EfficientInvestor wrote: ↑Mon Aug 26, 2019 6:40 am
One of my favorite investing books is William Bernstein’s, The Intelligent Asset Allocator. If you haven’t read the book, I highly recommend you grab a copy. In the book, Bernstein introduces us to Uncle Fred and his coin that determines the annual return of an investment. He then shows how if you have two identical coins that have similar but uncorrelated return and risk profiles, you receive a diversification benefit that results in better risk-adjusted returns.
So now you are probably thinking, “I really like this concept in theory, but what asset could represent the second coin.” If we do a quick overview of various asset types, we can see that there isn’t really anything else that provides the same kind of returns as the stock market while being completely uncorrelated.
Bonds lack correlation with stocks, but they don’t generally have the same kind of return potential.
It just so happens that if you apply leverage to bonds (via the use of futures, options, leveraged ETFs, etc.), you can approximate the return and risk profile of stocks.
I just wanted to say that I really appreciate and enjoy the way you present this. I also second your recommendation to ask Bill Bernstein about this on the Bogleheads podcast.
However, at MotoTrojan has pointed out, there are actually bonds that can have high returns, high volatility, and low correlation to stocks (without leverage): long duration bonds. One way to get the most duration exposure is with an extended duration Treasury fund, such as EDV, which for the period available in Portfolio Visualizer, Feb 2008 - Aug 2019, had higher return and higher standard deviation that SPY (S&P 500). A 40/60 SPY/EDV portfolio had higher return, lower SD, and higher Sharpe ratio than 100% SPY, which illustrates the coin flip analogy as well.
Portfolio 1 = SPY
Portfolio 2 = EDV
Portfolio 3 = 40/60 SPY/EDV
Source: PV backtest
However, it's also important to note that the 20y Treasury yield dropped from 4.35% at the beginning of the period to 1.78% at the end, and the 30y dropped from 4.35% to 1.96%. So this is why EDV returned 9.37% instead of something closer to 4.4% if yields had not changed.
I was going to say that yield curve rolldown return might have been expected to add to the return even with a static yield curve, but since the 20y and 30y yields were the same at the beginning of the period, there might not have been any rolldown component of return in that maturity range, since there was no positively-sloped yield curve to roll down.