Liquidity Risk in the Integrated Prudential Sourcebook: A Quantitative Framework

Comments by The Building Societies Association on the Financial Services Authority DP24


1.   This paper sets out the brief comments of The Building Societies Association on the Financial Services Authority's Discussion Paper 24 Liquidity risk in the Integrated Prudential Sourcebook: a quantitative framework issued on 30 October 2003. DP24 seeks views on the design of a single framework of quantitative requirements to apply to all financial firms with significant liquidity risk. Liquidity risk is the risk that a firm, though solvent, either does not have sufficient financial resources available to it to enable it to meet its obligations when they fall due, or can secure them only at excessive cost.

2.   The Building Societies Association represents all 63 building societies in the UK. Those societies have total assets of over £215bn, about 15 million adult savers and over two and a half million borrowers. Building societies account for over 18% of both outstanding residential mortgage balances and retail deposit balances in the UK. In recent years total building society liquidity has been around 20% - 22% of total shares, deposits and debt securities (although there is a considerable range among individual societies). Almost all building societies have their head office outside London, providing much needed regional diversity, and their 2,100 branch network covers all areas of the UK.

3.   The objective of the proposed framework is related to the ability of a firm to survive a serious liquidity stress. The FSA considers it sensible to use a framework based on stressed, rather than normal, behaviour. The FSA envisages a limits framework built up by starting from the pattern of inflows and outflows due contractually in a firm's current business profile, and applying "stress factors" (or discounts) to the different components making up that profile. These stress factors are designed to reflect, on a prudent basis, the pattern of flows in a temporary serious stress. The FSA has derived its current proposals for them in most cases in a two-stage process: first to consider how the normal behavioural pattern of flows might vary from the contractual, and then how their stress behaviour might vary from that in normal conditions.

4.   The proposed new framework would replace the four current separate liquidity risk regimes – those for the largest retail banks (the sterling stock regime), other banks (the mismatch regime), building societies, and the liquidity risk treatments within their capital adequacy requirements for ex-SFA firms.

Current Liquidity Risk Requirements for Building Societies

5.   The current FSA rules and guidance for building societies on the management of liquidity, including guidance on what is meant by "adequate resources", in respect of liquidity, are in Chapter 5 of the Integrated Prudential Sourcebook for Building Societies. That Chapter replaced, but includes most of the substance of, the relevant Prudential Note issued by the former Building Societies Commission. Material on liquidity risk systems and controls is to be removed from Chapter 5 of IPRU(BSOC) at the end of October 2004, when the relevant provisions (PRU 1.4 and 5.1) of the Integrated Prudential Sourcebook are expected to take effect. The current evidential provision concerning the maintenance of adequate liquid resources by a building society (IPRU(BSOC) 5.2.4 E), essentially a "stock" approach – a society should keep at least 3.5% of total share and deposit liabilities in 8 day liquidity – is to be retained for the time being.

Q1. What reactions do you have to our views on the appropriate main features of a quantitative framework?

6.   The Association appreciates the desire by the FSA for a single integrated quantitative framework to be introduced for assessing liquidity risk for all banks, building societies and certain investment firms.

7.   Some building societies have indicated that, in principle, the proposed framework – calculating the gap ratio between stressed outflows and inflows – while significantly more complex than the current requirements for building societies, seems an appropriate measure of a firm's ability to survive a serious liquidity stress. However, the majority of building societies that have expressed views to the Association consider that the framework outlined in DP24 is too complex and theoretical, that some of the stress factors are unrealistic or inappropriate for building societies, and that a more tailored framework for building societies should be considered.

8.   A number of societies believe that the current liquidity requirements for building societies remain satisfactory and it is not clear that they are inappropriate for the future.

Q2. What are your views about our current ideas of the detailed elements to realise such a framework, their definition and derivation on input values?

9.   The Association has a small number of comments on particular aspects of the proposed framework.

10.   In general terms it might be helpful for further explanation to be given as to how some of the stress factors have been arrived at.

11.   Paragraph 4.28 suggests that an option for firms to simplify their gap ratio calculation would be to exclude components from the calculation on materiality grounds. It is proposed that the materiality threshold would be based on 1% of a firm's gap ratio. However a firm could not calculate that ratio until it had worked out all the figures, including those that might, in the event, be immaterial. So it is not clear that any simplification would be achieved in practice.

12.   A similar position would appear to arise with the proposed definition of a material currency (Annex 3 section V, and Annex 4) – one in which total stressed inflows or outflows are more than 10% of all-currency total stressed inflows or outflows. For a firm that had significant currency mismatches, the calculations would need to be done in any event to determine what the material currencies were.

13.   Presumably, it should not be necessary to make adjustments for such detailed matters as instant access retail funding (with penalty) (Annex 2 item 13) on which notice of a withdrawal has been given?

14.   It is presumed that all fixed term retail funding that does not mature (or allow access) within one week or one month would be included in retail funding: other (Annex 2 item 15). In which case should not the stress factors for such funding be lower than for instant access retail funding (with penalty), not higher?

15.   The suggested separate item for large retail funding (Annex 2 item 16) would be very difficult for many firms to determine for a number of reasons. The £35,000 maximum protected deposit under the Financial Services Compensation Scheme applies to each individual, not to each account held with the same firm. Accordingly, in the not uncommon event that an individual has one or more sole accounts with a firm, and/or one or more joint accounts (potentially with one or more different joint investors) with the same firm, it would be extremely difficult for most firms to track the amount actually covered by the FSCS in relation to individuals on a day-to-basis, or at all. Accordingly, we suggest that this separate category be dropped.

16.   Paragraph 4.47 notes that the treatment of commitments granted by a firm to unconnected parties is built on the general assumption (with which the Association concurs) that the level of draw-down would be little affected by the firm suffering a liquidity stress. Accordingly, the FSA suggests that the calibration would mainly be a reflection of assumptions about normal behaviour. It is a little surprising therefore to see that the one week stress factor for mortgage lending commitments to retail customers (Annex 2 item 41) is proposed to be as high as 20% (whereas the one month factor of 40% looks more reasonable). Average building society mortgage commitments during 2003 equated roughly to three months lending. (A proportion of mortgage commitments do not go through to completion.)

17.   In addition, there is a further item, under "I – Continuation of business", for mortgage lending (Annex 2 item 67), which is described in Annex 4 as new business added to the mortgage pipeline. However, it would seem highly unlikely that a new mortgage commitment could become an actual outflow within such a short time as one week (so the stress factor should be even lower than 10%).

18.   The stress factors for buy-backs of commercial paper (CP) and certificates of deposit (CDs) (Annex 2 items 69 and 70) seem unduly high in the context of a building society.

19.   In paragraph 4.53 the FSA states that is considering whether there should be limitations on the extent to which holdings of CDs (and perhaps other instruments) should be counted as marketable assets for core (or other) firms. Any such limitation would be of concern if the FSA were to consider applying it to building societies (such assets are a key source of funding and liquidity for may societies).

20.   The FSA proposes (in paragraph 4.48, example A) that the percentage allowance to be given to potential inflows from commitments held by a firm (Annex 2 items 99 to 107) would be increased as the number of commitments in place (of the same type) increased, as a reflection of the supposed prudence of diversification. The Association considers this to be inappropriate for a number of reasons –

  1. It is not clear why a firm that had 10 committed facilities of £1m each would be in a better position than another firm that had 5 commitments of £2m each, yet the former would benefit from a higher stress factor.
  2. li>The proposed arbitrary boundaries between 5 and 6 commitments, and between 10 and 11 commitments, would give firms an incentive to take one additional small commitment, so as to achieve a higher stress factor on all commitments (of the same type).
  3. A firm that had 12 commitments, equally spread between the three types – committed facilities, covenanted funding and standby facilities – would suffer the lowest stress factor in each category; but if all commitments were of the same type it would benefit from a much higher stress factor.
  4. The distinction between the three types of commitments seems overly complex in any event.
  5. A single committed facility might comprise a number of counterparties in a syndication, which would benefit from a higher stress factor if broken down into bilateral arrangements. This seems to create an inappropriate disincentive for syndication.
  6. In an actual stress situation it is thought likely that a building society's core long-term counterparties would be most likely to provide the more robust support.

21.   In Annex 2 sections O and P – marketable assets – it is assumed that mortgage backed securities should be included under corporate debt.

22.   It would be interesting to know whether the reference in paragraph 6.8 – that the FSA plan to provide considerable scope for a firm whose business is relatively simple to benefit from the options for simplifying its gap ratio calculation – would have any particular application to building societies. Apart, that is, from the suggestions in paragraph 4.28 that, to avoid certain detailed analysis, a firm may exclude immaterial components from the calculation (but see the comment in paragraph 11 above), and may combine a number of related outflows (or inflows), as long as the most prudent stress factor (or discount) is used.

Q3. Using Annexes 2 and 3, can your firm now provide input on our current proposed stress factors and data, to assist the finalisation of our proposals for a calibrated framework?

23.   The Association understands that a number of building societies have provided direct to the FSA their views on the proposed stress factors and discounts (Annex 2) and data to assist calibration (Annex 3).

Q4. Do you believe at this stage that the requirements will mean a significant change in your firm's balance of business and liquidity costs of complying with the regime? What other input relevant to assessment of the main areas of cost would you give at this stage?

24.   It is not clear whether there would need to be significant changes by building societies generally in their mix of liquid assets under the proposed new requirements. This might depend on a number of factors, including planned changes in future business mix. There might be increased liquidity costs of complying with the proposed new requirements compared with the current provisions.

25.   There will clearly be systems and resource costs to ensure the necessary data collection under the proposed new regime. At this stage quantification of such potential costs is difficult.