Are Australia’s banks about to collapse?

Bank cracking photoUPDATE: In this post I repeated Business Insider’s mistake of attributing the presentation I criticise to Steve Keen. While Steve considers it an excellent presentation, he did not write it and I apologise for not confirming the source before publishing this post. I have now struck out the incorrect attributions. My criticisms of the presentation itself still hold, which is why I am leaving the post up in its edited form.

Steve Keen and his forecasts of a property market collapse have received plenty of local media coverage over the years. Now he has come to the attention of the international press as well.

In April, Keen hiked to the top of Mount Kosciuszko after losing a bet about the direction of property prices with Macquarie Bank strategist Rory Robertson. This event was enough to prompt an extensive review of Keen’s concerns in the New York Times. Curiously, Robertson himself did not receive a mention, despite winning the bet.

Now the US business site Business Insider, which has a penchant for drama, has published one of Keen’s presentations a presentation, incorrectly attributed to Keen, under the headline “Here’s What You Need To Know About The Major Property, Debt, And Banking Crisis Brewing In Australia”.

One of Keen’s central concerns is the size of private sector debt in Australia. This is a legitimate concern and should receive more focus than misguided fears about Australian government debt. However, I am far less pessimistic than Keen about the outlook for Australian property prices.

As for the Business Insider presentation, Keen takes his concerns it goes too far, to the point of unsupportable alarmism. The final slide of the presentation is evidence enough of this, not to mention being in extremely poor taste. This slide appears to have been added by Business Insider! If that is not enough to convince you, I will consider just one of the arguments offered by the anonymous author Keen.

On slide 22 of the presentation, he writes:1

Look at CBA 2009 annual report—Leverage ratio is almost 20 times (total assets of $620.4 billion against $31.4 billion of equity). Of $620.4 billion of assets, $473.7 billion are loan assets. If around 6.6% of CBA’s loans go bad (any loans not just mortgages), 100% of its shareholder equity will be wiped out!!

(the bold italics are not mine, they appear in the presentation). Here the implication is something like “6.6% is not very much. Wow! CBA could easily collapse!”. But, that line of thinking does not stand up to even moderate reflection.

Crucially, we must understand what “going bad” means for a loan. It does not mean losing everything, which is in fact very rare for most types of bank loans.

Over half of CBA’s are home loans and these are secured by the property that has been mortgaged. According to their half-year presentation2, based on current market valuations, the average loan-to-value ratio (LVR) for CBA’s portfolio is 42%. This means that, on average, the value of the property is more than twice the loan amount. This gives the bank an enormous buffer against falls in property prices. Of course, this average conceals a mix of high and very-low LVR loans. Even assuming that loan defaults occurred on a higher LVR section of the portfolio, say with an average LVR of 70%, and allowing for Keen’s oft-quoted figure of a 40% decline in house prices, CBA would still only lose 14% on their defaulting loans3. Even then, this does not take into account the fact that, like other lenders, CBA takes out mortgage insurance on loans with an LVR of more than 80%.

But we can be more conservative still. In their prudential standards, the banking regulator APRA considers a severely stressed loss rate on defaulting home loans to be 20%. To suffer actual losses of 6.6% in their mortgage portfolio, CBA would have to suffer a default rate of at least 33%! This would be astonishingly unprecedented. Currently, the number of CBA borrowers late on their mortgage payments by 90 days or more is running at around 1%. Most of these borrowers will end up getting their finances back in order, so for actual defaults to reach 33% is inconceivable. A default rate of a “mere” 2% would be extraordinary enough for CBA.

As for the rest of the $473.7 billion, it includes personal loans, credit card loans, business loans and corporate loans. The loss rates on some of these loans can be higher than for mortgage portfolios, but losing everything on every defaulting loan is still highly unlikely. So to suffer 6.6% in actual losses on these loans, defaults would have to run at a far higher rate. Furthermore, since the dire prognosis for the banks is rooted in the view that the property “bubble” is about to burst, presumably the argument would not simply be based on everything other than the home loan portfolio collapsing.

If property prices do fall sharply and our economy has another downturn, will bank earnings be affected? Of course. Are they teetering on the brink of collapse? Of course not.

1 While there is a footnote on the slide referencing this post, what is not made clear is that the whole paragraph is a direct quote rather than Keen’s own words. Presumably he agrees with it though!

2 Page 84.

3 If property prices fall to 60% of the original value, the loss on a 70% LVR loan would be (70% – 60%)/70% = 14.3%.

Possibly Related Posts (automatically generated):

34 thoughts on “Are Australia’s banks about to collapse?

  1. zebra

    Keen is losing the plot big time. Not only is his analysis on this wrong, his analysis on RSPT in yesterdays ABC was wrong as well.

    And that slide isn’t just in bad taste – it’s deeply offensive. If he gave that as part of a presentation at a company he worked for he’d be hauled up before HR. If it was for clients he’d be sacked.

    Keen has become the Jason Akermanis of Economics and Finance writers.

  2. (theOther)Bernardk

    What a tool!
    Was he moonlighting as BJoyce’s policy advisor?

    I’m 10 years out of touch with the figures; are those “loan asset” figures gross or nett? Does it include securitised, or otherwise swapped risk, or is this straight on sheet lending?

  3. Bill Mitchell

    Very nice analysis Sean. I hope Steve reads this and stops making alarmist comments that may reflect badly on heterodox economists generally.
    Perhaps you should make a bet with him that he needs to walk to Antarctica this time if none of the big four banks go bust in 2 years!

    best wishes
    bill

  4. Pingback: Following one link too few…mea culpa | Stubborn Mule

  5. Stubborn Mule Post author

    Juiced: thanks for pointing that out. I have just written a mea culpa and will not mark corrections in this post. While it is now clear to me that Steve did not write the presentation, I am still a little surprised that he describes it as an “excellent presentation”.

  6. Stubborn Mule Post author

    Bill: While the analysis still stands, the attribution does not. Perhaps I am giving heterodox economics a bad name! I hope not.

  7. Stubborn Mule Post author

    Note that the final slide, which is in such poor taste, appears to have been added by Business Insider! It was not in the original article linked to by Steve.

  8. Chris

    Your stats are totally wrong and misleading. Average LVR is meaningless. It is the distribution of LVRs that matter. EG:

    Loan 1: 95% LVR $650k loan ($684k house)
    Loan 2: 65% LVR $650k loan ($1m house)
    Average LVR: 80%

    If house prices fall by 20% under your logic the ‘loss’ should be zero. In fact the loss would be ~8%

    Your ‘analysis’ of mortgage loans vs personal and SME loans is actually not an analysis at all – I am unhappy to have wasted 30 seconds of my life reading it. Keep living in la la land for all care but you should rename your blog “the Stupid Mule” if you are proven wrong.

  9. Stubborn Mule Post author

    Chris: I do understand the importance of LVR distribution, in fact I wrote

    this average conceals a mix of high and very-low LVR loans

    It is precisely because you cannot go straight from an average LVR and an average price decline to a loss amount that I turned to APRA’s portfolio average stressed loss given default rate of 20%.

    Still, CBA’s low average LVR does mean that the portfolio has some form of buffer. To see this, imagine a very nasty LVR distribution that still has an average of 42%. Since CBA has mortgage insurance for LVRs over 80%, let’s cook up an distribution with as many 80% LVR loans as possible, and the average is pulled down by some very low LVR loans, say at 10%. A mix of 46% of loans with 80% LVR and the rest at 10% would give a portfolio average LVR of 42%. A 40% decline in prices would result in a 25% loss on any of the 80% LVR loans that default and zero on the 10% LVR loans. To suffer 6.6% losses at the portfolio level you’d have to have 58% of the 80% LVR loans default or 26% of the whole portfolio, which is an enormous and unprecedented default rate.

    You are right that I barely touched on SMEs, but the point remains that “going bad” is unlikely to mean losing 100% on all defaulting loans.

    I don’t think I am living in la la land on the basis of this analysis. Do I think that there will be a property crash in Australia? Unlikely, but possible. Do I think that if there is a crash that one of the four majors will go bankrupt? Very, very unlikely. I do think they are well-capitalised.

    On that basis, if one of the big four does go bankrupt, I will put a “Stupid Mule” banner across to the top of the blog.

    P.S. I’m sorry you feel you wasted 30 seconds reading this post, but at least it shows you are a fast reader!

  10. Stubborn Mule Post author

    Chris: checking those figures in the previous comment, they were actually under-done at the portfolio level. I have corrected the comment. With that hypothetical skewed LVR distribution, there would have to be a default rate of 26% at the portfolio level (with 58% of the high LVR loans defaulting).

  11. Chris

    I dont think property crash or banking crisis is likely, but I think it is a much more likely outcome than many people think. People underestimate the risk partly becaue of convoluted logic like yours in this post.

    If you recoognize that LVR distributions matter why on earth would you create two ‘sub groups’ where the ‘high LVR’ group is at 80% LVR? Clearly (as in the US in their crash) it is the 90-100% LVRs that cause the most loss and the most defaults (in fact when mortgages go underwater default rates increase exponentially). How many first home buyers who borrowred 95-100% of the houses value, with no savings, do you think will keep paying their mortgage if they are 20-30% underwater?

  12. Chris

    And one more thing – what do youmean by ‘default rate’ – do you mean cumulative over many years? You should, as this is how losses will accrue to the bank balance sheets.

    EG: if the LGD of defaulted mortgages (remembering that the high LVR loans will be much more prevalent amon defaulted loans) is 20%, then to get to 6% loss rate on mortgages you need 30% defaults.

    But it is not 30% defaults at any point in time or in one year, it is ovr the LIFETIME of the portfolio. If you imagine a deep 4-5 year recesssion driven by China cash, as well as housing market crash impact on employment and domestic economc activity, you would need ~6% new defaults per annum. As you say this would be much higher than historical default rates but not impossible by any means, esp when the ‘parabolic’ default tendencies of underwater mortgages are considered.

  13. Stubborn Mule Post author

    Chris: the reason I cut off the high LVR at 80% in that example was simple (and I thought I made that clear before). CBA, like the other major banks, takes out mortgage insurance on mortgages originated with LVRs higher than 80%. Since the blog post focuses on how heavily a property crash could impact on the banks, I was cooking up an exaggerated example with maximal direct impact on CBA. Of course, the other angle would be to assume that the portfolio did consist of a lot of mortgages with LVR over 80%, that these mortgages suffered losses to such an extent that all the mortgage insurers went bankrupt, leaving to the remaining losses with the bank. Of course, then you have to start making assumptions about recovery rates from insurers….

    As for the question of timing of the defaults, it would have to happen in a relatively short timeframe (i.e. not “many years”) for the bank to collapse, otherwise there would be time for the bank to improve their capital position, for example by raising additional equity. Certainly having cumulative losses over the lifetime of the portfolio would not send them to the wall.

  14. Chris

    not true stubborn mule…. why would equity investors put more equity in if they can see there is zero value in the company? Look at example of Citibank in the US…. the only ‘investor’ will be the government…

  15. Chris

    and there is no doubt that mortgage insurers will be in default if this scenario transpires…

  16. Stubborn Mule Post author

    There is a big difference between a gradually deteriorating portfolio (which is the scenario you suggested with defaults being realised over the whole life of the portfolio) and a bank sitting on enormous volumes of structured assets whose traded values have collapsed in the market. If the damage to their mortgage book was such that investors would value the bank at zero, I’d argue that it would have to be more than a drawn-out problem over the life of the portfolio.

    As for mortgage-insurers, even if they did default, it would not be before they absorb some losses, so again you have to come up with assumptions about the loss given default on the insurers. Again, it comes back to the original point that if a loan “goes bad” it does not mean that the loss incurred is 100% of the value of the loan. The whole point of the post was to criticise that implication. Of course, you can come up with a scenarios which destroy the equity of the bank, but my point is to do that your scenario has to be significantly more severe than “6.6% of the loans going bad”. Can banks default? Yes. They can and do. Are the four major banks as vulnerable as the presentation suggests? No.

  17. Marco aka Cracticus

    Hey Stubborn and Chris

    I don’t mean to be nosey, but I found your discussion quite instructive, so I would like to offer some comments.

    At one hand, I believe Chris has a point when he says your logic is convoluted, and a bit deceptive, Stubborn. In fact, ironically, you might have been a victim of it. I suspect your numerical example, for instance, has some wrong numbers. I believe that, to achieve a 42% average LVR, the low LVR loans need to rise from 10% to 30%. This, it seems to me, favours your argument.

    If I am right and you change this, you’ll find that with a 25% fall in housing prices, the portfolio losses reach 3%. With the 40% fall in prices you talk about, losses reach 12%. And this, I dare say, works against your argument.

    If you are interested, I’ll be happy to send you the worksheet (Excel, sorry!) with my calculations. Hopefully, the one with his argument torn to pieces won’t be yours truly!

    Regardless of the calculations presented (which I believe we all agree are just numerical simplified examples, in any case), on the basis of your calculations, you seem to conclude that only a sustained and deep fall in housing prices, caused by an equally prolonged period of defaults could have a significant effect on the banking industry. On this particular, I believe both yourself and Chris seem to be more or less in tune.

    Here I differ from both of you; and I will allow myself to repeat a comment I made about the “Non-performing Housing Loans” chart presented a few weeks ago at the Mule Stable (I believe there’s no way I can reproduce it here; if I’m mistaken, let me know, so that Chris can see it): like I said before, what that chart shows to me (based on US, UK and Spain) is: (1) the quality of a loans portfolio can deteriorate pretty quickly and (2) it doesn’t need to deteriorate too much, to have devastating effects on the whole of the financial system. Here I would add that a loans portfolio can look quite safe, until people realize it is toxic.

    At the time, and I hope not to be misrepresenting your position too much, your answer was that US sub-prime borrowers intended to default at the first sign of trouble.

    Here, although I strongly disagree (for several reasons, in fact) I am not prepared yet to argue. However, I would like to leave this comment for the record.

    But, independently of the ability of home buyers to service their debt, there is a more fundamental issue both yourself and Chris overlook. The analysis you both made seem to be concerned exclusively with home-buyer/occupier loans. And even if this analysis were all good and well (which, as I have argued above, there are reasons to suspect it’s not entirely true), it’s not the whole picture.

    What about investors? And on the supply side, what about builders?

    Now, allow me a bit of cheekiness: man, those Aussie banks are clever little devils… They get all those mortgage loans and insure them, so in case something goes wrong, someone else pays for it. In other words: bye-bye risk!

    If only the Yanks had thought of that, instead of clutching all their mortgages close to their chests…

    Oh well!

    PS: Is there any data on the LVR distribution?

  18. Stubborn Mule Post author

    Marco: I’m happy to have a look at your figures. You can post the spreadsheet on drop.io or alternatively put them into a Google docs spreadsheet.

    I will have another look at the CBA presentation…it may have some LVR distribution information in it.

    While I do think that US subprime mortgages were more likely that most Australian borrowers to default in the face of falling prices. One reason is that the practice of taking out unaffordable loans at low “honeymoon” rates and “flipping” the property–i.e. selling it–before the rates stepped up was a common practice in the US before 2007 but one that falls in a screaming heap if prices fall. It may happen here, but not to anything like the same extent. Another important difference is that the practice of securitising the US mortgages and then in turn packaging the resulting securities into so-called ABS CDOs had the effect of (1) levveraging exposure to the underlying mortgages and (2) spreading them all around the world.

    As for mortgage insurance, the mortgage insurers tend not to like insuring sub-prime mortgages for obvious reasons. It is common enough for prime mortgages to be insured in other countries around the world.

  19. Stubborn Mule Post author

    Marco by the way, my LVR calculation was simple:
    0.46 x 80% + 0.54 x 10% = 42.2%. Also, if you have 30% LVR mortgages and property prices fall by 40%, the property is still worth more than the loan (double in fact), so even if you were to replace all the 10% LVR mortgages with 30% LVR mortgages, the losses in the face of a 40% drop would not change.

  20. Marco aka Cracticus

    Hey Stubborn,

    I’ve done what the drop.io site said. What now?

    They’ve sent an email, as well. Do you need the info it contains?

    “by the way, my LVR calculation was simple: 0.46 x 80% + 0.54 x 10% = 42.2%”.

    Yes, I know. I think it’s wrong. 46% and 54% are the percentages over the total of loans. I think the average LVR should be “total of loans” divided by “total market value”

  21. Stubborn Mule Post author

    Marco: your advisers gave you the correct advice, but unfortunately it looks like you used the wrong link: it’s saying “this drop is empty”.

    In the meantime, your approach to calculating average LVR would explain the differences between us. However, the problem with the “Total Loans”/”Total Value” approach is that it gives too much weight to the low LVR loans. If you have two loans secured by property A and property B and borrowed collateralised by B defaults, you cannot draw on the value of property A to mitigate your loss. For this reason, this is not the method used by banks to report portfolio average LVRs.

    “Total Loans”/”Total Value” is effectively an average LVR weighted by property value (you can see this in the case of just two loans in this equation). Portfolio average LVRs are instead calculated as averages weighted by the loan amount not the property value, which is the approach I used.

  22. Stubborn Mule Post author

    Marco: I got your email, thanks. I have posted a modified version as a Google docs spreadsheet. The reason you got a loss of 12% at the portfolio level was because you were assuming that all the loans were defaulting, whereas I worked backwards to see how many would have to default to get a loss of 6.6% at the portfolio level. You’d get this if around 58% of the high LVR loans defaulted, or 26% of the whole portfolio.

  23. Marco aka Cracticus

    Thanks for your patience, Stubborn,

    “Portfolio average LVRs are instead calculated as averages weighted by the loan amount not the property value, which is the approach I used.”

    Although this may be common practice, it seems kind of weird, doesn’t it?

    However, once we accept this, the rest of your numbers certainly make sense: for instance, the 12% loss (with 100% default) vs 6.6% loss (with 58% default) makes sense: twice the default rate, twice the loss.

    I see what you mean with the weighted average thing, as well.

    BTW, I believe Chris interpreted the “average LVR” in the same way as I, for his numbers seem to match mine.

    Regardless, I trust you to know your finance stuff, better than I.

    Still, this does not address what I consider the most pressing matters:

    (1) As a consequence of an external shock, could mortgagors start defaulting at rates high enough to destabilize the system? This is critically dependent of the ability of people to service their debt and on the sensibility of the system to defaults. You might have heard of recent renewed reports on home owners using their equity as ATM.

    As you’ll imagine, this ultimately, is motivated by the chart I keep insisting on.

    (2) Can investors and developers start selling off housing stock as a consequence of their own endogenous dynamics?

    If investors count on capital gains, any slowdown in price growth could diminish their demand, and selling off stock.

    If developers count of capital appreciation, they could find themselves unable to service their debts, and selling off stock, as well.

    The combination of the two things could could feedback on slower prices.

    And recently there have been reports on lower clearance rates at auctions and even prices starting to fall. Just last night Allan Kohler was talking about house prices increasing only in Victoria and that because of the Victorian state government’s very own “FHBB”, by the way (you’ll remember Mark and I have talked about this at the Stable).

    (3) Is it possible some kind of interaction between (1) and (2)?

    (4) Even if banks are properly insured, the amounts we are talking about are quite considerable. Are we sure insurers can actually service them?

    Insurance, as we all know, only spreads the losses among others. It’s the proverbial fan. In the US case, the crap fell largely overseas (it put the global into GFC).

    As I said before: someone could end up with a lot of loose and bloodied “cojones” in their hands.

    Let me be frank: I am not saying that we should run for the hills, just yet. But I think there are a whole lot of reasonable questions (at least they seem reasonable to me) that need to be answered. And although you made a gallant attempt, I don’t see anybody answering them.

  24. Stubborn Mule Post author

    Marco: I do think that the conventional averaging approach is reasonable. If you invest $50 in something low grade with a “score” of 10 (whatever that may mean) and $50 in something high grade with a “score” of 20, then I think it’s reasonable to define the “average” score as 15. That’s really all that’s going on with the average LVR calculation. Still, it doesn’t actually make much difference to the example we were looking at: all the action was in the 80% LVR loans not the 10%/30% LVR ones.

    As for your other questions, they are all good ones. I should emphasize that, while I do see various differences between Australia and the US when it comes to the property market and home lending, I am certainly not of the “the only way is up” school. Although looking at property prices in terms of rental yields paints a less dramatic picture than in affordability terms, it is hard to imagine rents will keep rising at the rate they have over recent years. I don’t expect to see a 40% crash (although it’s certainly possible), but I would not be at all surprised to see volatility in prices with not much in the way of net gains and indeed possible net declines in coming years.

    As for the specific questions, here are some thoughts. (1) a spike in defaults would certainly put a big hole in bank earnings (although I don’t think the majors would collapse) and the economy would be very sick. It would be very bad news for Australia, but would not cause ructions for the rest of the world (there are plenty of other things that could do that, of course).

    (2) The scenario you paint could certainly cause nasty feedback.

    (3) Without a doubt!

    (4) If the downturn is severe, there could certainly be insurers going out of business (it’s happened before). At that point, any losses on mortgages not yet covered would come back to the banks. Overall though, taking probability of default and loss given default into account, I would say an 80% LVR loan without mortgage insurance is riskier for a bank than an 85% LVR loan with mortgage insurance.

    So, I am by no means a blind optimist. My main argument with bubble-proponents is that I don’t see a bursting as inevitable. Possible certainly. My main argument with the presentation discussed here is that even if there is a bursting of the bubble, it is not inevitable that the major banks will collapse (although they would certainly have some disappointed shareholders looking at shrinking dividends).

  25. johnnyboy

    Gents (particularly the author, which I believe is Stubborn Mule),

    I consider myself to have been educated significantly by discovering Steve Keen’s ideas some years ago, but what would I know…. I’m just a regular old Engineer who can barely hold back his laughter when he reads mainstream economic theories that have the gall to suggest that macroeconomic systems are linear mean reverting systems…

    I can’t even get the simplest of engineering systems to operate in linear modes at all times…. pray tell which cigarette are these guys smoking when they think a massively more complex macroeconomic system will behave linearly. Who needs comic fiction when you’ve got these comic gems!?

    Anyway, I digress….

    I’ve enjoyed the to and fro on this subject of Australian banks.

    I’d be interested on your responses to looking at this problem from a different angle.

    Step back in time to 2008-2009, during the first wave of the “GFC”.

    Pray tell, if our Aussie banks are so safe and comfortably well buffered, how come they went on bended knee to the Australian Federal Government to secure Federal Government guarantees on the debt they needed to issue/rollover? I know this question is not directly related to house prices, but (I think) it goes to the heart of the matter of the true picture of the strength and solvency of the Aussie banks, and sheds an interesting light on same.

    The knee-jerk answer would undoubtedly be “because the the debt markets had seized and the banks couldn’t get debt funding at all”. But this is not true. They could always get the funding, but at a much higher interest rate. It should be obvious (but I can’t prove it) that there are only two reasons why the banks took the action they did. It was either because:
    (a) These effect these higher borrowing costs would have on their financial position was so serious that the banks found themselves desperately seeking lower borrowing costs via Fed guarantees to avoid potential calamity, OR
    (b) The potential damage from the higher borrowing costs were not potentially calamitous to the viability of the banks, but the execs were so cocky that they went hat-in-hand to the Feds anyway merely to minimise the reduction in profit.

    If instead you’re answer is along the lines of “well yes they obviously recognised they would find it difficult financially to secure funding at, but it wasn’t going to cause a critical situation by any stretch”, then this is very similar to asserting that the Australian government put taxpayers on the hook merely to maintain banks enormous profit levels rather than saving them from going to the wall (or something very close to it). If that were the case and the word got out then a whole bunch of Australians would be justifiably frothing at the mouth with anger that the Feds would serve the interests of the banksters so egregiously.

    In short, if the banks are comfortably cushioned via low LTV’s and safely solvent, how come the Fed guarantees on their debt?

    Remember, these Fed guarantees were requested while the Aussie CRE market was still relatively OK and the Residential RE market was still chugging upwards. So absent any major RE issues, the banks still had this major problem that required the remarkable degree of intervention, presumably to keep them viable?

    So, one can’t help but wonder what position the Aussie banks may quickly find themselves in if we have another round of debt deleveraging (if it even has stopped at all since ’07) and another flight out of international debt markets AND even minor declines in the Aussie RE market….

    I readily accept that Aussie banks are, generally speaking, better cushioned than most US and European banks. But the picture, to this poor engineer, appears under the surface to be (on balance) more akin to regarding the Aussie banks as ‘relatively less sick’ than as ‘relatively healthy’.

    Counterpoints welcomed.

  26. Stubborn Mule Post author

    johnnyboy: thanks for the comment…you raise some good questions. Before getting into some of them, I should point I that I am also not a fan of excessive reliance of linear models in a largely non-linear world. Perhaps that’s because my training is as a mathematician not an economist!

    Going back to the thick of the financial crisis…I remember it well! Here is a presentation about the market turmoil back in May 2008. Keep in mind that I wrote this after Bear Stearns collapse but before Lehman Brothers. Fast-forwarding a bit, here’s paper about the government guarantee for Australian banks, written back in October 2008. As you can see if you have a look at these, I was giving these topics a lot of thought at the time! Those two are somewhat sanitised versions of documents written for another audience, but I did also write a few posts on the subject for the blog as well. Australia and the Global Financial Crisis is still generating more traffic than most, and there were also posts more specifically focused on the guarantee such as this one.

    Nevertheless, I don’t expect to you read all of those pieces so I’ll put a few thoughts in this comment as well. One point to make is that, while we will never really know what discussions took place behind closed doors between banks and the government, the sense I get from people who should have some insight into the matter is that the four major banks were not agitating for the guaranteed and were in fact taken by surprise when the announcement came out. The banks who stood to benefit the most in my view (they may dispute this of course!) were the smaller regional banks such as Suncorp and Macquarie Bank. I believe their funding difficulties were far more acute than anything the four majors were experiencing, although of course wholesale funding was very expensive for everyone at the time. To really go out into the realms of speculation, the cynic in me can’t help thinking that some in the four majors might have been disappointed by the guarantee as it robbed them of the schadenfreude of seeing some other players collapse! Who knows?

    Whatever the real story behind the guarantee, the problem being addressed was fundamentally one of liquidity not solvency. Here’s a quote from the second document referenced above:

    Banks can get into trouble if they make bad investments, lending to companies or individuals who default on their loans or if they investing in securities which suffer large price drops. This is solvency risk, the risk on the asset side of the balance sheet. However, even with strong assets, banks can still get into trouble on the liability side. Banks buy their assets using funds taken in as deposits and raised by issuing debt securities in the wholesale financial markets. Most of the deposits are at call or in short-term deposits and much of the wholesale debt is also short-term. On the other hand, their assets are mostly long-term. If a bank starts to lose deposits or struggles to issue new debt as wholesale securities mature, they cannot demand that borrowers repay their loans early. The worst case is a run on the bank, with every depositor trying to withdraw their money at the same time. This is liquidity risk and liquidity risk can trump solvency risk: no matter how good their assets are, no bank can survive a run.

    Of course solvency risk and liquidity risk are intertwined: one reason a bank can have funding problems is precisely because of concerns in tha market about their solvency. However, this is not always the case. A good example is RAMs. They got into trouble not because their mortgage book had deteriorated but because they relied so heavily on short-term funding in the US CP (commercial paper) market and they simply could not refinance their debt.

    Contrary to your suggestion that it is always possible to get funding–sometimes funding is available but expensive, sometimes it’s not available at all. You might argue that there’s a price for everything, you might just not like the price, but what if asset-backed commercial paper is trading at, say, 75 cents in the dollar (not out of the question at the time). That means to refinance $75 million dollars worth of debt, you would have to take out a loan that will require you to repay $100 million in the future, but would only give you $75 million today. What if you had assets worth $95 million, debt of $75 million and shareholder equity of $20 million (leverage of 4.75 times, which is not out of the question). If that debt was repayable, you simply could not borrow at 75 cents in the dollar as you would incur a debt of $100 million, which would exceed the value of your assets and you would automatically be bankrupt, no matter how good your assets were!

    That may seem like an extreme and artificial example but it was essentially the situation faced by RAMs and potentially some of the smaller banks were not far from the same fate. So, the mere fact that the Australian government felt compelled to step in with a government guarantee does not in and of itself say very much about the health of the banks’ loan portfolios. I should also add that once the guarantee was in place, all the banks (majors included) naturally took advantage of it. After all, the guarantee fee (payable to the government) was 0.7% for the majors and at the time they would have had to pay a significantly higher premium over guaranteed debt than that if they tried to issue bonds without a guarantee.

    Long comment….hope it makes some sense!

  27. Stubborn Mule Post author

    David, I’m not sure which “exposure” you are referring to. The original quote referred to the share of total assets represented by lending assets (which stands at around 77% on the page you linked to), but that is hardly surprising as lending is the primary source of assets for a commercial bank. More likely you are referring to the share of loans represented by mortgages, which is over 60%. That’s certainly high, which means that if there is a significant downturn in the mortgage sector, CBA is likely to be hit harder than the other four. However, it is still extremely unlikely that this means that losses on the mortgage book would wipe out their equity. If you look at page 41 of the pack from December 2010, you’ll see that losses on corporate lending are larger proportionally than consumer and (although not shown on this slide), within consumer, losses on credit cards and personal loans are higher than on mortgages (the former being unsecured). If CBA lost, say 0.5% on their mortgages, that would be an unprecedentedly bad result…getting to the 6.6% referred to in the Business Insider slide would be phenomenonally unlikely. If they blow up their balance sheet, it would be much more likely to be the result of a number of very large corporate loans defaulting.

Leave a Reply