Market abuse requires a dynamic response to a changing risk profile

Speech delivered by Julia Hoggett, Director of Market Oversight at the FCA, at the AFME Implementation of the Market Abuse Regulation in the UK event, London.

julia hoggett 240 180.jpg

Speaker: Julia Hoggett, Director, Market Oversight
Event: Implementation of the Market Abuse Regulation in the UK, London (organised by AFME and hosted by Linklaters LLP)
Delivered: 13 February 2019
Note: this is the speech as drafted and may differ from the delivered version

Highlights

Compliance with MAR requires a whole series of situational judgements to be made.
A regulatory system that relies on controls that work by detecting when an event has happened, will never be as effective as a system that also helps ensure that misconduct does not happen in the first place. Whilst it is important that both regulators and participants have controls to detect suspicious behaviour, it is critical that we consider how we protect the market from such behaviour occurring on our watch in the first place.
Firms need effective risk assessments and must consider for themselves the manner in which their systems and controls evolve as the risks within their businesses evolve.
Control of information leaving a firm is as important as control of information moving within a firm.
Firms need to take responsibility to ensure staff understand the consequences of unlawful behaviour and must be vigilant to ensure it is not happening within their walls. Greater awareness of the risks of market abuse is needed.
Access controls, surveillance capabilities and general mindset in investment banking and advisory platforms is not yet as evolved as it should be.
There is still more for the industry to do to improve its capacity to surveil for market manipulation.

Just over a year ago I gave a speech on the Market Abuse Regime where I said that effective compliance with MAR was a state of mind – an unusual thing, perhaps, for a regulator to say.

What I meant was that compliance with MAR requires a whole series of situational judgements to be made, including by many of you in the audience today.

The decisions you are required to make and questions you are required to answer, are often not easy, but they are critical to the integrity and health of our financial markets. Be they about what constitutes inside information; how it is it managed and released; how you ensure your traders are following the rules; or how you identify inappropriate trading behaviours within your firms.

A number of these judgements are extremely time sensitive and it is my suspicion that this pressure of time will only increase.

Let’s be clear - a regulatory system that relies on controls that work by detecting when an event has happened, will never be as effective as a system that also helps ensure that misconduct does not happen in the first place. To that end, a market relying on post hoc controls will never be as effective as a market with appropriate controls accompanied by heightened awareness - and a common understanding - of the risks and issues posed by market abuse.

When it comes to mitigating the risk of market abuse, you could say that we are not seeking to be in the business of closing the stable door after the horse has bolted.

While my last speech on this topic focused on laying out the role that minimising market abuse plays in enhancing trust in our markets, this year, I want to delve deeper into some of the specific issues that we worry about at the FCA. To help frame how market participants should think about the risk of market abuse taking place and hopefully add useful colour to my colleagues’ recent publications on the subject in Market Watch.

To help frame today’s conversation - let me briefly touch on the FCA’s 5 Conduct Questions approach. The approach serves as a conduct risk mitigation framework that is increasingly used across FCA Wholesale Supervision and which helps regulated firms enhance the manner in which they conduct business.

The first three of the 5 Conduct Questions focus on the identification of conduct risk; ensuring that staff within regulated firms feel and are responsible for managing the conduct of their business; and the nature of the support (including systems and controls) that firms put in place to improve the conduct of their business.

The 5 Conduct Questions approach is sufficiently broad-and principles-based that it can apply to almost every manifestation of conduct risk within a wholesale business and I am going to use it today to talk about market abuse.

Question 1: market abuse risk identification

We begin the 5 Conduct Questions approach with the question: how do firms identify the conduct risk inherent within their business? We do this because it is often the most elemental - but also the most difficult - thing to do.

Some risks appear blindingly obvious once they manifest, but do not occur to people beforehand. That is why developing the effective ‘conduct risk identification muscles’, is critical to protecting your businesses and ensuring our markets remain clean.

In service of the first of the 5 Conduct Questions and a point I often make as a regulator - that you maximise the chance of those you regulate meeting your expectations, if you make it clear what your expectations actually are - I want to highlight some areas in the market abuse space where we remain concerned that the risks we worry about at the FCA have not been appropriately considered.

Criminal versus regulatory offences and how to think about wrong-doing

All forms of behaviour can be found in our market abuse investigations, from deliberate criminal wrong-doing, to regulatory wrong-doing, to errors of judgement that might not meet the definition of serious misconduct. While our teams that police the market abuse regime are keenly aware of the risk that these behaviours may exist in our markets - we are concerned that such an awareness is not present in all market participants.

This lack of awareness can manifest in two forms.

Firstly, a failure to properly assess the risk that your institutions could be used to facilitate a financial crime taking place.

Secondly, the risk that your staff are not sufficiently conscious of the risk that their own behaviours may pose.

Inside information controls and potentially criminal behaviour

Let me start by addressing the risk of firms being used to facilitate criminal behaviour.

The life blood of all well-functioning markets is the timely dissemination of information, without which effective price formation cannot take place. The malignant form of that same life blood is the misuse or inappropriate dissemination of that information.

In the investment banking advisory and sponsor businesses, each and every day, dialogues are held with listed companies seeking to acquire other firms, unlisted companies seeking to acquire listed companies or listed companies seeking to undertake radical strategic transformation.

Ideas are developed, pitched, explored in painstaking detail and, whilst sometimes rejected, on other occasions, those kernels of thought turn into market defining events.

We view these dialogues to be critical to the effective functioning of the UK markets.

However, we are concerned that sometimes, the true value to our markets that these dialogues create, is not reflected in the way some firms think about their need to manage the inside information – and the broader risks that can arise from their work.

A great deal of focus is, understandably, placed on ensuring that firms have information barriers to mitigate the risk of inappropriate flows of information from the private side to the public side. Indeed, I would observe that a great deal of training and work on this has taken place for many years.

I will always remember that sobering moment as a young investment banker, when I was reminded of the criminal sanction for insider dealing and then was taught about my firm’s information barrier policy.

However, we worry that firms do not have as much discipline when surveilling for information leaking outside their building, as we know they do when looking for information being passed within their building.

Over time, we have come to realise that the focus firms place on information leaving the building tends to be on the controls around legitimate disclosure of information, such as through the soundings regime. Whilst we encourage firms’ efforts in this space, we need to see firms consider other ways that information could be leaked outside their building – and thereby ensure that their control framework is adequately mitigating market abuse risks holistically – be that through for example, a cyber-attack, or through their staff inappropriately sharing information outside the office.

Question 2: supporting individuals to conduct business appropriately

Question 2 of the 5 Conduct Questions approach, focuses on ensuring that individuals within institutions understand their responsibilities to conduct business in an appropriate fashion.

In the context of market abuse, one would be tempted to focus on trader behaviour – but I will come to that later as it is not the only place where behavioural risk exists.

It could be argued that the lack of focus on information leakage outside an institution’s walls tends to come from an instinct that the greatest risk comes from people not following the rules correctly. The assumption is that staff are ‘trying to do the right thing’.

Whilst, I would love to have a view of the world where this was the only risk, I fear I cannot make that assumption. In fact, it would be irresponsible for me to do so.

Wherever we have individuals in possession of highly valuable inside information, we should always be mindful that there is a risk of those individuals acting unlawfully with that information - for any number of reasons.

This could occur where individuals make use of the information to trade on it for their personal account – far from the watchful eyes of compliance teams – or where they illegally share that information, enabling others to profit from it.

Whatever the reason behind someone acting in this manner, it is likely that when the person is caught, they will have a career in the industry ruined, risk being prosecuted for committing a crime and ultimately, they will have contributed to damaging our markets.

Firms need to take responsibility to ensure their staff understand the consequences of this type of behaviour and must be vigilant to ensure this behaviour is not happening within their walls.

This as a risk that applies to all staff, no matter their experience. We are equally concerned about the risk posed by senior or extremely senior members of staff, who have the most access to information. These are the people who are much more likely to meet with senior investors and the people much more likely to be allowed to speak with journalists.

Unless you, and firms generally, are aware: That these risks exist; that such events do take place; that it is your responsibility to be alive to these risks; and ultimately, that you must ensure appropriate systems and controls are in place, then there is the potential that this behaviour increases rather than decreases, raising the risk of regulatory action and a damaging effect on our markets.

While firms and individuals being aware of this risk is one component of mitigating it, a firm’s own systems and controls are also a critical element – this goes to Question 3 of the 5 Conduct Questions.

Question 3: inside information Systems and Controls

At the FCA we look at the controls firms have in place across their investment banking and advisory platforms. As a theme, we believe that the access controls, surveillance capabilities and general mindset in this part of the business is not yet as evolved as it should be.

More importantly, it has not had as much focus as the risk profile would suggest it should have.

Firms should avoid overly relying on the permissions process for insider lists and deal-teams. Essentially, move away from the assumption that if someone legitimately has access to information, they will always act legitimately with that information.

Firms should ensure that they proactively review who is permissioned to access inside information and whether access is still required. This will go some way in helping firms mitigate some of the risks associated with inside information.

Taking a step back, firms should also consider who in their institution is able to access inside information outside of specific deal teams.

Some of you will have recently seen ITV’s recent Wednesday night TV show, Cleaning Up, focusing on an insider dealing ring in London. While I recognise the producers’ need to dramatize events – I am not suggesting office cleaners are going around bugging your offices - it is not unreasonable to expect that those working in a range of different functions, from cleaning staff to your head of compliance, but also IT support and other functions, should be considered in your firm-wide assessment of the access risks to inside information. These members of staff often require broad access – or are able to gain access – to information held across your firm.

Nefarious behaviour by staff is not something new and it is not something contained to Wednesday evening TV – many of you here will remember that we have already prosecuted print room staff.

In our ever-increasing technological age, it is also important to recognise that individuals outside core compliance or investment banking roles may also have access to this information and that indeed print room staff may feel somewhat ‘old school’.

We encourage firms to think critically about the front-to-back information management that they need to have in place.

Do firms think appropriately about the conduct risks that exist in holding inside information?
Do they have the right protocols in place to determine who the insiders should be and do those insiders fully understand their responsibilities?
Have they thought carefully about who should actually have access to that information?
Finally, do the systems and controls genuinely mitigate the risks of the misuse of that information?

Financial crime and market abuse

Within the theme of systems and controls I would like to return to a topic I highlighted in my last speech; that certain forms of market abuse are in fact a financial crime and therefore the necessary interplay between your financial crime and market abuse systems and controls.

To further build out good practice regarding compliance with SYSC 6.1.1R, last December we published an updated Financial Crime Guide following a useful consultation process.

This piece of work marks an important step for us. Not because the update to the Guide provides any new rules for firms to comply with, but because it provides a clearer understanding of our expectations in this area.

In some firms, we have already observed financial crime policies and procedures which cover insider dealing and market manipulation. Others however, do not appear to do so.

But there is a wider theme that I would like to pull out of this work.

Fundamentally, we recognise that by being more explicit about the fact that certain forms of market abuse are a financial crime – and that therefore you need to think about the inter-connectivity of surveillance tools, STORs and financial crime systems and controls – we can trigger a helpful industry response.

I can now tell you that since we began engaging more overtly on this subject, we have seen a decrease in the number of STORs filed by certain firms. I can also tell you that the driver behind this is that those institutions have offboarded certain clients who were repeatedly behaving suspiciously.

This also reflects the importance of firms undertaking effective risk assessments and considering for themselves the manner in which their systems and controls need to evolve as the risks within their businesses evolve.

FICC, commodities, FX and manipulation

This leads me on to another point I made in my last speech, but unfortunately, I feel needs to be made again: there is still more for the industry to do to improve its capacity to surveil for market manipulation - as opposed to insider dealing - and there remains a need to improve surveillance in the non-equity space.

I may as well also repeat my line; that we continue to believe that the assumption that market abuse only happens in equities, still needs to be thoroughly broken. There is more to be done here – it is not, after all, as if we have not seen significant conduct issues in the FX and Fixed Income markets within living memory.

Now, I can’t be completely churlish, when looking at the latest STOR data[1]. One could say that we have witnessed some improvement in this space with some increases in fixed income reporting.

However, when we look across all markets, the spread of STORs is still heavily biased towards equity-related STORs, and of those, the bias remains towards reporting insider dealing over market manipulation. Insider dealing STORs in 2018 accounted for 86% of all STORs - only down marginally from 87.9% in 2017. Whilst we have seen some progress in relation to fixed income and commodity STORs, we think that the total number of STORs received in both of these areas remains low, suggesting that firms with significant business lines in these asset classes need to do more to detect and report suspicious activity. This is where we will continue to focus our supervisory attention and I would argue we still have quite a lot of work to do here.

We also recently published MarketWatch 56[2] on surveillance approaches which reflected some of these concerns and I would like to pull out some of the themes relevant to today’s discussion. Particularly those relating to undertaking appropriate risk assessments and ensuring that systems and controls, meaningfully and thoughtfully, evolve not just as risks evolve, but also as surveillance techniques improve.

In MarketWatch 56, we highlighted our concerns that firms are looking at the list of indicators of market abuse in MAR and treating that list as exhaustive. We are also concerned that firms are not considering these typologies through the specific lens of the particular business that they undertake.

we only care about systems and controls for one very simple reason – so that they work effectively. The existence of a system or a control, in and of itself, may not necessarily be interesting to us, particularly if it does not work to mitigate the actual risk profile of an organisation.

I have often said that the FCA cares about systems and controls. You would expect us to, we are the regulator after all. However, we only care about systems and controls for one very simple reason – so that they work effectively. The existence of a system or a control, in and of itself, may not necessarily be interesting to us, particularly if it does not work to mitigate the actual risk profile of an organisation. If anything, if it doesn’t work, it would much more likely make me concerned about the firm.

Firms need to undertake a proper assessment of the nature of their businesses, the market abuse risks that may arise as a consequence, and the systems and controls that are most suited to mitigate those risks.

Let me highlight two areas where Market Watch 56 was calling out the need for more of that thinking.

Firstly, in the context of firms using off-the-shelf calibration settings for their alert parameters and using ‘average peer alert volumes as a measure of the appropriateness of their calibration’.

We observe firms taking comfort from the perception that ‘others are also failing’ in the same way that they are. As a regulator, I must say there is something rather depressing about that logic.

We observe firms taking comfort from the perception that ‘others are also failing’ in the same way that they are.

This can become an incredibly self-referential cycle – ‘I match my standards to others and then say I am not failing any worse than any others because I match my standards to theirs’. As a regulator, I must say there is something rather depressing about that logic.

Secondly, in the specific case of fixed income, we observed institutions using inappropriate parameters when assessing Fixed Income alerts.

I fully appreciate that the management of false positives in alerting is a challenging and time-consuming process. However, checking that the design of the alerts is actually meaningful to the risks in the markets being surveilled is an important, dare I say it, elemental starting point to having effective controls.

Observing proper standards of market conduct

The need to consider how alerts need to be adjusted given the nature of the markets in question, reflects the fact that the evolution of controls will necessarily require nuanced discussion. We believe this nuance is equally important when thinking about how you support staff to fulfil their responsibilities under the SMCR.

Under the SMCR staff ‘must observe proper standards of market conduct’. It is a broad, but important statement, but it rationally leads to the question ‘how do you define proper standards of market conduct’?

We are well aware that there is a continuum of behaviour in the markets and the consideration of how to define appropriate standards of behaviour is important. There are several elements I would like to highlight.

The first relates to Question 2 of the 5 Conduct Questions: How do you encourage individuals who work in your institution to take responsibility for managing the conduct of their business?

This fundamentally goes to the ability of the people who work in wholesale markets to know what behaviours represent good conduct and to uphold those behaviours.

However, Question 3 of the 5 Conduct Questions, regarding the support a firm puts in place to enable staff to improve the conduct of their business, is equally important and leads to the question: What steps are firms taking to ensure that those operating in the markets understand what ‘proper standards of market conduct’ actually are?

Where that line is easy to draw, then it is relatively straightforward: If you know you have inside information, do not trade on the basis of that inside information or share that information with someone to encourage them to do so. Simple, one might say, clearly on the wrong side of the line.

We know however, that life is not always that simple. In most professions, it is the ability to identify the shades of grey that is most critical.

Sometimes large tracts of regulation may not be the most appropriate place to define where the right and wrong side of the lines are. Indeed, the industry or industry bodies may be better placed to do so. This is one of the reasons, in the context of our work to ‘give codes teeth’, we have established a mechanism to recognise particular industry codes where we feel they support greater clarity regarding expectations.

We have recently closed our consultation on the recognition of the FX Global Code and the Money Market Code – and we will be publishing our decision on recognition in due course.

If recognition is granted, one way for firms to show compliance with the SMCR’s Conduct Rule 5 – ‘observing proper standards of market conduct’ – will be to demonstrate that they comply with the principles set out in the Code. Merely being a signatory is not sufficient – we will expect firms to embed the principles in their business models – and demonstrate this when asked.

Where there is no alternative to the recognised codes available in the market, firms and Senior Managers who choose not to adhere to the recognised codes, should carefully consider how they would demonstrate compliance with Conduct Rule 5.

There will be other areas however, where it will be necessary for institutions to determine how they wish to support their staff in identifying appropriate behaviour. I want to highlight a couple of areas where we have seen the challenge arise in recent years. The first is regarding the question of what we call ‘following’.

We observe instances where individuals within regulated firms will ‘follow’ the behaviour of their clients either by trading on their own account, or by giving tips to other clients or friends.

It is important that your firms have thought about the risks and conflicts which can arise from this sort of behaviour and we would recommend doing so through at least 3 lenses.

The first thing that firms should consider is why would your employee do this?

It will not surprise you to know that following occurs most where the follower believes the client is more likely than not to make a profit so they can make a profit by following their trading.

Secondly, why does your employee believe that a client trade will be likely to result in a profit?

It may be because they have observed that the client’s trading tends to be highly successful – perhaps suspiciously so, including where STORs have been filed about that client in the past - or because there is sudden demand from a large number of clients for a particular stock – which cannot be explained by publicly available information. There is also a risk that the client has indicated or suggested that they have inside information when placing the trade.

Thirdly, firms should then consider, given the obvious objective of ’following’, whether it is within their risk appetite for their staff to engage in such behaviour, given the issues which may arise.

For instance, if your firm identifies that client trading – which your own staff have followed - is suspicious, because for example, a takeover announcement is made a few days later, then your firm will need to consider whether to file STORs in relation to both the clients and the employees.

If your firm considers it necessary to file a STOR about an employee, it will also need to consider whether further action is necessary to mitigate the risks arising from that employee.

Firms should also consider whether an employee who has followed a client becomes conflicted in relation to their own regulatory obligations.

It is normal for firms to have policies in place which require front-office staff to escalate concerns about a client’s suspicious trading to compliance. This is required so that the firm can consider, for instance, whether to file a STOR. Now let’s imagine that one of your employees has been following a client’s trading, which later becomes suspicious. Would the employee not be discouraged from reporting that trading to compliance? It would, after all, shine the light of suspicion on the employee’s own conduct. This is a significant risk.

Firms should also consider client confidentiality, the duties under COBs and our Principles for firms to consider the best interests of their clients.

The dissemination of information about trading behaviours to other clients or indeed the mismanagement of any conflicts of interest by trading on that information for one’s own gain, is something that we would expect firms to have thought carefully about. Furthermore, firms should be alive to the risk that staff who are supposed to be operating on an ‘execution-only’ basis might actually be advising clients what to purchase - by for example telling one client what another client has traded in and encouraging them to also deal.

You should carefully consider these issues and whether your PA dealing policies appropriately mitigate the risks to which your firms are exposed – do those policies fully consider and map through to the appropriate standards of market conduct you wish your staff to uphold? How do your staff think about following clients?

Our view is that individuals involved in the receipt and execution of client orders should be focusing on ensuring that they act in accordance with regulatory obligations, such as delivering best execution and reporting suspicious behaviour. They should not be seeking ways to enrich themselves through copying suspicious client orders or sharing them in any way that encourages others to trade on those orders.

The other element I would highlight is the potential that firms provide data to the market or their clients, that incorrectly reflects the market in a given instrument. We have highlighted this before and have recently publicly addressed two concerning practices.

The first is ‘flying’ – which is essentially where clients are told that an institution has bids or offers that are not supported by, or sometimes not even derived from, an actual order or a trader’s instruction.

The second is ‘printing’ – which involves communicating that a trade has been executed, at a specified price and/or size, when no such trade has taken place.

Again, we would expect firms to have considered the risk that their staff may be misrepresenting the market to their clients. If false prices and/or trading activity are advertised to the market, then there is a risk that recipients of that information may make trading decisions based on misleading information. This can cause obvious harm to market participants and undermine market integrity.

MAR and MiFID II define what constitutes unacceptable conduct and practice, including behaviour that amounts to giving false or misleading impressions to the market. There are some occasions when this behaviour will be clearly abusive and where it may represent a breach of the FCA’s Principles for Business.

These are nuanced issues but they reflect that the cleanliness of our markets is not just founded in the avoidance of clear acts of criminal or regulatory breaches which produce enforcement outcomes.

The standards of a market are made up of the behaviours of its participants, each and every day. Where and how we set those standards and support staff to uphold those standards, is equally critical to maintaining a clean market.

Whilst as the securities regulator, we may enforce against those actions that are clearly and demonstrably in the wrong, we also have an absolute interest in ensuring that market participants meaningfully and carefully consider how they operate within the shades of grey.

Looking to the future

By highlighting some of the areas that we are concerned about, we hope that this can support more effective risk assessments within the industry. However, markets are not just made up of thousands of individuals making decisions. Increasingly they are also made up of algorithms and other forms of technology making decisions, albeit ultimately governed by human beings.

The risk profile of the markets therefore does not stand still, it evolves as market structures, market sentiment, macro-economic and geo-political triggers change and as technology develops.

Our risk assessments must seek to keep pace with developments and recognise where opportunities to put controls in place now exist where they might not have done before.

As a consequence, I want to flag a few areas where we are increasingly starting to focus and think further about how risks in these areas may evolve.

Evolving market structures

I am always wary of mentioning the B-word however, I think that it is important for us to think about the implications of Brexit on the longer-term functioning of our markets.

MiFID II has been, amongst many things, an exercise in bringing together a common rule book and a standardised set of transaction reporting in service of greater surveillance of the European markets.

Whilst the FCA will work to ensure we continue to have the most robust and collaborative regime possible after Brexit - firms must consider how their institutions could be used for market abuse in this evolving context.

Whilst activity may move, it is extremely important that the quality of controls remains robust.

Firms must not have gaps in their oversight, and equally, must remain confident that they can still see the big picture of the behaviours they are facilitating – even across multiple borders.

We have seen examples in the past where firms leave themselves exposed to the risk of facilitating financial crime by failing to ‘knit back together’ activities taking place within their institutions. There is no doubt that where activities are shifting and moving, this risk has the potential to intensify.

We wish to continue to work closely with market participants to understand and mitigate this risk.

New forms of behaviour and new forms of risk

It was once said that there is nothing new under the sun and perhaps that is right. Mark Twain commented that there was no such thing as a new idea, just a mental kaleidoscope where old ideas could be turned to make new combinations indefinitely.

In many respects, that feels like a good analogy for the evolution of the markets and the evolution of behaviours within them – the ideas may not, fundamentally, be new, but their application might be. This is one of the things that, as regulators, we are focused on – are we sufficiently attuned to any shifts in that metaphorical kaleidoscope?

One of those shifts is the potential for the creation of manufactured credit events.

We have observed behaviour in the CDS market in other jurisdictions that appears to involve intentional, or ‘manufactured,’ events. As Andrew Bailey said in a Bloomberg interview last year, we feel this behaviour is on ‘the wrong side of the line’. It goes against the intended purpose of these instruments. Manufactured credit events may also constitute market abuse by the involved parties.

We have been speaking to firms, ISDA and other regulators and support the work being undertaken. We look forward to swift and effective action being taken by the industry. Where necessary, we will take appropriate action to uphold our statutory objectives.

I will leave you with another area that we feel requires our growing focus: How one thinks about both the risks and opportunities that new forms of data and technology afford the markets through the lens of market abuse.

We have traditionally focused on company results and announcements, news flows and analysts reports to determine how one might think about inside information and publicly available information. However, a world where we create data at an ever-increasing pace, must lead us to consider how we view these new data sources, especially when they have the potential to be market moving.

A Forbes magazine article in May 2018 stated that ‘2.5 quintillion bytes of data [are] created each day’ but that this pace will only grow. It also reflected that ‘over the last 2 years alone, 90 percent of [all] the data in the world was generated’.

Whilst facts like these are almost mind-blowing, they also raise some very interesting questions with which we as regulators, but also you as market participants and practitioners, must grapple with.

Using such data to provide new insights into the performance of companies and markets is a valuable addition to the price formation process. However, as a result, there will no doubt be interesting regulatory questions we will have to ask ourselves in the future. We will be exploring how data is used in wholesale markets as we explore that question.

Equally important for us now is how the technology that consumes this data evolves – and ensuring that it can do so in a safe manner. Implicit in this is that the quality of this data has to be continuously assessed to ensure accuracy.

Algorithmic trading is a thoroughly embedded part of how markets function now and is it continuing to evolve. How machine learning and AI are applied to trading activity is something that we must closely follow, not just to understand how markets function and ensure that they function well, but also to scan for potentially unintended consequences.

Currently, artificial intelligence and machine learning techniques, whilst proving highly successful in certain fields, tend to lack the ability to explain how they derive their results. There should be caution in blindly accepting answers just because the computer says so.

Recent analysis of our Markets Data Processor data indicates that approximately 1,800 algos are active in a given week on FTSE100. This is reflected in regulatory data which show 75% of proprietary trading decisions in equities are triggered by algorithms, contrasting with less than 25% for debt instruments.

I have long remembered two pieces of information about the evolution of technology that have stuck with me. The first from a renowned American AI specialist who starts many of her speeches by stating that she can prove empirically that you cannot build an unbiased AI. And by the argument made to us by one high frequency trading firm that whilst we already have enough data points in the equity markets to inform increasingly effective machine learning, in the fixed income markets we would need data going back to the time of Elizabeth I’s disagreement with Mary Queen of Scots provide the same depth of data.

technology is only as good as the data that feeds it and the requirements and constraints to which it is put.

These points highlight that this technology is only as good as the data that feeds it and the requirements and constraints to which it is put. I can see a world where seemingly ‘rational’ AI, unconstrained and exposed to certain markets and data, would deem it entirely rational to commit market manipulation. Now, the FCA cannot prosecute a computer, but we can seek to prosecute the people who provided the governance over that computer.

The application of new technology is of course future driven. We are currently standing on the brink of a new era – that of the Quantum Computer which brings game changing technology for processing extremely large amounts of data at extreme speeds to the financial world.

This is just one example of how the market and its behaviours do not stand still, they evolve and may possibly continue to evolve at a faster and faster rate in the future. Firms and as well as regulators need to keep pace with these developments. This will require us all to remain fully focused and operating with the correct mindset in order to identify the new challenges that face our market and its integrity.

Conclusion

Without the right state of mind, I continue to believe, we will not have effective compliance with the market abuse regime.

Equally, without the ability to constantly evolve our assessment of the risks, we will not be able to achieve a market that relies less on post hoc systems and controls and more on the pre-emptive recognition of those risks. Without this, we will not be able to support the good behaviour we need in our markets as effectively.

Put simply, whilst it is important that both regulators and participants have controls to detect suspicious behaviour, it is critical that we consider how we protect the market from such behaviour occurring on our watch in the first place.

First published: 13/02/2019 Last updated: 17/05/2019