Do bonds do anything for your investment portfolio at all?
Do you have bonds in your investment portfolio? Do they do anything for the returns on your investment portfolio at all? Despite all the well-meaning advice to hold bonds to reduce risk in our investments, the typical 80:20 or 60:40 portfolio is still overwhelmingly dominated by the risk of stocks. And evidence from the 2000s show that they end up doing worse than 100% stock portfolios.
But things are a little different in the 2010s. Bonds still have a place in the investment portfolio, but as part of a risk-parity portfolio.
Introduction
“Put a portion of your investments in bonds. They help to control the risk of the investments to a level which you are comfortable with, based on your risk appetite“. So goes the usual advice we receive for investing. In fact, some may even go further and make it sound more scientific – “The amount of equities you should have is equal to 100 minus your age“. But in an era when bond yields at their all-time lows, hovering just above zero, is this sort of investing advice still relevant? Do bonds do anything for your investment portfolio at all?
Considering the low yields bonds have in almost every developed country over the past 10 to 15 years, one cannot help but wonder whether bonds are still relevant. Or are they just stuffed into the investment portfolio to make it safer, no doubt, but also to pull down returns? Especially when “risky” assets like equities have risen sharply with lower and lower levels of realised risk.
So let’s rewind a bit and try to figure out from the data of the past decade or two to see whether bonds really do anything at all for our investments.
Do bonds do anything for your investments at all? Or do they just lead to poor returns and little risk reduction?
Why do we need bonds in our portfolios?
The idea to have risk free assets in our investment portfolios actually goes a long way back, to 1958 to be exact. In this paper, the future Nobel laureate James Tobin suggested that by holding the market portfolio or risky assets, and either having deposits or borrowing at the risk-free rate of interest, we can tailor our portfolios to suit our risk tolerance. This has come to be known as the two fund separation theorem, a key foundation of the view that you cannot do better over the long term than holding the market portfolio instead of doing stock picking.
An illustration of the two fund separation theorem
Now, if you know anything about financial theory, you’ll realise that what Tobin talked about was lending or borrowing at the risk-free rate of interest. That is, holding a risk-free asset, and not bonds. Bonds, even government bonds, are actually considered risky assets because their prices are subject to interest rate risk. But this simple idea got modified over time into a rule-of-thumb about putting a part of your investments into bonds and a part into stocks. Try as hard as you may, you’re not going to find any rigorous theoretical justification for this. Except that empirically, in the 1960s and 1970s, the 60:40 portfolio seemed to work well as the most balanced portfolio allocation across stocks and bonds, which gave the best returns for a moderate level of risk.
In theory, if we consider bonds as another risky asset, we can still work with a 3-fund separation theorem. This would consist of the risk free asset, risky bonds and stocks. But investment management and advice did not quite work in this manner, partly because it was not easy back in the day for the average investor to borrow at anywhere close to the risk-free rate of interest. And especially not to invest in risky stocks and bonds. So this did not appear to have become a popular idea.
How has having bonds in the investment portfolio worked in the past?
In the past couple of decades, the idea of the 60:40 balanced portfolio, and its close cousins, like the 80:20 portfolio or the 50:50 portfolio have come under questioning for whether they are indeed sound (see here, here, here and even here). This is not surprising, as most of the received wisdom for portfolio construction has come from an era some 40 years in the past. Back when bond returns were higher, and stock returns were not so high.
Returns on US Stocks and Bonds by Decade
| Period | US Stock Returns (SPY) | US Bonds Returns (AGG) |
|---|---|---|
| 1970s | 4.89% | 10.23% |
| 1980s | 11.08% | 17.35% |
| 1990s | 7.38% | 15.72% |
| 2000s | 6.31% | -2.93% |
| 2010s | 14.82% | 4.06% |
| 2000-2020 | 10.48% | 0.50% |
What stands out clearly from the table above, is that there is world of difference in stock and bond returns before, and after 2000. It is not clear at the moment whether the post-2000 period will be the norm going forward, as far as stock returns are concerned, but it certainly looks that way for bond returns, given how low bond yields are at the moment. Hence, it is also clear why the old 60:40 portfolio has struggled in recent decades, something which retirees and believers in the 4 percent rule need to take note of.
Going one step further, how exactly have the stock-bond portfolios performed in the recent years, if we look at them from the perspective of the 3-fund separation theorem? Would an analysis of the efficient frontier still tell us that the 60:40 portfolio, or some variant of it would have performed well? To answer that, we are going to need some further information on the performance of US stocks and bonds in the past 2 decades.
Volatility and Correlations of US Stocks and Bonds
| Period | US Stock Returns Volatility (SPY) | US Bonds Returns Volatility (AGG) | Correlation |
|---|---|---|---|
| 2000s | 14.69% | 4.39% | 12.15% |
| 2010s | 13.57% | 3.09% | -1.86% |
| 2000-2020 | 14.34% | 3.59% | 5.24% |
Interestingly, we can see from the tables above that over the long term, volatilities are pretty stable and predictable. Correlations a little less so, but returns are all over the place! So while we may care most about the returns on our investments, it is only the risk (volatility) that we can be sure of!
To complete our model analysis, we also need to know the risk-free rate. We proxy by the Fed Funds rate over the past two decades.
Fed Funds Risk-Free Interest Rates
| Period | Average Fed Funds Rate |
|---|---|
| 2000s | 1.36% |
| 2010s | 2.16% |
| 2000-2020 | 0.63% |
Now that we have all the data needed to plot the Efficient Frontier for our 3-fund separation model, what do the results tell us about the value of holding bonds in our investment portfolios over the past 2 decades?
Efficient Frontier for US Stock and Bond Portfolios 2000-2020
From the chart above, we can see that for the entire 20 previous years, an investor in the US would have been better off holding a 100% allocation to stocks (proxied by the SPY ETF), and either having bank deposits to lower their overall portfolio risk, or borrowings to lever up their risk. Any holding of bonds (proxied by the AGG ETF) would have only led to poorer returns, both on an absolute basis, and on a risk-adjusted basis (i.e. measured using the Sharpe Ratio).
Simply put, why bother holding bonds over the past 20 years?
But we also know that both stock and bonds performed quite differently in the 2000s and in the 2010s. Would that have made any difference to the findings for the bond asset allocation?
Efficient Frontier for US Stock and Bond Portfolios in the 2000s
Looking at the chart above for the 2000s, the answer is “No“. The investor is still better off holding 100% stocks and zero bonds in the 2000s. Now, we can see why the 60:40 portfolio has come in for so much soul searching. Lacking a coherent theory behind it, it has underperformed massively when interest rates fell to their lowest for decades in the 2000s.
But is that the same case for the 2010s? After all, stocks outperformed bonds by a wide margin in the 2010s, especially in the last few years of the decade. Let’s take a look.
Efficient Frontier for US Stock and Bond Portfolios in the 2010s
We are in for a big surprise in the 2010s! Unlike the previous decade, and as a result of the incredibly low risk-free rate over the past decade, the efficient frontier analysis actually shows that the 3-fund separation efficient portfolio is actually one made up of an allocation of 80% to bonds, and only 20% to stocks! And this is so even when stocks massively outperformed bonds in the past 10 years! In fact, it appears that the past decade has been a time when bond performance has been more similar to the period before 2000. That is, when the Sharpe ratio for bonds was higher than that for stocks. The slight risk-adjusted advantage that bonds had in the 2010s meant that the efficient frontier as shown above has been tilted back towards bonds, although not in the ratio one would have expected.
Given the ease of investors getting financing currently at very low interest rates for investments, either through margin financing by brokers, or a pledged asset line with banks, the efficient frontier analysis above suggests that the 2010s were the era for successful use of the risk-parity portfolio (the AQR version, not the Bridgewater version).
In short, the idea behind a risk-parity portfolio is to allocate investments to stocks and bonds according to their risk, rather than their returns. For example, in a traditional 80:20 portfolio, although stocks account for 80% of the returns of the portfolio, they also account for almost 95% of the total risk. And in a 60:40 portfolio, stocks will account for 86% of the risk while only contributing 60% of the return. As a result, the traditional 60:40 or 80:20 portfolios, while marketed as helping to reduce the risk for the investor, simply do not do a good job of it.
A risk-parity portfolio, on the other hand, puts a greater proportion of the investments in bonds, something between 66% to 80%, and much less in equities. The idea is to equalise the risk between stocks and bonds. But it uses leverage to scale up the portfolio to the desired overall level of risk. In the chart for the portfolio in the 2010s above, the investor can allocate between stocks and bonds in a 20:80 ratio, and then use leverage to scale the portfolio up to a 10% volatility/risk level. This in turn, will achieve returns which are higher than a 80:20 or even 100% stock portfolio, for less risk!
This is also something which we have written about previously here.
Investing in Stocks and Bonds in Singapore
So far, everything we have talked about is US-centric. Given how much interest there is in investing overseas and in the US especially, this is not a bad thing. But curiosity leads us to ask: Does the same apply to our local market as well?
Now, there is far less data for the Singapore market, so all we’ll do here is to proxy stock returns using the ES3 ETF, and bond returns using the A35 ETF, from 2008 to 2021.
| Singapore Stocks (ES3) | Singapore Bonds (A35) | |
|---|---|---|
| Returns | 5.19% | 2.86% |
| Volatility | 17.61% | 4.54% |
| Correlation | 0.29% |
Using this data, and assuming that the risk-free rate (at which a pledged asset line is made available) is 1%, we chart out the following:
Efficient Frontier for Singapore Stock and Bond Portfolios 2008-2021
Voila! The results for the Singapore stock and bond investments over the past decade and-a-half yield almost exactly the same outcomes as the US data! In short, the optimum market portfolio to hold is one consisting of 80% in bonds and only 20% in stocks. Again, while stocks outperform bonds, on a risk-adjusted basis, bonds have a better Sharpe Ratio. Hence, they get a higher weight in the efficient portfolio.
Moreover, at a borrowing cost of 1%, the investor can, and should, leverage up to a level of risk/volatility he or she is comfortable with, in order to obtain better returns than a 100% stock portfolio! This is an interesting enough outcome, with actionable implications that we would like to share our thoughts on building a risk-parity portfolio using similar building blocks. But that is a matter for another post.
Conclusions: Use bonds to lever up the portfolio?
The efficient frontier analysis for US investments in the 2000s does not come as a surprise, showing that holding bonds would have done nothing for an investor . This is despite advice to hold a portion of an investment portfolio in bonds. However, the results for both the US and Singapore investments in the past decade does come as a surprise, showing that bonds still do have a role to play in our portfolios. However, it is not the traditional role of reducing risk, but for using leverage on a risk-parity basis to achieve higher returns with lower risk compared to a 100% stock portfolio. It also helps that bonds will give a higher loan-to-value ratio for financial institutions to extend loans against, and so are doubly useful for building a risk-parity portfolio.
Bonds are still useful in an investment portfolio. Not in the traditional sense of reducing risk, but as a building block for a leveraged risk-parity portfolio
A note of caution, though. What we have seen can explain the popularity of risk-parity investment strategies of the past decade, but who is to know whether these trends in bond and stock risk and returns will persist going forward? Perhaps the future looks more like the 2000s rather than the 2010s? In which case, investment strategy will have to adapt as well, perhaps going back to the era of 100% stock portfolios once more!
Finally, building such a leveraged risk-parity portfolio takes quite a bit of effort. But now, there is an ETF which allows you to do so quite easily!
There is line of thought that the 60:40 portfolio split between stocks and bonds is a way of diversifying the risks of the portfolio. But the problem is that it ends up reducing the return on the portfolio while also reducing the risks. If we define diversification more narrowly, it means getting the same return for less risk. For example, when we diversify a holding of a few stocks with more stocks, we are essentially maintaining the expected return on holding stocks while having lower risk, through the diversification between dissimilar stocks. So the 60:40 portfolio is not an efficient way to diversify risk, since it results in lower expected returns.
What the two fund separation theorem tells us is that we can split the diversification decision and the overall level of risk into two decisions or steps: Firstly, diversify with more risky assets to achieve the highest possible return for a specific level of risk – this is the optimal portfolio on the efficient frontier. Then use leverage or deposits to scale up or down the overall level of risk according to your risk tolerance.
For example, it may end up that the 60:40 portfolio is the optimal portfolio where the securities market line touches the efficient frontier. The the further step is to use leverage to scale the portfolio up or down to get to the desired risk level. The data from the 2010s show that the optimal portfolio turns out to be the 20:80 portfolio, with a large holding of bonds. Using leverage to scale this back up to a 100% stock portfolio level of risk gives use the AQR risk-parity portfolio. If we had constructed the optimal portfolio using more assets in additional to just stocks and bonds e.g. commodities as well, and then scaled it up using leverage, we would end up with something like the Bridgewater risk-parity portfolio.
Hence it is interesting how an old theorem in finance from the late 1950s ends up being relevant to the portfolio construction methods 50 to 60 years later!