This is the third post recapping the great disucssions from our October Investment Forum on Open Finance/Open Banking featuring panelists Tessa Lee from moneyinfo, Nick Eatock from Intelliflo, and Ross Laurie from Visible Capital. This section featured discussion about legal regulations, machine learning, and increasing adoption of open finance systems across the wealth management industry. Our next Forum session will occur on November 11th; tickets to attend can be accessed at the bottom of this page.
Open banking had implemented equivalent standards across a varied incumbent financial services industry, but that’s going to be very difficult for open finance without legal regulation. What impact will that have on delivery timelines?
Open pensions/pensions dashboard was pushed forwards by a petition signed by 200,000 people in the UK, but open finance will be more challenging to implement. There’s also the issue of where the data is stored. It would be good to have it mandated but I believe that the companies will follow what the market demands.
Ross Laurie:

I mean, it depends on what you’re referring to here. I mean, the interesting thing I said at the start open banking was sort of pushed forward from a regulatory point of view; strangely, open pensions and pensions dashboard– there was 200,000 UK people signed a petition to request that to happen (https://you.38degrees.org.uk/petitions/do-not-scrap-the-pensions-dashboard). So there’s an exact example of the market saying, “we want this, let’s see if you can get on and deliver it”. An open standard in open finance is going to be much more challenging than open banking, yes.

And there’s another part to that as well of where do you store that data? –  In terms of federated identity models – and also what the customer actually owns. What we’ve seen with the open banking rollout as we mentioned before, there was the CMA 9 banks that were mandated to do it but we’ve already got 21 banks that we could access as there are some other banks out there who are actively looking to give their customers access to their data because they know the customer wants that. And yes, you’re right.

It would be useful to have it mandated, but I’m a big believer in capitalism and how markets work. And if the consumer wants it, then the good companies will respond and give them that, and they will then accelerate. So it’s going to have to be probably a bit of both. Unfortunately I don’t think it will happen as quickly as I want, but thanks to this global pandemic, perhaps we might see some acceleration faster than I hoped.

Open finance is much more complex than a pensions dashboard.
Tessa Lee:

Just a quick point from me on that. I think you’re right. If pensions dashboard is setting the benchmark, we are in a bit of trouble because we’re going to be waiting a long time, aren’t we, for this. And you’re right.

The challenge with it is of course, open finance data is much richer and much more complex. We’re not just talking about account balances and transactions data. We’re talking about holdings and transactions and asset allocation and contributions and withdrawals and maturity. There’s a whole raft of data that needs to be considered.

So I think it is going to be more challenging, but as I iterated at the beginning, advisers and wealth managers are already in a unique position to provide open finance type portals now because you have access to so much of this data through the platforms and the providers and the agencies that you hold. So I’ll bang on about that point again, because I do think it’s a really unique opportunity for you out there.

How is machine learning making the use of open banking more efficient?
Oliver Bourke:

I think it’s great being able to pull bank accounts together for simple client scenarios. However, how does it work across several bank accounts and credit cards? How do we identify business expenses, non-standard transactions and movements between investment accounts?

And I think the way I’m looking for the question to be answered, is around machine learning; what are we doing in the machine learning space to be able to make this more efficient?

The systems attempt to categorize the transactions from your bank records when you add an account, and then the individual can come in and correct it.
Nick Eatock:

I mean, I guess it probably depends on how the vendor you use categorizes the expenditure. There are various different approaches out there, our own one is through machine learning for sure. And you can also categorize as an individual and it doesn’t forget your categorization. So when future stuff comes through it categorizes it like that.

I suspect both the solutions from Ross and Tessa work in the same way. Automatically these systems attempt to categorize on the first off, so when you link in with a particular bank account or whatever, depending on who the bank is, there’ll be somewhere between, I think it’s 3 to 12 months is the standard sort of range of data. It pulls that data down as a one-off kind of initial load. And it will automatically categorize that.

I think our experience has found that actually that categorization is pretty good, but it does at that point in time benefit, by the individual saying, “actually, no that expenditure should be categorized to that”, and it remembers that for every time going forward. And the interaction I think this speaks to what Tessa said, that the reality is that a lot of users are going in every single day now they have open banking and because they can see, and to have visibility over the information in there, even if you don’t necessarily do much about it, is quite comforting in itself.

And I think people start gaming with the system to try and say, “right, okay. I want to get that categorization as accurate as I can. So if I spot something that’s wrong, I will mark it. And I’ll know when I look a week later, when that same thing happens again, they’ll have categorized it correctly”.

The auto-categorization is very good, but there’s always going to be a bit of client interaction needed.
Tessa Lee:

Probably not much more to add to what Nick said, really. We take the same approach. It’s never going to be perfect because there are always going to be transactions that clients have to categorize themselves. But if the system remembers that next time, that’s brilliant.

We’ve learned a lot over the years about how to manage different types of transactions, transfers between accounts, what to exclude from income and spending because they’re transfers, or refunds, and actually screen scraping gave us the ability to handle that and learn from that.

And we applied that same logic with open banking. So I think as Nick said, it’s never going to be exactly perfect. But if I add my own accounts on which I do all the time, because I help out with the testing here sometimes. Then I probably get maybe 80+% of my transactions auto-categorize first time. And then I let them know that the Encore is a pub down the road that I frequent quite a lot, because they don’t necessarily recognise that like they do Tesco’s and then they remember that next time. So it’s fairly good as it stands, but I think there’s always going to be a little bit of client interaction needed.

We’re hitting 95% accuracy on initial categorization, and we flag uncategorized things for the adviser but we don’t usually let the customer manually input categories, though they can change between existing ones—we want it to remain standardized.
Ross Laurie:

Multiple accounts are actually relatively straightforward. It’s the same process in the aggregation, including credit cards as well, and the reconciliation thereafter. Our focus is really on trying to get to this to an enter it once or never approach – straight through processing. We’re hitting 95% accuracy on that first pass categorization. We’ve got over 180 categories, 27 alone on income.

So what we’re trying to get to position is that the customer can just look at it – if we’re hitting 95%, i.e. £70,000  income, £2000 uncategorized, fine, we’ll flag that to the adviser and pass it through. We don’t really allow the customer to manually input the categories. They can alter them from a drop down list if we’ve got something wrong, they can change it, but they can’t make up the category.

And the reason for that is we’re really focused on the data output standards. We can’t just have anything going into back office systems because then there’s no identifiable pattern there. If we can’t consistently categorize, we can’t look across the book and see what’s happening. So I think in terms of future standards, we’re working on that, but at the moment, give it a go, we’re at 95% accuracy at the moment, I’m quite pleased with that.

This is being sold as better quality for clients as well as beneficial for advisers? How do you get legacy people to participate?
Ross Laurie:

The client outcomes bit is exactly what we we’re driving at. The first thing the adviser does is to have a conversation where they can give the customer advice rather than “can you answer some questions that then allows me to give advice.”

Regulation is the natural way to get legacy providers to update their systems.
Tessa Lee:

I’d agree with Adam that that’s a huge challenge, but it’s a challenge now isn’t it because some of the legacy providers won’t give electronic data now to advisers, because their systems are dated and they don’t particularly want to give it. So I think regulation is the natural way to help with that. I think GDPR has its place, the data portability rules that that Ian has talked about, but of course, again, how you interpret that is down to you.

You could do that on an individual client by client basis; well, that’s no use for getting data on a bulk basis across hundreds or thousands of clients. So I think it’s a big challenge. And I don’t honestly know what the answer is to that Adam, other than over time, through regulation and legislation companies being forced to give access to that information. But those types of providers will be far further off than the more modern platforms and providers that we’re working with already today.

Public pressure through laws and competition will help. We saw a large discovery of held-away assets within the first few weeks of using open banking, which helped advisers a lot to be able to take that into account.
Nick Eatock:

I think they’re all excellent points. And Adam has picked on the challenge that open finance faces ignoring technicalities, which is just is the will there to do it. And I suspect without some form of being beaten up in public, whether that’s in one extent law or whether in another extent is regulation or in another extent it’s just competitive tension.

The States, albeit that they don’t have the same open banking infrastructure, they actually do things much more through players like Yodlee and players who do a mixture of screen scraping. Other data asset scrapes have shown that a huge level of discovery of what’s called “held-away assets”. Held-away assets are simply the assets that your client doesn’t mention to you as part of the advice process. Indeed in the UK, we saw in the first three weeks that 20 million pounds of assets was discovered within those first weeks. We’ll get an updated figure fairly soon, probably before our conference in November.

So I think you can see that there are capabilities here to give more access to the information for the adviser. Where the adviser takes that into account or starts to manage and provide advice services on all of that is obviously as a matter of time and a matter of appropriateness for the individual client, but we’re certainly beginning to already see the indicators that this is a great way of discovering some of those held-away assets.

I think things work best when businesses work together to better serve customers. I’m interested in how other countries have worked with the organizational aspect of introducing the new technology, how they’re learning from us and how we can learn from them.
Ross Laurie:

I suppose it’s actually about being inward looking, you know, we’ve done open banking and it’s worked out okay but there’s a lot of improvements that can be made. We’re now looking at the start of open finance, pension seems to be the thing that keeps popping up, but of course there’s open savings and investments and all the other things as well. I’m interested in the way that Australia are approaching this, and this is based on conversations I’ve had through events.

When we were still allowed at events, I spoke with Sarah Court from the ACCC and I was really fascinated about what they were learning from what we’d done. And their approach was very much to almost start with an open finance model and say, “okay, well, banking is one thing, but actually we need to get a bigger, complete financial picture”. And this was driven largely by the institutions asking this, like Australian Super and people like that, rather than the government saying, “this must be done”. And I rather like that. I think that’s how things work best when businesses work together, identify how they benefit their customers, and then sort of say, a rising tide lifts, all ships, you know? So I’m rather enthused by that model of going, “okay, well, how do we just embrace this and get on with it?”

So it goes back to the point that was made before, I think from Adam about the legacy involvements, there’s those classic three problems you’ve got to address in introducing any new technology or regulation. You’ve got the technology, which is actually relatively straightforward as far as I’m concerned, you’ve got the regulatory bit of it, which will help drive it forward. And then you’ve got the organizational bit, and it usually is the organizational bit that seems to take the most time.

So I’m enthused by how other countries have looked at what we’re doing and are trying to get on top of it. Other countries have a slight advantage over us because we never really adopted an identity card or identity system in this country. And I said that a couple of questions ago, we’re going to have to figure out where we’ve put all this data at the end of all this, if we’re going to look at it on a holistic open finance model. So there’s a bit of work for us to do on that as well. I think that I’m impressed by how people are learning from us and we should do the same.

Personal capital is a standout model for us to learn from. Supporting these services will help support customers and make it easier for them to do business with you.
Tessa Lee:

I’d agree with Ian that obviously personal capital, as you mentioned, is a bit of a standout model to look at, I think we can learn from these markets. We’ve also got to understand the different kind of culture and consumer that we have in the UK as well, and adapt to that in terms of awareness and communication.

But also, I think that the organisation piece, the parties that will be involved in open finance are the ones that are going to make it happen. And they have to be open to open finance because at the end of the day, within the advice and wealth management market, people will switch and move their money anyway, it’s just going to take longer without open finance, but it will still happen. So actually, to support consumers and to make it easy to do business with you by supporting these services is something I think that we can all learn.