willthrill81 wrote: ↑
Sun May 05, 2019 3:39 pm
randomguy wrote: ↑
Sun May 05, 2019 2:05 pm
Now the question is how often does the bond tent hurt. If we looked at the say the 1965 retiree with a 7 year bond tend end up with a lower SWR than the 1965 retiree that just held 50/50? I haven't done the math but I expect there are few cases where it actively hurts.
I'm inclined to agree that it's unlikely to help but also unlikely to hurt.
Using a bond tent definitely
hurts. In a kind of trivially obvious sense: anything that makes you hold less equities hurts by giving you a smaller portfolio in the overwhelming majority of scenarios. And the vast majority of real world retirees are going to have some kind
of link between portfolio size and spending, even if it isn't as rigorous as VPW. A bigger portfolio is going to result in higher spending (or donating to charity, or supporting family, or whatever) for many real world retirees. My tests show that a bond tent hurts in 70% of retirement cohorts by an average of $2,000/year[*]. And in 10% of cases by as much as $5,000-$7,000/year.
That's not even an especially controversial claim or interesting insight. Because bond tent advocates would say: "But those are all in 'successful' retirements anyway and when everything goes smoothly having extra money doesn't really matter, or at least it doesn't matter very much. If I retired expecting to live off of $40,000 a year and end up living off of $50,000 a year, I don't really lose sleep thinking that I could have had $55,000 a year by avoiding the bond tent. The bond tent is supposed to help with the worst case scenarios
not the 95% of normal case scenarios."
Which is fair enough. And a bond tent did
help out in 1929, 1930, 1931, and 1969 (among others). It only helped out in 30% of years but those are years that people really worry about!
In 1929, your lifetime income went from $35,000/year to $39,000/year.
In 1930, from $39,000 to $43,000.
In 1931, from $47,000 to $50,000.
In 1969, from $31,000 to $32,000.
(and in the other dozen or so cases the gap was progressively smaller)
These numbers just confirm what we already knew: the only thing
a bond tent helps with is a massive equity crash in the early part of retirement. Remember: the bond tent is about trying to deal with the portfolio size effect & nothing else. The numbers give a sense of the scale of the benefit. When there's an 85% equity crash then the bond tent gives us about $4,000/year in additional income during retirement.
For other scenarios, even 1969, the equity crash just isn't big enough for the bond tent to help.
Different people can look all of that and come to different conclusions about the value of a bond tent. After all, someone could certainly say "yes, there is only a vanishing chance of a 1929-style 85% crash again. But I'm willing to pay the cost for some protection against that." And who can argue with a personal choice like that, even if it isn't the one I'd make?
[*]: All income numbers are based on certainty-equivalent withdrawals assuming a VPW variable withdrawal scheme by the retiree. The bond tent assumes going from a 75/25 portfolio to a 100/0 portfolio over 30 years (7 years of bonds is 7/25 = 28%, so this is closely equivalent to holding 7 years of bonds to start with). The non-bond tent portfolio is a static allocation of 87/13. 87% because that's the average bond allocation of the bond tent over 30 years. 100+75 / 2 = 87%.