Mortgages are dangerous beasts

Barclays is complaining. The Prudential Regulation Authority (PRA) is proposing to test the bank's ability to comply with a proposed new regulatory target - the leverage ratio.

In corporate finance, the leverage ratio is the ratio of equity to total debt. But in the banking world, the leverage ratio is the ratio of equity to total assets. It is in theory a simple measure, although different accounting standards do make a difference to its calculation - for example, US GAAP and IFRS netting rules for derivative exposures are different, which leads Simon Johnson to make the mistake of claiming that Deutsche Bank's leverage ratio is too low relative to US banks, when actually under US GAAP it is rather good....However, I digress. The capital ratio, by contrast, is the ratio of equity to risk weighted assets. Risk weighting reduces the value of a particular asset or asset class in the capital ratio denominator: unsecured risky loans are weighted at 100% (i.e. the denominator includes the full value of the loan), but other loans are weighted at various percentages depending on the creditworthiness of the borrower, the value of any collateral and various other risk measures. Obviously, the lower the risk weighting, the lower the amount of capital needed to meet the capital ratio target set by regulators.

Barclays knew it couldn't meet the new leverage ratio target of 3% before 2015. So the bank's CEO, Antony Jenkins, warned that if the bank had to meet that target early, it might have to cut lending to UK households and businesses. This was dynamite. As the Sunday Times says (paywall), the PRA is incensed. A lovely spat is brewing between Jenkins and the PRA's head, Andrew Bailey, who has made it clear that cutting lending to the UK economy is unacceptable. But Barclays is by no means the only bank that is complaining about the new leverage ratio.

The biggest complaints have come from building societies and other lenders who specialise in mortgages. They argue that the leverage ratio penalises them because it fails to take into account the "lower risk" of mortgage lending. Nationwide, the UK's biggest building society and fifth largest lender, says it will have to find at least an additional £1bn of capital, which won't be easy because as a mutual it can't easily tap the financial markets for new equity.

Frankly I am amazed that anyone can seriously suggest that mortgage lending is "lower risk". Charles Goodhart recently commented that there have been four UK banking crises in his lifetime, of which three were caused by property market collapses. When house prices fall - which in the UK they do about every 20 years - even supposedly safe mortgages become risky, and risky ones become toxic.  And commercial property is even more vulnerable. Collapsing commercial property values was one of the biggest causes of UK bank insolvency in the financial crisis, and several of the UK banks still have significant amounts of impaired commercial property loans sitting on their balance sheets.

The idea that mortgage lending is "lower risk" comes from the fact that it is secured on property. Mortgages don't usually cover the entire cost of the property at the time of purchase. What is known as the "loan-to-value" (LTV) ratio is the ratio of the amount of the loan to the value of the property, and in a good quality mortgage it will be not more than 80%. The remainder is covered by the borrower from own funds and/or unsecured borrowings, often from family and friends. In financespeak, the property is the "collateral" against the loan, and the difference between the amount of the loan and the value of the property is the "haircut". In the event of the borrower defaulting on the loan, the bank can seize the property and sell it to recover the loan amount. Therefore the risk that the bank will lose money on the mortgage is regarded as low.

As I explained above, capital adequacy rules take into account lending risk. In the case of good quality mortgages, risk of default is negligible because of the value of the collateral. I should make it clear that collateral is not bank "capital" - it is the borrower's capital. The bank doesn't own the property, the borrower does, even though the borrower has bought it with the bank's money. But because the loan is secured on the borrower's capital, the bank doesn't need much capital of its own to support it. That's the traditional view and the justification for low risk weighting of mortgages in regulatory capital rules.

But mortgage risk rises spectacularly when there are housing bubbles - as there have been in the run-up to EVERY property price collapse in the last 50 years, not just 2007-8. Encouraged by the prospect of high returns and blinded by rising property prices, banks water down their lending standards. They lend larger proportions of the value of the property, even lending over 100%*. And they lend to poorer credit risks - people with no steady income, people who already have high debts, the recently self-employed, people who don't want to disclose anything about their financial circumstances. Capital adequacy rules do adjust to take account of increased credit risk, but in a rising property market this is still offset by the value of the property. Basically, as long as property prices are expected to rise, the view remains that the bank can't lose and therefore doesn't need much capital of its own to support even risky mortgages. In the run-up to the 2007-8 financial crisis, many banks' and building societies' capital margins were paper-thin.

When property prices fall, the "haircuts" on real estate collateral shrink. Obviously, for good quality mortgages this is annoying but not disastrous: a 75% LTV might rise to 90% but it isn't going to be wiped out. But for higher-risk mortgages, the effect on the bank's balance sheet can be terrible. If property values fall by 11%, all mortgages with LTVs of 90% or more are underwater (the loan amount is more than the value of the property). If a proportion of these borrowers then default on their loans, the bank takes losses. If the bank's own capital is insufficient to absorb those losses, then its creditors are at risk and it is insolvent. This is why apparently sound banks can suddenly become insolvent when asset prices are falling. I've shown this here with mortgages, but the same applies to other forms of lending and purchased securities, including the infamous CDOs whose value crashed to almost zero in the financial crisis.

What I describe above - property market collapse leading to widespread bank insolvency - has happened three times in the last 50 years in the UK: in 2007-8, 1990-92 and 1973-5. And it has now happened in the US for the first time in 70 years, with a much bigger correction than the UK experienced. Each time it was preceded by bubble lending and increasing risk in mortgage books. Each time, banks failed and were bailed out - in the first two crises (especially the 1973 crash) they were bailed out by the central bank, and in the most recent crisis they were bailed out by the Government. And each time, bank capital calculated using risk weighting rules failed to ensure that banks had enough capital to withstand losses on mortgages - including commercial property and construction loans.

This is why, in my view, the leverage ratio is essential. Risk weightings are not an adequate measure of risk when asset prices are falling, so reliance on risk weighted capital ratios can result in banks having insufficient capital to support losses. And furthermore, I think it is essential that capital rules are NOT watered down for supposedly "safe" specialist mortgage lenders. There is nothing "safe" about a specialist lender that is lending 95-125% LTV to poor credit risks - as Britannia Building Society and Northern Rock were both doing. Specialist lenders should have to comply with an unweighted leverage ratio just as universal banks will have to.

There is still a role for risk weighted measures, however. The problem with a leverage ratio is that it can encourage high-risk lending. After all, returns are better when risks are higher, and if regulatory capital rules ignore risks, banks might as well go for the really high-risk, high-return stuff. This can make their balance sheets more unstable and increase the risk that they will actually fail, which is not really what we want. So in my view we need both risk weighted capital ratio and leverage ratio. They do different jobs.

And there is one other matter too. Bank lending is fundamentally pro-cyclical: higher returns and (apparently) lower risks encourage excessive lending when times are good, but when times are bad low returns and high risks (especially if asset prices are or have been falling) discourage lending. Therefore macro-prudential regulation needs to be counter-cyclical. There has been some discussion of adjusting capital requirements according to the business cycle - relaxing capital adequacy rules in downturns to encourage lending, and tightening them in booms to discourage excessive lending. This is a start. But it is not enough. Lending standards are always watered down in the run-up to a crisis. Macroprudential regulation should be looking at asset quality, not just at capital levels.

And we also need to reconsider the effect of interest rate policy in booms and recessions. Standard macroeconomic interest rate policy cuts rates in recessions and raises them in booms. This affects demand for lending, because borrowers don't want to borrow at high rates but are often happy to borrow at low rates. But the converse is true for banks. Low interest rates discourage lending: high interest rates encourage lending. There is clearly a conflict of interest here which means that relying on interest rates alone to control bank lending is doomed to fail. I don't have a simple answer to this problem, but we need to look at ways of influencing bank behaviour that don't involve clobbering their margins when they are already damaged, or encouraging them to lend far too much at far too high a risk when times are good. And that involves building macroeconomic policy models that include banks as active participants rather than passive intermediaries, so that the effect of bank behaviour can be factored into forecast effects of macroeconomic policy. Unless we do that, we are going to continue to get macroeconomic policy horribly wrong, encouraging banks to behave badly and cause intermittent crises at terrible cost to our economy. As we have done persistently over the last 50 years.

Related links:

Bank lashes Barclays for loans threat - Sunday Times (paywall)
Risk insight: The Weakness in US GAAP Netting Assumptions - Nicholas Dunbar (Bloomberg Brief) (h/t Dan Davies)
British Banks' Comedy of Terrors - Simon Johnson (ProSyn)
Deutsche Bank's capital raising highlights leverage ratio bite - Euromoney


* High LTV mortgages are certainly not a recent phenomenon. My first mortgage in 1988 was 100% LTV. The property market crashed less than two years later and I was "underwater" (in negative equity, as we say in the UK) for the next five years.

Comments

  1. “…macro-prudential regulation needs to be counter-cyclical”. I have big doubts as to how realistic that is: the authorities just didn’t know they were in the middle of a serious house price bubble prior to the crunch. And in 1928 very few people realised there was a stock market bubble.

    “we need to look at ways of influencing bank behaviour that don't involve clobbering their margins..” I take it the “clobbering” you are referring to consists of imposing safer leverage. However that argument falls foul of Miller and Modigliani. They pointed out that ultra conservative leverage ratios do not increase banks’ funding costs. Reason is that the costs of bearing the risk of a given bank are a given. Thus the number of shareholders or pounds worth of shares between which that risk is shared has no effect on the total cost of bearing said risk.

    Mervyn King actually pointed out at a recent House of Commons Banking Standards committee hearing that depositors at mutual building societies are in effect shareholders. I.e. the leverage ratio at those building societies is 1:1, which is far cry from the 33:1 that currently prevails at Lloyds, Barclays, etc. But amazingly, those mutuals compete with Lloyds, Barclays, etc.

    One can criticism M.King’s point. E.g. one can argue that it’s actually taxpayers who bear the risk at mutuals. But it’s a point that is nevertheless food for thought.

    ReplyDelete
    Replies
    1. Ralph,

      No, I don't mean "safer leverage". You have picked that phrase out of a paragraph which is about interest rates. Low interest rates squeeze bank profit margins, which is the last thing damaged banks that are short of capital need. Perhaps you would like to rethink your comment, which completely misses the point.

      On countercyclical macroprudential policy - the Bank of England warned about the danger of an unsustainable credit bubble in domestic mortgage lending in 2003 and repeatedly thereafter. Clearly it did see the collapse coming. But it was powerless to do anything about it from a macroprudential point of view.

      Delete
  2. I was wondering if it was considered possible to not have interest rates set by appointed committees and politicians [Base Rate]. Or committees of banks [LIBOR]. The use of rates in recent years seems to have been immensely damaging to all bar an elite and done nothing/little/made them worse to improve the banks but has I believe drained to an extent pensions and savings.

    ReplyDelete
  3. Hi Frances,
    thanks for this post.
    As much as I agree with the need for an unweighted "leverage ratio" as the ultimate backstop for Balance sheet elephantiasis and especially as a safeguard against "capital arbitrage", which the banks have become way too good at for their own good, I have some reservations on the riskiness of mortgages.
    While every financial assets will bear some risk, there is quite a bit of evidence that RESIDENTIAL mortgages, which make up a substantial part of the mortgage portfolios, are indeed safer than other forms of lending.
    I'll refer to Spain as an example, which you will agree is a typical pricked real estate bubble. Down there, most of the banks' losses have stemmed from loans to real estate developers and construction companies, not households mortgages. Indeed, the current NPL ratio (ie several years into the crisis and with house prices falling by more than 30% in real terms) for households mortgages is still below 5%. For comparisons, NPLs to non profit sector and industry (ex construction) is high single digits, hotels NPLs approach the 20% mark etc (for the details, come visit me @ http://euro-banks.com/spain-non-performing-loans-are-still-increasing-and-not-just-in-real-estate/#more-431). I remember having seen similar ones from Ireland, (although I never bothered to blog about them because not many people care about Irish banks as investments anymore ;)

    On a separate note, I just wanted to point out that both the backstop leverage ratio and the countercyclical capital requirement are part of the Basel 3 package, and are indeed going to happen, although maybe with some delay. In particular, under Basel 3, national regulators will have the power to toy with a 2.5% extra capital requirement as a way to try and take the punch bowl away when private credit creation goes too fast and vice versa allow banks lower capital when they want to stimulate growth (lower capital requirement means higher RoE for banks, hence it is a great investment to extend lending to borrowers which would be unprofitable clients with the higher capital requirement).
    I have argued in the past (in my previous life as a stockbroker) that this is the most interesting bit of the Basel 3 package as far as the Eurozone is concerned as it gives back some monetary control to national central banks.

    Apologies for the long comment
    Marco


    ReplyDelete
    Replies
    1. Hi Marco,

      In the US and UK (among others), residential mortgages are anything but safe. The UK crises I mentioned were all triggered by housing market collapses, not construction bubbles (we don't have those, UK is supply-constrained). Residential mortgages were the cause of the 2007 sub-prime crisis and, ultimately, the financial crisis itself, since that was largely made up of runs on various types of residential mortgage-backed security. Yes, in Spain the problem was commercial property construction lending - but in Ireland and the US, construction bubbles were mainly residential property. Spain also has with-recourse mortgages, which make it far less likely that borrowers will default on loans because they remain liable for the payments even if the property is seized.

      There is zero evidence that nortgage lending is safer than other types of lending in countries with high home ownership levels and mortgages without recourse. On the contrary, the evidence is that residential mortgages are responsible for more bank failures than any other type of lending.

      I know that Basel III includes the leverage ratio - though UK regulators are calculating it differently - and countercyclical capital buffer. I'm saying it isn't enough. Regulators must concern themselves with asset quality: poor asset quality and bad management are the two biggest causes of bank failure, by a LONG way. And regulators really must ensure that lenders don't water down lending standards in booms.

      Delete
  4. As I've said before : property bubbles cause financial crises and so rather than armour plating the banks with high reserve ratios , it would be easier to stop property bubbling with a simple JS Mill Land Value Tax that cuts in when land values rise( and stops when the land price rise stops).The banks should be lobbying for Mill's LVT instead of staging a lending go-slow and bickering. The Mill tax is a very old economic solution but the relationship between land values and finance has been known about for a very long time : Frederick Harrison is credited with having predicted the Credit Crunch to the year and month using an eighteen year cycle for property boom/bust utilised by adherents of Henry George.

    ReplyDelete
    Replies
    1. I do agree that more countercyclical measures are needed to dampen the procyclicality of bank lending. They could include taxation, although we should be aware that it is ALL forms of bank lending that are procyclical, not just property lending. One of the most dangerously procyclical forms of lending is commercial lending - especially to SMEs, because they are also procyclical in nature so the effect is amplified (bigger booms, worse busts). LVT would do nothing to dampen that.

      I've heard of Harrison's work, but I didn't know he was using an 18-year cycle. That fits nicely with what I observe, which is that in a reasonably free mortgage market like the UK's the period is about 20 years.

      We do have to be somewhat careful about dampening cyclicality. I'm of the opinion that the reason the US's housing crash was so bad is that the market had been deliberately prevented from adjusting for 70 years. We do need market corrections from time to time, and I would suggest that frequent and small is better than infrequent and catastrophic.

      Delete
    2. The UK housing market should have re-set after 2008 but did not do so very thoroughly and is now being deliberately reflated by the Coalition. The reasons are not economic but political in the worst sense: since the abolition in 1964 of the old Schedule A on residential property which took house price rises straight out of Income Tax (no kidding), a bloc of UK voters has been formed amongst homeowners (or mortgage payers) which appears to determine elections. Politicians are scared stiff of hinting at anything which would curtail their soi disant "right" to see house prices go up. Believe me: I once had to lobby my Labour MP to join a projected IPPG on LVT: I was shocked at the violence of her language!
      As regards the cyclicality LVT does n't touch, especially commercial lending , you have to remember that land taxers of all stripes (there are different versions, principally those that tax any property price" gain from here on"[Mill] and those bent on "socialising the full rental value of land" [Henry George] ),all believe that high rents and mortgage payments are going to depress spending on the High Street with obvious commercial ill effects. Descriptive quotes in previous sentence from Martin Wolf's "Why we must halt the land cycle." FT 8 July 2010 (Wolf is in the "from here on" schools as am I)

      There is plenty about Frederick Harrison's eighteen year cycle on the Net, not least his Wikipedia entry which has an interesting footnote ( No 7) which accesses Dirk Bezemer's paper "No one saw this coming"(2009) which has a handy chart on page 9 listing those who warned of the crash .As well as Harrison there is also Michael Hudson who published in 2010" Land Value Taxes to Restructure Economies".
      Land value based Economics is not the orphaned ,fugitive tradition it once was with support now from the IMF and the Economist (this week).

      Delete
  5. There is a simple solution to most of the above problems. It’s the solution advocated by Laurence Kotlikoff and by John Cochrane in this recent Wall Street Journal article (both of them are economics profs. for what that’s worth):

    http://online.wsj.com/article/SB10001424127887324412604578513143034934554.html?mod=WSJ_Opinion_LEFTTopOpinion#articleTabs%3Darticle

    Plus their solution is very close to the solution advocated by Positive Money, Prof Richard Werner and the New Economics Foundation.

    The solution is to impose a 1:1 leverage ratio. I.e. those who want their money loaned on or invested by their bank bear the full cost when it goes wrong. Plus those depositors have a choice as to what is done with their money (i.e. they can have it put into ultra-conservative mortgages, NINJA mortgages, commercial mortgages, etc etc). In contrast, depositors who want 100% safety have their money taxpayer backed, but nothing is done with the money: e.g. it could be lodged at the central bank.

    The advantages of that are as follows.

    1. Banks as such cannot suddenly go bust: though they can shrink to nothing over a period of time. So no more Alistair Darlings being rung up in the middle of the night by the head of RBS to say RBS will be bust in 48 hours unless it gets £10bn.

    2. Ergo no more credit crunches.

    3. No more need for bank subsidies (TBTF subsidies included).

    4. No need for ringfencing: the same rules apply to investment banks as to retail banks.

    5. No need to break up large banks.

    6. Banks would be forced to charge realistic rates to different types of borrower. E.g. loans to mortgagors who have say at least a 30% equity stake would very near 100% safe. Thus they’d be able to borrow at rates no higher than they currently pay, and perhaps even lower rates.

    If the above isn’t the banking equivalent of E=MC2, I don’t know what is.



    ReplyDelete
    Replies
    1. Ralph,

      There is already a rate differential for funding supposedly safer forms of lending. But when the property market collapses - as it does about every 20 years in the UK - one of the first things that happens is that bank funding costs rise as their asset values fall. Safe mortgages become risky. Less safe mortgages become toxic. The point is that forms of lending believed to be "safe" ARE NOT SAFE when there are exogenous financial shocks. That is actually what went wrong in the financial crisis - assets that were believed to be safe because they were backed by residential mortgages turned out to be anything but. I did explain this in the post.

      You do realise of course that even if banks survive a property market collapse, afterwards all existing lending is much riskier? There are no 70% LTV loans when the property market has collapsed.

      Delete
  6. So you seem to imply that the causality goes from asset price bubble to financial crises, but this is not right, you get a financial crisis only when the crashing asset price bubble leads to high unemployment/recession. This is not a given, and the central banks can maintain spending, as in the DotCom boom and crash, which avoids unemployment and therefore avoids large default rates - which are the real trigger of a financial crises.

    IT is set up to maintain the future value of past claims on consumption, which means that when there are `unexpected' losses, they are borne by labour, rather than by capital. If a CB maintains spending, and hence, employment, losses of output are borne by owners of capital (savers) through higher inflation. This would make the financial system more stable because there would not be the episodes of mass unemployment that trigger high default rates. Hybrid targets are available - e.g. set a trend for inflation and NGDP and say that you will bring them back to trend symmetrically, e.g. if NGDP is below trendn then inflation should be above trend and vice versa.

    House price bubbles are the easiest bubbles to spot, because there is a pretty clear historical tendency for people to want to spend a certain fraction of their income on rent/housing, so when the housing market diverges from this trend it tends to return. However, this means that leverage rations should be much higher than average right now, that we are at the trough in the House price cycle. Now is that time when 100% LTV loans will turn out well. :)

    Its common knowledge that as soon as a bubble has burst people start worrying about the next bubble, when in fact, an asset price crash is never followed by another crash, until people have forgotten about the first crash. :)

    ReplyDelete
    Replies
    1. You have the direction of causation wrong. Financial crises precede high unemployment & recession, not the other way round. The events of 2007-8 demonstrate that clearly: banks started to fail due to collapsing asset values in mid-2007, with of course the main crash coming in September-October 2008, and the recession followed in 2009. If you look at the course of the Great Depression it shows the same pattern - bank distress and failure preceding economic recession.

      I think you confuse banking crises and stock market crashes. They are not the same thing. The Dot Com burst was a stock market crash: there was a banking crisis at around the same time, but it was caused by the LTCM failure. Stock market crashes don't necessarily have any great effect on the economy, but banking crises do.

      It is in theory possible to have a banking crisis that does not cause high unemployment or major recession. In fact if we manage the credit cycle properly, they should never have this effect. There will still be market corrections from time to time - we need those - but they should not have catastrophic effects on the economy.

      Delete
    2. Oh, and I suggest that the main reason why the LTCM crisis and Dot Com burst didn't have catastrophic effects on the economy was that they didn't involve residential property.

      Delete
  7. TBH, I don't see how you get from "mortgages are not safe if they are 95% loan to value" to "and therefore we need to use a leverage ratio, which specifically does not distinguish between mortgages at 20% LTV and at 105% LTV". One of the big things about Basel 2 was precisely that it began to make this distinction and it's not true that high-LTV mortgages were given low weightings - the low average weightings seen in the UK were mainly due to the fact that their average books were well-seasoned and generally below 50%.

    It's also not true in principle that "Risk weightings are not an adequate measure of risk when asset prices are falling". Risk weightings are always linked to the mark-to-market LTV on property loans and they are much more sensitive to the collateral value than a leverage ratio standard. The trouble is that they do vary procyclically - what would make sense would be to vary them according to departure from historic valuation ratios, but for some reason the supervisors have this holy fear of making forecasts.

    The UK's crises have always been associated with residential property busts, but not really with credit losses on mortgages - looking at the releases from Northern Rock Asset Management (the bad bank for NRK and B&B), it really bears out that the realised loss rates have been much more like the original planning and underwriting numbers than the more excitable marks-to-market from 2008. The trouble is that in the UK, because of the importance of the mortgage market, general liquidity and collateral values are highly correlated. It's a problem from hell that can't really be solved by short ways with ratios.

    ReplyDelete
    Replies
    1. What a very confused comment.

      Mortgages are not safe if they are 75%, 85%, 95% LTV. In fact they are not really "safe" at all - it is merely a matter of degree of risk. When asset values fall, a supposedly "safe" 75% LTV mortgage can quickly become a much riskier 90% LTV mortgage. I would have thought it was obvious, therefore, that a capital measure that weights loans according to their level of riskiness fails to protect when collateral values fall and/or defaults increase. Because the leverage ratio does not distinguish between risk levels, treating ALL loans as equally risky, it protects against supposedly "safe" loans suddenly becoming more risky. That is a no-brainer, frankly.

      Your remarks about risk weighting of high-LTV mortgages are contradictory. Firstly you say that they weren't given low ratings because of rising collateral values. Then in the next paragraph you say risk weightings are linked to the MTM LTV on property loans, which is of course defined by changes in collateral value. You can't have it both ways. Either collateral MTM does influence risk weightings, or it does not. The evidence actually supports my argument: in the run up to the financial crisis, risk weightings actually fell as loan risk rose.

      It is frankly ridiculous to separate mortgage loans from their collateral as you do in the last paragraph. Residential property busts result in greater credit losses on mortgages, because when house prices fall, defaults are not always fully covered by collateral. If banks are short of capital, then failures follow even if no more borrowers than usual have actually defaulted.

      Delete
    2. What a very confused answer. You don't understand how risk weights work as well as you think you do.

      I would have thought it was obvious, therefore, that a capital measure that weights loans according to their level of riskiness fails to protect when collateral values fall and/or defaults increase.

      A capital measure which weights loans according to their level of risk, increases the capital requirement as collateral values fall and defaults increase. As a commentator above notes, residential mortgages, even in Spain, have seen very low default and loss rates; as I mentioned myself (and you did not respond to, because it's true), the realised default rates from Northern Rock Asset Management also show that the risk weights on the majority of the loan books were, in fact, about right.

      The risk weights did not "fall as loan risk rose" and my points are not contradictory. You are confused between average and marginal risk weights. The marginal risk weight on UK mortgage lending rose slightly throughout the property boom. The reason that the reported average risk weights fell was that the loan books were seasoning (getting older) and property prices were rising, meaning that by 2008, a loan from the 2003 vintage would have seen its LTV fall from 80% to below 50%.

      Under a leverage ratio standard, of course, the incentive would be to securitise the older, seasoned and safe mortgages and load up on the new and higher-risk ones, which is what the US banking system actually did.

      You make two other points which are also really badly wrong about how banking works.

      1. In terms of the loss-given-default, a bank loses exactly the same amount on repossession of a 90% LTV mortgages as a 75% LTV mortgage - nothing. The loss is taken if the mark-to-market LTV is greater than 100%. You claim that a performing loan gets "more risky" with every move in the collateral value but this is not true, and it is not how the accounting standard works. Since you apparently regard the mathematically equivalent claim that a loan gets safer when house prices rise as ridiculous, I don't see how you can regard this as a "no brainer" unless you haven't really been thinking about it.

      2. It's just not true that the greater losses on mortgages in recessions are mainly caused by lower realisations from sales of repossessed property. Anyone who's analysed a mortgage bank knows this. They're caused by higher default rates, and the inability to refinance. Since the total fluctuation peak-to-trough in mortgage provisions is of the order 1-2% in mainstream UK mortgage lending, it's obvious that it can't be mainly driven by the large fluctuations in property values.

      Delete
    3. You have completely missed the point. You may know how risk weightings work, but you haven't understood either my post or my comments.

      Yes, capital requirements increase as collateral values fall. But the banks concerned don't have any more capital. Increasing the requirement doesn't increase the available capital. Therefore, when collateral values fall, banks can end up failing to meet regulatory capital requirements that they previously met.

      I do indeed say that risk weightings fell as risky lending increased in the run up to the financial crisis. They did. And I'm certainly not confused about the difference between marginal and average risk weightings. Yes, the marginal weight increased slightly - but the average weight fell due to the unsustainable house price increase, masking the increase in higher-risk lending and giving the impression that loan books were safer than they actually were. It was a self-reinforcing vicious spiral.

      I did make the point in the post that relying only on a leverage ratio encourages riskier lending, and that therefore there would still be a role for risk-weighted capital ratios. You seem to have missed that. Or perhaps you ignored it, because it didn't fit with your evident intention to take me apart and expose me as ignorant.

      You say that I made two other points that are "really badly wrong about how banking works". Sadly you have misunderstood the points I was making.

      Point 1: LTV changes. You note yourself that risk-weighted capital requirements rise as LTV falls. If, therefore, there is a house price correction that results in a 75% LTV loan becoming a 90% LTV loan, that loan has - for risk weighted capital adequacy purposes - become "riskier". If there is a further fall in house prices, that loan is more likely to end up underwater. Really that is a no-brainer. The reason why it is ridiculous for loans caught up in an unsustainable housing bubble to be regarded as "safer" is that bubbles burst. That too is a no-brainer. "Risk" and "safe" here apply to our existing measures, though, which I regard as deeply flawed: a 90% LTV mortgage at the bottom of a house price correction is actually "safer" than a 75% LTV mortgage at peak price, but it will carry a higher risk weighting. That is the third no-brainer, and a large part of the problem with risk weightings: they are inherently pro-cyclical.

      Point 2: your remark is simply wrong and frankly suggests that you don't understand secured lending. For all secured lending - not just mortgages - the risk to the bank is not borrower default but falling collateral value. If a mortgagee defaults, the recourse for the bank is to repossess the property. If the realisable value of the property is sufficient to cover the loan, the bank has not suffered a capital loss, although it does of course lose the NPV of future income streams. However, if the realisable value of the property is insufficient to cover the loan, the bank suffers a capital loss. Increasing defaults themselves do not increase the risk of capital losses unless collateral values also fall. Put simply, risk of default is only risk of capital loss if collateral is insufficient. However, as most recessions are associated with house price falls, the two tend to occur together - rising defaults AND falling collateral values. Perhaps that's why you keep talking about default when I am talking about falling collateral values.

      I should add that as I spent quite a lot of my career doing capital adequacy and risk management for various banks, I do actually know how risk weightings work, and how banks work. I would advise you to read what I write a little more carefully before accusing me of ignorance.

      Delete
    4. Just to be clear; I have not accused you of "ignorance". I have said that you don't appear to know as much about risk weightings as you think you do, and that you aren't thinking clearly about what you are saying about risk. (You've tacitly admitted the second point by saying that "risky" was meant to mean "with a higher RWA risk weight").

      On the other hand, you've called me "very confused" and "ridiculous". Let's be clear here about who's using the offensive language.

      On point 1, it's not true that risk weightings are "inherently pro-cyclical". The current Basel rules happen to be somewhat pro-cyclical, but it would be easy to draft risk weightings which (as I suggested in my original comment) " vary them according to departure from historic valuation ratios". A leverage ratio does not have this property and substitutes procyclicality for an incentive to increase risk at every stage in the cycle.

      On point 2, thank you ever so much for the dictionary definition of collateralised lender, but I was actually making a point about the empirical facts about mortgage banking in the UK and elsewhere. Losses due to insufficient realisation of collateral are not an important driver of the cyclical variability of loan impairments. Even at the depths of recessions, repossession rates are not high enough for this to be the case. Most of the cyclical fluctuation is driven by the loss of future income and administrative costs, which is why it is driven by the default rate and not by collateral values. This was even true of the 1990s house price slump.

      Delete
    5. Your first comment was confused, because you didn't bother to read my post properly before weighing in with your comments. You said I don't understand risk weightings and I am very wrong about how banks work. That adds up to accusing me of ignorance. And now you add to that a claim that I am not "thinking clearly" because YOU have misunderstood something I wrote. Priceless.

      Delete
    6. My first comment was in fact correct; it was only later in comments that you decided that "risky" meant "having a higher risk weighting". Reading the post a thousand times wouldn't have helped me know that you hadn't written it properly. This compares to your own decision to accuse me of being "contradictory" based on your own failure to read the word "average".

      I am not going to gainsay your claim to understand risk-weightings and will accept that you do. But the fact is that you said something untrue (that risk weightings on new lending fell during the boom) and then doubled up on the assertion by accusing me of being confused and contradictory when I pointed out the opposite was the case. So I'm not going to apologise for that as you are clearly the author of your own misfortune.

      The statement that you're very wrong about how banks work is just a fact. Changes in realisation values of repossessed collateral aren't the main, or even a particularly important driver of loan losses in residential mortgage business.

      Delete
    7. 1) No, I always meant that. Evidently I didn't express what I meant clearly enough for you to understand - but that's not the same as "not thinking clearly".

      I did read the word "average" and I knew what it meant.

      2) I did not say that risk weightings on new mortgage lending fell. I said that risk weightings fell as loan risk increased in the run up to the financial crisis - which is true - and I then in a comment to you explained how that happened. I also reminded you that risk weightings don't just apply to residential mortgages.

      3)Changes in collateral values together with default risk are the main driver of CAPITAL losses on loans, which is all I was talking about. I know that NPV losses from loan default can make a large hole in the P&L, and there may be a balance sheet impact if there is insufficient provisioning, but as capital ratios are calculated in relation to asset values not income streams it was not relevant to this particular post.

      Delete
  8. All excellent points, but is it really the banks fault in the case of the UK? Take for example the following statements made by David Miles at the Bank of England:

    Statement One: "The worst of the housing market crash is now over" July 2009

    Statement Two: "House prices to rise for years" March 2012

    Statement Three: "It probably never made sense for there to be 100pc mortgages. There may be no price at which it makes commercial sense for such a loan to be available." March 2012

    If those at the top who ultimately advise on regulation can't get it right - how can the banks be expected to?

    More on Miles statements here: http://www.ipinglobal.com/ipin-live/406319/solid-proof--the-banks-dont-understand-the-housing-market

    ReplyDelete
    Replies
    1. The Bank of England warned in 2003 about the unsustainable growth of credit, particularly for mortgages. And it repeated its warnings several times over the next four years. The banks took no notice.

      It absolutely is the banks' fault that they increased their leverage to dangerous levels in the expectation that house prices would continue to rise. There are enough people working at senior levels in banking who are old enough to remember the late 1980s housing bubble and 1990 crash. They should have learned from their history. The trouble is they thought "this time was different": admittedly this view was encouraged by foolish statements from politicians, and regulators weren't exactly enforcing discipline, but it really isn't reasonable to exonerate senior bankers. They should have known better.

      Delete
    2. This is also not right. Even HBOS, the most irresponsible of the UK banks, did not suffer particularly serious problems in its Retail division, even though it rapidly grew the high-LTV and buy-to-let "specialist" segments of the loan book. The TSC report concluded:

      "Although experiencing some deterioration as a result of the crisis, particularly in respect of non-standard mortgages, HBOS's Retail impairments were substantially less than either the Corporate or International Divisions incurred and were not a material factor in the failure of HBOS. We estimate that total Retail impairments would have been some £7 billion between 2008 and 2011. The Division generated profits before impairments of £3.5 billion in 2008.[348] Even allowing for significant pressure on this figure in subsequent years and for charges against the mis-selling of payment protection insurance, the Division's pre-impairment profits would have allowed the Group to absorb the likely level of impairments and still generate profits."

      http://www.publications.parliament.uk/pa/jt201213/jtselect/jtpcbs/144/14415.htm

      The UK banking crisis had very little to do with mortgage impairments. It was a combination of bad liability management in the context of a prime mortgage lending operation (Northern Rock), massive corporate lending irresponsibility (HBOS) and acquisition megalomania (RBS). It happened at the same time as the house price crash, but these were two phenomena with a common cause (the global liquidity squeeze and the UK recession), not a case of one causing the other.

      There was a country where mortgage lending defaults caused the crisis, but it wasn't the UK. It was the USA. Where banks are regulated on the basis of a leverage ratio and where Basel 2 was never implemented.

      Delete
    3. I did not say that HBOS failed because of risky mortgages. However, its Retail division was far from healthy - after all, your quotation indicates that Retail was facing significant impairments. It's just that other divisions were even worse. I have actually read the HBOS report from end to end and am very familiar with its findings.

      Nor did I say that the UK's 2007-8 banking crisis was due to residential mortgages. Indeed I specifically said that commercial real estate was a much bigger problem, along with commercial lending.

      You are incorrect that "acquisition megalomania" was the main problem in RBS: that was the proximate cause, but the real problem was appallingly risky lending across all divisions. RBS would have failed even without ABN AMRO.

      You are also incorrect that Northern Rock's problem was just bad liability management: Northern Rock was insolvent, not just illiquid. Bad liability management doesn't cause insolvency - it causes liquidity problems. Bad asset management causes insolvency.

      You also ignore the effects on the building society sector, which I discuss in the post. However, I was not just talking about this crisis - indeed the point I was making was that mortgage lending periodically causes banking crises. And I was not just talking about the UK, although obviously I was using UK banking examples. I was talking generally - and I specifically mentioned the US.

      Your reading of my post has been somewhat selective, to say the least.


      Delete
    4. I don't think it's been selective at all! You wrote:

      It absolutely is the banks' fault that they increased their leverage to dangerous levels in the expectation that house prices would continue to rise. There are enough people working at senior levels in banking who are old enough to remember the late 1980s housing bubble and 1990 crash. They should have learned from their history. The trouble is they thought "this time was different": admittedly this view was encouraged by foolish statements from politicians, and regulators weren't exactly enforcing discipline, but it really isn't reasonable to exonerate senior bankers. They should have known better.

      How can this be read other than as saying that the crisis was made worse due to, and bankers should be blamed for, large realised credit losses on UK mortgage business? My point is that there have been no such large realised losses on UK mortgages.

      Bad liability management doesn't cause insolvency - it causes liquidity problems. Bad asset management causes insolvency.

      This isn't true either as a general statement or specifically about Northern Rock. The Savings & Loans, for example, went insolvent because they had not managed their liabilities properly (funding 30 year fixed rate mortgages out of short term deposits). Lots of S&Ls went bust without any major credit problems.

      Delete
    5. 1) No residential mortgage failures. Really? What about Bradford and Bingley? Dunfermline Building Society? Kent Reliance? Please don't tell me all their problems stemmed from commercial lending or from wholesale funding. And what about Britannia - which not only has toxic commercial loans but seriously impaired residential mortgages?

      Impaired residential mortgages featured in the internal "bad banks" of RBS, Lloyds and the Co-Op (ex Britannia), along with commercial real estate. And s for Northern Rock - if its asset quality was so great, how come NRAM is still managing ex-Northern Rock residential mortgages transferred there when it was nationalised? I have already agreed that residential mortgages were not the main cause of the UK banking crisis. But to say they weren't involved at all is going way too far.

      RBS was involved in the US mortgage market not only as a securitiser but also, through a subsidiary, as an originator. As I said, they should have known better.

      2) You don't go insolvent when liabilities reduce relative to assets, because all that does is increase your equity. But you do go insolvent when asset values fall to the point where the available equity is insufficient to cover the losses. The structural mismatch you mention is a LIQUIDITY problem. I worked in an ALM department long enough to know the difference. I even specifically produced "Structural Liquidity" reports to show those mismatches.

      Delete
    6. I really wish you wouldn't be so aggressive, because you are once more posting things which with a bit of consideration you'd realise aren't true.

      If you're funding a fixed-rate asset (or a fixed-spread asset) out of short term liabilities, and you have a positive spread of 130bp (as, IIRC, Bradford & Bingley did), then one thing that can kill your business is if your funding spread increases by 135bp). You then go from a position of making a positive to a negative net interest income, and obviously if your portfolio has a negative spread, taken as a whole, you can never make money and are insolvent. This is actually, practically, what happened to the S&Ls

      And it is the reason why NRAM is managing Northern Rock and B&B loans. Not because they are experiencing credit losses - I gave you the link above and you can see they aren't. But because the bust banks couldn't *fund* the book on an economically viable basis. The loans (like those of Dexia Public Finance, which also has a bad bank with no credit risk these days) were "good" when they were made, and would have remained "good" if overall credit spreads hadn't risen.

      Having negative funding carry isn't a liquidity problem, it's a solvency problem. You must certainly know this if you've worked in Alco.

      Delete
    7. Aggressive? You've really been pretty discourteous to me - you didn't read my post properly, you accused me of saying things I didn't actually say, and you told me I was wrong and ignorant.

      Since you want to talk Alco stuff.....the boundary between illiquid and insolvent can be very blurred: it's largely a matter of timing and perception. Strictly speaking, a negative funding carry for a short period of time is illiquidity, not insolvency. It becomes insolvency if it can't be resolved - which is very often because asset quality is poor. The problem that central banks have is deciding when acute funding distress is a transient problem (due to market conditions, for example) and when it is actually indicative of insolvency. The Bagehot Rule is rather difficult to apply in practice! Also, illiquidity morphs into insolvency if left untreated.

      However, as I'm not writing for a professional audience in these posts, I try to keep the distinction between illiquidity and insolvency much clearer. And when I am writing about complex things such as risk weightings I have to simplify. Your comments are not helping matters.

      Delete
  9. Somewhat repetitively ,I would say that the UK political parties are locked into a Homeownerist agenda of bribing voters with rising house prices,(while real wages decline) so the political class, as embodied by successive governments, cannot be absolved of responsibility.
    Although written in a rather tabloid style by the very young Jesse Columbo The Bubble Bubble.com website is a good source. Its Emerging Markets Bubble lists housing bubbles in China, Hong Kong, Taiwan, S. Korea ,India, Singapore, Malaysia, Indonesia, Philippines, South Africa, Israel, Russia ,Serbia , Argentina, Columbia and Uruguay
    Elsewhere details are given of the usual suspects: London and UK housing bubble, Paris and France housing bubble, Germany , Canada, Australia , Belgium ,Netherlands........
    With this mountain of evidence it is incredible that the penny has not dropped.

    ReplyDelete

Post a Comment

Popular posts from this blog

Calculus for journalists

Game theory in Brexitland

Tariffs, trade and money illusion