Harry Browne Lives in My Head, but Pays Plenty Rent

Introducing a three asset + cash strategy that employs both risk parity, and momentum

(A Golden Doodle thinking about financial matters, courtesy of Gencraft A.I.)

Harry Browne’s Permanent Portfolio has been in my mind for years. It’s quite simple, and says: to invest well in all chapters of the economic cycle, you only need four classes of assets.

As they were: stocks, for the good years; treasuries, during deflationary recessions; gold, in times of inflation; and cash (for when none of the other assets are working).

The most basic Permanent Portfolio strategy involves allocating 25% your capital to each of these four categories. You re-balance them yearly to get back to your share of 25% each.

That means you spend 10 minutes a year on your investments, with this resulting return (and note well — this is from 1972 to November 30, 2025):

CAGR: 8.49%; Maximum monthly drawdown: -13.19%; Worst year: -8.44%;

Negative years: 7 out of 53.

Just look at that equity curve! What the heck happened to the “lost decade” of 2000-2010?

The Sickening Seventies, the Ailing Eighties, the Awkward Aughts: mere blips on a quite pretty long-term equity curve.

Link here.

And yet…

The question that has been concerning me for some time: is a CAGR of 8.49% the best we can do?

(Especially considering this has shrunk to a mere 6.24% from the year 2000 onwards).

In the nothing-goes year of 2022, the Permanent Portfolio lost -8.44%. I find that result to be somewhat concerning. In some ways, the present time is similar to the end of 2021. Stocks are expensive, gold is more pricey than ever, and inflation is sticky, which doesn’t help treasuries. If 2026 or 2027 turn into something similar to 2022, then I’d certainly like to have a better strategy than this.

Let’s start improving the Permanent Portfolio, by using Risk Parity

I have written about Harry Browne before (here, and here), as have I about Risk Parity. The point of the latter is first of all: all assets have their inherent leverage — stocks are highly leveraged when the underlying company has high debt, but a debt-free corporation can have very low leverage. So, common investing allocations such as 60/40 stocks/treasuries are naive, and ineffective. Better to allocate assets according to their volatility, which is a much better measure of risk than any simple formula.

And anyway: volatility usually means bad times. If the stock market is getting wobbly, you’d better de-leverage. The VIX is usually inverse to the SPX.

Now, let’s look at what happens to stocks/treasuries/gold if we set them up in a standard risk parity framework, from 2000 onwards.

CAGR: 10.61%; Maximum monthly drawdown: -28.64%; Worst year: -23.29%

Negative years: 4 out of 25.

This strategy’s recent performance, since 2023, has been remarkable, with a CAGR of 22%.

And as poor as the drawdown number looks: it’s much less negative in most years. The only time the maximum drawdown crossed the 15% threshold was in 2022.

Nonetheless, that -23.39% yearly result for 2022 is intolerable.

As I’ve said in the past: my main problem with Risk Parity is that it doesn’t go to cash. Which leads to my next tweak…

Final addition: A simple momentum filter

To hope to diminish losses like in 2022, I added a filter: Assets are selected based on whether their price is above its 200-day moving average. Anything below gets kicked out for the month.

The rationale for this is: not much good happens below the 200-day moving average. It normally makes little sense to own an asset that has become so unpopular.

So — first level: determine which of our three assets (QQQ for tech stocks; TLT for long-term treasuries; GLD for gold) is above its 200-day moving average.

Second level: use risk parity to determine the relative allocation of each asset as a percentage of total.

The allocations and selections are carried out at the beginning of each month.

This gives you a wide array of possible allocations. In some months, we could be 100% invested in gold, or in the Nasdaq. In others, we ‘d be 100% in cash, if none of the three assets were above their 200-day moving average.

In most months however, we’d have a varying combination of two or three assets, like 54% QQQ and 46% gold, or 39% QQQ, 31% gold, 30% TLT.

The results I will now present were not optimized. I didn’t try using different assets, or playing with various market entry/exit schemes. All I did was test a hunch. But the results were not bad, for a start.

From January 2000 to end-November 2025:

CAGR: 11.53%; Maximum monthly drawdown: -14.5%; Worst year: -6.4%

Negative years: 5 out of 25, of which 4 were under -5%.

Positives:

+ This strategy managed the crisis years of 2008, 2011 and 2020 remarkably well. Looking at monthly results, we see why. In the last quarter of 2008 for instance, we gained 27% by being 100% invested in TLT. For March of 2020, we were up 0.86%, despite that month’s crash in equities, but once again, treasuries saved the day.

+ I like how a combination of just three unleveraged assets can perform quite strongly in varied years such as 2002, 2013, 2017, 2024 and 2025.

+ Simply buying and holding stocks performed weaker in both absolute and in risk-adjusted form. SPY b&h from 2000: CAGR 8.01%; max monthly drawdown -50.8%, worst year -36.81%.

+ In buying TLT, you’re owning insurance (and getting a current forward yield of 4.33%).

Insurance against what? Well, say a natural catastrophy, such as (heavens forbid) a west-coast earthquake: treasuries usually do quite well in such circumstances, as they are perceived as a safe haven. In case of a man-made crisis, like a hot war, TLT is also a good place to be. TLT also performed well in the half man-made, half-natural catastrophy of Covid.

+ Buying and holding the S&P 500 or the Nasdaq makes little sense to me at the current valuation levels. Any strategy worth its salt needs to have an exit point, and needs to know where to go after the exit point. I have strategies that exit to cash, and another one that goes to a market-neutral anti-beta fund. The former is always merely defensive, and doesn’t create value in itself. The latter is not time-proven, at least not for the long term. To use gold and treasuries as out-of-market assets is a useful addition.

+ Which other strategy has provided more than CAGR 10% with <15% drawdowns for the last 25 years, without being overly complicated and optimized? Which such strategy uses assets that fit each kind of economic situation? Please let me know in the comments section if you know of something that can perform as well, or even better!

(Honestly: I do have a few strategies that work almost as well, although most of them require leverage — which I dislike — to reach the level of 11% CAGR. Nothing like a margin call to blow you up! Anyway, this is the subject of my next post; make sure you’re subscribed to not miss it!)

Negatives:

- Getting re-positioned every month is a hassle, and it doesn’t come cheap, either.

- Three of these 25 years were a real dog’s breakfast. 2014 to 2016 were proper underperformers, and this is the time we suffered the -14.5% drawdown. (No other drawdown was larger than 9.6%, by the way).

The problem with moving averages is that you can suffer from bad timing luck (buying and selling at the worst possible time), and these whipsaw years had bad timing luck in spades.

In this strategy’s defense however, the five years from 2013 to 2017 averaged out at +9.32%, which is at least not miserable.

Own treasuries now, really?

For this strategy, the allocations for December 2025 are 51.38% TLT, 28.22% QQQ, and 20.41% GLD. Really? Who would want to be that strongly invested in treasuries right now?

As coincidence would have it, Callum Thomas of the excellent Topdown Charts service just a few days ago published a plea for owning treasuries.

“Key point: Investor allocations to bonds at are an 18-year low”

“Bond allocations reached major lows at both of the last two major stock market peaks (2000, 2007), and basically served as a bear market harbinger.

Aside from giving clues on the stage of the market cycle, this chart also served as a contrarian bullish indicator for bonds — with treasuries turning in strong double-digit returns after those two big troughs (and doing so while stocks dropped).

So I think this chart says as much about the stage of the market cycle, as it does about the importance of asset allocation (bonds performing their role as diversifiers and risk dampeners), but also about the big bullish setup in bonds in general.

(…) bonds have all the makings for a contrarian bullish setup (cheap valuations, bearish sentiment, cycle-low allocations) — and, for now, lack only the technical and macro confirmation (the tactical/timing element).”

Finally: even though TLT is officially in a bear market (20% beneath its highs), it’s been trending up recently, and according to this post I wrote three years ago, most treasuries are currently quite buyable.

This needs further testing.

This is only an initial trial; a proof of concept, so to speak. I carried out this backtest by use of time-consuming manual data entry into a spreadsheet. I can’t use my backtesting platform of choice — Portfoliovisualizer.com — because it doesn’t seem to be able to integrate moving averages into adaptive allocations. Perhaps Jaewon Jung’s excellent Quantmage platform would be good for this? Any tips would be much appreciated. Comments, please!

What I would like to evaluate is, how much better / worse would this approach be

• with other assets (SPY instead of QQQ, IEF instead of TLT)

• a different amount of assets (e.g., 2 max instead of 3 — in which case we would currently, for December 2025, be invested in 41.97% GLD and 58.03% QQQ, which may “feel” more appropriate

• how would we have performed in 1987?

• how about using a different moving average filter, such as EMA 200d, or a 50/200 crossover?

• you name it! Let me know how you think this approach could be improved, and I’ll try to test it. Thanks for reading!