SERVING THE QUANTITATIVE FINANCE COMMUNITY

 
User avatar
BerndSchmitz
Topic Author
Posts: 242
Joined: August 16th, 2011, 9:48 am

SABR vs Heston in rates modelling

September 9th, 2015, 12:36 pm

Hi,sry if this questions has been asked before but I'm wondering why nobody seems to use the Heston model for rates. I mean the model has nicer features than SABR (less explosive) and can be solved in closed-form (and that exactly and not only approximative).So is the Heston model not flexibel enough to match the smiles? Or is the solution simply too slow? Or is there some other drawback?Thanks,Bernd
 
User avatar
Jhammer
Posts: 37
Joined: February 24th, 2014, 8:05 pm

SABR vs Heston in rates modelling

September 9th, 2015, 1:30 pm

 
User avatar
mtsm
Posts: 352
Joined: July 28th, 2010, 1:40 pm

SABR vs Heston in rates modelling

September 15th, 2015, 2:25 pm

One argument is that the Heston model has some kind of CIR backbone, while the SABR model offers more flexibility.Might be cultural. Closed form if somewhat approximate, not to say wrong, solutions are quite important in market making systems.
 
User avatar
BerndSchmitz
Topic Author
Posts: 242
Joined: August 16th, 2011, 9:48 am

SABR vs Heston in rates modelling

November 12th, 2015, 5:29 pm

Hi mtsm,what exactly do you mean by "SABR offers more flexibility"? I thought you can always choose some external backbone (e.g. MLV/Bartlett) so that the model implied backbone isn't that crucial. Or maybe I don't get something here.
 
User avatar
mtsm
Posts: 352
Joined: July 28th, 2010, 1:40 pm

SABR vs Heston in rates modelling

November 13th, 2015, 1:30 pm

What is MLV??Yes, but you are looking at this the wrong way round maybe. These external backbones were introduced, because the models weren't performing. Vice versa, introducing an external backbone and not using the model implied one, obviously invalidates part of the model. Then at some point the question becomes, why do you want to use a differentially specified stochastic vol model in the first place??Anyway, so if you do want a backbone, then SABR accomodates a continuous range of exponents.
Last edited by mtsm on November 12th, 2015, 11:00 pm, edited 1 time in total.
 
User avatar
rh303
Posts: 11
Joined: March 12th, 2013, 9:28 am

SABR vs Heston in rates modelling

November 26th, 2015, 4:53 am

See this too for why some prefer Heston http://www.wilmott.com/messageview.cfm? ... TARTPAGE=6
 
User avatar
BerndSchmitz
Topic Author
Posts: 242
Joined: August 16th, 2011, 9:48 am

SABR vs Heston in rates modelling

December 7th, 2015, 4:08 pm

@mtsm: The Bartlett backbone is sometimes also called minimum local variance (MLV) backbone
 
User avatar
Pat
Posts: 1207
Joined: September 30th, 2001, 2:08 am

SABR vs Heston in rates modelling

December 10th, 2015, 7:59 pm

Strangely enough, one can analyze the Heston model (or a generalized Heston model with an arbitrary backbone and with underlying/variance correlation), and show that at any given expiry date, the implied volatility smile is given by the same(!) explicit formulas of the SABR models. (The derivation determines the effexctive alpha, rho, and volvol in terms of the native Heston parameters). So if we had initially analyzed that one, we'd be talking about the Heston formulas ...
 
User avatar
BerndSchmitz
Topic Author
Posts: 242
Joined: August 16th, 2011, 9:48 am

SABR vs Heston in rates modelling

December 11th, 2015, 8:10 am

That's interesting! Are there any papers out there on this?
ABOUT WILMOTT

PW by JB

Wilmott.com has been "Serving the Quantitative Finance Community" since 2001. Continued...


Twitter LinkedIn Instagram

JOBS BOARD

JOBS BOARD

Looking for a quant job, risk, algo trading,...? Browse jobs here...


GZIP: On