Are Stocks Really Riskier Than Bonds?
Updated for 2019 data
Stocks are generally considered to be more risky than bonds. This article provides the data, in graphical form, so you can see and decide for yourself if stocks have really been riskier than bonds.
First we will look at portfolios that are 100% stocks vs. 100% bonds or 100% cash. Then we will look at balanced portfolios.
For short-term investors, stocks are indeed riskier than Bonds. But for long-term investors the evidence from actual historical returns indicates that Bonds were actually riskier than stocks. But it all depends on having a proper definition of what risk means.
When it comes to long-term investors, virtually the entire investment community is focused on the wrong definition of risk. Much of what is written about risk is at best inappropriate and at worse completely wrong for a long-term investor. This is caused by an over emphasis on short-term volatility.
For long-term investors we need to have a proper definition of risk. Financial academics and the investment community generally define risk as the short-term (annual, monthly or daily) volatility of returns from an investment. The volatility of returns is measured by variance or standard deviation.
From the perspective of a long-term investor, this definition of risk is flawed for two reasons:
- The analysis and conclusions almost always focuses on the volatility of annual (or even monthly or daily) returns. An annual focus might be appropriate for many investors, but long-term investors should be mostly concerned about risks associated with their long-term wealth level and not primarily focused on the bumps along the way.
- The analysis and conclusions are almost always based on nominal returns and ignore the erosion of purchasing power caused by inflation. For short-term investors, inflation may not a be a big concern but it has a huge impact in the long-term.
As to the second point above, it seems self evident that better conclusions will be reached using real (inflation adjusted) returns rather than nominal returns.
As to the first point above, under the short-term volatility definition of risk, stocks are considered much more risky than long-term Bonds or Treasury Bills (T-Bills). Yet, it is a fact (based on United States data since 1926) that stocks have outperformed both 20-year government Bonds and T-Bills over all historical calendar 30 year periods. In all of the 65 different 30 calendar year “rolling windows” starting from from 1926 – 1955 then 1927 – 1956 all the way through to the most recent period being 1990 – 2019, stocks have provided a higher return. But due to higher annual volatilities, stocks are considered more risky! (Admittedly, there would be some 30-year periods based on non-calendar year start and end points where long term bonds out performed stocks.)
To avoid risk (defined as annual or daily volatility) you may be advised to put substantial portions of your investments into into Bonds or T-Bills even though the historic data suggests these are in fact are almost guaranteed to under perform stocks in the long run. This kind of thinking on risk, while it may allow you to sleep better, may be hazardous indeed to your long-term wealth. (That is, if your goal is wealth maximization at some distant point like 20 or 30 years in the future, as it is for many investors, and particularly younger investors).
The following graph shows the actual annual volatility in Stock, Bond and T-Bill returns from 1926 to present. In regards to stocks, this discussion deals only with the performance of the S&P 500 large company stocks as a group; it does not deal with the risks of investing in a non-diversified portfolio of stocks. The bonds are 20-year U.S. government bonds and the investment is rolled over into a new 20 year bond each year to keep the maturity always at 20 years. The data here is for U.S. returns as published in the Ibbotson yearbook entitled, Stocks, Bonds, Bills and Inflation. The returns are total “real” returns including dividends and capital gains or losses adjusted downward each year by inflation. (In the depression years the real returns are adjusted upwards by deflation since “inflation” was negative.)
Sure enough, the stock returns (the blue bars) were clearly far more volatile on an annual basis. Long-term Treasury Bond returns (the red bars) were also quite volatile while T-Bill returns (in green) were more stable. It’s also fairly obvious that the average stock return is much higher than the average Bond return which in turn is much higher than the average T-Bill return.
On a real return basis, stocks have had calendar year losses of over 30% in four of the 91 years from 1926 to 2019, with the latest occasion of course being 2008. And two (1930/1931 and 1973/1974) of those occasions included an adjacent calendar year with a loss of at least 20%, meaning the total compounded loss was over 60%! Using daily data there would be more occasions when stocks have plunged at least 60% from a previous peak. That is certainly real risk and is extremely hard to stomach. Yet we know that despite this, stocks have clearly out-performed bonds in the long run.
The age-old question for investors is whether or not the (highly probable but perhaps not certain) higher long-term average return from stocks justifies the extra risk (short-term volatility).
In judging the risk of Stocks versus Bonds, you must consider more than the annual volatility. The following illustrates this.
Imagine your rich uncle offers to play a coin toss game with you. If you lose he gets half your net worth. If you win he gives you an amount equal to twice your net worth.
Your expected return is 0.5 * (-0.5) + 0.5 * 2.0 = 0.75 or 75%.
So on average you will win 75% of your net worth but you have a 50% chance of losing half your net worth and a 50% chance of tripling your net worth.
Should you play this game? Simple expected value math says yes, but most people would consider it too risky and would not play. It would be a real downer to lose half your net worth on a coin toss. (If in doubt, a male could ask his wife, she would likely have no doubts).
How risky is this game? It’s very risky unless you are allowed to play several times. But imagine now if your starting net worth were $100,000 and your rich uncle said you could divide your money into ten piles and play the game 10 times, each try based on $10,000. If you win 5 times and lose 5 times, you will win $100,000 and lose $25,000 to net $75,000 ahead. If you win only 2 times and lose 8 you win $40,000 and lose $40,000 to break even. So now you can only lose if you only manage to win less than 2 coin tosses out of 10. This changes things drastically. It now seems much more sensible to play the game since you have little chance of losing. Your expected return is still the same 75% but your risk is much reduced (though not eliminated).
This illustrates that in looking at any risky venture it is important to ask how many times you get to play the game. If the average return is positive, the risk declines with the number of times you get to play and if you are allowed to play many times then the risk approaches zero. Technically speaking, the standard deviation of your total return over “N” tries is equal to the standard deviation of each individual try divided by the square root of “N”. As “N” becomes large your total risk becomes very small.
Similarly, the stock market becomes much less risky, if risk is defined as a lower possibility of failing to beat bond or cash returns over your full holding period, the longer you stay in.
In regards to the stock market and the range of possible returns you simply cannot analyze your risk unless you first consider how many years that you will be investing. It is generally accepted that the average investor is concerned very much with annual and even daily returns and has perhaps a one year time horizon. Most of the discussion you will ever see about stock market risks will focus on the one year or even daily volatility. That might be fine for the mythical average investor but it leads to completely wrong conclusions, for true long-term investors.
For a long-term investor, I would argue that the more relevant risk is clearly the risk of insufficient growth in long-term real purchasing power. Analysis which focuses on the risk of short term volatility in wealth or returns is quite simply looking at the wrong risk when it comes to the long-term investor.
ACTUAL RETURNS IN THE MARKET STOCKS vs. LONG GOVERNMENT
BONDS vs. T-BILLS
Each point on the above graph shows the compounded annual return from holding each asset class for 30 year periods ended in the year shown on the “X” axis. For example, the leftmost points on the graph indicates that the average compounded return from holding stocks (the S&P 500 index stocks) for the 30 year period started in 1926 and ended in 1955 was about 8.5% while holding the twenty year treasury bond (and selling the bond at the end of each year to purchase the latest 20 year bond to maintain a constant maturity of 20 years) earned about a compounded 2.1% and holding treasury bills for that 30 year period earned just less than 0.0%. In all cases the returns are after accounting for inflation, omit taxes, omit trading costs and assume reinvestment of all income.
The above graph shows that for the 65 different 30 calendar year rolling investment periods ranging from 1926 through 1955, all the way to 1990 – 2019, real (after inflation) stock returns were higher at the end of 30-year holding period than 20-year government bonds (Although for the 30 year period ended in 2011 it was a virtual tie). In most of the 30-year periods, stock returns were very significantly higher.
Note that stock portfolios that were set up in 1928, just before the massive stock crash of 1929 – 1932, and ending in 1957, still beat bonds – and by a huge amount.
Stock (S&P 500 total return index) real returns for 30-year holding periods ranged from a compounded return of 4.4% to per year to a maximum compounded return of 10.6%, with an average of 7.2% across the 65 different 30-year holding periods.
T-Bills which are supposed to be safe are almost a guarantee that your return will at best barely outpace inflation in the long run. Treasury Bills always returned less than a compounded 2% (after inflation) over 30 year periods, and often returned less than 0.0% as T-Bills failed to even compensate for inflation. The average real T-Bill return for the 65 different 30-year investment periods was 0.4%.
20-year Treasury Bond real returns for 30-year holding periods ranged from a compounded return of minus 2.0% per year to a maximum compounded return of 7.8%, with an average of 1.8% across the 65 different 30-year holding periods. Treasury bonds provided compound average real returns over 4% in the 30-year periods that ended after 2000 mostly due to the high interest rates that prevailed in the late 1970’s and the 1980’s and due to the huge drop in interest rates over those 30 year periods.
To my way of thinking, this graph illustrates that over 30 year investment periods stocks have not been riskier than long-term government bonds given that they always outperformed bonds over these historical 30 calendar year holding periods. Also, the real, after inflation, compounded annual return from the S&P 500 index was always at least 4.4% while bonds provided negative returns in the 30-year periods ended from 1964 to 1985. If the future is like the past, then we might be justified in concluding that stocks will likely continue to outperform bonds over the next 30-year investment period. We are therefore justified in expecting stocks to continue to outperform over future 30-year periods.
Admittedly, what the above 30-year rolling returns graph does not show is how volatile the annual returns were. For example, in any given 30 year period, stocks always won out in the end, but they did subject the investor to horrifying volatility along the way as illustrated in our first graph above.
The 30 year stock returns are indeed riskier in terms of annual volatility of results but clearly are not riskier in terms of what really should matter most (to long-term investors), long-term growth in wealth and purchasing power. It seems wrong-headed to consider a real return that historically varied between 4% and 10% per year, after inflation, (this is over 30-year periods) to be riskier than the return on treasury bills, which never got above 2.0%, after inflation.
Most investment theory teaches that the risk versus return trade-off is a matter of personal preference. The stock market offers higher average expected returns on stocks but at the cost of higher annual volatility. To their credit, the industry encourages those with longer term time horizons to use a higher equity weighting but still advises that all investors allocate some funds to Bonds and Bills. But this really offers little guidance to investors.
My conclusion is that the risk return trade-off is more a matter of time horizon and education, rather than personal preference. If you are virtually certain that you will not require the funds prior to 20 years or more, then history teaches that stocks have not been riskier. Stocks will almost certainly return more than Bonds or Bills (based on historic data).
If investors are educated about this then most of them can become more comfortable with the daily, monthly and annual volatility of stocks safe in the knowledge that high returns in the long-term is their almost certain destiny. It’s a bit like taking a drive on a mountain road with a lot of switchbacks. If you don’t know the road, the switchbacks and back-tracking might be highly stressful (as you think you are going in the wrong direction). But if you have studied a map carefully then you can relax and the switchbacks will not bother you since you know that the road to your destination will be circuitous.
Also, those investors who are in the savings phase of their lives are not reliant on any one thirty year period but instead typically invest a certain amount each year. This reduces risk through time diversification. If in any individual 30-year holding period stocks are highly likely to beat bonds, then in a long series of 30 year holding periods it may be virtually certain that stocks will beat bonds.
Shorter time periods
For 15-year holding periods there are a few periods where bonds beat out stocks. Since stock portfolios began in the late 20’s (and ending 1940 to 1944) or that began at the end of 1993 through the year 2000 (and ending 2008 to 2016) did not out-perform long-term government Bonds over the next 15 years, it might be wise to refrain from investing 100% in stocks if the money is needed in 15 years and especially during times where stocks seem to be clearly over-valued, if one can identify that. However, another consideration is that long-bonds did well during periods of declining interest rates. An investor in 2020 may wish to consider the probability that interest rates will continue to decline as well as the risk that they may rise.
For 10-year holding periods there are still not very many periods where bonds beat out stocks. However, we do see that in the five 10-year periods ending 2008 through 2012, bonds did beat stocks by a significant amount. Overall it appears that the range of real, after inflation, bond returns is relatively large from minus 5% compounded for ten years to over 10% compounded per year. It is apparent that the average return from stocks over the many 10-year holding periods has been significantly higher. Interestingly, the worst case scenario for stocks was not quite as bad as the worst bond scenario.
Before concluding that a 100% allocation to stocks is a reasonable approach even for very long-term investors, we need to take a look at how diversification through balanced portfolios can affect the outcome.
Most investment advice advocates holding a balanced portfolio of stocks, bonds and cash. It is sometimes claimed that due to dollar cost averaging balanced portfolios can provide both higher returns and lower risks. So let’s take a look at the average returns over 30-year periods using balanced portfolios.
The above graph illustrates that over the 65 rolling 30 year periods ending in 1955 through 2016, Balanced portfolios have under-performed 100% stock portfolios in the earlier decades. However for time periods ended in recent decades, the balanced 70% stocks / 30% bonds portfolio often marginally beat the 100% stock portfolio and never lagged stocks by much.
The strong performance of the balanced approach in the more recent periods leads many to conclude that this result is to be expected in future. That’s probably not so. Balanced portfolios nearly kept pace with (or even beat) stocks in more recent years only because of the dramatic drop in interest rates. Now that interest rates are at historic lows, that situation will almost certainly not occur for 30 year periods that are starting now.
Proponents of Balanced Portfolios often argue that you will only give up a small amount of return and will get lower annual volatility. But note that a $100,000 stock portfolio growing at say 8% (after inflation) grows to $1,006,300 in 30 years, while a Balanced portfolio growing at say 6.0% grows to only $574,300. So, the stock portfolio in this case is worth a hefty 75% more. So much for giving up a “small” amount of return!
Balanced approaches, however may be advantageous for shorter periods, as the following illustrates.
The above graph illustrates exactly what would have happened to an investor following a partly balanced approach of 70% stocks and 30% bonds (the yellow line) or following a more fully balanced 50%, 25%, 25% stocks, bonds and cash (bills) allocation (the purple line). This assumes U.S. data and that stocks are represented by the S&P 500 index stocks, bonds by 20-year U.S. government bonds and cash by 30-day U.S. government Treasury bill investments. This assumes annual re-balancing to maintain the balanced allocation and to maintain the 20-year maturity of the bonds. There is no speculation in this data, only historical reality.
The result is that the partially balanced portfolio significantly under-performs a 100% stock portfolio in many of the 80 different rolling 15 year periods. The partially balanced approach was superior only for 17 periods which were those that that started in 1926 through 1931 and ending in 1940 through 1945 and as well in the 15 year periods ended 2002 and in 2007 through 2016 and the period ended 2018 (so a lot of recent periods).
One thing that the graph illustrates is that for 15 year holding periods, all of the asset classes gave poor real returns for the 15 year periods ending approximately 1975 through 1986. And being balanced among three poorly performing asset classes was of no help.
Overall, the graph above provides good support for the notion that a balanced portfolio such as 70% stocks, 30% bonds is a good approach for investment periods of around 15 years.
The balanced 70% stocks / 30% bonds portfolio sometimes outperforms the 100% stocks approach over these ten year holding periods – especially in the more recent periods. And it rarely trails the 100% stocks approach by much. The 50% Stocks, 25% bonds, 25% cash portfolio appears to consistently under perform the 70% stocks approach. It would be reasonable to conclude from this graph that the 70% stocks approach is a good approach compared to 100% stocks for ten year holding periods.
In regards to stocks, this discussion deals only with the performance of the large stocks comprising the S&P 500 index as a group it does not deal with the risks of investing in a non-diversified portfolio of stocks.
For shorter-term investments the stock market is very risky compared to Bonds and short term treasury Bills. The average return from stocks has been consistently higher over long periods but over shorter periods (anything under 10 to 15 years) the results from stocks are hugely uncertain. It would be most unwise indeed to invest money needed next month or next year or even prior to about 10 to 15 years in 100% stocks
As the time horizon lengthens, we reach a point where stock returns are almost (but never quite) certain to exceed Bond and Bill returns – at least based on historical calendar year U.S. results from 1926 through 2019. For time horizons exceeding 15 years it seems quite unlikely that stocks will under perform Bonds and virtually certain that they will out perform Bills. With a 30 year time horizon it seems virtually certain (based on history) that stocks will outperform Bonds. A 100% (diversified) stock portfolio seems virtually certain to outperform, over 30 year periods, portfolios with any portion of the funds allocated to Bonds or Cash. And the case for stocks is all the stronger if you consider that people don’t typically invest a single lump sum for 30 years. Rather they invest on an annual basis which greatly reduces the exposure if one is unlucky enough to run into the odd period where stocks do trail bonds over a 30 year holding period.
This analysis was based on making an initial investment and letting it grow over time.
Of course, if one is capable of expertly timing the markets then it would be possible to beat the 100% stock approach in the long-term by “simply” being in the highest returning asset class each year. This will be attractive perhaps to psychics and clairvoyants. Mere mortals investing for 15 years or longer might wish to consider the 100% stock approach. However, investors that are uncomfortable with short-term volatility should use a balanced approach. And it may be realistic for long-term investors to move some money out of stocks if stock prices are in an obvious bubble.
Virtual certainty is not quite 100% certainty there is always some small chance that Bonds will outperform even in a 30 year time horizon.
You don’t have to agree with my conclusions. You can also study the graphs above and draw your own conclusions.
Self-described long-term investors need to be sure that they really have a long time horizon before they act accordingly. For many investors, there is a chance that they will need to cash out their investments early. This could be caused by illness, job loss, disability, legal problems and other reasons. But, if an investor is virtually certain that they have a very long time horizon then it certainly appears that stocks (based on a U.S. large stock index) are not riskier than bonds, in terms of achieving the highest ending portfolio value.
The above analysis shows that stocks tend to significantly outperform bonds and balanced portfolios as of the end of a 30-year holding period.
But what about the volatility along the way? And what about for more realistic scenarios like saving so much per year for 30 years rather than just making a one-time deposit and waiting 30 years. And what about retirement scenarios?
We have all of that covered in two related articles:
The first article shows what happened to portfolios for all the possible 30-year savings periods from 1926 to 1955 all the way to 1982 through 2011 invested in either 1. 100% U.S. stocks (S&P 500 index fund in non-taxable account) or 2. Invested 60% in stocks, 35% in corporate bonds and 5% in cash.
The second article shows what happened to one million dollar portfolios for all the possible 30-year retirement periods from 1926 to 1955 all the way to 1982 through 2011 invested in either 1. 100% U.S. stocks (S&P 500 index fund in non-taxable account) or 2. Invested 60% in stocks, 35% in corporate bonds and 5% in cash.
Shawn Allen, CFA, CMA, MBA, P.Eng.
President, InvestorsFriend Inc.
Article originally created in June 2001 and last updated October 20, 2020