Using the Kolmogorov–Smirnov Test on Swaption Market Data

Summary

The document asks how to compare the distributions of swaption market observations, such as volatility or liquidity spreads, across two time periods. It considers whether all USD swaption observations should be treated as two samples or whether observations should first be grouped by dimensions such as expiry, tenor, and strike. It also raises the concern that bucket-level samples may be small and make test results difficult to interpret.

No answer or empirical analysis is included, so the document does not establish which sampling design is appropriate or report any test result. The question highlights that distribution comparisons depend on how instruments are grouped and that small samples can limit the strength of conclusions. It supplies no guidance on handling dependence among observations, multiple bucket tests, or whether the standard two-sample test assumptions hold for the market data in question.

Key ideas

• The question concerns comparing swaption market-data distributions across two periods.
• It asks whether to test pooled samples or separate groups defined by expiry, tenor, and strike.
• Small bucket-level samples may make results harder to interpret.
• The document contains no answer, test results, or recommended methodology.

Tags

# Kolmogorov Smirnov test on derivatives market data


# Kolmogorov Smirnov test on derivatives market data












If I want to test market data( say volatility or liquidity spreads) for swaptions in a particular currency (eg. USD) for 2 different periods of time and ascertain whether their distribution is more or less the same, I believe KS test would be a good idea.

What would be the right way to do so? Would all market data for USD swaptions from the 2 different periods be tested as 2 different series? Or would the right way be to bucket them by expiry/tenor/strikes from each period and test at bucket level? If latter, given the sample size would be very small in each bucket, would there be any restrictions on the interpretation of the KS test results?

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.