I was thinking about the "ultimate backtest" of simulating different worlds for 500 years. Essentially what I think would be really cool to do is to look at the last 100 years and simulate the next 500 but to add in some extra variance by including the probabilities of things like: -A greater depression (something worse than the 30's) -A stock market bubble bigger than the 90's -A propesrous time like the 20s but even more so -Bigger inflation than in the 70's -A Great Moderation but even greater -WW3 -Huge tech boom -World plague wiping out populations -Global warning getting a lot worse -Other extremes Essentially, take whatever price extremes we saw in the last 100 years in stocks, bonds, commodities, gold, real estate, CPI but add the chance we could see something even more extreme in them. And that probability would not be normally distributed in a Gaussian fashion but rather, there would be a real chance that they could happen. And that got me thinking, in that simulated world, how robust is to be 100% invested in equities in the same country? Isn't Buffett just underwriting the risk of a greater depression or something really bad? Because it if happened, he could very well be wiped out, just like Graham was in the 30's I think its this sort of exercise that made Ray Dalio put most of his money in his All-Weather fund and not on the Pure Alpha fund. We just simply don't know how extreme things can be on the future but odds are, they will be more than the past, not less because time is infinite, and Murphy's law says that whatever can happen, at some point will. We just don't know when those extremes will show up
I'm not saying he is necessarly wrong in investing 100% of his money in the US stock market. If you think you know something that other don't, it might pay to go for it. Especially, because I'm talking about the kind of variance that affects time intervals that are way beyond a human's lifetime. At some point you might say "I rather have a 99% chance of a good life and 1% of being wiped out, than a having an average life". I'm addressing in general the robustness of the idea of being 100% in stocks in the same country. Greek investors learned it can be pretty devastating. Overall, given that issues I mentioned, it seems a pretty lousy investment policy. But even in the US it could turn out to be lousy, the US is only 200 years old, time is infinite. An event like Cuba/USSR/Greek is pretty much a terminal event to a 100% stock investor (you lose 99%-100%). Who can guarantee that in 500 years there WONT be a terminal event in the US stock market?
Thinking about this simulated world, I find it rather ironic that a guy that Taleb criticizes pretty heavily (Markowitz) is the guy that provides the solution (diversification) in how to cope with this sort of 'Extremistan' problem in asset prices
So lets say I simulate the next 500 years with some extremes in asset prices bigger than seen in the past. Then lets say I simulate 500 "worlds" with 500 years in each them following the same principle. Then I simulate 500 universes with 500 worlds each, and 500 years of simulated asset prices in each. The amount of extremes that would be possible as a result of this backtest would freeze in fear any investor. Like, the things that would be in it would shock anyone beyond their widest dreams. Yet, it would still all be realistic because we keep getting more and more data in financial markets and the distribution is not Gaussian. Bond yields in Germany were pretty low before the hyperinflation in the 20s, there was 1987, 99.99% loss in purchasing power of Brazilian currency etc, etc It would be pretty scary to even look at the spreadsheet, most people would run away in terror. But how cool it would be to know the optimal portfolio as a result of such simulation?
I won't be around to put my money where my mouth is, but I would pretty much guarantee that there WILL be a terminal event in the US stock market in the next 500 years. As a student of history, I'd point out that the greatest empire the world has ever known, The British Empire, saw its demise after WWII. Nothing is permanent. America will go into decline and you will see that event.
If there is an optimal portfolio for this, it could be -A perfect 'Tangency' portfolio that is balanced against all the extremes -A Taleb type allocation of 90% safety and 10% in stuff that benefits from the extremes Maybe thats why Taleb recommeds that 90/10 allocation. If you think about it, his portfolio would fare quite decently because it would not suffer as much from the extreme surprises of so many simulated realities. The issue is what is '90% in safe assets', in some extremes, the safe might turn into dust overnight. So, thats where Markowitz comes in and saves the day with diversification
This leads to the idea of, in addition to a balanced portfolio, one should "invest" in extremes as an asset class. it provides further diversification. its not as simples as going around buying options, but buying them in seleced situations, investing in venture capital, biotech, etc
And if the simulation ever were to be done, one thing is almost surely going to be found -every asset class (all stock, bond, currency market plus gold) will lose all its value at some point, not likely at the same time though. by stretching out time like that, its a certainty at some point they will run into trouble. heck, thats true even in a gaussian distribution. in a non-gaussian it just happens a lot faster -but if you invest in different countries, you decrease your risk dramatically
The thing about this 90% of the Taleb portfolio is that it can't be 100% in short-term fixed income of the same country. That would go against the history of financial data plus it would leave one pretty vulnerable to extremes (and it just flat out doesn't work). As a result, I think it makes quite a bit of sense that the 90% to resemble more a All-Weather/Seasons type allocation (with country diversification), but perhaps tilted as not to carry too much stock market risk. Perhaps a better built version of my 'widow' portfolio. Then with the 10% the investor can try to gun for extremes. So essentially, its an All-Weather but with an added asset class, the extreme fishing asset class. As a result, I really dont see much difference between what Taleb advocates and what Dalio advocates with the exception that Taleb tries to add a component to benefit from extreme events. Dalio doesn't but when you got $170B to manage, going around looking for startups is not really going to make much of a difference. With that kind of size its hard to get any money invested in options, biotech, etc But for small guys like us, its a nice idea to add to a diversified basket
Found this awesome article by Bridgewater talking about their fund after the 2015 turmoil http://www.ahwilliamsco.com/includes/OurThoughtsaboutRiskParityandAllWeather.pdf Its providing further clues about Dalio's 2014 portfolio he gave to Tony Robbins "Isn’t this bond heavy portfolio especially vulnerable to a selloff in bonds that seems likely now that it’s pretty clear that interest rates will rise? No. While the amount of money invested in bonds is generally greater than the amount of money invested in more volatile assets in the portfolio, this is done to achieve balance in the portfolio so that it would not have a systematic risk. Because the portfolio is well-diversified, it has no such systematic bias to do better when interest rates are falling than when they are rising. For example, from the point that bond yields were at their lowest in the 20th century to when they were at their highest (from March 1946 to September 1981) the average annual return of this diversified portfolio would have been 8.7%, in comparison to a 7.6% return for a 60/40 stock bond mix. In the inflationary 1970s when bonds were a bad investment, this diversified portfolio did about as well as during the disinflationary 1980s when bonds were a good investment. That is because the portfolio is diversified so as not to be exposed to any particular economic environment. The chart and table below convey this picture."