Who pays wins: a reply to Mark Capleton
Julian D. A. Wiseman
Abstract: An essay dated 18th November 2008 by Mark Capleton of
The Royal Bank of Scotland entitled Who pays wins (in several ways),
refers several times to an essay by this author, Gilt Asset Swaps: DMO Should Profit, dated October 2008.
Between the lines of our broad agreement are the various disagreements discussed below.
Publication history: Only here.
Usual disclaimer and copyright terms apply.
On 18th November 2008 Mark Capleton of
The Royal Bank of Scotland published a research essay Who pays wins (in several ways)†1 referencing, and mostly agreeing with, this author’s essay
Gilt Asset Swaps: DMO Should Profit. However, there were several disagreements, overt and implicit, minor and major, old and new.
[Wiseman’s] second bad effect is one we struggle with.
He feels that gilts trading the wrong side of libor at the long end is a cause of 5y UK sovereign CDS spreads trading at worryingly wide levels, thereby damaging perceptions of UK government creditworthiness.
Our problem with this argument, mapping libor spreads to credit spreads, is the maturity difference.
We would be less resistant to his case if we were talking about similar maturities, but he is talking about 5y CDS spreads.
5y gilt swap spreads are historically rather expensive to libor, so causation is not obvious. We will leave this second “bad effect” for the reader to chew on, …
This is a fair observation, but omits a vital fact: it is very difficult to trade UK CDS longer than 10 years.
If the CDS were causing gilts to trade wide, one would expect all gilts, or maturity-matched†2 gilts,
to trade wide.
But if the causality were the other way, that some gilts trading wide has affected the CDS, one might see something close to current pricing.
My guess is that the cheapness of gilts is part of, a non-trivial part but not most of, the cause.
Hopefully this was understood from the phrasing in the original essay: “Partly because UK gilts are trading so cheaply, …”.
— the market has long been in such a distorted state that although long real and nominal gilt yields have been very low in absolute terms (suggesting strong demand),
long nominal and linker auctions have not been consistently well covered. And this could be a concern should the DMO want to step-up in supply …
This confuses a macro-economic state (low yields) with a micro-economic problem (auction mechanisms).
Yes, bidding at a government bond auction involves short-term unhedgeable risks.
But these risks are relative to the current market price, and exist whether demand be high or low.
These risks can be removed with a better auction mechanism†3,
but it is unlikely that the DMO would introduce a much improved auction mechanism for the swap auctions, so switching from gilts to swaps would not necessarily improve the number or quantity of bids.
Living in New York, I had missed Charles Goodhart’s article in the Telegraph†4, but the Professor’s suggestion would introduce a new danger, in return for a benefit that can be had far more cheaply.
If a government funds the deficit by selling 30-year bonds, each year the government would need to refinance about a thirtieth of the debt.
But if the government funds solely in ½-year T-Bills, the debt needs to be refinanced sixty times as often.
For the rich world, the 1998 crisis was about hedge funds, and the 2007/’08 crisis about banks.
The next might or might not be about countries.
For the UK to be putting itself in a position in which it would need to sell a significant fraction of the national debt every few months is needlessly risking trouble.
Hence my suggestion, contrary to that of Professor Goodhart and of RBS, but like that of Lombard Street Research†5, that the DMO shorten to 3-year gilts rather than to ½-year T-Bills.
(But Professor Goodhart’s desideratum that the authorities “forcibly liquefy the UK's financial system” contains a nugget of truth, a large nugget that can be had far more cheaply.
However that requires better action from the Bank of England, which should reform its money-market operations, its implementation of monetary policy, to make a price rather than to add a volume, as described in a recent letter to Paul Tucker†6.
That suggestion would make T-Bills, gilts, debt of some foreign governments and indeed a range private-sector assets immediately and instantly convertible into the medium in which payments are made, without an auction, without a queue, and without scaling-down of any requests.)
The two other possibilities would seem to be for a floater to reference either an ois rate or a TBill auction clearing rate (as with Italian CCTs).
Imagine that, as some coupon is being fixed, the floater is owned mainly by the banks, the same set of banks that might bid for the T-Bills.
Those banks would then have a very strong incentive to ensure that the T-Bill auction goes very badly.
And the sterling financial ecosystem lacks investment funds specialising in T-Bills (because they have historically been very small and illiquid), so there are a dearth of non-bank bidders.
Hence the DMO’s interests would be ill-served by relying on a T-Bill fixing.
There is a similar problem with an OIS ‘fixing’: there is little other business dependent on such a fixing, incentivising banks to act one way.
Libor is different: banks are positioned both ways, some long, some short, and the fixing is widely watched and hence reasonably well policed.
Libor would be better in this sense, but, as Mark Capleton also says, the object is for the DMO to receive floating Libor, not pay it. Further the complexity can’t be good.†7
But we agree that the DMO should fund short (even if we disagree on how short), and pay long.
Julian D. A. Wiseman
New York, November 2008
†1: Who pays wins (in several ways) (alas an RBS client password needed), Mark Capleton, 18th November 2008.
†2: A better match, as discussed in On The Plotting of Yields, would be the ratio of true Macaulay convexity to true Macaulay duration, but that detail isn’t important here.
†3: See A Better Auction Mechanism, And Why Governments Should Sell Futures Rather Than Debt, written as long ago as December 1997 having been first discussed with the Bank of England in April 1995.
†4: Pencil in 2010 as the start of recovery from the recession, Professor Charles Goodhart, Telegraph, 31st August 2008.
†5: Debt management policy is in need of adjustment, Lombard Street Research, Financial Times, 7th November 2008.
†6: The Implementation of Monetary Policy: The Next Attempt, published in early November 2008.
†7: The complexity of floaters has form, as described by a 1994 Bank of England Press Notice: “The Floating Rate gilt, first issued in March 1994 with a further tranche in June, pays interest at 1/8% below the London Interbank bid rate for three-month deposits.
The rate is reset every three months.
By mistake, the interest payment to be made on 9 September was set on the wrong date, on 10 June rather than on 9 June.
The error came to light when the next payment was being set, in September.
Had the rate been correctly set, it would have been 0.01562% higher and the recipients of the interest payment on 9 September would have received an extra 0.39 pence per £100 nominal of stock.
These persons are being individually notified and extra amounts are being paid with interest.
The Bank is also contacting holders of the stock at the time the incorrect rate was announced who subsequently sold stock and may have suffered loss.
The Bank regrets this mistake. 20 October 1994”.
Obviously the Bank hadn’t done its private Libid fixing on the correct day, so must have estimated the answer: note that 3-month GBP Libor dropped 1.563bp between the 9th and 10th.
Floaters are needless complexity; whereas swaps are easy because SwapClear does all the work.