Tuesday, November 6, 2018

Bonds can Outperform Stocks for Very Long Periods

It’s widely believed that over 30-year periods, stocks have always outperformed bonds. However, recent research says this isn’t true. U.S. bonds beat stocks over the 30 years ending in 2011, and it happened many times in the 1800s according to retired professor Edward McQuarrie at Santa Clara University. However, the important question is what should we do with this information?

Jason Zweig at the Wall Street Journal reported on this research in his 2018 Nov. 2 article Sometimes, It’s Bonds For the Long Run, a play on the title of Jeremy Siegel’s excellent book Stock for the Long Run. It’s doubtful that Zweig is recommending that investors consider whether bonds are the best source of long-term returns, but his readers could be forgiven for thinking this.

The next chart shows McQuarrie’s data on 30-year rolling periods. This means that the return shown in a given year is actually the average returns over the previous 30 years. So, near the end of the chart where it shows the bond return higher than the stock return, this means that bonds beat stocks from 1982 to 2011 (by a whopping 0.15% per year).

It’s hard to get much of a feel for this information when we talk of average returns. So, let’s switch to dollars. Suppose an investor makes two equal-size investments in stocks and bonds, and lets them ride for 30 years. At the end, what is the ratio of the ending value of the stocks to the ending value of the bonds? The next chart shows this.

Consider our hypothetical investor who made two equal investments in stocks and bonds from 1982 to 2011. If the bonds had finished at $100,000, the stocks would have finished at $95,900. If we repeat this experiment for any other 30-year period in the past century, we get a much different story. The most extreme case is 1942 to 1971 where if bonds ended at $100,000, stocks ended at $2,130,000! For all 30-year periods ending between 1948 and 2001, stocks always more than doubled bonds.

The older data tell a different story. During much of the early 1800s, bonds beat stocks over 30-year periods. Most of the time, stocks lagged by less than 1% per year, but there was one period where bonds outperformed by a cumulative 94%. That would have been a good time to own bonds, but it’s nothing close to the dominance of stocks during most of the past century.

Zweig warns that history may repeat itself, possibly “in a way that investors who have all their money in stocks should hedge against before it’s too late.” Unfortunately, while hedging improves worst-case returns, it hurts long-term expected returns.  Perhaps Zweig's warning is at least partially a commentary on how the bull run in stocks since 2009 has to end in a crash at some point.

So, what should investors do? Should we be increasing our portfolios’ bond allocations? Should we be ready to switch to bonds if we see a repeat of the conditions that led to bond dominance in the early 1800s? For my own portfolio, my answers to the latter two questions are no and no. I already have sufficient fixed income in my portfolio.

Nothing has changed. Stocks are a better bet than bonds over the long run. Bond allocations remain a useful way to control a portfolio’s volatility. For any money I won’t need for a long time, I’d rather ride out stock market volatility than give away expected returns by owning bonds. Your mileage may vary.


  1. Interesting article. My portfolio is 50/50 or 60/40 depending on whether one considers my $90K in BMO ZPR Preferred Shares ETF to be FI or equity. Regardless, I will keep my portfolio of bond,prefs and high dividend ETF's the same for the foreseeable future.

    1. @Marko: You and I choose different investments, but I agree that this new information shouldn't be a reason to choose a new asset mix.

  2. In a sense we are too focused on recent history (everything since 1950). The nature of the economy is changing and we may see times that are more similar to the 1800s or different from anything we've experienced in the past.

    However the stock market was not nearly as well developed in the 1800s and that probably contributed to the differences. It's amazing that the returns actually ended up being so close to what we've known lately since it was much harder to get information or trade stocks! Despite relying a lot more on blind faith investors actually did well according to this data.

    In the end if bonds have higher returns that does violate some basic common sense since investors wouldn't have a reason to take on the risk of stocks. That might be moderated to a degree if bonds are extremely risky (possibly the case in the 1800s) or stocks are highly regulated (possibly the case going forward). Otherwise it should be a sign that changes are coming.

    My personal position at the moment is summed up by the headline of a newsletter the WSJ sent out this morning: "The Strategy of Combining Stocks and Bonds Backfired This Year".

    1. @Richard: I think the reason bonds beat stocks during the 30 years ending in 2011 is that interest rates dropped steadily during that period. It's not possible for interest rates to drop that much from current levels. So, the only way for bonds to beat stocks again in the coming decades is if stocks perform spectacularly poorly. Of course, this is one possibility out of many.

    2. Another way would be if investors start with expectations of high defaults on bonds and then it doesn't happen. That would essentially be the same thing (yields go from high to low) for a different reason. The expectations of defaults don't appear to be high now.

    3. @Richard: That could contribute, but even junk bond rates would have to go negative to have interest rates fall by as much as they did from 1982 to 2011.