What will an investment portfolio earn over the long term? That issue is important to individual investors, state pension agencies and corporations offering defined benefit pensions. State pension agencies have been lowering their assumed returns. A decade ago, 8.0 percent was the dominant assumption, with some states higher and some lower. Now the most common assumption is between 7.0 and 7.5 percent. The sub-seven assumption was never used as recently as 2011 but is now embraced by several pension authorities.
What assumption should a family, a government agency or a corporate pension fund use? For a long time, it's been best to go back to the long-term averages, but the current outlook is less rosy. I personally have revised down the estimate I use in planning the Conerly family's spending and saving, and I concur with public bodies who do the same. I'm not fully convinced that I'm right; I just think the pain from erring on the low side will be less than the pain of erring on the high side.
The traditional approach is to look at long-run returns, and the book of numbers for that analysis is the SBBI Yearbook covering stocks, bonds, bills and inflation (hence the SBBI name). This research is based on pioneering work done by Roger Ibbotson and Rex Sinquefield.
Dr. Bill Conerly based on data from SBBI Yearbook.
Long run investment returns.
Since 1926, when their dataset begins, U.S. common stocks have rewarded investors by 10 percent per year, counting capital gains and dividends, before taxes. Corporate bond returns averaged 5.6 percent returns. An investment portfolio split 50 percent in stocks (the Standard and Poor's 500) and 50 percent in corporate bonds would have earned 8.3 percent per year over 1926-2016. That justifies a long-run expected return around 8.0 percent as was common.
But don't stop reading yet! Remember two important points. First, past returns are not guaranteed in the future. Second, even if the past points the way to the future, the past includes whole decades with negative returns to stocks, albeit just slightly negative.
On the first point, the structure of the economy has changed substantially since 1925 when the good data begin. Jeremy Siegel in his book, Stocks for the Long Run, shows stock market data back to 1802. He finds a seven percent annual return from 1802 through 1925. This suggests that we cannot take investment returns fixed in stone; they can be higher or lower over long periods. (Siegel's book is one of my top two picks for the average person making investment decisions. The other is Burton Malkiel's A Random Walk Down Wall Street.)
Stock market returns have been pretty good recently. Look at the S&P 500 since 2012 (counting capital gains and dividends, before taxes):
But high returns can be due to overly optimistic speculators rather than economic fundamentals. We know that economic growth has been below normal in recent years. We also know that interest rates have been well below long-run averages. That suggests - but does not prove - that returns on capital are lower now than in the historic average.
Low returns on capital might trigger a stock market gain in the short run, as lower interest expense makes corporate profitability look better. But in the long run, stock market returns must reflect the returns of investing capital in a business. So if low corporate bond interest rates today reflect low returns on capital, then stock market returns should be low in the future.
The story for low returns on capital now is simple: much of our new production requires very little capital. A steel mill or car factory requires lots of capital. A Google or Facebook requires far less. With less need for capital, the owners of capital will earn lower returns. And the global supply of savings is rising, partly due to aging baby boomers around the world and partly because a larger share of world income is in countries with weak social safety nets. I provided more detail in "Returns on Capital - And Interest Rates - Will Be Low In The Future."
The second caution mentioned above is the tremendous variability of returns. The long-run average for stocks may be ten percent, but the entire decades of the 1930s and the 2010s had negative returns. An investor ended a ten-year period with fewer dollars than at the beginning. And don't forget the spectacularly bad years: 1931, -43 percent, and 2008, -37 percent. The long-run average tells you little about next year's return. If the next bad decade starts just as you retire, you may feel pretty uncomfortable waiting for the long-run average to return. And if you can't stomach the occasional bad year, then you're likely to shift into a low-return investment when the stock market rebounds.
If I have to make a best guess as to how the next 100 years will look, I roll with the long-term average and say that stocks will return about ten percent. But I have arranged my personal affairs so that long-run returns can be much lower and I'll still be able to eat.