Bob French, CFA

Some People Are Saying Bonds Don’t Provide Diversification Benefits, and They Couldn’t Be More Wrong

Recently, one of our clients asked our opinion on an article he found. We get these types of questions pretty frequently, and normally we just walk clients through what the article means for their portfolio (and often, why it’s wrong).

But this one caught my eye because it is wrong in a new and different way. Plus I’m on a statistics kick, and this article plays into that really nicely. (You can see my last two articles on understanding risk and uncertainty here and here. I’m on this kick because I’m actually in the middle of developing an investment eCourse)

The basic gist of the article is that bonds don’t actually provide any diversification benefit when you use them with a stock portfolio. The reduction in both your portfolio’s returns and standard deviation are really the result of reducing your equity exposure – not the diversification benefit from the bonds.

The article also claims that “bonds merely serve as a return booster to cushion the opportunity cost from a lower equity weight” and that bonds went down with stocks during the global financial crisis (I’m just going to ignore the suggestion that you buy bonds on margin).

Let’s walk through these claims one by one and talk about why they are so wrong.

Before we begin, I just want to stop and make sure we’re all on the same page. Diversification is a good thing. It’s the only free lunch in finance. It helps you reduce your portfolio’s standard deviation, and also allows you to focus on the risk factors that have positive expected returns. We like diversification.

OK, back to the article at hand. Its most “damning” piece of evidence that bonds aren’t good diversifiers for stocks is that the correlation between an all-stock portfolio(1) and an 80/20 stocks/bonds portfolio(2) is really high.

Put mildly, this is wrongheaded. While the correlation is really high – from January 1976 to March 2017,(3) the correlation between the two portfolio types was 0.996 – this doesn’t actually tell us much from a diversification or portfolio construction perspective. It’s tempting to think that the returns between the two portfolios must be basically equivalent since they’re so highly correlated.

This is clearly not the case:

⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀

  Annualized Returns Standard Deviation
100% Stock Portfolio 11.66% 15.18%
80% Stock / 20% Bond 11.03% 12.41%
Data from 1/76 – 3/17. Stocks represented by the CRSP 1-10 Total Market Index and bonds represented by the Bloomberg Barclays US Aggregate Bond Index. Indexes not available for direct investment. Past performance may be higher or lower than future performance.

⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀

The portfolios don’t seem all that different from a return and standard deviation perspective, but they are clearly not identical. So what’s actually going on here? Or better put, what do correlations actually show?

Most people assume that correlation measures how different two sets of numbers are from each other – a correlation of 1 would mean that they are the same series, and a correlation of -1 would mean that you just multiplied everything by -1. This is incorrect. Correlation actually measures the relationship between two different series (though it still doesn’t imply causation). Basically, it asks this question: “If I give you a number from the first series, how much can you tell me about the corresponding number in the second series?”

When you look at it from this perspective, it becomes pretty apparent why an 80/20 portfolio has a really high correlation with the stock market. The volatility of the stock market monstrously swamps the volatility of the bond market. Not only do bonds not move around as much as stocks, they only comprise one-fifth of your entire portfolio.

So if you tell me what the stock market did in May 1997 (it was up 7.23%), I can have a pretty good idea what an 80/20 portfolio did in May 1997 (it was up 5.97%). The returns of the 80/20 portfolio will probably be closer to 0 than that of the stock market, but the bond portion won’t move the needle all that much in most cases.

In terms of diversification, what you really want to look at are the component pieces.  How related are the pieces of the portfolio? How much do their movements explain each other?

This will tell you whether you are getting any diversification benefit for combining the two asset classes. You want stuff that will zig while everything else is zagging.

Stocks and bonds have a pretty low correlation with each other. Again, from January 1976 to March 2017, stocks and bonds had a correlation of 0.20. They’re the classic example of diversification.

An easy way to check this is to look at the monthly returns distribution. Since we’re working with an 80/20 portfolio versus 100% stocks, we would expect the returns to be closer to 0, but simply reducing the equity exposure won’t cause a month with a negative stock return to end up positive for the 80/20 portfolio.

If there is a diversification benefit from adding bonds to an equity portfolio, then there should be some months where stock returns were negative, but the portfolio had a positive return. In other words, the 80/20 portfolio should have fewer negative months than the 100 percent stock portfolio, which is exactly what we see:

⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀

  % of monthly returns below 0%
100% Stock Portfolio 36.57%
80% Stock / 20% Bond 35.15%
Data from 1/76 – 3/17. Stocks represented by the CRSP 1-10 Total Market Index and bonds represented by the Bloomberg Barclay’s US Aggregate Bond Index. Indexes not available for direct investment. Past performance may be higher or lower than future performance.

⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀

The difference may not look all that big, but this is a big deal. By merely incorporating a small amount of bonds, nearly 4 percent of months that showed losses now showed gains.

This takes a lot of wind out of the sails of the claim that the 80/20 portfolio’s reduction in standard deviation compared to the 100% stock portfolio is based solely on the lower amount of equity in the portfolio. If the difference between the returns was a direct result of reduced equity risk – plus a little bit of a “return boost” from the bonds – then we wouldn’t expect to see this much of a reduction in the number of negative months (there would be some effect, but it would have been smaller)

But let’s think about that “return boost” for a second. The article tosses this in pretty casually – that bond returns “cushion the opportunity cost from a lower equity weight” – but like most of the article, it flirts with some interesting ideas but ultimately fails to make the right connection.

The idea of making trade-offs between different parts of your portfolio (specifically the different risk factors you want to target) is incredibly important when thinking about portfolio construction. Capturing more of the value premium means you can’t capture as much of the term premium.

But that’s not what the article was getting at. It was simply saying that the returns from a total bond portfolio are higher than sticking with cash – as if this is free money, so why not take it?

There’s still a problem: the difference between the returns of cash (or, more specifically, really short-term US Treasury notes) and the broader bond market comes down to the level of risk. The market’s not just going to give you better returns because you’re foregoing stock returns and deigning to buy bonds.

Just like everywhere else in the market, risk and return are related. You can’t get returns above the risk-free rate (or the cash return we are talking about) without taking on some level of risk. It just doesn’t work.

Aside from the diversification that the article is urging people to forego, there are no free lunches in finance. All returns come at the cost of taking on the corresponding level of risk.

The last piece I want to address is one of my personal pet peeves. Admittedly, this is not one of the main claims pressed by the article, but it suggests that bonds aren’t good diversifiers because they don’t protect you during a crisis. The specific claim that the article made was that “during the worst five months of equity returns (all of them fell between June 2008 and February 2009) the bond index lost value as well.”

First, this claim is completely accurate. The Bloomberg Barclays US Aggregate Bond Index was down during the five months (I presume) the article was talking about – June, September, and October 2008, and January and February 2009.

But here’s the thing: the bond losses compared to the equity markets were tiny. The bonds lost on average 8.85 percent of what the stocks lost during those five months. And those were pretty close to full-on market meltdowns.

People wanted to get out of anything risky (including corporate bonds, which make up a really big chunk of the aggregate bond index) and into the safest stuff they could get their hands on (US Treasuries). The term “flight to quality” doesn’t properly convey the fear most people were feeling.

But just like it’s a bad idea to legislate based on exceptions, it’s a bad idea to invest based on them, too. So let’s take a longer term look at how often both stocks and bonds are down.

⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀

% of months when stocks were down 36.77%
% of months when bonds were down 31.11%
% of months when stocks and bonds were down 14.55%
Data from 1/76 – 3/17. Stocks represented by the CRSP 1-10 Total Market Index and bonds represented by the Bloomberg Barclays US Aggregate Bond Index. Indexes not available for direct investment. Past performance may be higher or lower than future performance.

⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀

First, it’s striking to see just how often bonds had negative returns – they were down almost as often as stocks. Of course, the bonds’ negative returns weren’t nearly as severe as the stocks’ losses.

But notice they didn’t line up – they were both down only 15 percent of the time. In other words, bonds were only down a little less than 40 percent of the time when stocks were also down. So, a little more than 60 percent of the time when your stocks were down, your bonds would have been up. And this makes sense, since so much of the US Aggregate Bond index is made up of corporate bonds. If a company is running into trouble, both its stocks and bonds will take a hit. The company is worth less, and it’s at least a little bit more likely that they won’t be able to pay off their bonds in the future.

And this gets even more apparent when you just look at short-term government bonds. Over this same period, one-year US Treasury notes were only down 10 percent of the time when stocks were down. If we go really short – one-month US Treasuries – they were only down 0.55 percent of the time when stocks were down. In other words, they zigged when the market zagged.

This is a really longwinded way of saying that, yes, bonds do provide diversification to a stock portfolio.

To find out more about how to actually build a portfolio that works for you, read our eBook 12 Principles of Intelligent Investors.

(1) The article uses Vanguard’s Total Market Index ETF. I’ll use the underlying CRSP 1-10 Total Market Index, which is the fund’s benchmark.

(2 & 3) The article uses the period from 7/31/2006 – 7/31/2016, but since we are using indexes, there is no reason to limit the data we are working with.

Retirement Researcher is a SEC registered investment adviser. The content of this publication reflects the views of Retirement Researcher (RR) and sources deemed by RR to be reliable. There are many different interpretations of investment statistics and many different ideas about how to best use them. Past performance is not indicative of future performance. The information provided is for educational purposes only and does not constitute an offer to sell or a solicitation of an offer to buy or sell securities. There are no warranties, expressed or implied, as to accuracy, completeness, or results obtained from any information on this presentation. Indexes are not available for direct investment. All investments involve risk.

The information throughout this presentation, whether stock quotes, charts, articles, or any other statements regarding market or other financial information, is obtained from sources which we, and our suppliers believe to be reliable, but we do not warrant or guarantee the timeliness or accuracy of this information. Neither our information providers nor we shall be liable for any errors or inaccuracies, regardless of cause, or the lack of timeliness of, or for any delay or interruption in the transmission there of to the user. RR only transacts business in states where it is properly registered, or excluded or exempted from registration requirements. It does not provide tax, legal, or accounting advice. The information contained in this presentation does not take into account your particular investment objectives, financial situation, or needs, and you should, in considering this material, discuss your individual circumstances with professionals in those areas before making any decisions.