Financial Services Regulation Committee 

Uncorrected oral evidence: Regulatory sandboxes in financial services

Wednesday 28 October 2026

10.05 am

 

Watch the meeting 

Members present: Baroness Noakes (The Chair); Baroness Donaghy; Lord Davies of Brixton; Lord Eatwell; Lord Hollick; Lord Lilley; Lord Sharkey; Lord Turnbull; Lord Vaux of Harrowden.

Evidence Session No. 2              Heard in Public              Questions 12 - 25

 

Witness

Anna Wallace, Chief Executive Officer, Centre for Finance, Innovation and Technology (CFIT).

 

USE OF THE TRANSCRIPT

  1. This is an uncorrected transcript of evidence taken in public and webcast on www.parliamentlive.tv.
  2. Any public use of, or reference to, the contents should make clear that neither Members nor witnesses have had the opportunity to correct the record. If in doubt as to the propriety of using the transcript, please contact the Clerk of the Committee.
  3. Members and witnesses are asked to send corrections to the Clerk of the Committee within 14 days of receipt.

19

 

Examination of witness

Anna Wallace.

Q12            The Chair: Welcome to today’s meeting, which is the second oral session as part of the committee’s inquiry into regulatory sandboxes. Thank you, Ms Wallace, very much for attending. This session is open to the public. It is broadcast live and subsequently accessible via the parliamentary website. A verbatim transcript will be taken of the evidence and will also be put on the parliamentary website. Ms Wallace, could I ask you to introduce yourself for the record? I believe you will be making a few opening remarks and then we will move into questions.

Anna Wallace: Yes. Thank you very much, Chair. I was responsible for establishing the FCA’s original regulatory sandbox back in 2016, and I led it until 2020. I now lead the Centre for Finance, Innovation and Technology, also known as CFIT, which was established following the Kalifa review to help turn financial innovation into practical, scalable solutions. That was funded by the Treasury and the City of London.

The original sandbox was designed to promote competition in the interests of consumers—that was shorthand for consumers and users of markets as well—while maintaining appropriate safeguards. However, our ambition was broader than regulatory experimentation alone. In 2015, we also recommended an industry-led sandbox, recognising that some of the barriers to innovation were not regulatory. They related to access to data, infrastructure, and the ability for firms to collaborate. That initiative did not materialise. One lesson was that building that technical infrastructure is not enough. Industry needs to engage on the problem it wants to solve and to commit to working together. This is the approach that CFIT, which I lead, takes today. Our coalitions bring together industry, regulators and government to tackle shared challenges, develop and test minimum viable products—that is, prototypes and small-scale products—and build the evidence needed for wider adoption.

Ten years on, I see two opportunities: first, to evaluate the sandbox against its original objectives, looking at its long-term impact on competition and consumer outcomes; and secondly, to revisit how industry-led experimentation can complement this regulatory sandbox and support market-wide adoption. Thank you.

Q13            The Chair: That is very interesting. Thank you. Could you explain how CFIT interacts with the sandboxes?

Anna Wallace: Yes. CFIT interacts more broadly with the regulator in a range of ways. It is involved in our coalitions. That can involve anything from identifying market failures around innovation, areas where industry could lead action, and thinking about some of the recommendations that may come out of this coalition work. It is less involved but can have oversight of some of the prototyping that industry can do together. We will always inform it, but it is not actively involved. The conversations that take place are much more commercial.

However, what we have seen in developing the CFIT model, and CFIT is still relatively new itself, is that there is a pathway from the CFIT coalitions into the FCA’s regulatory sandbox. We have one example now. We were commissioned by the Treasury and the Kalifa review to look at corporate digital identity, open finance, and SME access to finance. In one of those coalitions, Digital Company ID, we have identified a consortia—that is probably a better word—of market actors who will be testing a novel way, or a digital way, of verifying a company, its directors and the activities that it carries out through the corporate digital verification ecosystem. That involves banks. It also involves some players who work as orchestrators of that data, and it works with companies that do the front-end engaging with consumers.

We have been able to bring together a proposition that is not just one firm. It is a range of firms that are looking at how a novel ecosystem, for want of a better word, could work. The FCA has agreed to take that sandbox forward. It is a collection of firms, I think probably six firms, in that group that will be testing collectively the use of this novel approach to identity of companies.

The Chair: Okay. You bring what you call coalitions together. How representative are those coalitions of the whole market?

Anna Wallace: The coalitions are open in terms of who can join the coalition. There is due diligence that we will undertake in those coalitions. Do they have the capability? Are they interested in public value? Out of CFIT, one of the things everyone has to sign up to is that the outcome of the work is a public good—not all participants are interested in that—and that they are strategically aligned with the public policy goals: in other words, the corporate digital verification, and enabling financial innovation.

The Chair: If I put it another way, are there parts of the market that do not engage with CFIT?

Anna Wallace: Yes, there are sections of the market that do not participate with CFIT.

The Chair: What would they be?

Anna Wallace: There are not whole swathes of companies. The first would be a company that is not interested in the coalition, so it is not interested in the topic. The second might be that it has a commercial interest and does not want to engage with this collaborative effort, so it may not want to get involved. Another may be that it is not aware of the work. As much as we try to publicise the work and be open in terms of firms engaging with us, I am quite sure there are some entities that may not have heard of us—we are only three years old—and have not engaged with CFIT.

Q14            Lord Lilley: My question is prompted by Lady Lilley, who asked me this morning, when I said I was going to discuss sandboxes, what a sandbox is. I was unable to explain it to her because no one has ever given me a concrete example, not of your coalitions but of a basic sandbox. What happens? Does someone say, “I have just invented this new product, a digitalised pension” or something? Could you take us through the steps of what it involves, please?

Anna Wallace: Yes. I can take you through the steps as I knew them. I think it is fair to say that things have evolved. There is no such thing as one regulatory sandbox. Many have been spawned since that time, but I can talk you through the original concept.

The original concept was that after the great financial crisis there was significant reluctance to bring innovation into financial services. We saw a number of drivers for that. We saw that the regulator was hesitating in allowing new innovations into the market because of what had happened under the great financial crisis. We saw small entry firms, start-ups, struggling to get access to capital. When they said to venture capitalists, “I have this great innovation. Please fund it”, they said, “What market is it in?” They said, “It is financial services”, and we saw the VCs looking to markets that were potentially less regulated as more attractive ways. We also saw large firms reluctant to develop new products based on what had happened as a result of the financial crisis. That was the context within which we were operating.

Our goal was to encourage greater innovation in the interests of consumers. That was what we were looking to do. In the original concept, we devised a number of approaches that we could take. We said that for small firms we could try to get them to market quicker than going through a normal authorisation process. That was the tool of restricted permission. That enabled them to not go through the full licensing or authorisation process, but they could go through what we called a limited permission process, a restricted permission, which said you are allowed to operate in the market with real customers under certain conditions for a limited period of time and at that point we will consider whether we can give you full authorisation or whether you do not get full authorisation. To be honest, the firms had that choice as well. That was the limited permission route that enabled them to get into the market quickly in a restricted way, to prove both to us as a regulator and to their investors that this actually could work within a regulated sphere. I think by far that was the most used in my time.

We were also able to give individual guidance to firms. Again, they are in the sandbox. The core concept is that they are testing with real customers what they are doing, but they may need a range of tools to enable them to do that. The other tool to enable them to do that would be guidance, and that would be individual guidance to that firm, about the application of FCA rules and guidance specifically to that context. There was no need for that firm to take general rules and guidance and try to interpret it within its specific context.

The third was a no action letter, which basically was around, I guess, the concept of safe harbour, when large firms say, “We need safe harbour to operate”. Safe harbour is an industry term. I do not think it is really used by regulators. We developed a no enforcement action letter approach. That one we actually did take from the US CFPB, which came up with the concept of a sandbox before the FCA, but I do not think it was used by the CFPB and it was not really used by us. The point is that they can test this innovation with real customers. They can get regulatory feedback. The devices to enable them to do that were, broadly speaking, authorisation, as in licensing; policy, as in individual guidance; or enforcement—a no enforcement action letter. They were able to carry out that test.

I think I should impress upon you the significant evaluation and due diligence that the team would do and input into that test. Often, what we found was that firms did not necessarily consider up front the impact on consumers. We would often build in protections for consumers for that test. It was an opportunity to test out a business model with real consumers, which in lots of unregulated markets is not something you need help with, but in a market like financial services, in the wake of the great financial crisis, this felt like an intervention that was needed to support innovation and getting into the market.

Lord Lilley: What would happen if, having got restricted authorisation and flogged the product to a number of customers, it was then found to be a bad thing?

Anna Wallace: One of the things we would do as part of the test is think about the potential consumer harm that could be caused. There were a number of interventions we could make at that point. We could determine what types of consumers were engaged—for example, we could say, only sophisticated investors who have already practised this before. We introduced a number of interventions to prevent that happening. That is the first point.

However, this was not a risk-free environment. We would also think about the harm that could be caused by the consumers from that very specific product for that very specific test. We would often require the company to hold effectively a pot of money for compensation for any consumers that were involved in that test. We would decide with the company what that compensation might be. It may be consumers made whole; it may be something over and above the consumer being whole from the failure. If they could not do that, because obviously, that was effectively a form of regulatory capital we were asking them to hold, then they were not able to test the product.

Lord Lilley: Thank you. That is very helpful indeed.

Q15            Lord Eatwell: I am intrigued by the fact that the idea of the sandbox, I suppose, is about 20 years old. I do not mean the FCA one, I mean in general. Yet somehow the whole concept has not been seen as a success. For example, can we identify some extraordinary innovations that have come through the sandbox process? I cannot think of any. I am sure there may be some. I am wondering whether the problem is that the very nature of the sandbox is to present new ideas and new ways of doing things and to test them with the regulator.

People get sensitive about what they deem to be commercially sensitive information. For example, you talked about groups of companies doing things. That already suggests some form of strange competitive or anti-competitive relationship between them. Am I right in thinking that this is a problem? The concept seems to be a great idea. There was innovation going on. Innovation was expensive. You test it out with the regulator so that you do not spend an enormous amount of money, and then have your legs cut off when you finally introduce whatever the new product might be. It is a very sensible idea, but it has not really taken off, has it? Or do I have a false perception in this respect?

Anna Wallace: I think probing that is incredibly important. In between my time at the FCA and at the Centre for Finance, Innovation and Technology, I joined the Gates Foundation to help it in thinking about expanding digital financial services into emerging markets. Many said, “Let’s just industrialise sandboxes”. In fact, that was already under way by the time I joined the Gates Foundation, but it clearly was not an appropriate regulatory tool for the challenges that those regulators were facing. They had innovations coming into their marketplace that included 23 million customers a year. I said that the problem our sandboxes were trying to solve was that there were innovations but there was no customer acquisition. So it was a very different context that we were trying to apply.

I think you are right to question the validity and the market impact of the regulatory sandboxes. I have been reflecting on that since I was asked to come to this committee. Our measurement of success was around these firms coming into the market and the impact of that on the consumers. That was very early days. I think the 2017 Lessons Learned report that we published attempted to engage with the market on how we saw success. Obviously, at that point, that was very short-term, firm-level outcomes.

In terms of attributing the work of the regulatory sandbox to broader market outcomes and broader consumer outcomes, I think there are still some questions to be answered. I believe a longitudinal study could be very helpful in not just the UK but the global understanding of how and where these tools can be beneficial to financial innovation and the outcome for consumers and markets.

Lord Eatwell: Yes. The other thing that puzzled me in your initial statement was this emphasis on benefit for the consumer.

Anna Wallace: Yes.

Lord Eatwell: A lot of innovations have, of course, taken place in market infrastructure and in wholesale markets, yet they can affect the consumer. For example, the innovation of credit default swaps transformed the mortgage market and therefore transformed the experience of the consumer, but nobody would have thought that credit default swaps were something to do with consumers. They are clearly a wholesale market product. How do you define what is good for a consumer, whether a product is consumer related, whether it is directly sold to consumers or whether it will affect the consumer indirectly, as credit default swaps did?

Anna Wallace: I think that is an excellent question. The look-through is often missed. We can often index on what a particular sector does or delivers, but in terms of how the whole of the market has evolved in response to that, certainly, we did a number of wholesale propositions, in particular as crypto was coming into the market. I think at that point it was called Bitcoin and cryptocurrency. The regulatory sandbox was able to say that it thinks the benefit probably comes from market infrastructure around the distributed ledger technology. We were able to use the sandbox trials to then inform a piece of policy work, which was around crypto assets and their interaction with MiFID, the market infrastructures directive as it was then and which is now a piece of UK legislation.

I think the point you are making is a good one, and I do not have the evidence to give you that says, okay, if the market structure changed and enabled certain activities in the wholesale market, what impact did that have writ large on financial services and what does that mean for the end consumer? In that circumstance we could talk to, and we often talk to, access to financial instruments direct to retail consumers. That was one of the claims by these innovators, that it could disintermediate banks and brokers and that would be more efficient. It was also a way of trying to cut through the rhetoric of innovators and understand whether that was a feasible claim to be made.

Lord Eatwell: Could we characterise this whole endeavour as something that is a proposal to try to reduce the risk in innovation, and this is particularly regulatory risk?

Anna Wallace: Yes, absolutely.

Lord Eatwell: It has not really worked, has it? That is the puzzle. That is where I go back to my initial point about commercially sensitive information or whatever. I am trying to work out why it has not worked. Are there any other thoughts, given that you are in this position now that looks across the industry? Do you have any other thoughts as to why it has not really worked?

Anna Wallace: It has not worked. I think it is a good question to ask. There is some emerging evidence that shows that there are elements that have worked. For example, even just my own perspective in the market, working for the Centre for Finance, Innovation and Technology, when I set up the regulatory sandbox the banks would not even engage. They effectively saw it as a way to bring in intelligence that we would take enforcement action against. That was their attitude. Whereas now the Centre for Finance, Innovation and Technology has broad engagement from banks, which are now engaging with the regulatory sandbox.

Similarly, if you look at market sentiment—and I appreciate the perspective on market sentiment—there was a sense that the regulator was attempting to support financial innovation. I think you have seen the evidence so I do not want to labour the point. Did they get into market quicker? Yes, they did. Were they able to get access to greater funding? Yes, they were. That was evaluated by the Bank of International Settlements as well. Were the consumer safeguards—

Lord Eatwell: Is it BIS evaluation of that?

Anna Wallace: It is, but only on the firm side.

Lord Eatwell: I see.

Anna Wallace: I think that is an important, credible piece of work. The BIS may not want to go into impact on consumer; that is not really its bag. However, it looked at the counterfactuals. It looked at those who came in and those who did not get in, and they compared the two cohorts. It was an exciting piece of work, and that is a fundamental cornerstone of why other countries have copied it.

Lord Eatwell: Thank you. That is very helpful.

Q16            Baroness Donaghy: Like Lord Eatwell, I was intrigued by your initial statement that in a sense the sandbox is a methodology and there are other things that are necessary to achieve what the initial intention was. I hope I am not putting words in your mouth. It seems to me what you are doing now is more about culture change, co-operation, building trust, which is a bit more difficult to get your teeth into but it may lead to a bit more progress in this area.

On the original intention, we keep hearing about the cost, that it is a very high cost to an applicant to come into the sandbox and get their proposal considered. Can you confirm or deny that cost is a big factor? There is also, of course, the cost to the FCA, presumably. How do you avoid the favoured applicant? Is there evidence that those who succeed in this experiment actually also receive funding that an unsuccessful company would not? Which comes first? Sorry, that was not very articulate.

Anna Wallace: Please correct me if I get the last one wrong. In this 2017 Lessons Learned report that I talked to, we tried to set it up in a way that could be continued through the years, and you would begin to get a greater sense of market impact. One of the indicators was reducing time and cost to market, and that was a specific cost to firm. To us the counterfactual of that was twofold at that point. I think there are more ways to get into the market today than there were 10 years ago. The first was being directly authorised by the FCA, which has a cost to get authorised but also the cost of developing the application and understanding the requirements of the regulator. The second was an appointed representative regime, which was where a firm maybe was not big enough—and it only exists in some markets—but is able to be licensed to operate through another party that is regulated. This is quite common in the adviser market. We compared what it cost for the cohorts that had come through the sandbox to that point with those counterfactuals, and we found that we had reduced that cost. I think that is an important test for the sandbox. As market conditions change, as regulatory models change, that may have to be recalibrated over time. There is not one cost of getting into the market, so I think that is something worth while but I cannot confirm or deny that through my role at CFIT.

Our costs were very closely monitored. I think there was a freedom of information, but we put out our costs as well for the regulatory sandbox. We were very conscious of whether that met the balance, and we did that through our audit co and board. At that point, we were satisfied against the Lessons Learned report that there was a public good and a public value. The point that is so incredibly important, which I think Lord Eatwell made, is that this is about the regulatory risks to innovation. The FCA cannot do everything to support innovation, nor would we want it to. It is a hierarchical, bureaucratic, independent regulator. We do not want it driving all innovation. I think we would be in a poorer position in the UK for that. I have not reviewed the current cost of the regulatory sandbox and how that compares to outcomes, so I would not be able to give a position. All I can say is that it was constantly monitored.

Then on the anti-competitive side, what I think you are asking is whether the benefit for those that get into the sandbox could be unfair, that the regulator is providing unfair support. This was something we were very conscious of. Our prior, when we went into it, was that innovators were at a disadvantage, and they were at a disadvantage in a number of ways. From a licensing perspective, if I was doing something quite common, I would probably have some guidance that told me how to apply. I would have people who worked in that licensing department who knew exactly what to do and had guidance and best practice to do that. I probably had some policy out there that helped me, some guidance that helped me think about how to apply it. So one of the anti-competitive measures that we reflected on ourselves, as the FCA, is how we approach policy-making, guidance-making and licensing having a perverse effect. It was intended to redistribute that.

Then on the support for small companies, whether the banks will always agree with this perspective, at the time we viewed that having a supervision team could be helpful because you would have a big team that could answer your questions about whether it should be this product or that product, or doing it this way or that way. You could get this informal advice. We felt also that the sandbox could support those small companies that did not have that close contact with the regulator where they were doing something innovative. We saw it as a means of redistributing unfairness, but you are correct that where that line is constantly needs to be monitored and evaluated. I think it is an exceptionally good point to make.

Q17            Lord Sharkey: I think in your opening statement you mentioned that the FCA has not published an updated analysis of the sandbox impact since 2017. I wonder why that is. I also wonder why there are not any cost-benefit analyses available for us to look at. The question arises: in the absence of CBAs, for example, how are the learnings from sandboxes distributed or shared among those outside the sandbox and, in particular, consumers?

Anna Wallace: My opening statement was not intended to say that that has been the only evaluation of it; it is the evaluation of the original concept as I knew it. I believe since then there have been reports. I think the 10 years since the regulatory sandbox provides insights. I think there has been a range of ways that the FCA has reported on the progress and evaluation of the sandbox. This is the one, for me, that relates most closely to the original intentions of it and one that I see could be the basis of a longitudinal study on broader market impact and consumer outcomes.

Sorry, excuse me: what was the second part of the question?

Lord Sharkey: How are the results of the sandbox distributed or shared, and to what extent and among whom, including retail consumers?

Anna Wallace: There are a range of ways. I guess transparency is probably the—there was the publishing of anyone who was in the sandbox and the concept of the trial. There was always a publication of who was in it and what they were doing. There were publications—I can provide you with further details; sorry, I do not have it in my mind—on the results of cohorts. That is where there were cohorts. We were able to give a broad range of outcomes and themes, I guess. As a result of that, there was also consideration of whether regulatory intervention was required off the back of that.

At that point, that is where something like a cost-benefit analysis would be triggered because that would be a necessary part of policy-making or making guidance. However, as the regulatory sandbox was an operational function as opposed to a part of the legislative framework, there was no cost-benefit analysis. The 2017 report was intended to be an attempt to look at what we normally would from a cost-benefit analysis, so it is a CBA by proxy.

Lord Sharkey: Are there any learnings from these sandboxes that end up being proprietary—in other words, restricted to the outside? Obviously, one of the things we are worried about is the creation of unfair advantage by presence in the sandbox. There is a slight logical paradox in all this—

Anna Wallace: There is.

Lord Sharkey:—but the pointed question is: what is restricted to the participants and what is not shared, therefore, outside?

Anna Wallace: Certainly, individual guidance was not a huge part of the regulatory sandbox. That being said, there was regulatory feedback, is I think the term, or informal guidance that was given. To say that no information was provided to that firm that could have been useful is probably unfair. However, there was a test as to whether this would be useful to the rest of the market and, if so, what the approach is to get it out into the marketplace.

Probably the one example I would give is on the robo-advice function. This came out of the Financial Advice Market Review. I think it was a cross-policy from the Treasury, the FCA and the Bank of England. It identified that there was additional support required for entities that were looking to bring—there are many terms for this, so apologies if I am crude and not totally accurate—limited advice for consumers. At that point it was robo-advice because that was a very long time before we knew what AI could do. That was guidance. We identified that because being on the wrong side of the advice/guidance boundary was so significant we provided independent advice.

Lord Sharkey: It is more significant now than it was then.

Anna Wallace: Yes. That was an effort to try to put things into the marketplace. That is a good example of how we were asked to consider it for the sandbox. We said it is not appropriate for the sandbox, so we will devise a new innovation service that is much more appropriate and can get lessons to the market.

Lord Sharkey: As an observation, I am struck by the fact that nobody has told us of some absolutely radical finding generated by sandboxes that has benefited everybody that has gone near it. Is this because there is not any?

Anna Wallace: In a market as deep and as complex as the UK’s, I would be amazed if there were one silver bullet that unlocked all activity. Ultimately, this was targeted at trying to find where the flowers bloomed and bringing that up to the next level. I think what we need to look to is the association of what came through the sandbox. Crypto-assets is a very good example, where there was really no other way for us as a regulator to get our hands on what people were talking about. Ultimately, it was seen as an unregulated activity, and it was the first place where we could begin to see that. There were a number of wholesale share certificate experiments that were taken. That is the early days of what we have at the moment in terms of the Wholesale Digital Markets Champion review.

The question is not necessarily, did nothing come from it, but why is it not happening fast enough and at the scale that we need it to? That is why, for me, working at CFIT is incredibly important because I think there are some missing elements in terms of trust and collaboration that a sandbox and that market infrastructure piece does not necessarily unlock. It shows the potential at a product or firm level, but we also need to think about what market infrastructure we need to enable that to scale to the next level and get that into the hands of consumers and users of financial markets.

Q18            Lord Hollick: One radical agent of change in the financial services market is the rapid adoption of AI. It is reshaping the relationship between the consumer and the producer of financial services. It is posing very significant questions to the regulator. How does it regulate AI? With agentic AI coming into the market as well, the whole relationship and the whole regulatory basis upon which the financial services market operates is being changed. What is the sandbox doing to investigate this and to make recommendations about how AI should be regulated?

Anna Wallace: Obviously, I do not work at the FCA any more, so I do not have an inside track on this one.

Lord Hollick: What is your organisation’s perspective on this, then?

Anna Wallace: As I was preparing for this, I was thinking about it in three ways. I am not saying the sandbox is not the regulator—of course it is—but what do government need to do; what do the regulators need to do; what does industry need to do? That is the way I always think about it. That relationship and what those various players need to do changes with innovation, from the very nascent to the point at which it is systemically important. What different players need to do alters throughout that innovation lifecycle.

I think globally, the FCA would be recognised as being on the front foot when it comes to artificial intelligence. It has sought to adopt the regulatory sandbox model to think about what is required from artificial intelligence. From that context, it has tried to evolve and adapt the offering for the needs of artificial intelligence, but that is obviously limited only to that innovation service.

From my perspective, trusted AI adoption will need shared standards, assurance methods, data infrastructure, and arrangements for identity, consent and liability. I think the FCA’s AI sandbox will give us some lessons and insight into what that might look like and what the market practices are. It is an excellent way of us understanding what the regulatory risks are and what is likely to come into the marketplace. It gives the public an exposure to what is happening in the market, and it enables the regulator to think about whether its rules, licensing and permissions are appropriate or whether it is something that it needs to alter. It does inform policy-making in a way that is more practical, potentially, than the normal policy-making consultation process.

I think they are very focused, through the Treasury, on looking at agentic payments through AI initiatives as well. Again, that is a policy intervention. That is not a regulatory intervention or a sandbox intervention. What you are seeing is the regulator responding appropriately to what is required from a sandbox, what the limits are of what a sandbox can do, and where a policy could support that.

From my perspective, what I am hearing from industry—and this has come from the AI Champions’ report to the Treasury—is that there are things they can do and they are excited about the commercial opportunity, but there does need to be some standardisation across the industry around assurance and use of data. I think the industry is quite willing to work together on this. There are a few players that are obviously trying to get ahead. Personally, from a public value perspective, and the reason why CFIT is interested is that, we want to see those that maybe do not have the level of investment required for these assurance models benefit from that thinking, and for there to be a UK approach that can protect consumers.

I am also concerned about SMEs in the UK. While it is not a financial services issue, if financial services begin to adopt agents for payments, I am concerned that UK SMEs that are not necessarily engaged with that digital economy may lose out as well. A CFIT model that is looking at the public value of AI, is thinking about catalysing the innovation but doing it in a safe and sustainable way and thinking about where the winners might be and how to redistribute that, I think is a helpful precursor to potentially further regulation as well, or a sandbox. While I think there are guardrails that are required, like any regulation of a market, understanding what can be done at an industry level first and then where the regulatory intervention is required, whether that be through a sandbox experimentation or through policy and guidance, is a helpful way of us collaborating. For me, I see that it has to be a joint effort from government, regulators and industry. It is not all sitting on a sandbox.

Lord Hollick: The sandbox can play an important part in evaluating it.

Anna Wallace: Yes.

Lord Hollick: Another innovation over the last 10 years has been the advent of influencers whose business model is to promote financial products for which, assuming they are adopted, they are rewarded. They seem to sit outside a lot of the regulatory perimeters. Have you looked into the role of influencers and does your organisation have any view as to how their role should be properly regulated?

Anna Wallace: It is not something we have looked at in CFIT, but I did lead the financial promotions team for a while at the FCA so it was a big part of my job. That is when Twitter was only just starting. We tried as much as possible to capture anyone who was doing anything for material gain. It is technology agnostic. Regardless of the platform, regardless of whether it looks nice and friendly and cute or it looks like an interview, if that person is doing it for financial reasons—it is called “by means of business”—they are captured. It was incredibly challenging at the time, in terms of both supervisory and enforcement, to capture these people on these new platforms and the sheer volume of it. Observing what the FCA has done since I left, it has been deploying AI tools to get greater market understanding and intelligence.

I have not thought about what we could do from a regulatory sandbox perspective, but you are right. The digital economy and the drive of innovation and the use of these platforms is certainly a unique, complex and very fast-moving risk to consumers.

Q19            Lord Vaux of Harrowden: In one sense, sandboxes have been incredibly successful in that they have been copied pretty much all over the world now. Can you tell us if there are any lessons we can learn from what other countries are doing in respect of sandboxes, where things are being done better or perhaps being done worse?

Anna Wallace: I think what the UK has done incredibly well is to attach the sandbox to the market failure and stay true to that. Where other sandboxes have struggled has been with entry into market. I think that is too broad. It being well-resourced in the early days, as we were honing it, has also made it a more successful and productive function when compared with others, and that has continued. In the offering we saw in particular two examples where there was a need from innovators to work across sectors. With AI, certainly from the CFIT coalitions that we have been working on, this cross-sector engagement from regulators is incredibly important. We are working on one at the moment on open property, which covers a range of regulated markets. Regulators working with each other on trials, which we called cross-sector sandbox, was an idea that we looked to. Then we also looked to international sandboxes, so cross-border sandboxes. I believe the lesson learned was they were both too early. I think we were further ahead and we were better resourced.

The other point to make is that comparisons are not always comparing like for like. Often the FCA is likened to the Monetary Authority of Singapore in terms of success. It has a very different mandate when it comes to economic growth. The FCA, while it has a new requirement, it is quite distinct, so sometimes those international comparisons are not necessarily doing like for like. Potentially, for the UK to deliver that outcome, it would require government as well as the FCA, so something like the Treasury’s Office for Investment is a response to that comparison.

Lord Vaux of Harrowden: Given the new growth and competitive objectives, is there anything that comes to mind? Singapore is a good example. Is there anything we can learn from that? Is it doing things? What is it doing there with sandboxes that is perhaps genuinely driving innovation or growth?

Anna Wallace: Genuinely, on the numbers of companies coming out of sandboxes, the FCA was streets ahead of any other regulator.

Q20            Lord Vaux of Harrowden: Also on the international thing, we were given the example of Canada last week in the evidence, where people go into the sandbox and they have a limited authorisation. They come to the end of that and rather than forcing them on to the full authorisation process, they end up with an individual waiver that seems to last for ever. Is that something that happens in this country as well?

Anna Wallace: You can have waivers. That is a really good comparison. I did work with a number of countries on their regulatory sandboxes because sometimes they could not be persuaded otherwise. Some of the design questions that came to mind were: how long were you in the sandbox and how long would these interventions last? At the FCA we were mindful to keep it quite short, but Indonesia is an example where you could be in the sandbox for a number of years. Yes, unlimited waivers was another example. Ultimately, the purpose of the sandbox was to test whether it could operate in a regulated market. If a waiver is required ad infinitum, then the policy should be changed.

Lord Vaux of Harrowden: Or the regulations.

Anna Wallace: Yes, that is what I mean—sorry, the regulations; exactly. If there is a need to continue that, then to me that moves into a policy or regulation change.

Lord Vaux of Harrowden: Are there examples of that here?

Anna Wallace: Not to my knowledge, but this would just be from my time. Yes, nothing comes to mind. If I do recall something, I will provide evidence.

Q21            Lord Davies of Brixton: I am wondering to what extent sandboxes are just about facilitating new developments as opposed to encouraging them. I am sure those two concepts overlap, but what we have heard in practice is that it is really just a process that delivers innovation. By making it easier, maybe that encourages but it does not inherently create new ideas. That is a comment. Am I off base?

Anna Wallace: No, I think it is an excellent point.

Lord Davies of Brixton: We were told last week that fintech has flourished in the States and sandboxes do not have a significant role.

Anna Wallace: No.

Lord Davies of Brixton: So there is innovation and facilitation.

Anna Wallace: Yes. The US is a very different market, as a comparison. The facilitating and not encouraging point is a good one. It absolutely facilitates, so I think that is a good point.

Lord Davies of Brixton: You can see that?

Anna Wallace: Yes. In terms of encouraging, it is certainly a factor. From the CFIT perspective, when we talk to coalition members around what we can do to drive innovation, and that is address systemic barriers to innovation, the prospect of using the regulatory sandbox is welcomed. That is large banks, not just innovation functions. In respect of Digital Company ID, the prospect of getting feedback on the interaction of policy and rules regarding this practice, this particular selection of firms, has been seen as a positive.

Q22            Lord Davies of Brixton: Something else that came out of last week’s discussions was the suggestion that the real benefit of sandboxes is regulator learning. Do you think that is an important element?

Anna Wallace: I do think it is an important element. I think that is a market-wide element. It has led to, for example, the AI survey that has just come out today from the FCA and the Bank of England. That was as a result of the innovation services. We saw that we did not have line of sight. We could see the deployment in small pockets, but we did not really understand what that looked like across the marketplace. That was a recommendation from the innovation services nine years ago. The policy-making around crypto assets is a result.

While there is learning, that learning then produces regulatory interventions. With the Financial Advice Market Review, that learning then went on to market activities, which then went on to further guidance that went beneath that. Yes, learning is incredibly important, but that learning also does result in action by the regulator that the market benefits from.

Q23            Lord Turnbull: In much of the questioning so far there has been an underlying tone of scepticism—that some are possibly not all they are cracked up to be. I will continue in the same vein. Maybe you can help me under two headings. One is progression. A firm spots an opportunity with customers, an asset or technology. It negotiates on certain waivers of the regulatory regime and takes that forward in the sandbox. What we do not hear much about is the people who go through that process and graduate from it. Is this more like Peter Pan, a place where no one grows up? Where are the cases of someone who says, “Ten years ago, I went in the sandbox. I developed something and I have taken it on from there, and it is now a roaring success”? We are very short of that information-sharing—where it has made a difference and how, if your experiment goes well, you can then develop it still further.

The other question is the one that has come up about conflict of interest. A firm gets into discussion with the FCA and it comes up with a bespoke package that relates to it and its proposition. It may fail, in which case it just gets timed out, but supposing it is doing well. What does the FCA then do? What does the company do? “We have developed this and we should retain the benefits of the effort we have put in, the intellectual property we have developed.” Whereas the FCA may say, “No, our duty is that we should be taking that knowledge and spreading it, making it available to other players in the market”. Which way should the FCA go? Having helped someone discover something that works, does it just sit on it and let that company continue with it, or does it have a duty to try to find someone else who could use it?

Anna Wallace: On the first point, around progression, the participants in the sandbox are a matter of public record. We could certainly do a study that could look at the impact and where they are now. I believe the FCA is—

Lord Turnbull: We have 200 cases of firms. There must be someone there.

Anna Wallace: I did a quick check outside and certainly, one of the early people in the FCA sandbox is valued at more than £100 million and has adverts on the TV. I would be very cautious, without doing my own personal due diligence, without having the insight that the regulator does into those companies, to laud them as the specific use cases. However, that work can be done and there are companies.

One thing I was always conscious of and one thing we always spoke of was what success was. It was the valuation and the impact that that one company could potentially have. It was also whether it encouraged greater innovation within that area. One of the things we also observed was that large financial institutions began to innovate in spaces that were like the ones that were coming from the sandbox. Our assumption was that these were product ideas that they had on a shelf somewhere and decided not to do. We also looked at the offerings of large financial institutions and we looked at the partnerships between those fintechs and banks as well. There are examples of companies that went through the sandbox that are now large financial institutions.

Lord Turnbull: Do you think more should be done to make that knowledge available?

Anna Wallace: That is why it is the change in the context, but also the benefit of us really understanding the impact of this and ensuring that, when we think about AI, we ask, “Is this the route we want to go down or are there other ways we want to go down?”—that we look 10 years on and have a proper longitudinal study to understand the impact and inform our next 10 years of financial innovation offering. Yes, there are small companies that were being cited £60,000 a month to be regulated under a principal in a market and they decided, “No, we want to become directly authorised through the FCA and we are going to use the regulatory sandbox”, and they are now valued at over £100 million.

Yes, I agree: I think more needs to be done to establish that and understand whether those companies have gone to market directly. Remember, one of the contexts of this was the challenges that we had with the neobanks getting into the market. One of the challenges we were also facing was the peer-to-peer market and getting that novel innovation into the market. There were years and years when we stalled as a country in getting this new entity into the market. What the sandbox has tried to do is to get more product innovation, rather than saying, “I am a bank”. It is more, “I am providing a certain service of a bank” or, “I am providing a certain service of an insurance company”. What will be interesting to see is whether those companies have now broadened out their offerings and are becoming banks themselves or are becoming the underwriting insurance company themselves. That route to market will be important.

The joint venture point will be important. One of the concerns we had in the early days was that the innovation would go out. It would all sound great at the end of the sandbox, they would be bought by a large institution, and the innovation would be taken out of the market. That was something we were also very concerned about. Looking at those lessons and that progression is important.

On the conflicts of interest, I guess that speaks to the points I have been making. Once someone has something that we think is of value to the rest of the market, the FCA would attempt to put that into the market in the most appropriate way, whether that be through the publication of the tests and the outcomes and the FCA’s view of it, or through—forgive me, I am forgetting what they were called. They were not policy documents. They were almost White Papers or something like that, in which we would put out insights into themes as we saw them. We also put out guidance and rules. We attempted in a number of different ways to put out the lessons, in a way that was appropriate for a regulator that could support the rest of the market and say, “Okay, this innovation is something that could work in a regulated market, or, “I am now benefiting from the guidance that one firm spent a lot of time and effort working with us on”. Because remember, when they are coming to the FCA to do the sandbox, it is also costing them time and money. If they get guidance and policy advice or regulatory advice that is helpful, there is also a test of when to put that into the market. Is that calibration always right? That is an individual decision taken within the context at the time and in light of the other information. Maybe looking back 10 years, with hindsight we are able to say we should have got that guidance out sooner. That was clearly an emerging area that other market participants could have benefited from.

Q24            Lord Sharkey: If I could just ask one short question, who owns the data generated by the sandbox?

Anna Wallace: This is interesting because this is what I spend my time at CFIT and the Gates Foundation thinking about, in terms of intellectual property. One of the challenges around bringing many players to the table is understanding what is background technology and what is funded development. Intellectually, what can we say belongs to the group and how can we ensure we maximise public value from that testing and that collaboration? That equation is slightly different as a regulator. I do not have the up-to-date thinking on this. The FCA may have evolved its thinking beyond that.

Ultimately, as I see it—and I may evolve my thinking after this—the IP of the regulatory advice was ours and that was ours to share with anyone as we so wished. Before you were in the sandbox, you had to sign up to a certain amount of information being shared about the test. That was our way of ensuring that, without giving away the proprietary information necessary for them to take it to market, the benefit they got from what they were doing in the sandbox was transmitted. We had the rights to transmit that to the rest of the marketplace. Again, the devil is in the detail on that one.

Q25            The Chair: Finally, from the perspective of the work you do at CFIT with the people in the market, is there an area that you would like to see improved or changed in the way the sandboxes are run, for the benefits that you are trying to achieve?

Anna Wallace: It does not sound like this committee is of the view that the FCA should be doing all the driving of innovation.

The Chair: We are natural sceptics.

Anna Wallace: A sceptical audience. This is not a uniform view. There is a significant amount of pressure on the FCA, and probably the Bank of England as well to a lesser extent, to be entirely responsible for the level of innovation in UK financial services. There is a broad view of it. I take the view that there is a place where a sandbox should begin and end. I hope that my evidence today has tried to show where I personally saw those lines. We believed that there were a lot of sceptics in the market who would not want to take innovation to an FCA. We also believed that not everything needed regulatory feedback. They did not need to de-risk the regulator’s perspective. Not everyone needs a new licence or authorisation to get into the market, so it could just be a product innovation in a bank.

What we are seeing at CFIT is that we are now at this point of innovation where we have seen the innovators come into the market. We have seen the capabilities, other than, obviously, artificial intelligence, which I think we are still at a very nascent stage of understanding and we need more information. There are now some structural barriers that may be impeding this innovation benefiting the market, the users, and getting into the hands of consumers and helping them to benefit from it as well. As such, what CFIT tries to do is to bring those players to the market, with government and regulators, and identify ways and give a space for industry to do that collaboration before getting into a sandbox or policy changes.

There is a time when a sandbox is too early. There is certainly a time where regulation is too early. If we move too quickly to regulate, we may well inadvertently stifle that innovation. Something like Digital Company ID—where it is not really about what one player does, it is whether a group of players can trust and rely on information that is provided in a certain form at a certain time for a certain purpose—requires the market to take a view on it. CFIT is not representative of everyone. We are looking to drive and scale that innovation and that is why we have a very close relationship with entities like UK Finance, like Innovate Finance, like government departments and regulators, so that when we begin to see something emerging from the market, they can begin to think about it from a whole-of-market perspective as well.

I am hearing scepticism in the room on whether the sandboxes are being used in an inappropriate way—that maybe more formal, traditional regulatory tools should be used—and a thirst for, I guess, evidence to back that line up. I would encourage the committee to also look at that piece before the regulator touches it and think about whether we are actually generating the right conditions. It was the Treasury that set up CFIT to try to support that market co-ordination problem before regulation and policy took hold. Looking at it as a whole ecosystem could be helpful, to think about where the line begins and ends.

The Chair: Good. Thank you very much for your evidence this morning. It has been very interesting. You have been very generous with your time and very fulsome in your replies, so we are very grateful to have had your experience today. Thank you.