APRA gets it right on RMBS liquidity

From the AFR, APRA head John Laker last night snuffed out hopes that RMBS may be included in an expanded basket of qualifying assets for Basel III liquidity requirements for banks:

“The discretion to add additional assets is qualified by the fact that these assets … must have a proven record of a reliable source of liquidity in markets even during stressed market conditions…So the discretion is carefully qualified by the fact that these assets need to demonstrate liquidity in duress.”

Chris Joye, Mark Bouris and Nicholas Gruen, all mortgage interests, have argued that the inclusion of RMBS  is warranted because it will release the tax-payer burden apparent in the RBA’s committed liquidity facility, known at MB as “cash for coconuts”. They all have a point.

But the problem is that RMBS in the US and Australia froze solid during the GFC. The reason is simple enough. When mortgages melt down in unforeseen ways, the investors that trade the floating securities that package the mortgages can lose faith in the market. You get a kind of modern bank run, investors from an entire market, as opposed to depositors from a bank.

So it doesn’t make a whole lot of sense to let the banks, which will also be under pressure at that point, to carry reserve liquidity assets that could very well also freeze.

It’s as simple as needing a lender of last resort. Which basically is what the RBA is offering in the committed liquidity facility.

Now, of course, there are some big problems with the RBA solution. Moral hazard is one. Another is loss of market transparency. When banks do get squeezed it becomes very difficult to know who is insolvent or who is not when they’re all feeding at the RBA teet together. Another is diminished competition as small banks have to pay more for money through a less liquid RMBS market. I admit these are very serious drawbacks.

But it’s surely better than spinning the credit roulette wheel again. Especially since we have a very under-capitalised banking system vis-a-vis mortgages and even more so for the mortgage insurers which are a crucial part of the credit enhancement of many tranches of the RMBS.

David Llewellyn-Smith


  1. “Chris Joye, Mark Bouris and Nicholas Gruen, all mortgage interests, have argued that the inclusion of RMBS is warranted because it will release the tax-payer burden apparent in the RBA’s committed liquidity facility, known at MB as “cash for coconuts”. They all have a point.”

    Well yes, but you’ll never see them criticising for buying RMBS (unless it’s for not buying enough) so it’s pretty clear that they’re only against tax-payer burden when it doesn’t help their business model.


    “The federal government has earned profits of more than $600 million from investments in mortgage-backed securities over the past four years.”

  2. I’m probably not understanding the issue here, but mortgages held on their books are part of the Teir II structure. So does this mean that RMBS tranches sold don’t qualify as Tier I because they will need to be rolled over at some point in time?

    • The issue is the banks obligations to meet their Liquidity Coverage Ration (LCR) requirements (defined by the structure of their liability portfolio). The volume of Level 1 assets in Australia is probably insufficient to cover the banks LCR investment requirements. Ordinarily Covered bonds and RMBS would be allowed as Level 2 assets (subject to exposure constraints), but the caveat on these assets has always been one of demonstrable liquidity – and it is fair to say that there is insufficient evidence for “stressed” market liquidity for these assets in Australia.

      In the absence of sufficient available assets to cover liquidity needs, the fallback is to either modify liability structures (more retail deposits, smaller balance sheets) or to carve out an alternative approach to liquidity support for banks.

      This has been done through the use of a Level 3 category for assets (largely at the discretion of local regulators). Banks pay for the facility and the RBA published a note outlining the decisions underpinning the pricing for the committed liquidity facility. In summary though, the banks have the option to deposit cash with the Reserve Bank, it counts as a liquid assets and defines a least/worst cost option for liquidity.

      • Thank you for the reply – do banks have sufficient cash to lodge with the RBA or will they need to undertake a capital raising?

        I know that question may be premature.

        • When banks raise liabilities, (deposits – retail/wholesale, term debt, equity etc) most of it gets lent back out as an asset. The balance is invested in liquid assets to assist in managing movements in liabilities. These liquid are coins and notes, deposits with the central bank and sovereign debt (for Australian banks).

          So in the sense that you ask, they by definition have enough cash….they just used some of the cash to buy RMBS instead of holding it (cash) or buying sovereign debt.

          • I’m not sure that I understand your point – banks will minimise cash holdings due to risk and lack of return, so it’s not a solution for them.

            Are you referring to RMBS purchased or sold?

          • If they’ve bought RMBS they don’t have the cash.

            They have a marketable security/interest in a trust which recent history shows is not liquid when liquidity is most needed. There are also multiple layers of counterparty, documentation and enforcement risk that does not exist with cash. EG robo signing scandal

            Issuing covered bonds are the building of a potential disaster for long term unsecured creditors (most likely to get bailed in) and the Australian government in the event of an Australian house price collapse.

  3. Is this tacit acknowledgment from APRA that our housing stock is very vulnerable to a fall in value and therefore that bank capital requirements are inadequate?

    It also seems odd to want to have RMBSs as collateral for big banks who will be in distress due to weakness in the housing market and therefore RMBS market.

    • Not really, I think that it just reflects the run down in sovereign debt prior to the GFC and the realities if trying to meet global guidelines for liquidity.

      As for whether or not weakness in the housing market will materially affect banks – that’s an area wide open to debate. I do find however, that the discussion is too often focussed on the risks for loss given default (ie, the cost of a defaulted loan) rather rather the probablity of default. In my view, much of the commentary assumes that a house price > mortgage = default. Further, there is often (in my view) an implicit assumption that any a large house price adjustment will put many loans under water. I think both these points warrant good discussion – after all, losses are a function of both a default event and the cost of resolving the event. Focussing on one aspect of the cost at the expense of the other can somewhat arbitralily distort discussion.

      • Good points – as I understand it most underwater mortgages in the USA continued to be serviced. The ones who couldn’t adjust their borrowing rates through a refi were the vunerable group, whereas here most mortgages self adjust as rates fall.

        • Arrears data from the Mortgage Bankers Association (the charts can be obtained from http://www.huduser.gov) show that most borrowers in the US prime mortgage space have continued to service their obligations – arrears rates are much lower those seen in the sub-prime market.

          • And check for differences between states where there is no risk of bankruptcy from walking away from an underwater housing loan, just a hit to the credit score.

            There are also state variations in enforcement or rights of foreclosure, some states require judicial oversight making the process longer and more expensive and difficult to be arbitrary, while other states do not require judicial oversight and so the process is much faster.

            In judicial oversight states with no personal liability for shortfalls you can just stop paying on an underwater mortgage and sit there for 12 months rent free until you are finally foreclosed and evicted.

            At a 3x median earnings housing multiple 12 months rent free is a good start to buying a replacement home!

  4. Deus Forex Machina


    Laker and your assesment of secondary market liquidity is correct.

    but the second part of your argument misses a crucial point I think which is that the market causing the run on a bank may just be wrong.

    So I have to disagree with you on this one with regard to housing assets and the RBA – a bank which may be otherwise healthy or able to survive a run may suffer a shortfall in the short term if they lose market confidence.

    But in saying that banks cant use balance sheet assets such as internally rated RMBS for the RBA LCR to get through the liquidity crisis means that the person or people pushing the bank have the upper hand and the bank will fail without the ability of the Central bank, RBA, Lender of last resort to evaluate the solvency and long term outlook for the instituion.

    Bagehot said you should lend genrously but at high rates to solvent banks and that is what we should do because in the absence of an ability to collateralise and borrow against balance sheet assets the banking system becomes more fragile not more stable.

    The assests in an interenally held RMBS are therefore a crucial bridge for the institution that may have hit aa short term issue which is not going to threaten it long term.

    The RBA’s LCR can buy time at a crucial time not only for the bank but ultimately for the depositors and the Australian Government (who has guaranteed all deposits under 250k) and crucially the rest of the banking system.

    The alternative is a banking system that needs to carry 30-40% liquidity in its day to day activity which ultimately would be a tax on potential growth or the rest of the economy.



    • Those liquidity levels would take us back to the pre Keating days when FRB lgs and srd ratios added up to over 40% of funds – it would thump the economy.

    • “a bank which may be otherwise healthy or able to survive a run may suffer a shortfall in the short term if they lose market confidence.”

      The definition of surviving a run is not suffering a shortfall in the short term.

      The RBA can still do repos for banks based on purchase of the RMBS, it’s just a liquidity ratio that should suffice say 99% of the time. When it is not sufficient the RBA can do repos of their RMBS to provide additional liquidity. The RMBS might only secure a gain proceeds of say 50% of face/market/estimated value, giving the selling bank a big incentive to repurchase

      To my mind the only entities that ought be able to get security over bank assets is RBA and Government and their ought be clawback provisions against all but insured depositors who get cash out from funds provided by government and RBA if the bank fails or wipes out its shareholders funds.

      In other words a traditional insolvency clawback should apply to creditors paid out any time that the bank was insolvent.

      By the way, how can you tell in say a week (you need to be making fast decision at times of crisis) whether a bank is solvent or not and more importantly whether it will be solvent next week so that you don’t put in good money after bad? That is why additional funds to provide liquidity in crisis when there is a risk of the bank being insolvent ought be highest security.