I'm using classic Hull-White model for short term interest rate dynamic: $$dr(t)=[\theta(t)-\alpha(t)r(t)]dt+\sigma(t)dW(t)$$ (Notation is quite intuitive, anyway I am using the same as Wikipedia whether any doubt should occur).

I would like to have some feedback from users about the price sensitivity to constant volatility of a callable fixed rate bond priced under such model.

Let I have a fantasy bond with following features:

  • termination date $=$ 1,825 days starting from today
  • coupon payment tenor $=$ annual
  • face amount $= 100$
  • redemption $= 100\%$
  • coupon $= 3\%$ of face amount

That above is a callable bond whose call schedule says it can be called at $100$ every year at the same date of maturity, hence we have $4$ scheduled call dates (e.g. if maturity were 28-Sep-2020, we would have following call dates: 28-Sep-2016, 28-Sep-2017, 28-Sep-2018, 28-Sep-2019).

Hull-White model parameters:

  • $\alpha=0.03$
  • $\sigma=132\%$

For the sake of this example, term structure employed in Hull-White model is a snapshot of (linearly interpolated) current euro swap curve - which I attach here below:

-0,135%     0D
-0,145%     1W
-0,113%     1M
-0,040%     3M
0,029%      6M
0,012%      11M
0,025%      1Y
0,029%      18M
0,051%      2Y
0,122%      3Y
0,230%      4Y
0,354%      5Y
0,484%      6Y
0,619%      7Y
0,748%      8Y
0,860%      9Y
0,963%      10Y
1,058%      11Y
1,142%      12Y
1,331%      15Y
1,481%      20Y
1,516%      25Y
1,520%      30Y
1,523%      35Y
1,522%      40Y
1,498%      45Y
1,475%      50Y

No credit risk!

Regardless of days conventions, day counters and trinomial tree engine grid size - which should account just for small variations in final results - my questions are:

  1. what is the clean price Hull-White model is supposed to return with such parameters (mean reversion, volatility and term structure)?
  2. Is it possible that halving volatility has a huge impact on clean price - about 15 ~ 20 points?
  3. If volatility drops from $132\%$ to, say, $7\%$ does clean price go from very small default-like values (below $10\%$) to "normal" values (around $100\%$)?
  4. If (2) and (3) were correct, how would you qualitatively justify such huge volatility sensitivity of the clean price?

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.