I wound up asking myself the question, after I bumped into the following while working up an article on Prudential Financial (PRU). From the 10-K:
There it is: long term equity expected rate of return: 9.2%. This article explores the evidence, and leaves the answer to the reader.
S&P 500 Total Return Index
The S&P 500 (SPY) sports impressive returns, when dividend reinvestment is considered.
From June 1986 to January 2012 the S&P 500 Total Return (SPTR), which includes reinvested dividends, appreciated 9.16% per year. The period is arbitrary, but includes the October 1987 Crash as well as two severe down markets occurring in this century. The data can be downloaded from the CBOE website. It rounds to 9.2%, and could be cited in defense of Prudential's expectations.
Economists in academia have looked carefully at the historical data. Most of them present real returns, leaving the end-user to provide an inflation rate. Jeremy Siegel has looked as far back as 1802 and his data shows a long term geometric average of 7%, which could reasonably be adjusted up to 9% to allow for inflation.
The above was excerpted from "Estimating the Real Rate of Return on Stocks Over the Long Term," a collection of papers by John Y. Campbell, Peter A. Diamond and John B. Shoven presented to the Social Security Advisory Board in 2001.
The papers, while dated, are worth reading. The authors suggested that market level, which was still elevated in the wake of tech bubble, should be considered when establishing a long term (75 years) rate of return. The point is, the 7% real annual rate isn't realistic unless you start with the market at a mid-point value.
At the time, Peter Diamond estimated that stock prices would need to decline some 55% before a 7% real return going forward would be realistic.
3M (MMM), Johnson & Johnson (JNJ) and Procter & Gamble (PG) like other companies that have legacy pension exposure, report their expected rate of return on the 10-K. MMM reports that their plan assets, diverse but including both private and global equities, returned 8.1% and 10.4% for the past 10 and 25 years, respectively. JNJ has used a 9% rate of return for plan assets for the past 3 years. PG cites a range for equity returns, 8% to 9%.
My impression is that most major company pension plans are using an expected rate of return of 8.5%, to include a large proportion of fixed income securities at rates averaging substantially less. This implies expected equity returns in the 9% area.
The long term average rate of inflation in the US, from 1914 to 2010, was 3.38%. Adding that to the above real returns, nominal rates of return around 9% are indicated.
Investors who put money into the market at generational high points like 1929 or 1999 will not achieve anything like the long-term average rate of return. Similarly, long term returns will look extraordinarily good if the starting point is a low such as occurred in March 2009. It isn't possible to discuss long-term returns without developing a clear idea of current market level.
Eddy Elfenbein cites an average P/E (since 1950) of 16.4. Jeremy Siegel has argued that under modern conditions an average P/E of 18 is in order. Most writers use 15 as an average or typical figure. The most conservative estimate I can locate is 14.5 (Shiller and Campbell 2001), roughly the inverse of the 7% long-term average real rate of return. Today's multiples are atypical: the market is under compression.
In any event, TTM earnings for the S&P 500 were $97.05 as of the end of 2011. Multiplying that by 14.5, a conservative midpoint valuation of the S&P 500 would be 1,363. My estimate currently vacillates between 1,400 and 1,600.
Thinking About the Future
History is fine, but it's a good idea to consider forward looking estimates to determine if historical norms are likely to continue, and if any current outliers are likely to reverse. As a way of thinking about this, I developed the following spreadsheet:
There are several assumptions that merit discussion - assumption consistent P/E, margin reduction, and dividends/buybacks.
On margin reduction, corporate profits as a percentage of GDP are as high as they have ever been. They can be expected to decline toward average levels: however, wages are unlikely to reclaim all of the lost ground. With that in mind, a moderate reduction in margins is the most likely outcome.
Normal P/E is more difficult. Informed estimates vary from 14.5 to 18, as discussed above under market level. It makes sense to use a P/E that is consistent with the other assumptions: if inflation is high, the P/E should be low, for example. Looking at a world where inflation is low and steady, GDP growth is moderate and steady, companies are using their cash to increase shareholder value, and labor is regaining its fair share of the pie, all is well and a 16.4 P/E fits.
Many large corporations have cash well in excess of what is required to conduct their business in a prudent and orderly manner. Analysts frequently point out that P/E's are low when the excess cash is considered. If it is systematically returned to investors as either dividends or buybacks, P/E's can be expected to rise accordingly.
The following quote from Campbell's paper is relevant:
To the extent that expected future equity returns are not constant, but change over time, they can have perverse effects on realized returns. Suppose for example that investors become more risk-tolerant and reduce the future return that they demand from equities. If expected future cash flows are unchanged, this drives up prices and realized returns. Thus an estimate of future returns based on average past realized returns will tend to increase just as expected future returns are declining.
This paradox works both ways. Investors who give thought to expected future rates of return and relative market level can do much to improve their long term results, particularly if they are in the accumulation phase.
My thoughts on the 9% question: 7% nominal is more likely from where things stand right now. If the S&P gets back to the 1,250 level, 9% would be more realistic. Margins and inflation are both soft points.
Coping with Volatility in Retirement
From the foregoing, it is evident that historical averages center around a real return of 7% for US equities, and that such returns are possible going forward, provided an appropriate entry point is selected.
The extremes of volatility seen over the past 5 years create a challenge for retirees who need the yield available from equities but have limited ability to absorb volatility related losses. The smaller the nest egg in relationship to the required living expenses, the more acute the problem.
The high road, although it does require substantial assets, is blue-chip dividend payers. If dividend payments are sufficient to fund the bulk of required living expenses, market fluctuations can be tolerated.
For those less fortunate, strategy involves holding a portion of retirement assets in stable-value investments, sufficient to fund withdrawals at times when the sale of equities would be ill-advised, and sufficient to enable the purchase of equities when market levels are consistent with the desired rate of return going forward.