Speaker: Peter Andrews, Chief Economist, FCA
Location: The Howard Centre for Financial Analysis, Imperial College Business School and Financial Conduct on 4 October 2016.

I am going to discuss consumer decision making in the context of the FCA’s incipient Mortgage Market Study. Let me explain my perspective on decision-making and why it is a problem or, at least, a challenge for conduct and competition regulators like the FCA. I will also say how I think economics can help address this problem.

Consumers in the UK mortgage market have not experienced systematic mis-selling of the kind that has blighted some markets for investment and insurance markets. At the same time, mortgages are the biggest financial transaction that many consumers make. As a result, even seemingly small differences in the price of mortgages chosen by consumers can have material impacts on household budgets. Thus mortgage regulators are concerned with how consumers make these decisions so that, if appropriate, they can be influenced.  

It turns out that this decision-making by consumers may be very complicated, and working out how to influence it effectively may also be complicated. Moreover, in these endeavours the FCA of course needs to proceed in accordance with its statutory objectives and duties and with the general requirements of Administrative Law. In other words, the problem is to handle a tricky issue in the way the law requires.

Ensuring markets work well

I will start with the overarching statutory provision. The Financial Services Act, 2012 gives the FCA one ‘strategic objective’, namely: ‘to ensure that the relevant markets function well’. What does this mean?

Ensuring well-working markets could mean…

First, from the perspective of regulatory risk, to try to ensure that things that adversely affect our statutory objectives - consumer protection, integrity, and competition, as set out at the start of the Financial Services Act, 2012 – do not happen, whatever their cause.  

Second, from the perspective of regulatory economics, to counteract traditional market failures (market power, information asymmetry, externality) or behavioural ‘imperfections’ since markets affected by these failures are generally not considered to be well-working and their presence adversely affects our statutory objectives.

Conceptually, these perspectives are complementary. The first is about identifying and preventing possible bad outcomes in markets. The second is about addressing aspects of the operation of markets that are inconsistent with effective competition and that a ‘state actor’ can improve, in principle leading to better outcomes.  The second might be a subset of the first, although its focus on causes of harm and whether these can be addressed successfully is distinctly different. 

I am not convinced that it is worth agonising about nuances of meaning here, though the meanings are certainly debated by some stakeholders in regulation. If we see harm, whatever its nature, and can counteract or otherwise remedy it, why would we not do so?

Questions for mortgage regulation

A more interesting question, it seems to me, is how do we use these concepts in mortgage regulation? To illustrate, let me give a commonplace example.

‘Hyperbolic discounting’, effectively choosing a small reward today over a larger one tomorrow, can lead to choices not reflecting an individual’s true preferences. Perhaps when initial mortgage savings are much more than offset by later high charges.

This can justify consumer protection measures under the ‘risk’ concept which clearly allows ‘benevolent paternalism’ and it is for sure a behavioural imperfection, the presence of which can be construed as a risk that people will damage their own interests.

On the other hand, some argue that if the information was clear, and the consumer was not misled, then the market did not fail: it catered for a plausible preference, perhaps implying that there is no case for intervention.  

Equally, one can say that the demand side of the market failed to conduct itself in the way required by a well-working market. Or that the market failed to supply a de-biasing service for which there was true demand.  This looks like a market failure that provides a case for intervention under the standard notion of regulatory economics above. On the other hand, it could be that de-biasing was practically impossible.

But I want to avoid being side-tracked into a perhaps circular debate about whether demand-side weaknesses are a true market failure. We obviously do not want them to be a general, unfocussed excuse for intervention which could be costly and wasteful. Yet, if the regulator sees a genuine harm arising from demand-side weaknesses, and has powers that realistically can make things better, then, again, it should probably take action.

There are hard questions here, though, namely:

• (a) whether there really is harm: how can the regulator identify the consumer discount rates that determine whether mortgage deals that are initially cheap and are expensive later on, do or do not align with mortgagors’ true preferences and
• (b) whether any intervention will be successful if it relies on strengthening a demand side that really is weak. For example, the disclosures that formed the core of the 2004 regime of mortgage regulation by the FSA were extensively tested on consumers but are not widely believed to have changed consumers’ choices materially for the better.

In our brief discussion of a commonplace mortgage product, we have already seen that there is room for debate about the case for intervention depending on the precise facts of the case and on how one conceptualises regulation and its role. There are also difficult practical questions about whether there is harm and, if there is harm, about whether and how it can be remedied.

We have not even begun to discuss my main area of concern, the complexities of consumers’ decision-making, to which I will turn later.

Anyway, it is clear that there is a regulatory problem to address and before exploring it further, let’s see what the law has to say about how we should proceed to deal with it.

Administrative law constrains regulators in ways that are likely to influence the process, and content, of policy-making and thus influence how the regulator addresses the problem of changing consumers’ bad decisions in the mortgage market. 

First there is the concept of ‘Wednesbury reasonableness’: this is the fairly low hurdle of not making decisions that are so unreasonable that no reasonable regulator would make them. The judges in this case used dismissal of a teacher for having red hair as an example of such unreasonableness.

More challenging is the notion of ‘abuse of discretion: taking into account irrelevant considerations or failing to take into account relevant considerations’. Lawyers Harwood and Wald note that: “A dismissal of a red-haired teacher [for being red-haired] could also be challenged on the basis that it took into account an irrelevant consideration.”

I observe in passing that our biggest constraint is probably our wish to do a great job. That is, to implement only remedies that will really change markets for the better. This is something that can be very hard to tell in advance and opens up an important debate about evidence, which I shall not engage today.

What are the ‘relevant considerations’ that we regulators must take into account in order to avoid abuse of discretion? In the present case, we might say the following.

• to decide whether to intervene, the FCA needs to know how well the mortgage markets are working. This is the point of the research on consumers’ choices which is to be presented after this talk.
• To decide how to intervene, should this be necessary, the FCA needs to understand how consumers actually make these choices and then use this understanding to work out how to change what consumers decide. It is important to note that decisions might be changed by seeking directly to influence consumers’ decision-making processes or by constraining in some the choices that are available. Product regulation is an example of such a constraint.

Consumers’ decision-making

What then does science say about decision-making? In terms of the legal discussion about how public bodies should proceed, what is ‘relevant’ to understanding decision-making?

At the Bank of England’s 2016 Chief Economists’ Forum, Andy Haldane presented some disturbing statistics on the extent to which economists quote other sciences, and the extent to which other sciences quote each other and economics. It turns out that economists use other sciences markedly less than other sciences do. 

In the case of regulatory analysis of decision-making, a narrow, economic approach to understanding what is going on seems wrong as it is obvious that many sciences have something to say about the topic.

I should quickly add that I do not contend that regulation is synonymous with economics. Regulation can and does look far beyond economics for its inspiration and methods. At the same time, though, it is true that most regulatory thinking about decision-making draws almost exclusively on neoclassical economics and behavioural economics – meaning, if you like, economics which takes some account of psychological insights – backed by consumer surveys. 

Now, I would say that extensive use of economics in regulation makes perfect sense because it is the science that, linking back to the perspectives with which we started, addresses whether or not markets are working well. The only issue is whether economics should and can make more use of other sciences in its analysis of decision-making.

Anyway, here are some ideas relevant to understanding decision-making and the names of the authors of the papers in which they may be explored. Some of the papers were written by economists but many of them were written by experts in anthropology, neuroscience, marketing, psychology, law, or sociology. It is worth noting explicitly that psychology offers far broader insights into decision-making than the ones so far used in behavioural economics.

1. Consumers use information but not always very well (Peters, Hibbard, Slovic & Dieckmann)
2. For example, consumers may be biased in their use of information (De Meza, Irlenbusch & Reyniers, FSA; FCA OPs)
3. Consumers struggle with complex calculations and terms (FCA paper in prep; Dellaert & Stremersch)
4. Consumers struggle with uncertainty (Preusschoff, Mohr & Hsu)
5. Information, complexity and uncertainty can interact in nasty ways and firms can increase complexity to exacerbate this (Ben-Shahar and Schneider)
6. Decisions may be solo, in couples, or influenced by family or other ‘social’ advice (Capuano & Ramsay)
7. Or be influenced by experts – typical, given rules, in UK mortgage market (Meshi, Biele, Korn & Heekeren)
8. Or be based on emotional factors (Lerner, Li, Valdesolo & Kassam)
9. Or be based on ‘Darwinian’ factors (Gigerenzer)
10. Or be based on ‘stories people tell themselves’ (Akerlof & Shiller)
11. Or they may be best understood through an anthropological lens (Boholm, Henning and Krzyworzeka)

In other words… it’s complicated. 

For example, consider uncertainty – which is just a part of item five above. Smith and Stern list four types of uncertainty:

1. Imprecision – the outcome is not known precisely but decision-relevant probability statements are possible
2. Ambiguity – also widely known as Knightian uncertainty; we cannot make probability statements
3. Intractability – we cannot formulate or, if we can formulate, we cannot execute the relevant computations
4. Indeterminacy – with respect to quantities relevant to policy-making for which no precise value exists.

Tuckett and Smith in a new, cross-disciplinary project on decision-making under uncertainty, that is being undertaken with ESRC funding, argue that neoclassical and behavioural economics – on both of which we at the FCA have worked hard - help very much with point one and not very much with points two to four (for example, because bias is implicitly relative to a knowable ‘right’ outcome).  This may be important because several categories of uncertainty are relevant to many financial decisions.

Yet it is a challenge for regulators to engage with all the sciences – anthropology, neuroscience, marketing, psychology, sociology, and so on – cited above, even though they appear to be relevant.

Let’s discuss some implications of all this.

Consumer decision-making is clearly very complex. It is not simply rational use of information, nor rational use of information allowing for one or more of a list of psychological biases.

Moreover, our experience so far mostly tells us that regulatory information and nudges tend to have small or unexpected effects when trialled. This is a huge pay-off from trialling. The underlying reason for the results could easily be that decisions are affected by multiple drivers.

Regulators also need to be cautious in setting remedies because of course in real-world markets firms can respond to any changes observed in response to our demand-side interventions. And firms largely control the context and nature of their interaction with customers.

Another issue is responsiveness of consumers to nudges and carefully crafted information. It is striking that when Ryanair’s business model involved aggressive marketing of travel insurance the relevant web page contained up to about nine linguistic behavioural tricks. 

Presumably, Ryanair’s own continuous experiments on its website found that each trick had an incremental effect on sales. Perhaps then public policy makers should not be surprised when single nudges (as opposed to ‘fiat’ defaults and the like) fail to transform markets. Perhaps public policy makers should develop their own websites for live experiments too. 

I suspect also that even where we cannot see specific linguistic devices designed to influence purchases, it may well be that sophisticated firms have deployed unobservable design features arising from retina scanning and other science. These considerations influence the positioning and colour of text, choice boxes and the like.

So, clearly, we must somehow simplify our problem. The question is how.

First, I would say this means that we should not give up on economics. We can use economics to learn, and it can help simplify our problem. For example, the issue of influencing complex decision-making may call, ideally, for natural experiments in live markets.

We can also use economics to work out an efficient equilibrium for the relevant market and then – here is a key simplification - design remedies that mimic the outcome of competition, ideally leaving some space for the process of competition to continue.  

Such remedies may be especially useful where the demand side is too weak or too unpredictable or just too hard to model. Or where, even if we can get a reasonable understanding of the demand side, working out how to influence it does not yield reliable results or shows that it’s likely that it cannot be done effectively. In fact, cases of an activist demand-side prone to informed switching seem rather rare in financial services, whereas the benefits to consumers of the price cap in high cost short term credit markets are very substantial.

To be sure, though, the FCA’s mortgage market study is only just starting so its implications for remedies are completely unknown.

Conclusion

I have briefly described some constraints on the FCA as policy-maker and some ways in which consumer decision-making in the mortgage market is a challenge for the FCA in its statutory role. This decision-making is hard to understand and hard to influence sufficiently, through regulations, to make markets work well. All academic help from relevant sciences will be gratefully received.

I have argued that economics can help us understand consumers’ decisions, especially where we can carry out realistic field trials.

I have also argued that economics can simplify our problem by showing what a competitive equilibrium would be and consequently what interventions would mimic the outcome of effective competition.

Challenging issues to consider are, first, whether economics can utilise more psychology and other science of decision-making in its analysis of market equilibria. Second, whether consumers will be best-served in mortgage markets by attempts to empower them or by more intrusive regulations that mimic the outcomes of ‘competition in the interests of consumers’.