Philip Robinson – Director, Deposit Takers Division, FSA
Keynote Address
CML/GEMI CREDIT RISK MANAGEMENT WORKSHOP
18 September 2003

Good morning, ladies and gentlemen. I was very pleased when Michael Coogan asked me to give the keynote address today, since this is an opportune moment to be reviewing the mortgage market as it stands now, and the implications of future regulation.

My background is in accounting, and I have limited direct experience of credit risk management. However, I understand that bankers talk about their revered "canons of lending" and I see that these still feature in the syllabus for the Institute of Financial Services practitioner exams. For those few of you who have trouble remembering what the "canons of lending" comprise, the IFS syllabus suggests a handy mnemonic: CAMPARI – not the alcoholic drink customarily mixed with soda but the initial letters of Character, Ability, Margin, Purpose, Amount, Repayment and Insurance. You may remember a different mnemonic, but that spelt something too racy to form the basis of a public speech. So instead I intend to use CAMPARI as the framework for this morning’s address.

Now, of course, the CAMPARI canons do not form part of the FSA’s rule book, but they have historically provided the Banker with a tried and trusted model for credit analysis. So, always mindful of the Principle of Good Regulation that requires me to recognise the responsibility of senior management, I thought I would use your canons as the basis for a review of how the industry is managing the credit risks inherent in mortgage lending. In addition, I want to use the canons as a framework for thinking about the consumer issues that arise in the mortgage market, and how those will need to be addressed once they become regulated from October next year. I hope to raise some issues that you could usefully be thinking about over the next 12 months, and provoke you to review your plans!

Character

Times have changed from the days when lenders knew their customers by name, and had personal knowledge of their finances. Today’s assessment processes (using computer searches of credit reference data, CCJ records and electoral rolls) are a highly efficient means of checking the data supplied by the prospective borrower, and help you gain an overview of the individual’s financial circumstances. They also help to meet some of the anti-money laundering requirements. Nevertheless, there are still some questions to ask about the way this data is used by firms.

There are suspicions that credit scoring make it more difficult for individuals of good character who don’t happen to have credit histories to get credit on reasonable terms. And there must be some doubt whether the deliberate targeting of individuals with poor credit histories as a profitable "sub prime" segment of the market would have found favour with the original drafters of the canons. They would have understood the economics of risk based pricing, but not necessarily the idea of lending money in the sure knowledge that a significant proportion of it would not be repaid. After all, the canons were designed to be rules for good lending, and the ideal would have been a 100% success rate in achieving repayment. No doubt you would argue that "good" sub prime borrowers would be deprived of access to credit but for the approach taken, but I can’t help thinking that, from a customer viewpoint, the more the industry can do to refine its scoring systems to separate the simply unlucky from the serial defaulters, rather than just playing the risk/reward percentages, the better its reputation will be

From a credit control viewpoint, it is clearly essential that you control sub-prime portfolios especially carefully, since the economics are finely balanced. The sub-prime theory has not yet been properly tested by a recession, and I’m sure you wouldn’t want to bet the bank (or building society) on it working 100% successfully. ; Everything in moderation, as the authors of the canons would say.

Ability (to repay)

For the purposes of today’s speech, I have chosen to define the second canon as meaning ability to repay. This should, of course, be of particular interest to you, since our new rules on responsible lending applying to mortgage lenders from October next year include a requirement for you to "be able to show that…account was taken of the customer’s ability to repay the loan". And assessing compliance with this rule will clearly be one of the important elements of the supervisory regime we apply to mortgage lenders from next year.

So, how far do mortgage lenders currently take due account of ability to repay when making their lending decisions? It is certainly the case that the credit scoring systems used to establish character also provide a good guide to a customer’s propensity to repay a debt. But propensity to repay is not necessarily the same as ability to repay – in order to assess the latter, a lender would have to take a view on whether the amount of the repayment was realistic given the borrower’s other financial commitments. A scoring system based mainly on evidence that a borrower has in the past been able to meet his or her credit commitments will not necessarily show that a new mortgage is affordable, since an additional loan might actually be the trigger for financial overload.

And we all know that the rate of growth of consumer credit, including mortgage credit, suggests that the risks of such financial overload must be increasing. The FSA said back in January that we thought the rate of growth was unsustainable, and we highlighted in our Financial Risk Outlook for the current year the problems already being experienced by a large number of borrowers. If anything, the position has worsened since then, with figures from the Bank of England showing consumer credit up 14.2% in the year to July, and mortgage borrowing up 13.9% over the same period. Subscribing to the adage that "a picture is worth a thousand words", here is a graph showing the growth in consumer debt gearing.

The point to note is that, at 130% of income, total debt gearing is now higher than it was in the late 1980s, and that was before the record levels of borrowing seen this year, including in each of the past two months. I simply do not believe that individuals can continue to borrow at such a rate without unpleasant financial consequences further down the line. Indeed, only a couple of weeks ago, the Citizens Advice bureau reported that, with debt problem cases up 50% over five years, one in five people who have credit use it to cover everyday household bills, a quarter of them struggle from time to time to keep up payments, and a tenth have more serious financial problems. These difficulties may not be evident in mortgage arrears numbers – well, not yet anyway – but the sharp increase in personal bankruptcies, to their highest level since 1993, is a clear indicator that not all is well. And that is before the effects of the new Enterprise Act have been felt. So it would be wrong to become complacent about the risk.

Against that background, and with the prospect of the new rules, it would seem timely for you to ensure your credit appraisal systems, whether automated or not, take account of other outgoings of the borrower, credit related or otherwise. For example, given the current divorce rate and trend, family support payments must be a significant figure for a number of borrowers.

But of course assessing ability to repay implies that it is not just outgoings that need to be reviewed. I can see why lenders have responded to competition and cost pressures by cutting back on positive verification of income, though I’m not convinced that the authors of the canons would be so sanguine. And clearly the corollary of responsible lending, responsible borrowing, requires that prospective borrowers are truthful in the dealings with lenders - they certainly could not claim a mortgage was irresponsibly lent if they had lied in self-certifying their income. But how can you ensure ability to repay, if you have not performed at least a "sanity check" of the information given, to see whether there are no reasonable grounds for doubt, before accepting an income figure for credit decision purposes. If your credit appraisal systems are not already attuned to spotting such anomalies, you may have some work to do before October next year.

What is less clear to me is how well your assessment processes take account of the possible impact of future interest rate changes on ability to repay. This is especially relevant when you are offering incentives such as short-term discounts or fixed rates at below your prevailing standard variable rate, since an increase in the monthly instalment at the end of the incentive period is much more likely to occur. At the very least, a "canonical" lender should perform a stress-test of these borrowers’ capacity to repay at the standard rate, but there is a strong argument for stress-testing all new borrowers against a higher rate that might realistically occur during the first few years of the mortgage. After all, a 2% increase in the current level of base rates could translate into a near 50% instalment increase for a number of variable rate, interest only mortgage borrowers. For people who are already fully borrowed this could be financial disaster. Our new "Key Facts" document that lenders will have to provide to all prospective borrowers will highlight the risk to borrowers, but if you do not already do so, you might also want to think further about stress-testing disposable income levels. Especially given that, in the case of some borrowers, these have actually fallen over the past few months as a result of National Insurance and Council Tax increases.

Margin

The next of the canons is one near to your heart – Margins. So how well does the modern mortgage industry manage the risk to reward ratio through margins?

I have already mentioned the sub-prime market, so I won’t labour that point – other than to note that, whilst it does show that some progress has been made from the era when all mortgages were offered on the same terms, it also demonstrates the trend of some lenders to look for additional margin by taking higher risks.

In the prime market, it must be said that much of the risk pricing we see now seems still to be based on the margin of security offered by the property value, rather than on the relative strength of the borrower’s covenant. This suggests that the value of the security held may wield a disproportionate influence on the risk management approaches of lenders. I will return to this point later.

There are also some signs that the long-running business strategy, of attracting new borrowers with discounted rates financed, largely, by the inertia of existing borrowers, may be reaching the point where there are insufficient numbers of inert borrowers to make it work. The overall implications of this for margins and rates have not yet worked through, but the competition appears, if anything, to be intensifying. From a reputational viewpoint, firms that communicate the existence of lower rates to their existing customers are more likely to be well-regarded than those who do not. And given the cost of acquiring new customers, it has long baffled me why lenders don’t look harder for ways of encouraging loyalty rather than promiscuity in their customer bases.

Of more concern to us as prudential regulators is the growing tendency of lenders to avoid the competition for prime business and seek higher margins elsewhere. The issue for us is whether those lenders that have chosen the higher risk, higher margin route have adequate systems for pricing the risk. Those with long memories may well recall how the mortgage lenders of the late 1980s that topped the building society league tables were those with high growth rates and apparently high profitability, usually based on above average margins and income boosted by arrears fines or other charges for default. Many of those firms did not survive the downturn of the early 1990’s because their business models failed to price adequately the risks they were taking. What looked good in boom conditions was found wanting once house prices stopped rising – and seriously misjudged when house prices fell.

In the current market, the jury is still out on whether buy-to-let lending is really the banker’s holy grail of a low risk, high margin pursuit. Or whether a squeeze on rental yields, coupled with longer voids (such as has been seen in the London market) will undermine the financial viability of transactions.

The push of some firms into commercial mortgage lending raises similar issues, but I don’t intend to comment on these today.

Purpose

On, then, to "purpose". When the canons were first devised, I have no doubt that the purpose of a home mortgage was inextricably linked with house purchase. Even quite recently, additional borrowing for improvements or extensions was almost always documented as a further advance, subject to different terms and rates. Of course, there were tax-related reasons for keeping house purchase borrowing separate that no longer apply, so it is not surprising that the market for secured loans has become more generic.

What is perhaps more surprising is the extent to which borrowers have taken advantage of market liberalisation to the extent they have. Another picture:

This shows the actual amount of mortgage equity withdrawal, rather than withdrawal as a proportion of GDP or retail spending. However, mortgage equity withdrawal was equivalent to 7.3% of household consumer spending in the first quarter, compared with the record of 7.7% in 1998. The recent rate of growth is quite simply staggering, and although some mortgage equity withdrawal has simply replaced other forms of retail borrowing - a CML/Bank of England survey suggests that nearly a fifth of equity withdrawal was to pay off or consolidate other debts – overall consumer debt has continued to rise. So what is clear is that consumers have been cashing in the increasing value of their properties and spending the proceeds on the high street (possibly in reverse order). Which is all well and good from a macroeconomic standpoint whilst it supports GDP growth through a downturn in other sectors, but if spending falters before the rest of the economy is in a position to take up the running, the process could go into reverse.

So have mortgage lenders seem to have been paying less and less attention to the purpose of new loans granted? And why is the purpose of a loan a key factor in the traditional approach to lending appraisal represented by the canons? Does it matter?

At one level, it is easy to show that the purpose of the loan has direct implications for repayment prospects: if we do suffer a general economic downturn and unemployed borrowers need income support, any interest payments from the DSS will be based on the amount lent for house purchase purposes and nothing else. This may seem a remote possibility at present, but the value of income support in staving off repossessions in the early 1990s should not be underestimated – even if the current scheme is less generous and only cuts in after nine months.

At another level, it seems to me quite likely that the risk profile of a borrower seeking a to borrow for a house extension represents a very different risk to one who is looking to pay off credit card borrowings. The fundamental aim for both lender and borrower should surely be trying to obtain some sort of match between the term of the loan and the life of the "asset" financed. In the case of an extension, it clearly makes sense to look at repayment over a long period, especially since the value of the property will be probably enhanced by the process. And there are clearly reasons why a retired person could see merit in releasing equity from his or her home to boost income, though trading down would seem to be a much safer option than a lifetime mortgage. But is it sensible to finance a car that will be kept for three to five years with a loan that lasts twenty-five years? And if that is the case, how can anyone justify using a long-term mortgage loan to pay for a holiday that lasts two weeks, or last week’s grocery bill? Except, of course, if the borrower is desperate.

Amount

Given all I have said already, I’m not proposing to spend long on "amount". Clearly the size of the loan requested is a key component of the credit appraisal process discussed under ability to repay, and also is connected to its purpose.

My main observation under this heading relates to the tried and trusted income multiple calculation that is often used to determine how much can be lent. Whilst recognising that this is a fairly crude measure, lending at higher multiples (over 3 times single income or 2.75 times joint income) roughly doubled over the four years to 2002. The most recent figures show some signs of moderation, but it is probably too early to say whether this supports the theory that we have seen a one-off adjustment to take account of the low inflation environment or whether it simply reflects the current dearth of first time buyers.

But if income multiples are going to stabilise, house prices ought to do so too.

Consider this graph:

Leaving aside that the last time the blue house price index line was above the red earnings index line was in 1989 (need I say more?), does it really look likely that the blue and red lines can continue to diverge? Especially given how closely they correlated in the 1970s and early 1980s. I will return to this when I look at the implications for security values.

Repayment (source)

The penultimate canon, source of repayment, raises significant issues for a regulator. I have no doubt that, for the authors of the canons, having a clear source of repayment was absolutely critical to the decision to lend. Yet, as I survey the current mortgage market, I am struck by how little importance you seem to attach to this. Since I have already dealt separately with "ability to repay", I want to concentrate now on capital repayment rather than meeting monthly instalments.

You will probably argue that, since the majority or mortgages are now arranged on a repayment basis, the source of capital repayment is explicit within the mortgage terms. But in how many cases are these repayment mortgages agreed for terms that long exceed the realistic working life of the borrowers. I have read that some of you are now offering 50 year mortgages, which suggests either that you think the retirement age is going to be lifted to a minimum of 68, or else that you don’t really expect overall repayment to be achieved from income. For some customers there may be a good reason to take out such loans. But will the majority of borrowers understand that the loan isn’t really "cheaper", since they will pay a lot more interest over 50 years than when borrowing for 20 years.

And it is not just the period length of repayment mortgages that raise such issues. I don’t want to dwell on the use of endowment policies as repayment vehicles – I know that is a sore subject. But I am struck by the thought that, twenty or so years ago, it would have been customary for a lender to register an interest in an endowment policy, even if a full assignment of the policy was not taken. That way, at least, the lender would be aware if premiums on the policy stopped being paid or its projected terminal value fell, and could take action to convert the mortgage to a repayment basis. Of course, the administrative costs of the process put paid to noting interests years ago, but, as a result, how many of you can say for certain that you know there is actually a repayment vehicle in place for your interest only loans, let alone whether it will be adequate to repay the capital? Or that there is life cover on the borrowers that will pay off the loan in the event of their premature death? And what about all those pension, PEP and ISA mortgages that were so popular a few years ago. How are those repayment vehicles doing? Do you know? Do you care? Putting a disclaimer notice on the back of the annual mortgage statement reminding interest-only borrowers of their responsibilities will not ensure repayment, and experience shows that early action on a potential default is the best way to minimise loss, and manage reputation.

And remember, an often forgotten consequence of low inflation is that the burden of repayment does not reduce much over time, The adjustment of expectations to a low inflation environment is a theme we at the FSA return to time and again, because it seems to us that neither consumers nor firms have fully reflected the consequences of low inflation in their behaviours.

Insurance (security)

So if the industry’s performance against the "repayment" canon raises doubts, what about the final one. For the purposes of the mnemonic, this is "insurance", although it is generally taken to mean "security" – rather than being an early reference to the practice of cross-selling insurance products to borrowers! The point, of course, is that security taken as insurance against failure of the borrower to repay. No self-respecting banker would make a decision to lend just on the basis of the value of the security held – that would be pawnbroking rather than banking.

Now, I confess that I might be pushing it too far to claim that current mortgage lending has more in common with pawnbroking than banking. Though, as you will have gathered from my remarks on assessing ability to repay and ensuring repayment of capital, it does seem to me that the industry puts too much emphasis on the current and potential future value of the property held for security and not enough on the borrower’s covenant to repay. It is interesting that LTV ratios of new loans have remained relatively constant, whilst income multiples have risen.

However, if your lending strategy is based on your expectations for the future direction of house prices, and your borrowing customers are betting on a rising market, at the very least you need to stress-test your mortgage book for stability in adverse circumstances. Whilst a fall in house prices may look unlikely, it is by no means impossible. I have already alluded to the risks to the economy from overindebtedness, and potential changes to the relative attractiveness of buy-to-let property. Add to that the relative dearth of first time buyers, who underpin the market, and you have to wonder when house prices will begin to comply again with the laws of supply and demand. The shortage of supply may ensure that the market equilibrium price is above that which would normally be considered reasonable, but there is no law that all employment has to be in South East England, whatever the cost of living there. And there is plenty of empirical evidence to suggest that markets have a habit of overshooting – on the downside as well as the upside. Another graph:

The question is whether the recent increase in house prices is supported by economic fundamentals, or is really part of a series of speculative booms. The graph shows the Nationwide house price index plotted against the FTSE 100 index for the same period. If we lag the FTSE 100 index by two years, this is the picture

If we lag the FTSE by three years, we get this picture

And if we are as positive as some industry commentators we can lag the FTSE 5 Years, and we see two more years of growth:

What do you think ?

Whatever you think, we can probably agree that whilst there is strong anecdotal evidence that some consumers may see their property as an alternative to "traditional" pension investment, the recent capital performance of residential property is clearly unsustainable. And concentration of investment in a single asset class (or even a single asset) breaks all the portfolio diversification rules – especially when you have borrowed the money to buy the asset, rather than saved it. (i.e. you are heavily geared, or buying on margin!)

So is your security capable of withstanding economic shocks such as rising unemployment, or falling house prices, or both simultaneously. And, do you know how and when to scale back new lending if there are signs of stress – history suggests that, in a recession, most bad debts arise from the newest lending, so being able to turn off the tap soon enough is important.

Finally, if your strategy is based on security values, you need to resist temptations to cut back on staffing in your collections areas, since you know as well as I do that early contact with a borrower in difficulties is vital – both for you and for them. The FSA is pleased to note initiatives such as that between the BSA and Shelter to encourage borrowers to contact their lenders, and it is a message we will be seeking to reinforce under our consumer education remit. As you know, under the new regime you will be required to give our leaflet on arrears to all borrowers that get into arrears.

Summary

So, what conclusions can we draw about the relevance of the canons to today’s lending practices? Well, it certainly looks to me that like new drivers passing their test, new bankers can forget what they were taught, and I do wonder what mark your current practice would be given if tested against IFS examination standards.

Moreover, I think there are other lessons to be learned from applying the canons to the consumer dimension of lending policy. As I have said in previous speeches on this subject, the FSA’s view is that the senior management of a firm are ultimately those responsible for ensuring that the interests of consumers are properly considered in corporate strategy. This is not a responsibility that can be abdicated, or delegated to the marketing department. And it extends across all parts of the business, including all regulated subsidiaries.

Putting the consumer at the heart of your business, designing products that meet his or her needs rather than those of your sales force, setting realistic targets, and monitoring service standards are all part of a customer focussed strategy. We will be looking for evidence of such an approach in all regulated firms, including from next year in those mortgage lenders that are not currently regulated.

Thank you for your attention. If you have any questions, I would be pleased to try and answer them now.

More Speeches: