The Permanent Portfolio: Skeptical, But No Dismissive

I will admit, when I first read about the Permanent Portfolio in the late-80s, I was somewhat skeptical, but not totally dismissive.

Here is the classic Permanent Portfolio, equal proportions of:

S&P 500 stocks
The longest Treasury Bonds
Spot Gold
Money market funds

Think about Inflation, how do these assets do?

S&P 500 stocks – mediocre to pretty good
The longest Treasury Bonds – craters
Spot Gold – soars
Money market funds – keeps value, earns income

Think about Deflation, how do these assets do?

S&P 500 stocks – pretty poor to pretty good
The longest Treasury Bonds – soars
Spot Gold – craters
Money market funds – makes a modest amount, loses nothing

Long bonds and gold are volatile, but they are definitely negatively correlated in the long run. The Permanent Portfolio concept attempts to balance the effects of inflation and deflation, and capture returns from the overshooting that these four asset classes do.

What did I do?

I got the returns data from 12/31/69 to 9/30/2011 on gold, T-bonds, T-bills, and stocks. I created a hypothetical portfolio that started with 25% in each, rebalancing to 25% in each whenever an asset got to be more than 27.5% or less than 22.5% of the portfolio. This was the only rebalancing strategy that I tested. I did not do multiple tests and pick the best one, because that would induce more hindsight bias, where I torture the data to make it confess what I want.

I used a 10% band around 25% ( 22.5%-27.5%) figuring that it would rebalance the portfolio with moderate frequency. Over the 566 months of the study, it rebalanced 102 times. At the top of this article is a graphical summary of the results.

The smooth-ish gold line in the middle is the Permanent Portfolio. Frankly, I was surprised at how well it did. It did so well, that I decided to ask, what if we drop out the T-bills in order to leverage the idea. It improves the returns by 1%, but kicks up the 12-month drawdown by 7%. Probably not a good tradeoff, but pretty amazing that it beats stocks with lower than bond drawdowns. That’s the light brown line.

Results S&P TR Bond TR T-bill TR Gold TR PP TR PP TR levered

Annualized Return 10.40% 8.38% 4.77% 7.82% 8.80% 9.93%
Max 12-mo drawdown -43.32% -22.66% 0.02% -35.07% -7.65% -14.75%

Now the above calculations assume no fees. If you decide to implement it using SPY, TLT, SHY and GLD, (or something similar) there will be some modest level of fees, and commission costs.

What Could Go Wrong

Now, what could go wrong with an analysis like this? The first point is that the history could be unusual, and not be indicative of the future. What was unusual about the period 1970-2017?

Went off the gold standard; individual holding of gold legalized.High level of gold appreciation was historically abnormal.Deregulation of money markets allowed greater volatility in short-term rates. ZIRP crushed money market rates.

Federal Reserve micro-management of short-term rates led to undue certainty in the markets over the efficacy of monetary policy – “The Great Moderation.”

Volcker era interest rates were abnormal, but necessary to squeeze out inflation.

Low long Treasury rates today are abnormal, partially due to fear, and abnormal Fed policy.
Thus it would be unusual to see a lot more performance out of long Treasuries. The stellar returns of the past can’t be repeated.