Interpreting Lognormal and Normal Trading in Swaption Volatility

Summary

The document explains a market comment that short-dated, front-end swaption volatility has traded more lognormally than longer-tenor tails. It describes a practical interpretation based on regressing changes in implied volatility against changes in interest rates. A positive, statistically significant relationship—rates and implied volatility tending to rise together—is taken as evidence of more directional, lognormal behavior. Weak or absent correlation is associated with a more normal relationship.

Applied to the quoted observation, the short end of the volatility surface is described as more responsive to rate moves, while longer tails such as ten-year to thirty-year swaptions are less so. This is a brief explanatory answer rather than a full account of normal and lognormal pricing models. The regression is offered as a heuristic for interpreting market behavior; the document gives no data, sample period, model specification, or evidence establishing that this relationship holds generally.

Key ideas

• A regression of implied volatility changes on rate changes can help interpret the volatility regime.
• A positive and statistically significant relationship is described as more lognormal behavior.
• Weak rate-volatility correlation is associated with more normal behavior.
• The stated contrast is between the front end of the swaption surface and longer-tenor tails.

Tags

# Swaptions vol trading lognormally


# Swaptions vol trading lognormally












What does this mean:

```
"Front-end vols have been trading lognormally while longer tails have traded normally."
```

I read this in a research report, in the context of swaptions, which does not explain this further. My limited knowledge tell me it has something to do with correlation of vol with rates but I'm not sure. Could someone please help?

## Answer by Helin (score 3, accepted)

https://quant.stackexchange.com/a/12698

Typically, strategists run a regression of changes in implied vols against changes in rates. If rates are highly directional with implied vols (regression coefficient is positive and statistically significant), then it would imply a more lognormal relationship. If the two series are not correlated or very weakly correlated, then the relationship is considered more normal.

So what this guy is saying is that the "top-left" corner of the vol surface has been more directional with rates (rates go higher, vols go higher), while longer tails (10y-30y tails) are less directional.

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.