Permanent Portfolio using Vanguard LS40

[1nvest]

Vanguards LS40 (Lifestrategy 40/60 stock/bond) 75% weighted, along with 25% gold, approximates a Permanent Portfolio (given that generally the PP is permitted to drift between 15% and 35% lower/upper weightings for each of the stock/gold/short term bonds/long term bonds assets before rebalancing)

Vanguards web site shows LS40 historic yearly total return data back from 2012 and combining that in 75/25 LS40/gold proportions ...


Taking a 4% SWR (inflation adjusted 4% yearly income) and as of the end of 2021, ten years into retirement, the portfolio value = 91% of the inflation adjusted 2012 start date portfolio value (down from 95% at the end of the 9th year).

The general consensus with SWR is that of sustaining 30 years of income/withdrawals. Personally I think 25 years is a more appropriate range, as that sees a 65 year old retiree through to age 90 when perhaps you're more inclined to sell your home to fund all-inclusive care/retirement home for your remainder years. Having 90% of the inflation adjusted 2012 start date amount still available after 10 years, with another 15 years of drawdown expectation would seem well on course to likely be successful.

[vand]

There are a lot of problems with using LS40.

- LSx equity allocation is 25% UK, rather than the global cap-weighted 4%
- PP strategy deliberately calls for a barbel strategy between T-Bills and Long Bonds. The bonds in LS are a blend of all duration fixed income which reduces the effectiveness of the barbel
- LS automatically rebalances daily, so you never get discreet rebalance. This is fine if the rebalancing period for the whole portfolio is daily too, but otherwise you have one rebalancing rules for some portions, and another for other portions

[1nvest]

The differences between a 1 and 20 year barbell and a 10 year bullet broadly wash. As might the higher yield from corporate bonds minus defaults tend to wash. Total bonds rather than short and long dated gilt/treasury barbell is perhaps OK.

UK has around 70%+ of earnings sourced from foreign, so is far from being a pure domestic stock index.

Rebalancing is more of a risk control mechanism over that of a reward enhancement mechanism. Rebalanced or not broadly makes little/no difference other than non-rebalanced tending to see higher interim volatility.

Looking at US data and for 30/45/25 stock/bond/gold the risk/reward was reasonable, and running Monte Carlo simulations also indicated reasonable outcome. I set that to 4% SWR 25 years (rather than the more usual 30 years), and also set it to a bad earlier years sequence of returns risk i.e. the 'worst five years first' setting, despite which it still had a 99%+ success rate. And a 50th percentile (average) case of ending with 80%+ of the inflation adjusted start date amount still available at the end of 25 years.