If one imposes a form $r(t) = \text{...}$ on the cc. short rate, and aims to fit the short end of a SOFR (or another modern RFR) using futures, how would one best go about this within a "curve-fitting" paradigm e.g. least-squares on fitting error a la fitting a bond curve by minimizing error of $\sum_b (\sum_i c_{i,b} \exp(-\int r(s)ds))$ (where b is a bond, c a cashflow).

I have seen this done by converting a 3m compounded SOFR quote into a forward quote, how does that sound in principle (convexity adjustment aside)?

It should be noted that the form $r(t)$ would be chosen to answer a specific question (it could be piecewise flat around meeting dates, it could be a smooth hump around the cycle) — i.e. the question is how to best do it for an arbitrary but well-behaved function.



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.