Portfolio Stability And A Good Night's Sleep

If one goes to a place like portfoliocharts (strongly recommended for the interested asset allocator), there's a concept that we have struggled with.

It's the idea of Start Date Sensitivity.

For us, this is not at all intuitive, so let's try to go all the way to see where we are now in our understanding. We have discovered that it's key to quite a lot of insights around investment portfolios.

It's a funny and unusual measure. It took us quite some time to start to understand this way of looking at a portfolio. 

One way to define it is as a measure of how good a guide the 10 years last history has been for the 10 years that lay ahead. By pointing out when this difference is as large as possible (both in the postive and negative sense), it focuses on the year when the previous time period was as maximally misleading for the upcoming time period, and how big that effect was. 

Is history a good guide for the future?
(Reconstruction of the West Pediments of the Parthenon, CC BY-SA 2.0 Tilemahos Efthimiadis)

Going with one asset class

Let's look at some examples:

• Total US Stock Market: Luckiest 10 years = 14.9 percent points better, per year, than the preceding 10 years, Unluckiest 10 years = -18.1 percent points worse per year then what the preceding 10 years annual average return hinted at.

So this mean that a happy US investor could sit and look at the stock market and think that: well, this is sure looking good. And at some point, the investor would say that, what the heck, the last 10 years have been going really, really well. Let's jump in and put my savings in the stock market.

The most unlucky the investor could have been when doing that decision, was the period when the stock market performed on a yearly average -18.1 percent points worse per year, for 10 years, than it had done the last 10 year period.

So that's a measure of how bad a guide for the future the last decade was.

For the US stock market, that would have happened in 1999, back when we were in our teens. The market had had a return on around 14 percent per year for ten years, but that would not continue. The average yearly return for the next 10 years would be -4 annually. This gives us the most unluckiest start date sensitivity for 1999, with those 18 percent points .

As both of us experienced and at least vaguely remember 1999, this negative experience colored our view of the stock market and probably at least partly explains why we treat it with caution. 

Of course, time would have partly fixed it for our unlucky investor, but it would take a very long time indeed. Still today, that 1999 investor would have got only a modest 5% return per year from that initial decision, and a lot of volatility on the way. Not that much of an issue if one is a teen and just started to accumulate money. More nerve-wracking if one is deeper into one's career and accumulation, and put substantial money on the table.

On the other hand: the luckiest investor would have jumped onboard the stock market in 2009 and then been in for ten tremendously good years, a lucky strike we're still part of.

Just for fun, let's have a look on a much more volatile asset class. No, not bitcoin. Gold, of course. 

• Gold (USD): Luckiest 10 years = 21.0p.p. , Unluckiest 10 years = -30,7p.p.

One would probably be close to insane to put a significant part of one's money in an asset class that is as volatile as gold, but let's play with the idea. 

In that case, an investor, after witnessing a terrible performance for gold during 10 years and then by some miraculous inspiration buying into it anyway, could be 21 percent points better off, per year, for the upcoming ten years. 

So the start date sensitivity goes both ways, and the bigger it is, the less guidance we seem to get from the last ten year period.

What happens if we mix two assets?

Let's instead mix a healthy portion of the stock market with a substantial amount of gold. Let's say by using our hypothetical 80/20 split.

• TSM (80%), Gold (20%), looking on performance in USD:  Luckiest 10y= 10.6 percent points, Unluckiest 10y = -11.3 percent points difference per year

What does that mean? Well, our investor would still be in for a surprise if she hit the unluckiest year, but after yearly rebalancing in and out of that gold, she could comfort herself that she would be much better of than with going totally into the stock market, and she would have come out quite ok 10 years later and would be back in black numbers much quicker.

Good for her.

Several asset classes

What happens if we go to our more conservative portfolios that we've been using ourselves, where we mix gold, long term bonds in different currencies, cash and small and large cap stocks, both domestically and abroad?

• Permanent portfolio, our take on it, in SEK: Luckiest = 5.1 p.p. , Unluckiest = -4.9 percent points difference per year.

The results are roughly the same for the US market, but with even smaller sensitivity. You can read more about our take on the portfolios here, here and here.

What does this mean? Well, there was a  year, where the last 10 years had an annual average return of 10%. If our unlucky investor jumps in that year, she would be in for 10 years where the average annual return would be 5.1%, or 4.9 percent points worse than indicated by the preceding 10 year period. All this after inflation.

And that was the most unlucky difference that ever happened during the last 50 years.

So in the worst case it would be roughly the same as for the stock market, but with much lesser volatility. And the permanent portfolio is back in black much quicker.

That's a glimpse of why we, who both were into our careers and had a bunch of money, decided to start our investment journey with the permanent portfolio. 

The permanent portfolio with its more narrow start date sensitivity, in contrast to the stock heavy alternatives, almost entirely avoids the question of 'is this the right time'? 

That's an important, stress-inducing question that can be avoided.

Finding our path

But of course, one pays a price with the permanent portfolio when it comes to where return. So now the fun part. 

• Our Pathfinder Portfolio, our take on it, in SEK: Luckiest= 6.5 percent points, Unluckiest= -7.3 percent points difference annually
• Our Pathfinder Portfolio, our take on it, in USD: Luckiest = 4,1 p.p. , Unluckiest = -4,2 p.p.

Our pathfinder portfolio of course has a higher start date sensitivity, measured as the difference between the luckiest and unluckiest points, than the permanent portfolio.

But still, it's much, much lower than the stock market. For our domestic set-up, the average return for the pathfinder portfolio was 9.2%, and a US adaptation was 7.3%. 

The total US stock market for the same period had an average return of 8.3%. 

So by mixing a combination of individually high volatility assets, one gets a lot more safety, quite small sensitivity in regards to when to buy into the portfolio, and an average return on par or even higher than the stock market. 

Just saying.

A Good Night's Sleep

That could be the end of this article. But it's not. 

What we then slowly realized was that the question if one should put all one's money in this or that portfolio is not only a question about start date.

It's not a one-time question.

It's a question that comes up all the time, nagging with these little thoughts that can keep anyone who is not a Stoic God awake at night.

• Have the last years been booming? How much could we realistically loose by sitting still?
• For how long should we accept that the portfolio is lagging behind? 
• Should we sell tonight? Or should we sit still and hope that the boom continues for a little longer?
• Do we dare to buy into an asset that has been lagging for a long time?

All investors take the decision if we should stay in an asset allocation or leave it every night, irregardless of if we try to pretend that we don't. There's no real way of avoiding this, as taking no decision is still a decision.

At the bottom of it, start date sensitivity is about the stability of an asset mix. Having a good stability in the mix of assets means that one has to think much less about time periods, paradigms, bull and bear markets and timing in general, and all those little nagging questions that timing entails.

Because the effect of being lucky or unlucky with timing has much less impact on the performance of the portfolio.

The decision to stick with the portfolio can then truly become permanent, with less second-guessing, and better sleep at night. 

Farewell,

//antinous&lucilius

Where to go now?

• Go over to portfoliocharts.com and play around for yourself.
•  Read about the Pathfinder Portfolio
• ... or the Permanent Portfolio: A Portfolio for Accumulation Part 1 and Part 2.

Note on the numbers: As usual, this is back-testing from 1970, cumulative annual growth rate, with inflation removed.