Singapore Banks - Maybank Kim Eng 2020-01-07: Data Is King


Singapore Banks - Data Is King

Access to data will determine winners

  • In a media release, MAS claims to have received 21 digital banking applications – 7 for digital full bank (DFB) and 14 for digital wholesale bank (DWB). Hong Kong received 33 applications when they opened up for digital banking in 3Q18. The applicants are largely made up of consortiums with members including e-commerce firms, technology and telecommunications companies, FinTechs and financial institutions, according to MAS.
  • The publicly known applicants are:
    1. Grab and Singtel.
    2. Razer Youth Bank – a consortium with Sheng Siong Holdings, FWD Group, LinkSure Global Holdings, Insignia Ventures & Carro.
    3. Beyond – a consortium with V3 Group, EZ-Link, Far East Organisation, Singapore Business Federation, MSIG & Heliconia.
    4. A consortium with iFast Group, Yillion Group & Hande Group.
    5. Ant Financial.
    6. ByteDance.
    7. A consortium with Sheng Ye Capital, PhillipCapital and Advance.AI.
    8. A consortium with AMTD, Xiaomi Corp, SP Group & Funding Societies.
  • All applicants have claimed they would focus on previously ignored customer segments in Singapore, namely the underbanked, falling in to four primary categories:
    • Low income individuals and families
    • Early income millennials
    • Startups
    • Micro-SMEs
  • Some of these applicants have stated regional ambitions with establishing in Singapore as a gateway to other ASEAN economies. Most are planning to leverage on existing product and service capabilities of consortium partners to develop cross-selling opportunities through these platforms – especially for wealth management and insurance products.

Up to 5 licenses available

  • As a recap, the MAS digital banking framework sets out provisions for the following:
    • Two digital full bank licences that can take deposits from the public, while not having any physical branch presence. The initial minimum capital requirement of SGD15m will be raised to SGD1.5bn over time.
    • Three digital wholesale bank licences where deposits will be sourced from business banking accounts of SMEs and corporates. Public deposits are allowed only for fixed deposits over SGD250,000. The minimum capital requirement of SGD100m will be raised to SGD1.5bn over time.
  • While the starting capital requirements are low (vis-à-vis full bank requirements of a minimum SGD1.5b), the MAS will adopt a sandbox approach with restrictions on the size of deposits (maximum of SGD50m) and product features (plain vanilla lending) in the first 1-2 years. These restrictions will be progressively eased as the banks ramp up.
  • The banks will be subject to the same prudential regulatory oversight as all nt players under the framework.

Access to data will determine success

  • Success in winning one of the 2 retail licenses or 3 wholesale licenses and, more importantly, ultimate operational success may likely be determined by the depth, quality and access to data the consortium will have. Data will be the key driver that can help lower transaction costs and manage risks effectively in customer segments that traditional banks have historically shunned as unprofitable.

Legacy infrastructure costs and limited returns has led traditional banks to de-prioritize low yielding segments

  • We believe the reason incumbent banks have not focused on the underbanked in Singapore is that credit underwriting, compliance, risk management costs do not justify returns.
  • Allocating credit scoring, KYC procedures or risk management resources to a large volume of low yielding accounts is unlikely to be returns positive – especially when trying to maximize returns on risk weighted assets.
  • Besides, increasing strategic focus on large, high quality corporates, cross-border capabilities and leveraging on wealth management have enabled the sector to drive ROAs up 15bps between 2011-2019E.

Automation, lack of legacy infrastructure provides digital banks a path to profitability in low yielding segments

  • Digital banks are uniquely geared towards solving this problem as they are unburdened from historical infrastructure costs. They can achieve this through automating onboarding, KYC, underwriting, monitoring and risk management using big data, AI analytics, machine learning etc., on their virtual platforms.
  • China’s internet giants led MyBank and WeBank have shown how processing costs for previously unbanked customers or low yield customers can be lowered.
  • According to Reuters, MyBank’s cost of approving a SME loan is around CNY2 compared to CNY2,000 for a traditional bank. Since launching, MyBank has been able to extend credit to over 7 million customers, using their big data and AI driven algorithms. According to Management, 80% of MyBank’s clients are first time borrowers and 90% of customers are from lower-tier cities in China.
  • WeBank leverages on data generated from their sister company, WeChat, to gain insights on user behaviour and finances. This has enabled the bank to manage their credit risks effectively and deliver a NPL ratio of just 0.64% compared to the national average of 1.89% (according to the PBOC).

Domestic market impact limited

  • We combine MAS’ digital banking guidelines, their existing regulations of qualified full banks (QFB) and scaling trajectories of digital banks in the UK (which was one of the earliest markets to liberalise and encourage digital banking challengers in the banking industry) to run scenarios on the impact of digital banking in Singapore. Please see Singapore Banks: Not so Digiciting, 10 July 2019 for a more detailed discussion.
  • For our retail digital bank modelling purposes, we take the examples of Atom Bank and Monzo Bank. Both English banks kicked off in 2015 and have seen rapid growth in their user base to around 2 million each.
  • According to Cruchbase, Monzo has raised around SGD550m of funding from investors including Y Combinator and Stripe. Atom Bank has raised around SGD630m from investors including BBVA.
  • For our wholesale digital bank model, we take OakNorth (not listed) as an example. OakNorth launched in 2015 to serve SME clients using AI driven credit analysis and underwriting. The company has raised around SGD1.3bn from investors including Softbank and Singapore’s GIC and EDBI.
  • Using a starting point of SGD15m of equity and a low asset-to-equity ratio (see by Atom Bank and Monzo when they started), we expect only a small loan-to-deposit ratio of 18%. We expect loan-to-deposit ratios to rise rapidly to levels seen on average by other QFBs in Singapore while we also expect gearing to rise to 15x by Year 3 – this is the upper limit of gearing we have seen in domestic QFBs.
  • We estimate that the overall loan book in Year 3 can increase by 10x y-o-y. For Atom Bank, the loan book increased 12x y-o-y in 2018, while for Monzo this was 7x y-o-y.
  • Assuming both retail bank licensees follow a similar trajectory, by Year 3, they will account for SGD2bn of loans - this is 0.8% of Singapore SGD system consumer loans as of end-November 2019.

Synthetic virtual bank -Wholesale

  • With a starting point of SGD100m of capital, we assume lower gearing levels (maximum 6.5x in Year3) relative to the retail operations. Given the focus on wholesale banking, we expect a bigger proportion of funding not to come from deposits, but from the interbank market. When we look at the wholesale banking operations of the domestic banks, we see similar funding structures with a typical loan-to-deposit ratio of 120%.
  • Assuming all 3 wholesale banking licensees expand similarly, by the end of Year 3, these will account for SGD6.2bn of loans. This is equivalent to 1.5% of Singapore SGD system business loan as of end-November 2019.

Limited impact in the Medium Term

  • As our analysis shows, even with UK style aggressive growth from inception, on a combined basis the digital challengers will only be able to command around 1.2% market share of total system loans. This is on the benign assumption that all incumbents – which includes 33 local, QFB and full bank licensees and 99 wholesale banking licensees - sit still in the face of this competition.

Asset quality, growth more relevant in near term

  • We believe near term drivers for the sector should be asset quality, earnings growth and dividend yields.

Stable asset quality

  • In terms of asset quality, we believe tougher macro conditions may have an adverse impact. However, increased focus on higher quality corporates as well as retail wealth management segments should help balance out some of these risks, in our view. High levels of provisioning and regulatory capital also provide a significant buffer.

Potential NIM upside?

  • In their final FOMC meeting for the year on 10-11 December 2019, the US Fed indicated a pause in lowering the Fed funds rate in 2020E. This is different from Street (and MKE) expectations of further cuts. We believe such a pause may potentially provide upside surprise to tightening NIM expectations for the sector as SIBOR will likely level off.
  • Our sensitivity analysis shows that a 1bps increase in NIMs should result in a 0.7%- 0.8% rise in sector PBT. UOB should benefit the most with a PBT increase of 0.8%.

Strong, visible dividend yields

  • The Singapore banks offer the highest defensible 2020E dividend yields in SE Asia. In this context, UOB and DBS are our top picks given strong, cautious balance sheets and regional growth.
  • See attached PDF report for complete analysis.

Thilan Wickramasinghe Maybank Kim Eng Research | https://www.maybank-ke.com.sg/ 2020-01-07
SGX Stock Analyst Report BUY MAINTAIN BUY 30.500 SAME 30.500