Designing a Tail-Risk Put Sleeve for a Barbell Portfolio
Summary
The document considers how to allocate a limited portion of a portfolio to puts as a tail-risk sleeve. It asks how option maturity, renewal frequency, and distance out of the money affect the strategy, noting that longer-dated options have greater vega exposure. The response says these choices require examining historical option prices and understanding option sensitivities, while recognizing that past market behavior may not repeat.
As a practical reference, it describes a public tail-risk ETF's staggered put maturities and positions that grow with maturity, with strikes around 8% and 12% below the index level at the time cited. These details provide an example of implementation rather than evidence that the structure is optimal. The fund is described as a complement to a base portfolio, so its standalone performance would not establish how the combined allocation behaves. No method for selecting renewal intervals or optimizing strikes is specified.
Key ideas
- A tail-risk sleeve can allocate a portion of capital across put options.
- Maturity, renewal frequency, and strike distance are design choices that need historical option data and sensitivity analysis.
- Longer maturities increase vega exposure, but do not by themselves establish better strategy performance.
- A public fund example uses staggered maturities and puts at different distances below the index.
- Evaluate the option sleeve as part of the combined portfolio with its base allocation.
Tags
Full text
# Questions regarding a “lite/kindergarten” Barbell investement strategy implementation # Questions regarding a “lite/kindergarten” Barbell investement strategy implementation The idea for this question is more or less taken from a slight hint regarding how Universa Investments L.P. functions from Taleb's Antifragile (obviously the real case is far more complex but this is just a home implementation). Suppose you have a portfolio X, where 20% of X gets diverted into put options, more specifically in a 1/N distribution (hinted at in some page in Antifragile). Let's assume that N = 5, then we divert the capital (4%) into 5 put options. What is a good way to determine if it's better to buy the options closer to expiry (and subsequently renew the purchase in set-intervals, and how to determine the intervals) or farther away from expiry and how far out of the money should they be? I know that if I buy them with a bigger TTE (time to expiry), the Vega Greek will be more pronounced and thus the payoff from a change in volitility (on which this set-up is banking on) should be better. Any other tips how to compute, at least in principle, how out of the money the options should be to maximize potential gains (I know this isn't exact science, I would just like to have some dialogue on this topic)? Tyvm for the answers. ## Answer by nbbo2 (score 1, accepted) https://quant.stackexchange.com/a/65485 The design of such a strategy is a complex thing. It involves a trial and error process of looking at historical data on option prices, as well as an understanding of measures such as Vega. All the while knowing that the future will not be exactly like the past. If you look at existing funds that follow such a barbell strategy perhaps it will give you a starting point for iterating your own implementation. Here is the composition of Cambria Tail Risk ETF, a fund designed by Meb Faber who I consider a very good quant and who has published widely (unlike some others who are tighter with information). Cambria Tail Risk ETF (scroll to bottom of page) As you can see they have Puts expiring in 3 mo, 6 mo, 9 mo and 12 months with the size of the position increasing with maturity (0.275 million, 1.2 mm, 3.8 mm, 7mm). With the S&P at 4220, the strikes they own are 3900 and 3700, approximately 8% and 12% OTM as of now. This ETF is a publicly traded vehicle so there is good data on performance. (Of course the ETF is not intended as a standalone holding, it would be held together with a base portfolio and the combined performance is what is relevant).
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.