Do Stocks Always Outperform Bonds?

A professor challenges stocks for the long run
 

By: Greg Obenshain

Stocks for the long run is as close to commonly accepted wisdom as there is in investing. The concept predates Professor Jeremy Siegel’s book, Stocks for the Long Run, but that book solidified the concept as iron truth. And since World War II, this has certainly been the case in the United States.

But since Siegel’s first edition of the book in 1994, a significant amount of new data has become available thanks to a large project led by Richard Sylla at New York University to collect stock and bond trading data back to the 1790s as well as the digitization of old newspaper records. And since claims about long-run performance of equities rely on having accurate long-run data, this old information is very relevant to Siegel’s thesis.

A retired professor in California, Edward McQuarrie, has devoted himself to incorporating this data to update and extend Siegel’s original data. McQuarrie thinks that long-run stock and bond returns might be closer than Siegel’s research has suggested. 

Siegel’s book was first published in 1994, and he used the best data available at the time. Below, we show the famous chart from the 2014 edition of Siegel’s book that shows the dramatic outperformance of stocks.

Figure 1: Total Real Return on Stocks and Bonds from Stocks for the Long Run, 1802-2012

Source: Siegel (2014)

McQuarrie both adjusted and extended Siegel’s data back to the early 1790s. McQuarrie details his efforts in a paper and shares his data on his website (you’ll need to use a Microsoft browser to download it). You can also listen to him describe his efforts on a podcast with Meb Faber.

For stocks, McQuarrie made major revisions to the pre-1871 data by including more stocks, cap-weighting returns, and, most importantly, correcting significant survivorship bias in the original data. According to McQuarrie, “Banks failed during panics, turnpikes and canals succumbed to railroads, and struggling railroads went bust in the 1840s and 1850s to an extent not previously understood. In short, Siegel’s sources had left out the bad parts, producing an overly rosy picture of antebellum stock returns.”

For bonds, McQuarrie engages in an impressive forensic effort to build a new and more accurate data set of investment-grade bonds available to the public. Previous efforts at debt data series had relied on prices imputed from yield and questionable choices such as choosing the bonds with the lowest yields to represent the risk-free rate. McQuarrie instead uses actual corporate bond prices to “discover how an investor in a hypothetical bond index mutual fund would have fared across the centuries.” The Treasury market has been far less consistent in structure and attributes than the corporate bond market, so the corporate bond market may be better suited to a long-term examination of stock versus bond returns. The resulting long-dated (15+ year maturity) bond series more closely approximates the long-dated high end of investment grade today (15+ year AA and AAA bonds). The total returns of this series are close to today’s long-dated Treasury returns, as we will show later. So Professor McQuarrie is comparing equity returns to long-dated very high-quality corporate bond returns.

Below is the fruit of Professor McQuarrie’s labor, a revised chart of the stock and bond return series over the same period as the chart shown above.

Figure 2: Total Real Return on Stocks and Bonds Using McQuarrie Data, 1802–2012

Source: www.edwardfmcquarrie.com and Verdad analysis

The first observation is that stock returns are lower (5.9% versus 6.6%) and bond returns are higher (4.1% versus 3.6%) in the revised data. McQuarrie also points to international evidence from Dimson, Marsh and Staunton, authors of Triumph of the Optimists, whose most recent world ex-USA equity return data show annualized real USD returns of 4.5% from 1900 to 2021 and 5.5% since 1972.

Figure 3: World ex-USA Real USD Returns, 1900–2021

Source: Dimson et al.

In all these charts, stocks do indeed outperform bonds, supporting Siegel’s thesis, but the returns of bonds and stocks are closer than Siegel’s original work found.

A more revelatory observation is that, in the first 150 years of McQuarrie’s data, bonds and stocks perform about the same. And they also perform nearly the same from 1982 to 2012. In fact, what emerges is that the equity outperformance that seems so evident in Siegel’s data is in fact concentrated in the period from around World War II to 1982.

To make this clearer, below we show a table of sub-period returns. And to alleviate concerns that the corporate series is dramatically different from the Treasury series, we also show long-dated Treasury returns for the periods in which they are available.

Figure 4: Sub-Period Annualized Real Returns for Stocks and Bonds

Source: www.edwardfmcquarrie.com, Treasury data from Morningstar Direct, and Verdad Analysis.
Data extended from 2018 through 2022 using CRSP US Total Market Index, US Gov/Credit Long Index (to approximate corporate AA), and CPI-U from Bloomberg.

For the first 150 years, stocks and bonds perform about the same. Then for 40 years stocks perform well and bonds lose money. And after that the data seems to revert to “normal,” with stocks outperforming by 2.7% through 2022.

McQuarrie argues that the modern period of equity returns beginning around World War II in the United States is not repeated in the long US history nor in international history. Below we show trailing 10-year returns for equities over the entire history and circle the period of equity outperformance to make it visually clear how different that period is.

Figure 5: Trailing Annualized 10Y Stock and Bond Returns, 1802–2022

Source: www.edwardfmcquarrie.com and Verdad analysis

Stocks have outperformed over the long term, both in the US and internationally, but there are long periods where that has not been the case. According to McQuarrie, “sometimes stocks beat bonds, sometimes bonds beat stocks, and sometimes they perform about the same.” If we look at the 10-year stock minus bond returns, we can see just how frequently stocks have returned less than bonds over 10-year periods, even in recent history.  

Figure 6: Trailing 10Y Annualized Stock Minus Bond Returns, 1802-2022

Source: www.edwardfmcquarrie.com and Verdad analysis

Professor McQuarrie points out that while the period of stock outperformance that began around WWII is not replicated internationally or in US history, there is no instance when debt outperforms stocks to the same degree. We believe equity has more ability to deliver surprisingly high returns. As volatile as the stock outperformance is, it does seem to have a central tendency somewhere between 0% and 5% for the last 120 years. Bonds seem to serve as a lower bound to stock returns, but sometimes stock returns fall below even that. And for taxable investors who pay taxes on coupon payments, there is a significant tax advantage to equities.

The strongest takeaway is that, while our best long-run estimate of the real equity return may be closer to 5% and our best estimate of debt returns is the real yield, neither is guaranteed. Stocks will probably outperform bonds over the long term, but in the specific long term that matters to most investors (the next 10, 20 or 30 years) we cannot be certain.