Duncan’s Economic Blog

The PBR in One Picture

Posted in Uncategorized by duncanseconomicblog on December 9, 2009

As I type Osborne is ranting again about ‘markets losing confidence’. He is totally and utterly wrong.

The chart below shows the yield on ten year government bonds. If the markets are worried they demand a higher interest rate and the yield goes up. If the yield falls it shows confidence.

Here’s what happened as Darling spoke.

Looks like a credible deficit reduction plan to me.

61 Responses

Subscribe to comments with RSS.

  1. Paul said, on December 9, 2009 at 1:44 pm

    Brilliant – have retweeted.

  2. Peter Kenyon said, on December 9, 2009 at 1:45 pm

    Agreed. One picture…..

  3. Daniel said, on December 10, 2009 at 12:06 am

    I’m afraid you are totally and utterly wrong – the chart you are showing has Gilt yields barely moving. 5bps? The bid/offer spread is 1-2bp. Show a long term chart or not at all.

    Apart from anything else, no-one is going to aggressively sell Gilts whilst the BOE keep buying them through QE.

    When QE is over however, expect a huge rout….rates are already as low as they can go, the fiscal situation is dire and suddenly the BOE will have to come to market and sell huge amounts of paper, with the UK not far off downgrade status.

    What really is amazing is how little QE and rate cuts have actually moved yields lower, despite simply huge purchases. Its testament to just how profligate the government have been and what a horrible mess the UK is in.

    • duncanseconomicblog said, on December 10, 2009 at 9:15 am


      5bp is abig move for one day in the gilt market. And the fact remians that whilst Darling spoke, gilts became more attractive to buyers.

      Happy to show a long term chart, the yield is near record lows!

  4. […] Duncan asks if the market knows more than George Osborne. Yes, yes it does. […]

  5. Daniel said, on December 10, 2009 at 12:40 pm

    Gilt market can move a lot more than 5bp in one day. I’ve worked in it. For example, 10y yields are roughly 12bp HIGHER today.

    Is it surprising that Gilt yields are at all time lows, with rates at all time lows, huge Gilt purchasing by the BOE through QE (equivalent to this entire years new debt stock)?

    If memory serves, when rates were cut from 1% to 0.5% and QE started, 10y Gilts were at 3.70%. Which is roughly where they still are, despite 185bn of QE purchases. What does that tell you? It tells you the market is already gearing up for these things to sell off. People know another 15bn of purchases are to come, and I highly doubt yields will move a great deal before the end of QE, though traders will happily get long the free option (so short Gilts) Darling/Brown have handed them.

    Please do show a long term chart though.

  6. Daniel said, on December 10, 2009 at 1:14 pm


    Firstly, for housekeeping, you never state *which* Gilt yield you are talking about. I’m assuming 10y though.

    Your chart shows exactly what I’ve said though – yields are in fact now higher than when rates were cut from 1% to 0.5% AND after 185bn of bonds have been purchased by the BOE. By your logic that would suggest the markets have NO confidence in the government to sort this problem out….which in this case is completely true.

    185bn is not far off the entire amount issued this year, especially when you account for re-investment of coupons/maturities. Effeectively the government has funded itself this year by printing money!

    Your charts are also very misleading to the uninitiated – yields will always be higher if the base rate is higher. Its not as simple as saying low yield = good, high yield = bad. Its a moving target.

    To show a better description of the percieved value of Gilts, you should really be charting the return-on-investment chart of real yields. I.e. the chart of the spread between Gilt yields and the base rate adjusted for inflation. You could of course just look at the CDS which also shows the same story – people are demanding more protection and higher yields for holding UK government debt. Relatively speaking, McDonalds is now a safer investment.

    Not that the UK can ever really default (it can print money), but the effects are devaluation of the currency and increased debt servicing costs. From my position in the market, I’ll tell you for free that everyone already thinks it’s stuffed. Rates can’t go any lower and QE will at some point end and have to be reversed. Pension funds already are lightening up the Gilt holdings. The only thing holding yields in at the moment is the other 15bn the BOE have to spend (and the chance they’ll be forced to spend more afterwards, extending QE) and the chance the that the Tories, when they inevitably come to power, do a Lenihan and cut significantly, as Ireland have been forced to.

    • duncanseconomicblog said, on December 10, 2009 at 1:50 pm


      This argument is getting silly.

      Gilt yeilds fell whilst darling spoke. What does this suggest?

      • Daniel said, on December 10, 2009 at 3:08 pm


        Gilts yields have gone much higher today, and are higher now than before QE started.

        What does this suggest?

        It must also be noted that there were 1.7bn of Gilt purchases by the BOE yesterday, which could easily explain any move lower in yields, which have now been more than reversed.

        The link someone else posted to the spectator blog makes the point very clearly. Its now 3 times more expensive to insure UK debt than German debt. With UK 5y CDS at 85 bp roughly, that is half way to emerging market territory

        I’ve looked through your “economics” blog. You should probably rename it. It’s economics light, Labour party-political heavy.

        As such, I doubt you will ever accept that the Gilt market is trading extremely heavily. Or infact that Labour have done anything at all wrong in their management of the economy. If you are so sure though, i’ll happily put on a friendly little wager with you.

  7. Alex Ross said, on December 10, 2009 at 1:21 pm

    Hi Duncan,

    What do you think of this post?


    You’re the man to argue against it – not me!🙂

    • duncanseconomicblog said, on December 10, 2009 at 1:55 pm


      Thanks – I’ll have a look this afternoon.

  8. Bill le Breton said, on December 10, 2009 at 1:22 pm

    The second chart at that June posting compared UK, US, Swedish and German yields. Are they still moving together?

    • duncanseconomicblog said, on December 10, 2009 at 1:54 pm


      Broadly – yes. The usually do!

      • Bill le Breton said, on December 10, 2009 at 2:39 pm

        And each would have difffering positions on money creation/QE, budget deficits etc

  9. duncanseconomicblog said, on December 10, 2009 at 2:50 pm



    The point I’ve laboured over a few times is that interest rates on gilts are determined by loads of factors.

    • Daniel said, on December 10, 2009 at 3:15 pm

      Ugh, again, wrong.

      That chart shows rate from 2007 onwards – i.e. the start of the global recession.

      OF COURSE rates are going to move roughly together in a huge global recession. The point is though, the example DOES NOT MAKE THE RULE.

      If you take longer term charts, you will see that sometimes rates will trend together, and sometimes they don’t.

      Even in this recession, you can show that spreads in bond yields have diverged greatly – check out UK/Bund spreads for a quick example.

  10. duncanseconomicblog said, on December 10, 2009 at 3:14 pm


    10 yrs yield 3.80% today. They yielded over 4% in june. They yielded over 12% in the last recession.

    What wager are you thinking?

  11. Daniel said, on December 10, 2009 at 3:25 pm

    Yes, they yielded over 4% in June for a day as a rumour went around that QE might end early.

    When QE start in March, they yielded 3.6%. Now they yield 3.8% AFTER 185bn of purchases by the BOE. That is equivalent to about 15% of the ENTIRE debt stock of the UK, and they STILL can’t rally.

    Sure, they yielded 12% in the last recession. But the base rate wasn’t 0.5% and inflation wasn’t zero either. The REAL yield in the last recession was lower than the current REAL yield. My chart only goes back 10y so I can’t give exact figures, sorry. The point is, you are not comparing like with like.

    As for the wager, something along the lines of a spreadbet? I bet yields will be higher than the start point, come close of business on a particular date in the future. 3 months or something.

    • duncanseconomicblog said, on December 10, 2009 at 3:38 pm


      Most global asset allocators see G7 sovereign bonds as an asset class.

      A previous post of mine – will dig out link – has real yeilds going back to about 1880. Think the chart is soured from the IMF.

      In terms of such a spread bet. I think bonds offer decent value at the moment. A lot depends on how other assets do – if the FTSE wobbles, expect bonds to be bid.

      I shalll think up some terms and email you.

  12. freethinkingeconomist said, on December 10, 2009 at 10:23 pm


    What do you think of this FT story?




    • duncanseconomicblog said, on December 11, 2009 at 9:08 am


      Let’s see where yields are in a month’s time!


  13. freethinkingeconomist said, on December 11, 2009 at 9:30 am

    You’re right – and 4% for 10 year borrowing is still marvellous. But some of the other headlines are getting nervous. May blog on this.

  14. Liam Murray said, on December 11, 2009 at 11:22 am

    A quick observation & a question.

    Your original post Duncan, typed minutes after the Chancellor finished speaking, was premised on the idea that instant market reaction was important. You had a powerful visual that let you make a nice partisan point – fair enough.

    In the course of the discussion that followed – and with another couple of days market data to inform it – you’re position shifted and we had everything from ‘let’s see where yields are in a month’s time’ to references to charts going back to 1880! Again, all interesting stuff but there does appear to be a fondness for whatever data series or timeframe suits your politics rather than an objective analysis of them all.

    My question is simply this – can we all see the terms of the wager with Daniel? Would be interesting to follow…

    • duncanseconomicblog said, on December 11, 2009 at 11:30 am


      Fair comment as ever.

      I usually argue and stand by the reasoning that long term trends matter more.

      My point yesterday was to show instant market reaction to the PBR.

      Trying to phrase a wager at the moment. I shall share the details.

      • Liam Murray said, on December 11, 2009 at 12:05 pm

        Thanks Duncan. The reason you’re on the blogroll is there are far too few people who understand economics among the myriad of those happy to comment on it. This is one of those very rare blogs where you actually learn something most times you visit so please take any comments of mine in that context!

        I was prompted to revisit your PBR post by that piece in this morning’s FT (http://www.ft.com/cms/s/0/7f9daa08-e5ce-11de-b5d7-00144feab49a.html?nclick_check=1) and the gossip about differences between No.10 and Darling over how explicit to be about tackling spending (http://news.bbc.co.uk/1/hi/uk_politics/8407318.stm). Whatever the detail of how that played out I think Darling should’ve been far, far more explicit about spending restraint and that electoral considerations prevented him from doing so. I also think whichever party triumphs next spring we’ll need to see something far more radical pretty soon (let’s hope not quite as radical as Ireland needed).

  15. Daniel said, on December 11, 2009 at 1:12 pm

    Hi again,

    I’ve not heard anything from you yet about the wager, but I would suggest a simple spread bet, with expiration at some point in the medium (3 months ish) future, to eliminate some short term noise yet not have the bet last forever (and thus be pointless).

    I see you were an economist at the BOE and are now a fund manager, so we are both “proffessionals”, so I see no harm in doing this trade for some friendly amount.

    • duncanseconomicblog said, on December 11, 2009 at 1:43 pm


      10yrs to yield less than 4% at end of Feb?

  16. Is this the end? « Freethinking Economist said, on December 11, 2009 at 1:21 pm

    […] Duncan may have gone off too early. If he is looking for a bet, why not just buy this Gilt Future (disclosure, I used to work at IG). […]

  17. newmania said, on December 11, 2009 at 1:54 pm

    Fight fight fight ….hey can outside punters take some of this Duncan action up?

  18. Daniel said, on December 11, 2009 at 6:01 pm

    No Duncan, you can’t skew the market 20bp in your favour before you take a spread bet. If you are offering me an odds bet that is different, but you aren’t.

    Pick a particular bond or future, and lets do a straight 3 months contract at X GBP/bp, spot starting. As in, the starting price of the contract is where the current market is today.

    • duncanseconomicblog said, on December 14, 2009 at 10:36 am


      What odds will you offer that the ten year will yield less than 4% at the end of Feb?

  19. Daniel said, on December 14, 2009 at 1:27 pm

    Being realistic, we are never going to agree odds. That is why I suggested a spreadbet, where odds don’t matter.

    I’ll try and find out what the option price for a 4.00% put ending Feb is, so I can calculate some odds from it, but at a guess, you won’t like them.

    That said, I know why you don’t want to take the spreadbet – you’re unsure about this trade now, and you are hoping to get something like evens with a 20bp headstart!

    Man up! Spot starting spreadbet, straight 3 months duration.

  20. Daniel said, on December 14, 2009 at 1:29 pm

    Put it another way;

    Why would I trade with you at a worse price than I can trade with the market?

    • jdc said, on December 14, 2009 at 2:01 pm

      To be absolutely pedantic, we’re not talking about current 10-years are we, we’re talking about 10-years for a ten year period starting in just over two months. Given that we’re currently in a period of exceptionally low short-term interest rates, and that period is very unlikely to last 10 years, one would assume the rate will shift anyway.

      What I mean is (all else being equal, obviously this is simplistic) if you expect a year of 1% rates and then 9 years of 5% rates, a 10-year would, if equivalent, would be offering 4.6%. On the other hand, six months into those 10 years, if you assume the 6 months at the end will be like the 9 years not like the one year, then an equivalent 10 year would have shot up to 4.85% without that meaning a change in the market’s perception of anything other than how close we were to the end of extraordinary policy.

      So Duncan is not being as selfish as it appears in not wanting to choose today’s spot price as the mid-point for a spread.

  21. duncanseconomicblog said, on December 14, 2009 at 1:41 pm


    Equally – why bet with you when I coukld just buy some gilts?🙂

    My point is not that gilts are a must buy right now. It’s that we are not likely to face any sort of gilt crisis in the next three months.

    Given that, I’m suggesting 4.00%. I’m certainly not saying that gilts can only go up from here.

  22. freethinkingeconomist said, on December 14, 2009 at 1:52 pm

    As an ex-spreadbetting options market-maker, I think the problem is that the gilt options market is a bit sh*te. Duncan is not saying so much that 3.80% yields will fall: just that he would not be a buyer of calls on the yield at 4% or 4.5%. I wish I had a good curve to indicate what they are at present.


    is some help. But it is only about the shorter rates. 1 year rates are 1.72% on 11 Dec, and were 1.62% on 1 Dec – but 2% on 1st Nov.

    If a crisis were taken to be the 1 year rate hitting 2.9%, then there seems to be a 15% chance. Which, to me, seems too high. I’d sell that call.

  23. Daniel said, on December 14, 2009 at 3:23 pm

    ok where to begin;

    Duncan – the bet i am offering you is exactly that – you would be effectively buying Gilts at todays price. The only difference is that we would choose a date to settle up in advance.

    We might not suffer a Gilt crisis in the very short term, but the way the Labour party is going, we will eventually. Hence the bet.

    jdc – I’m sorry, but you are totally wrong. We are talking about where you could buy/sell a current 10y yield, and are betting on where that will be in 3 months time (market expectation = price). We are not betting on the future path of interest rates over 10y. The current price of an asset includes amongst other factors, a reflection of the expectation of future prices – i.e. the yield curve.

    fte – Duncan was saying originally he thought Gilts were good value. He seems to be backing up on this assertion though. A spread bet is fairer as it requires no agreement on, or judgement of any kind of odds.

    • jdc said, on December 14, 2009 at 3:30 pm

      I disagree. Duncan, at least, is very clearly talking about the 10-year gilt at any given point. Frankly so are you, otherwise when you say “when QE started the 10 year was yielding” etc etc, you should be comparing it with a gilt expiring in 9 years as of now, not the current 10 year.

  24. Daniel said, on December 14, 2009 at 4:25 pm

    OK disagree all you want, but it doesn’t make you any less wrong. Those charts all use a generic 10y Gilt. Not a specific Gilt (in which case you would be right).

    So, when I do my comparisons, I am comparing like for like.

    • jdc said, on December 14, 2009 at 5:06 pm

      “Those charts all use a generic 10y Gilt. Not a specific Gilt”

      That’s my exact point. When you compare the yield on a 10-year gilt today with the yield on a 10-year gilt in March, you are including market expectations about a different period in the future in one case but not in the other. Specifically, April to December 2019.

      The same goes for the proposed bet with Duncan. A 10-year gilt today includes the inevitable period of ultra-low short term rates in January and February in 2010. By the end of February that time will be banked, and replaced in a 10-year horizon by the introduction of January and February 2020.

  25. Daniel said, on December 14, 2009 at 6:45 pm

    OK, oo, tell you what, you hold on there for a sec whilst i get my crystal ball out and determine the effect on bond yields of the months dec-feb 2019.

    I’ll tell you for free (given the yield curve is pretty smooth). None at all.

    Time does not change bond yields. Information might, like for example, a terrible budget deficit. QE can change them (supply/demand). Legislation can. Base rates obviously affect them. Inflation can, as can currency strength. Time in itself, as the only variable, can NEVER affect bond yields.

    Your logic is confused, because in a years time, a 10y Gilt, is now 9y paper. That has a different yield, sure, and will sit somewhere else on the yield curve. If nothing else has changed though, the 10y Gilt will still be at exactly the same yield.

    Remember – 10y yields are not divroced form short rates as you suggest – of course people are taking that into account when they price them. All it says though, is that if 10y Gilts can’t rally when short rates are so low, they are never going to…..probably because Labour have handed the country roughly 1000bn GBP of debt before budgets can be balanced. People are asking a higher risk premium for UK debt, because of the shafting Brown has given the UKs finances.

    • jdc said, on December 14, 2009 at 10:25 pm

      You appear to be saying over and over again that I’m wrong, at the same time as explaining why I’m right. I therefore give up.

  26. Daniel said, on December 15, 2009 at 12:59 pm


    OK re-read what I’ve said. You are *totally and utterly* wrong. The generic 10y Gilt today is THE SAME as the generic 10y GIlt in 3 months time. It will also be DIRECTLY comparable to the generic 10y GIlt in 10y time, or whatever else.

    A SPECIFIC Gilt, with specific maturity, will NOT be the same from one day to the next, as it rolls down the yield curve. That is what you, in some half-assed fashion, are talking about. Not that it seems you understand what a yield curve is, or the maths behind it.

    The bet I was trying to make with Duncan related to the 10y Generic Gilt yield. Which is what all the charts show.

    I’m not sure what you do for a living, but seriously, if you are going to pick an argument about bonds and yield curves, don’t do it with someone who has been trading and modelling them and their more complex derivatives for the last decade.

    • John said, on December 15, 2009 at 1:23 pm

      No it won’t. I can’t believe you are allowed to work in the industry if you don’t understand something as basic as this. You don’t have a crystal ball, right? Therefore you assume that the distant future will be ‘about average’ – it’s all you can do in a situation of uncertainty. That’s a large part of why the short end of the yield curve swings around so much during the economic cycle, whereas the long end doesn’t move as much. The present is, for most practical purposes, more certain than the future, that is simple enough and I presume you would accept it.

      The 10 year gilt today includes an expectation of exceptionally low short-term rates continuing for a little while. Those don’t determine long rates, but they are a part of the factors which feed into them. The attractiveness of a certain yield over 10 years is greater the longer it is the case that the yield over a shorter time frame is even lower. The further we get out of that expected period of short term rates, the smaller a percentage of the 10 year period is composed of that period of time (fairly obviously). Therefore we can, initially, expect the 10-year yield to rise when short term rates rise. Otherwise there’d be huge arbitrage in the markets whenever anything significant happened and long rates didn’t respond.

      Or, to put it more simply, the market will demand a lower yield for a ten-year period which contains a period where there is a known risk of recession and deflation which are assigned a high probability than for a future period where the assumption is a return to trend growth. If you think I’m wrong, then I bet you £200 index-linked that 3-month T-bills have a higher yield in December 2019 than they do today. Since neither of us have a crystal ball you must think it’s an unknowable and therefore roughly 50/50. The odds of 2 to 1 that I am going to offer you should, therefore, look like a steal. Are we on?

      • John said, on December 15, 2009 at 2:18 pm

        Jeremy Warner makes the same point as me, but much more eloquently and concisely, in today’s Telegraph. The key quote being “If short-term money is priced at virtually zero, it will inevitably force the price of long-term money lower too, however irrational this might seem.”

        As short-term rates rise as confidence grows that the recession has ended, long rates will go up too.

  27. Daniel said, on December 15, 2009 at 2:41 pm

    John, jdc and whoever else.

    The yield curve is constructed from tradeable bonds. The price of a bond might contain a lot of other information in it, but ultimately it is the price someone is willing to buy/sell at which determines the yield.

    You are correct in saying short rates are more volatile that long rates, for the reasons you give. I don’t dispute either that short rates will effect long rates. BUT they don’t always the way you say. In fact, a hike in short rates often can lead to long rates falling, as people expect inflation to fall and better financial governance (tightening).

    What you call an arbitrage is commonly reffered to as the carry trade. Borrow short, lend long. There is no arbitrage there. No one knows the path of rates, so there is more risk involved in buying (lending) long. Hence you are compensated more for it (at the moment).

    I am making no direct observations about the future at all. I AM making observations of TODAYS prices of assets.

    As for your bet: well, it really proves the point that you just don’t know what you are talking about.

    I don’t know where rates will be in 10y time. I DO know where 10y Gilts are NOW, and Gilts around that point. From that (using 10y and 11y Gilt prices), i can find out mathematically where the price of a 3m T bill, 10y forward is. With the curve upward sloping, my rough estimate will be around 3.95%, but feel free to work it out accurately. I don’t have a crystal ball. I DO have current market prices for long dated instruments, and from those I can work out where the market CURRENTLY prices rates 10y hence. NOT where they are going to be….but where we *think* they are going to be.

    I’ll be happy to take a 2:1 bet around the 3m rate 10y forward. I doubt you will though….

    • jdc said, on December 15, 2009 at 7:17 pm

      Same person, different computer.

      You have done an excellent job, yet again, of proving my point. You can construct the forward rate from gilts which are in the future. Do that exercise for 10-year gilts from the beginning of March. Is it the same as the spot price today? I bet it isn’t. Which is the original issue regarding whether Duncan was being unfair in wanting to centre his bet on a 4% yield or not.

      I agree that the long rate might fall *if the short rate goes up faster than market expectations because of a policy decision*. That’s very different from what it will do if the short rate rises *in line with current expectations*. So as the bet goes, as I believe a great man once said, I’m in if “the starting price of the contract is where the current market is today”. For 3-month bills today, not synthetically constructed future ones😉

  28. […] of course, explains the ‘as Darling spoke graph provided last week by Duncan, but which information was conveniently ignored in the ensuing days as commentators rushed to […]

  29. Daniel said, on December 15, 2009 at 7:50 pm

    OK jdc. You really are a tool. You clearly have no concept of how bonds/yield curves work.

    Let me illustrate. Again.

    The bet i was doing with Duncan was for a 10y Gilt. Generic. In 3 months time, we will still be looking at a straight 10y Gilt. It is as if the bond is constant maturity (because it is a generic bond).The yield curve is already priced into the bond – the market knows short rates are low.

    IF we took a particular 10y issue, in 3 months time it would be a 9.75y issue. The yield curve is upward sloping, so if nothing else changes, you would expect the yield of the bond to roll down the curve. So really, by your logic, it should be ME getting compensated, not Duncan.

    Neither of these situations are forward starting. The Gilts I am constructing forward rates from ARE NOT IN THE FUTURE. They are well defined (cashflows), and SPOT (i.e. today) STARTING.

    Let me just say that again.


    Go find a price on a 3m Bill or Future starting in around 10y time. You will find its somewhere near 4%. Probably higher. Not near 0.5%.

    If you were right, you would be able to go into the market and sell 3m futures starting in 10y at 0.5% and buy 10y Gilts at 4% (a la Duncan).

    But you are not right – I am. The very fact that you can’t get the prices you are talking about in the market, yet I can, shows it.

    • jdc said, on December 15, 2009 at 8:01 pm

      The fact that you can’t get those prices for the future was *precisely my point*. People who do not have crystal balls nonetheless make markets based around the future, and they guess at it being “roughly normal”.

      If your bet with Duncan had been around a fixed maturity, I agree that is already priced in. It was not. It was based on what 10-year gilts would yield in three months’ time, not what 9.75-year gilts would yield.

      You can construct a synthetic 3-year in the future and find that it yields around 4%, evidence that the market does not thing the future will resemble the present forever. Similarly then, the market does not expect that the period March 2010 to Feb 2020 willl be exactly the same as the period December 2009 to November 2019 – it thinks it will include one more short period of ‘normal’ rates, and one fewer short period of ‘abnormally low’ rates.


  30. Daniel said, on December 15, 2009 at 8:45 pm

    OK, you are pretty moronic. Your point keeps changing, as badly articulated as it is. As far as I can tell, you don’t have one baby-shits worth of a clue as to what you are talking about.

    The chart you see of 10y Gilts is a GENERIC construct. It is as if a new 10y Gilt was issued every day. Hence, the yield of TODAY’s 10y is DIRECTLY comparable of the same in 3m time. It is the SAME instrument. It has NO specific maturity.

    As for your markets “guessing” that things will be roughly normal in the future. No, all forward prices are constructed from SPOT rates, which themselves come from MARKET PRICES. For example, the fwd rate for 2010-2020 is created from the 10y price and the 20y SPOT prices. The fwd price is a PURELY mathematical construct. It does not “think” anything about rates.

    But lets stop this argument now. I have no idea what you do for a job, but I can tell that it isn’t in any rate markets. I’ve worked at 3 different banks on several different rate desks, including sterling, and frankly, I’ve had interns with a better knowledge of bond maths than you.

    • jdc said, on December 15, 2009 at 8:56 pm

      Good grief. People buy or sell these things based on their expectations. Your “purely mathematical construct” is constructed from people deciding to buy or sell the 10-year and the 20-year and whatever might exist inbetween, it’s not invented from first principles in a university maths class. If it came out with a rate which was very far at all from the expectations of the average market actor, there would be a wave of buying or selling. This is unbelievable, it now turns out you actually don’t even know how a market works, never mind a bond…

      It doesn’t matter whether the 10-year is a “generic construct” or not. The 10-year “construct” starting at the end of Feb covers *a different ten year period* from the 10-year “construct” starting today. Do you get that yet? It is not the same ten year period. It contains different economic events, therefore interest rates are likely to be different. Only slightly, because 9.5 of the 10 years are identical, but enough that when .25 of those years are seriously abnormal it is likely to make a difference.

      Otherwise, a thought experiment. I will offer you another bet. £100 at evens that the 10-year will yield over 4.15% on 1st December 2012. There you go, I’m giving you a head start of over 25bps over the spot price. And as you say, it’s a generic construct – it’s therefore DIRECTLY COMPARABLE. Or is it, perhaps, not directly comparable for some reason?

  31. Liam Murray said, on December 15, 2009 at 8:51 pm

    I love this blog and – to begin with anyway – this thread started off very informative & educational.

    Now it has the dubious honour of being the first post I’ve ever unsubscribed for… get a room please.

  32. Daniel said, on December 15, 2009 at 9:45 pm

    Re: thought experiment. Am I buying or selling bonds here? You aren’t clear. Remember, I am a *seller* expecting yields *higher*. Not sure where you get 4.15% from either….your forward price should be calc’d from 3y and 13y Gilts. The point is though, you are NOT trading the same thing I am. You keep failing to see that.

    Regardless though – If i hedge with my generic Gilt, it will settle in 3y time against….the generic 10y Gilt price. QED.

    So yes, it does matter if it is a generic construct. If nothing else happens in between, the 10y today will be indentical to the one in 3m time. Its only events that have any effect, NOT time itself. What you keep reffering to is new bond vs old bond. Which would be different.

    As for forward prices, i’ll tell you again. You calculate them PURELY from 2 spot prices. It IS purely mathematical, to the point where odd shaped curves throw up some very strange forward prices. The spot prices themselves are determined from the market, and all the things that drive it. Simples, yes?

    I am right in saying you don’t work in the markets aren’t I?

    • jdc said, on December 15, 2009 at 10:04 pm

      You are correct, I work in Parliamentary lobbying. Although over the course of the last two years I have made five bets of any substantial amount with people who do work in the markets, about market events, and I have won four of them. I don’t think we disagree as much as you think we do, but to summarise;

      * The 10 year you buy today will yield the same until maturity if you hold it until then. Agreed (well, in reality of course it will yield a different amount from the imputed yield and you’ll either overpay, or get a windfall at the end, but that would also be accounted for all else being equal if you sold it in the interim). That’s a fairly indisputable fact – its price may change if you wanted to buy it again ab initio at the shorter maturity, but that’s down to a whole other set of factors, of which time is not one. It’s also not what people generally mean when they talk about the change in the yield on 10-year gilts.

      * The 10 year yield ‘in general’ starting at whatever day now or in the future you pick is the one where we disagree. I think time alone can shift it – the market has expectations of a period of exceptionally low interest rates and indeed Quantitative Easing, and an expectation of how long those will last. The closer we get to the end of those, the more (assuming the market is proven right and there are no major policy surprises) the curve will flatten, but it will flatten by going up across the curve, just up by more at the short end than the long end.

      So, the specifics.

      “Not sure where you get 4.15% from either….”

      It was an example picked from thin air to be 0.25% or thereabouts above the current spot price, to demonstrate that the spot price, for an item in the future, might not be the appropriate centre point on which to base a spread bet. This is true across a number of markets other than bonds, from “Consumer Price Index” to “goals scored”.

      “your forward price should be calc’d from 3y and 13y Gilts.”

      Exactly!! And your forward price in your spread bet with Duncan (which, pace the above, *cannot* be on the fixed maturity because as you say it is as of right now and its yield is therefore technically unchangeable, except insofar as some of the profit compared to short rates would have been banked by March), so you need to calculate it from 3m gilts and the implied 10.25 year price, not take the spot price right now. Because 3m gilts are yielding very little and the implied 10.25 year price is nearer the long-term average, the 10-year starting March (I should have saved this to cut and paste it) will have a higher anticipated yield than the 10-year does right now.

  33. Big fight report « Though Cowards Flinch said, on December 16, 2009 at 2:36 pm

    […] But no blog fight to date has measured up to this one over at Duncan’s place.  […]

  34. Daniel said, on December 18, 2009 at 4:20 pm

    OK jdc, last post. You are getting closer to what I was trying to explain to you, but you still don’t have it quite right. I think you are mixing up yield (which is just another word for price in bondspeak) and return on investment (ROI).

    Lets do a thought experiment.

    We have an upward sloping yield curve, and 10y Gilts yield 4%. The DMO issues a new 10y Gilt with a 4% coupon (i.e. at par). Let us now journey forward in time and in this particular experiment, nothing at all happens in the world for a year, and the yield curve is totally unchanged. Our bond is now a 9y bond, and yields 3.80%. The DMO issues a new 10y with another 4% coupon….and the yield is again 4%.

    The new bond represents the generic 10y people chart. It is contructed as if there was a new bond issued every day (in fact, its more like an IRS but…). My bet with Duncan shouldn’t take into account any forward prices. If we do it with the generic bond, we are looking at like for like. A specific bond (like our 9y) actually works in his favour (in an upward sloping yield environment, as we have at the moment) because all things being equal, if nothing else happens, as the bonds’ maturity shortens, its yield will fall down the curve (in his favour). Both ways, he shouldn’t be getting a headstart. You are simply barking up the wrong tree when talking about forwards in this context.

    Both bonds take into account the period of low interest rates at the front of the curve. People know that front end yields are low, so it is implicitly priced into longer dated bonds. I’m mot entirely sure what you’re getting at, but I think you are trying to say something along the lines of “we know short rates can only go up from here so you have to compensate me for it in the price of a long bond.” Well no, I don’t. The market won’t either. The price of the 10y is where it is. The market has come to that price after all factors are considered (short rates, inflation, govt policy, supply demand etc). That might change every day with new information, but at ANY GIVEN TIME that is the market’s best collective idea of where the compound yield over 10y will be.

    If 9y yields are 3.80%, a bond with a 4% coupon will trade ABOVE par (say at 105). A 9y bond with a 3% coupon will trade BELOW par (say at 85). BOTH bonds offer a yield of 3.80% though…but your ROI for buying either will be the same (this also explains why people use yield instead of price to describe the “price” of a bond – it allows better comparison of like maturities). One has a bigger coupon, but costs more upfront. The Yield of these bonds will change naturally over time, as they fall down the yield curve, even if, as in the experiment, nothing else happens. What you have done though is locked in your ROI.

    To sum up. The yield of a fixed maturity asset like a bond ALWAYS changes (unless the yield curve is perfectly flat). Every day. The cash return you might get for holding such an asset is fixed on the day you buy it.

    • jdc said, on December 18, 2009 at 9:31 pm

      Yes, I’m talking about the yield to maturity – I assumed we all were, because as you say the running yield is fixed on the day you buy it. On the rest, we’re not going to agree. You say that the price is where it is and will change based on new information, and you’re right, if you were spread betting based on what the yield of a 10-year today will be in three months when it’s a 9.75 year, that would be the most important point – surprises in either direction vs the passage of time.

      If, however, you are betting on what a 10-year now will be compared to a 10-year in three months, then the factors you cite – short rates, inflation, etc, can be relevant even if they are entirely in line with market expectation – the market has an expectation about how long short rates will last and therefore an anticipation of the 10-year in March 2010 that is different from the 10-year now.

      I am convinced – 110% convinced – that anyone talking about comparing changes in the yield on the 10-year for the purposes of a spread bet is talking about constant maturity – 10 years now, and 10 years at the close of the contract, not talking about the specific notional gilt as it travels down the yield curve.

      To pick a simpler example which makes the case more starkly, suppose the market expects short rates to stay at 1% for 10 years. The 10-year would probably trade around 1%. Now suppose it expects short rates to power up to 20% for the 10 years after that. Would you take a spread bet on the basis of a 10-year right now, centred on the current price, as against a 10-year in five years’ time? Of course not. Fair enough the current situation isn’t as extreme, but it’s logically the same.

  35. Daniel said, on December 18, 2009 at 11:20 pm

    Paragraph 1: The price/yield of a bond will change over time even if NOTHING happens. If nothing happens to the yield curve a new bond of the same original maturity will have the same yield as the old one.

    Paragraph 2: Like I say, expectations of short rates etc are priced into the value of the bond. A bonds value is determined by traders buying and selling it. Traders are not stupid. Fair value is where a bond last traded. You are trying to look into the future to tell me the price of a bond isn’t fair. I am trying to tell you that people have already done it and come to an agreement (a trade) about where it is fair. You might disagree, at which point you want to buy or sell the bond today. BUT there is no “anticipation” or forward pricing of anything involved.

    Paragraph 3 is what I’ve been trying to get into your skull for ages. The bet i was trying to make with Duncan was based on the generic, or constant maturity Gilt for simplicity.

    Paragraph 4: Your example is nonsense unless you happen to have a crytal ball. Even if the real path of rates was as you described, the yield curve would never look like that because you don’t know what is going to happen in anything more than a probabalistic fashion. So whilst rates are low now and might stay low, the 10y would never be at 1% because you can never predict the true path of rates with any certainty. See my explanation of paragraph 2 again.

    • jdc said, on December 19, 2009 at 12:38 am

      “Like I say, expectations of short rates etc are priced into the value of the bond. A bonds value is determined by traders buying and selling it. Traders are not stupid. Fair value is where a bond last traded. You are trying to look into the future to tell me the price of a bond isn’t fair. I am trying to tell you that people have already done it and come to an agreement (a trade) about where it is fair. You might disagree”

      I don’t disagree at all. I just don’t believe that a trader would, always and everywhere, agree that a fair price for a bond, today, maturing in December 2019, is necessarily a fair price at which to agree in advance the purchase of a similar bond, in March next year, maturing in March 2020. That’s not complicated and I indeed so obvious that I don’t see how it can even be in dispute.

      While the future is uncertain, it is stonkingly likely that short rates between March 2010 and 2020 will be higher on average than between December 2009 and December 2019, because, given we know nothing about Jan, Feb and March 2020, on balance they are likely to be ‘average’, whereas we know quite a lot about Jan Feb and March 2010, they are likely to be spent around the ZIRP zone.

      The yield curve *now* is an indicator (in large part) of what the market expects the *yield* to be during the future, not an indicator of what the market expects the *yield curve* to be at a given point in the future.

      “expectations of short rates etc are priced into the value of the bond”

      In fact, it’s more than that – the absolute nailed-on certainty of short rates is priced into the value of the bond. It will *no longer* be priced into the value of a bond of equivalent maturity in three months’ time, because we will be three months closer to the end of the period of short rates.

      You don’t need a crystal ball to know that – I am a year closer to my death than I was last Christmas. That’s not a statement about when I will die, I haven’t got the faintest idea – but it’s none the less actuarially valid when pricing up the cost of selling me life insurance.

      So if you were betting on me dying “in the next ten years”, then the specific gilt (John to die by December 2019) would be getting worse value each day that I do, in fact, stay alive. In contrast the generic gilt (John to die at some point in the next ten years) would get gradually more expensive with the passage of time.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: