Response

BSA response to the FSA’s CP 09/17

A specialist sourcebook for building societies: enhanced supervisory guidance on financial and credit risk management

Contact: Jeremy Palmer Date: 4 Sep 2009

Print page  |   Email

Introduction

The Building Societies Association (BSA) represents mutual lenders and deposit takers in the UK including all 52 UK building societies. Building societies have total assets of over £370 billion and, together with their subsidiaries, hold residential mortgages of over £245 billion, more than 20% of the total outstanding in the UK. Societies hold nearly £240 billion of retail deposits, accounting for more than 20% of all such deposits in the UK. Building societies also account for about 36% of all cash ISA balances. Building societies employ approximately 50,000 full and part-time staff and operate through approximately 2,000 branches.

General points

We are grateful for the opportunity to provide the FSA with a detailed response to CP 09/17. At the time of publication, we welcomed the positive and sensible messages about building societies that FSA put up front in the opening paragraph of the consultation paper(1), while reserving our position on the detail. We are pleased that FSA designed its proposals to encourage the maintenance and promotion of a strong, vibrant mutual sector, though we strongly question below whether that will in fact be the outcome.

We also welcomed the assurance that this sourcebook was in no way intended to restrict societies’ ability to compete with banks – though again we now doubt that the result will be as intended.  We said that we looked forward to working with FSA through the consultation period and beyond, on improving and refining this first draft. This response is an important part of that process. We also welcome the direct engagement of FSA’s staff in explaining the proposals, first at a preview seminar at the Association’s annual conference, and more recently, at a seminar we organised on 16 July.

Our main overall concern is that much of the detail of the proposals simply fails to match the high level objective that FSA set itself – of promoting a strong, vibrant mutual sector – or the assurance of maintaining societies’ ability to compete.   In many areas, the proposed sourcebook will lead to unintended consequences for   individual societies.  And  we question whether the aggregate impact of the introduction of this sourcebook in the current market conditions has been appropriately considered and analysed.

Some of the proposed limits are so restrictive that it will be difficult to prevent breaches, a situation in which societies would not wish to find themselves.  There is no mention of what penalties could be applied in the event of a breach.

In short, there is a significant risk that, far from strengthening mutual building societies, the initial effect of the sourcebook may be to weaken them, in particular by uniquely fettering their ability to compete.  We expand on these central concerns in the following paragraphs.  But the consensus among our members favours suspending the finalisation and implementation of the sourcebook at least until the FSA’s own mortgage market review is complete and the liquidity regime agreed. It cannot make sense now to proceed with this sourcebook  which is based in part on prescriptions decided in ignorance of the findings of the mortgage market review.

While we resist in particular the discriminatory regulation of societies’ mortgage lending in advance of the conclusions of the mortgage market review, moves to clarify the framework for the prudential supervision of societies could be helpful, especially where this mirrors the familiar financial risk management approaches.  CP 09/17 explains(2) that it hastens the convergence of all societies onto sound, well-controlled strategies – faster than would have occurred solely as a result of FSA’s case by case supervision. More important for societies, it should bring welcome transparency and clarity of expectation, and should protect them against arbitrary and inconsistent supervision – a particular concern in the current febrile climate.

Discriminatory and anti-competitive effect in mortgage market, and pre-emption of FSA’s own review

But this improvement risks being completely overshadowed by societies’ major concern that a new, more narrowly focused, and potentially quite restrictive regime is directed only at them.  Societies fear that the imposition of this sourcebook on just their sector will place them at a competitive disadvantage - unless equivalent provisions apply to banks that are active in the mortgage market, and even to such non-bank mortgage lenders as remain. Otherwise, competing mortgage banks will continue to operate outside this guidance, unaffected by the various restrictions which would apply to societies. They would not be required to “comply or explain” against this or any other equivalent guidance.

The introduction of this guidance would therefore benefit those mortgage market participants who continue to underwrite their chosen risks freely in their lending activity, can rely more heavily on wholesale funding, and can take market share at the expense of the more heavily regulated societies. Arguably, it is some of those participants, who would remain outside the scope of such regulatory guidance, who are, and have been, most in need of it. 

And it is on this point that we must highlight the internal inconsistency with FSA’s policy stance on mortgages generally. As explained in the Turner review(3), FSA will shortly begin a thorough review of the case for mortgage product regulation weighed against using comparable policy levers such as tighter regulation of selling or more aggressive use of differentiated capital requirements. Any such moves should, quite rightly, be evidence-based – as the Turner review says: “The FSA’s paper on regulating the mortgage market will assess the strength of the arguments for and against. It will analyse the extent to which customer defaults and bank losses are correlated to either high initial LTV or LTI……”.  However, the proposals in CP 09/17 take these and other correlations for granted, thereby pre-empting (but in relation to building societies only) FSA’s analysis before it has even begun.

The FSA paper will moreover cover the whole mortgage market, and any conclusions would be applicable to building societies, mortgage banks and non bank mortgage lenders – so potentially requiring, or arguing for, substantive change to the CP 09/17 regime in relation to lending at an early date. We do not see, in these circumstances, how the imposition in the interim of a detailed and prescriptive regime purely on building societies comes anywhere close to being justified.  We note that the regulator considered applying the sourcebook to mortgage banks but did not proceed. In the light of the evidence of reckless mortgage lending emerging in the published H1 2009 results of the failed or nearly-failed mortgage banks, and their extensive reliance on the Government’s asset protection scheme to cope with their toxic assets, the decision to apply a new lending regime to societies only looks especially perverse.

We suggest that any lending constraints should emerge – and only if justified by the evidence - from the mortgage market review and then apply to all lenders.  The track record of building societies in mortgage lending does not warrant immediate additional regulation over and above other lenders. So- as stated above – the implementation of the sourcebook should for the time being be suspended.

In principle, the FSA’s focus on the themes of improved control over lending, liquidity management and funding processes, is uncontentious. However, the production of detailed prescriptive guidance/rules and its application to one subset of mortgage market participants appears to be completely premature. If a set of detailed limits is seen as desirable, it should at least be applied simultaneously to all participants, and not just to building societies in advance of the rest of the market.  A cynic might suggest that it is relatively easy to put an early tick in the box against the need to overhaul the Building Societies’ Sourcebook when compared with the much more problematic exercise of regulating these issues across the whole mortgage market and its kaleidoscope of participating lenders.

We see this as an example of a wider policy inconsistency on FSA’s part.  We note from the Turner Review(4) that FSA rejects, with reasons, the separation of casino from utility banking – in favour of retaining “broad” banks. However, in Lord Turner’s letter to the Chancellor of 16 April on FSA’s supervision of Dunfermline, we see the following on page 8:

“It is worth noting how this issue relates to the ‘narrow banking’ versus ‘investment banking’ debate, which is discussed in The Turner Review. That debate relates to whether large complex commercial banks should also be allowed to be engaged in risky proprietary trading activity. But it is worth noting that building societies already operate under a ‘narrow banking’ regime which restricts them not only from investment bank style trading activities, but from investment commercial banking activities. There is no debate that these constraints on societies should exist: the issue is how tight they should be.”

Societies may therefore wonder why on the one hand the large retail commercial banks are allowed, indeed encouraged, to remain "broad" while the effect - even if not the stated intention - of the sourcebook is to make societies, already "narrow" by statute, even "narrower" still. Indeed - given the contrast with the prescription for banks  (including the failed or nearly-failed mortgage banks that have just reported egregious losses), there needs to be  a rather better justification why societies are uniquely to be "guided" to an even "narrower" and more traditional identity than legislation already requires.

The disadvantage is also apparent in funding. Mortgage banks and wholesale lenders will not have their funding, liquidity and financial risk operations fenced in by the same framework, and will therefore gain a new competitive advantage. Longer term wholesale funding and/or committed facilities are particularly difficult to find in the current market conditions, and given the limited market for further increases in retail funding, restriction of mortgage lending would appear to be an inevitable result. This is contrary to the Government’s stated objective of encouraging the banking sector to responsibly lend money back into the economy again, and thereby risks a further prolonging of the “credit crunch”. If applied to building societies alone, such wholesale funding limits would restrict the ability of societies to compete in the mortgage market.

Any such regulation should not be restricted only to building societies, and other participants are arguably more in need of regulation in this area. It was larger mortgage banks such as Bradford & Bingley, Northern Rock and HBOS whose heavy reliance on short term wholesale funding resulted in (or at least contributed to) their recent failures. Any prudential wholesale funding limits should therefore only be introduced across the whole mortgage lending sector, to apply irrespective of whether the lenders are building societies, specialist mortgage banks or mortgage lending divisions of complex banking groups.

While the risk management models are familiar, the three tier approach to lending represents a major change, which generally reserves any kind of diversification in lending (including higher LTVs, book purchase and widely-defined commercial lending) for the middle or higher tiers: the range of business available in the lower, “traditional” tier is limited.  Many societies will have to decide whether to discontinue such areas or opt for a higher tier and incur the systems, controls and expertise costs – even where the business in question, even if notionally “commercial”,  is in fact traditional to that society and justified by its favourable loss experience.

 We can see that FSA is drawn to a romanticised and nostalgic view of the market in which smaller societies operate – simple, wholesome, low-risk lending at administered rates. This “cottage industry” vision of smaller societies is no longer valid, and for FSA to attempt to fence small societies in to this vision simply endangers their survival. One genuine feature that many small societies aspire to offer is a more personalised underwriting approach that can cater for good-quality but unconventional loan applications that fall out of automated scorecards. The proliferation of cross-cutting limits and sub-limits threatens this flexibility, thereby undermining one of the few natural advantages that smaller societies possess. This matter was ventilated with FSA at the July seminar.

But the adverse effects will not be limited to societies – they will extend to first-time borrowers(5)  requiring higher LTVs, specialised borrowers such as those operating residential properties as guest houses/B&Bs, and so on. Even if in the abstract the sourcebook’s prescriptions were the right ones, there would still be a strong case for assessing their aggregate impact on the mortgage market, and certain specialised niches, when considering the timetable for implementation.  Here, however, FSA appears to have fallen into the very trap that was clearly marked out in the Turner Review – of taking measures in a downturn that enhance, rather than countering, procyclicality.  The plans for the sourcebook take no account of the need to transition sensibly from today’s macroeconomic position.  We consider that this is just as necessary in relation to new lending strictures as in relation to higher capital requirements(6) . Similar concerns as to aggregate impact relate to the funding side: the sourcebook, in conjunction with FSA’s new liquidity policy, will “steer” all societies towards a higher level of retail funding. The CP 08/22 proposals, taken together with supervisory pressure, will affect banks similarly.

The aggregate effect will be that, at a time of exceptionally low interest rates when savings mobilisation is already extremely difficult, all societies and other deposit-takers will be forced to compete for the very limited pool of extra retail savings that exists. We have not seen any evidence that FSA has addressed the macro-prudential implications here (which are, inter alia, that societies end up paying very high and unsustainable marginal rates for retail savings to an extent that is far more damaging to their core business than a more moderate pace of reduction in wholesale funding). Here too a sensible transition path is needed.

Sourcebook text as guidance: contradictions with wider FSA statements

We also envisage practical difficulties in applying detailed limits in what is essentially guidance. The draft sourcebook states that the guidance constituted by the various levels or approaches is ”neither mandatory nor exhaustive”.  Indeed, it could not be otherwise, given FSA’s own definitive pronouncement(7) as to the status of guidance in the Handbook:

“Whatever guidance is used for, it is not binding on those to whom the Act and rules apply, nor does it have ‘evidential’ effect. It need not be followed in order to achieve compliance with the relevant rule or other requirement. So a firm cannot incur disciplinary liability merely because it has not followed guidance. Nor is there any presumption that departing from guidance is indicative of a breach of the relevant rule.”

We commented in passing above that the proliferation of prescriptive limits can prove counter-productive and self-defeating. We are also aware of a number of apparent inconsistencies. But we believe there is even greater potential for inconsistencies and illogicalities to emerge only at a later date as societies grapple with the limits etc on a day to day basis. That is one reason why an important safety-valve discussed at the July seminar needs to be hard-wired into the text. 

Where a society for good reason plans to exceed one of the limits applicable to its current approach, it is clearly not sensible for this automatically to necessitate a move to the next higher approach across the board. Instead, in keeping with the promise “neither mandatory nor exhaustive”, the smart and efficient response is for the supervisor to agree a customised in-between situation. 

We regard it as imperative that this flexibility is fully described in the sourcebook text (not in the policy or feedback statement, which becomes irrelevant). A similar point arises from the welcome clarification that societies will not be expected to liquidate legacy positions in particular instruments that do not fit within the funding/ liquidity/ financial risk management approach to which they self-classify. This, to be effective, needs to be set out as a transitional provision to the sourcebook itself, not merely remarked on in the policy or feedback statement. The same also applies to societies’ transition to the final percentage limits on wholesale funding - this, too, cannot sensibly be done overnight. Finally, there needs to be a general safety-valve for inadvertent breaches of all kinds of percentage limits that result not from transactions that are subject to the limit but from shrinkage of the overall balance sheet and disproportionate shrinkage of other asset or liability classes.

There are many areas of overlap between sections in this sourcebook and provisions that would be more appropriate in an ICAAP document on risk appetite. Indeed, it has been pointed out that the risk appetite statement on lending is to be reviewed semi-annually or quarterly, while the ICAAP needs to be reviewed less frequently. It also seems odd that the liquidity policy is to be reviewed at least annually, less frequently than the lending risk appetite. The periodicity of all such reviews might benefit from a consistency review by FSA taking account of the CBA consequences of any individual review being “required” more often than annually.

We address the problem of the sourcebook text having ostensibly, though probably not in practice, the status of guidance further in our responses to Questions 1 and 4.

Responses to individual consultation questions

We set out below our responses to the specific consultation questions, and following that – in Annex A - a series of comments on the draft Handbook text (that do not fit easily with the consultation questions). 

Q1: Do you agree that it is desirable to structure these provisions as guidance to allow sufficient flexibility for them to be tailored to the individual circumstances of individual societies?
Yes, definitely, given the formal status of FSA guidance as summarised above. The tailoring to individual society circumstances can, if required, be formalised as individual guidance which involves far less bureaucracy than rule waivers.

Nevertheless, concerns remain about the practical application of the guidance by individual supervisors.  Notwithstanding the clear statement about the effect of guidance in the Reader’s Guide and the recent FSA statement, many societies have experience of their supervisor treating existing guidance as if it were rules.  In part, this may be due to the hangover from the IPRU era when the outward form of guidance was used for what were clearly intended to be binding requirements – the so-called “guidance plus” approach.  While this may have been a useful expedient at the time, it has unfortunately blurred the distinction between rules and guidance – some supervisors indeed appear ignorant of the true position.

So the inclusion of what appear to be hard limits in what is essentially guidance is potentially confusing for these same supervisors as well as for societies - and so risks being unduly restrictive in effect.  While the guidance may not (when in its final form) be mandatory, the use of explicit numerical limits will inevitably draw individual FSA supervisors to apply the guidance as if it represented a set of hard rules. It could be very difficult for an individual supervisor to accept internal society policy limits which are outside the guidance, and consequently we believe that internal society policy will rapidly come under pressure from the FSA to conform to the guidance. The guidance would therefore become a set of de facto rules which would restrict the activity of building societies.

The CP states ( paragraph 2.2) that guidance will be “neither compulsory nor restrictive”, but also repeatedly states (paragraphs 2.2 / 2.3 / 2.4) that each society must “comply or explain”. However, there is no clarification of what “explain” means. If it means “explain and convince the supervisor of the explanation”, and supervisors take refuge in the guidance as a default position, then the guidance will soon become both become compulsory and restrictive.  Indeed, the “comply or explain” approach unwittingly violates the very status of guidance as the FSA has reaffirmed it – because it reveals that the sourcebook text is intended to have evidential and presumptive effect: failure to comply with the guidance creates a presumption of failure to comply with the underlying rule, so there is something to explain, and an onus on the society to convince the supervisor to the contrary.  But this is incompatible with the FSA’s 12 August statement(8) –

“Guidance and supporting material do not set out the minimum standard of conduct needed to comply with a rule, nor is there any presumption that departing from guidance indicates a breach of a rule. If a firm has complied with the Principles and other rules, then it does not matter whether it has also complied with other material the FSA has issued.”

There is also a minor presentational point about the tabulation of these limits – in order to fit onto one page, the way the limits are expressed becomes compressed and elliptical– to the point where ambiguity is introduced.  Such tables may provide a useful overall summary but the detail should be properly spelt out in narrative text.
 
Finally, for societies, there is a question mark over what justifies the “steer”(9) .  There is a concern that their FSA supervisors may force them to move to a simpler – and less profitable – model even though they have individually proved themselves to be prudent and conservative entities before and during the current financial crisis.  Societies fear that general comments such as “building societies diversified without having the requisite skills and systems to manage the risks" show that the FSA appears intent on “steering” all mutuals to simpler models that are commercially unsustainable – the “cottage industry” mentioned above.  This runs counter to the current government’s policies of promoting sustainable growth among mutuals and encouraging resumption of residential lending in general, and may have a significantly anti-competitive effect on many, possibly the majority of, building societies.

Q2: Are there any details within the three treasury management tables which are inconsistent, either within a particular table or between tables? If so, please explain.

One inconsistency, which we have highlighted previously, is in the eligibility criteria for the standardised buffer / simplified regime, which differs between two recent consultations: this paper sets it as those societies currently using the matched, administered or extended approaches, whereas CP 09/14 Strengthening Liquidity Standards 3: liquidity transitional provisions appears to exclude those on the extended approach.  We are now clear that extended approach societies can be eligible for the simplified approach.

The inclusion of both a standardised buffer ratio or ILAS requirement and a minimum call/ overnight holding of 3% of SDL imposes an excessive burden on extended approach societies.  If the buffer deals with the stressed event and a society manages its remaining liquidity prudently, there will be no need for the apparently arbitrary 3% minimum.

More generally, the FSA has always maintained that the supervisory approach to treasury management categories was more of a continuum than five discrete steps, in response to the widely diverse business models in the sector.  Based on the fact that most of the wording in Section 1 of the draft handbook text is unchanged from IPRU(BSOC) we would expect this approach to remain unchanged.  However it appears from the various tables showing guidance (BSOCS 3.3.11G, 4.2.3G and 5.7.3G) for key areas linked to the approaches that this principle has been weakened to the extent that the FSA seem to be applying specific onerous limits.  This would therefore thwart the “continuum” approach which the FSA also are keen to maintain.

Q3: Are there any suggested limits details within the three treasury management tables which you believe have been set at the wrong level? If so, please explain.

Our members’ biggest overall concern relates to the limits on wholesale funding, as mentioned above. If the detailed guidance were to be adopted, the 5% < 3 months and 15% < 12 months limits in table 4.2.3G (for wholesale funding as a % of SDL) would present particular problems. If supervisors were to insist on these as hard limits, societies would immediately need to start substantially reducing their short term wholesale funding even further. This is already at much lower levels than before the “credit crunch”, and if anything, good treasury risk management practice would now indicate that more (rather than less), and more diverse wholesale funding would be desirable.  One society has suggested a net funding position “ladder” (liquidity minus wholesale funding) over time would be a viable alternative.

In general, building societies believe that the funding limits are too low and run the risk of lowering margins to unprofitable, even unviable levels.  On the matched approach, for example, there is concern that the drop in the maximum wholesale funding (below which those societies will then need to build in a margin of safety to prevent inadvertent non-compliance) from the current 25% to 15% is far too sharp and inflexible; 20% would be more appropriate.

Some of table 5.7.3G appears contradictory.  As mentioned above, a table is not the most appropriate way to present limits – other than in summary form - since it gives rise to questions that cannot be answered easily.  Also in table 5.7.3G, there are proposed limits on fixed rate lending / funding activity, which have been expressed as minimum thresholds for administered rate activity. This is presumably on the basis that the use of the term “fixed rate” in the 4th column header is taken to include variable rates such as base rate trackers and LIBOR-linked rates, which are fixed to benchmark rates and are outside the discretion and control of the relevant society. It would be useful for this terminology to be clarified.

Societies believe the limits on fixed rate products in the matched approach will put them at serious competitive disadvantage, particularly at times – as now – when fixed rate savings and mortgages are in demand.  Societies need flexibility to adjust to demand, for example, when market rates are high, borrowers prefer variable rates and the proportion of fixed rates mortgages slows.  Higher proportions of fixed rate products, suitably hedged, should not compromise societies’ ability to manage their margin. 

Finally on this issue, we would seek clarification of the CP text set out below in the context of comments made by the FSA at the July seminar:

“Overall, we intend that societies who can demonstrate that they have the requisite systems and controls should have complete flexibility to operate within the statutory limits set by Parliament.” (paragraph 1.8), and

“Publication of guidance specific to building societies is not intended to restrict their ability to compete with banks. Societies in the most sophisticated category will have complete flexibility to diversify up to the statutory limits set by Parliament.” (paragraph 1.19).

The recent FSA presentation on the matter then reinforced this by explaining that for societies on the “comprehensive” and “trading” approaches, the wholesale funding maturity structure limits set out in table 4.2.3G were likely to be replaced with a structure “self-determined as established in setting the ILAS”. We look forward to seeing this in any final version of table 4.2.3G rather than the specific numerical limits in the CP table as currently drafted.

A leading society has also contributed the following suggestions:

“We agree with the principle of societies setting such minimum thresholds, to ensure that each society always has room to manoeuvre when managing its margin, through adjustments to administered rates. Measurement of administered rate assets as a percentage of total commercial assets (rather than total assets) appears sensible, but we would expect that on a consistent basis, administered rate liabilities would be measured as a percentage of total retail savings liabilities (rather than SDL). Furthermore, we would expect to be able to take into account (perhaps using a weighted calculation) retail savings liabilities which will shortly revert to administered rates (say within the next 12 months). In any case, we believe that societies should have the option to develop alternative measures to monitor and control their margin flexibility. The numerical limits shown in the 4th column for the Five Approaches, should be illustrative rather than prescriptive if societies have suitable alternative measures / thresholds which are compatible with their risk appetite.”

Administered approach

Do the treasury investment limits exclude current accounts?  Where the 50% maximum for term deposits is mentioned, does that mean half the total investment limit or does it mean half the treasury investments held?

Matched approach

Under this approach, a society will no longer be allowed to invest in FRNs.  While their marketability, like CDs, can be questioned, societies find it hard to understand why CDs can be bought on the basis they are held to maturity, FRNs cannot be bought on a similar basis as the counterparties are often the same.

Clarification on “tracker products” would be appreciated as they appear to be non-administered but are not specifically mentioned in the table.  They appear to be the cause of significant market pressure currently.

The table states that the minimum administered rate assets are 75% of commercial assets, which suggests that when inverted, the maximum level of fixed rate loans is 25%.  Similarly, minimum administered rate liabilities are 75% SDL when inverted means the maximum level of fixed rate savings is 25%.  However, for the maximum stock fixed rate (> one year) it is 20% of CA + 20% of SDL.  That seems to contradict the first two limits and impose even lower limits for fixed rate loans and savings.

Does this mean that, while the overall limit is 25% in each case, societies will be required to manage their book to ensure that - within that limit – they have only 20% of loans/savings with a (residual) fixed rate period greater than one year? One victim of the text compression in the tables is uncertainty as to whether it is the original or residual period until maturity or repricing that is meant.

The next limit for the matched approach states that the maximum fixed rate lending/funding is 25% loans advanced/retail funding per annum (a limit societies find constraining).   Does this mean that only 25% of loans advanced in any one year can be at fixed rates, and similarly only 25% of new accounts can be at fixed rates - presumably by value not number in each case?  Are these limits pre- or post hedging: if the latter, there is a question whether a limit is needed at all (except to control implied basis risk).

The final limit, on maximum loans reversion to variable in any quarter, is 5% CA.  Does this mean that societies must attempt to manage fixed rate lending so that no more than 5% of the loan book will come off its fixed rate in a given quarter - presumably to restrict the proportion of borrowers potentially experiencing payment shock?

Extended approach

Societies question the removal of credit derivatives.  They argue that purchase (only) of credit protection is both prudent and desirable even for medium-sized societies. (It is already established by virtue of section 9A of the Building Societies Act 1986 that societies cannot assume risk by writing credit protection to sell to others.)

Why is the interest rate view (8.2G (2)) apparently mandatory (when virtually everyone gets it wrong)?  Surely this part – indent (2) – should be permissive?

Some societies feel that the minimum administered rates and liabilities limits are set too high.  The objective of this limit is to allow societies flexibility to change their rates in response to stressed market conditions.  Such limits on their own are inappropriate.

Q4: Do you agree that it will be useful to building societies for us to extend our published guidance on how they might comply with SYSC 4.1.1 R and SYSC 7.1.2 R to cover control of the lending book?

Yes, in principle, for the reasons of transparency and consistency mentioned in the summary. But we should also point out that the situations in which building societies lend are too omplex to be shoehorned into rigid stereotyped blocks, as was demonstrated at the BSA seminar in July.  Underwriting remains the requisite skill – as explained by a practitioner at that seminar, LTV is undeniably important in decision making but far from the sole factor.  Building societies also consider – for example on so-called “commercial” lending:

• investment vs owner occupied
• nature and financial health of the purchasing entity
• nature and covenant of the tenants
• nature of the premises and its re-saleability
• flexible use on resale
• market demand
• structure of the deal/ directors guarantees, collateral charges.

It would be unfortunate if the sourcebook’s focus on certain limited features such as LTV, fixed rate etc, discouraged societies from deploying their wider knowledge and experience in underwriting.

This concern relates back to the wider issue about the practical status of the “guidance” in the sourcebook, and whether the detail is prescriptive or illustrative. Specific instances relating to lending approaches are set out below.

In paragraph 2.6.1, the draft handbook text states that “the FSA would expect the lending limits which societies following each of the three lending models have in their lending policy to resemble the following table”.  The use of the word “resemble” prompts the question of whether the table is purely illustrative, and whether it is completely open to individual societies to determine any relevant limits for themselves. If this is the case, it would be useful and unambiguous for the FSA to state that the contents of the table are purely illustrative. However, if the table is intended to be prescriptive, then presumably societies would need their lending policies to “follow” rather “resemble” the table. The use of the word “resemble” unfortunately promotes ambiguity rather than clarity.

In table 2.5.2G: The confusion as to whether the guidance is prescriptive or illustrative is further evident in table 2.5.2G. For instance, against the row header of “Risk Mitigation” (and under the “mitigated” approach) it is unclear whether:

• the four lending risk mitigation methods listed (or any particular combination of them) are in any way compulsory (i.e. prescriptive guidance), or
• other methods of lending risk mitigation are also acceptable alongside the four methods proposed (i.e. they are purely illustrative examples). Careful underwriting of the lending risks would be the obvious alternative strategy, avoiding excessive risk in the first place or pricing fully for it, rather than taking on the risk and subsequently mitigating it using the methods listed in 2.5.2G. These methods are often expensive, and always retain some of that risk (albeit transformed into a counterparty default risk, or risk of capital losses on debt portfolio sales etc.). However, it is unclear whether a careful underwriting strategy on its own, would still be considered sufficient in this context, as it is unclear whether the methods listed in the table are intended to be prescriptive or purely illustrative.

If the sourcebook is interpreted strictly in accordance with the FSA’s own rubric on the status of guidance, then such detail can only be illustrative, not prescriptive, but the drafting of the relevant paragraphs suggests otherwise.
 
Q5: Do the three models provide sufficient granularity of controls over the mortgage book? If not, please explain what further gradation could be introduced.

Yes, in principle, subject to one important caveat mentioned in the summary.  We doubt that the added complexity from the introduction say of one or two more tiers of model would be matched by real benefits for societies. Far more important is the point discussed at the July seminar – where a society requires to go beyond its current approach in one area, it should not be remotely necessary to move up to the next approach – instead it should be able to agree a customised intermediate model with its supervisor. Although this is implicit in the concept of “neither mandatory nor exhaustive” we repeat that it is imperative that this flexibility is fully stated in the sourcebook text.

The absence of prescribed limits for certain types of lending (for example, self build, bridging finance, development finance, second charges, debt consolidation, concentration of large loans) is taken as leaving the setting of any such internal limits to societies’ discretion. 

Societies question whether the prescribed LTV and portfolio limits flex - in the countercyclical manner recommended by the Turner Review  - for prevailing market conditions.  We presume that any change of limit must be discussed and agreed with the supervisor, a potentially time-consuming exercise.

Q6: Are there any particular aspects of our proposed guidance which you believe risk placing building societies at a competitive disadvantage when compared to other financial institutions? If so, please describe what they are and which institutions you believe will be advantaged by this.

We have explained in the summary the overall risk that smaller societies are treated, and patronised, as “cottage industry”. And lending does not fit so well into a multi-tier approach as funding, liquidity and treasury management generally where there is a more obvious link between sophistication of management and sophistication / riskiness of product  - whereas in lending smaller building societies in particular tend to focus on niches and in doing so become specialists in managing and mitigating the associated risks.

The FSA might also wish to reflect on the responsible lending requirements it imposes on lenders and how these are often addressed by fixed rate mortgages.  These ensure borrowers have a level of certainty in their mortgage payments.  Unsurprisingly, societies therefore find the proposed restrictions a shock, especially in light of the Government’s recently stated aims to provide more longer term fixed rate mortgages.

If the proposed limits remain in place, societies will have to compete more aggressively for mortgage business up to 80% LTV, with a resulting negative impact on margin, or move into riskier non-traditional areas.  Both options appear perverse, particularly in the current climate – and the notional justification for the restrictive stance on LTV could vanish completely when the findings from the mortgage market review become available.

We also question why the setting of sublimits on lending appears to have had no regard to the actual profile of current mortgage lending across the market, which FSA will be aware of from the product sales data. Such a comparison would, we think, be illuminating – substantiating our case that these proposals risk fencing many societies in to a small part of the overall market.

Moreover, the limits appear to give no credit for successful “seasoning” of new mortgage loans. In the past, the FSA’s predecessor, the BSC accepted that mortgages that had performed satisfactorily for five years, however apparently risky they may have been at origination in terms of LTV etc, were demonstrably low risk. Why has FSA not accepted, similarly, that performing loans after a suitable seasoning period can and should be free from these various risk limits?

The risk is that these proposals could lead to such societies having to withdraw from areas of lending over which they have acquired experience and established good risk assessment  practice, because they find themselves assigned to a tier that places their particular specialism, or at least the business volume necessary to sustain it, apparently beyond their capability. This could apply to high LTV lending to first time buyers, shared ownership, buy to let, lifetime mortgages and semi-commercial loans for specific businesses or sectors. There are also three areas of particular concern to building societies generally:

a. Automated valuation models

A number of building societies use AVMs in conjunction with traditional valuation methods to assess the value of properties that are being used as security for mortgage loans. Any restriction on the use of AVMs could put building societies at a competitive disadvantage to organisations that are able to continue to use them.

Building societies only use AVMs as part of their underwriting and risk management processes and where there are good business reasons for doing so. They tend to be used to value properties of similar design and construction within a particular location, as this is where automated models are at their most robust.  Even in these circumstances they only form a part of the assessment of the societies’ security. They are not used where the property concerned is in some way unusual or when it is in a remote location.

Societies also only use AVMs on transactions where the loan to value ratio is low.  In transactions where the LTV is high, a physical inspection will be carried out, although this may still be supplemented with an AVM if it is of standard construction and there are a good number of comparable properties nearby.

We also note that the average age of chartered surveyors is high, with relatively few new professionals entering the profession.  If this trend continues and AVM development continues and becomes more widely adopted the sector could – if compelled to adhere to traditional valuation practice across the board -face capacity issues in the coming years.  This would not only put them at a competitive disadvantage in terms of cost (an AVM being less expensive than a physical inspection) but also, potentially, in terms of service (fewer surveyors to carry out physical inspections). 

AVMs are currently rarely used for mortgage originations, rather they are used more widely for low LTV remortgage transactions.  This could change in the coming years as confidence grows in the accuracy and reliability of AVMs and it would seem inequitable to prevent some societies from taking advantage of innovative products that do not dilute their risk appetite.

b.  MIG(10)  for loans over 80%

The proposal for higher loan to value mortgages to be covered by mortgage indemnity guarantee insurance could put building societies at a significant commercial disadvantage to other lenders who did not require such insurance.

Competitive market pressure means that even if there were sufficient supply of such insurance to meet increased demand (which we currently believe not to be the case) building societies would have little choice but to bear the costs of providing this insurance themselves if they wished to continue to price their mortgage competitively.   It would be a shadow tax. And in the past the regulator has expressed considerable scepticism about the robustness of MIG cover, so it is somewhat surprising that its use becomes quasi-mandatory.

The FSA mortgage market review is currently being carried out and a discussion paper is due to be published in the autumn. This review will be investigating the extent to which high LTV products are correlated with customer default and lender losses. But the requirement for MIG at > 80% pre-empts this review. Any such restrictions should be considered only after the review, and then in the whole market context and apply to all lenders, not just building societies.  The track record of building societies’ balances in arrears is consistently below the mortgage industry average and that of the banks.  Societies should not be subject to further immediate regulation in this area pending the findings of the review.

There could also be a severe restriction in consumer choice as many societies may choose not to offer high LTV loans to first time buyers and other groups.  This could further stifle the availability of credit to these groups so important to a fluid housing market. This may be particularly serious for first-time borrowers who are good quality but unconventional where many, especially smaller, societies are prepared to invest the extra time in personal underwriting of these cases.

One society has asked if the FSA would confirm that other classes of collateral security permitted in existing building society legislation could continue to be acceptable as an alternative to MIG.

c. Limit on buy to let and shared ownership

We disagree that buy to let lending and shared ownership is inherently unsafe compared with the generality of mortgage lending. As long as the lender makes the necessary checks on both the borrower and the property to make sure that the loan is sustainable and, in the case of buy to let, is confident that the rent will cover the mortgage repayment, this type of lending can represent a good business proposition for both borrower and lender. In the case of shared ownership, the lender typically has additional security because of the subordination of social housing grant, and the availability of housing benefit to cover the rent element.   Therefore, societies consider rental cover for buy-to-let, for example 130% for the limited approach (and only slightly less restrictive than the traditional approach), too high and out of step with commercial realities.

As mentioned previously, these issues are being considered as part of the FSA mortgage market review.  We believe it is more appropriate for this to be considered as part of that review and any conclusions and changes applied to the whole of the market.

Q7: Do you agree with our proposal to ask societies to pre-notify the FSA of material diversification into other areas, and supply the requisite risk information and ICAAP? If not, please explain why not.

We do not oppose a simple pre-notification to FSA of diversification above a fairly high threshold of materiality, as we understand the regulator’s need to know. Moreover, such pre-notification, where it is genuinely material to the business and prospects of the society, is arguably already required under Principle 11. 

There is some debate among societies and others both about the definition of diversification (see also Annex A) and the thresholds chosen.  Some activities listed in chapter 2.28, such as estate agencies, are not always seen as real diversification since, like mortgage lending, they are dependent on the housing market. (In support of this view, societies can quote the authority of section 92A of the Building Societies Act, where the test for diversification - reflecting the statutory principal purpose- is whether or not the proposed activities have a connection with loans secured on residential property.) While additional prescriptive detail is not welcome, some debate on the nature of true diversification for building societies is called for.

Chapter 6.1.1 (2) says that where a planned diversification requires investment of more than 5% own funds, or where the diversification is forecast to deliver more than 10% of pre-tax profits in the future, this will be material and should be “pre-notified” (again an area that needs definition.  At what stage in the investment appraisal process does the society get in contact with the FSA?). 

On what basis were the 5% and 10% figures chosen?  Is the investment as referred to above represented only by share capital or equivalent in subsidiaries, or total funds committed including loans?  Should it be calculated on the basis already specified in section 92A or on some other basis? More generally, this requirement also strays into the province of the ICAAP. The footnote to paragraph 2.30 suggests that under existing GENPRU rules significant diversification already triggers the requirement for a new ICAAP. But this is not what GENPRU 1.2.40 says – the requirement for a new ICAAP (in between the routine annual or more frequent assessments) is only triggered when “changes in the business, strategy, nature or scale of [the firm’s] activities or operational environment suggest that the current level of financial resources is no longer adequate” – a quite different proposition, and a much higher threshold.

We would like to know how the FSA intends to use this data.  For example, are the estimated exit costs to be included in the ICAAP?  If so, this could limit diversification in the future.

Q8: Are there any additional instances where you believe societies should pre-notify us?

This depends on what diversification means.  While we understand the regulator’s current need to know the detailed operations of a building society, we are concerned that such “pre-notification” will be seen as a barrier to societies’ operations and at best will delay business opportunities and at worst, will stifle them. It has potential anti-competitive effect too. Consider a small business that a bank and a building society are both in the bidding to buy. The building society has to go through the laborious pre-notification process, while in the meantime the bank – free of such constraints- secures the deal.

Q9: Do you agree that it makes sense to restructure this material in this way and move it to a specialist sourcebook? If not please explain your concerns.

As mutual entities, societies’ legal structure, operations and ethos vary enormously from shareholder-owned banks and it is only right that the regulator should recognise this.  Their conservative and prudent approach to lending has meant the sector has been less adversely affected by the financial crisis than banks.   As the consultation itself says, “the extent of that weakening has to date been less than that experienced by the banks – mainly because of the lower exposure to wholesale funding and complex financial instruments.”

But there is a concern about the potential for building society lending to be the subject of greater regulation than banks and other lenders through the proposals outlined in this consultation.  We note that the regulator considered applying the sourcebook to mortgage banks but did not proceed.   We strongly resist any measures that place building societies at a competitive disadvantage compared to other lenders.

Furthermore, we suggest that any lending constraints should be considered by the mortgage market review and if justified by its findings then applied to all lenders.  We do not believe that the track record of building societies in mortgage lending warrants immediate additional regulation over and above other lenders while the review is under way.

Q10: Do you agree with our assumption? If not, please identify those areas where you believe societies will rather strengthen their control systems, and explain why you think they will make this choice.

Without contacting the 52 societies individually, it is hard for the BSA to comment authoritatively.  But the assumption in chapter 4.5 that cost benefit considerations will lead societies to choose to amend treasury, liquidity and lending aspects of their business models and strengthen control systems appears correct.

Q11: Do you agree that data is available and IT costs will be minimal? If not, please identify the major changes which might be necessary, and an estimate of the associated costs.

Again, such figures can be provided only by individual building societies.  It is difficult for societies to quantify the costs attached only to the sourcebook implementation as they are also having to budget simultaneously for the significant costs attached to the new liquidity regime and forthcoming changes to the CRD   Most would not divide the three.  Building societies are also paying the price of failed banks in the form of FSCS levies. 

Societies are already under pressure to produce a seemingly ever-growing number of regulatory returns, including ad hoc information (some based on returns which the FSA had discontinued).  And some are attempting to grapple with apparent inconsistencies between this consultation and CP 09/17 over issues such as repo capability (see Adrian Coles’ letter to Sally Dewar of 6 August 2009).

Societies find that the sourcebook costs of £17.5k for a smaller society, rising to £55k for one of the larger ones, are low and do not comprise mainly one-off costs.  Moving to a different treasury risk management position, as some will have to do to continue the range of lending they have always done, and intend to keep on doing, has significant ongoing systems, control and staff costs.  Moving to a more risk averse profile of, for example, liquid assets will have an adverse effect on borrowers.

Despite the FSA assertion to the contrary, costs appear disproportionately high for those smaller societies which may have to strengthen controls and recruit the necessary additional resources to adopt the approaches most suitable to their business model.  The costs will represent a greater proportion of their assets than the larger societies, and certainly more than banks which can spread their costs more widely. 

Those societies on the traditional approach would be most restricted and yet the aggregate share of the UK mortgage market of the smallest 10 societies is approximately 0.0005%.  In fact, all societies, excluding the top 10, have a very modest aggregate market share .  It begs the question why a specialist sourcebook is being pursued this way.

Q12: Are there any other material general overhead costs which you would face as a result of our proposals? Please detail both what would cause the costs and an estimate of their value.

We would like the FSA to clarify whether the new BSOCS, taken together with the final version of the new liquidity policy, will replace the 8 day liquidity regime.

Q13: Do you agree with this analysis? If not, please explain why you believe we have under estimated the costs or over estimated the benefits.

Depending on its exact interpretation, the change in inter-society lending could lead to a reduction in funding for some, or driving up their costs as they attempt to raise funds elsewhere.  There are few sources of wholesale funding available to smaller societies: only a limited number of banks and local authorities lend to them and these numbers are falling in the wake of the current financial crisis.  There is a risk of a funding shortfall for smaller societies if the limits are tightened further – the CBA assumes that societies will reduce their inter-society holdings only to 100% of the new limit – i.e with no headroom – which seems implausible, so the figure of only 0.7% of funds at risk is also to be questioned.
Building societies support limits on how much they can lend to each other.  They wish to reduce the possibility of systemic contagion within the sector: what affects one society can easily – and quickly - affect another.  But they face further restriction in their sources of wholesale funding and a clampdown on inter society lending could stop some smaller societies lending.

More generally, just as FSA is unable (paragraph 4.26) to disentangle the incremental benefits of the CP 09/17 liquidity proposals from those attributable to the CP 08/22 proposals, so societies cannot meaningfully disaggregate the extra costs exclusively attributable to the CP 09/17 proposals from the substantial overall costs they expect to face from the implementation of the new liquidity regime.

Q14: Are there any other material costs (actual or opportunity) which you believe will result from our proposals? Please identify what they are and the magnitude of cost that will result.

Please see above.

Q15: Do you agree that our proposed enhanced supervisory guidance on financial and credit risk management for building societies is compatible with our statutory objectives and principles of good regulation?

The guidance is compatible with the FSA’s statutory objectives, but we challenge both whether sufficient regard has been had to some of the principles of good regulation, and also whether the sourcebook as proposed is the most appropriate way of achieving those objectives. 

As we have explained earlier, we have grave concerns that the proposed sourcebook will place building societies at a further competitive disadvantage compared to mortgage banks, the mortgage operations of clearing banks and also non-bank lenders – especially when the resulting restrictions will be implemented solely for societies in advance of a review that will provide the evidence base to justify –or not – wider product regulation. We therefore disagree with paragraphs 5.7, 5.8 , 5.10 and 5.11. By anticipating the full mortgage review, FSA is unable to substantiate that the lending restrictions to be imposed uniquely on building societies proportionate to the benefits expected to be obtained (and it is clear from paragraph 4.44 that FSA’s attempted justification is pretty thin). We find that paragraphs 5.10 and 5.11 show FSA simply in denial about the effect on competition. Finally, we consider that the lending restrictions in particular are highly likely to discourage innovation (again, paragraph 5.8 simply evidences denial on the part of FSA). We therefore conclude that, as presented, the FSA’s proposals fail against the 3rd, 4th, 6th and 7th principles of good regulation.

 
ANNEX A

Detailed comments received from member societies. These comments are supplemented by further detail contained in member societies’ individual responses to FSA – which, for brevity, we have not automatically included in this response.

Para 1.2.1

Building societies, like most lending institutions, exist to achieve maturity transformation - borrowing short and lending long. Paragraph 1.2.1G reinforces this, by rehearsing that societies’ principal purpose is (long-term) residential lending financed by (much shorter-term) members’ savings – but then says, completely illogically, that “societies should therefore adopt a risk-averse approach to maturity mismatch” . We detect some muddled thinking here: this section seems to be confusing maturity transformation with aspects of liquidity risk and interest rate risk. One society commented: “How are societies to interpret section 1.2.1(1) which “requires that risk limits are set ....to ensure that ...such exposures are minimised?”

To state that where interest rate positions are taken (and in practice all firms do) they should restrict potential reductions in economic value estimated under robust stress testing scenarios is too general. After the latest round of ARROW visits one society has said that it is tired of trying to comply with general guidance and then being told it has misinterpreted it.  It would be more helpful if it were spelt out that using technique x that a y% parallel shift beginning next month and persisting for 3 years should not give rise to a change in either net present value of cash flows of z% of general reserves and not reduce interest margin by more than w%. On the other hand, other societies have argued against numerical limits generally.

Para 1.3.2 (1)

To say that “Boards should aim to minimise ... credit risks from lending are minimised through careful underwriting” can be easily achieved through not lending at all: risk is zero, risk has been minimised, but the result damages both consumers and the wider economy.   A better formulation would be “Boards are expected to set lending policies which prudently balance the risks inherent in lending against the return obtained, taking account of the quality of property pledged as collateral, and most importantly the strength of the borrowers covenant to repay both interest and capital when required.  Boards should not assume that problem loans can be foreclosed without considerable loss to their society and the borrower.”

Para 1.14.2 (3)

Given that quantitative techniques have proven to be deeply flawed as far as credit risk is concerned (cf the ratings agencies) their mandatory use  - this paragraph requires econometric analysis - is of concern - particularly when no mention is made of confidence intervals to be applied, validity of supposed causal relationships, degrees of correlation etc.  The Turner review stated: “The very complexity of the mathematics used to measure and manage risk..... made it increasingly difficult for top management and boards to assess and exercise judgement over the risks being taken”.

Para 1.15.1

This needs to be clarified as it reads as if administered or matched building societies must have a floor on their SVR, when we presume that this is intended to give societies the option to have a floor on the SVR.  Any attempt to enforce a floor on building societies would have a detrimental effect on their competitiveness. 

Para 2.2.1 (3)

Is what is meant here “variable” rather than “discretionary” for example, bonuses and overtime? 

Para 2.1.11 (3)

Complexity may also arise on syndicated commercial loans.

Table 2.5.2G on lending policy approaches

The table does not make it clear whether 100% risk mitigation is achieved using MIG or whether it is envisaged that a stop loss policy is required on top of any MIG cover. 
In view of the relatively small size of building societies likely to take advantage of the traditional approach, we suggest that there is no benefit in segregating credit policy and underwriting (as with the limited approach).

Para 3.3.7

No mention is made of the difference between fixed and floating rate instruments (they carry different risks) and degrees of market liquidity for different instruments.

Table 3.3.11 on liquidity policy

Qualifying money market funds have no maximum allotted to them and yet can be sources of considerable valuation loss and lack true liquidity.

The gap between the limits applicable to the extended approach and those on comprehensive/trading seems a cliff edge - strict limits on extended, nothing on comprehensive/trading.  One society says that seems out of kilter and a possible source of structural unfairness/ uncompetitiveness - this is precisely the sort of two tier market structure which was one of the causes of the crunch: that the risk averse were squeezed out by the reckless.

Table 4.3.2G on wholesale funding

The maturity structure of wholesale funding appears problematic, as highlighted in the general responses above.  Due to short duration of current wholesale markets, this effectively limits wholesale funding to shorter duration limit (<12months), and is likely to cause many societies to have to replace wholesale with expensive marginal retail.  Not always easy, and possibly counter-productive.

Also, are there transitional provisions to help societies converge smoothly with  the applicable percentages? If not it is likely that some societies may be in immediate breach of these and liquidity provisions.

Para 5.2.6

What is the “financial risk audit”? Who conducts it, how often, do the “auditors” have to be independent? Can internal audit conduct it? If not, this could be a source of additional expense and duplicate work already being done.

Para 5.4.2

This does not mention regulatory/legal inflexibility in constraining a society's ability to manage rates due to TCF/unfair terms or other legal or quasi-legal considerations.

Para 5.5.7

With regard to reverse stress testing the FSA says: “...This provides the benchmark.”  Of what?  It may be useful to regulators but boards would regard a test to destruction as a catastrophe, not a benchmark – nor is it in any sense established that this is “good practice” – we suspect it is merely a fashion to which FSA currently subscribes. Single dimensional stress tests can be, but often are not illuminating.  We made this point in our response earlier this year to the FSA’s consultation on stress and scenario testing, CP 08/24.

http://www.bsa.org.uk/policy/response/stress_scenario_testing.htm

Para 5.6.3

Should the word ”treasury” be inserted before ”credit risk committee” to avoid confusion with retail (mortgage) credit risk committees where the skill set is quite different?

Para 5.6.9

We are concerned that the text on use of committed facilities again betrays muddled thinking, and also that this paragraph contains an ambiguity as to “vulnerability to withdrawal of individual deposits”. On the latter, any society is vulnerable, at all times, to a large-scale withdrawal of individual deposits. We think what is meant here is “vulnerability to individual withdrawal of large shareholdings or deposits” – i.e. following on from the previous subsection – can this be made clear please?

Obviously committed facilities have a price, and holding “higher than average” liquidity can also be costly – incidentally, is “average” the right concept here – surely what is meant is “higher than would otherwise be necessary” – i.e. in the absence of the vulnerability).  Section 5.6.10 then seems to undermine the role of committed facilities and seems to contradict the guidance in 5.6.9.  We think these paragraphs need to be re-written to give a more consistent message.

• Policy
• |
• Consumers
• |
• Media Centre
• |
• Publications
• |
• Events
• |
• Membership
• |
• Site Map
• |
• Contact Us
• |
• Website T&C's

© 2011 The Building Societies Association (BSA)