I know there are various types of yield curve twists and shifts - one of these being the butterfly twist. Is it typically the case that this would need to be either a positive or negative twist? What form does it take if is just a butterfly twist (i.e. not neg or pos)?
- Long and short, - middle, just like a butterfly.
a butterfly flaps its wings in both directions
In my opinion, the main objective of hedging is to eliminate any negative impact and not to make profit from positive impacts.
So, the most appropriate form is the one with least convexity.
Good point Pierre, I think I agree with your reasoning. I may have reasoned differently when I first came across this problem.
I just realized my question relates to butterfly twist s, not twist in the singular, so this would cover both positive and negative…ugh. Lower convexity would be correct form to hedge…
A butter fly twist is a movement in long and short term rates in the opposite direction of medium term rates. If you wanted to hedge against these moves you could go long both a bullet and long bond with unequal weights, maybe 75/25 - since the long bond has larger duration the opposite movements would be off setting. In my humble opinion.
Butterfly twists (a plural) mean that the direction is not certain. When yield curve’s direction is not certain a high convexity gets more valuable, in my humble opinion.
Butterfly twists (plural) mean high interest rate volatility which would indicate wanting to be long convexity or a barbell portfolio. Just my two cents.
Nope. A twist like this leaves an investor with significant structure risk. To minimize this you need low dispersion. (Convexity). Matching duration with low convexity will appropriately hedge the risk but if you have high dispersion, the structural risk would not be appropriate for ALM.
Yeah, but isn’t that only applicable for the single liability immunization? For immunization of multiple liabilities, a convexity should be larger.
Convexity should always be minimized slightly above the convexity of liabilities for immunization. Liability convexity 100, you need 101 but you do not want 150.
Even if you had a barbell structure whose present value of assets rose from the butterfly twist, the Portfolio IRR will fall, indicating structural risk, from the standpoint where even if the PVA>PVL the IRR portfolio
Positive butterfly twists mean something negative --> yields go up at the short end and long end, whereas yields in the middle fall.
Negative butterfly twists mean something positive --> yields go down at the short end and long end, whereas yields in the middle go up.
Isn’t a single liability considered the equivalent of a zero coupon bond, which would make matching the durations a priority over minimizing convexity?
In the real world, you can never match duration exactly so it’s a bit of a tradeoff. You try to match duration “close enough” . What close enough means is based on judgment and whatever other factors you can/can’t match.
I also think using the term “minimize convexity” is potentially confusing/misleading. Even if you know what you are going to do, what you really mean is keeping asset convexity above liability convexity (assuming we are in an asset/liability investment space), but by as little as possible (given other constraints). But, I’m sure there are situations where you might make asset convexity a little less than liability convexity because your other choices make asset convexity “too much” more than liability convexity so the keep convexity close “rule” overrides the keep it above in specific practical situations. (they’re really guidelines more than rules)
What the CFAI would test and what they would expect for an answer in a specific situation is, as with many situations, hard to tell. They say they don’t test you on exceptions, except of course, when they do.
A lot of the theory is based on assuming the bond market is ‘continuous’ and you can get things to work out on whatever matrix of things you decide you want them to work on, but the reality is that the bond market is very discontinuous as you look at credit quality, maturity, on the run or not, duration, coupons, issuer concentration, etc. (why do you think a lot of adjustments are done with futures and not with actual bond trades? Market is too lumpy and illiquid)
Like a lot of things they test, if you really understand an area, you have to ‘dumb down’ how you respond to things and do so in the right way to respond the way that they want. (or at least that’s what I hear from various experts in different practice areas).
Correct, duration is the first derivative of bond price, convexity is second. Hence most price impact is from general rate movements, not from curve steepness or curvature.
Ss 11 reading 23 practical problems 1-6