An audit of Irish Debt

My UL colleagues Sheila Killian, John Garvey, and Frances Shaw have produced a valuable report Auditing Irish debt. In particular, pages 13–19 will be of interest to the readers of this blog, as will tables 9 and 10, where the total debt guaranteed by the State is collated. Table 9, which I reproduce below, just looks at the covered institutions.

Author: Stephen Kinsella

Senior Lecturer in Economics at the University of Limerick.

36 thoughts on “An audit of Irish Debt”

  1. @John Foody

    Rumour has it, that’s not all he said. I am so glad I’m not his PR Guy today.

    Er, he was quoted out of context mein Frau Doktor?

  2. So total liabilites are €371bn or around 2.5x GNP. It would be interesting to see a similar analysis of the offsetting assets – i.e. how much the net debt is.

  3. Carson.

    Theres only about 5bn left in the NPRF isn’t there? Not sure about remaining cash balances.

    The banking shares should be worth something.

    We could argue (possibly optmistically?) the Nama assets are worth par.

    Also the ELA and ECB funding should reduce as deleveraging takes place.

    Finally, and possibly being pedantic here, but the Exchequer Figures for end-March actually say that the first 3bn was paid out by then so the promissory note should probably be 3bn smaller in the above table.

    (Should the 10bn plus in interest we have to pay on the promissory notes be included here or would that be an accounting fail on my part?)

  4. @Carson,


    Oh for a set of pro forma balance sheets and income and funds flow statements rather than the ‘tennis club’ accounts the Government uses.

  5. @ Carson

    A guarantee is not a liability, otherwise every insurance company would be hopelessly bust.

    Re Table

    I’m confused by the treatment of the promissories. Are these not similar to sovereign bonds, why are they appearing in a list of guaranteed liabilities?

  6. “A guarantee is not a liability, otherwise every insurance company would be hopelessly bust.”

    That is true to an extent but I don’t think your comparison with insurance companies is entirely correct. Note that the authors of this report refer to the €371bn figure as the ‘Scale of Irish National Debt’ as of 31/03/11 (table 10).

    In an exercise where you were working out the net debt, you might reasonably assume that the deposit guarantee has a net value of zero as the small expected value of a loss in a catastrphic scenario is offset by the fees the banks are paying for that guarantee. Similarly, you would add back your assumption for the fair value of the NAMA assets etc. In such an exercise would our net debt position look much different, if at all, from the headline debt number we see frequently cite in the debt/GDP ratios used by the IMF etc?

  7. The paper states that the sharp rise in National Debt over the last couple of years (from 70Bn to over 90Bn) is primarily due to the banking crisis. We are running a fiscal deficit of c. €20bn p.a. so I think that observation is misplaced. The banking crisis has been mainly funded by promissories and raiding the NPRF.

    Whilst the tone of the paper is not quite Kellyesque and it does produce verifiable figures rather than wild estimates, it still comes across as tendentious. It would have greatly benefitted from a “putting in perspective” of the contingent liabilities. In fact we now know that in our type of socio-economic system the people in reality underwrite their banks, its just that we have been forced to be explicit about it. The other side of that coin is of course the assets of the banking system. If these are worthless then the guarantees are also worthless for there would be no economy to fulfil them.

  8. Thanks Stephen for an interesting read. It is good to see all the various liabilities of the State pulled together in a single source. Some coverage of the assets offsetting many of these liabilities would be useful to complement this.

    It is a debate we have had elsewhere and I am pleased to see that the authors have netted off the value of the Promissory Notes when determining the potential liability created by the ELA from the Central Bank.

    I am less enamoured of Table 10 which takes the €279.3 billion of liabilities from Table 9 reproduced here and adds €91.8 billion of government securities to get the “scale of Irish national debt at 31 March, 2011” of €371.1 billion.

    First of all I think the national debt (lowercase) is much higher once all debt owed by residents in the State is summed. However, the indication given here is that the figure of €371.1 billion is some measure of public debt that must be paid by the State. That is patently untrue. Many of these liabilities are hanging over the State, some with signiificant risks, but in my view there could be no realistic expectation that the State will find itself in a position where it will have to repay/rollover all of these liabilities.

    I do not think the difference between debt and contingent liabilities is given sufficient weight by the authors. What really matters is not the size of the contingent liabiities but the amount of them that may have to be paid by the State.

    In many instances the State is only providing insurance cover for these liabilities. The actual liabilities can only appear on one balance. For example, customer deposits are either owed by the banks or they are owed by the State. They will not be repaid twice.

    I think a better way to look at the banking debacle is to look at the asset side. The State will never be repaying deposits or bonds or any liabilities in the banks. We are on the hook for losses in the covered banks and we will provide the money to meet these losses that will then allow the banks to meet their liabilities. The amount of money we have to provide to the banks will be determined by the losses on the asset side of the balance sheet rather than the size of the items on liability side. We’re at €63 billion and running on that count!

    Of the items in the €279.3 billion from Table 9 we are fully on the hook for the €30.9 billion of Promissory Notes and I would view them as comparable to government bonds. All the other items will have assets that will offset, and in many cases do so entirely, the amount of the liability to be met by the State.

    There are also some elements left out of the analysis. It is acknowledged that there is €13 billion of retail debt outstanding. This is money that the State will have to repay. At the 31st March we had also drawn down about €18 billion of funds as part of the EU/IMF programme. There is another €30 billion that could have been added on, and rightfully so in my view.

    I also think the reasons given under Table 10 to suggest that this €371.1 billion figure is a “conservative estimate” are less than stellar.

    I presume the first reason is referring to ECB liquidity. We know the remaining assets in the banks are crap. That’s why they were told to raise €24 billion of capital in the stress tests. The only way the ECB liquidity will become a problem for the State is if the assets are even worse. Possible, but the €24 billion gives us a good start.

    Second very little of the €91.8 billion of government bonds were issued because of the banking crisis. Up to the end of 2010 the Exchequer had pumped €4.7 billion directly into the banks. There was also €4.5 billion put into the NPRF in 2008/09 that also be said is part of the bank bailout. This only total about 10% of the outstanding stock of government securities. I am not sure why “it is reasonable to assume that a large part
    of the €91.8 billion raised by the sale of government bonds was also due to the banking crisis.” We also built up a substantial cash buffer by issuing bonds and this stood at €16 billion at the end of 2010. Most of the bonds were issued to finance the ballooning deficit from 2008.

    The report is useful but I would have liked to see a little more consideration given to the asset side of the balance sheet.

    In March of this year the amount of money the State actually had to repay/rollover was around €165 billion (bonds, retail debt, promissory notes, EU/IMF loans). This is going to increase over the next few years but will not approach €370 billion. The annual deficits will continue to push up the debt as well as any losses in NAMA or additioal losses in the covered banks. The key is that it is actual losses not simply contingent liabilities that will determine our public indebtedness.

  9. Thanks for the detailed reads, all. Apologies for not responding in detail – am travelling without a laptop for a few days, and I just have a few minutes here in the airport

    Seamus, I’m sorry you don’t think the difference between debt and contingent liabilities is given sufficient weight. Oddly enough, one of the things that has been praised about the audit is that it made that distinction quite clearly. That’s obviously the idea behind separately tabling the 92 of long-term bonds, and the 279 of internally-generated and contingent liabilities. So I don’t accept that there is an indication that 379 needs to be repaid. This is an audit, not an economist’s forecast. As such it lists real and contingent debt, without any assumptions about the strength of the contingency. You’ll see that in any set of financials

    I also don’t agree that very little of the €91.8 billion of government bonds were issued because of the banking crisis. We were running comfortably below 40 billion up to the international banking crisis in 08. We’re now more than double that. It seems clear enough

    I’d love to see the work expanded to look at the asset side. We’ve detailed all our sources, and I really hope you can pick it up and run with it along the lines you suggest. Let me know if you want any of our data – it would be great to see your ideas developed

  10. Good read on both sides.

    As Seamus said we stopped issuing bonds in September 2010 and the 91.8bn cited above does not include 30bn+ promissory notes.

    So…given the majority of the bank recapitalisations took place after September 21st it seems difficult to argue that that portion or our debt is due to the banking crisis.

    If however, it is meant in a more indirect sense, that a property crisis/banking crisis led to a shortfall in Government revenues and subsequent deficits then that would make a little more sense..although I wouldn’t really agree with labelling fiscal deificts as banking costs.

  11. @ Sheila

    Table 10 clearly states that the “scale of Irish national debt at 31 March, 2011″ was €371.1 billion. I would be of the view that a debt is something that has to be repaid or rolled over. The Irish State did have have debt commitments totalling that amount on that date. As I said above our debt on that date was around €165 billion, which in itself is horrendous.

    There is a whole heap of other liabilities, and the collation of these in the report is a very useful addition to the debate, but there is the potential for full or partial repayment from another party before the State has to step in. NAMA may be able to repay the NAMA bonds. The covered banks may be able to meet their liabilities (albeit because we have already pumped €63 billion into them).

    I still don’t think we can use the banking crisis to account for the increase in government bonds from 2008 to 2010. We issued €12 billion of bonds to build up our cash reserves from €4 billion to €16 billion in 2008 and 2009. Excluding bank related payments, the annual deficits for 2008, 2009 and 2010 came to around €45 billion. These account for the more than doubling in the amount of government bonds outstanding from 2007 to 2010.

    The only bank-related transactions that can have had an effect are the direct cash injections to Anglo and INBS from the Exchequer Account and the 2008/09 transfers to the NPRF. All other monies used for the banks up to 2010 came from the creation of Promissory Notes and the destruction of the funds built up in the NPRF over the previous seven years. We issued no government bonds with the explicit purpose of recapitalising the banks.

  12. Some interesting developments today alright. It is hard to know if the interest rate reduction is on top of what was already factored in or is just confirming what was previously indicated.

    Any changes to the Promissory Notes could be significant. How would a 30 year repayment schedule work if the IBRC is wound down by 2020 as Alan Dukes has recently suggested? A change to the interest rates on the Promissory Notes would have a greater effect. The final €9.1 billion tranche from January of this year has an annual interest rate of 8.6%. Halving that could save €400 million a year in interest costs.

  13. @Seamus Coffey

    ‘… we are fully on the hook for the €30.9 billion of Promissory Notes and I would view them as comparable to government bonds.’

    This citizen dissents; on the former the citizenry was not consulted and at an opportune time Blind Biddy will gladly slip off that ‘hook’ and BURN them in the FIRE.

    On the latter, you are entitled to your personal, if dangerous, opinion. Rotten Toxic Mushrooms may be comparable to next year’s Fresh Spuds- but only in the sense that they are both forms of food. Dodgy Promissory Notes are NOT genuine Sovereign Debt, albeit both may be compared in the sense that they are both forms of debt.

    That said, keep up the good work.

  14. “The authors acknowledge the support of the Debt and Development Coalition, Afri and UNITE”

    This audit was commissioned by the above and although useful I believe it will crop up without the clarification of contingent liabilities versus actual debt.

  15. @ Seamus Coffey

    Any changes to the Promissory Notes could be significant.

    Surely this is smoke and mirrors. We own Anglo, so the promises are to ourselves. The only way reducing those promises would ever be a real saving is if we were at some stage to walk away from Anglo. But Noonan is clearly paving the way for a U-turn to a position were we will honour all outstanding Anglo liabilities.

    I agree with your critique of the paper which is really quite flawed by commission as well as by omission.

  16. @BWII
    Not really – the promissory notes are used to get funding from the ECB. They are subject to a particular haircut from the ECB. If the ECB will give more generous terms on the haircut, it will reduce Anglo’s interest expenditure so they will require less interest income. Currently, we are effectively paying interest to Anglo to pay to the ECB and the ICB…

  17. @ Hog

    I was wondering about that. Are you sure you are right? My understanding is that the haircut on the collateral does not affect the interest on the borrowings, for example I might put my house up as collateral for a €1K loan. But I am a bit above my paygrade here so I may be wrong. If it is used as backing for ELA from the ICB does my argument that it is simply going from one pocket to another not still stand, even if you are right?

    In any case, even if it is as you describe the extra interest to the ECB is surely small beer compared to defaulting on the unguaranteed seniors. Noonan is trying to wiggle off this hook, quite pathetic.

  18. @BWII,

    It’s all to do with the interest rates. It will take more radical action if we are to benefit on the capital side and that seems very unlikely at the moment.

    The annual coupon on the Promissory Notes ranges from 4.2% to 8.6%. As a result of the 2011/12 “interest holiday” the annual interest payments from 2013 will range from 6.1% to 11.8% across the four tranches.

    As hoganmahew says if the terms of the Promissory Notes can be changed we can make savings on the interest costs of the Promissory Notes through their use as collateral with the ICB (I don’t think the ECB accept the Promissory Notes as collateral).

    When Anglo’s half-year results were released at the end of August, Mike Aynsley came out and said that the final capital cost of Anglo could be between €25 and €28 billion.

    That is lower than then €29.3 billion of capital we have put into it so we may get some back. However it should be noted that an important reason for the reduced capital requirement is because we are due to pay €11 billion of interest to Anglo over the lifetime of the Promissory Notes.

    If there is to be a small surplus from the huge amount of money we have poured into Anglo, it is better to claim it through reduced interest payments now rather than wait for a capital return when the bank is wound down in a decade or more. This is not really good news just news that is less bad.

  19. @BWII
    To take your last point first – not over the 10 years the promissory note will be in place.

    Well, the collateral haircut increases the price (i.e. the amount of money you get for a fixed interest amount). The collateral haircut is in part dependent on the terms of the promissory note – it currently has to bear a market rate of interest (no idea what that is at the moment). A reduction in the interest paid could increase the haircut.

    You’re right, though, the promissory note is at the ICB rather than the ECB. It may be on ELA rates? Which are about 3.5%, I think.

    Namawinelake had a post a while ago on Anglo’s results showing that interest from the promissory note constitutes most of Anglo’s income, I think.

    I don’t know any of these things for sure – the structure of the promissory notes s a little bit opaque – not on the face of them, but how the ECB/ICB considers them. For sure there is something odd about the interest bearing element of them. It came as a surprise to many of us that they would bear interest and while the reasons why have been explained, the reason for those reasons have not – to me it makes sense that national bad asset agencies get, effectively, free funding to workout bad loans (in the absence of a european resolution mechanism).

  20. @ Seamus & Hog

    I suppose I can see that there are real advantages to reducing the interest rates on the Promissories but I think Noonan is chancing it when he suggests that this would be superior to welching on the 3.5Bn unguaranteed bonds, which has been his mantra for some time.

  21. @ BW II

    Of course they are not mutually exclusive. Engaging in one does not eliminate the other.

    The saving generated by reneging on the unguaranteed bonds would more than likely be less than €2 billion. The saving on the interest rate could exceed that amount.

    However, it is important to remember that the gain on the interest rate is going to happen anyway. If we leave the interest rates as they are there should be a capital surplus left over when the bank is wound down. The saving, such that it is, exists in the system. We are just changing the timing of it.

    Defaulting on the senior bonds would be an actual saving above what will happen anyway and would see some of the cost of this debacle shouldered by someone else.

  22. @Seamus Coffey
    “If we leave the interest rates as they are there should be a capital surplus left over when the bank is wound down. The saving, such that it is, exists in the system. We are just changing the timing of it.”
    This depends on the interest rate that the ECB/ICB is charging on lending against the promissory notes – the government can reduce the payment on the promissory notes, but this is going to result in an operating loss at Anglo if the interest cost it pays to the ICB/ECB is not reduced.

    At the moment, by NWL’s figures, Anglo is making an operating profit which is offsetting the capital losses it makes. As Anglo is not active in the debt markets, it’s interest expenses must be down to the ECB/ICB. No?

    If the price for getting reduced rates at a long term from the ECB/ICB is not haircutting senior bonds, then that is a tradeoff that needs to be examined in terms of which is the better price, not in terms of which is the most popular or indeed the most moral.

  23. @ Hog

    For example, United States Treasury bills, which are seen as fairly safe, might have a haircut of 1%, while for stock options, which are seen as highly risky, the haircut might be as high as 30%. In other words, a $1000 treasury bill will be accepted as collateral for a $990 loan, while a $1000 stock option might only allow a $700 loan.

    This is from Wiki. It suggests to me that the collateral haircut does not have any implications for the interest on the loan.

    @ Seamus

    Mr Noonan said he was “committed to raising burden sharing again”, but that the promissary note issue was “more valuable to Ireland than any burden sharing on residual bondholders.

    Agree that both could be on th etable but Noonan is clearly preparing his exit here. I think you are agreeing with me that Noonan is wrong in his assertion that the promissory note issue is more important than burning bondholders.

    @ Seamus

  24. @ hoganmahew

    I looked briefly at the Anglo half-year results here.

    In the first six months of 2011, Anglo recorded an operating loss of €105 million, but this was really only because it collected/accrued €644 million of interest from the Promissory Notes while paying €519 million of interest to the Irish Central Bank for ELA. This difference is a “profit” for Anglo for borrowing off one State institution (ICB) and lending to another (DoF). It now appears that this can be reduced.

    Who gets the profit the ICB makes on the ELA?

  25. @Seamus Coffey
    “Who gets the profit the ICB makes on the ELA?”

    I would imagine the ICB will need that “profit” to write it against the losses they incur on some of the 56b of crap the ECB won’t take.
    I see from the link above that you estimated the eventual cost could be 60b.
    The spin and bs coming out of government gets worse.

  26. @Seamus
    Apologies – it was indeed your good self not NWL…

    Well, I *think* that the Irish government gets the profit of the ICB. On the other hand, companies that make large profits tend to “pay to get the best, because clearly we’re worth it, look at the paper profits we are making”.

    The difference, though, is that we have to borrow now to pay Anglo who then pay the ICB who then realise a profit the following year. In the meantime, we’ve paid a year’s interest on that profit (since we have a fiscal deficit).

    Even assuming that all the interest paid comes back to the state (which I think is doubtful), it is an inefficient use of money.

  27. @BWII
    Supposing you need 1000 in cash, though. A 1% haircut means you have to put up collateral worth 1010ish. A 30% haircut means that you have to put up collateral worth 1300.

    Given that the collateral is promissory notes that the government is paying interest on at face value (not at the discounted value that the ICB would use), this makes a large collateral haircut expensive for the state to provide Anglo with the amount they require (at a minimum the deposits they have transferred and the cash collateral they are required to hold on their derivative positions).

  28. @ Hog

    My understanding is that the lender gets none of the return on the collateral unless of course the collateral is called.

    I think it works as follows: Anglo needs 100M liquidity. ECB (let’s say) says fine that will cost you 1.5% and by the way we need 110M of those promissory things as collateral security.

  29. @BWII
    Yes and the state pays 6% on the 110million – the state has to provide the extra collateral.

    So if the haircut is increased as a result of the state paying lower interest on the promissory notes to Anglo (i.e. as marketable instruments they move to a lower rate and so a smaller NPV), the state has to increase the collateral to Anglo.

    That is why the state needs agreement from the ICB and behind them, I’m guessing, the ECB on the haircut to be applied. As I understand it, NCBs are free to set collateral requirements, but within the overall framework of the ECB’s collateral requirement system, if I am not misquoting M. Trichet.

  30. A very good paper that dispels the smoke and mirrors.
    A much needed eyeopener, the authors and the University of Limerick have performed a valuable public service.

  31. I agree with Seamus Coffey that “some” contingent liabilities are included. Well if that is the case why not provide in the figures the massive amount of potential liabilities in guaranteed Pensions that have and are continuing to be doled out to Civil and Public Servants by our Governments during the last decade. Over the next 4 months alone between 8 – 10,000 Civil and Public Servants will be given average Pension pots of €1 million each without a rex provided by Government available to meet it. The irony is that some here spend millions of words berating Politicians for not burning bondholders but at the drop of a hat have no problem with the same Politicians taking on to the balance sheet massive unfunded Pension liabilities for the taxpayer. You can also throw in the 1000s of millions in Pension liabilities of University Pension Schemes taken over by Government in 2009. University Academics are very critical of Governments taking over bank liabilities but have no problem when it involves themselves.

  32. I am from the USA and have been eagerly following the Irish story as I am absolutely sure that there is far too much funny business going on in government finance here as well. I have some basic questions…

    1. What were these debts incurred for?
    2. Was it ever truly determined who is holding all this paper?

    Thank you. Keep up the good work, Ireland. Remember that Ecuador was able to really make a dent in their debt through their audit.

Comments are closed.