Making competition king – the rise of behavioural economics at the FCA

Promoting competition 

Thank you ASIC for inviting me to offer some insight into our work on behavioural economics over the last year. It is a pleasure to join you in Sydney.

On the way over, I was reflecting that if you asked someone on the streets of London for a list of important cultural exchanges between Australia and the UK, it’s unlikely regulatory philosophy would feature.

Shared history, values, brands and services – yes. Political frameworks – yes. The arts and media – undoubtedly. Bar staff… in all these areas the ties between the two countries are clearly strong.

Yet arguably, the most important area of crossover in recent history (certainly in terms of moving the UK economy forward) is also one of the least publicly recognised: our adoption of the Australian regulatory model. The prudential/conduct split.

So in the UK, as in Australia post-Wallis Inquiry, we have a new voice in regulation, with more emphasis on moving things forward in key areas like ethics, crisis prevention and consumer protection.

And it’s in this context, with this new voice if you like, that we’re using behavioural economics to support increased competition in financial services. Effectively, the challenge here is how do we use this science to encourage firms to compete more determinedly on price and product quality?

For leaders in regulation, as well as politics, this is a debate that’s been snapping at our heels for some time.

If I think back to the 1970s and 1980s, when the UK was seriously absorbed by analysis of whether governments should set supermarket prices, the standard response to monopolistic behaviour in financial services was almost invariably de-regulation.

Today there’s clearly a more nuanced, balanced approach to this issue and progress has been made. But it’s still true to say that many of the concerns relating to competition in financial services remain ‘sticky’.

So, even with the significant, post-crisis structural changes we’ve introduced into banking over the past five years, the British are still more likely to seek a divorce than a new current account. And a similar story holds true in areas like insurance, where 60% of people buy annuities from the company they’re saving with, as opposed to shopping around.  

Now, some argue this lack of competition is all about too little disclosure. Others insist it is all about numbers and quotas. The UK has relatively few banks compared to, say, Italy, which has close to 700.

But the evidence here is not clear – or at least it’s only clearly ambivalent. So, in the financial sector more than perhaps any other industry, we know the addition of more data and more market players is not always reflected in the mark-up or quality of retail services.

In other words, prices do not always respond efficiently to new information. Yet too often in the past we assumed the opposite. That prices do respond efficiently (or if not efficiently, that they do at least respond); and that decision-making is rational, not always, but certainly rational for sophisticated decision-makers with access to data.

An important question for the new FCA has been to understand ‘why’ this doesn’t happen. Effectively, why do so many consumers make the wrong decisions, or at least ‘sub-optimal’ decisions, when selecting financial products? Why do they sometimes reward poor products with their custom?

Payment protection insurance

In the UK, the classic example of this asymmetry was payment protection insurance (PPI). An insurance product sold alongside loans, mortgages and credit cards to millions of consumers from the 1990s onwards. In structure, broadly equivalent to Australia’s consumer credit insurance.

Now, clearly taking out insurance against future loss of income is not, in and of itself, indicative of any competitive failure or breakdown in economic models.

But PPI as a product should never have been the gravy train it was: generating £45 billion premiums or a 490% return on equity. In a perfectly rational market, it would have been both too expensive to succeed, with premiums adding 20% to the cost of a loan, and too ineffective to succeed. The claims ratio only averaged around 16%.

The lesson here, as well as in relation to other products like mini-bonds in Hong Kong and subprime mortgages in the US, is that consumers (even so-called sophisticated consumers) sometimes struggle to make rational decisions when faced with the ‘black box’ of finance. The ‘hidden gears’ described by Michael Lewis.

So, unlike other industries, where giving your customers a metaphorical thump in the stomach before running off with their money would reduce turnover, businesses in financial services can still be rewarded. And behavioural economics offers two key explanations for this.

The first is the idea of ‘bounded willpower’. The suggestion that we have trouble following through on rational plans.

Now, clearly this might just mean we don’t go to the gym for a few days. But it could equally mean we spend money we don’t have on credit instead of saving money for the long term. Instant gratification. 

The second explanation of this asymmetry relates to the idea of ‘bounded rationality’. Put simply, when presented with complex, difficult maths or many different moving parts, we often make decisions that aren’t in our best interests.

An example from my previous role in Hong Kong. There, individuals who had term deposit accounts maturing were invited to meet the bank manager – the banks would then offer them a new deposit rate paying 1% or an alternative ‘safe’ product paying 7%.

‘Why does one pay 7%?’ was the question that consumers didn’t ask. They didn’t because it was offered to them by their bank and they trusted that the bank wouldn’t sell them a ‘risky’ product.

Had they known the 7% product was a complex structured product that was effectively writing credit insurance, they might have thought twice. But the product was so opaque consumers didn’t know the right questions to ask.

Right problem, wrong solution

A key recurring issue here is the fact that this buy-side challenge is made more difficult by significant sell-side and regulatory issues.

So, on the former we know firms have long understood they can profit from human mistakes. Sometimes inertia; sometimes our inability to forecast or understand issues like compound interest; sometimes lack of confidence; sometimes over-confidence; sometimes a combination of some, all or more.

And this is why we see so many businesses – including non-finance businesses – making a virtue of complexity: using well-worn ploys like drip pricing; under-emphasised fees; teaser rates; complex policy exclusions; reference pricing; pressure selling and so on to shift products. In the worst cases to people who do not need them.

On the latter point and the impact of regulation, the problem in Britain has been the tendency of the official sector to see a problem coming – perhaps worries over exclusions in insurance contracts – but to deploy ineffective counter measures. Right problem, wrong solution, if you like.

So, in the UK there’s historically been a lot of focus on rules over ethics – sometimes too much – as well as a heavy reliance on instruments like unsophisticated disclosure to support consumers. Effectively, when faced with a breakdown of price efficiency or rationality, the standard response was to provide more information and provide it more quickly.

If someone did not appreciate the risk of a product, we extended the description; if it looked too risky, we pushed people to the risk profile that allowed a sale. People were required to tick the boxes – that they had high-risk appetites; that they had read and understood the terms and conditions; and that the decision was their own, that it was a non-advised sale.

This is why, in the UK at least, we’ve seen a proliferation in the number of terms and conditions – typically associated with products like bank accounts and insurance contracts – that are longer than Hamlet.

So, effectively, what we’ve had over the years is a perfect storm of conditions eroding competitive pressures.

Difficulties, if you like, on the buy side – bounded willpower and rationality. Moral failures on the sell side – knowingly exploiting consumer biases. Systemic failures within the official sector, focusing too exclusively on areas like disclosure.

Behavioural economics: one year on

A key issue in all this for the FCA, is that behavioural economics is quickly becoming a game changer.

In order to move things forward in this space, it’s clearly an imperative to seek new solutions. And this is where the FCA work on behavioural economics enters the picture.

Since launching our occasional paper[1] last year – effectively opening up the debate around the use of interventions like nudges in financial services – a significant amount of work has been set in train.

Generally speaking, this activity is embedded into traditional regulatory analysis: so to some extent it flies under the radar.

But there’s now little doubt behavioural economics could have a profound impact on many of the most serious challenges facing policy makers today, including that key question: how do you help households discipline markets more effectively?

From the FCA perspective, we’re already seeing significant possibilities across a range of UK markets like cash savings, general insurance and retirement income, with more to follow shortly.

But I think there’s also opportunity here for behavioural economics to support more specific issues like complexity; consumer inertia; marketing and the impact of firm communications to consumers.

The FCA itself is active in all these areas: using behavioural analysis to help collect better management information (MI).

A good example: we last year published research[2] of our work with a firm that was writing to almost 200,000 customers, offering redress for a mis-sold product.

The firm’s original letter, written in good faith but designed without the use of any behavioural economic ‘nudges’, had a remarkably low response rate. Around 1.5%. And we suspected an issue here was how the information was presented in the letter.

So, working with the firm we did a randomised controlled trial, testing different versions of the redress letter with interesting results:

• so, we found reducing the amount of text more than doubled the response rate – increasing it by 128%
• explaining to readers that the claims’ process would only take five minutes – again almost doubled the response rates – by 91%
• sending out reminders increased the response by 93%…
• and finally, putting the main thrust of the message in bullets at the top of the letter was the out and out winner, almost tripling response rates. The increase here was 271%.

In a nutshell, we found very subtle alterations were able to effect very considerable improvements.

So, overall, we were able to move the level of response from a frankly unimpressive 1.5% to a more respectable 11.9%. Equivalent to an additional 20,000 people responding. Or over half a million pounds in extra compensation.

Now, clearly these are very positive numbers. But in terms of how we move things forward from a regulatory perspective, some of the most useful data actually relates to what didn’t work. So what drove down response levels during testing.

One area of interest for us was whether any official endorsement from the regulator would encourage more consumers to engage with the material. Perhaps increasing trust.

So, we tested this in the trial by adding our own logo alongside that of the firm logo. Sadly, this had precisely zero impact on response rates.

Even more dispiriting from my own perspective, we found that using the signature of the firm’s chief executive in the letter (rather than a generic customer service team sign-off) actually reduced response rates, particularly among female customers.

What I think is interesting here, however, and useful for the FCA going forward, is that we’ve drawn on those insights in a dozen or so further communications from firms. In each case our supervisors have used the evidence from that first trial to help business leaders create better communications.

In some cases a simple change in presentation. In some cases using the evidence to convince firms that sending a reminder letter is not an option – but a practical necessity.

Now, a key issue in all this for the FCA, is that behavioural economics is quickly becoming a game changer. Not just for firms, not just for consumers, but potentially for the shape of regulation for many years to come.

In other words, we’re not simply interested in its potential to change corporate behaviour or help consumers. It also offers opportunities for self-reflection.

So, one of the areas we’re now investigating is whether behavioural economics can offer us insights into how individuals within organisations behave and respond to regulation.

A priority question here: what does this teach us about the way we intervene within markets? Are the day-to-day interventions we rely on as effective as we imagine?

Behavioural economist Daniel Kahneman famously talks about the ‘illusion of understanding’ here - and its impact on areas like stock picking, making the point that for a: ‘…large majority of investors, taking a shower and doing nothing would be a better policy than implementing the ideas that came into their minds.’

Does the illusion of understanding affect regulators in a similar way? Or are we more effective than Kahneman’s investors? Either way, it’s clearly important to at least ask the question so we move things forward as effectively as possible.

GI-add on study

Finally, a word on our most recent and high profile use of behavioural economics: an FCA competition market study[3] – published this month – into the UK’s £1bn add-on general insurance industry.

Today we released the background analysis[4] to this work – giving a sense of the scale and significance of the scrutiny here.

The study itself focused on whether there are common problems in how well competition was working for consumers across different markets for general insurance add-ons.

Put simply: are there questions over the actual mechanics of the add-on sale model – so selling insurance on the back of another purchase, like travel insurance cover offered alongside a holiday? GAP insurance offered alongside a new car and so on and so forth.

Some key questions here for us were: how do consumers behave in these markets? And how does the add-on sales format influence their buying decisions?

Today’s occasional paper goes some way to answering these questions and gives more granular detail into the design of the behavioural research.

It also offers an indication, I think, of the potential importance of experimental evidence to global regulators. 

So, working with academics from University College, London, we designed a simulated online shopping experience in which participants were asked to hunt around for a main product – so a holiday, car hire, boiler, laptop, tablet or the like – and they were then given the opportunity to buy related insurance products.

At this point, the insurance was offered to the study participants in very different ways – perhaps earlier or later in the purchase of the main product, perhaps priced monthly or annually, and so on and so forth.

To test the full impact of the add-on mechanism, we had one group shopping for insurance alone – the other buying both main product and optional add-on insurance. 

In both cases, the groups were given very clear prices. But comparing the two, side by side, showed that keeping track of multiple prices without a total is a major barrier for consumers comparing offers from firms effectively.

For standalone insurance with a single price there were almost no mistakes in choosing the best deal. It is arithmetic we can handle. When it came to comparing offers with prices for a main product, as well as a separate price for an add-on without a total, we start to see problems emerge.

One in five participants here could not identify the cheapest combination.

At this point, we tested the impact of transparency. So, the add-on was only unmasked, if you like, when the individual had started buying the main product.

Unsurprisingly, these ‘point of sale’ offers had a more powerful impact on consumer behaviour than selling add-ons transparently. 

So, compared to a standalone insurance purchase, consumers were four times less likely to shop around for insurance. In fact 65% did not look at any alternatives at all before buying and were six times more likely to make mistakes. 

Also clear, within all this data and detail, is the significance to sellers and buyers alike of framing.

So, when monthly prices for an annual policy were presented to consumers, instead of the total cost for the year, there was more confusion about the cash outlay, with less shopping around; higher prices paid; and larger losses made.  

All of which tallies with our broader understanding and analysis of the industry numbers we’ve been crunching over the last year.

Key figures here: around 25% of consumers who bought add-on insurance were not aware they could buy it separately elsewhere. 58% did not make comparisons with other policies on the market – compared to 22% of standalone buyers. And around one in five could not remember buying the product three months after purchase.

On top of this, we see claims ratios falling below acceptable levels: Guaranteed Asset Protection insurance at 10%; personal accident at 9%.

So, overall not an encouraging picture of competition, with evidence of persistently low value, as well as consumers being collectively overcharged, to the tune of some £200m a year, for products they may not need or use. Reflecting, once again, the asymmetric relationship between reward and profit, and asking serious questions of policy makers.

Our response here has been to change our response. There is less reliance on classical economics – or dependency on rationality. More focus on behavioural economics and areas like complexity, bounded rationality, willpower, biases and so on and so forth.

Next steps

So, among other areas we’re looking at: a ban on pre-ticked boxes to challenge consumer inertia; publication of claims ratios to reduce information asymmetries; consumers confirming purchase of GAP insurance in writing after the sale; and improvements to the way add-ons are offered through price-comparison sites.  

At this stage these are proposals, not policy. So we’ll only be taking final decisions on direction after formal consultation with the industry. But for the FCA, this is a quantum leap forward in the competition debate.

For the first time, we’re seriously considering how we can support consumers to discipline markets more effectively, rather than attempting to reverse engineer market outcomes. The ‘supermarket price setting’ option, if you like.

Conclusion

One word of caution, however. Human behaviour and nudges, in particular, are a notoriously complex equation. Inputs and outputs do not always correlate. Nor are they predictable.

In the UK, for example, the tax authorities have had success driving up revenue by using nudge techniques in correspondence to taxpayers.

Compliance on tax payments increased from 68% to 83%, simply by telling non-payers that the majority of people in their local area had already paid. Yet similar trials elsewhere have led to a significant decrease in response rates.

So, this is not a silver bullet solution. Nor is it a like-for-like replacement for traditional regulation. Preserving choice is a political imperative – it’s one of the reasons behavioural economics is so popular – but it’s not always a practical imperative.

If consumers are consistently harmed by a product, firm or market, it makes sense that a ban or enforcement may be the best option.

Having said this, it is clear that behavioural economics – as part of the regulatory toolkit – could become profoundly important.

Charles Dickens famously described economics as: ‘a mere skeleton unless it has a little human covering’. Nowhere is this more true than in our asymmetric world of reward for bad outcomes, punishment for good.

And firms in the UK at least, should be under no illusions as to how serious we are about breaking that link between poor products and high financial reward.

Competition will be king as we move forward. And that means tomorrow is unlikely to look like today.