Serving the Quantitative Finance Community

 
User avatar
Jericho
Topic Author
Posts: 25
Joined: August 22nd, 2007, 3:14 pm

Bond PnL

May 3rd, 2022, 8:05 pm

Hi All,

Wanted to get your thoughts on Bond Pricing - we have a system which calculates PnL upto the next trade settlement day of the Bond (e.g. T+2) and discounts this back to today using a USD Libor Curve - is this standard?  Surely the PnL should be (Price Today - Trade Price) * Notional / 100 + Carry.  What is the need to add the discounting portion to it?

Thoughts would be welcome as I have not seen this approach up until now

Many Thanks

J.
 
User avatar
bearish
Posts: 5186
Joined: February 3rd, 2011, 2:19 pm

Re: Bond PnL

May 4th, 2022, 12:34 am

This gives me flashbacks to the 90’s. If your entire business settles on a T+2 basis, you might as well do your accounting on that basis, too. If you have a mix of products that settle with different conventions, and you run a very big or very leveraged business, you may want to standardize everything to a common basis. It doesn’t have to be T+0, but that is one reasonable choice. Of course with short term interest rates being zero, it doesn’t really matter, but we seem to be lifting off from that floor as we speak (at least in USD).
 
User avatar
Jericho
Topic Author
Posts: 25
Joined: August 22nd, 2007, 3:14 pm

Re: Bond PnL

May 4th, 2022, 3:57 pm

Thanks Bearish - appreciate the comment - just trying to think of the mechanics of how this would work.  If we have a 10M Bond which shows a price of 100 and was traded at 99, if we use T+0 convention (and ignore FX and Carry) then my PL is simply 10M x (100 - 99) / 100

If we introduce the settlement date approach then is there a corresponding formula you would use in order to calculate the PL for T+2 and discount back to today using a USD Libor Curve?  

Appreciate the help here, still trying to come to terms with this new approach..

J.
 
User avatar
DavidJN
Posts: 242
Joined: July 14th, 2002, 3:00 am

Re: Bond PnL

May 4th, 2022, 4:23 pm

This issue has its origins in now outdated settlement practices, particularly the surely unnecessary 2-day LIBOR settlement lag. Swaps in more enlightened currencies (GBP, CAD, etc.) have no floating rate settlement lag because there is no reason to have one, at least not anymore.
 
I too argued with the accountants about this in the 1990s. They seemed to think it was important to have everything settle on the same day for P&L and other reporting purposes. I see no such need, and I beat back the accountants by throwing their own logic back at them.
 
Allow me to explain… Say you decide to settle a bond on a date different than regular settlement. Unless short-term rates are identically zero your transaction price will necessarily be different than that for regular settlement, meaning you cannot simply observe the market price anymore and must result to using a model to price a what is normally a vanilla instrument. Financial instruments valued using models have higher accounting and regulatory hurdles to clear, literally more diligence to do. Why would anyone want to turn vanilla trading into more complicated trading? End of argument.
 
Mercadian
Posts: 39
Joined: July 24th, 2020, 4:22 pm

Re: Bond PnL

May 5th, 2022, 4:40 pm

.... What is the need to add the discounting portion to it?
This has several points of view, but to me at its core its the simple yet fundamental principle that you need to take into account the time value of money for the future cashflow that will result from the settlement.

Others come from the Finance and Middle Office realm where you have risk policies or regulatory requirements for how to compute your realized PnL, Liquidity GAP,  VaR, PFE/CVA, RWA, so on and so forth, but they can all be traced back to the same basic principle TVM.

The fact that we have a settlement delay is just to give time to Back Offices around the market infrastructure to do their job without putting to much strain on operations trying to avoid fat fingers or operational errors, nowadays seems like a crazy old dated thing, but its just the way the system self organized... that is until the blockchain crowd (not crypto) gets their way with near instant settles, but even then I believe we would start to use the same some form of continuously compounded interest rates for intraday calculations.

Finally think about a case were you bought the bond for 99, and now 2 weeks have passed after you bought it... you have 99 in your nostro/bank/transit account and have a bond with MV of say 100.... if you were to sell the bond again today (2 weeks after you bought it) you would settle in T+2... it would seem incorrect to simply say 100{@T+2} -  99{@T+0} = 1 is your pnl right away (it will be correct two days from now, not today).

Jericho:
....If we introduce the settlement date approach then is there a corresponding formula you would use in order to calculate the PL for T+2 and discount back to today using a USD Libor Curve? 
Again, in line with TVM, I would say the formula is just the present value, future value discount factor relationship.

PV=FV*DF

Now, the choice of discount rate/curve is up to how the bank/desk and how they would think about the liquidity/funding in each currency, the creditworthiness of counterpart, FTPs and all the other fancy adjustments you want to put on it.

Rgds,
M
 
User avatar
Jericho
Topic Author
Posts: 25
Joined: August 22nd, 2007, 3:14 pm

Re: Bond PnL

May 5th, 2022, 8:30 pm

Thank you for all your contributions its great to get all your thoughts and experiences on various topics i really appreciate it