Supervisory Judgement – A Fresh Perspective?

26^th June 2025

An article by Sam Woods, Chief Executive of the Prudential Regulation Authority and Deputy Governor of the Bank of England, in conversation with Professor Sir Andrew Likierman of the London Business School

Introduction

The job of the Prudential Regulation Authority is to keep banks and insurers safe and sound, while also supporting competition, competitiveness and growth. We’ve built judgement right into the core of our approach to this task, and in particular into our supervisory activities – for which our strapline from the first day of the PRA has been “judgement-based, forward-looking supervision focused on the key risks”. We’ve got judgement in that strapline explicitly, but also strongly implicitly in the choice of which direction to look in and which risks to focus on. We mean by this that our job cannot be done well with a compliance, tick-box approach because that would lead us to miss the woods for the trees. 

But have we in the PRA really thought enough about what we mean by “judgement”, and how we avoid pitfalls in taking an approach which relies on it so heavily? Well I thought that we had, but then I read Professor Sir Andrew Likierman’s excellent book “Judgement at Work” which made me think again. So I’ve put some questions to Andrew to help illuminate this vital area.

Interview with Sir Andrew Likierman

(SW – Sam Woods)

(AL – Sir Andrew Likierman)

SW: Thanks for joining us Andrew. Perhaps a good place to start would be to ask you: what is judgement?

AL: There is no standard dictionary definition, and I have combined the entries of many dictionaries to define it as “The combination of personal qualities with relevant knowledge and experience to form opinions and take decisions”. Most people only think about decision-making in relation to judgement, but it is also very much about forming opinions, as the PRA has to do all the time.   

SW: That will serve our purposes for today. Now in your fascinating book you break this down into six elements. What I’d like to do is explore these in the context of prudential regulation. Your first element is about relevant knowledge and experience. How should we think about this? Some our most able supervisors are ones who have joined us from fields as wide-ranging as scientific research and the armed forces. Of course, we have other staff who’ve worked in banks and insurers, and others who have been in the regulator for many years. How should we think about this?

AL: Understanding whether knowledge and experience is relevant is the key here. Specialists are people with a lot of knowledge about a particular topic.  But it takes judgement to apply that knowledge to particular situations. So it’s not surprising that those with good judgement who come to the PRA from different backgrounds can and do make the necessary connections. Good judgement also means, for example, knowing when to ask for advice and who to trust if they don’t have the necessary knowledge, say on a particular technical point or AI. 

SW: So this is partly about knowing what the limits of your own knowledge and experience are, and where and when you should look for help. So say a new supervisor thinks a bank’s liquidity ratio looks a bit low – they should pursue this hunch but seek input from a liquidity expert before reaching a firm view.

AL: I’m sure that this is what PRA supervisors would do once they realised that they didn’t have the knowledge to back up their hunch. The difficulty is sometimes not being aware that one doesn’t have enough knowledge. Which is where the next element of the judgement framework comes in.

SW: That takes us neatly into your second element, which is on how aware you are of what is going on. That does seem a fair point – I hope I have some vague idea of what’s happening out there! But tell us what you mean by this Andrew – in our world would this be about being across financial markets, credit conditions, cyber threats and things like that? How do we know if we are aware enough?

AL: There are two elements to awareness. One is how aware we are of what is going on. This might be in an interview with a firm where there is an awareness of what is not said as well as what is, that information is missing and is being withheld, or when the person should know the answer but doesn’t. 

Another part of awareness is self-awareness. This also applies to other parts of the judgement framework when we are aware of our feelings and beliefs, including our biases and attitudes to risk. More on that when we get to it.   

SW: I would say that illustrates the importance of supervisors having enough entry points into a firm. The CEO might be very polished and convincing, but does this tally with what we hear further down in the organisation? Of course whistleblowing and other channels are also important in this respect – and experience definitely helps supervisors in calibrating what they hear.

AL: In terms of improving our knowledge and experience, as in all other aspects of the judgement process, one of the ways of getting the most of learning as we go is to review what went wrong and what went right.   

SW: Now we come to a very interesting element for prudential regulators, number three: who and what you trust. I’m tempted to say that in our line of work we should trust nothing and nobody! But perhaps that is too reductive. Of course we place some reliance on capital ratios, liquidity ratios, and how convincingly people in banks and insurers can explain to us what they are doing and why it’s OK. Talk us through this one Andrew.

AL: As with so much when dealing with judgement, it’s getting the balance right. I’ve worked with the audit profession on identifying professional judgement and as you know, professional scepticism is an essential part of their role. We also know that auditors get into trouble if they don’t use enough of it. Asking about who and what we trust and why is a good start to getting the balance right. 

SW: So this is partly about having a sceptical, questioning mindset. I think that a degree of sceptical curiosity is a vital quality in a prudential supervisor, just as it is in an auditor or a journalist. We shouldn’t just take what we hear from firms at face value, nor should we take the numbers at face value – we should ask what is behind them.

AL: There are many ways of establishing trust. For example, using a legal context one can look for internal consistency, consistency with what was said before, consistency with what other people say, the reputation of the person or firm and their demeanour (“conduct, manner, bearing, behaviour, delivery, inflection”). 

SW: Element number four is also interesting, and I suspect that we significantly underestimate this in our line of work. Number four is feelings, beliefs and values you have. I say we underestimate it because we consider ourselves to be very analytical and evidence-based in our work. But of course, we have priors, and – contrary to popular belief – we also have emotions! Maybe we get too attached to some beliefs, or too inclined to believe some things and not others. What do you make of this for a line of work like ours Andrew?

AL: As someone pointed out to me, one of the most difficult biases to overcome is the no-bias bias - “Me?  I’m completely objective”. My working assumption is that we all have feelings and beliefs, including our emotions as you say. When I’m working with executives, I suggest that it’s not realistic to assume that they can get rid of these. In some organisations people are encouraged to declare their biases when making difficult judgements. What matters is that they are recognised and taken into account and that includes our values and those of the organisation. 

SW: I can see how that is very relevant to supervision Andrew. In particular, if a supervisor has experienced a serious problem with a firm in the past they will naturally bring the feeling of that experience with them to today, and it will probably make them more cautious. Or it can be the other way round – positive experience in the past makes the supervisor less sceptical today. Sometimes this can be good but at other times that feeling might simply be out of date, and distract the supervisor away from different issues which have become more important to the firm’s safety and soundness today. Just like the market, the supervisor must be willing constantly to update their view and be aware of biases or feelings they have that might impede that process.

AL: I can imagine how difficult it is to identify shifts of this kind, as when the firm that has been rock solid turns out to have issues which are very difficult to identify. Information may be deliberately withheld, or – even worse –the management itself may not be aware of the issues. The quality of detachment, which includes awareness of feelings about a firm, is clearly crucial here.   

SW: Now element number five is more firmly on our screen: the way you make a decision or choice. We have different methods for different decisions: we have wide consultation and decisions taken by our top committee for big policy choices, whereas decisions are taken at a junior level in the line for minor supervisory issues. But there is more richness in how Andrew has thought about this – is it about more than delegations Andrew?

AL: My assumption on choice is that we need to make sure that, whatever the level, we are in a position to have a high-quality choice process. This means that we understand the way groups function well and not so well so that we get the best out of them. This includes the quality of the information provided, well-framed choices and ensuring that people feel free to speak up. We equally need to avoid the dangers of dysfunctional groups such as factions or groupthink. 

SW: I think this is particularly relevant for how we approach decision-making for decisions which are both very important and also extremely complicated and technical. Decisions around whether or not to approve firms’ own models for setting capital requirements is a good example. On the one hand this is very important, which might lead us to escalate it to our most senior committees; on the other hand a more junior committee of technical experts might understand the detail more thoroughly and make a better decision. In practice we have to strike a balance.

AL: Yes, and whatever the level at which it’s taken, the way the discussion takes place needs to get the best out of those round the table, for example by making sure that group operates well. Good judgement relies on a well-functioning group and the committee Chair has a big responsibility here.   

SW: Last but not least we come to element number six, which is whether you can actually deliver what you’ve set out to do. I must admit that I have set out to do some things, like reforming parts of Solvency 2, which it turned out that I was spectacularly unable to do! So this is an area for development, as they would say in a performance appraisal. I know that you debated whether or not this is part of judgement Andrew but decided it had to be included – how do you think it applies to work like ours?

AL: I included delivery as part of the judgement framework because it seemed to me that a good judgement meant that it could be delivered. It’s the firms who are responsible for delivery and under this heading I would point out that the judgement framework applies just as much to the firms regulated by the PRA. Thus it might also be used by PRA staff to help to assess whether those running the firms have judgement. This doesn’t just apply to an assessment of their ability to deliver, but to an assessment of whether they too have what we have been discussing – relevant knowledge and experience, whether they are aware of what is going on, whether they trust the right people and so on. A firm being run someone who you assess as having poor judgement must surely be subject to higher risk, and not just in their ability to deliver. 

SW: Many thanks Andrew. Do you have any final bit of advice to us on judgement?

AL: A key takeaway from my work is recognising judgement as a process, particularly when faced with important choices in conditions of particular uncertainty and complexity. Doing so can help in the quality of the choice, whether it’s on capital, on risk or on individuals. There are many ways to improve judgement for each of the elements we’ve discussed. This might be identifying gaps in our knowledge and experience and finding ways to fill them. It might be about making sure that we are aware of biases. Recognising the fact that it is a process also provides us with the basis of explaining to others what we have done. 

To help in the process, here is a link to my list of possible actions to improve judgement. It’s an a la carte menu, so not aligned to PRA-specific needs, but I hope may be helpful in provoking thinking about practical steps that might be useful. 

Closing thoughts

We’ve built judgement into the core of the PRA’s approach to delivering its statutory objectives. All prudential regulators do this to varying degrees, and there is an interesting debate emerging in the USA about the approach to supervision including how far it should go beyond simply checking key financial metrics. My experience has been that there is a very important role for judgement in our work, and that a tick-box approach is ineffective. However, my conversation with Andrew does make me think that it is worth regulators testing more rigorously their own capacity to make good judgements. We do this all the time in a live setting in that the consequences of good or bad calls on supervisory issues often manifest quite rapidly – but we can also stand back and use a framework like the one Andrew has developed to test in a more methodical way: do we have good judgement?