Page 1 of 1

Rates - Pricing Model Limitations/Issues

Posted: March 7th, 2017, 8:30 pm
by zmfactor
What are some pricing model limitations/issues you have seen, with respect to rates products?

I know this is a very broad question. Just wanted to get a sense of what people have seen.

Re: Rates - Pricing Model Limitations/Issues

Posted: March 8th, 2017, 6:46 am
by Samsaveel
Inconsistent evolution of rates generated by a calibrated model to reality on ground.

Re: Rates - Pricing Model Limitations/Issues

Posted: March 8th, 2017, 9:35 am
by Gamal
They don't bring you coffee.

Re: Rates - Pricing Model Limitations/Issues

Posted: March 8th, 2017, 5:37 pm
by mtsm
Inconsistent evolution of rates generated by a calibrated model to reality on ground.
Second that. Basically all quantitative finance models going back to Bachelier, Samuelson, Black-Scholes and the whole industry on top of that are wrong. Quantitative analysts are a dying breed.

Re: Rates - Pricing Model Limitations/Issues

Posted: March 8th, 2017, 10:49 pm
by zmfactor
Inconsistent evolution of rates generated by a calibrated model to reality on ground.
Second that. Basically all quantitative finance models going back to Bachelier, Samuelson, Black-Scholes and the whole industry on top of that are wrong. Quantitative analysts are a dying breed.
Any specific product types where this is most pronounced? Is this a recent problem with the current markets/rates environment, or a long-running issue?

Re: Rates - Pricing Model Limitations/Issues

Posted: March 9th, 2017, 4:52 am
by Samsaveel
Inconsistent evolution of rates generated by a calibrated model to reality on ground.
Second that. Basically all quantitative finance models going back to Bachelier, Samuelson, Black-Scholes and the whole industry on top of that are wrong. Quantitative analysts are a dying breed.
Any specific product types where this is most pronounced?

Any stochastic model used in finance for pricing and risk management must be in-line with observed reality. Common sense is a key dimension here in the application of models to achieve a market consistent end result

Is this a recent problem with the current markets/rates environment.
Yes, the negative level of Interest rates in some markets requires a model that mimics that behavior. Many posts here on negative interest rate environment and the most up to date modelling techniques that are currently implemented within financial Institutions.

or a long-running issue?

this is a judgement call! this requires a deep analysis into the policy direction of future interest rates.

Re: Rates - Pricing Model Limitations/Issues

Posted: March 9th, 2017, 6:06 pm
by zmfactor
Any recommended articles/books on this topic?

Re: Rates - Pricing Model Limitations/Issues

Posted: March 9th, 2017, 8:46 pm
by Gamal
Any. Start with something basic.

Re: Rates - Pricing Model Limitations/Issues

Posted: March 9th, 2017, 9:34 pm
by mtsm
It's actually more tricky than this... Those who write about this tend to be overly optimistic and always reassure us that financial asset prices follow a random walk, that the market is efficient, etc..., that one can take these as safe starting points to further developments. The proof is often assumed to lie in the work of Bachelier, Samuelson, even Hagan. You can't read about the failure of these models easily in my opinion. 

Besides, their failure has nothing to do with low or neative rates. They never worked actually.

Your best bet would be talk to a market maker in a serious franchise, provided he/she accepts to talk to you. I think that market makers are those that come closest to having a real keen interest in a risk-neutral world. But they also have totally different problems. They often sit on huge inventories of really toxic shit and are moreover subject to real and varying market constraints, liquidity, etc... None of the work I know of addresses this, because I don't think that any quant I know of is aware of the way a desk or book actually works. 

Re: Rates - Pricing Model Limitations/Issues

Posted: March 10th, 2017, 9:06 am
by Gamal
Mtsm, all yield curves predict future short rates. Is this prediction correct? Of course not, it not their purpose. So - why do we expect interest rate models to predict anything?

Re: Rates - Pricing Model Limitations/Issues

Posted: March 10th, 2017, 9:20 am
by Paul
Mtsm, all yield curves predict future short rates.
No they don't. They tell you how much you have to pay now to get something in the future. 
If you want to take that as a prediction of something then go for it! But that's the start of all sorts of confusion and misconceptions.

Re: Rates - Pricing Model Limitations/Issues

Posted: March 10th, 2017, 10:51 am
by Gamal
Mtsm, all yield curves predict future short rates.
No they don't. They tell you how much you have to pay now to get something in the future. 
If you want to take that as a prediction of something then go for it! But that's the start of all sorts of confusion and misconceptions.
Exactly. The same applies to all arbitrage models. All. They have other purposes than prediction.

Re: Rates - Pricing Model Limitations/Issues

Posted: March 10th, 2017, 11:27 am
by Paul
Maybe compare and contrast with weather forecasting?

Edit: I started to make such lists but quickly realised that the similarities and differences are far too subtle to do in a hurry!

Re: Rates - Pricing Model Limitations/Issues

Posted: March 10th, 2017, 2:14 pm
by mtsm
Not sure why you talk about prediction. Did I say anything about that?

Re: Rates - Pricing Model Limitations/Issues

Posted: March 10th, 2017, 3:24 pm
by Gamal
Yes. Maybe not prediction of exact values but at least of the shape of the evolution: random walk etc.

Arbitrage models are nothing more than an a bit tricky interpolation of prices.