Good morning, everybody. I really appreciate all of you coming here. For some, like Nick, he says he stays close to here, so it is a short distance. Some of you had to schlep from the other end of the town. We wanted to do it in this facility because it gives us the opportunity to use ours. This is our innovation hub. So, it gives us the opportunity to use our demo rooms, et cetera, and showcase some of the things that we want to talk about in the course of the day. I also appreciate that many of you spent most of last week doing investor visits with some of our competitors, and a few of you actually stayed on over the weekend. So, thank you for that.
So my job here, I'm gonna spend the next 30, 40 minutes. I'm going to do a bit of an executive summary of the things we're talking about. But also I'm going to use this to tee up the agenda for the day, the various speakers. And basically, I'm going to focus on what you should expect to get and take away from each of the presentations that are going to follow through the course of the day. But before I get into that, I thought it might be useful to actually start from here, which is, you know, our investment thesis. Why we think you should, as investors, be interested in owning DBS. And I'll start with the extreme left.
You know, when 2009, 2010, when we first started, or I first joined the company, we spent some time trying to think about what could we create, which is a differentiated investment proposition for investors. And we decided that essentially, there's a real opportunity for a bank to offer a different way to play Asia. And the different way was effectively this: One, we would not focus on a subregion. We would not be just an ASEAN bank or a North Asia bank, and that's what most banks were focused on at that time. We would try to play across the three major subregions. We'd do a North Asia, Southeast Asia, but also subcontinent. So we'd connect the major flows across the region.
We also decided that while we would build a network of connectivity, we would go deep in the big markets, so we would not actually try to do multiple countries. Our general belief, and still continues to be, that if you're inch deep, mile wide, it's very hard to build scale or relevance. So, our premise was that we took the four biggest countries in Asia, ex-Korea, Japan, so China, India, Indonesia, Taiwan, and tried to go deep in those countries. We could actually differentiate ourselves from a lot of the global banks. If you could do this, this would give us two or three big advantages. One, we're different from the globals, because in the big markets, we'd go deeper. We'd be different from the locals because we would be able to do the connectivity trade, the trade investment flows around the region.
Anchored in Singapore, Hong Kong, that's about 60%-70% of the business. That would give a stable anchor in the two big financial hubs, both triple A countries. And the presence in China, India, Indonesia would allow us to gain diversification. And frankly, that would have been useful. If you look at China, in the first half of the last decade, the banks were booming. The last 2 years, the banks have got hammered, and the story's been the reverse for other countries. So, the notion was, because emerging markets go through cycles, and sometimes the cycles are not aligned, this would allow somebody to get a diversified access to Asia.
So, I sometimes get asked, "Why would we not just buy banks in each country?" Well, the reason for that is that you get diversification when you go through a focused regional play, as opposed to making bets in individual markets. So that was the thesis, basically. In the 2013, 2014 period, when we got this epiphany about leveraging technology and digital, we figured that we could really leverage technology to add a differentiated, competitive advantage that will supplement that, Asia. It's not a pivot but add on to the Asia positioning.
And that was really this, that if you focus hard and try to learn from what tech companies do, understood the model of tech companies, how they acquire customers, how they transact without paper, how they do things instantly, how they use data, you could wind up creating a financial outperformance. So that was the overlay on top of the Asia piece. And then more recently, in the last three, four years, it's been quite evident to a lot of people, including to us, that the sustainability agenda, apart from being good for the planet, is going to be a massive trillion-dollar income opportunity. And therefore, if you can get onto this early, learn how to figure the sustainability opportunities, solutions, et cetera, you could wind up distinguishing yourself. And Asia is going to be a big part of the sustainability play.
So the banks which are able to do this well and do this early should be able to add incremental value from that. So that's basically the notion that if you invest in DBS, it gives you diversified access to Asia, it gives you deep access to the big growth countries, gives you a technology play, which gives you some lift, and then hopefully can leverage the sustainability agenda. So what I'm going to do, the bulk of today is down the middle piece. You know, so that's why it's called Digital Transformation 2.0. It's really around why we think our technology and what we do gives us a distinction.
But I'm going to double-click on the left and the right, right up front for a couple of minutes, because obviously some questions around Asia and, so on and so forth. So, if we start with this, the diversified access to Asia's growth, and this question comes up often, I mean, is this still valid? Is Asia still the place to be, and is the investment proposition still, valid? So, I thought three points are interesting and useful to make. One is that, you know, people ask, "Is still the Asian century or not the Asian century?" And you've heard enough about this in the last week from our competitors. I'm on the same camp, that I think the structural, drivers of Asia's long-term growth are intact. A slow Asia still grows circa 5%.
You know, China this year will do 4.5%-5%. India will do 5%-6%. Indonesia, irrespective of every year, does ±5%. Thailand, does 5%. So I think Asia's secular growth in the 5% range is still very, very much intact. The export engine sometimes slow down, but consumption demand is picking up. And this notion of an integrating Asia, trade and capital flows in Asia, are still pretty much intact. Asia's consumption today, I saw the 2020 data, it's about 0.96 for every $1 that you see in the US, which is not bad. And if you look at the fast-forward projections, in the next decade, Asian consumption is going to outstrip US consumption.
So the mega trends that drive Asia as a structural area of opportunity and promise, I don't think those are taken away. The second big thing is, you know, the large Asian markets. We still love the view that our premise that if you go and build, in the big China, India, Indonesia, Taiwan, that opportunity is still, intact. If you look at the growth rates of the banks in India, or the banks in Indonesia, those are solid. The Chinese banks have been slow, but certainly in pockets, we think that growth is there, it's not going to disappear. So, one of the things that we've done, as an example, through COVID, we took the opportunity to actually double down on these big markets.
So, we did the LVB acquisition in India, and that's given us a large-scale footprint. We're going to talk about that. I have Surojit talking later about the India opportunity. But we've now got the platform and the scale to really grow in India. We bought Citi's business in Taiwan, and that makes us by far the largest foreign bank in the country. And while it won't have the same returns profiles as in India or Indonesia, we think we will generate decent return and scale out of the Taiwan market. We built a position in Shenzhen Rural Commercial Bank. It's a minority position, but we know we have the opportunity to increase that stake over the next two, three, four years, and we're going to use that to continue to anchor Greater Bay Area.
So again, not dissimilar to many of our competitors, but we will have a platform to be able to anchor and grow that at scale. So, this notion that the big Asian markets still offer significant upside, and the cards that we have been able to play over the last three, four years allow us to be able to leverage that, I think is an important point to keep in the back of your mind. And finally, the geopolitical question. So, what happens if there's a complete drift, there's an unbundling, decoupling, et cetera? Now, again, my base case, which is what we as a company are baselining everything on, is that a complete unbundling is very hard. You know, China is just the biggest trade partner for 68 countries.
It's too deeply integrated into global supply chains, into global consumption patterns. It's the biggest market in many ways. So, unlike the Russia situation of Cold War, where you could draw an Iron Wall, we think that's very hard to do. You'll find pockets, you know, semiconductors, you might find an activity which people try to bifurcate, but fundamentally not easy to bifurcate. And actually, we looked at the G7 announcements in the last couple of days. You know, we don't want autarky or we just want resilience and diversification. And Biden's comments last evening about how he's willing to see rapprochement and warming up in ties and so on. I think you'll see noise. It'll keep going up and down, but we don't see an unbundling entirely.
We do think that you'll have to be nimble because it'll create tensions, and you will have a multipolar world, and people will form alliances and partners for different issues at different points in time. But ironically, this is helpful for Singapore, and it's helpful for banks like ours. So, we are beneficiaries of this thing for alliance for diversification alliance. So, the China Plus One, I don't see anybody pulling investment out of China, but I do see people making a marginal investment in somewhere else place, Thailand for autos, Penang for electronics, some in Indonesia, some in India. We're seeing a lot of that. Most of it, 90% plus, from what I can see, maybe 100, is still in the region.
So, the friendshoring, nearshoring, other than semiconductor, I'm not seeing too much of it going back to the U.S., Mexico. Maybe some will happen. But the fact that it's in the region helps us as a country and helps DBS. And we've seen that in the flows, both the corporate flows, so companies, Chinese companies, American companies, Western companies, still look at Singapore as a place to domicile for the regional treasury centers, for the trade centers, trading hubs. And that's true for private flows as well. And again, we've all seen in the last year the spate of private flows that have come into the country from all directions. So, like I said, in an interesting kind of way, I think we are beneficiaries of some of the tensions.
So net-net, I'm not going to do more of this, but, like I said, you heard enough about this last week, but I'm basically in the same camp. I think the Asia story exists. It is consistent. I think the way we are trying to play Asia is still a smart way to play it, and I think actually we will be beneficiaries of some of these events in the short, medium term. Someone, a quick double-click on the sustainability agenda. It's very early, so I have to be clear about that. But if you look at the McKinsey report, right, they did one recently, they said effectively that the sustainability-linked opportunity is about $100 billion a year of the banking revenue pools for this decade, of it, Asia is about $30 billion.
And, I think that's about right. We sort of did back of the envelope. We can see those, opportunities as well. So what we've done in the last couple of years is that we've actually been ahead of the game in terms of defining these transition pathways for the big sectors and industries aligned to all the global standards, et cetera. But we've been able to figure out how we get from a high carbon-intensive sector and make that sector less carbon-intensive. So how you go from brown to light brown, to light green, to green. And we've created a lot of solutions to help in that process. So how do you get people to improve? So we believe that the capacity and capability we have to help in this transition process should allow us to put some runs on the board.
As a case in point, so as you know, one of the things that the world is looking at right now is the JETP partnerships. So, is it okay to invest in coal if you can accelerate the shutdown of the coal, investment you make? And the South Africans started it. There's some discussions going on in that regard within Indonesia, Vietnam, et cetera. Now, we are advisors to INA, the Sovereign Wealth Fund in Indonesia, to think about this. That, you know, how do you actually work to exiting some of the assets earlier than you would? And we can do that because we build capabilities to be able to create solution sets of that sort. Our sustainability-linked revenues have been moving quite sharply. So, we made about a small $75 million two years ago. That doubled in 2021.
That doubled again in 2022, so we made about $300 million last year. We have total originations of north of $60 billion. We did about $25 billion in fixed income. We have a book size of north of $30 billion. So I think the revenue pools are there, and our ability to be able to track those revenue pools are there. But I'd be the first to say, you know, over the next eight, ten years, what happens is not entirely clear. I think we're positioned well. So let me get to the heart of what we really want to talk about today, which is this financial performance driven by digital transformation.
You know, before I dive into that in the agenda, I thought it would be appropriate to address the technology issues we've had in recent times. Our board and actually all of us as management, we take these issues very seriously. And so the way we've dealt with it is we've got a special board committee which is running an investigation into entire technology processes. We've got external independent advisors who are helping to review the entire technology process and what we have. And we will be submitting a report to MAS when the entire investigation process is complete. You know, whatever comes out of that, if there are gaps or there are areas of improvement, I mean, we're committed to investing what it takes to make those improvements and to address those gaps.
I know a lot of you would like to get more details about this at this point in time, but we, I'm very happy to discuss this at the appropriate time. That will only be when the review is completed, when we've been able to get the end-to-end pieces, and it would be inappropriate to try and discuss it piecemeal at this point in time, and so we're not going to do that. As and when the report gets done in the next few weeks, then we're happy to come back and talk to people about what we've learned through that process. But on the financial outperformance and the digital transformation, so let me take you through a deeper dive on some of these things. First is this, this idea of financial outperformance.
If you look at, you know, this is end December 2022, the top 100 banks in the world by asset size. On the extreme left, our top line has grown at about 6%, and that puts us in the top 20 of those 100 banks. If I remember, we're probably number 16 or something in terms of top line growth. If you look at the profits in this period of time, by the way, I've gone back to 2015, because that's when we started our digital transformation journey, and this is the only chart where we've benchmarked it to 2015. If you look at the profits, we've grown 9%, and that again puts us in the top 20 of these banks around the world.
But if you look at our return on equity, and that's really, material, we're actually in the top 10 banks in terms of improvement in ROE in the last, seven, eight years, in terms of, outperformance. Actually, for last year, we were number eight in the world in terms of ROE, and of course, as you saw, the first quarter of this year, our ROE was north of 18%, 18.5%. Now, it's quite clear some of this outperformance is cyclical. You know, we are beneficiaries of the interest rate environment and the credit cycle. But we think that there's some structural stuff going on here as well, because the cyclical factors impacted the whole industry.
So the whole industry saw an improvement in returns, and therefore, the fact that our position has gone up from 44 to number 8 means there's something more than cyclical going on over here. So the first presentation after me is actually Sok Hui, the CFO, and she's actually going to try and unpeel some of the stuff. So we're going to try and unpeel where we think this structural, secular improvement in our returns is coming from, how you can actually get some line of sight and see the breadcrumbs. And she will actually then go on and try and give some sense for what is the medium-term outlook, so where we think this is going and what we can expect to see. So that's the first presentation after mine.
Sok Hui will basically focus on both the financial outperformance but also the outlook of what we hope to be able to see. Then after Sok Hui, the rest of the day is architected into two big sets of presentations. So on the left, these presentations refer to what I think of the enablers, right? So we'll talk a little bit about what we've continued to do in the last 6, 7 years to create this sustainable advantage with technology. So we're going to have... Jimmy's going to talk a little bit about- I have slides on each of these, so I'm going to follow up. The second big thing is, we really think that what we've done in the last few years has changed the way we work.
Again, I have a slide to tell you about it, but this, the method of working, including a lot more use of things like AI, et cetera, is really important. So that's the second key enabler we're going to talk about. And the third we're going to talk about is our credit process. So we've also upgraded and redefined the way we run credit, again, a use of data, et cetera. So these are the three enablers. Basically, the way we manage is fundamentally different. Now, it's important to understand this because I think about, you know, up till COVID, I think competitive differentiation technology just came from digitizing better, right? And so being able to change your tech stack, digitizing, et cetera. Through COVID, there was a lot of catch-up. Almost everybody had to cater to working from home and consumer markets and so on.
I still believe that we're competitively different in terms of the range and comprehensiveness of our digital capabilities. But I also believe that as you go forward, the real differentiator is in this. You know, how can you use data? How do you create systems? How do you manage differently, et cetera? And so, it's worth spending some time being able to explain that to all of you. And then after that, we have 5 presentations, which are on the right, which really focus more on specific business outcomes. The first of these is the consumer and SME business that we showcased 6 years ago in 2017, the Singapore and Hong Kong business. So, we will show and demonstrate how we've continued to make progress in that business, how we've been able to achieve what we thought we could, and where we are.
This is interesting, obviously, with all of the new banks in Singapore and et cetera, so it will give you a good sense for how we're positioned, where we're going... The next three are businesses we did not cover in the past, because in the past, our focus was consumer retail. In the last 6-7 years, we've taken this whole digital capability to large regional businesses, private bank, transaction banking, and the T&M business in particular. So we will showcase for each of the businesses, what we've been able to do over this period of time, as well as some sense of, therefore, how big is big? You know, where do we go with this? And finally, we have also, in this period of time, been able to use the capability to start scaling in some of those growth markets we talked about.
We're going to do a use case, a showcase of India, because, again, people have had questions about post-LVB, what is India looking like? We'll take only one. We didn't have time to cover lots. We'll cover the India piece of the story in there. That's the second set of presentations is going to be around the outcomes related to some of the businesses. Let me take, like I said, the sneak preview on each of these. On the technology thing, there are five things that Jimmy, our CIO, is going to talk to, things that we've done. One is the use of cloud. You know, I think we mentioned it the last time, that we've actually built a virtual private cloud, which is world-class.
It continues to be world-class in many capabilities in terms of cost, point, scale, et cetera. But in the last few years, we've been able to create a hybrid cloud capability between a good match between the private cloud and leveraging the use of public cloud. So Jimmy will talk a little bit about that. He will talk about data and what we've done with data, and this is probably the single most important thing we've done over this period of time. I'm personally convinced that the future is around, forget ChatGPT, irrespective, future is around the leveraging and user data for AI. But the trick to that, and the biggest moat in this game, is the data. So how do you actually get the data so you can actually constructively use AI at scale, with this?
It's taken us 6-7 years to be able to do that, ingest our data and create the appropriate data infrastructure and architecture. So Jimmy will talk a little bit about that. APIs. Again, in 2017, we already started talking about application programming interfaces, a different way to plug and play, both with the external world and with our internal pieces. We've continued to build on that. So the API infrastructure and architecture, which allows us to plug and play, this is really important if you want to build ecosystems and partnerships. He'll talk about that. He'll talk about how we continue to mature our Site Reliability Engineering practices, and then he'll talk a little bit about the experimentation thing that we have, use of blockchain, use of other kinds of capabilities.
So after Jimmy, the next one I'm going to, we'll have Kwee Juan. Kwee Juan, actually, till last month, used to head up our strategy and planning functions, a lot of these things in the transformation office. In the last month, he moved. He's now the, the country manager for Singapore. But he will talk fundamentally about three big capabilities and ways of managing that we have built over the last six, seven years. So, the first is, you know, how do you manage horizontally? And people call it agile at scale, and where some of you went and saw the Trust Bank walk around last week as well. If you think about most tech companies and the way they manage, they all manage horizontally. And the problem is that most legacy companies like us are organized vertically.
You'll have HR, you'll have marketing, you'll have, you know, finance, et cetera, but workflows horizontally. In reality, I mean, we, as organizations, have known this for a long period of time, that organizing work vertically is done because it's efficient. You know, Max Weber came up with the pyramidal hierarchy 150 years ago, with middle management and, you know, separate specialization of work. That was appropriate for Henry Ford assembly line. Over time, companies have tried to switch that to follow the actual flow of work. It's been hard to do. You get more effective outcomes, but you can't scale. It's not efficient. The tech companies started doing this 20 years ago, figuring that you can actually organize work horizontally now because technology lets you do it very differently.
And so they brought agile practices first to tech development, but then eventually, they brought agile practices to other pieces and everything in the company. ING, as you know, was famous for having tried to do this now 10, 15 years ago, 8, 10 years ago, trying to create this, and they had some success. The problem is that most legacy companies which have tried to do this have failed. Many of them have failed because it's not easy to convert a company from vertical to horizontal management. We think we've succeeded. Today, 60% of the company is running through journeys and running it horizontal.
So, the first thing, Kwee Juan is gonna talk a little bit about what we've done, how we do it, and why we think that we've been able to create a way of managing, which is somewhat different, the concept of agile. The second thing he's going to talk about is how we've been able to industrialize AI. I told you, the AI is going to be super critical. Half the battle in AI is the data, maybe 60%, getting the data right, and I said, we'll talk about that. It's not easy to do. It's taken us many years. But the other big chunk of the battle in AI is not the modeling. In fact, with ChatGPT, it's become even easier. But even though the modeling is important, but the real trick is institutionalizing.
How do you get AI into the operating rhythm of the company, so your people at the front doing the journeys all, adequately equipped to start using AI in the day-to-day business flow? So it's not centrally driven, but, federated, democratically driven. So he's going to talk a little bit about how we've tried to industrialize AI to get broad-based outcomes. In fact, there's one theme you'll see, and we'll go down later, across this thing. The one thing that you'll probably pick away is the amount of AI we're now using across everything in the company. The third thing that Kuijuan will talk to you about is ecosystems. So one of the other changes that we've successfully done in this period of time is being able to leverage partnerships and ecosystems for a lot of our origination, underwriting, creating product value propositions, et cetera.
Obviously, there's architected, again, of the technology we have in APIs, but there's been a real shift in how we go and approach market in many in many products and in many countries. So Han Kwee Juan is going to cover all of that.... The third piece in the enablers, Tan Teck Long, who's our CRO, he will talk briefly about how we've been able to use all of these tools to rethink the way we're doing credit. And in particular, of course, he'll talk about the efficiency we have in the process, but in particular, he'll talk about how we're using data and information to do credit to originate differently. And that's important because this allows us to scale in growth markets and in segments that we have found, credit thin segments, that we found difficult to be able to underwrite in the past.
So he'll talk a little bit about how we can do that, how we can underwrite in with our ecosystem partners, and how it gives us access to new segments and new markets. He will also talk about how we're using the same data and capabilities to do more effective portfolio management, right? So early warning signals, how do we do our stuff at the portfolio level? We believe, and Sok Hui will point this out, that our overall management of our credit capabilities is actually showing up in reduced NPLs and reduced provisioning levels relative to the market. So Tan Teck Long will show you a little bit about what we do and how it's beginning to make a difference. Moving on to the right side, the outcomes.
So like I said, the five pillars, and we're going to start with the consumer and SME in Singapore, Hong Kong. Tse Koon, who ran Singapore till a month ago, he's now Head of Consumer and Private Bank. He will take this, and, I think he'll do a deeper dive into Singapore, but the numbers on top relate to the whole, consumer and SME. So, he'll demonstrate that we continue to get solid growth. So, this one, again, from the time we started digitizing, we've got solid growth in income. Our cost income ratios have improved. Our return on equity is up. Again, some of it is flattered by rates, but even if you X out the rates, about half of the difference is structural, half of the difference is rates. So, he'll be able to demonstrate that.
He will focus on these half a dozen things that we've been able to do. One, you know, when we met 6 years ago, about 33% of our customers had been converted to digital. And we said then that over the next few years, we expect this to get to 50%-60% digital. We've actually converted 60% of our customers to digital, which means three-fourths of the activity with the bank now happens digitally. So, we've been able to do that. We've been able to close most of our last mile gaps, so even what we had the last time, and COVID helped. So, we essentially let everybody do everything from home.
But what that means is that we think we have the most comprehensive digital solution in the country today, bar none, because you can do everything digitally on your, this thing. We think that's important because there are no specific things that you still need to, you know, I can do payments, but I can't do a mortgage. We, we pretty much cover everything. We further rationalized our distribution, so our branch system and our cost of distribution have got rationalized further in the last 5-6 years. He's going to talk a bit about that. We've broadened our product range. What in the last 6 years we've, we've been able to do is, because of digital, take and create some product capabilities which are really important game changers and pump that into the retail mass market. Democratizing of wealth is one.
He's going to talk about that. We made digital marketing more effective, so use of AI and nudges to be able to get better outcomes for our digital marketing. And finally, gain access to new segments. So for example, we weren't a big player in the micro-SME unsecured segment in the past. Because of the digital capability, we've been able to get into segments like Shee Tse Koon is going to give you a sense for all the things we've done, which have continued to drive dominant market shares in the Singapore market, actually increasing market shares. Then, private bank. So we'll have Joseph Poon come and talk about private bank. In a private bank, the big question is always at the high end, ultra-high net worth material, does digitization actually help?
As Joe will show, you know, we've continued to scale our private bank really well. We are today we're top four private bank in Asia. Our AUMs over the last this thing have grown at 9%. The market has grown at 3%. So our AUM growth has been much, much faster than the market. Our income growth has been solid from $1 billion to $1.7 billion. Our cost income ratio at 46% is world-class. I don't know most private banks who've got a cost income ratio in the sub-50% rate, we are in the 46% range, and our ROE is at 30%. So it's a very solid, high-performing business. But what Joseph will show to you is the value proposition, the model that we've created, which we call Phygital.
At the high end of the market, you still need the RM. But how do you couple that RM with digital capability? How do you inform the RM, and how do you take the digital capability and also make it available for self-execution and self-direction to the customer? We've been able to stitch that end-to-end together really well. And one of the easy ways to think about that is our return on AUM is about 1%, right? And he will show you how it's much higher than the return on AUM, everybody else. We think that the way we manage the client and bring it together, we get much better outcomes from doing that. After private bank, the next big business we'll talk about is transaction banking. Transaction banking, again, we've really scaled.
The business has gone from $1.9 billion to in 2017, and when we showed you last, to $3.3 billion. Cost income ratio is low, at 32%, also coming down. ROE is up. Again, like, the earlier thing, about half of this is explained by rates, but half of it is explained by structure. And Su Shan, who's going to do this section, she runs our institutional bank, she will point that out. The three things that she will try to make sure you take away: one, that in this period of time, we've really improved market access. As the region has gone into real-time instant clearing system, we have plugged and been able to become a member of most of these environments.
So we're a direct member of CIPS in China, we're a member of UPI in India, so we've been able to get market access across the region. And we've been able to create capacity, because one of the big shifts in transaction banking is high volume, low value, right? And so the volume considerations are very, very different from what they were 5, 7, 8 years ago. So we've been able to create the capacity to be able to do high volume and very low latency. Now, that means that we can capitalize on these two mega shifts which are happening. One, more and more of commerce is moving online. And when commerce moves online, the way you do your transaction banking, your collections, payments, receivables, it all changes. It changes because you don't have—you can't batch and aggregate.
You know, earlier, your salesperson went, he went to the shops and, you know, collected a lot of checks, but you gave the bank one single transaction to do, which is all the sales proceeds for the day, you hand it to the bank once. In a real-time world, you have to deal with every transaction in real time, so you've got to change the whole orientation for how you run this. Similarly, supply chain shifts. COVID helped, but even before that, the supply chains are changing quite fundamentally in the way they operate. Much bigger use of things like platform, much bigger use. So, Su Shan will point out how the technology and capabilities we have been able to leverage on this disruption in the way transaction banking capabilities are needed and given us some reasons to be able to scale what we have.
And the last business we'll talk about is treasury and markets. Treasury and markets, again, you can see we've had very stellar outcomes. We've grown from SGD 1.9 billion to SGD 2.7 billion in income. Our cost income ratio has come down. Our return on equity is 8 to 4. By the way, this is not just the trading income. This includes trading and sales, so some of this is a duplication from the earlier slide, but overall, we're getting a 14% ROE in the markets business. In some ways, I think in this area, we've actually just played catch up with what some of the big banks in the US had. You know, so but relative to Asia and the Asian banks, I think we've created capability which is quite differentiated.
So, Andrew will talk about this, how we've been able to create pricing engines so we can price and structure at speed, how we've been able to create electronic distribution, so both internally, externally, all our distribution now happens electronically, and how we've been able to create trading efficiency, including the risk warehousing that we do. So, Andrew will cover up, this piece. Actually, the one key takeaway, which I think, Sok Hui will point out, if you look at these three businesses, the transactional business, the private banking business, the T&M sales business in particular, these are all capital-light businesses and high-return businesses. And if you look at the relative share, therefore, of these businesses in the mix of DBS, this has been increasing.
We think that this is a phenomenon that we will continue to see in coming years. Following the three business lines, back to India, we want to talk a little bit about India. If you look at India, we think we finally cracked the puzzle and started getting scale. Our income between 2017 and 2022, we did about $550 million last year. Our cost income ratio has started to come down, 78%. But most importantly, from a loss-making franchise, it's now started making money. Now, 5% is not great ROE, but the switch from a negative to positive ROE is material, and we think the outcome as we go forward into this thing is that we've got the trajectory right.
So, Surojit, who runs our India business, will talk about some of these things. He'll talk about expanded franchise, how, again, bringing phygital has turned out to be the best way to play this game. So we migrated from pure digital to, phygital presence. What we have now is a full service platform, so we, we can do everything that a Kotak or HDFC or any local bank in India can do. We have a complete platform, which lets us do everything like them. And how we're leveraging digital capabilities and ecosystem partnerships to really grow in the high return segment, consumer and SME. As you know, our India franchise was typically a large corporate franchise, which is a lower return franchise.
The big returns on consumer and SME, and we're beginning to see traction and growth, and therefore, the better return part of the country. So Surojit will cover that. So like I said, an update on consumer SME, Singapore-focused, the three big regional businesses and one, case study of the India opportunity. So, to bring it back, the overview is my section. Sok Hui will do the group financials after that. I'm going to come back at the end for some closing remarks. The enablers, the technology capabilities, the new way of working, credit risk management, and then, the stuff at the bottom. Now, as you notice, I've got some italics in there. The italics in there represent five showcases that we have around the room.
So we're going to have one showcase on how we use data and AI, which is over there. We have a showcase on how we're using workflow and work engines to drive this horizontal working. We think that that's the route of being able to get it right, so we'll showcase on that. We will showcase on how we've been able to use Singapore open banking to democratize wealth solutions, which is what's giving us this big thrust in the retail segment, wealth segment in Singapore. So we'll showcase on that on that side. We're going to do a showcase on how we've rethought transaction banking to help companies transform for this e-commerce and new world. So we will showcase on that.
Then finally, we will showcase behind the, that corner on how we have used data and AI to reimagine how we do treasury markets, on the five showcases. Now, I think Mona's going to come and give you some instructions later on the showcasing, the five of them. Each of you have a badge which is color-coded, and basically, depending on your color, you're going to go into a different showcase. Each showcase is about 8-10 minutes. So, it'd be useful if you go there, because then you can hear the full story as opposed to different. Mona will give you more specific instructions around that.
At the end of all of the presentations and in the afternoon, we'll leave about an hour for Q&A, where a few of us will get up on stage and happy to answer any questions that you might have. That's what I have to tee off, so you hopefully get a good sense for what we think we've done, why we think we're different, and what you should expect to see and take away from each of the sessions in the course of the day.
Good morning, everyone. I'm Chng Sok Hui, Group CFO. Earlier this month, we announced our first quarter results. Net profit, SGD 2.6 billion, ROE 18.6%, and we gave guidance for the full year, with profits to hit SGD 10 billion and ROE to be above 17%. But I know you are looking past 2023. As you calibrate your valuation models, I hope to help you build some convictions around the following questions: To what extent has DBS outperformance been the result of a higher interest rate environment and a more benign credit cycle? To what extent did structural factors play a role in growing our ROE? What might be the medium-term outlook ROE be in a more challenged interest rate environment? And what are the opportunities and drivers for future growth? You have heard Piyush unveil our investment thesis.
My colleagues will provide proof points of how our digital capabilities, our new way of Managing Through Journeys, leveraging data, artificial intelligence, ecosystem partnerships, and the momentum in our growth markets, will help us drive the medium-term ROE from to the range of 15%-17%. This chart shows you that our successful digital transformation since 2015 has led to a higher ROE and better shareholder returns. The top chart shows you the ranking of the top 100 banks by asset size. DBS has leapfrogged the position from 44th position in 2015 to the 8th position. We are among the top 10 today in terms of ROE.
And you might be wondering, "So where would you place the best U.S. bank or the best European bank?" In 2022, they delivered ROE of about 14%, which placed them in the mid-teens. Turning to the bottom part of the chart, that is where you see total shareholder returns since 2015, and you see that DBS has delivered 16% annualized TSR, which places us fourth place among our global peers. What does this mean in concrete terms? It means that you invested a dollar in the DBS stock in 2015, you'll be up almost three times, 2.8 times, to be exact, in 2022.
This annualized shareholder return is formidable, and I hope you are all shareholders, happy shareholders, and you have invested from the investment returns that DBS provided in the last few years. The transformation has not only helped us to leapfrog our global peers, the transformation has helped to widen our structural ROE advantage. In this chart, you'll see DBS ROE profile from 2015 to 2022. We show on this same chart our Singapore peer average, only for the reason that we operate in very similar markets and have the same and face the same macro environment. You'll see from this chart that our ROE has been widening since 2017, and that we are currently at 3%-4% ROE higher than our average peers.
This has happened during a period where actually Fed fund rates have been cut to almost zero in 2020 and 2021, which hurt DBS net interest margin the most, and yet we outperformed. There is something structural that is happening here. Now, this chart shows you where we have improved and what are the drivers for our structural improvement. You'll see that in 2015, our ROE of 11.2% has been lifted by 3 percentage points to an ROE of 15%. Structural drivers have contributed to a 2.5 percentage point improvement in our ROE, and I'll go through point by point in the later slides.
You will note that financial leverage was a drag of 0.9 percentage point, and that really means that in 2015, our CAR ratio was 12.4%, and it was 14.6% in 2022. So, the 2.2 percentage points of capital buffer was the main reason causing the financial leverage factor, which was a drag on ROE of about 0.9 percentage points. The external drivers you are familiar with, and you have built it into your models. We did benefit from a high-interest rate environment, a more benign credit cycle, as well as the market growth factors. Turning to the structural drivers, which I've labeled A, B, and C, I will unveil in the subsequent slides how these three drivers are quantified.
The first driver is really extending our superior deposit franchise from Singapore dollars to foreign currencies, and how we have been able to grow our deposits faster and to grow it more cheaply. That lifted ROE by 1.6 percentage point. The second item, which is shown as B, is actually the growth in our high ROE businesses, and that would be in wealth management, in transaction services, in treasury sales income. And you see the slide of how we have outperformed and quantified the improvement at 0.6 percentage points. The third factor is new capabilities in credit risk management, which our CRO will talk to, but we have quantified the impact of 2.3 percentage point structural improvement to ROE. The first factor, extending our superior deposit franchise to foreign currency.
As you know, deposits are the crown jewel of any financial institution. I would like to make four points on this chart of how our digital capabilities have boosted our deposit franchise. The first thing to note is that our Sing Dollar deposits market share, savings market share, has actually continued to grow, and today, our market share, as at April 2023, is 54%. The second point I would make is that our foreign currency deposits have now been deeply entrenched, and we also have a superior deposit franchise. Foreign currency deposits now make up 60% of our total deposits of SGD 522 billion. Foreign currency deposits are growing faster than Sing Dollars. Foreign currency deposits growing at 8% CAGR over this period, compared to Sing Dollar deposits, which grew at 6%.
What is even more pleasing is that as a proportion of total foreign currency total deposits, foreign currency CASA accounts, cash operating accounts, which you will hear my colleagues talking about, the CASA operating accounts, which are very sticky, that has also grown from 35% in 2015 to 54% in 2022, a lift of 19 percentage points. Not only have we grown our deposits faster, we have grown our deposits more cheaply. You'll see from this point that our measurement of this, which is used by the industry, is deposit beta. And during this period, where there's been unprecedented increase in Fed Fund Rates, we measure how much of the increase is actually passed through to customer rates. In the case of DBS, it was 29%.
For our local peers, it was 34%, and for the average of the US banks, it was 43%. And again, what is more pleasing is that our CASA, our GTS CASA accounts, the deposit beta is actually half the 29% at 15%. So all these drivers, the ability to grow our deposits faster, to grow our deposits more cheaply, and to deploy them in high quality liquid assets, which are very capital efficient, actually have led to an ROE impact of 1.6 percentage points. Second structural driver, the growth in high ROE fee income. And here you see we have listed wealth management, transaction services, cards.
We didn't put in some of the treasury sales income that you see under the other income line, but they have grown at a similar pace, only because it's hard to compare and outstrip that component versus peers. You just look at the fee income, it has grown from SGD 2.1 billion in 2015 to SGD 3.5 billion in 2021. That is a CAGR growth of 9%. Very impressive growth, and it is actually double that of our peers. Of course, in 2022, wealth management income took a dip as customers turned more cautious. You see that even if I measure from 2015 to 2022, the CAGR growth is 5%, which is three percentage higher than peers' growth of 2%.
You can see that this structural improvement is how we have used to quantify the 0.6 percentage point lift in ROE. Third structural drivers, the new capabilities in credit risk management that's driving structurally lower allowances. You see from this chart that between 2015 to 2018, our credit cost was actually higher than that of our peers. From 2019 to 2022, our credit cost was about five basis points lower than our peers. And that is something that we have also communicated to you in our previous results briefing, that with all the investments we have made in more effective credit portfolio management, through AI, through early warning signals, we have already quantified that there would be a structurally lower through the cycle credit cost that we can expect, which was also at five basis points.
So that gives us a lift of 0.3 percentage point in ROE terms. So in total, our superior deposit franchise, being able to grow faster, grow more cheaply, deploy more profitably, as well as outperformance in our fee income, as well as structurally lower credit cost, has led to a 2.5 percentage point structural improvement in our ROE. I'll now turn to the near-term outlook. Our fundamentals are sound. Global Finance magazine has ranked DBS number 1 or 2 in the last 2 years, among the world's safest commercial banks. We have also been named Asia's Safest Bank for 14 years in a row, from 2009, when we started the ranking, till 2022. So I will not go through the metrics that are typical to assure you that we are a very safe bank.
High capital adequacy ratio, strong liquidity position, strong asset quality position, and you can see it in our results announcement. What I would like to quickly recap is the base case for 2017, 2023, the ROE is likely to be above 17% from a higher net interest margin, from fee growth in the high single digits, cost-to-income ratio below 40%, and credit costs of around 10-15 basis points. That is sort of still below our through the cycle credit costs, which we estimate to be more in the 15-20 basis points range. And we have done a lot of stress testing, given the external environment. We have stress tested our portfolio, and our asset quality continues to be very resilient. We have set aside substantial allowance reserves.
General provision allowance reserves are now at SGD 3.8 billion, of which SGD 2 billion are in the form of overlays. It means that we have a basic general provision model, we have created a lot of stress testing and set aside SGD 2.2 billion of overlays that can be released in times of stress to buffer the allowance line. Our NPA coverage are at very high ratios of 127%, if you include collateral, at 229%. Our CET1, we are still operating at above our guided, range of 12.5%-13.5%, and this capital buffer is expected to increase during the Basel IV transition period. We have ample liquidity to support our business operations, even in stress, funding conditions.
We are not only a safe bank, we are a bank poised for growth, poised for scale in the new businesses that we choose to operate in. This is the guidance that we released this morning. Our medium-term ROE will be in the range of 15%-17%. Some of you have asked, "How long is this medium term?" You can take it to be around 3-5 years, and we intend to operate at a CET1 ratio of 12.5%-13.5%. With a 2023 ROE at above 17%, you must have built in all these external drivers. Interest rates are expected to be peaking or have peaked, and therefore there'll be some downward pressure on NIM. Credit cost is expected to be higher than what we have put in to 2020...
In 2023, of 10-15 basis points. The third factor, which some of you may or may not have incorporated into your valuation model, is BEPS. This is the G20's implementation of tax rules globally. And most countries will begin to put in place a 15% minimum tax rate. Singapore is no exception, and we do have a tax rate that is slightly below 15%, and therefore, there will be some additional tax costs to be funded. If you look at the structural drivers, which I will go through in more detail, the first one is business mix, right? You'll see that we are able to grow, be more profitable, low capital, high ROE businesses faster. That is the first driver.
The second driver is growth markets, and here we would show you, even, in subsequent presentations, of how in each of our growth markets outside Singapore and Hong Kong, in India, Indonesia, China, and Taiwan, we expect to benefit from scale, which will double our ROE in our growth markets. The third point is capital management. Just as in the previous period, 2015-2022, a higher capital ratio led to a 0.9 percentage point drag on ROE, because we maintained a higher capital ratio during this period. Likewise, as we move towards our guided management operating range of 12.5%-13.5%, you should expect that there will be some lift to the ROE, and we'll have enough capital headroom to support profitable growth and higher distributions.
The first lever, faster growth in capital light, high ROE businesses. You'll see that this is similar to the chart that Piyush showed you earlier, in the three main lines of wealth management, global transaction services, and treasury and markets income. In 2015, they were about 39% of group income, rising to 42% in 2017, and 47% in 2022. We expect this business, business mix to continue to improve going forward and to exceed 50%. Second lever, increasing scale in growth markets.... Just so again, to repeat, in this chart, we show you the growth market outside of Singapore and Hong Kong. It will be India, Indonesia, China, and Taiwan. Look at the opportunities in these countries, which you can see on this chart.
In the business lines that you know we are pursuing, they have extremely high CAGR growth. In transaction banking, in wealth management, in SME lending, the CAGR growth for the entire industry in these four markets are averaging 7.5%. In unsecured retail lending, the CAGR growth is almost 10%, but we are only capturing a very small sliver of the potential banking revenue in these sectors for these countries, and therefore, we have the opportunity to scale and to grow much faster than implied by the CAGR growth you saw on this chart. So Surojit will show you how we are using our differentiated business model to deliver the scale we need for India. Piyush has showed you that India has already transformed with an ROE of minus, from negative to a positive 5%.
The business models that not only India, but also the other countries will be adopting or have already been adopting to drive scale, are in three main areas. In digital ecosystem, Surojit will explain to you how our ecosystem partnerships in, with different, with different partners in, in unsecured lending, in gold loans, in co-branded credit cards, will have actually led to an increase in customer acquisition. The other point to note is, it is not easy to acquire these customers and originate at scale if we do not have the credit infrastructure in place.
So the second factor is that we have already developed these models, the Artificial I ntelligence, the machine learning models, that our CRO will explain to you later, and that allows us not only to originate, but also to originate at scale, comfortable with our risk appetite and to our own standards of effective portfolio management. The third point I would make is that we are building local depth, but via our phygital model. Piyush has talked to you about LVB. We have understood how digital capabilities can be married with an enhanced network to drive growth. We have already seen it in India. Taiwan Consumer Bank will benefit from Citi acquisition, which will be completed in the third quarter of this year.
In the Greater Bay Area, we have an associate, Shenzhen Rural Commercial Bank, and in collaboration with Shenzhen Rural Commercial Bank, we also see the opportunity to build local depth via this phygital model. So in all, the growth market, we believe, will lead to net profit tripling in the next few years and ROE to double to the medium term. And this is understandable, given that in these markets, the ROE today are still subpar and have plenty of opportunities to actually grow to at least meet our cost of equity. We are already seeing results. The customer data in 2015 was 6 million. We had 6 million customers in 2015, and that has grown to 12 million in 2022. Third lever, capital headroom to support profitable growth and higher distributions.
At the AGM in March this year, shareholders approved a special dividend of SGD 0.50 and also an increase in the fourth quarter dividend of SGD 0.06, from SGD 0.36 to SGD 0.42. This brings the total 2023 dividend per share to SGD 1.68. Our baseline planning assumption is that, barring unforeseen circumstances, we should be able to continue to step up the dividend payout by at least SGD 0.24 per year, and that's sustainable in the medium term, given our ROE trajectory. We also believe there would still be further upside of, say, SGD 3 billion, that we would be able to distribute because we would still be operating in a range that is above the 12.5%-13.5%.
The distribution could be in a further step up of more than SGD 0.24 in ordinary dividends or, special dividends or buybacks could feature more regularly or more prominently. The pace will depend on the business conditions and the macroeconomic outlook. So to recap, I'm sure many of you, and I've read your analyst reports, that you have factored in the external drivers into DBS's ROE projections. I hope that you would also be convinced that these three levers that I have just laid out for you in terms of faster growth in capital light, high ROE businesses, higher ROE with scale in our growth markets, and the capital headroom to support profitable growth and higher distributions, will enable us to deliver ROE of 15%-17% in the medium term.
We have proven that we are able to deliver with digital transformation 1.0, which we unveiled six years ago. We have done it, and with digital transformation 2.0 we will be able to do it again, and to bring DBS to a new level.
Hello, good morning, everyone. Very good to welcome you here again. You heard earlier on that Piyush mentioned about how we have used digital transformation to drive growth across all the business segments and across all the markets. And you heard from Sok Hui, how that has translated to our financial outcomes. And we believe that a big part of this was really about technology capabilities. It's one of the key drivers of the entire growth. What we also believe is that these technology capabilities have created, also give us a sustainable advantage. And why is that so? One, the transforming tech capabilities is not easy. It takes time.
You know, when we started our whole entire journey of digital transformation, we went down to the valley, and we actually learned from, you know, the likes of the tech giants, like Google, Amazon, Netflix. And we also found that their digital transformation that makes them the tech giants today takes them about 7-10 years. We are in the 9th year of our digital transformation, and that gave us a little bit of advantage over the peers that have just started, because building technology capabilities takes time. One of the many key takeaways that we actually have learning from these tech giants; there are quite a number of them. And Piyush mentioned briefly just now, you know, the journey to the cloud, the massive use of data, API-driven architecture, Site Reliability Engineering, as well as experimentation with new tech.
And what was really important for to, you know, in this one data transformation, is the talent bench strength that they have. So what we did when we actually learned from all these tech giants, we came back and actually distilled all the learnings, and we actually created this whole thing called Being the D in GANDALF. As you can see, the whole acronym, you know, it's actually the tech companies logo. You know, Google, Amazon, Netflix, and D is, of course, DBS, Apple, LinkedIn, and Facebook. We take all this, actually take this and run into, you know, a big program of transformation since 2015. Let me give you a very quick tour in terms of what we have done, you know, since then.
Now, digital transformation, you know, is really about taking, you know, turning boulders into pebbles. Taking big, rigid systems, you know, the likes of IBM and running on monolithic applications, and actually transform them into nimbler modern tech stack using microservices as well as API architecture. We have been able to actually progress quite steadily, and today, our whole entire tech stack, you know, is about 99%, running on cloud optimized. Now, the thing is, we still have our mainframe. Now, the thing is that, you know, we will actually get off the mainframe, you know, in the next 3-4 years, and this has actually translated to benefits for us. Our modern tech has enabled us to scale with agility.
You can see our total deployment releases, you know, since 2017 to 2022, has actually increased four times. If you compare ourselves to the leading banks, we are actually slightly ahead, and of course, we actually, you know, closing the gap with our tech company partners. So let me actually move on to actually look at one of the key takeaways that we have on the journey to the cloud. Like all the technology companies, they use massively the public cloud, and we too use the public cloud for advanced capabilities and cutting-edge technologies that they offer. But what is different from them is that we have a hybrid multi-cloud strategy, that allow us to actually have on-demand computing power to be able to use on the public cloud.
What we have built, you know, in the private cloud is an in-house infrastructure as a service, that we can achieve efficiency below the price point of AWS. This is really important. You know, this allows us to actually run persistent workload as well as storage at a lower cost. How were we able to achieve that? Well, you know, you're using very simply, capitalizing on Moore's Law. You know, when we first started our whole entire building out of our VPC, which is our virtual private cloud, in 2016, we have a capacity of 250 VMs, virtual machines. Think of VMs as compute power, right? This is the 250 VMs that we have.
Our cost efficiency ratio, at the point of time, was about $330 per VM. Was above the price point of what AWS was quoting. But comes two years later, you know, in January 2015, we upgraded our whole entire hardware stack. We doubled the number of VMs to 500. Our cost efficiency has dropped by a third, to a third at $120, below the price point of what AWS is offering. And in 2020, two years later, we upgraded our hardware again, and we this time we actually moved from 500 VMs to 1,000 VMs. And what is really, really amazing is that, you know, the price efficiency dropped to $70. Now, below the price point of AWS, in less than half of what AWS is offering.
And what does that give us? I think that previously, I think in the previous Investor Day, we talked about, you know, having our data center downsized by 75% and creating a 10, 10 times growth capacity. You look at this growth, and today, you know, this growth continues without having even to incur additional costs on our side. And this cost efficiency is actually very important for us. If you actually look at our tech CRR, it has actually maintained, you know, quite constant at 10%, you know, over the years. Our operate CRR, which is the productivity I have read that explained just now, enable us actually marginally decrease, you know, to about 5%.
That allow us to actually put our investment and our money and dollars into the build capability, to actually build more, you know, for our—not only just for Site Reliability Engineering, for cybersecurity, for data, and so for our customers. And you can see that, you know, in 2017, you know, we—our OpEx is about SGD 1.1 billion. You know, in 2022, it's SGD 1.6 billion. And the split between operate and build has actually shifted drastically. So coming back to, you know, the second key takeaways, you know, data. This one's a more important initiative we have run.
We have built key capabilities in ADA, in our data platform, and they allow us to actually ingest data at scale, and allow us to actually have instant access to permitted data, understand data in minutes, very, very quickly. And we have on-demand capacity and the state-of-the-art tools. I'm not going to go into details in that, because I think later on, you will have a booth there that they will actually give you a deeper explanation of the capabilities of ADA. But what I'd like to touch on is a few key points. Now, we are able to actually take quite a number of data warehouses that we have, you know, from each of our core systems. And you can imagine, you know, with all the core systems in each of the data warehouse, you will have duplicate data.
You have many, many data sets sitting, you know, in across the entire bank. We are able to take that and actually consolidate them into a single source of truth, into one data lake. And this is not easy. You know, to actually get data, you know, to clean them, to actually define them in the metadata, to actually look at the data lineage, to how data has been transformed and enable, you know, quick discoverability. It takes years, you know, for many of the banks and our competitors to achieve this. And we did that in the last five years. And this is really, really the moat that people talk about, the ability for us to actually get data and enable AI, ML at scale. And what's the other differentiator for our data platform?
Now, most of the data platforms that you have actually rely primarily on the public cloud. As I mentioned earlier on, our whole entire spectrum of cloud facilities, you know, from private, you know, to public cloud, actually plays with the way that we architect our data platform today. Now, we can see that, you know, in our private cloud, you know, we have a data storage as well as a compute, our GPU farms. And persistent workload as well as data has been stored there and run there at a lower cost point. But we also make use of the entire capabilities in the public cloud, you know, with AWS, Azure, as well as Google. We use their models as well as compute.
The ability for us to move data from private cloud to the public cloud using one of the tools that we have built, we call Beetle. It's a data orchestration layer. Also build another tool, which is called Oculus, which provide data security, you know, across the two, private cloud as well as the public cloud. Now, this is a very unique differentiator. For most of the other banks that you think of, they're actually worried about and concerned about data security. But when we actually look at this, the way they've architected, most of the data actually reside on-prem, and it brought up to the cloud for us to actually do all this compute and modeling. This is very different from our other competitors. Moving on to API. What really is API about?
Well, API is really about, you know, taking our core systems, building what we call plug-and-play or hooks, that enable internal as well as external players to be able to actually use and, you know, to actually interrogate the core product systems easily. Now, over here, as you can see, we have actually exposed a lot of our internal API to external API to actually create ecosystems. I will again not touch on the external APIs ecosystems, which my colleague will actually touch on later, how they use this infrastructure to create the ecosystem play. But enabling that is really our whole entire API marketplace.
It's a portal, a portal that allow us to actually publish, to actually develop and publish and provide the API, to allow people to actually discover what API is available, and then to be able to subscribe and use it. And as you can see, all these have actually translated to actually massive use. You know, our external APIs today actually gone up 10 times. Our external API volumes has also gone up about 100 times, and our internal API volumes has gone up about 20 times. Massive increase in the use of these APIs to actually, you know, access our core product system. What I'd like to do is just actually move on to actually look at how the internal APIs has enabled us to develop our workflow and workbench.
Now, in the past, when staff come into to work, they have to actually log into multiple systems to get the job done. Today, what we have done is to actually create a workbench that allows the staff to have a single sign-on to the particular workbench, where their whole-- all the work they have to do is actually residing. And how does this workbench perform? It's actually connected to our core banking system through our whole entire API layer. That allows us to have a single point of access from anywhere. And once the work is done, they can actually flow their work seamlessly to another part of a bank, whether it's in credit or operations to be completed. And that, by itself, you know, gives us ability to actually create what we call the horizontal organization.
And my colleague later will actually showcase to you how that has actually helped in the whole entire Managing Through Journeys. We do have a booth later to actually showcase, you know, how this has been done, you know, with one of the examples that we have. Moving on to the next key takeaways and capability. And it's really about maturing our Site Reliability Engineering. Again, we went down to the valley and now learn from the likes of Google, AWS, Netflix. We recruited one of the Google Cloud SRE head, and we created our own internal SRE framework. As you can see, it's actually very comprehensive.
It starts from the way that we architect, the way we build and develop, the way we test, the way we do product launches, you know, the way we operate, you know, with observability, and the way we handle incidents, and so forth. Now, you know, this has actually been a key part of our whole entire building capability building process. We have actually started building some of these key capabilities, you know, within our whole entire organization, and give us the foundation for us to actually scale this and embed it, you know, within our organization. Next. One of the key things about all these technology companies is the ability for them to actually experiment and adopt new technologies. We are no different.
We started blockchain and smart contract technologies and actually experimented with them even when the technology was nascent. We have actually built up quite a number of foundational capabilities. We have tokenization as well as wallet management, we have compliance, legal, and surveillance, and we can operate now efficiently, you know, for both public and permissioned chain. And as a result, all these capabilities that we have built over time, we have been able to capture and build new business as and when the business opportunities present themselves. In the area of origination as well as issuance, now we have our FIX Marketplace, a tokenized digital asset capital market issuance. For example, now, trading and execution, you know, we have our Climate Impact Exchange. It's tokenized carbon trading exchange.
In the area of clearing, settlement, and custody, we also have our digital custody servicing platform. Now, there are more of that. And the point I've been trying to make is that because we are already playing in these new technologies when they were nascent and building capabilities, when the opportunities presented itself, we were able to actually build all these new technologies that will enable these new businesses. Last but not least, I'd like to come back to this. Now, all these capabilities can only be built when we actually have the talent, and today we have about 12,000 technologies. We have people mainly based in Singapore and Hyderabad, and we are now also looking into China, which we will establish in 2024.
We are also hiring from all the big tech companies, you know, from the likes of Google, Apple, Huawei, AWS. These are all the tech companies that actually join us because of our whole entire DNA. We also have more technologists than bankers, and also have the breadth of talents and skill set, you know, from data, you know, and the depth of them, you know, in terms of analytics, data technology, to deep tech, you know, in the likes of cloud infra, SRE engineer, you know, cybersecurity, and so forth, and as well as in the area of design. You know, we are actually also attracting 1000 young talents from the polytechnics and the whole, the universities as well.
We have a pipeline of our young talents, you know, coming through these universities. So in summary, you know, I've actually gone through quite a number of technology capabilities, you know, with you, and it just showcase some of the things that we have done in the past 8-9 years of transformation. But these technology capabilities remain only as technology capabilities. It only becomes real when we actually be able to exploit these technology capabilities by the new way of working.
Good morning, ladies and gentlemen. My name is Han Kwee Juan. I am responsible for DBS Singapore as a country head. Jimmy has just shared our nine-year journey in building these technology capabilities, as you have seen on the screen. I'm excited to share with you three new ways of working that leverages on this technology to lift the trajectory of our growth. Managing Through Journeys is a permanent shift in the way we manage and work within our organization, focusing on customer journey, pretty much like how the big tech companies would do it.... Second, industrializing and infusing AI, ML across the entire bank. And third, in 2017, we shared with you our ecosystem approach on acquiring, transact, and engage at scale.
Today, I'm delighted to share with you the progress that we have made and share with you some of the proof points on how this is now very effective for us to launch a different growth that we will see through ecosystem partnership. These are the three new ways of working that we will embark on going forward. Managing Through Journeys. You may wonder and say: What is Managing Through Journeys? In Managing Through Journeys, multifunctional teams come together, working as one with one single focus, increasing customer satisfaction with a differentiated customer experience. It is a shift from vertical, siloed organization, as Piyush has described, to an integrated horizontal organization, all working together with one main ultimate goal of delivering the outcome that differentiated customer experience.
We believe we have achieved agile at scale with this approach, as MTJ now covers more than 61% of all of our CBG and IBG revenues in DBS. And finally, we are seeing impact of Managing Through Journeys in our operations. It will drive revenue growth and productivity gains. Let me just share with you... Sorry. All these are enabled and anchored on what you see here on Performance Cells and how the teams are organized, and as well as the workbench workflow or Workflow Workbench at the heart of the Flywheel, which I will explain. These are the secret sauce towards how we organize ourselves and get the impact of revenue growth and productivity gains. Let me start with Performance Cells and give you an understanding of how these multifunctional teams are organized.
Within the Performance Cells, it is led by two co-leads, one, a business lead and the other, a support unit lead. They come together as a multifunctional team, focusing on the original journey, origination journey, engagement journey, or service journey, where the team members comprise of product specialists, operation or technology, specialists, as well as bringing in expertise such as design experts or data scientists into the entire multifunctional team to deliver the experience for our customers. How they will come together for targeting the improvements and what we have is through the performance management architecture, what we call PMA, looking at customer outcomes such as system availability and customer satisfaction, business outcomes and financials, as well as acquisition and many more, and finally, on the employee metrics on how the experience would be for our employees and how they're working in this new environment.
Workbench workflow or Workflow Workbench is the heart of MTJ. This is where the magic happens. We will stitch together the work and orchestrate it from end to end from a customer journey perspective and allow the teams to collaborate and work together. This is best illustrated through a video that I will now play. The video, please.
For every customer, many departments in the bank will work together to complete the job to be done. But our work, people, information, and systems are largely organized within each individual business and support unit in the bank, resulting in gaps and manual handoffs and disjointed information flow with little, if any, end-to-end visibility. Take Mr. Tan, for example. A simple loan request that exceeds his credit limit can result in many steps of manual handoffs and duplications. Along with information gaps that exist, the customer journey is easily broken. The end experience for both customer and employee alike can certainly be improved, and so we did. Using data from past transactions, we've first identified the personas and mapped out the sequence of steps which the customer job to be done needs to go through, from initiation to completion.
These are workflows that help break down silos and stitch journeys by automating what needs to be done and connecting whom it needs to be done by. We then built common widgets that enable the real-time sharing of relevant information across employee personas involved in the job to be done, a common source of truth, visible to all involved, and placed it all within a workbench, a single unified platform that enables employees to get their day-to-day tasks done in a more efficient manner. In the case of Mr. Tan's loan request, with workbench workflow, all relevant personas are informed of what needs to be done when, in real time, at each stage, and decisioning processes are improved with relevant information readily available. The entire transaction is now efficiently and swiftly completed.
To date, we've created 13 workbenches and mapped out more than 120 workflow automations, growing our reusable asset marketplace, which can be tapped upon bank-wide for rapid, scalable implementation. Eliminating manual handoffs and duplications to reduce processing time for better collaboration, greater productivity, and an integrated, seamless, end-to-end customer and employee experience. Streamline customer and employee journeys via seamless and efficient workflows and workbenches.
... I would encourage all of you to visit the showcase right at the back here, to really see the power of how Workflow Workbench delivers the promise of a better customer experience for our customers. Value Maps and Control Tower are the brains of MTJ and let me share with you on their definitions. When we look at a value map, we often start with the outcome, the customer experience, and in there we determine what are the drivers that will deliver that differentiated customer experience, and what are the different levers that we can pull to make sure that we can deliver that. We believe that a differentiated customer experience will want customers to keep coming back for more, and that results in better financial outcome for ourselves. I'll use an example of the everyday MTJ that we have within the consumer bank in Singapore.
These are the Control Towers that you see online. In the everyday banking PC, we look at the interactions that we see with customers at the branches, at the call center, and digitally. One of the main things that we would like to achieve is to get customers to experience the superior digital experience that we have from all of our services, and work to move them from offline to online. These are the burndown charts that you see from the branches and call center and digital, down to the branch, down to specific activity, in this case, account opening. We looked at a metric called negative customer impact. The definition of that would be, when a customer goes through digitally, there are drop-offs. We saw 40% of our application dropped off, and that is not good.
So what we did was the team focused on bringing this down, and happy to share with you, it's now 20% dropped off and has led to a higher number of applications being processed through the bank and a lift on digital adoption for account opening to now 94%. This is the power of how a Control Tower help focus the teams to deliver a differentiated customer experience. And now, this is the exciting part where I'll share with you on the whole Flywheel effect that you will see using these various elements. Through the Flywheel, we will be able to generate positive impact through continuous listening, getting insights, employing behavioral science techniques, and using AI, ML, and experimentation to innovate new products. Let's see what they all mean. Continuous learning is a ritual of interviewing customers, determining their experiences with DBS.
Second, they are all trained in behavioral science and observation techniques, and we've done more than 346 immersions last year. With the insights that they gain, we will start to develop the hypothesis on what we could do to deliver, to deliver that better customer experience, come up with experiments, and then go through the whole cycle of observation, experiments, results, to fine-tune the outcome. Here is an example of what we have in transactional FX that focuses on corporate cross-border payment, with FX embedded in it, and with messages such as, "We missed you," on top of a price change. Actually, gives a far better response compared to just pure price changes. What an interesting discovery that we had through experimentation! Satisfaction has also gone up as a result of that.
We have now embedded innovation at scale within the MTJ, where through discovery or pain points, we begin to innovate on what we can do with our services and products. This has led to more than 600 new ideas being generated, and these are two examples that we have launched in the TFX and as well as consumer bank space in Hong Kong. Next up is on AI, ML, where we will industrialize AI, ML across the bank. These are the use cases that we have now used right across the entire bank. Everyday banking and wealth management, in consumer banking, and in institutional banking, we apply AI, ML in SME lending and as well as early warning for large corporates.
This has also been applied into our T&M business, on algo trading and sales, and in audit, we have used AI, ML to enable always-on audit and NLPs to generate the audit reports that we have at the end of the audit, generating that out much earlier than the previous way of doing it. Kian Tiong will share on how credit risk now is powered by AI, ML on the back of how we do credit evaluations. In operations, we use it for demand management, in HR, for recruitment and career development, and in finance, for outlier detection and financial forecasting. These lists are not exhaustive.
In actual fact, we have more than 600 models and 300+ use cases right across the entire bank, and therefore, when we evaluated economic value from AI, ML, we were very disciplined to look at a control group and compare that to those where we applied AI, ML, and determined that we have now gained SGD 178 million of incremental economic benefits, either revenue growth or expense saves in 2022. These are two examples that I'm gonna go through with you. First up, on digital payment fraud. As you know, real-time fraud prevention is critical, and to win the war against the scammers, what we have done is we've deployed AI, ML models on the back of it.
Looking at more than 300 data features that includes the customer's behavior, that we will see when they use their account, together with financial and non-financial transaction information, to enable us to launch an AI, ML model at the back of every transaction that we process. In less than 10 milliseconds, we would be able to determine an outlier in a particular payment that our customers initiate. And this has resulted in the lift of more than 24% of customers saved in transactions, and this model is actually far more effective than the rule base by more than 5x. We've also gained productivity as a result of it and received more than 10 times compliments from customers as a result of this change five months ago. In CBG, through cognitive banking, we have applied AI/ML right across the entire business.
As consumers, you would demand your bank to be able to speak to you as and when you want us to. Through AI/ML models, more than 200 plus of them, looking at 16,000 data points, we're able to speak to customers through nudges, giving them the right information, be it for investments or the next best restaurant they should dine in whenever they're near a particular neighborhood. This has resulted in an uplift for us in the consumer bank from AI, ML models being launched. We are looking at doubling that to $200 million this year. It is implemented right across all six markets, covering 7.5 million customers, and last year alone, we generated more than 344 of such nudges to both the RM and as well as to the customer.
Lastly, on ecosystem as the third new way of working. I'm delighted to share with you the progress that we have made. As you know, in ecosystem, we have three key objectives. We have now achieved acquiring customers at scale through ecosystem and acquiring those customers in the partners platform that we work with. Secondly, we have leveraged the data to enable us to be able to develop good credit models, but more importantly, we've used that for credit monitoring to enhance the risk management that we have in lending within ecosystem. And finally, the AI/ML models that we have are now embedded to engage the customers very differently as we work with them in those platforms. Our strategy remains the same. For CBG, it's about where consumers work, live, and play, and we embed solutions such as lending, wealth management, and payment in those platforms.
In IBG, it's about the procure-to-pay and as well as the cash to order cycle, embedding supply chain financing solutions, and as well as cash management and FX into these platforms. Let me share with you two examples on how we differentiate ourselves in ecosystem partnership against the rest of the banks. CRED is a leading payment platform in India for credit cards. More than a third of the credit card population in India use CRED to pay their credit card bills across India. So it is a very important platform in India. What we have worked with them is to look at the customer base that they have, to pre-approve the list of customers with their data and our model, and with that pre-approved, when the customers are making payment, we will prompt them on a potential loan that they could take up.
And when they decide to pick up the loan, we will do KYC, AML, and fraud checks immediately on the fly through the API connectivity. So within minutes, the customers will be approved and onboarded, and when they take up the loans, it'll be dispersed through our API connectivity into UPI payment infrastructure in India, and they receive their money in another bank's account. The credit monitoring that we would do is by looking at a payment behavior throughout the whole loan period that they have, and we will detect that there are any potential delinquencies that could come up because of smaller amounts that they pay or delay in payment.
We will use that to be able to talk to the customer and help her to mitigate the potential, you know, losses that could come because of weaknesses that we see. These APIs are implemented right across all other partners that we work with in our six markets. The second example is on supply chain financing partnerships, and this is where we work with JD Logistics in China and Hong Kong. I would point out some of the cool features that were deployed in this ecosystem. Within this platform, we are supplying financing for both AR and as well as inventory financing to customers of JD Logistics.
What we do is we would use that information, that they have on the transactions that they are doing on the logistics platform, and there are banner ads that we would serve using our AI/ML models to target the right customers for financing. When they're interested in taking up the loan, we will use the information within their platform, such as accounts receivable and inventory levels, to determine the right size loans to be given to them, and once they agree to pick it up, we would credit it into their account. The other cool feature that exists in this structure is we can actually see the inventory level and AR of each of their customers.
In event that we realize that their credit are deteriorating, we have the ability to tell JD Logistics to not allow them to pull out those inventory until they pay down the loan. So this is something that is really quite powerful through the API connectivity that we do and as well as the live interaction that we get out of it. This is also being implemented across all other partners for supply chain. The impact from ecosystem has been tremendous. The number of customers that we have acquired through ecosystem has grown up by 8x for consumer bank, and today it contributes to 30% of new customers that we acquire outside of Singapore. In IBG, it's about 2x over the last three years.
On the balance sheet side, it has grown by 8x, now contributing to 65% of all new secured and unsecured loans that we generate outside of Singapore for unsecured and secured assets. It's a supply chain financing assets that we generate, more than 30% of new bookings are now being pushed through from ecosystem. The AI/ML models that we put into the platforms allows us to engage the customers quite differently. So, on our own marketplace, PayLah!, which is the everyday app, we are now the number one bank or wallet that they would use with the highest NPS score and as well as the highest share of wallet from the customers.
In addition to just providing supply chain financing to our SME, we're able to now do cross-buy right across a broader set of customers and thereby getting greater value from ecosystem partnership. Looking ahead, Managing Through Journeys covers more than 61% of CBG and IBG revenues. It is pervasive across the entire bank, with 69 MTJs live, covering more than 6,000 customers, more than 6,000 employees, and SGD 8 billion worth of revenues. More importantly, the new way of working with PCs, the Workflow Workbench at the heart, and the whole Flywheel, will enable us to drive growth through MTJ. Secondly, the data and AI business impact has been tremendous.
We've been able to double every single year since 2020, when we started to measure, and now it is pervasive across the entire bank, and with the ADA and as well as ALAN architecture that is being set up, which you will see in the showcases. It will give us a very good platform to incorporate any new generative AI models that are coming through. We believe that AI will help generate an incremental SGD 1 billion of economic added value for us over the next 3-5 years.
Finally, on ecosystem, you have seen the financial impact that we have achieved over the last three years, and more importantly, we are growing the number of partnerships, increasing the number of products that we will infuse into these partnerships, and this is also pretty much a BAU in the way we run our businesses in the country, as what Surojit will share, and as well as in CBG and IBG. Ecosystem partnerships will also give us a lift of $1 billion in the next three to five years. In conclusion, these three new ways of working, leveraging technology that we have, will result in the uplift of our trajectory in both revenue growth and as well as productivity gains. Thank you.
Hi, good morning, everybody. Today, I'm here to share with you how we have transformed credit risk management. In a typical credit process, it usually starts with a customer requesting for a credit facility, and the credit facility is usually processed internally via a credit memo. So, once this credit memo is approved, it will actually be activated and then available for use by the customer. So yeah, post disbursement, what we will see is that the bank will monitor it for performance. This is the typical origination, underwriting, and portfolio management aspects of credit risk management. So how have we reimagined credit risk management? The enablers are the cloud-native architecture, nontraditional data, and AI/ML. And what are the outcomes that we have achieved from all these enablers?
First, we have customers increasing by 11%, from 11 million to 12.3 million, from 2019 to 2022. So 2019 being the time that we embark on this journey. Second, loans increased by 16% from SGD 362 billion-SGD 420 billion. NPL rate has improved from 1.5% to 1.1%. Next, let me share with you how cloud-native architecture has given us efficiencies at scale. So what we are able to build is a unified platform for which we can define 17 unique personas. For example, relationship managers, credit risk managers, as well as credit control personnel, and we are able to integrate all these roles across the entire platform.
So we are able to generate flexible and collaborative workflows, resulting in 48% reduction in the time taken to prepare a memo. And we are also able to automate financial spreading and financial analysis, and this has given us a 30% reduction in the preparation time for financial analysis. And because we have this unified platform, we also have digital connectivity from the front to the back, and so the back end has also benefited from this. So what we are seeing is a 71% STP for limit creation at the back end. Another added advantage of having this unified platform is that all the data that we have collected during the analysis phase, whether these are the nontraditional data or the traditional data, they are then able to be used in digital formats. So this data tends to be well-defined, well-organized, and ready for use.
In this case, they actually give us a lot of advantages in origination, such as underwriting new segments and markets, as well as strengthening our fraud detection for credit applications. Also in portfolio management, where we are able to deploy advanced analytics to drive early warning. Let me now share how we have powered new origination using data and analytics. In some of these new markets, we do have customers who have inadequate traditional credit data. So for example, customers who lack a credit history, but have a very, very rich digital footprint. Now, and customers who are also having untimely financial information, for example, as well as new economy workers, who are most likely freelancers and not full-time employees, and therefore they lack full-time employment data...
What we are able to do is, we are able to then combine traditional data, such as payslips, credit bureau data, tax information, together with bank transactional data to create credit scoring models. We are also able to use buyer information and lifestyle information to create income estimation models. We can also use supply chain, transactional data, as well as inventory information, to effect drawdown transaction validation. By drawdown transaction validation, what I mean is that we are able to assess the inventory level before we have to disburse a loan. This creates a very, very significant credit control for us when we are disbursing loans. So how have these models helped us? It has created the capacity for growth as we access new segments.
So in the consumer banking ecosystems, we have instant credit decisioning via real-time APIs, and we are able to scale consumer lending with minimal incremental cost. So what we have seen is the number of applications approved has gone up by 16 times from 2020 to 2022, and we have seen new volumes increasing by 28 times. In the corporate banking ecosystems, we now have the ability to onboard thousands of suppliers within minutes, and we are able to effect end-to-end digital design with very minimal manual intervention. And we see the number of anchors and platforms increasing by 1.8 times, and we have seen the book size grow by 2.5 times. Amidst all this business growth, we also recognize the need for us to strengthen our fraud management, for new credit applications. And so what we have done is exactly the same thing.
We basically continue to leverage on traditional data, such as credit bureau, as well as the rule-based detection mechanism, but we supplemented it with non-traditional data, such as the device ID and facial recognition, and we're able to deploy new tech to do electronic verification. So we're able to detect bot attacks, and we're able to identify multiple applications and geolocation spamming. What we then do is we overlay it with AI ML models that are capable of detecting false positives. And what we have seen is a 12% reduction in false positives and a reduced identity theft by syndicates. Let me now share how we have done the same thing in the area of portfolio management. Our early warning models are able to predictively identify for cases for early warning. So we use big data. More than 700,000 network linkages are traced.
More than 13 million payments and remittances are examined. We combine that with our internal early warning model called CRANE, which is able to iterate and learn with additional factors. For example, macroeconomic factors and industry indicators, and this has given us timely alerts for more than 73,000 borrowers. With all these timely alerts, and coupled with our internal monitoring, we are able to reduce instances of sudden death from 1% in 2019 to 0.4% in 2022. By sudden death, what I mean is a case that goes directly from performance status into NPLs without going through a watchlist process. Now, what we have also found is that we are able to improve with the industrialization of the model development. So this one, Kwee Juan just now talked about it.
And what I can do is to show you with an example. In the case of the COVID-19 moratorium, because the SME's cash flows were very adversely affected, we were actually able to enhance the existing CRANE model with a cash flow model that was developed within three months, and that can achieve a 95% accuracy in flagging potential MPAs, as well as 60 days delinquent customers. And what we have done is, we have identified more than SGD 80 million of at-risk exposures. So in the previous slide, the question will be, how did we develop all these early warning models?
So the models that were developed for proactive monitoring first leverage on the ADA that Jimmy talked about this morning, whereby it is filled with and ingested with traditional data and non-traditional data, and this is a very rich database for which we can extract features that are predictive for early warning. By deploying risk analytics methodologies, we are able to produce scoring models that are granular and segmented. So we can produce models for small enterprises, we can produce models for medium enterprises. And what we do then is we overlay it with customer linkages to produce a list of at-risk customers, and these at-risk customers are then subject to a more deep dive and internal credit review process. So in the previous slide, what do I mean by customer linkages? That refers to network-based analytics that captures risks that are very difficult to identify manually.
So in the left-hand side of the slide, you can see that what we do is we use transactional data together with the risk factors from the models, and then we overlay it with network-based analytics to flag at-risk entities that provides sharper insights. So as an illustration, if you look at the diagram in the middle, you can see that Auditor A serves multiple borrowers with high and medium risk. What this tells us is that potentially, even some of the existing low-risk customers may actually be at-risk borrowers. Similarly, if you look at PQR Private Limited, you can see that it has a high supplier concentration from ABC Private Limited and XYZ Private Limited, which are high-risk borrowers. So what this tells us is that potentially PQR can be susceptible to the failures of ABC or XYZ.
And so through this network-based analytics, we are able to identify 44 instances of potential roundtripping financing. In the last 10 minutes or so, I have shared with you all how we have transformed the credit risk management. What we will do looking ahead is to continue to power productivity through our cloud-native architecture, and to continue to penetrate new markets through our scalable origination capabilities, and to sharpen tools to drive portfolio effectiveness through advanced risk analytics. With that, I thank you.
Very good morning to all of you, ladies and gentlemen. My name is Shee Tse Koon. I'm the Group Head of Consumer Banking and Wealth Management. I'm gonna talk to you about three things today. In the first section, I'll talk to you about what or where we have left off since our last Investor Day in 2017. At that time, we specifically called out our consumer banking and SME businesses in Singapore and Hong Kong. We made a distinction then between our digital customers and our traditional customers, and we showed that our digital customers have outperformed our traditional customers. So in my first section, I'll show you what we have achieved since then. In the second section, I'll then move on to tell you about what we have done, how we have gone about delivering those, that outperformance.
The third section, I'll talk about where we then go next. After this outperformance, what other opportunities remain for us in Singapore? Let me first move on to the first section. Now, a quick recap of what we said in 2017 to the market. This was our thesis, that digitalization has led to superior financial and operating performance. We told you how we have changed the business along these three pillars, A, T, and E: acquire, transact, engage, which you have heard from Kwee Juan earlier on. How we acquire customers through ecosystem, through a data-driven digital marketing, and through digital onboarding, and how we continue to drive straight through, straight through processing, paperless processing through our digital channels. And how we engage our customers through being embedded in the end-to-end customer journey, and at the same time, use data for contextual marketing.
And as we go about driving our transformation, our business model, in this way, we believe that our digital segment, our digital customers, will deliver superior returns in lower cost-income ratio, higher ROE, and that these digital customers will transact more with us and be more engaged. And in so doing, likewise, they will deliver a higher income and a lower cost to serve. This was our thesis then. So what have we achieved since 2017? You've heard from Piyush earlier on, that we have transformed or driven adoption, such that now 60% of our customers are deemed digital customers, from 33%. From 1.9 million to 4.1 million customers during this period. We mentioned at the time that these digital customers are stickier with us, that they will have a broader and deeper relationship and engagement with us.
I'm happy and delighted to report that if you look at the results here on the top left of this slide, the product holding multiples between digital to traditional customers, where we told you in 2017, was at 1.6x in 2015. We started tracking in 2015. Now, this thesis continues to hold true today. In 2022, this product multiple has continued to grow from 1.6x now to 2x. If we then move on to the right, top right of this slide, this has then translated also to incomes, to real incomes. In 2015, when we started tracking, the income multiple of digital versus traditional customers was 2x. This has continued to grow through the cycle. At the end of 2022, it is now 3.1x.
If I orient you now to the bottom left of this slide, the cost-income ratio. Piyush mentioned that from 2015 to 2022, our cost-income ratio has improved from 49% down to 40% at a bank-wide level. Happy to report and highlight that the difference in cost-income ratio between our traditional customers and our digital customers has continued to have a gap. Right? This gap has shown from 2015 to 2022. Now we can see a 66% versus a 34% in 2022, from 58% versus 40%. And when we wrap it up from an ROE perspective on the right, bottom right of this slide, we can see that in 2015, there was a gap of about seven percentage points. 25% ROE for digital customers versus 18% for traditional customers.
Now, this has stayed through, stayed true throughout the interest rate cycles, up and down, and in fact, by 2022, that gap has now widened to a 15%, a 15% difference. And actually just want to call out 2021, where interest rates were at the lowest, actually, the gap widened, the gap widened even more to 18 percentage points difference. So we can see that our digital customers do indeed yield us superior returns, and it is very, very true through the interest rate cycles up and down. So our thesis holds true today. Let me now move on to the second section of my presentation. So what have we done, and how have we gone about driving these superior returns? How we have enhanced our digital proposition to reduce cost, to grow our revenue, and also to gain market share.
Now, let me use Singapore as an example to illustrate this. In Singapore, actually, we were already very dominant, as you all know. At the same time, there were new digital players, digital, banks, fintechs, that have also come into the scene. Now, as we observe our customer behaviors changing to adopt more digital, and with these new players, we have completely reimagined our digital proposition. We have built a comprehensive digital proposition in Singapore that's truly distinctive because it covers all segments, it covers all full range of products and services, and it is truly embedded in the customer journey end to end.
And therefore, despite the fact that we were already dominant and there are new players coming in, we have not only been able to retain our dominance, our market share, we have in fact been able to stay relevant and to penetrate new segments that we were unable or that we didn't do so in the past. On the left-hand side of this slide, this illustrates what we've done in our consumer banking space. We have very consciously adopted a companion app strategy. With our digibank app, it allows our customers to do a full range of what they have to do with their banking, whether it is their account services, right through to larger payments, larger value payments, to their financial planning and to their digital investments.
And then with the everyday lifestyle app in DBS PayLah!!, this allows smaller value, higher volume, more everyday kind of requirements they need to pay at the hawker centers, to book their movie tickets and so on and so forth. So we can see how we have designed our digital proposition to cover all aspects of our customers' journey, whether it is day-to-day or episodic. And this is through a single sign-on, interoperable between these two apps. On the right-hand side, you see our SME proposition, where through DBS IDEAL, our SME customers are able to initiate their transactions, they are able to collect their payments, they are able to make their payments, to receive information for their reconciliations, to book their FX. They can apply for loans. They can open a new account through our PWeb.
So as you can see, this is really a very comprehensive digital proposition that we've put in place for our customers here in Singapore. This is how we have stayed distinctive, stayed relevant, and continue to gain market share. Let me now dive down, dive into a bit more details to give you a flavor. Now, our digitalization has lowered our cost to acquire and cost to serve. On the leftmost of this slide, in our digital onboarding journey for consumer banking, we have spent the last few years continuing to enhance this digital onboarding journey. For example, we were the first bank in Singapore to launch facial biometrics onboarding of customers, leveraging Singpass to access our national identification agency.
We have continued to improve our journey, and now, with the lowest, the smallest, the shortest number of clicks, we can open an account in Singapore, and you can see the results for yourself. In 2017, 83% of our consumer customers opened their new accounts at the branches and only 17% through digital. Now, we've been able to drive this rapid adoption of digital onboarding, and by the end of 2022, 93% of all our newly onboarded customers are done through digitally and only 7% at the branches.
Now, if I move you to the middle of this slide on digital transactions, we have continued to leverage data capabilities to analyze our end-to-end processes to see where customer process breakdowns are, and we have enhanced our capabilities to allow customers to initiate more seamlessly, to access more information through our digital assets, through our digital banking. And that allows transactions, therefore, to come through more digitally. Customers have adopted digital means to transact with us. You can see the active rates of all our digibank customers now growing from 66.8% up to 78.3%... and you look at the massive 154% growth in the digital transactions that come through us now. So transactions have grown exponentially, and a lot of them has now come through digitally. What does this all mean?
If we look at the right-hand side of this slide, this has led to a massive 60% improvement in our cost to acquire and a 45% lower cost to serve per transaction. The rapid adoption of digital has also allowed us to transform our network and the way we service our customers, and this has led to further reduction in cost. If you look at the left-hand most of this slide, this shows the network or the touchpoints, the branches that we have in Singapore. We have what we call the full-service branches, which is what you will probably be familiar with, where you go in, and there are counters and tellers behind the counters serving the customers, and there will always be financial planning consultants or advisors there for to give advice to customers who are looking for investments.
Now, in 2017, we had 84% of all our touchpoints being these kind of full-service branches, and only 16% being self-service branches. Now, these self-service branches are not just ATMs. We have deployed, even at that time, what we call VTMs, virtual teller machines. Now, these are machines from which actually the tellers are set in our call centers. So when a customer presses a button, sticks the cards in, a real, not an avatar, a real human being appears on screen and can give guidance and help the customers through the transaction, just as a teller would have done behind a counter, but the fulfillment is then done through the machine. So we already deployed these machines. We were the first in Singapore to do so in 2017, but it was only 16% of our touchpoints that had these self-service capabilities.
Now, because of the rapid transformation or rapid adoption of digital, we saw a 67% reduction in branch arrivals during this period. Now, this has allowed us to reimagine and reconfigure our network, and in 2022, you can see now only 46% of all our touchpoints are now full-service branches, and the rest of them are self-service or financial planning branches. Financial planning branches meaning they have self-service machines, VTMs, and also some human beings to provide advice. In the middle of this slide, this is our call center. Now, likewise, we have used our data, our AI ML capabilities, to preempt, to proactively predict and intercept calls. So we're able to send information to customers because we anticipate it, so that they no longer need to call. We also use this to enhance our chatbots.
You can see, likewise, between this period from 2017 to 2022, that the main calls have also reduced from 5.18 million down to 4.11 million. Now, as we do these things both at the branches and at our call center, apart from a reduction in cost, it has actually also enhanced our customer service. Because now our customers are able to access our branch services beyond banking hours, because these VTMs are technically there 24/7. At the same time, likewise, as we preempt these calls coming in to the call centers, this is also enhanced customer service as well. So it's not just about reduction in cost, it is also an enhancement of our customer experience. And what does this all mean from a financial aspect?
If you look at the right-hand most, our call center, sorry, our call center and branch costs have been lowered by 26% during this time. Let me now move from the cost lens to the revenue lens. Some of you would have seen the showcase with EV on the left-hand side, and some of you will see it later. So for those of you who have seen it, you have probably seen a lot more, but let me take this opportunity to explain what we've done. Wealth management services has traditionally been availed only to high-net-worth individuals through our AMs. But through our digital transformation that includes data, AI ML capabilities, we have been able to democratize wealth, and now driven wealth management into a new mass retail segment and allow them access to wealth management. On the left-hand most of this slide, this shows our DBS Digital Advisor.
Let me do a quick introduction of what this is. This is really an online, a digital capability, which allows our customers to do their financial planning, to do their budgeting. It's a financial planning tool that is online. This allows them to do their financial planning, including their budgeting, setting their goals. This allows them to do their own risk profiling, which in the past used to be done on many, many pieces of paper. This is now done online, automated. At the same time, through this platform, we have also accessed open banking in Singapore. In Singapore, multiple agencies, public, private agencies, have come together, shared data through a platform called SGFinDex. And in this SGFinDex, they are able to pull data from public services such as CPF, the pension fund, IRAS, the tax authorities.
They're able to pull information from private players such as participating banks, as well as insurance companies. So through this digital advisor platform, our customers are able to consolidate their entire individual balance sheet. So not just their holdings with us in DBS, but all their holdings across various banks, their pension position with CPF, their borrowings from HDB, which is their mortgage loans with the public housing and so on. And with this consolidated view, together with the financial profiling they would have done in point one, we then use our AI/ML capabilities to send hyper-personalized nudges to them. Now, this includes content that they can read or they should read, and this includes potential investment recommendations or investment recommendations that they might want to consider.
So with these hyper-personalized nudges to the customers, they are also able then to fulfill their investment decision. They're able to do their investment now also through this platform. They can buy unit trust, they can increase their savings, and so on and so forth. Through this platform, likewise, we have availed for them what we call digiPortfolio, which is our CIO views, which is now availed to the retail clients, that through this CIO view, with a technology behind to help rebalance their portfolio, they have now these bite-sized chunks, investments, solutions that they can also avail themselves to. And finally, point six there, which is online to offline. When they need advice from a real human being, through this platform, they can make an appointment, they can ask for that advice as well.
So this capability that we have put in place has helped to democratize wealth, and you can see right at the bottom there, 3 million customers have been engaged through 280 million nudges during this time, resulting in 2.7 million transactions, generating $47 billion of trading volume. And in more recent times, as we continue to enhance this and put digital insurance on it as well, we have seen a 6x increase in premium for digital insurance. Now, what does this all mean, again, from a financial perspective? If you look at the right-hand side, this has resulted in a higher investment AUM between 2019 to 2022, at a CAGR growth of 20%, and a higher investment income during the same time, CAGR growth of 13%.
Bottom right of this slide, you see the market share that's relevant to this. We have already been number one for quite a number of years in bancassurance, but we continue to see that growth from 39% to 47% in our regular premium market share. And this CPF market share, this is really about the pension funds from which our customers can use to make investments. Likewise, through our digital journeys, our customers have been able to link their SRS and their CPF IA investment account to us, and from there, this balance sheet's consolidated. This can be used as a source of funding for their investments. We've got 62% and 44% now of this market share, respectively, which is also a growth, as you can see from here.
Apart from just the mass retail, we have also adopted this concept in more recent years for our mass affluent customers, which we call Treasures. In 2017, our RM or our clients per wealth manager for this segment was 500. So one RM handled 500 customers. Now, to be fair, it would not have been very effective for one RM to engage with 500 customers. And for this mass affluent segment, they are really higher net worth, they do need RMs, unlike the mass retail that we talked about earlier on. But for one RM to engage 500 effectively would not be that realistic. So what we have done in our transformation is to enable our RMs through technology, through what we see on the middle, the top of the middle, middle part, which is sending them nudges as well.
So you can see in 2022, 96% of our Treasures customers now receive these nudges, and on average, each customer receives about 100 nudges every year. Now, this has enabled our RMs to be a lot more efficient and also more effective. So now, in 2022, one RM can handle 700 customers, so they are more efficient, but actually, they are also more effective. Because if you look at the bottom, bottom left, the productivity per RM has grown from just slightly less than SGD 600,000 revenue per RM to now SGD 860,000 per RM, so a 44% increase. And the percentage of customers who have now started to own wealth products with us have grown from 34%-46%.
Again, from a financial perspective, if we look at it from the right-hand side, there's a CAGR growth of 11% in higher Treasures investment income. Beyond just the cost and revenue that I've talked about, our digital transformation in Singapore has also allowed us to partner with ecosystem partners outside of the bank, so this enables us to reach beyond the bank. On the left-hand side, you see our mortgage, our property marketplace. So we have launched this property mortgage marketplace, linked in with property players out there, and start to originate our mortgages from this platform. You can see in 2020, that the percentage of mortgage sales through this platform of 5.7% of all our mortgages originated, is now grown to 16%.
The mortgage loans that have been booked have grown 2.7x up to over $2 billion now, and we continue to dominate the market in mortgages with a 28.9% market share here. Now, moving to the right-hand side of this slide, this is our DBS PayLah!!, the everyday lifestyle app that I talked about. It really started as a peer-to-peer wallet for people to pay one another. Over the years, we have further embedded this platform into the lives of our customers by bringing on board a lot more ecosystem partners. Many, many more merchants. Now, through this platform, you are able to book a transport, whether it's a private taxi or a public taxi.
You can book your movie tickets, you can order your food, you can go and scan and pay at about pretty much all the hawker centers, 12,000 hawkers around Singapore and so on and so forth. So this has really driven ubiquity and made us very, very pervasive in the entire Singapore market. On the top there, you can see our average monthly logins now have grown to 23 million, and they are now 2.4 million as of end of 2022, a 3x growth in the number of PayLah!! customers in Singapore. And more importantly, at the bottom, a massive 96x growth. 96x growth in P2M payments, payments to merchants, and this is testament of the success we have achieved in bringing more and more merchants into this space and being really pervasive in our everyday lifestyle of our customers.
Now, this has really driven a much stickier relationship we have with our customers, and so you can see on the right most of this slide, that our deposit market share, our Singapore dollar deposit market share, which was already high at 52%, has grown to 54%, as Sophi mentioned on earlier on in April this year. This is really about being relevant, being pervasive, and our cards market share has continued to grow as well from 18% to 22%. Let me now move on to SME. I've spoken a lot about consumer, but what about SME? Now, we already have a strong and dominant SME market share in Singapore as well. But SMEs is actually a very broad spectrum, so we still have some of the really bigger SMEs, so to speak.
But for this slide or the next two slides, I really want to cover the smaller SMEs, those who have annual turnover of less than $10 million. So in this segment, these are really the smaller SMEs, including what we call the micro SMEs, which many fintechs would say that they are going after, is their target market. Through our digital transformation in SME, we are able to also access this market or this segment much more effectively and much more efficiently. Now, along these three pillars, again, A, T, and E. Now, on acquire, 99% of all SME accounts, operating accounts that are open today in DBS, are open digitally. And likewise, we leverage the same facial biometrics, we leverage the same API calls to pull data from ACRA, which is the registrar of companies.
This has allowed a very seamless, very easy way of onboarding SMEs digitally. Now, these customers, below annual turnover of $10 million, typically do not have RMs. And that's why the ability for us to acquire digitally is so important. And you can see also from here that we have continued to enhance our journeys, such that the straight-through processing of these account openings have now grown from 13%-53%. Now, on the bottom left of this slide, this talks about our digital loan application. Likewise, we have reimagined the entire loan application journey for these customers. Again, you know, these are really smaller value loans. So to acquire these loans and underwrite them manually, is just gonna be uneconomical.
But we have been able to originate, to allow them to apply for these loans, underwrite these loans, approve these loans, get the letter of offer, accept it digitally, and then finally disburse digitally. All of these have grown now from 33% of all of these working capital loans being done digitally from 33%, we've increased that to 48% now through reimagining all these journeys. Now, in the middle of this slide, you can see the way our SMEs transact with us. We can see this DBS IDEAL activity has actually grown with the logins growing to 28 million now, a massive increase, and the active rate is now 91%. Now, through this, our customers initiate their transactions, they can get their reconciliations, as I talked to you about earlier on.
On the engagement front, through DBS IDEAL as well, through our AI, ML capabilities, we have now been able to nudge our customers, and now, on average, they get about one nudge every month. You can see now 80% of these customers who do not have RMs, who traditionally probably would not have heard from the bank, will now be receiving regular nudges from the bank. Now, what does this all mean? Now, in this slide, you can see on the leftmost, our loan balances, just for this particular segment of customers, have actually grown 4x between 2017-2022. I'd just like to remind you, this was a segment that we traditionally were probably not as effective in penetrating. We have now been very successful in this.
On the middle of this slide, we talk about the digital engagement, and some of you have seen the DigiMarkets on the right-hand side as well. With some of these not just bringing together our colleagues from treasury, from FX, we have seen an increase in fee revenue per customer from FX grow 24% CAGR, 24% during this time. And on the right-hand side, the revenues for this segment has grown 48% between 2017 and 2022. Now, in particular, I want to call out the some words there on the bottom right-hand, which is the market share, particularly in the micro segment within this segment. We were at 18% of these loans, now we've grown to 34% during this time. So my third segment, where are we going from here?
I hope that I've illustrated to you over the last 20-odd minutes, how successful we have been in driving out performance in this digital segment. But where do we go from here? We believe that we can still continue to grow our customer wallet share. Let me explain. If you look at the left-hand side of this slide, it shows us that even as we have been very successful in driving digital adoption, but this all didn't happen in one day. This happened over the last few years. As customers continue this digital journey with us, we do believe that they will continue to be stickier, deeper, broader. But as we speak, only 64% of this digital segment today... Or, or sorry, should I say 64% of this segment today still hold only one product with us, right?
With the rest holding two or three or more products. So this actually means a lot more opportunities for us as we continue to nudge them and engage with them, we can push the 64%, or we can bring the 64% to deepen further with us, to broaden further with us. So that's the opportunity that remains for us. If we look at the middle of this slide, within this digital segment of customers, 87% of them still do not hold either an investment or insurance with us. While we have seen, we have seen those below 12% and 1% holding either investment or insurance, or some holding investment and insurance.
So this is really the next journey for us, how we continue to guide them through their financial planning, how we guide them through their retirement planning, in order to deepen that relationship with us in this I&I space. So it is about the opportunity both to broaden and to deepen. And on the right-hand most of this slide, we have seen encouraging growth, encouraging adoption of investment and insurance from those that we have been engaging over the last 12 months or so, a 7% growth during this time. So this, to me, is where we are going forward, the new opportunities that remain with us, notwithstanding the fact that we have already driven such massive digital adoption and superior outperformance in the digital segment. So finally, ladies and gentlemen, this is DBS digibank.
Truly distinctive, because it's comprehensive, it's left no segment behind, it's left no products and services behind. It is truly embedded end-to-end in our customer's journey. I hope that I have managed to convince you that our digital outperformance continues, that we have a robust digital proposition that has spurred growth reductions, revenue growth, and dominated market share. And that finally, there remains opportunity for us to deepen and broaden with our customers to grow our customer wallet share.
Good afternoon, ladies and gentlemen, my name is Joseph. I'm the group head for the private bank. I'm standing between you and lunch, so I'll make it snappy, and, hopefully to, usher away your hunger pains. I want to share with you today about our transformational journey over the last several years, where we had really redefined the way our customers have engaged with us, and in that, delivered better client outcomes, better RM efficiency, and better revenue for the DBS Bank. Let me just flick through quickly on the slide, which is about AUM. Before I go any further, let me just explain the full length of that bar is about our wealth segment. The pink bit at the top is really about the mass affluent that Ziku was talking about earlier.
What I'm going to be talking to you about is the red solid bar. As you can see, our AUM has been going up every year for the last several years. It has been remarkably consistent, and remarkable it has to be, given the massive gyration we've seen in clients' assets for high net worth individuals over the last several years. You may remember, briefly, the pandemic in 2020, with a massive downdraft in clients' assets. Also, last year, where major equity markets and fixed income markets also had a massive downdraft, losing between 15% and 20%. So it is remarkable, despite all those road bumps, that our AUM continues to go upwards every single year.
The magic of that is we have been getting record net new money inflows into our business for each of the last several years. Now, this is not just from new clients who are attracted to our platform, it's also from our existing clients who have now moved more assets to us and switching to make DBS Private Bank their key wealth advisor. This is an important point to note, and this really speaks to our very, very differentiated client engagement model. Let's just go down one more level on our financials.
You can see there, we certainly had more than our fair share of new high-net-worth clients, but more importantly, the highest end of the ultra-high-net-worth market, where the single family office sits, we have actually emerged to have a dominant market share over the last 3-4 years in Singapore, where 1 out of 3 single family office are banking with DBS Private Bank. That drives our income, obviously, to new heights the last several years. But more importantly, the quality of our income, the recurring revenue, has doubled over the last several years. Now, this speaks to a higher client engagement, high quality discussion with our customers, and to the quality that our client is looking towards us to manage their wealth more strategically over the long term.
As you said earlier, this translates to a continued low cost income ratio, one of the best in the industry, despite our growth in our business. That of course translates to a high ROE in our business as well. You can see in the bottom right. Our continued year-to-year outperformance has propelled us now to be a top four private bank in the Asia Pacific. This is important, because it means we've got a lot more AUM than we used to have. But far more importantly, the other half of our success story is the return on AUM. We are far and above the Asian average and our global peers.
This speaks very much to the high quality of our client engagement, how our clients are engaging with us, not just the sharing of the heart, but of their mind, of their share wallet, where they come to us now with a highly engaged, very, very involved and about their long-term investment with us, making them very sticky and also very engaged with us. You can see the difference there on the board. What has been the three key drivers of transformation over the last several years? On one hand, we have invested heavily in systematically training our people to make sure that domain expertise continue to rise with our sophistication of client base, looking towards them for solutions.
We've also improved our RMs' strategic advisory capabilities as an art and a science to measurably engage clients and help them to plan very complicated, multi-jurisdictional wealth planning structures. The middle part of proposition, we have made a decisive shift from a transactional-based business to a portfolio, long-term driven advisory business, capturing much more to share with our customers who want to really develop a long term and build their wealth beyond their generation with us. Secondly, we've also expanded our advisory process to include not just investments, but legacy planning for our clients and their business succession, but also including about more sophisticated of their, of their values, of their philanthropic aspirations, and build it into a holistic solution for our clients.
But the biggest, biggest part of our transformation has been how we take the best of a traditional RM-client relationship, built by emotional connection, intuition, and have then augmented it with our proprietary digital architecture. We call that phygital. What does that mean? It means now we've armed our relationship managers with data and digital tools that allows them to use the power of AI and ML to come through with structured solutions to clients on the run. It also means that we're now sending our clients systematic, timely, highly personalized, and actionable insights and actions, which is alignment with not just their profile, but with their preferences up to date.
What it means now is that clients can now see their portfolios in real time, and they can manage it together with the relationship manager dynamically to adjust their portfolio according to market conditions, giving them a peace of mind and a sense of very strong control over their destiny. So just going down the memory lane for a little while now. Traditionally, conventional engagement between the RM and the client has been very much paper-based, where the RM has to take a long time to find information, different systems, different sources of data, to come up with ideas to help the client to take the next step. This takes time, but it also means the advice given to clients traditionally has been fairly much generic and one-size-fits-all.
Then that will push the client to make sure they have many, many conversations to fine-tune, to zero down on the right conversation and the right solutions. And at the end of all that, the only way a client usually to transact is through the relationship manager, through multiple systems, causing a lot of latency between decision-making and execution. Then comes our phygital model. This is the model using the data that we have, using AI, ML, driven, powered advice, intuition, to systematically push out prompts to both our customers and to the RM, sometimes independently, sometimes synergistically, sometimes symbiotically. This allows customers to have very, very digestible, real-time, and relevant insights that the clients themselves, if they chose to, can transact immediately. The relationship manager also understands what's being sent to the customer. They have a far more contextual view.
It allows the RM and the customer to come together for more strategic discussions that are prompted by these nudges to have a more prepped, nuanced, and informed discussion, where decision now can be arrived far quicker, and the client can choose to implement or to have further discussions about other aspects. All this, by the way, gets fed back into the system as a new, yet new data point to enhance the further synthesis of the client's nuances, to allow even better nudges to come through going forward. And the circle, virtuous circle, then repeats itself. Just a little bit deep dive into this engine. We have over 16,000 data points for clients, which includes, by the way, not just the static data around their preferences, around their recent investment activities, but also around how they responded to recent nudges.
Their browsing history, the RM's conversation with the client captured in those notes, are all being now synthesized to match together with real-time data, our research house, of course, our recommendations through 30 AI ML models, to come through with hyper-personalized nudges that goes either to the client or to the RM, or concurrently. I want to illustrate all of this with a couple of very simple but powerful use cases. The first one is capturing the customer's holdings in equity. Our system actually prompt the client when a particular holding they're holding in their portfolio, which is either 52-week high or 52-week low. It's a simple nudge. It's a powerful nudge. It reminds clients that they're holding this particular equity.
They look at their push notification, straight away can see what's the holding they have in their portfolio, what was their cost price, what is the market price, do they, do they act or do not act? Do they add to position? Do they, do they average up? How does, how does it look in the context of the entire portfolio? And if the client chooses to take action, he can press the button, and it can transact it on the spot. Very powerful nudge, which our clients find very useful. Similar use cases are for foreign exchanges, when client uses that to transact at a favorable rate, when they want to do a, a transaction for, for regular purposes. A more powerful nudge, it's the second type of nudge, what we call the awareness nudge.
This is where clients who may not immediately take to an idea, particularly around strategic portfolio management. So, for example, this is a call that our CIO made in 2022 around gold, and gold as a portfolio hedge. For some clients, the idea didn't catch on immediately. The system provides a series of nudges to the clients, showing different types of usage for that, particular whether it's commodity angle, whether it's a portfolio, volatility mitigant, whether it is a tactical trade. Over time, the client gets to understand what the call means.
At the same time, by the way, the relationship manager, through their Client Connect device, sees straight away that prompt to the client, look at the client's client risk profile to make sure it's appropriate for that nudge, and also have a sense of how engaged that client's around the different research that's being sent to the client, in terms of the engagement, the time they spend those research. Get a much better sense how the client's feeling about that particular call. The RM can also straight away deep dive into look at the sales model portfolio, look at that particular client's actual exposure in all the asset classes, to make sure that gold call is relevant and how it would apply for that client specifically. Now, at that time, both the client is now primed, the RM is now prepped. They have a conversation.
This has actually drove some clients to actually take up the trade to hedge their portfolio, or in some cases, at least it trigger a thought in the client's mind about what hedging portfolio meant. By the way, by the end of 2022, the gold trade was the only asset class that not lose value last year. We all know equities lost 20% last year, fixed income lost 16% last year. So it was the right thing to prompt clients to do. And I think you guys would also know that gold has continued to move up, but it's done its job to trigger the client to think about how to protect the portfolio in market volatility. Just to summarize, really do that, nudges for certain clients. We do a nudge for every single client, hyper-personalized.
We actually triggered 1.4 million nudges last year across not just investments, but the client's legacy planning around the certain asset classes and around ESG that they show interest in overall. The client outcome has been very prominent for us. It allows clients to be actually managing a portfolio in terms of there's no, no cash lying around they want deployed. It tells clients when to their profits and where their losses in their holdings, so they will take some action around what they need to do rather than not seeing it. One of the most powerful results of these nudges was around legacy planning. Legacy planning means the client to plan for the succession of their asset, the part of their asset to the next generation. We all know that clients tend to not to deal with that until tomorrow. That's the reality.
What we have done is we send a series of nudges to the client, piquing their interest about what wealth planning means, what legacy planning means, over a series of nudges... And again, the arm gets to see the client's reaction or the client's engagement on those nudges, those research. Over time, when we felt the client is primed, a conversation is had. As a result, by the way, many more clients have taken their first couple of steps to do the planning for their legacy. This is a breakthrough, because we all know advice is good, but advice is only better when the clients take that advice on board.
So to summarize, we have found a space we call phygital, that really allows the best of a human-client interaction, human intuition, emotional connection, together with AI ML-powered insights, that really helps the client to take to the next level in how they manage their portfolio. So they can really articulate, and they can execute their long-term planning in their portfolio. So to summarize, we feel the best way to meet the client's ever-evolving needs in a fast-changing world, is to fundamentally change the way or redefine the way our clients engage. And in doing that, we've already found that to be driving improved client outcomes, improve our efficiency, and, of course, an uplift in our financials. With that, I thank you.
Good afternoon, everyone. I have the distinct honor of being your speaker after lunch, and my job is to keep you awake, energized, and excited about how our digital transformation has enabled the success of our transaction banking business. We call it GTS in DBS. My name is Su Shan, and I look after the IBG Corporate Banking, Institutional Banking business. In the next 20 minutes, what I will do is, as you're all analysts here, I'm gonna show you the money. I'm gonna start with a slide that shows you the outcomes of our digital transformation, what it means for our cost-income ratio, what it means for our ROE, what it means for being asset light and making returns.
Then I want to share with you, really, the depth of this is that we believe that there has been a paradigm shift in the world of transaction banking. We believe that the transaction banking of the future, you know, is really gonna be transformative. The success of transaction banking will be enabled by banks that can enable their customers' own digital transformation. It will be enabled by banks who can give their customers real-time solutions, real-time transactions, scale, speed, and at a cost that makes sense. I will unveil that paradigm shift. I will throw in what we see in the demand and supply chains in the world.
I will go through the pillars of success, what we've built upon our APIs and microservices, and then we'll do a deep dive on cash and payments, a deep dive on trade, and then I'll end again by showing you the numbers, the impact of GTS on DBS as a whole. And then I'll end... Hopefully, you'll still be awake. I'll end with a video. So here are the numbers. With the digital transformation that we did, we have enabled a lot of customers to do a lot of their cash transactions with us. In fact, as I look around in this room, many a workshop has happened in this very room where you're sitting with all our large corp customers.
And it's interesting, when we go and see them, in the past, it was just the RM and the GTS person, the GTS salesperson with the treasurer or CFO. Today, we bring what we call sales engineers. We bring technicians, technologists into the room to face off with their CTOs and their business heads. So it's really embedding ourselves in the customer's technology transformation. So what does that mean for our business? One, sticky deposits. GTS deposits have gone from SGD 130 billion to SGD 197 billion last year, end of year. Cash mandates, we're winning about 800 cash mandates a year, accumulating in 5,000 new cash mandates end of last year. And transactions, so with the cash mandates come sticky transactions. With sticky transactions comes multiple fees.
So we've enabled 110 million transactions on our own digital platform called IDEAL, and last year, we processed $1 trillion in transactions on IDEAL. I talked about the whiteboarding here in this room. Well, with the APIs we've given to our customers, that leap in API calls, that means customers using our APIs to do their transactions, has gone up exponentially. We started with 7 million in 2019. We are now on course to fulfill over 300 million this year. And what does that mean for our revenues and cost income ratios and ROE? Well, the numbers speak for themselves. GTS made SGD 1.9 billion in 2017. It made SGD 3.3 billion last year, a CAGR of 12%. Cost of income ratio, 38% in 2017. Today, 32%, down 6%.
ROE from 18% up to 38%, up 20%. Again, digital enablement. Most of you would have heard of Greenwich Associates. This is the authoritative survey that we use for large corporates, and here are the numbers, specifically for cash management and for trade. With the digital transformation that we did, we have really won market share in trade, so in cash and trade. So cash is about, you know, 600 customers here being pulled. We were 21% market share, fifth in our markets. Today, we are second in the market and 30% market share. For trade, and this is just shy of 1,000 customers being surveyed, we went from 28% and number three in terms of market share, to 39% and second in terms of market share. So what caused this paradigm shift?
What was going on in the world that made us pivot as we did? Well, first is the obvious. The demand side was changing. Consumers, like you and I, were buying things simultaneously on our mobile phones. Taxi drivers, delivery, you know, delivery truck drivers, everyone were doing simultaneous transactions at the same time. So demand shifted online, demand shifted to the mobile, and demand shifted to multiple, you know, millions of transactions per second. So it's voluminous, and it went from batch. In the past, all the processing was done batch, you know, end of day or after a few hours. Today, it's all real-time, instantaneous, instant fulfillment, instant reconciliation. So demand was changing. The supply as well. Supply chains in the world has been changing at pace, driven by both digital transformation as well as geopolitics.
The geopolitics meant, you heard Biden this morning, diversification, de-risking of supply chains. That speaks to our Asian connectivity. And secondly, with the COVID experience, everybody wanted to be more resilient in their supply chain. Resiliency in energy supplies, in food supplies, in chip or technology supplies. So with that resiliency, it meant you need to have real-time inventory solutions as well. You needed to go from just in time to just in case. And guess what? With the sustainability wave, the ability to track and trace, to be transparent in your supply chain, and the ability to check the provenance, to be able to, you know, to offer sustainable financing along the way, has been a great enabler for us. So here's what's happened. Jimmy talked about our APIs and microservices.
Well, with that backbone of API microservices, we were able to create with our APIs, over 200 APIs, a seamless and easy way to connect with our customers. Either they embed our APIs, and then they can fulfill their own customers' journeys, or they use our own platforms. So by and large, large corporates were using our APIs, as we've shown in the booth. And the three pillars of success that we have here are, firstly, giving them this 24/7 scale at speed at a cost that was cheap enough for them. It was giving them access to these massive product, fast product processes that we give them, from trade to cash, to payments, to loan applications. The second was giving them access to the real-time payment rails. In Asia, we know.
India started with UPI, Singapore has FAST, Hong Kong has Faster, Indonesia has BI-FAST. So together with these real-time payment rails, we were enabling e-commerce companies, insurance companies to do millions of transactions at the press of a button. So accessing that real-time rails, and also for cross-border, being able to optimize your corridor, so you're using the cheapest corridor available, that was enabled by our technology for our customers. And thirdly, intelligent information for our customers on demand. So if you're a taxi company, online taxi company, you want to know where and when to pay every driver. If you're an insurance company, you need to know where and when to collect premiums or to pay for claims. Stockbroking houses, online digital companies, everyone needed this instant reconciliation.
We were able to give them incoming notifications, whether in-app, whether SMS or multiple multimodal instant reconciliation, really saving our customers time and effort. So how do we connect with our customers? Well, I talked about two. First is us giving them our digital assets, and this is device-agnostic. They can go on web, they can use a mobile device, they can use our IDEAL app. And, and we have about 270,000 customers now using us, and my colleagues are really obsessed with the usability, making it easy for them to transact so they can onboard digitally, seamlessly with us. They can transact, they can do their loan application, their, their cash payments, they're paying their employees, payroll, et cetera. Heck, they can even update their addresses, passports, so all the admin, add a new account, et cetera.
We constantly engage them with a feedback loop. You know, Han Kwee Juan talked about that. We constantly engage our customers there. So what is the outcome of this? Number one, the usability. Transaction growth volumes in IDEAL has gone from 60 million to 110 million. That's 80% growth in the last 4 years. We have a checkout, you know, post, logout, pre-logout, web waiting, 4.7% for IDEAL and 4.9% for the mobile, and more than 20 million contextual nudges done so far. So really active, and SMEs actually use this a lot for everything. And then the other is us connecting into the customer's own ecosystem so that the customer's customer doesn't have to leave their platform to fulfill a transaction. And here we have embedded finance, what you call embedded finance.
We are like your Intel chip in the customer's own digital channels. Here, we've enabled companies to set up mass collections, whether it's governments, you know, collecting money from 40,000 families to enable their kids to register for school and collect school fees. It's e-commerce companies. Think about the big e-commerce companies. When you check out, you have multimodal checkout, you can use PayLah!, you can use your credit card, debit card. That web hosted web page is all done by us. So multimodal transaction as you check out. And what's the outcome of this? Well, 32 times growth in API calls. We even have e-wallet companies that need, like, you know, sometimes over a million calls a day, so 7 million to 240 million. And also important to note that we've increased our capacity.
We can now do 60. Most companies do 60 transactions per second. We can do over 1,000, and that's instant fulfillment. So over 300 million API calls this year and making the flow of funds smart. So we were the first bank to offer Swift GPI, which is kind of like a tracking for global payments, but it means you can track where your cross-border payment is going as well. So our channels and our co-connectivity has enabled us for the customers to enable their own digital transformation. Let me now unpeel the cash and payment side, and then the trade side. So I talked about the holy trinity of the architecture of the product processes, real-time payments, intelligent data. Here are some examples of use cases that I want to share. Firstly, the B2C companies.
All of us buy travel insurance, right? Do you buy them at the airport as you're about to take off? Yes. And if your flight gets delayed, normally you have to submit a claim. Today, all the insurance companies that we bank, it's instant claim payment on us, right? The moment your flight is delayed by more than 4 hours, we'll know, we'll pay you straight away. 2 million repayments were fulfilled using our platforms. Government transfers. I shared about collections. We all have the RedeemSG. That was 3.8 billion transactions enabled by us. This is the government vouchers. Many more there. And then payment platforms and aggregators. These are the e-wallet companies, taxi companies, having to do, you know, millions of transactions a day, and to reconcile every consumer payment. And then the NBFIs here, they are like the e-brokers, online digital wealth companies.
Here, real-time settlements is crucial 'cause you have to execute trades in U.S. time, European time, Asian time, and here, again, we help to fulfill their own real-time transactions at scale. What are the outcomes? Well, in terms of transaction counts, exponential pickup from 8 million transactions to 74 million for domestic payments, using those real-time payment rails I talked about. Cross-border, well, we've grown from SGD 3 billion to SGD 24 billion. That's a 7x growth in volume and still growing, like good fintech payment companies. And last but not least, in terms of fees, here's where the fee generation comes in, and it's sticky. From SGD 260 million to just shy of SGD 400 million in the last year. And this does not include, by the way, the cross-border FX. That is embedded in the journey.
Of course, our costs of processing anything paper has gone down. The use of checks alongside the cost is down 19% in Singapore alone. Let me talk about trade. Here's the supply chain. Now, we all know that trade today is moved from discrete nodes now to interconnected platforms. The rise of platforms, whether it's your own anchor, Western MNC, or Asian MNC's own platform, or an open trade platform, like an open supply trade platform, where everyone can connect with each other, like C2FO, or an industry-specific platform, like one that covers, you know, footwear, textiles, like Infor Nexus, or procurement platforms, logistics platforms, like what you just saw with JD Logistics. We put ourselves squarely in the middle of this whole supply chain ecosystem.
So we partner with the platform or the anchor, and with that partnership, we are able to see all the data of the flows. We are able to see the invoices from the suppliers, and that can be thousands of suppliers here. And with that data, and with the flow of goods on the logistics that we can see, even with the inventory, together, we can provide cash collections and loans. With the data we have, we are able to make instant loan fulfillments using what we call alternative lending for trade. So multiple supplier onboarding at scale and speed by the thousands. Multiple onboarding of buyers. So here we can do pre- and post-shipment, we can do supplier payment services, and we can do inventory finance. What are the outcomes?
Here are the outcomes: We have grown the number of anchors and platforms that we bank from 98 to 176. That's a CAGR of 34% in the last 3 years. As I said, the multiplier effect of that on the supply chain financing means we've been able to bank thousands of SMEs that we would not otherwise have banked, and mostly in high-growth emerging markets like China, India, Indonesia. Three point seven thousand to six point four thousand spots in the last 3 years, and very strong sticky growth in our supply chain financing assets from SGD 3 billion to SGD 9 billion, a CAGR of 28%. I will now end with how GTS and the work we've done on digital transformation has really increased the market share, both of IBG, big corp business, and also the relevance of GTS to the group numbers.
I talked about the cash and trade positioning just now. Overall, with the digital transformation, we have been able to increase our market share penetration for most overall relationship for Asia, going from 45% penetration rate to 52% penetration rate. This is a survey of 753 corporates in the markets that we bank, and I think we were one of the few banks that really bucked the trend, and we were tied with a big global bank at number one position last year. GTS contributes a significant portion, 20%. We were 15% at SGD 1.9 billion in 2017. Today, we're at 20% at SGD 3.3 billion. So an increasing importance to the group overall income. And sticky deposits.
We talked about how, you know, with the operating cash, you can really get a lot of sticky deposits and, more importantly, sticky CASA. So deposit growth, 1.30-1.97, two-thirds of that is about two-thirds of that is sticky CASA. And so we talked about the foreign currency mix, while GTS as a group contributes 55% of our foreign currency deposits, up from 51%. So I'd like to end with a repeat of what we just said. You know, the paradigm shift that we have seen means that for us to continue to succeed in transaction banking, we have to continue to embed ourselves and use whatever we have built to win, to win market share, to enable our customers' digital transformation, and to create seamless transactions at scale, speed, and at a cost that's never been seen before.
As long as we're being able to transform our clients' digital business, we will continue to win in transaction banking. Let me now show you a video of what I've just said. Thank you very much.
In a world where customer needs are changing, industries face disruption, and new business models emerge, Global Transaction Services is powering transformation across a wide range of sectors and clients. With our market-leading suite of digital solutions, we help transform customers' journeys by enabling instant fulfillment capabilities. Our digital capabilities have helped businesses become more agile by transforming their customer journeys across industries. Digital transformation across corporate and government bodies has helped improve productivity and enabled reach. With real-time information exchanges, businesses are able to track and manage their operational needs and risks better, streamlining their treasury management functions. Our cloud-based technology architecture enables businesses to scale exponentially with higher throughput and very low latency. In this ever-evolving digital world, DBS is at the forefront of transforming businesses.
...On the revenue, we have a CAGR growth of 8%, compared to the industry of 6%. If you look at the cost income ratio, the market is 67%, and we are 44% in 2022. And then if you look at the return on equity, we are 14% against the industry of 12%. So these are the financial outcomes that because of what we did, which I'll talk in the next couple of slides, what we did that we really get to this outcome. So today I'm going to share with you five things that we have done in last couple of years. Number one, is to digitize our multi-asset pricing and structuring capability. Number two, is how we digitize our product internally and externally.
Number 3 is how we use the algo to help us, you know, to warehouse some of our risk, and then to lay off the risk in a much more effective, efficient manner and in a much bigger scale. So number 4, we'll talk about how to leverage data and AI/ML so that we can get for more sales business, and also how to manage my trading business much better. And number 5, we are talking about. We tried to create an ecosystem for my DCN business on the origination side. So number 1, so on the pricing engine and the structuring capabilities, I think we did quite a fair bit of effort in last couple of years on sharpen our pricing engine on foreign exchange.
We co-locate the FX engine in London, Tokyo, Singapore, so that we can get much lower latency, and get a much faster feed of pricing into our engine. Just now, I think you heard of some of the business, right? The FX revenue actually has grown dramatically. It's because we have created such engine. You can see that the E-FX volume that we have grown in the last couple of years from 27,400 billion, but in 2022, it's 1.2 trillion. It's a 3 times growth, I think, on the E-FX volume. On the derivative side, we also tried to sharpen our engine. We managed—I asked my quant team to help me to build an internal pricing engine, so that...
They are smart enough to say that, "Oh, maybe we can use the GPU farm to help to price these things, can make us 60 times much faster than using the traditional way." And because of that, we've been able to streamline, you know, all our workflows into all this auto pricing and execution. I mean, in the old days, we always, you know, have the trader, they have a volatility pricing suite, they have a pricing suite, they get another feeds on that, and then they try to enter deals into so that to make the RMs to enter the deal. So with this engine, we can make this in a much faster manner. And then this is highly scalable with APIs, microservices, and cloud.
So you can see that, the request for quotes for the FX derivative, the volume in 2020 is only 210,000, but in 2022, it's 1.25 million on the request for quotes, 6 times more. And then on the volume for the FX derivatives, right? You can see that from $70 billion, now it's $100 billion, a 43% growth on the request for quotes. And the same applied on our equity derivative products. Before I enhanced the engine, 165,000 quotes, but now I reach over half a million of requests for quotes on that. A 3 times increase. And then if you look at my annual customer volume on equity, is from $14 billion, now is $24 billion. Seventy percent growth on the equity request volume.
Two, so we just we tried to build something that we can help to distribute my product internally and externally. I think just now, I think some of you may already have visited my booth there to talk about the DigiMarket. So on this DigiMarket, we created a platform for all my internal stakeholders, be it the IBGRMs, be it the CBGPB, the consultants, so and then also my salespeople. So they managed to use this platform, and they significantly reduce the time to execute a deal. In the old days, before I had this DigiMarket, they need to spend 6-8 minutes to close 1 deal, but now within a click, 1 minute, they can transact a deal.
And somehow it's also have a pre-trade track, also post-trade, end-to-end, and you increase efficiency, and at the same time, you mitigate all the operational risk. So this according the improve the customer journey. This is for our internal. And externally, right? Because of our engine efficiency, so we also managed to price our FX online using the portal and our FX API to plug into different kind of engines, platforms, and ecosystems. So you can see that on a DigiMarket, on a transaction count, in the old days, we only have 8,000 counts, but right now we have 150,000 transaction counts on the DigiMarket. And then if you look at the Treasury API transaction, we are only like, what? $700 million in 2019. Right now, we are talking about $12 billion.
Since our volume, FX volume has grown so much, then we also think of maybe on the trading side, how can I make myself to be more efficient and how to scale this, this business? So I asked my quant team to try to develop an algo program. Before we had this algo program, every 1 million of FX transaction, I only able to make SGD 13 million. But after we applied this algo, you can see that we average in last couple of years, we make SGD 19 million out of 1 million of notional of FX. And then, yeah, and then you talk about, you know, the kind of volume that we have, we have grown, right, on the e-FX is huge. So it's a 45% increase in the trading efficiency over the last four years.
To leverage data and AI/ML to drive the business, to drive business across both in sales and trading. I think on the sales side, we talk a lot on, you know, scale, how to do experiment, and then how to look at the customer data so that we can trigger some nudges, trigger some events-driven FX. And at the same time, we look at the behavior of the customer, look at data, you know, from that side, so when we can kind of tell them, "This is the right time to do an FX," and then we can also do the FX tiering according to the behavior. And then on the trading side, we also use the data to help us to manage our liquidity better.
And then at the same time, it also give us the data, also help us to know that our traders' pricing, are they good? So they can somehow they can adjust the pricing to enhance the RFQ quote on, on the hit rate. And then, of course, it also improve our trading workflow, so using all this kind of data. Using the AI/ML-driven experimentation and campaign, again, I think just now, if you visited the booth, I think my colleagues has kind of explained to you how we draw data, you know, from the system, how you look at the how we do a different pricing of optimization, and then how we activate the client reaction. We even send nudges to them, "Oh, the dollar thing now is reaching a level.
Today, you know, there are some events on the Fed. There's some events, you know, on the street, that trigger you, your sense of that. We even send you some messages, emotional messages, like, "Oh, we miss you," or "This is now the last time, you know, to come to hit us." So we said these kind of things and do a lot of experiments. And end up, we find that the kind of using this kind of event-driven FX, it really drive our sales volume on the FX. So I think this is a very good example to say that how we can leverage on data and AI/ML to drive the sales activities. So lastly, I think we talk about—I talk about the FIX Marketplace.
This is a... Originally, we started from an automated workflow from a DCM business. Because, you know, the origination time for DCM business in the past, right? It might take something like 5-8 weeks to originate a deal. So you have the lawyers, you know, they come to say, "Oh, I need to kind of do a different kind of documentation. I need to change a bit of wordings." I get the investor saying that, "Oh, I need this kind of range of yield to deal." And then I have the issuer, they are also like, "Oh, I'm timing the market, but whether I can issue lower." So we have all these kinds of request from the stakeholders. And somehow we managed to automate all this workflow without adding a headcount.
Then on this platform, I think since we launched, we have SGD 20 billion of trade volume on this platform. And at the same time, I think you can see that we also managed to get some innovation awards from Euromoney, from other awards companies. We have two global and one regional innovation awards in 2022. So I think the whole value proposition, right, on this, is we want to create something that maybe in Asia, we will create something is the best for the DCM origination platform, I think in Asia. And of course, because of what we have done, you can see that our CAGR of sales growth, right, in last couple of years is a 9% CAGR growth because of this digitization effort.
Of course, we also continue to use the data, so that using a data analytics, so that we can have a much more discipline on my capital management. So you can see that, you know, our last couple of years, we managed to increase our revenue to our RWA from 3.2% to 4.6%, 1.4 times growth. And we also managed to decrease our RWA versus to our SSI. So we are ranked second on the capital reduction vis-à-vis the asset growth, which I think this is a true reflection of our using data analytics to drive our capital management.
I think what I said so far, I think it just testify that what we have done last couple of years on the digital transformation, it really built my sales business, scaled my business, and make myself to be more efficient. Of course, in next two, three years, we'll continue to deepen our digital capabilities, I think in both on the structuring and the warehousing. Just now we talk about, you know, we do a lot of digital on the FX, on the equity, but now we're extending, you know, to maybe other asset classes, be it on credit bond or be it on rates. And then we also extend our algo in other overseas countries.
To externalize the DigiMarket, I think you have so far, you have seen the DigiMarket, I think it's a wonderful platform. In Singapore, if you see that in last 1 year, you have so many family offices are coming to Singapore to set up new offices, and then you have all the asset managers, all the people are coming in. So I think this is really a good time for us to launch our DigiMarket externally. And of course, to, on this, MTJ, I think just now you heard all my colleagues, they talk about this MTJ. I think on T&M, we also adopt this MTJ, and then so that we can continue to accelerate some of our agendas, and then we work collaboratively with other partners.
Then lastly, as I said, on the FIX Marketplace, so I think my dream is really to create a platform that we can be the best in class in Asia on the origination platform.
Good afternoon, ladies and gentlemen. My name is Surojit, and I am the country head for DBS in India. I've been in this position since 2015. So I was here in 2017. So before I go into what I have to say about what's happening in India, you've heard through the day the digital transformation across our businesses, across our various front, middle, and back office. I'm gonna talk about how this is playing out in one of our key growth markets, India. In 2017, we had presented our thesis for a growth market like India. In India, at that time, there were some very interesting trends which were crossing digitalization, which provided a unique opportunity. I'll talk a little bit about that.
I'll talk about the tests that we did in India, the learnings that we got, and how we have shaped our strategy in that country. Then I will talk a little bit about where our strategy is today, and what we are trying to do over the next 3-5 years, our three-pronged execution strategy, and finally, the early results that we are seeing and where we expect to be over the next 3-4 years. In 2017, if you were here, you would have seen this slide. I talked about the 3 things that was happening in India. Firstly, the government was really pushing digitalization through the creation of the India Stack. Many of you have heard about that. Aadhaar, the identity.
UPI, you've heard about that through the day, eKYC, Digital Locker, and that was really driving a significant push towards digitalization. Second driver was the fact that handset, smartphone handset prices were crashing, and India had probably the lowest tariffs, which was meaning that both consumers as well as SMEs were adopting very quickly. This was also creating an explosion of data, both data for cross-buys, but also for credit underwriting. So that was really the basis of the model that we saw was possible in a market like India, which was really to build a business, bypassing, if you will, some of the long periods that it takes to build SME and consumer businesses.
As a challenger brand, we were very keen to partner with customer-owning and data-owning partners, and then build a business which could be faster, thereby telescoping profitability in a market like India. Hence, again, in 2017, we talked about the fact that digibank, our mobile-only platform, we actually test-marketed in India as the first test market. In 2017, we did a unique partnership with an accounting software company, Tally, which had more than a million paying, fee-paying customers and almost 7 million users, and had a large network of partners who were deploying accounting software. So what we did was we built a e-payments-embedded solution within the accounting software, and that was our path to grow our SME business. The tests that we did gave us some learnings.
Let's talk a little bit about what we saw. Our thesis that we could accelerate customer acquisition was correct. We got 10 million downloads of digibank app in the first two years. We also got more than 400,000 SMEs who were interested in banking with us. But we also saw that the conversion ratios were lower because of some last-mile frictions. There was clearly a need for some assisted digital to be able to increase the conversions, which would then make our acquisition cost better than what we saw. Even the customers we acquired, about 3.2 million customers on our digibank platform, we found that there was a preponderance of younger, transaction-heavy customers with low balances, which meant that the path to profitability was much longer.
Also, for those customers who were in our target market, they used the very convenient digibank platform as a convenient secondary bank and did not move their primary relationship. There was clearly a need for some more assurance and some more touch and feel. And finally, even on our SME side, we got about 37,000-40,000 customers who converted. Many of them were small and micro-SMEs with low balances. We also saw that on our asset side. We launched a fully digital consumer lending platform, which was very, very good in terms of UX, but we saw that most of the people who were adopting it were younger and with smaller loan sizes. Slightly older and the larger loan sizes, again, needed some some assistance in terms of conversion.
So our conversions were great when there was younger population and smaller loans, but not so great as we got into higher loan sizes. So clearly, some need for last-mile intervention. And the more profitable secured loan products or transaction-heavy SME businesses clearly were not easy to deliver on a fully digital platform. Again, need for some amount of intervention. And as I talked about the fact that our growth model for India was that as a challenger brand, we wanted to build a successful partnership, both with customer-owning as well as data-owning partners. We found while the dialogue was good, the pace of adoption was not as good as we wanted.
So clearly, there were some key learnings which we built into our move, movement or evolution into our business model for India, which was really to provide more assisted digital, give more assurance, and for both secured loans and SME, provide last-mile connectivity and more assurance for our partners. So even in 2017, I had talked about the fact that we were converting our branch presence into a wholly-owned subsidiary. If you recall that, for branches, there were severe restrictions on physical expansion. So in 2017, we talked about the fact that we had already applied to become a wholly-owned subsidiary, which we completed in 2019. So that was our first phase of providing a physical underpinning to our phygital strategy. And in 2020, we acquired Lakshmi Vilas Bank, which allowed us to nationally expand.
And now with more than 500 branches across the entire market, we can access 90% of the addressable market, which at SGD 200 billion, growing at 10%, is very attractive. Similarly, during the same time period, over the last 4-5 years, we've completed our product suite. We've filled all the product gaps that we had and also all the customer segments we are targeting. So since 2017, we've added a credit card business, we've added analytics-based SME lending platform, and I'll talk a little bit about the results. We've become now an approved bank to handle government business, and very recently, we opened a branch in the international business, the GIFT City at Gandhinagar, Gujarat, which provides us some interesting opportunities, both for inbound institutional investors, but also for the overall market.
So filled up our product range such that we can now target 70% of the addressable market. So that's really the build-out of our physical underpinning for our evolved phygital strategy. So just a quick summary. I'm not going to go through each one of them. Our key business strategy is that we have a mature, large corporate and FIG business. We'll continue to grow that at between 10% and 15%. We will expand our mid-cap business, given that now we have a larger footprint. Over 30+ cities, we will be present with relationship managers, use a very sharp industry focus on-overlay that on the, on our midcap strategy. Grow that again, mid-teens.
Our key two growth businesses are SME and consumer, where we will use the phygital strategy to both build liabilities on the SME side as well as on the consumer side, and build out our loans business, both on the using analytics-based lending and also using our linkages with large and mid-corporates . Supply chain, Tan Su Shan talked about that. India is one of the markets we're seeing the best pickup of supply chain. We now can do it nationally. We'll grow our secured loan businesses. We acquired Lakshmi Vilas Bank with some interesting capabilities on gold loans. We are growing that, and I'll talk about how through partnerships, we grow both organically and through partnerships, we are growing that.
So both our consumer business as well as our SME business will grow at an accelerated 40%-60%, primarily because they were very small, so they're growing off a small base. But over the next three to five years, you'll see a much more balanced business. Let me talk about now what our execution strategy—execution plan is against this strategy. So three real pillars. Firstly, you've heard about our digital capabilities. We'll continue to deploy, grow our digital capabilities in growth markets like India. And we've already done. We've talked about digibank. We have IDEAL, we've got our SME financing workflow, which we launched in India, now scaled to the rest of the bank. Again, Kwee Juan talked about 60% of our growth businesses are using Managing Through Journeys.
That's true for India as well, and that will be a key part of creating the Flywheel for growth, using our digital assets. Second pillar, we will use our expanded physical footprint to deepen customer acquisition, deepen our wallet share, but also to launch some of the products which we couldn't launch when we were essentially digital. And finally, using both these strengths to build meaningful partnerships with the customer-owning and data-owning ecosystem partners. And I'll talk about all three of these and show you some proof points of where we are on each of these pillars. Firstly, let's talk about continuing to deploy our digital capabilities and Managing Through Journeys. I'll do this in three buckets.
Firstly, for consumer, then for SME and corporate business, and finally, all the digitization that Jimmy talked about, front office, mid office, back office, what's it doing to our overall digital adoption? Let's talk about Digi Bank. On Digi Bank, since 2016, we've built out the entire platform in India. We have over 1 million customers. I talked about the fact that we acquired 3.2 million customers, but we also saw a large number of them had very low balances. So we've been able to refine our customer base. Now we have 1 million customers where we think we can cross-sell and sell multiple products. So we have now a mature set of 1 million customers on Digi Bank before we acquired Lakshmi Vilas Bank. Now we are migrating 1.4 million customers from our acquired bank onto Digi Bank.
You're seeing some result upticks as well. In March 2023, 83% of our new accounts were done on digibank, fully digitally. We also seen over the last 5-6 years, clear increase in terms of our revenues per customer. Here we add our partnerships, because our partnerships are essentially digital. So we've seen a 5x increase per customer revenues, and we think it'll continue to grow. And we are seeing much better quality of acquisition because we can be more targeted and we are able to get better results. You can see that our average savings account, SA balance or savings account balance, is up over 2x, and our ticket size on our loans, again, with assisted digital, has gone up.
So clearly, this will reduce the number of years it takes to get a new customer that we acquire, become profitable, and that's a big driver to our growth strategy. Second, part of our continued digital, you know, ambitions is to continue to build out our SME platforms. We have now close to 12,000 borrowing customers on our SME workflow financing, which is our algorithmic lending platform. We are adding, and all of them are on IDEAL, which is our internet banking platform for SMEs and also for large corporates. We are migrating 70,000 LVB customers onto this platform, post-integration now available to all of them. Again, in March 2023, 85% of our new accounts was digitally initiated for our SME business, and the results are there to see.
Since we launched our SME financing workflow, both our asset growth as well as through IDEAL, our liability growth are quite impressive. And we are seeing this also in terms of our productivity. Our per RM revenues have gone up, as has our per relationship, you know, returns, because at 4.4%, it's almost about 50% more than where we were, say, 2-3 years back. Now, the last bucket, where we talked about the work that we are doing on the credit side, we're doing on the mid-office side, all of them are helping to make our platform lot more digital. We are seeing digital adoption product by product. Of course, we launched our card business, which was essentially digital, so we have the benefit of a 100% digital card business.
But even in our wealth business, where it was probably 50/50, 4-5 years back, we are up to 74% fully digital. And we're doing that for a much more physical business, because a large part of our, fixed deposit business came from, Lakshmi Vilas Bank, where we started with a low, digital adoption. We already got that up to 50%. We expect to take it up to 75% over the next few years. So again, continuing to build, our digital assets in the country, refining them, making sure that the results are better, and also making sure that Managing Through Journeys creates the Flywheel in each of these growth businesses. Let me now turn to the second pillar, which is leveraging our physical, expanded physical distribution.
Before I talk about what's happened to our branch integration, we just want to pause and let you know that since the acquisition in 2020, we've completed the full integration of Lakshmi Vilas Bank, not just in terms of branches, technology platform, people, branches, but also some of the portfolio issues that we had. We've significantly been able to resolve this. So now we have a unified platform with more than 500 branches, as I talked about that. We have upgraded all our branches. They are all rebranded to DBS now, and we've shifted and aggregated 90+ branches, and we'll continue to do that as we find gaps, if there are any, in our distribution. We expect 65% of our combined branches to become profitable by end of 2024, and 80% by end of 2025.
That would be very impressive, because just to remind you, that in a country like India, more than a quarter of our branches have to be in deeply rural area as per regulatory requirements. So if we get to 80%, effectively, even 5% of our rural branches are profitable by then. Let's talk a little bit about our asset businesses. We have, post our integration, now been able to deploy a new platform for all for gold loans, which is now growing attractively. It's now available over 300 locations. We've been able to, what I talked about earlier, assisted digital. So our unsecured loans now through our branches, are being able to be deployed, and we're seeing a 10x growth over one year.
This is just a one-year growth because we are being able to get some of the drop-offs that we were getting earlier at our branches. Our SME financing workflow, which is basically our algorithmic lending, now available at 100+ hubs covering 220 branches, which gives us access to 80% of the SME addressable market, in the country. And again, Tan Su Shan talked about it. With a national supply chain solution, we are able to now use our large corporate, mid-corporate, and they're both our dealers and suppliers, to accelerate our SME growth because we have a national presence. And you see the results on the liability side. Key part of our strategy, continuing to grow our, liabilities through SME as well as consumer, and you see the benefit that we are getting because of the physical, phygital play.
Already, our initial balances are up 2x on savings accounts, and we are seeing clearly that our month-on-month acquisition of fixed deposit from consumers is up 4x. And even on our SME, what we call operating accounts, where we are not lending, just opening current accounts, we are seeing significantly better results because of our phygital play. So that's our second pillar on continuing to get the juice out of our expanded physical footprint. And finally, using both of these, building strong partnerships with great results. So I talked about the fact that while we had a lot of interest on partnerships, the results were not still showing up while we had a narrower base. Now, with the phygital evolution, you see some of the results. On unsecured loans, we have a significant uplift.
68% of our new loans are coming through partnership. Han Kwee Juan talked about CRED partnership. That's one of them. We have two others, and we'll add two more before the end of the year. Our co-branded credit card. In India, we decided, because we were a late entrant, that we will go with a strong, large player, which had large card issuances month-on-month. So we started our Launched our card business, counterintuitively, with a co-branded card, and we already have a 7x increase month-on-month. We're adding 30,000 cards. We expect to exit the year at 50,000 cards per month, and get to by middle of next year, such that we could add 1 million cards on... just the partnership itself.
This gives us an expanded base, which allows us to then build our proprietary card business, which we've just launched in May, our proprietary card business, which will allow us to then deploy that across our 2.4 million customers. And we are seeing the results are great. Not only are the cards that we are originating very digitally active, even the results are better than market. And let's take the last example, gold loans, a very traditional product. You would assume that that's very difficult to partner. We have a partner which is adding now almost 50% of the new customer additions, and already within a year, 25% of our overall gold loan book is coming through a single partner. We expect to add another partner before the end of the year.
So good proof point, in March 2023, over 50% of our new customer acquisitions were through partnerships, and we want to replicate that for our SME business and all our growth businesses. So that's the third pillar. So what does this do to the shape of the franchise? We're already seeing that, if you look at where we are today. So if you see, where before we acquired Lakshmi Vilas Bank, we were, we were significantly overweight on our large corporate business. And Piyush talked about the fact that while we have a good position, that yields are lower. And clearly, post the acquisition, already by 2023, you're seeing that the share of the large corporate business is down to 64, our growth business is up to 36%.
And we expect that by 2026, it's much more evenly balanced, which will allow us then a NIM uplift of more than 100 basis points, because that will then allow us to sharply compete and do switch strategy, both on the SME side as well as on the consumer side, which will mean that our dependence on corporate deposits, which currently was 67 or was 67% before the acquisition, down to 44%, will get down to 33%. So those are the two big drivers, and that strategy has not changed since 2017 or 2015, where we said we need to have a much more balanced business, both on the asset side as well as on the liability side. So you're already seeing that the results are visible.
So from now till 2026, we expect almost a 2.2x increase in our overall balance sheet side on the asset, with a large growth coming from SME and consumer on a smaller base, albeit, but significant growth at 40% CAGR. And our overall balance sheet on the deposit side will grow at a slower pace, because with lower corporate deposits, a much more efficient book as we grow our liabilities. This is showing up also in our results. Again, Piyush talked about it, also Sok Hui talked about it. So we've seen about a 25% growth from 2017 to 2022 in our top line. Our income is up from SGD 183 to SGD 551, roughly about 25% CAGR, and from a loss to a profit in 2022.
We expect in 2023, this will also grow by about 25-26%, maybe a little higher, and then continue to grow within 25%-30%, such that by 2026, we'll cross SGD 1.4 billion on the top line, and our profitability go up to about SGD 375 million. That's where we are going. This will have a significant impact on our cost income ratio, tightening from 78% to close to 55%-57%, and it'll also mean that our ROE will go up. This will start to then contribute more meaningfully to the overall group. So that's a good way of looking at how digital transformation and digital assets are being deployed in a growth market. Now, if you come to what that does this mean in terms of our aspiration in India?
Clearly, our aspiration over the next 3-5 years is to be one of the top 10 private sector banks in India. And if you look at some of the names that are ahead of us today, they have all executed similar strategies of growing their SME and consumer business and aggressively switching on the liability side. We have all the elements, both with our digital as well as our physical presence, to be able to do all of those over this next 3-5 years. So that's our aspiration, profitably break into the top 10 private sector banks in India. Thank you very much.
So one of the things that we have done in the last 2-3 years, we've piloted a bunch of these things, and I call them piloted because they're nascent, by and large. The first two I'm going to talk about, but this cross-border, low-value payments, this is not new. This we've been doing for 8-10 years. This is the equivalent of the Wise business. You know, so, the digital asset ecosystem I referred to right now, blockchain-based payments are part of where we've begun to finally, you know, we had to get regulatory approvals. The last one, from the US came through a couple of weeks ago. So we're now building an ecosystem of, multi-currency players who can do that. The carbon exchange, and this that we launched, we've got a listing capability auction.
We're obviously 25% shareholder only in that, but we need to get traction with the Climate Impact Exchange. We launched a few mezzanine funds, both for growth and then we have this thing in the back of our mind, there's so much interest in the software that we built, that we could probably monetize that sometime in the future. The two businesses which are worth thinking about, we talked about, you know, what do we do? So one is the cross-border, low-value payment business. You know, last year, we moved about $24 billion. You saw that earlier. We've been doing mostly B to C in this space.
A couple of years ago, we started going into the B to B to C space, so we're taking this capability and giving it to other banks and other companies, to be able to then deliver to their consumer customers. So where we are now, we think we'll get this 24 billion to about 40 billion in throughput by 2025. So we are on track to do that. Last year, we made just a tad below $200 million in this business. We think we can get 50% growth in this business, so build it to be a $300 million business. At this scale, in terms of the volume we're moving, we are about one-fourth of what Wise does around the world in terms of volume.
But in terms of our monetizing capabilities, actually better than what they're doing. So you can, you know, figure this is something we could actually take. We've thought about how you could take it, and at the right time, it is appropriate, it's something that we could do and, and spin out. The other that I talked about, the exchange custody and securities token offering. I've been quite surprised, actually, myself. With this crypto winter, I thought this thing is going to be very moribund. In fact, I think our fact that we are an exchange which is offered by a regulated entity, DBS, has actually been helpful to us over the last year. So there's a lot of these global, you know, FTXs of the world have been coming under scrutiny and question.
We are getting a lot more interest because, I think of what we are. So like I said, in terms of coins, again, we haven't gone the same. We had about 1,000 investors, both institutions as well as individuals, who we invited on the exchange. So, even the 1,000 people, the volume of business they gave us doubled in the course of the last year. This year, our thing is that we're going to grow that by 20 times, but still in the accredited investor institutional space. I do think there's a further opportunity. If you look at Hong Kong, for example, the regulators are a lot more open about thinking about expanding it and making it easier to go into the retail space.
We'll still be cautious because I think, you know, suitability and appropriateness is not straightforward, but I do think over the next 2-3 years, the opportunity to scale this is quite material. So this one, I think by 2025, you could get, you know, $80 million-$100 million of revenue in this activity and business. So if you got that, then you might again have the consideration saying, "Can you take it and do something with it?" The environment has not been conducive for us to really focus very hard on this right now, right? Nobody's funding, nobody's looking to invest, and so, it hasn't been top priority for us, at this point in time. But it is something which is out there. It is something that, you know, we will still continue to work on, on the side.
I guess my last slide, it's just a summary, some upside, but, you know, we, we call it this thing about a different kind of bank. But if you look at what we've been able to achieve in the last 5, 6 years since we last came, so 1, we have built a set of best-in-class regional businesses, the transaction business, the private banking business, the T&M business. I think the data, the external, market share data speaks for itself. The fact that our private bank, you know, gone up to number 4, number 3 in Asia, our transaction bank, on Greenwich survey, we're number 2 in Asia. Obviously, we're gaining some traction, even if, you know, except our numbers are not, the market numbers are, are quite, visible.
But now we finally, I think we cracked the code to start building up scale in the local businesses we've been wanting to build in some of the market. It's not consistent, but I think we'll get India, I think we'll get Taiwan. I think the Shenzhen Rural Commercial Bank approach into the Greater Bay Area will take a little bit more time. Our equity stakes are low, and as it opens up, but I think we have line of sight, so I think we've been able to do that. So 3-4 big priority growth markets as well as the large regional businesses, it's, this and that. The other is that obviously, we're outperforming in our core market, and Singapore is quite clear, whether it's market share, whether it's external data, whether it's our, economics and our ROE, it's been very strong.
But the data-driven underwriting capabilities we built, I think it finally gives us the capacity to scale outside the core markets as well, build out the big customer bases and get into that. This method of working, this self-reinforcing Flywheel, and we've got the concept of Flywheel from Amazon. They talk famously about their Flywheel. We borrowed it from them. But if you take a look at the ability we built to be able to create a control chart, to be able to experiment, do customer observation, we have an experimentation engine, we do control testing, we innovate out of there, we take the feedback loop, and it's not done centrally, it is done in every journey.
So I think this unique capability we've built to be able to manage through Flywheel, I think that's actually extremely interesting. The one other thing we haven't talked about, but obviously, we've had the benefit of a lot of stability in our team, senior team. We've got very consistent team. All of the senior that you see have been around for a long time. We've had limited turnover. And I think that's quite helpful, having a senior, stable team which just continue to run the shop. But at the same time, we married it with a lot of new kind of tech talent. Our tech talent is very good. We've got 12,000 people, the design people, the deep tech people. So I think that blend has been actually quite a useful mix.
And like we said, I do think we've created some options in the long term through a portfolio of these new growth engines that we have. So when you put all of that together, I think we have some reasonable degree of confidence in the targets that Sok Hui have laid out for you. And if we can do that, as we think we can, I think this goes back to where I started this morning. I think the investment thesis is quite compelling. If you believe that Asia is around, if you believe that there's a smart way to play Asia, and you believe that being able to do this and leveraging a company which gives you some technology benefits, some ability to manage more like a tech company, then, DBS is a good, investment opportunity. So thank you.
Thank you all for your time today. I hope you find it instructive, and I'm sure we're happy to take further questions from you, offline, if you want to get back, either through Michael or Sok Hui. Thanks very much.