DBS Group Holdings Ltd (D05) Earnings Call Transcript & Summary

Good morning, everyone. Thank you for joining us on the -- on our first half results briefing. I'd like to pass it over to Piyush, our CEO; and Sok Hui, our CFO.

Actually, I want to pass it on to the Q&A because there's not that much I want to say over and above what we said in the media section.

Thank you. We can go to Q&A, please. Operator, can we have the first question, please?

[Operator Instructions]

There are no questions.

There are no questions. Operator, do you have any questions?

First, we have Jayden from Macquarie.

I guess I'm just trying to square the credit quality outlook. Just a couple of questions only if I can, Piyush. For the full year, you're saying that provisions won't be any more than $500 million per se, but that suggests up to $400 million for the rest of the year, given the first half has been very benign. And your comments earlier sounded very upbeat on the outlook for credit quality. So is it possible that this guidance is just too high and that we actually need a far lower credit charge for the next 2 quarters? That's the first part. And second of all, I know in the last quarter, a lot of the release had to do with, at the time, sort of model adjustments, which took, I guess, macro variables. Can I confirm if that was a big factor this quarter or if there was any sort of discretion that came into the general release? And how much of the write-back had to do with the repayments versus upgrades to the credit outlook? I'm just sort of curious to some of the mechanics and what we should expect for the rest of the year.

Right. I'll take the first question, and Sok Hui, you can take the next 2. The $400 million is -- I used the word carefully, that I'm hedging my position saying $400 million. I think it could be a lot less because based on, actually, everything we're seeing. I'm not seeing a pickup. I'm not seeing provisions come through. It's just that the final impact of the delta on consumer portfolios in Indonesia, Taiwan, et cetera, are still a little uncertain. So you could see some pickup from that. Second is as the loans are coming off moratorium, so far, I'm not seeing a pickup, but it could be that it's very early, that as you go down next 2, 3 months, you might see a pickup in NPLs and therefore, provisions related to the loans coming off moratorium. But if I had to make a bet right now, based on what we have, I think there might be some upside to that number. So Sok Hui, you want to talk to the model thing as well as the repayments question?

Yes. So I understand your question to be around the general provisions. We said that there was a net repayment -- net write-back of $275 million. And a large part of that came from repayments and credit upgrades. So the way the model works, you should expect some reduction when cases move into NPL because that's how it's supposed to operate. You'll release general provision to go into a specific provision. Within that category, you would get improvements also when, in our methodology, you move from, say, [ weak ] to red or to amber. That means there's an improvement in the credit quality of the borrower, and that corresponds to typically counterparty risk rating upgrade as well, or you put a write-back from maturity or simply repayments by the borrowers when they refinance out. So for a number of reasons, these were the ones that contributed to the $275 million, of which the -- both the institutional banking and retail banking, both contributed the larger component from the nonretail base. The downgrades were largely sort of offset by -- the downgrades were largely offset, and we had repayments totaling -- resulting in ECL of about $75 million reduction, upgrades about $37 million reduction and some due to just reduction in maturity and within the Stage 2, improvement in the rating from [ weak ] to red or red to amber. So these were the main drivers for the sort of $275 million write-back.

Let me actually give you a different perspective on that. Our total general provisions are about $4 billion and change. Out of that, about $1.5 billion are what we call modeled overlays. The modeled overlays reflect a downside scenario that we take, and we build provisions in case something can happen. So about $2.5 billion of that reflects what GP we keep because it gets built up through our model. The modern GP, of course, is a function of ECL and the probability of default, loss given default and the risk rating of our portfolio. So when the portfolio risk rating improves, either because a customer has upgrades or the customer's actual outstanding fall because they pay us back and so on, then that obviously results in an improvement on that model component, right? And so that's a part of what reflects in the GP.

Yes. Many years ago, I actually was scoring corporates for their credit quality. So a lot of it has to do with the outlook. If I take the comments carefully, it sounds that the actual write-back was $75 million and the improvement in the outlook for these customers is $200 million. Am I reading that correctly?

Regarding the outlook, it might be also the outstanding. So if a customer has got a $1 billion loan and he sort of pays it down, I don't count it in the payment if it was not fully paid. It's just that the outstanding for that customer in this current category reduces.

Or it could be that the maturity of the same loans actually shortened. So our methodology takes into account the maturity profile and the Stage 2.

Yes. Sorry, just to be really clear, do you have the breakdown between the 3? Like how much of it was repayment? How much was it because of maturity shortening? And how much was it because the opinion was that the customers' prospects has gotten better? That's what I really want to understand.

Yes. That's why -- okay. Of the $275 million, about $80 million will be -- $80-odd million will be due to retail. And retail has come off because a lot of the general provisions that were set aside last year was much higher given the outlook. And given they have unsecured loans, they have either actually sort of provisioned -- been moved into specific provisions or the loan base has declined because we are actually not doing -- some are not going from this portfolio. So the retail ECL has declined by about $80 million. So the balance is really from the institutional side. And on the institutional side, I mentioned repayment would contribute about $75 million, upgrades about $37 million. And then the others are all due to maturity and other sort of improvements from the modeling point of view.

Next, we have Melissa from Goldman Sachs.

Maybe just a little bit back in terms of the provisions. Did I hear right that you have still $1 billion as the overlay, management overlay, over the provision? And does it mean that you're still keeping it and perhaps you might consider maybe releasing it next year if things are really better? Then secondly, just moving on to the margins. Can you just give a little bit more of a split in terms of the 4 bps compression, what it's actually made up of? And perhaps in the second half, you mentioned that there will be pressure. Can we just get some indication of how margins might trend in the second half of this year?

So the overlay that we have is actually about -- Piyush mentioned it's about $1.5 billion, right? So that is the amount that we set aside, partly last year and partly built up in previous periods to cater for kind of stress events. So we are unlikely to write back the entire amount of overlay because in our modeling, there'll always be something that would cause us concern. And prior to COVID, we had also overlays to cater for, for example, the geopolitical tensions, et cetera. So a good gauge is the -- I said we have $800 million above MAS minimum requirement. So that's one sort of guidepost you can use potentially if things really improve. How much, we can actually prepare to release that. Melissa, does that answer your question?

Yes. So does it -- so then can I just check back in terms of when the -- what is really a normalized credit cost? I think [ exiting H2 ], what can we expect on normalized credit cost? So then I think kind of that looked like 800 offset that. What possibly can do things really, really better?

So we expect that this will be released in tandem with, I guess, the outlook for the economy, in particular, also whether travel restrictions are lifted. So we are going to take a conservative stance, and it's not like we're going to release an entire amount. So it will be gradual and will be in line with what we are seeing on the horizon.

Melissa, what we've said in the past is that about 22, 25 basis points should be our normalized cost of credit in the past. I think our normalized cost of credit will [ forget ] all of this noise from building up a provision or release of provision. It is likely to be sub-20 basis points. I think that reflects the quality of our portfolio and our credit processes as we are today. The cost of credit SPs in the second quarter was 14 basis points, as you can see. Now as Sok Hui pointed out, we do think that we've been very conservative in building up our buffers. And so if things really get much better, we do have the capacity to release some of those. Some, I think, will automatically reverse because of what the earlier question from Macquarie was. The models that push us to improve -- as the customers' credit quality improve, the exposure come down. Some of it will unwind because of that. This balance, $1.5 billion, which we've kept as an overlay, Sok Hui pointed out, we're not going to release all of that. We always like to keep some in our hip pocket because you never know what might come along the road in excess of models. But we have the opportunity to do that as well. So the way I would work your model is to look at standard cost of credit in that 18, 20 basis point range and then adjust it up and down depending on what's happening to provisions and reserves.

Right. And on the NIM, please?

Yes. On the NIM, I think you're referring to second quarter versus first quarter, where we saw a 4 basis point decline. And that's really a function of the surplus deposits that we have, which are then deployed to mainly Central Bank, [ sort of basically ] Central Bank. So I think you heard the response earlier. So these are 0% weighted. They do not pick up capital. They can be accretive to our income as well as bottom line. But it does mean that we have some volatility due to the interbank interest rates. So this quarter, it's down a bit. So you see some drag from the surplus funds that are placed with Central Bank.

Melissa, actually, I think that's the bulk of things. There are other -- 2 other reasons. One, LIBOR and HIBOR came off this quarter compared to the first quarter by 2, 3 basis points. So that has a little bit of an impact. And then obviously, the flow-through into the rest of our fixed resident portfolio, there is still some large resident impact of that. But our outlook for it, in the beginning we guided for rates -- NIM of between 1.45% and 1.50% for the year. I think our NIM for the year will be at the low end of that, 1.45%, which means that you will still see a little bit of a drag of a couple of 2, 3 basis points through the back end of the year.

Next on the line, we have Terence Chua from Phillip Securities.

I have one question. What is the percentage of loans coming off loans moratorium? And my second question is which sectors are they coming from?

Coming off loans moratorium or the loans in moratorium?

Loans in moratorium and loans coming off moratorium.

Well, it's quite difficult. Okay. Let me answer. We had $5 billion in the mortgage moratorium in Singapore. We had another $5 billion in the SME moratorium in Singapore. That's $10 billion. And at its peak, we had about $7 billion or $8 billion in moratorium in Hong Kong. So $17 billion or $18 billion out of our total loan book of about $400 billion. So that was about 4% of our loans were at its peak in moratorium, if you will. Right now, we're left with $0.5 billion in mortgages, $400 million in SME and about $1.2 billion, $1.3 billion in Hong Kong. So about $2.2 billion, $2.3 billion on $400 billion. So we're left with about 0.6% or 0.7% of our loans currently in moratorium. On top of that, we have the ESG loans which is 90% government protected, but that's another $5 billion, a lot of ESG loans. The risk on that is about $500 million for us. So we want to throw that into the pot, and I would say everything put together is about under 1%, 0.8% or something maybe.

Right. So just to make sure I get you right. So about 0.8% of loans are currently under moratorium. Is that correct?

No. It's only 0.6% under moratorium. The 0.8% is if I add the ESG risk loans. So I think about 0.6% are under moratorium.

Next on the line, we have Harsh Modi from JPMorgan.

Three questions. First is if we start thinking about '22, do you think -- is it fair to say that provisions should be in line with '21 or lower given the extent of GP and whatever is a central base case right now for asset quality?

So Harsh, it's tough for me to say. When I'm doing my budgeting for next year, I'm not planning to budget at this year's low levels. I am planning to budget at the through cycle level that we've had. And the reason for that is that, at the margin, the net impact of the government fiscal and monetary stimulus responses finishing in many of the countries is still relatively unknown. So -- and that's the only thing. I can't tell you what happens because right now, in every country, the government is still providing some form of support in one way, shape or form. And that's one uncertainty. The other uncertainty is the taper. So if you figure that liquidity starts draining out and because the taper -- and you start seeing some pickup in rates, that's another uncertainty. So I won't rush to say it will be as good as this year. It could be. You could be right. But it's just that there are some unknowns yet about next year.

Got it. Okay. That's fair. Second one on capital, again, kind of similar question, Piyush, 13%. There are multiple paths to it, returning more capital, reinvesting the capital as you have been doing. Most likely, it's some combination thereof, but how should we think about time line towards that 13% and the path towards 13%?

It might happen tomorrow if we wind up doing all the Citi deals. So Harsh, as I said, that's a cute way saying that part of the answer is a function of do we do any M&A or not. And if we wind up not doing any M&A, the Citi deals are -- don't come through or we don't want to do them, which is equally likely, then we'll go back and take a more aggressive view on what we want to do with capital and capital management.

So going with that logic by, let's say, next 6 to max 12 months, we should see some conclusion on the 3 markets you talked about: Indo, India, Taiwan. So then is it fair to say that by the time we're thinking about final dividend are at max interim dividend next year, we should have a much better clarity on payout -- on the dividend per share and payout numbers?

Well, Harsh, our view on the dividend question, [ we are acknowledging ] that we will continue to increase it steadily over time. And that policy still remains. So if you are making more money, you could obviously pay out more. The M&A thing, Citi is the immediate possibility. But of course, over the next couple of quarters, I still want to keep my eyes and powder open if there are other of these bolt-on deals that come on -- become available because of the macro environment. And those could be banking or nonbanking deals. But yes, I think over the next 12 months or so, it's reasonable to say that we should have a good sense of if anything else comes along our way in the back of this crisis. I remember the 2 deals we've done basically were an outcome of the crisis or because some -- we got an opportunity. And so I'm keeping some powder dry if something else might come along. If it doesn't and things become normal, then we go back and take a fresh look.

Right. And just on that, what kind of internal hurdle rates you are solving for, for any of the deals? Like you get a broad guidance on IRR or payback? Anything on those sorts.

I said before, I want to make sure that the deals become accretive in about 3 years or less, which means they've got to return cost of capital upwards of that.

Okay. And final one, Lakshmi Vilas, how is the progress there? And any milestones that you would want to suggest that could be reached or that you are aiming for, let's say, by end of the year for that deal?

So Lakshmi Vilas is good news and bad news. The good news is that things are proceeding on plan in terms of integration and in terms of credit quality. So we've actually been getting repayments on some of the stuff we have provided for earlier. Integration is smooth. Deposits are up. Cost of funding is down 60 basis points. Gold loans are up for the year. So all of that is proceeding well. The bad news is that because of the India pandemic and COVID situation, we are about 2, 3 months behind in terms of actually being able to sweat the franchise. And so it's deliberate. I don't [indiscernible] -- the franchise is a large consumer, SME, et cetera. And I'm reluctant to push the pedal on that just yet, until I'm more confident and secure about what the overall macroeconomic situation in India is. So that part is a little bit slower. In terms of, therefore, the thing to watch out for by year-end, I think year-end might be a little early because of that. We're still -- we're beginning to dial up a bit, but we've been very careful about dialing up. So I think the chances are we probably have better line of sight to what pickup we should expect next year rather than in the next quarter or 2.

Okay. So just a final question on that, Piyush, that, let's say -- by, let's say, end of next year, do you get to a point where you become comfortable enough to inject more capital, where now it is genuine growth capital rather than precautionary capital, so to say? Like where -- by when do you sense a steady state there?

No. No. So I won't say steady state. If you look at our current view, I mean, we think that over the next 5 years, we will actually put a lot more capital into India because the expanded franchise we have gives us very significant growth opportunities. And so we have a very ambitious growth agenda for India on the back of this acquisition, which will take capital for at least the next 2, 3, 4 years before it starts leveling off.

Right. No, exactly, exactly. So that's what I'm trying to get, that in terms of time line, when do you get enough confidence to go back to the Board and say, "All right. This is the amount we are -- we have enough of a track record now to commit this much of capital"? And that's the state I'm trying to figure out. When do you reach that?

We've already done the work. So our plans are in place. We actually haven't taken it to the Board yet. We will do that in the next couple of months to get Board blessing. But informally, we run it through this thing. I mean it is an ambitious plan, and the Board is fully supportive of it. I mean the Board basically took line of sight to that when they approved doing the deal.

Next is Nicholas Teh from Credit Suisse.

Just have a couple of questions. Firstly, I wanted to ask on Hong Kong, China, specifically. I guess with the rising cases in China and the shutdowns there and also the delays in border reopening in Hong Kong, have you started to see any impact on the accounts there? The second question I had is on the deposit side. Deposit growth has been very strong. And any sense or any thoughts on how to think about whether these deposits could be stickier than what we initially anticipated, and hence, starting to think about deploying some of those deposits into higher-yielding assets rather than with MAS? Those are my questions.

On the first one, the short answer is no. I mean the increased cases in China and the border is not creating any incremental downside or headwinds there. You've got to remember that in Hong Kong, we've been through [ stresses ] now for 3 years, well before the pandemic. It was because of the China-U.S. tensions and the technology supply chain. So these issues have been with us for years and we've been managing very tightly to that. So I think we've actually corralled ourselves quite well. I don't see -- anticipate any further downside. In China, this incremental increase in cases, back to my earlier comment in the media discussion as well. So if you look at the translation of that into the macroeconomic indicators, this is sectoral. And if you, like us, don't have a consumer book in China and don't have an SME book in China, it's unlikely it will make any material impact to our portfolio and our kind of business that we do. On the deposit, I think the answer is some. So I fully expect that some of this massive deposit increase will run off, partly when the taper happens, partly when interest rates start picking up because there's too much money sloshing around. Now how long that takes is anybody's guess. It's not clear to me that our Central Banks will be able to eliminate this liquidity anytime soon. If you think about the GFC, right, the Fed [ increased ] the balance sheet to about $4.5 billion. They wanted to bring it down to sub-3, 2.5, I remember. They got to 3.7, 3.8 and then turned turtle, so it went back up. Now this trend, they put out $12 trillion. So how long it takes to squeeze the liquidity out of the system is anybody's guess, but I don't see it disappearing soon. And nevertheless, at some stage, if you assume that liquidity -- that surplus liquidity starts going, then how much -- the second part, how much of that is sticky with us? I do think that there's a fair element of that because a lot of the liquidity that we've got is CASA. It's not FD. It's CASA and it's operating accounts liquidity. Our CASA ratio is 73%. All of that is strictly the operating balances that we're getting. And a lot of that has to do with the digitization and the API connectivity and the engine and supply chain integration and all the work we've done. So if I had to hazard a guess, I would say 30%, 40% of the surplus liquidity will stick. But again, this is a guess more than anything else. Now how do you use that money? Of course, you can put the money to work. But on the other hand, you also got to have a view on the risk. Who do you want to give the money to? And it's part of the earlier question from Harsh. I do think eventually, we can put the money to work with growing the SME and the consumer franchise. But we've got to be thoughtful about the time and when you want to start doing that. So for the time being, we've really been continuing to grow our business in the large corporate space. I've been very careful about the risk we want to take in the more risky segments of the market. We will get there when the economy stabilize some more.

Okay. Got it. Can I just slip one last one in and ask what your stake in the carbon exchange is?

Our stake in the carbon exchange is -- I'm trying to remember. We're originally 25. I think we might be a tad bit below that now.

Next on the line, we have Aakash from UBS.

Can you hear me all right?

Yes.

Yes. Great. My first question is on the revenue opportunity, like the $200 million number that you talked about earlier. I'm just wondering, like what is the expense growth associated with this group of businesses? What would be the cost-to-income ratio for this group of businesses? Do you have any idea on that?

Yes. To unbundle it, actually expense associated with everything other than Lakshmi Vilas Bank is not material. The -- on the Lakshmi Vilas Bank, expense is material. So right now, the cost-to-income ratio of the bank is close to 100% -- that's not 100%, actually, it's lower. Including the provisions and credit, it's about projected breakeven.

Breakeven. It's close to breakeven, yes.

Okay. So LVB which is, mainly, that we would be thinking about from an expense perspective, not the other businesses?

LVB because of the existing expense. Actually, the expenses are going to be reducing because the LVB expenses come, to a large extent, from franchise. We're already looking at rationalizing the franchise. We need a lot of it, but not all of it. Also, head count. When we got LVB, we started with 4,000 people. We're already down to 3,300, 3,400. We're already down 500, 600 people as part of our integration and rationalization. So we will see improvement in the cost-to-income ratio for LVB as well. But right now, the cost-to-income ratio is still high.

Okay. Got it. So if I can just also ask you about the impact this has on your long-term ROE target. Like you said in the past, 13% to 13.5% is where you think the business should be when rates recover to the pre-pandemic level. So does this new growth initiative have an impact on that target? Or is it still broadly similar?

I've given this indication some years ago, that as we continue to try and find alternative avenues of growth, structurally, the ROE of the company will keep improving. But it takes time for that structural improvement to trickle through the system. If I have to look back over the last 5, 10 years, without a doubt, the fact that wealth management today is a $3.5 billion income business for us or cash management makes us $2 billion, all of that improves the structural ROE. And that's why our ROE has gone up from 8.5% to 13-odd percent. A lot of that is the structural shift in the nature. So all of these activities will continue to help that. And -- but in a given year, we are lucky to see 0.1%, 0.2% from these activities. It's collectively over a period of time, we can look back and say, okay, I've got 0.5% or something up from these activities that we've done.

Okay. Understood. And then just on a related note, we've talked about this 40% CIR in a steady state, again, something that we discussed in 2017. What's the progress towards that? And how are you thinking about that number now?

Well, we continue to work at this. Some of that, of course, is a function of income. And the 40% is the biggest -- the thing on that is interest rates. So the interest rate impact on the income line between last year and this year is almost $3 billion, $2.8 billion. And it becomes much harder to get a 40% ratio when your denominator got suddenly impacted by that part. As you know, from the income, we clawed back most of the 3 from the other things we've done. But let's say, the income is still down from that level, so we need to see some benefit from the interest rate to improve the cost-to-income ratio as well.

Yes. So the best sort of indicators to look at is what we disclosed in our annual report were the cost-to-income ratio for the digital sort of segment that we track in consumer banking and SME in Singapore and Hong Kong. Excluding the sort of benefit from rates, you actually see the cost-to-income ratio is actually improving year-on-year or at least stable in the period where rates are [ going low ].

Okay. Understood. And then the last one I had is on the crypto business. So what's the progress on that? And this latest crypto license that was granted by the MAS, how critical is that for your business? Is that a positive in general for the ecosystem, do you think?

Let me answer your second question. Frankly, I don't know the answer. I've just asked my people this morning to try and understand why the crypto license is any different from the crypto license we have or if it is different at all because the FT reported that this is the first crypto license. And I know that the 3, 4 exchanges before us, and we certainly have a license to do everything. So I'm not entirely clear what the difference is. I'm going to find out. So I don't know the answer short. On the first one, I'm actually quite happy. So we've now got to a tad under 400 customers, investors on the exchange. We did about $170 million, $180 million of trading volume in the quarter. We have about $130 million, $140 million of assets under custody in that business. And as you know, we've been selective. We're not going mass market because just the total noise around this is a challenge. My target is to get to about 1,000 customers by this year-end. Now these customers obviously are active and trade well. One of the things that we need to do, which we're going to do by the next month or 2, is make it a 24/7 exchange. We started off with an exchange which operated in the Asian time zone. A large part of this activity actually happens outside the Asian time zone. So once we do that, I expect the volume of activity to pick up a lot. But I'm quite pleased. It's good going quite well.

[Operator Instructions] Next, we have Robert from Citi Research.

I've got some bits and pieces questions, but I'll start with my main question. Just thinking about the math around capital. So let's just start with the 13% CET1 ratio. You've got a return on risk-weighted assets, I think, something like 1.9%, I think. And maybe a sustainable RWA growth might be a high single digit or mid- to high single digit. So with those sort of parameters, what would be the optimal or neutral payout ratio if you were starting at a 13% CET1 rather than the current CET1? I'm just trying to get a sense of the math.