Thursday, 14 March 2013

Risk versus safety, bank reform edition

The Parliamentary Commission on Banking Standards has produced its second interim report. Predictably, the media homed in on its proposal to include provision in primary legislation for full separation of retail from investment banking across the entire UK banking industry if ring-fencing turned out to be a dud. Not that that would mean much - the only UK bank that still has a major investment banking arm is Barclays, and even that is being scaled down in favour of renewed emphasis on retail banking. In fact the way things are going, by the time the ring-fencing scheme is implemented it will resemble a plan to repair the door on an empty stable. The horses will have long since become Tesco burgers. Yawn.

But the report is actually far more interesting if you ignore ring fencing and look at the rest of it. It affords an extraordinary snapshot of the conflicting agendas of Government and Parliament at the moment. Government is in a hole, largely of its own making: the economy is stagnating, inflation is rising and the prospects of a Conservative victory in the 2015 election are getting slimmer by the day. The Chancellor's strategy for "getting the economy moving" rests almost entirely on increasing the amount of private sector borrowing in the economy - from households, mostly in the form of new mortgages (though there is growing use of payday loans too), and from businesses. This despite the considerable evidence that households remain over-indebted and the squeeze on real incomes is depressing discretionary spending, with significant consequences for domestic business. Lack of sales remains the number 1 concern of UK industry, particularly small & medium-size companies (SMEs). However, the popular perception is that "banks aren't lending" not because people don't want to borrow but because banks won't lend to reasonable risks. In vain do the banks - and the Bank of England - argue that the main obstacles to lending are a lack of creditworthy borrowers and subdued demand for loans: in vain do businesses argue that the main obstacles to their borrowing are lack of investment opportunities and subdued domestic and global demand. No, banks must lend. The Government's deficit reduction strategy (and their election campaign) depends on it.

But the Parliamentary Commission has a different agenda. Its objective is reform of the UK banking system to make it "safer" - or, more accurately, to reduce the risk that taxpayers will have to bail out depositors due to bank failures. The entire report focuses on measures to improve safety: nowhere does it recognise the Government's desire for banks to increase risk lending. So there appears to be a fundamental mismatch between the objectives of the Commission and the Government's current focus. The Commission wants banks to be "safer": the Government wants them to take more risks. It seems unlikely that this impasse can be resolved to mutual satisfaction unless both sides are prepared to compromise. And so far there are no signs of compromise. The Commission adamantly refuses to accept the Government's watering-down of some of its earlier recommendations: the Government rewrites some of the Commission's proposals in order to make it easier to dismiss them. The headline finding of the Commission was "More work needed". It's not just more work that is needed. It is agreement on the fundamental aims of the entire exercise.

The Commission is equally uncompromising when addressing the question of bank capital requirements. And actually its recommendations on this subject are both more interesting and more useful than the media-friendly ring fencing scheme. The Commission proposes early adoption of a leverage ratio as a backstop to the Tier 1 capital ratio, and recommends increasing the leverage ratio from Basel III's 3% to 4% in line with the higher capital ratios for both ring-fenced and investment banks already agreed.

The background to this is the mess that successive Basel committees have managed to make of capital adequacy regulations. Basel I's belt-and-braces approach was at least comprehensible, though there were some bizarre effects such as 50% weighting of domestic mortgages across the board regardless of the loan-to-value ratio (LTV) or the creditworthiness of the borrower. Basel II, in attempting to resolve some of these anomalies, vastly increased the complexity of the risk calculations, forcing regulators to allow banks to develop and use their own risk weighting models for more complex instruments. Among the banks allowed to use their own risk weighting models were HBOS and RBS. Need I say more?

Basel III has tightened up on capital adequacy, leverage and liquidity ratios, though the Commission is proposing even tougher requirements. But it has failed to resolve the question of complexity and still provides opportunities for banks to game the system, not least because it still allows banks to use their own models in some circumstances. Unsurprisingly, the Commission is unimpressed. In its view Tier 1 capital ratios calculated according to Basel III risk weighting rules can be fudged in many of the same ways as the Basel II ratios were. So the Commission wants another measure which does away with the suspect risk weightings and simply looks at the ratio of shareholders' funds to total assets - the "leverage ratio". To be fair, the Basel committee is proposing to introduce such a measure.....but not any time soon. The Commission wants it now. And it also wants the Bank of England to be responsible for reviewing how risk weightings are calculated. Somehow I get the impression that the Commission doesn't trust the Basel committee. Or banks.

Using a measure of total leverage as a backstop to the capital ratio seems eminently sensible to me. After all, the risk of any asset is at best a subjective measure, and it is subject to change without notice. Prime mortgages become sub-prime when unemployment rises and property prices fall. Sovereign debt becomes high-risk when fundamentals in the countries concerned (or a poorly constructed political experiment) give cause for alarm. The ability of supposedly "safer" assets such as mortgages suddenly to turn toxic is in my view a reason NOT to have a lower leverage ratio for building societies. In a property market collapse, building societies are as much at risk as banks.

Risk weightings cannot easily adapt to sudden changes in asset risk due to economic shocks, and banks do not like to admit that their assets are riskier than they appear so are reluctant to recalculate weightings in the light of changed circumstances. Therefore Tier 1 capital ratios can give a misleading impression of strength. The relationship of the risk-weighted capital ratio and the leverage ratio is itself a key indicator: increasing divergence between the two should be a cause for alarm. Although so should convergence.....after all, we don't want banks having nothing but high-risk assets on their books. Regulators should take a view as to the appropriate distance between leverage and capital ratios - as indeed the Commission is doing.

Higher capital ratios, a new leverage ratio, ring-fencing and separate capitalisation of retail operations.....all of these are supposed to protect taxpayers from bailouts. And the Commission is suggesting other measures too: depositor preference (where insured depositors rank ahead of unsecured bondholders in the queue for payouts after insolvency), greater use of instruments such as contingent convertibles (CoCos) that can be converted from debt to equity under distressed circumstances, a voluntary insurance scheme for deposits in excess of £85,000. None of these measures would prevent banks suffering serious losses and in some circumstances becoming insolvent: the bulkheads simply won't go all the way up to the deck. But what they may do is make banks easier to resolve and reduce the likelihood of depositor losses and taxpayer bailout. They won't stop the ship sinking, but there should be more lifeboats.

However, all this emphasis on safety comes at a price. Safe banks are risk-averse. They will only lend to good risks. The effect of higher capital ratios is to encourage banks to reduce asset risk - for example by requiring more and better collateral, and lending only to the best credit risks. And depositors like this. They do not want banks "taking risks with their money". They want banks to lend prudently. This conflicts with the needs of SMEs: lending to SMEs is about as risky as it gets, because most SME's are asset-light and cash flow is always uncertain. And it also conflicts with the desire of savers for good rates of interest. Lending only to the best risks does not make for good returns on savings. Savers may be protected from acute losses, but they won't make money on their savings. If savers want good interest rates, banks need to be able to take risks - and that means that savers (or rather taxpayers) must accept some risk of loss.

Protecting savers (or rather taxpayers) at all costs inevitably means reduced lending to SMEs unless unweighted capital levels are much higher. But more fundamentally, even a bank lending only from its own capital, or lending only to to the best risks, is not "safe". There is no such thing as a "safe asset", and therefore no such thing as a "safe bank". All lending is intrinsically risky, however it is funded and whatever collateral is pledged against it. The path to safety in asset portfolios lies in active management of risk, not in its avoidance. That applies to banks just as much as investors.

Even if risk weightings were not suspect, there would still be a fundamental problem with relying on them to define the safety of assets and, by extension, of the banks that hold them. Capital ratios based on risk weighted assets encourage banks not to manage risk but to avoid it. Just as reliance on agency credit ratings encouraged investors not to bother to manage risks properly, so defining "risk" by means of risk weightings encourages banks to rely on regulatory definitions of risk instead of understanding the nature of assets and actively managing risks themselves.Lending decisions start to be driven by their effect on capital, rather than by the balance of risks in the asset portfolio: no surprise that banks seeking yield would try to "game" the risk weightings to enable them to take more risk without appearing to do so. I would rather have a bank that was accepting higher risks and managing them properly, than a bank that was seeking out higher yields on apparently "low risk" assets. That's how banks ended up exposed to periphery sovereign debt. It was never "low risk", but the risk weightings masked its true riskiness. Capital ratios are no panacea. We rely on them at our peril. The leverage ratio is perhaps less easily fudged, and has the advantage that it does not discourage risk lending. But it too is no substitute for professional asset portfolio management by banks.

Real safety will not come from tougher regulations. And it certainly won't come from a ring fence imposed far too late to make any real difference. All these measures will do is decrease risk lending by the regulated sector and drive riskier enterprises, and savers looking for better returns, towards unregulated forms of financial intermediation. Real safety comes from professional risk management and commitment to customer service on the part of the banks themselves. Until these are established throughout the banking industry - retail banks as much as investment banks - there can be no real safety.

Related links:

Banking reform: towards the right structure - Parliamentary Commission on Banking Standards
Osborne should accept commission on banking standards' advice - Nils Pratley, Guardian
Supermarket banking - Coppola Comment
The illusion of safety - Coppola Comment
On risk and safety - Coppola Comment
A really scary story - Coppola Comment


  1. Frances,

    I don’t accept there is any conflict whatever between the need for bank safety and the need for stimulus, which is what you seem to suggest (unless I’ve got you wrong).

    Government just doesn’t need banks if it wants to effect stimulus: to put it crudely, it can just print money and spend it into the economy. The latter would work even if banks didn’t exist, and everyone kept their spare money under their matress.

    Indeed the latter is more or less what Positive Money and other advocates of full reserve advocate. That is, they advocate forcing those bank depositors who want their bank to lend on their money to accept the risks involved. That would result in less lending. But the deflationary effect of that is easily dealt with, to reiterate, by having government / central bank create and spend extra money into the economy.

    That way, everyone has a larger stock of money, thus banks don’t need to lend so much.

    Next, you claim that “Real safety comes from professional risk management and commitment to customer service on the part of the banks themselves.” I suggest that is pie in the sky. The idea that the crooks, fraudsters, spivs and carpetbaggers who run banks will VOLUNTARILY suddenly become responsible is plain unrealistic.

    Banks need a clear set of rules to follow which make it impossible for them to have taxpayers pay for the risks they take when the risks go wrong. As Mervyn King rightly pointed out in his last appearance before the Baking Standards Commission, we (i.e. Vickers) have so far failed to do this.

    1. Ralph,

      yes, you have misunderstood. I was pointing out the conflict between the Government's economic strategy and the objectives of the Parliamentary Commission. It is not possible simultaneously to both reduce and increase risk in banking.

      I fundamentally disagree with you that safety can be externally imposed. There are no such things as "clear rules". And the more rules there are, the easier it is to find loopholes - as a glance at the tax system should tell you. Real safety can only come from within - through cultural change and comprehensive reform of banking practices.

    2. Obviously you are right to say that “the more rules there are, the easier it is to find loopholes..”. In fact two Dallas Fed economists argued in the Wall Street Journal a day or two ago that Dodd-Frank is near useless because of its complexity. See:

      However there is an EXTREMELY simple rule that can replace the hundreds of thousands of words that make up Dodd –Frank, Basel III, and Vickers. The rule is: “if you want your bank to lend on or invest your money, you carry the can if it all goes belly up”. That way, banks as such cannot fail, thus there is no taxpayer exposure. As to money that depositors want to be 100% safe, the money just ain’t invested or loaned on: it’s deposited at the central bank, or in the case of the UK, put into National Savings. Again, there is no taxpayer exposure.

      That system is full reserve banking. Full reserve is to economics as the theory of relativity is to physics: both ideas are beautifully simple and they solve numerous problems.

  2. Spot on.

    I find the Goverment behaviour irresponsible while i agree with the Parliamentary committee.

    The level of subsidy that the goverment is pushing towards the housing market is very high and should be stopped.
    NewBuy, FLS and all the similar proposals from the goverment are only inflating the over-inflated housing bubble.

    I dont want to pay for these as a taxpayer.

  3. Please expound upon "managing risks."

  4. I agree with your main points (that the problem is demand, and the policy contradictions re more/less risk), but some nitpicking:

    Insured savings (deposits) are synthetic gilts (synthetic CB reserves) really, so they should return the same. All kind of things break down (monetary policy transmission for instance) when you have different wrappings of the same asset behave differently. Hard to explain to retail though... A good first step would be to make deposits above the insured limit illegal, so max 85K per customer per group in savings accounts, and if someone wants to have uninsured savings with their bank, they can buy bonds of said bank in the brokerage section of their account, and if they want insured, they can buy gilts (that's your voluntary insurance scheme, it's there already, no need to duplicate it).

    "Lending only to the best risks does not make for good returns on savings." I think that's just a reiteration of the empirically incorrect "high risk high return" meme. Over the cycle all risk categories will tend to return the same net of losses: you're not going to get paid for risk that can be mutualised (it follows from the scarcity of free money theorem).

    Also what proportion of bank asset does SME lending form? I bet it's tiny, so even if you say doubled it it would barely move the needle re the banks' risk (or return if it did return more).

    One way the government could help SMEs without the banks getting involved is to promote more equity via the tax treatment, to encourage savers to fund SMEs, or even direct equity investment from the state in SMEs. Hard to do though: VCT/EIS schemes do the tax thing and seem a mixed blessing, direct government investment is a minefield, but then so are loan guarantee schemes.

    1. cig

      I agree generally, though I too can do some nit-picking...

      - there is a huge problem with the £85k limit for current accounts. Current accounts are simply a form of deposit account from which banks will make payments and which may have a revolving credit facility attached (overdraft). Large payments such as property sales go through current accounts. Unless you distinguish between current accounts and other forms of deposit account, your £85k legal limit would make receiving payment for a house sale impossible. Not everyone immediately puts those funds into buying another house, you know.

      - I agree that the higher price of higher risks is simply a representation of expected losses across the totality of loans in that asset class. But in a given portfolio, there is a possibility of higher returns because losses may not be as large as the price suggests. Conversely, they could be larger of course - that is the risk you take. If you had no prospect of making higher returns (a gamble, if you like) there would be no point in taking any risk, would there?

      - I personally think the preferential tax treatment of debt should be ended. I don't see any real objections to this.

  5. Tweet:
    Frances_Coppola The Cyprus bail-out: Unfair, short-sighted and self-defeating | The Economist

    Recent History on the Matter:

    Was this a False (Insider?) Prediction that possibly Cost People Money According Following Link?:

    "The government has managed to avert a haircut on bank deposits and the issue is no more on the table, according to a close associate of President Anastasiades.

    DISY part deputy chairman Averof Neophytou said that the government's efforts are now centred on keeping the corportate tax at its present 10 percent level."

    Does This Fit?:

    "Mr. Averof  Neophytou, Deputy President of DISY has supported the campaign to save the Cyprus Grass Snake..."

    Does he have a sense of humor?

    1. Thanks. I'm going to re-post your comment on my new post about Cyprus, so that others can read it.

  6. - agreed, transaction float is a practical problem for an 85K limit, but probably fixable, e.g. by mandating the moving of any overnight (or weekend) surplus to short gilts automatically.

    - yes higher returns are numerically possible, it's just that you don't see them in the data empirically. And it's true of almost all asset classes, from loans to options through horse racing, one of the few reliable results in finance! Then, why take risk? When 2 products are interchangeable the demand will split among them (two similar restaurants opposite each other on the same street will split the custom). So even if investors have no return-based reason to choose between a portfolio of large cap loans or one of SME loans, or stocks, or whatever else mid-risk thing where the specific risk can be diversified away, they have to put their money somewhere and will spread around according to non-risk randomness, like the punters and the restaurants. There does seem to be a premium above totally risk-free, so you do usually get compensated for making the jump from gilts to anything else, but that's all, apart from returns collapsing again with extreme risk, e.g. lottery tickets or AIM mining stocks.

    1. cig

      The implications of your comments about risk and return are important. I'm wary of fallacy of composition though. There is an element of gambling in the decision to invest in higher-risk stocks, which suggests that although the data imply there is no risk premium overall, the distribution of returns is sufficiently uneven to create a reasonable chance of doing better than the price would suggest. (Or doing worse, of course.) Although this could all be in investors' minds.