AST market share of the Big Two (including online)12.4%10.7%8.9%7.3%5.8%5.0%5.9%6.9%2.0%2.4%0.9%1.1% 1.3%1.1%3.2%1.6%1.6%4.3%2.3%2.0%

47、3.1%2.6%4.0%12.4%10.7%8.9%7.3%5.8%5.0%5.9%6.9%2.0%2.4%0.9%1.1% 1.3%1.1%3.2%1.6%1.6%4.3%2.3%2.0%3.1%2.6%4.0%3.3%4.0%4.8%5.5%10%5%201420152016201720182019E2020E2021E2022E0%201420152016201720182019E2020E2021E2022EFigure 9: EDU/TAL: Revenue and Y/Y Growth+32% +24%+29%+35% +27%+27%+51%+25%+38%+15%+42%+21

48、%+65%+31%+65%10,000+32% +24%+29%+35% +27%+27%+51%+25%+38%+15%+42%+21%+65%+31%+65%8,0006,0004,0002,0000EDUTALEDU: OfflineEDU: OnlineTAL:OfflineTAL:OnlineSource: Company data, Frost & Sullivan; J.P. Morgan estimatesSource: Company data, J.P. Morgan estimatesTable 3: K-12 AST: Industry revenue and our

49、forecasts(RMB bn)201420152016201720182019E 2020E 2021E2022EOffline K-12 AST285318354393426466510552594YoY12%12%11%11%8%9%9%8%8%EDU8.211.716.422.329.537.6Market share1.1%1.3%1.6%2.1%2.8%3.5%4.4%5.3%6.3%TAL8.611.814.918.222.026.0Market share0.9%1.1%1.6%2.2%2.8%3.2%3.6%4.0%4.4%Online K-12 AST8.010.815.

50、521.929.742.360.383.5114.9YoY31%35%44%41%36%43%42%38%37%Online penetration2.7%3.3%4.2%5.3%6.5%8.3% 10.6% 13.1%16.2%EDU (Koolearn)-0.00.40.71.2Market share0.2%0.3%0.4%0.6%0.7%0.9%1.0%TAL10.015.722.8Market share1.1%1.4%1.9%3.9%8.6%13.2% 16.6% 18.7%19.9%Industry revenues293328370415456508570636709YoY12

51、%12%13%12%10%12%12%11%11%EDU8.311.916.622.730.338.8Market share1.1%1.3%1.6%2.0%2.6%3.3%4.0%4.8%5.5%TAL9.414.420.528.237.748.8Market share0.9%1.1%1.6%2.3%3.1%4.0%5.0%5.9%6.9%Source: Company data, Frost & Sullivan, J.P. Morgan estimatesHighly stable, highly visibleThe AST business has a simple cost st

52、ructure, with most items being fairly steady and predictable. This, along with non-cyclical demand and strong pricing power, makes their earnings highly visible, which is a boon for long-term investors, in our view.OFFLINE: A matured learning center (3+ years of operation) generates GPM 50-55%, with

53、 only three major cost items i.e., teacher compensation (20-25%of sales), rents (12-13%) and material & others (around 10%). SG&A varies per scale operation, but major players can print OPM of over 15% for a matured center, after netting Sales & Marketing (10-15% of sales), R&D (5%, including conten

54、ts development) and General & Admin (15-25%, as per the size of HQ) expenses. Most of these are fairly stable and transparent, which, coupled with strong pricing power, leads to superior earnings visibility for offlineASTs.ONLINE: In terms of GPM, online courses can generate much higher margin than

55、offline (especially for large-class format), ranging over 70-80% for major players. Despite much lower unit price (about half of offline), online superior margin thanks to its massive scalability (where one teacher can accommodate thousands of students) and absence of rental. Teacher compensation ty

56、pically accounts for 15-20% of sales, while other CoGS (e.g., broadband, materials costs) is a meagre 5% ofsales.But the biggest swing factor is actually the level of Sales & Marketing (incl. traffic acquisition), which hovers at a substantial 40-50% of revenues for most companies. This reflects the

57、 nascent nature of online AST, where most players focus on expanding user-base vs. making money (and the fact that most companies are unlisted does not help either), hence the trend will change when the market matures (more on this in P28-30). Nonetheless this makes it very difficult to assess the t

58、rue earnings power of online business.Figure 10: Offline AST Cost structureSource: J.P. MorganWe understand most online players are barely making profits (except GSX; more on this in the company note), hence we lay out a simplified PnL below based on break-even OPM.Figure 11: Online AST Cost structu

59、re (large-group dual-teacher)Source: J.P. MorganCash is beautiful, and they have plentyAST companies churn out cash like most others cant.Figure 12: EDU & TAL: Historical cash conversion cycle (days)-200-250-300-350FY19-400FY19EDU TALSource: Company reports.For starters, the business itself has a ne

60、gative working capital structure, as well as negative cash conversion cycle with upfront fee collection. This allows ASTs to enjoy stellar cash conversion ratios on both operating and net levels. For example, EDU/TALs operating CFs have averaged at 170% of EBITDA, and FCFs have also been well above