300 Year Sim: Permanent Portfolio vs Bogle, Ivy, Momentum, Risk Parity, et al.

[Sylph]

Hi guys,

Long time listener, first time caller.

I read with interest the discussion regarding the frustration some users are experiencing with the Permanent Portfolio. Short-term, results-oriented biases are the culprit, but what else do we really have to go on?

Well, this egg head programmed a 300 year Monte Carlo simulation to answer just that:

http://www.nasdaq.com/article/the-most- ... n-cm279719

tl;dr?

"The Permanent Portfolio ("PEP") defended its reputation, to be a good and simple investment concept as it has delivered the third highest Sharpe Ratio and the third highest overall score."

What I found curious was the standard Bogle portfolio (60/40) performed the worst, followed by the Momentum strategy. This is not the final word, by any means, but food for thought. Incidentally, the CAGR for the top two (risk parity and diversified) was about 5.3%, the PP was 5%, and the Ivy was at the bottom with 4%.

I think this sort of simulation has value in that might help re-align expectations as to what a safe, simple, balanced approach to the market can offer on the passive investment front. Returns of 10% or more a year are simply not the norm. Certainly if you are heavy in stocks, bonds, or gold and one of them takes off you can make out like a bandit. But to capture them, you will need to either get lucky, or have specialized knowledge.

We all wish were were in 100% stocks since 2009. But had we been 100% in stocks since 2000, we would have had about half the return of the Permanent Portfolio.

And probably a few ulcers.

[barrett]

Sylph,

Welcome to the forum. Alas my investment horizon at the moment is only 40 years!

Yeah, the tone on here would be different if we were getting 2014-type results, but keep in mind that there are folks who use the PP who don't even get worked up enough to post here.

[ochotona]

This is crap. The heart of any Monte Carlo analysis are the variances and co-variances. To take 1980-2013 variances and co-variances and pretend to extend them from 1700-2013 is utter rubbish. Ignore it. The author's own proprietary allocation "wins", and then he sends you to click on a link to subscribe. Commercialized horse-apples.

"Furthermore, we have used the historical variance-/covariance matrix from 1980 until 2013 for the...."

[MachineGhost]

1968-2015 would have been fine.  I may do it some day when I can be bothered.

[dualstow]

Great first post, Sylph. Thank you!

[Reub]

I'd give dualstow's left nut for a 300 year time horizon!

[dualstow]

I'd give dualstow's left nut for a 300 year time horizon!

Hey, I'm still using that!