Performance                                InvestorsFriend.com                             Home

 Stocks, Bonds, Bills and Inflation  - Asset Class Performance

This article examines graphically the long-term performance of the three major asset classes of stocks, bonds and cash. The results are truly enlightening and amazing! The results are based on U.S. data going back to 1926. The data source is a well-known reference book called "Stocks, Bonds, Bills and Inflation" 2009 edition. The book is published annually by Morningstar. (Ibbotson SBBI classic yearbook)

Note that most analysis of historic returns that you have seen is horribly flawed in that it is based on "nominal" returns before inflation. The graphs and figures below are based on "real" returns after inflation. That is, this analysis shows the real increase in purchasing power generated by each investment asset class.

The graph below shows the real (after inflation) returns on large capital U.S. stocks (The S&P index), long term U.S. Treasury bonds, Long Term Corporate Bonds and 30-day cash investments (represented by U.S. Treasury Bills). 

Isn't that amazing? In real-dollar terms (adjusted for inflation), large U.S. stocks have absolutely walloped long bonds and short-term cash investments in terms of total return. Each $1.00 invested in stocks at the end of 1925 is now worth $174.75 (83 years later at December 31, 2008). The same $1.00 invested in long U.S. government treasury bonds for those 83 years is now worth $8.45. $1.00 invested in U.S. long-term corporate bonds in 1926 did slightly better than treasury bonds and is now worth about $9.82. $1.00 invested in T-Bills in 1926 is now worth just $1.75. Remember, all figures are after inflation and also assume tax-free and no-fee investment accounts. (With taxes the growth would be less dramatic but would be even more in favor of stocks given the lower tax rates on capital gains and dividends.)

This means that if old Grampa had foregone just 1 case of beer at the end of 1925 and invested the money in the S&P index of large stocks (and reinvested all dividends and rebalanced to stay with the index over the years), his grandson, at the end of in the year 2008, could go out and buy 1 case of beer and still have enough left to buy 173.75 more cases! This is truly amazing and is really a case where you can in fact have your cake and eat it too, if you just delay eating the cake and instead invest the money for a long time. (Later I will show that there are some pretty good returns over 20 year periods, so you don't have to actually invest for 83 years!). Note that the year-end market data peaked at the end of 1999 when the $1.00 in stocks had grown to $303.

This amazing out-performance of stocks (which beat corporate bonds by a factor of 174.75/9.82 or 17.7 to 1, in the 83 years has occurred in spite of the two huge stock crashes that have occurred since the year 2000, not to mention the stock crash of the great depression.

The Corporate Bond investor with only $9.82 after 83 years has only 5.6% of the amount that would have occurred in stocks. And the T-Bill investment at $1.75 has just barely kept ahead of inflation. 

The above graph which has a normal linear scale does a great job of showing the huge difference in the ending portfolio values but unfortunately is distorted in three ways.  First, the results from the earlier years are not really visible, Second, it looks like the percentage rate of growth for stocks was increasing toward infinity until 1999, and third it also looks like the early 2000's stock crash was by far the biggest market crash ever.  A logarithmic scale solves these problems because a constant percentage growth appears as a straight line and the percentage gains in the earlier years are much more visible. Unfortunately a logarithmic scale tends to somewhat obscure the huge differences in the ending values. When viewing a growing data series it is probably best to view it with both logarithmic and linear scales to better understand the results.

The same data presented in the above graph is presented below with a logarithmic scale.

In this graph (with the same data as above) you now have to look more closely to realize the amazing extent to which stocks outperformed bonds over the 83 year period. But this logarithmic scale allows you to view the volatility over the years.

A constant slope on this logarithmic graph represents a constant annual percentage growth. By the nature of logarithmic graphs, a dip or gain on this graph of a certain height represents the same percentage change whether it happened in the 1920's or the 2000's. Any dip or gain visible on this graph is actually large since a logarithmic scale tends to make percentage changes look small.

This graph reveals that stocks (blue line) were much more volatile than bonds, particularly  1926 - 1932, the mid 70's and in the last eight years. Again, remember that the graphs show real returns, adjusted for inflation. It is Interesting that the big stock market crash in 1987 is not apparent on this graph. The reason for that is the fact that the graph here shows only year-end figures. The big crash in 1987 was actually an event that happened within the year as stocks soared until October that year and then crashed. On a calendar year basis stocks were actually up slightly in 1987.

The Graphs below take the data above and break it out into 20 year periods and reveal some very interesting insights into asset performance in different periods. Note that the scales below are linear and that all the scales go to $8.00 (A 700% gain, from the $1.00 starting point). By using the same scale it is easier to visually compare the performance across the different 20 year periods.  Also note that in the graphs below, Treasury Bonds are included but not corporate bonds, this is because as illustrated above, the corporate bond return is quite similar to the treasury bond return and was only moderately higher. 

Bonds look like the place to be in the 1926 - 1945 period. Stocks beat out Bonds in the end but it was a rough ride indeed. The stock index returned 291% after inflation in the 20 year period while the Treasury bond index returned 148% and Treasury bills eked out 22%

The (relatively) unique thing about this time period was the huge stock valuation bubble in the late 20's followed by a bursting in late 1929, which was then exacerbated by poor government policies that led to the Great Depression. Note that the full extent of the crash is not visible in this graph because it uses only year-end, rather than daily data.

 

Wow 1946 - 1965, what a run for stocks, while bonds and cash (T-Bills) failed to even keep up with inflation. It's interesting to note that stocks would be considered to be much more risky, they increased in a volatile fashion while bonds were pretty flat and went nowhere. But if this is what people call risk, I'll take it! The stock index returned a whopping 664%, after inflation, while the long bond index investment lost about 21% and even so-called risk-free Treasury bills lost 16% after inflation, over the 20 years.

Whenever you look at long term data that shows the huge margin by which stocks have beaten bonds, it is wise to remember that a huge chunk of that came from the 15 years after 1949.

During this period there was moderate inflation, in contrast to the deflationary 30's. Long-term bond rates did not appear to reflect an expectation of even moderate inflation. Stocks were able to keep up with inflation, while bonds got hammered.  The post war years also saw unprecedented gains in productivity and the birth of the consumer society. This benefited stocks, hugely. We should not expect these factors to be repeated in future.

 

Ouch! 1966 - 1985 was an ugly time to be an investor, no place to hide. Note that the scale extends to $8.00 so that the graph can be easily and properly compared to the 1946-1965 graph above. These were the really big inflation years and both stocks and bonds as well as treasury bills had a very hard time keeping up with high inflation. It does not look like much, but stocks returned a total real portfolio gain of  53% over the 20 years while long bonds lost 7% and Treasury bills made 20%. Both stocks and bonds were volatile and both had periods where they dropped about 50%. T-bills were looking good with low volatility and reasonable returns compared to the other assets.

Next we look at the period from the end of 1985 through the end of 2005.

Wow, every form of investment is up but $1.00 in stocks invested at the end of 1985 was worth $5.29 at the end of 2005 for a gain of an even 429% (in spite of the stock crash of the early 2000's) in real after-inflation terms at the end of 2005. The corresponding figure for Long Bonds was $3.15 (215% gain) and for T-Bills is $1.36 (36% gain).

A very distinctive thing about the period of 1986 through 2005 was a huge drop in interest rates. This provided a huge boost to bond returns. It also contributed to higher P/E multiples being justified for stocks, which boosted stock returns. Another relatively unique thing about that period was the huge stock valuation bubble of the late 90's which then deflated.

In order to cover the 20 years ended March 7, 2009, our last graph will significantly over-lap the one just above.

 

In all the other 20-year periods shown above, stocks beat out bonds. However in the 20 years from the start of 1989 through 2008, (and it gets much worse when we extend it to March 7, 2009) government bonds beat out stocks. And they did so with tremendously less volatility.

Many analysts will use this last 20 years or so to conclude that a balanced position between stocks, Bonds and T-Bills is always best. I'm not convinced that it's best for everyone. If you are investing for at least 20 years, a good case can possibly be made for a 100% allocation to stocks. I would reconsider though if I thought that stocks were way over-valued at a particular point in time. At this time however, it appears that stocks are far from being in a bubble, but long-term government bonds are arguably at bubble levels and therefore it may be wise to avoid them, despite their great performance in the past 20 to 25 years.

The above graphs demonstrate that the market looks very different in different time periods and it is therefore very dangerous to make assumptions about the relative performance of stocks and bonds in the next 20 years.

Conclusions 

By studying these graphs, you can draw your own conclusions about the relative returns and risks of Stocks, Bonds and T-Bills.

Note that the return indexes ignore taxes (effectively assumes a non-taxable account) and also ignore trading costs.

Stocks (as measured by the S&P index) out-performed Bonds and Cash by an absolutely staggering amount over the last 83 years.

Stocks even out-performed over the 20 years from 1926 through 1945, in spite of the crash of 1929-1932. Bonds also did reasonably well. T-Bills were basically the after inflation equivalent of stuffing cash under the mattress.

For the 20 years from 1946 to 1965, stocks were far superior. Bonds and Bills imitated mattresses (but did protect against inflation).

The 20 years from 1966 through 1985 were ugly all around. Stocks came out slightly ahead but were the best of a dismal lot.

During the 20 years ended 2005, Stocks did very well but with high volatility, Bonds did unusually well compared to stocks and with a lot less volatility. Cash (T-Bills) continued to only slightly out-perform inflation. In the 20 years ended 2008. stocks were extremely volatile and bonds turned out to be a better investment.

A major learning from the above graphs is that the markets look very different in different time periods. It would be foolish indeed to base your investment decisions solely on the results from the last 20 years or so. Those two decades were unique due to a combination of low inflation and declining interest rates. Long-term interest rates probably won't get much lower (and have been heading up) and so the next 20 years is sure to look far different than the most recent two decades.

The above data and graphs focus on just four investment periods beginning at the end of 1925, 1945, 1965, and 1985 (plus the over-lapping period beginning a the end of 1988. Given the significant differences in the performance of stocks, versus bonds or T-bills over those four+ periods, it is also very useful to look at the comparison over all the possible 10 to 30 year holding periods beginning each year since the end of 1925. My related article does this by graphing the average annual returns over all those possible holding periods and attempts to answer the question of whether stocks are really riskier than bonds.

Originally written Summer 2001 and last updated March 8, 2009

Shawn Allen, CFA, CMA, MBA, P.Eng.

President, InvestorsFriend Inc.