Covered by whom? Bonds on what?

When you’re abducted by aliens, there’s little cause to be cynical. The sucktitude of a painful probe is straightforward. There’s no sugarcoating to see past, things are exactly as bad as they seem. Between the screams you realize that, in a way, you have been offered a kind of innocence. When you are abducted by aliens, you savor the silver linings.

After such an ordeal it’s a bit depressing to be dropped off on the Planet of the Covered Bonds. Cynicism levels are off the tricorder here as, alas, they should be. [See Yves Smith, Michael Shedlock, Maxed Out Mama. Less cynically, David Merkel offers a very nice description of what covered bonds are and how they work.]

Covered bonds sound nifty. They’re designer drugs. They’re just like the mortgage-backed securities that gave us such a fine party, except the nasty hangover inducing components have been engineered away. They are on-balance sheet loans, look Ma, no Enron! (finally…) Covered bond issuers have “skin in the game”, skin, bone, and sinew actually, as they guarantee the loans. That problem of “misaligned incentives” is solved, ‘alleleujah! (…though intrafirm agency problems are not addressed.) These are old-fashioned, full recourse, secured and overcollateralized loans, just packaged into tradable securities. What could possibly go wrong?

Formally, the only way anything could go wrong would be if the issuing bank fails and the pledged assets turn out to be worth less than originally estimated. Do you think those two events might be correlated? Covered bonds can certainly be no worse, from an investor standpoint, than the nonrecourse asset pools they are intended to replace. A guarantee by the issuing bank has gotta be worth something. If it were 2002 again and the banking industry had adopted this originate and guarantee model (rather than the originate and forget model they chose), perhaps we wouldn’t be in the current mess. But it is not 2002. These bonds will be offered by banks that would already have collapsed without vast support to the financial system by the Fed and the US Treasury. Guarantees by money-center banks are no longer bonds of confidence in the prudence or skill of bank managers. The value of such guarantees comes from a different place, from the notion that it is unthinkable the state would permit these banks to fail. A covered bond offered by Citi or Bank of America would only default if a titan collapsed. Investors might reasonably believe that would not be permitted to happen. If they are right, then these bonds are indeed covered. They are covered by you, dear taxpayer.

The great credit crisis of 2007-2008 is slouching towards its Bethlehem, a full faith and credit crisis for the United States of America. This die was cast at the first TAF auction, when the Fed chose to pull private credit risk onto taxpayers’ already strained balance sheet, rather than endure any unpleasantness. Covered bonds may prove to be a success with investors. But, careful what you wish for. The more banks sell, the more we’re all on the hook, if the loans go bad. Covered bonds issued by “too big to fail” banks are basically equivalent to mortgage backed securities guaranteed by Fannie and Freddie. It’s just another way of putting private-sector bells and whistles on a public sector assumption of risk.

These bonds are seen as a way of “unfreezing the housing market”. The housing market seems frozen because in many areas, relationships between home prices, rents, and incomes are still out of whack. Assuming relatively stable rents and incomes (bad assumption, I know), mortgages in “stuck” markets made at or near current asking prices are likely bad investments. That suggests the implicit taxpayer guarantee won’t expire unused. The more covered bonds are sold, the more extreme measures or hidden subsidies will be required to prevent household names from failing.

The committee to save the world is you, and we will be grateful for your contribution, although we will never thank you, or admit that anything other than the skill of our red knuckled, fabulously wealthy financiers had anything to do with the eventual recovery. That period of commodity inflation and steep yield curves was just a market outcome, a fact of nature. Of course our proud financial institutions were always going to weather the storm. They are the best and most sophisticated in the world. Thank goodness for private enterprise.

[HT Yves Smith on the slouching towards Bethlehem thing. BTW, my use of the term “taxpayer” is imprecise, state guarantees are really backed by “taxpayers and/or those most vulnerable to inflation (low bargaining power workers and those on fixed incomes)”. My guess is that we will use tradables inflation more than outright taxation to save the whales. FD, I’m short long-term Treasury futures and long precious metals, going with that whole full faith and credit crisis scenario. As always, this ain’t investment advice, I frequently lose my shirt so go copy Warren Buffett or something.]

Update: Felix Salmon directs us to an excellent new blog on which John Hempton writes:

[I]f you have lent to people in a currency where interest rates suddenly go to 50 percent (as happens in some devaluation crises), your funding cost (deposits) will rapidly go to 50%. However if you pass that on to your borrowers they will fail. You will suffer credit risk and possibly go insolvent. If however you have offered fixed loans to your borrowers you will wind up with huge funding mismatches – and possibly go insolvent. For small moves there is a difference between credit risk and interest rate risk. For large moves there is no effective difference. The same analysis applies to currency and credit risks.

This goes some way towards adressing Tyler Cowen’s demurral

I cannot see that the credit of the United States government is in danger. There is a) the printing press, and b) our location on the left side of the Laffer Curve.

I agree with Tyler that the US government is more likely to print than default outrught, if those are the alternatives (though do recall this from the generally levelheaded Accrued Interest). But, for large moves, I think the distinction between credit risk, inflation risk, and currency risk is largely academic. (Regarding Tyler’s second point, Robert Olson hits the nail on the head in a Marginal Revolution comment, “Being on the left-side of the Laffer Curve doesn’t matter much if the political situation makes it impossible to raise taxes.”)

Update History:
  • 29-July-2008, 11:09 a.m. EDT: Fixed misspelling of Warren Buffett’s name. Thx Nemo.
  • 31-July-2008, 3:03 a.m. EDT: Added the word “by” somewhere where it was needed. Added the update regarding credit risk vs inflation/currency risk.

15 Responses to “Covered by whom? Bonds on what?”

  1. Alessandro writes:

    London Banker on RGE Monitor has an interesting take on ‘covered bonds’ issued by smaller banks, worth at least some thought. In the event of a failure would covered bond holders get the good assets before the FDIC insured deposits? (I’m asking, I don’t know)

    How covered bonds affect the liquidation of a failed bank?

  2. Alessandro writes:

    Ouch! here is the link to London Banker post

  3. Anon from Convenience Yield writes:

    My first intuition on the convenience yield was options related. My first on covered bonds is that it is the balance sheet equivalent of credit default swaps used as asset insurance (implicitly, also option inspired).

    Here is an early, not necessarily complete or entirely correct, train of thought:

    a) A loan is equivalent to a credit default swap issued by a lending bank to a borrower, where the equivalent CDS is over-collateralized at the maximum risk exposure and where the reference credit is the borrower (i.e. “self-referencing”).

    b) The borrower has no counterparty risk, because the “CDS” is collateralized to the maximum. The lending bank is exposed to the credit risk of the borrower, of course.

    c) The lending bank issues a covered bond, essentially to “insure” its credit risk.

    d) Thus, a covered bond is issued, referencing a group of loans that is equivalent to a group of maximum collateralized credit default swaps, sold by the lending bank to the borrower, and where the reference credits are the borrowers.

    e) The covered bond “monetizes” and “insures” this group of loans. This “monetization” is structurally equivalent to an “embedded” balance sheet securitization, where there is a specific linkage between an asset group and a newly issued liability class.

    f) The effect of the covered bond monetization is to establish the bank issuer and the bond buyer as counterparties in a second transaction in which the borrower’s credit is now a third party credit reference.

    g) The bond buyer is the “writer” and the bank is the “buyer” of this second maximum collateralized CDS (i.e. covered bond).

    h) Thus, the bank has hedged or effectively insured its original position as lender / maximum collateralized CDS writer.

    i) The position of the bond buyer is isomorphic to that of a CDS insurer that posts maximum collateral.

    j) The big difference compared to a bank buying normal CDS is that the bank issuer of covered bonds has no counterparty exposure to the bond buyer because of the effective maximum collateralization.

    k) The bond buyer as writer is exposed to both reference credit risk and the issuing bank’s credit risk. The bond buyer as writer is exposed to more credit risk than a normal CDS writer because of maximum collateralization with respect to counterparty (bank issuer) credit risk.

    l) So covered bonds are the balance sheet equivalent of credit default swaps. The difference from normal CDS is that the covered bond buyer has effectively posted maximum risk collateral. This eliminates the counterparty risk normally faced by a CDS buyer, but increases the counterparty risk normally faced by a CDS seller.

    m) Thus, caveats applying to writing CDS insurance, regarding underlying and counterparty risk, also apply to covered bonds, and more so because of maximum collateralization. On the other hand, the covered bond issuer (buyer of insurance) is in a more secure position with regard to counterparty risk (say, in comparison to counterparty risk in the case of CDOs or CDS insured by “bond insurers”).

  4. Nemo writes:

    Steve —

    You got it. The plan is to convince timid investors to buy these things because they are implicitly guaranteed by the U.S. Government. Heck, worst case, the FDIC can make good on them… And thus will the moribund mortgage market get “kick-started”. This is basically adding a bunch of new Fannie/Freddie type arrangements.

    Because the problem with the mortgage market, you know, is that we haven’t socialized enough of the losses yet.

    Paulson during the press conference:

    “As we’re all aware, the availability of affordable financing is essential to turning the corner on the current housing correction. So we’ve been looking broadly for ways to increase the the availability and lower the cost of mortgage financing to accelerate the return of normal home buying and refinancing activity.”

    If you thought the key to higher demand was lower prices, you are waaaay behind the times. The key is cheap credit. Duh!

    P.S. His name is “Warren Buffett”, not “Buffet”. :-)

  5. Benign Brodwicz writes:

    Steve —

    Especially creative and useful post. I really liked the alien abduction metaphor.

    “Covered” = covered by more fiat money. LOL. Major sweeping under the old magic flying carpet.

    But I still think, at least for the short term, that long rates are coming down.

    However, there is a general expectation of something catastrophic happening later this year. My research area used to be “animal spirits,” and my models do not show us hitting bottom yet–about 60% through a typical cycle–despite generational lows on surveyed confidence measures. A crisis would almost certainly send long rates up, at least for a while. There seems to be a major DOD and media push to back down from Israeli demands to bomb Iran, but madness remains in high places in this country, IMHO, so that’s an open question.

    But I’m still a deflationist intermediate to longer term.

    Lots of people do seem to think gold will spike.

    FD: I’m long a gold fund, intermediate term Treasuries, and money, and will probably short the NDX after the turn of month rally.

    None of this should be construed as investment advice.


  6. Anon writes:

    Alessandro: I think the answer to your question is Yes, teh FDIC is second in line. “Under U.S. bankruptcy law, collateralized claims like advances are settled first during failure resolution. ”

    see here

  7. Nels Nelson writes:

    Given the recent commentary by Bill Gross on the need to stop the decline in real estate prices, the panic surrounding the problems with the GSEs, the recently passed housing rescue bill and now Paulson’s plan to use covered bonds to provide loans for real estate, it’s balls to the walls to stop the decline of house prices and patch the leaking bubble. If this is the case, is it just my imagination or are these leading lights saying the future of the U.S. and its financial system depends on something as unproductive as our personal residences and their use as places to store all of our accumulated consumer detritus. If this is the case, it’s more frightening than the thought of being abducted and probed by aliens.

  8. Anon writes:

    Alessandro: This in the FT today: “For example, the FDIC said banks should be restricted from using covered bonds for more than 4 per cent of their funding in order to avoid depleting the assets available to repay ordinary depositors and other unsecured creditors if a bank failed.”

  9. Anon from Convenience Yield writes:


    SRW – I said what I wrote above was “not necessarily complete or entirely correct”.

    In fact it was entirely incorrect in at least one gigantic way. However, the error can be corrected and the sequence completed I think by inserting just one extra step, with the rest remaining as is.

    (What I was attempting to do was to translate the balance sheet situation of covered bonds into an array of equivalent credit default swaps with a “maximum collateralization” characteristic – a task of risk analysis made more difficult by the fact that, as of 10 a.m. this morning, I knew virtually nothing about covered bonds. Now that it is later in the day, I know at least epsilon about covered bonds, which translates to – this may still be wrong, but should be closer to being right.)

    I think what I actually described was a bond that was exposed to the credit risk of the underlying assets independent of the credit risk of the bond issuer. As I understand it, a covered bond is exposed to the risk of the underlying assets only if it is exposed to the risk of the bond issuer.

    Therefore, what I wrote can be corrected by an additional step in which the bond issuer effectively sells a “normal CDS” on the underlying assets to the bond buyer (i.e. normal not being the” maximum collateralized type” that I was describing as equivalent to balance sheet transactions). This step now reverses or hedges the risk of the underlying credit to the bond buyer, as previously described in the sequence, except for counterparty (bond issuer) risk exposure in the amount of this hedge.

  10. Benign Brodwicz writes:

    Re: the inanity of “stabilizing” house prices, quick & dirty

    This helpful blog posts the following:

    Year: Median HH Income Median House Price Ratio

    1965: $6K; $14K = 2.3X

    1970: $9.5K; $23K = 2.4X

    1975: $12K; $35K = 2.9X

    1980: $18K; $62K = 3.4X

    1985: $24K ; $75K = 3.1X

    1990: $30K; $92K = 3X

    1995: $34K; $110K = 3.23X

    2000: $42K; $139K = 3.3X

    2005: $46K; $219K = 4.75X

    The 1965-2000 average ratio is just under 3.0. If we take 2006 as about the peak, about 8% above 2005 on average according to Case-Shiller, that implies a ~40% decline in house prices for the ratio to revert to mean, and the Case-Shiller index shows only about a max 20% drop so far. Moreover, real median household incomes are falling pretty rapidlly, implying a larger drop in prices will be required.

    It is ridiculous to think that housing prices can be stabilized at current levels. The primary effect of pumping porous reserves underneath the magic flying carpet of fiat money will be to make the ride more treacherous via monetary multiplier effects. Even if the stuff is marked to market correctly, which I doubt is happening, given the talk of stabilization, people can’t afford houses at these prices. And it will get harder.

    N.B. The reason I think we’ll deflate: The US is going to confront what we used to call a “collapse of effective demand,” meaning that most people simply won’t have enough money to spend to maintain a level of AgD. Rising prices from oil, weak dollar, as well as increased precautionary saving, etc. will simply take demand from domestic items. Unless the distribution of income becomes much more equal in a hurry, or some huge fiscal stimulus is applied (it would be nice if both occurred, with the fiscal stimulus aimed not at another war but domestic human and nonhuman capital development), aggregate demand will collapse. At 20% of GDP, our federal government’s budget is actually *small* by many developed nations’ standards.

    Sorry, Ben, but you don’t have the silver bullet for this one. If your federal government partners don’t pick up the ball and run with this one, y’all will have succeeded in turning the USA into one big banana republic.

  11. Anon from Convenience Yield writes:

    Something about the analysis troubles me.

    Financial institutions deemed “too big or interconnected to fail” have an implied 0 probability of failure, and an implied non-0 probability of taxpayer bailout.

    Covered bonds as per the discussion are subject to two types of coverage – explicit asset coverage in the case of failure, and assumed taxpayer coverage in the case of bailout.

    But such taxpayer coverage applies to all bonds and liabilities in the case of a bailout, whether or not such bonds or liabilities are “covered” explicitly by assets.

    So asset coverage for bonds is worthless in the case of a “too big or interconnected to fail” institution.

    Asset coverage only has value in the case of an institution that is allowed to fail.

    In the event of failure, asset coverage of bonds does become a cost to the taxpayer to the degree that the assigned asset cover drains valuable collateral from the total pool that might otherwise be available for FDIC liquidation in order to fund its obligations to insured depositors.

    In the event of the bail out of an institution that is “too big or interconnected to fail”, there is no such incremental cost due to asset coverage. Asset coverage is redundant and has no value.

    There is an alternative approach to the idea of “too big or interconnected to fail”.

    Consider instead the idea of “almost too big or interconnected to fail”.

    So we can assume there is a very small probability of failure, q, for such a large or interconnected institution.

    And conversely we can derive (1 – q) as the probability of either continuing as a going concern or in the worst case being bailed out.

    Then the value of the asset coverage for the bond is:

    [q][v] + [1 – q][0] = [q][v]

    (Where q is the probability of failure, and v is the differential benefit of a contractually enforced collateralization or asset cover for the bond at the point of failure. q must be positive in order for asset cover to have value; i.e. the institution can’t be “too big or interconnected to fail”, but can be “almost too big or interconnected to fail”)

  12. a writes:

    “If however you have offered fixed loans to your borrowers you will wind up with huge funding mismatches – and possibly go insolvent. For small moves there is a difference between credit risk and interest rate risk. For large moves there is no effective difference.”

    Not true at all, at least not in the conceptual way explained here. Banks actively manage interest rate risk by hedging asset and liability sensitivity. It’s not like the old days of the S&L’s.

    In fact, the only interest rate risk mismatch that can’t be hedged very easily is the asset sensitivity offset to common equity funding. And in his case, higher rates on assets are a benefit to net interest margins.

  13. All — Sorry to have flaked for a coupla days.

    Anon from CY — So, in general I like the approach of analogizing or decomposing things into familiar instruments to get an intuition for what’s going on. But I’m having a hard time following the CDS story. Any non-risk-free bond can be analogized in part to a “fully collateralized” CDS written by the lender to the borrower (lender gets a credit spread, but pays in the form of a write down when borrower defaults, borrower faces no credit risk on this transaction). But that’s only a piece, there is the risk free interest component, to a covered bond as to any other. A covered bonds credit component is more complex, in that there are two sorts of credit quality that can impact the bonds’ values (quality of the issue, quality of the security pledge by the issuer). I’m sure there are ways of taking the credit part of the arrangement and decomposing into CDS-like pieces, primarily between borrower and lender, but also pieces involving borrowers from the underlying assets, albeit at much lower rates. But I’m still having a hard time using a CDS analogy to come up with a story that seems easier to understand or value than an analogy to old-fashioned collateral backed bank loans.

    Benign — The inflationist/deflationist debate is a tough call, I think. I’m much more certain of volatility than I am of direction. (I suppose I should be looking into straddles, then…) Overall, though, I think we’ll see oscillations but eventually inflation, because the US’ international and public debt position would be too painful under deflation, and though you can’t force people to borrow and spend newly printed dollars, you can force creditors to accept wet cash despite their lack of enthusiasm for holding it.

    Anon from CY II — If you let the probabilities of guarantee be 1, then any sort of credit enhancement is meaningless, here is no credit risk (at least no credit risk above Treasury risk). As you say, if an institution is absolutely TBTF, then it’s just the state. But if it’s “almost” TBTF, then enhancement matters. Covered bonds are just bonds issued by banks with a certain form of credit enhancement, which has real value if there is any doubt of a guarantee. I also think there’s a “behavioral” component that is understated by this analysis. Suppose investors have shunned an asset class, “irrationally”, or due to rational principal/agent incentives (lose now on mortgage bonds, you’re fired immediately… lose on something else and it’s an ordinary investment loss). Merely rebranding the class, even if the substantive difference is small when you do the math, might overcome the stigma and rekindle investor interest. I think this is the main reason for the covered bond push — persuade investors that these are the good, safe mgage bonds, the real thing this time, not what they thought was AAA two years ago….

    a — Your point is well taken that even for large moves, interest rate risk, currency risk, and credit risk are hedged differently, and it might be easier to hedge one than another. With the possible exception of credit, all three risks are easily hedgeable individually, by banks as well as by anyone else. But now we have a new risk to hedge, we can’t tell what form a large change in real valuations will take, so what do we do? Do we hedge all three? That’s expensive, in theory we should earn a risk free nominal rate if we fully hedge, in practice I suspect hedging all three risks well enough to cover extreme scenarios via any channel would leave negative nominal return. Traditionally, different parties are accustomed to hedging different risks. A US bank might balance maturity and interest rate risk quite adroitly and earn positive nominal USD returns under any scenario except outright default by the Treasury. But, it still faces inflation risk, so ignoring outright repudiation, its hedge is partial at best in real terms. (Sure, since its creditors and investors get paid in nominal USD, the firm is therefore hedged against the possibility of bankruptcy, but seeing through the firm to the claimants themselves, they are not hedged against inflation/currency risk.)

  14. RueTheDay writes:

    Steve said: “The housing market seems frozen because in many areas, relationships between home prices, rents, and incomes are still out of whack.”

    Bingo. Nothing more really needs to be said. House prices should represent the NPV of their rental value, which in turn should be driven by the incomes of those seeking the houses. There is no magic here.

    Regarding Tyler’s statement: “I cannot see that the credit of the United States government is in danger. There is a) the printing press, and b) our location on the left side of the Laffer Curve.”

    This is faulty framing, IMO. The US government is not going to just go bankrupt, obviously. However, we are reliant on foreign financing of our huge trade deficit. At some point, the “printing press solution” will cause our creditors to start demanding payment in something other than dollars. Once you have an external debt denominated in something other than your local currency, you can no longer inflate your way out of it, and you are in for some rough economic conditions ahead…….or war, as history has demonstrated time and again.

  15. Buy bonds? Hahahahaha the Mobgambu Guru would say. Franz Pick, 1898-1985 used to say, “Goverment debt certificates are certificates of guaranteed confiscation. They are either defaulted or inflated away”.