[Quantlib-users] Calibrating short rate model to cap volatility smile

[Jacek B. <jac...@gm...>]

Hi.
My question is not related to the library itself, but rather to market
practice.
I would like to price RFR caps with some short rate model, e.g. G2. From
what I know (and tried to simulate), neither G2 nor Hull White can produce
volatility smile. However, if you take a look at market cap/floor surfaces
for RFR, they usually have a smily shape. I intend to use short rate model
having Bloomberg as benchmark, who started to use HW 1 factor in their
pricer. Unfortunately, Bloomberg is not disclosing any details on the
calibration procedure they use, nor model parameters, thus my question: how
one can use short rate model to have prices consistent with market vol
surface across all strikes? Does anyone know how Bloomberg does it? Is it
different set of model parameters for each strike?
Any hint would be much appreciated.

Thanks,
Jacek