Building A Risk-Parity Portfolio: An Example
Summary
- Risk-parity portfolios weight asset classes by volatility, and use modest leverage to boost returns while keeping volatility manageable.
- This article will walk you through a simplified example of how to construct a risk-parity portfolio with 4 assets.
- We will use the S&P 500, long-term treasuries, gold miners, and high-yield bonds in our example.
- I will end with a prediction of how risk-parity portfolios will likely perform for the next decade: 5% CAGR, and 10% standard deviation of volatility.
The idea behind risk parity is simple: build a portfolio of uncorrelated assets, weighted according to their volatilities, and use modest leverage to boost returns while keeping volatility tolerable. Let’s try our hand at building such a construction.
We’ll use four asset classes, and for the purposes of being able to simulate the past 29 years of financial market history using portfoliovisualizer.com, we’ll only use mutual funds available since 1991.
- US Stocks – S&P 500 (VFINX)
- Long-Term Treasuries (VUSTX)
- Gold Miners (INIVX)
- High Yield Bonds (FHAIX)
Important note: I will use a lot of charts to illustrate portfolio performance in this article. In all of these simulations, all dividends were reinvested. Additionally, all portfolios were rebalanced annually.
A First Foray
If we knew absolutely nothing about these asset classes but had to assign each of them a portfolio weighting, we would probably resort to an equal weighting of 25% each. This is called the 1/N portfolio. Let’s simulate how this portfolio did (assume annual rebalancing):
Source: portfoliovisualizer.com
It looks very respectable – 8.9% annual returns with an 11.9% standard deviation for volatility. Assuming a bell-curved distribution of annual returns, this portfolio was and will likely be profitable in 77% of all 1-year periods. The maximum drawdown of 29.66% was caused by the 2008 financial crisis.
An Improved Allocation
You might be happy with this sort of portfolio already. But let’s ask for more: let’s try to improve it! Let’s dig into the mechanics of the portfolio and its statistical properties.
Source: portfoliovisualizer.com
These tables might look intimidating, but we’ll only focus on the parts that are boxed in green.
Firstly, notice that the S&P 500 (VFINX) had the highest CAGR, or annual return rate of all the assets. We are greedy, so we are tempted to increase the allocation. Let’s leave that for until later.
Secondly, notice that gold miners (INIVX) have a very high standard deviation – they are very volatile. To make their contribution to portfolio risk more proportionate, let’s cut their allocation from 25% to 10%. This frees up 15% of portfolio space.
Thirdly, notice that high yield bonds have a lower CAGR than the S&P 500. Since we feel greedy, let’s transfer 5% of allocation from FHAIX to VFINX, resulting in a 20% allocation in FHAIX and 30% allocation in VFINX.
Fourthly, looking at the monthly correlations of the assets, notice that long-term treasuries have a slight negative correlation with high yield debt and the S&P 500. Negative correlation helps reduce a portfolio’s volatility, so let’s do that by increasing the allocation to VUSTX by 15% to 40%.
Putting together the four points made above, this results in the following new (rather pretty) portfolio allocation:
- 10% Gold Miners (INIVX)
- 20% High Yield Bonds (FHAIX)
- 30% S&P 500 (VFINX)
- 40% Long-Term Treasuries (VUSTX)
Let’s see how this portfolio performed over the same time period:
Source: portfoliovisualizer.com
Portfolio CAGR increased to 9.05%, so we squeezed out a tiny bit of extra return – 0.15%/year, to be precise. But, the portfolio standard deviation was reduced to 8.09%, and the 2008 drawdown was reduced to 20.62%. Our reallocation of weightings reduced portfolio volatility by a third!
Similarity to Tony Robbins & Ray Dalio’s “All Seasons”
Notice that this portfolio is very similar to the “All Seasons” portfolio advocated by Tony Robbins, and sourced from Ray Dalio:
- 30% stocks
- 40% long-term treasuries
- 15% intermediate-term treasuries
- 7.5% gold
- 7.5% commodities
Our portfolio is nicer for the retail investor in one particular way: we don’t require a commodity exposure, and instead have a "leveraged" exposure to gold through the mining stocks. Commodities are problematic as gaining exposure through ETFs can be costly (very high fee structure), and trading commodity futures is not for the retail investor.
Introducing Some Leverage
Most readers might just stop here. But let’s keep going: how can we improve this portfolio? How would we tactically use some leverage to boost the portfolio’s returns? Let’s go back to this image we’ve seen before for ideas:
Source: portfoliovisualizer.com
Notice that high-yield bonds and long-term treasuries both have lower standard deviations (volatilities) than stocks. Let’s use leverage on them to increase their returns until their volatility matches that of stocks and gold miners. But how do we model this?
An investor “levers up” by starting with cash, borrowing more, and buying the asset. To simulate the margin loan interest, we’ll add a negative allocation to VFISX, the Vanguard Short-Term Treasury Fund.
Just eyeballing the standard deviations, let’s increase our exposure to high yield bonds 100% and our exposure to long-term treasuries by 50%. This will result in a portfolio consisting of:
- 10% Gold Miners (INIVX)
- 40% High Yield Bonds (FHAIX)
- 30% S&P 500 (VFINX)
- 60% Long-Term Treasuries (VUSTX)
- -40% Short-Term Treasuries (VFISX) – this simulates the margin loan
Let’s see how this portfolio did:
Source: portfoliovisualizer.com
Our leveraged portfolio had a CAGR of 10.48%. 1.3%/year of extra yield is nothing to scoff at: starting with $10,000 in November 1991, the investor in the levered portfolio would today have $172,599, as compared with $119,032 in the previous unlevered portfolio. As expected, our portfolio does have a marginally higher amount of volatility than before.
Replicating The Leveraged Portfolio… Without Using Margin
You might be wondering: is it possible to reap the benefits of using leverage, without actually using a margin account? My answer would be a definite “yes”! But we would need to use some rather niche assets.
Let’s discuss our levered long-term treasuries first. The VUSTX holds whole bonds, which consist of coupons and principal. Long-term bond funds that hold “zero coupon bonds”, or bonds without coupons, do exist, and have volatility that matches that of our 1/3 levered long-term treasuries. Those funds are the Vanguard Extended Duration Treasury ETF (EDV) and the PIMCO 25+ Year Zero Coupon U.S. Treasury Index ETF (ZROZ). Let’s see how they compare:
- Portfolio 1 is the 150% VUSTX / -50% VFISX
- Portfolio 2 is 100% EDV
- Portfolio 3 is 100% ZROZ
Source: portfoliovisualizer.com
It turns out EDV and ZROZ are quite good substitutes for levered long term treasuries!
What about our levered high yield debt? We can use BDCs (Business Development Companies) for that: they are essentially collections of high yield bonds, levered up to a 1:1 debt/equity ratio. That leverage profile perfectly matches that of our last portfolio.
To see whether BDCs are a good substitute for levered high yield bonds, let’s plot the following:
- Portfolio 1: 200% FHAIX, -100% VFISX
- Portfolio 2: Ares Capital (ARCC) (a “good” BDC)
- Portfolio 3: Main Street Capital (MAIN) (a “good” BDC)
Source: portfoliovisualizer.com
As we can see, things are not looking as consistent. On the bright side, this does suggest that “good” BDCs can easily beat levered high yield bonds. BUT, this is contingent on our ability to pick “good” BDCs.
Unfortunately, many BDCs are horrible investments (read: consistent NAV decline); however, what separates the good from the bad is clear: a history of good underwriting and few/no loan chargeoffs.
Putting all this together, we can finally arrive at a leverage-free risk-parity portfolio that retail investors can hold (with example funds):
- 10% Gold Miners (GDX), (RING), etc.
- 20% “good” BDCs (ARCC), (GBDC), (MAIN), (HTGC), etc.
- 30% S&P 500 (VFINX), (SPY), etc.
- 40% Long-Term Zero Coupon Bonds (EDV), (ZROZ).
What does the future hold?
Unfortunately again, the experience of the past 30 years will probably not be repeated. Indeed, two phenomena conspire against an expectation that future returns will match past returns.
First: The current stock market is richly overvalued, as measured by the cyclically adjusted price-to-earnings ratio (or Schiller P/E):
Source: multpl.com/shiller-pe
As we can see, the current stock market is as richly valued as it was at the time of the crash of 1929. This does not bode well for stock returns for the next decade (not just for this risk-parity portfolio, but for almost all portfolios). Fellow Seeking Alpha writer Georg Vrba estimated in his article Estimating 10-Year Forward Returns For Stocks With The Shiller CAPE Ratio And The Long-Term Trend - January 2020 that the 10-year forward return of the S&P 500 is just 5.9%.
The second headwind is that we are near the top of the bull market for bonds. In particular, the 30-year treasury yield is nearing its bottom:
Source: fred.stlouisfed.org/series/DGS30
Since falling rates cause bond prices to rise, and since current 30-year bond yields are just ~1.6%, close to zero, there is essentially no room left for bond yields to fall. Hence, we might surmise that the CAGR on EDV or ZROZ for the next decade will be about 1.6%.
Let’s summarize my expectations for the future:
- The S&P 500 is expected to return about 6%/year.
- EDV or ZROZ is expected to return about 1.6%/year.
- A good BDC has about an 8% dividend yield (per ARCC’s yield).
- Gold mining will return about 5%/year (per previous charts).
The CAGR of a portfolio is just the weighted average of the CAGR of the individual assets (plus a ~0.5% diversification and rebalancing bonus):
Asset | Weight | CAGR | Contribution |
Gold miners | 10% | 5% | 0.5% |
BDCs | 20% | 8% | 1.6% |
S&P 500 | 30% | 5.9% | 1.77% |
Zero coupon long-term treasuries | 40% | 1.6% | 0.64% |
(Diversification & rebalancing premium) | 0.5% | ||
Total | 5.01% |
Assuming that past volatility predicts future volatility, we might estimate that a risk-parity portfolio is expected to return about 5%/year, with ~10% standard deviation of volatility (comparable to an individual long-term treasury).
What do you think of this portfolio and my prediction? Let me know in the comments!
This article was written by
Analyst’s Disclosure: I am/we are long HTGC. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.
Seeking Alpha's Disclosure: Past performance is no guarantee of future results. No recommendation or advice is being given as to whether any investment is suitable for a particular investor. Any views or opinions expressed above may not reflect those of Seeking Alpha as a whole. Seeking Alpha is not a licensed securities dealer, broker or US investment adviser or investment bank. Our analysts are third party authors that include both professional investors and individual investors who may not be licensed or certified by any institute or regulatory body.
Recommended For You
Comments (46)
Own Baby Bonds of HTGC.
Own Common of MAIN.
Just spreading my momey & risk around. Looking at the common of ARCC once the price comes down.
Should give a long term (20 year) Sharpe north of 1.1, sorting north of 1.7 and a calmar north of 2.
Stay safe! And once more, it's a very well written article!
What's the difference between risk parity vs. Black Litterman scheme? I read risk parity does not depend on expected return. Is that a shortcoming?
For example if you expect bond return 1.6% going forward, why would you allocate anything? For me, investing long term bond going forward needs good timing since risk reward is not that better than cash.
