Long Term Risk and Return
Between 1926 and 2019, $1 invested in the S&P 500 with dividends reinvested would have produced a cumulative real return of $643. In comparison, $1 dollar invested in risk-free Treasury Bills would have grown to only $1.50 in today’s purchasing power. Equities dominated hands-down. Their long-term superiority is undeniable. The annualized total real return on the S&P 500 during this period was 7.1%. The annualized real return on the Treasury Bills was less than 1%. The difference is the so-called “equity premium,” the long-term reward for taking the risk of investing in equities.
The fact that there is a measurable, significant long-term premium for investing in equities is a fundamental building block for all investors. But what this data doesn’t tell you is how rough the ride has been along the way. Any investor who lived through the plunge in 1929, the decline in 1973-1974, the collapse in 1987, the bursting of the technology bubble in 2000, the credit crisis in 2008-2009, and the market crash brought on by the novel Coronavirus in 2020 knows that from time to time market mayhem will hammer portfolio returns. While at times volatility chips away at a portfolio through long painful downturns (for instance, Japan after 1989’s near triple digit Nikkei Price-to-Earnings ratio), statistically significant negative returns are almost always short-term, self-correcting events. Instead, the real risk to your portfolio is the failure to fulfill long-term consumption goals. This can occur as the result of one or several factors: severe prolonged inflation, economic depression, and a redistribution of wealth through taxation. While it’s impossible to completely insulate against all of these things, it is wise to design your portfolio with these real-world possibilities in mind, and also to have realistic return expectations. In other words, as William Bernstein succinctly points out in “Rational Expectations: Asset Allocation for Investing Adults,” the economic scenario and the rewards for risk that you expect to play out in the future are not necessarily what you are going to receive.
The Equity Premium
Assumptions about the equity premium drive allocations between stocks and bonds, as well as investment decisions among asset classes within those segments. Assumptions about the equity premium also reflect what returns investors can expect from equities and what the future risk-reward tradeoff is likely to be.
Because the equity premium has swung widely over periods of one, ten, twenty, and fifty years, the question of what equity premium we can expect is a source of controversy. Judgments about it should be based on the full extent of financial market history, not only in the United States but other countries as well. If long-term historical average returns are reasonable estimates of expected returns, the implication for long-term investors is rising real spending levels over time. When more conservative adjusted means are used, investors may have to adjust their future wealth expectations.
Prior to the end of the technology bubble in 2000, the most widely cited US source of the long-term US equity premium was Ibbotson Associates. Ibbotson Associates risk premium estimates are based on data from the Center for Research in Securities Prices (CRSP) whose data history starts in 1926. Based on CRSP data from 1926-2000 Ibbotson estimated that the annualized US equity premium relative to US Treasury bills is 7.3%.
In the spirit of using the longest possible data to analyze expected stock returns, Fama and French (Equity Premium, 2002) analyzed returns on the S&P 500 Index from 1872 – 2000 and calculated 5.57% for the equity premium. The average annual equity premium reported in the Fama French paper is approximately 1.75% lower than the equity premium reported in the Ibbotson study. Some of the difference arises from the choice of time frame. The Fama French study dates back to 1872 and includes the pre-1926 period for the United States when returns were lower, partly due to events leading up to, and including World War I. A comparison between the 1872 –1950 period and the 1950 – 2000 period in the Fama French research study makes the point that the equity risk premium could change over time. The equity premium for 1872 to 1950 is 4.40%. In contrast the equity premium for 1951 to 2000 is 7.43%. The difference in returns between these two periods is significant.
Jim Davis (DFA) says there are two schools of thought on how to explain the variation in returns. Some attribute it to rational variation in response to Macro Economic factors (Fama French, 1989), while others judge that irrational swings in investor sentiment are the prime moving forces (Schiller, 1989). Whatever the reason for variation in expected return, and whether it is temporary or partly permanent, the message from end-of sample dividend and earnings yield ratios in the Fama French research is that although the US economy has been remarkably successful, it may face a period of low expected returns.
To help put the US equity premium record in perspective, Elroy Dimson, Paul Marsh and Mike Staunton (London School of Economics) compiled a database of long-run international returns including reinvested income for 16 countries over the 103-year period from 1900-2002. In “Triumph of the Optimists” (Princeton, 2003) Dimson, et al. cover the U.S., the U.K., Japan, France, Germany, Canada, Italy, Spain, Switzerland, Australia, the Netherlands, Sweden, Belgium, Ireland, Denmark, and South Africa. The authors show that some historical indexes overstate long-term performance because they are polluted by survivorship bias and that long-term stock returns in most countries are overestimated, due to a focus on periods that in retrospect are known to have been successful. Over the entire 103-year period, the annualized equity risk premium, relative to T-bills, was 5.3% for the United States, and 4.40% for the world index. The US equity premium reported in the Dimson et al. study has a standard error of about 2% and is similar to the equity premium reported in the Fama French study. In the global study Dimson et al. show that only Australia, France, Italy, Japan, and South Africa had higher equity premiums than the US over the 103-year period. Australia led the field with a 6.8% premium, while the lowest premium was 2.2% for Belgium. Of the three market leaders by market capitalization in 1900 (US, UK and Germany), only the US managed to outperform the world during the subsequent period.
The Forward Looking Premium
Will the forward equity premium on the S&P 500 index deliver 5% to 6% real returns as it did over the past century? Or will the equity premium be closer to 3% or 4% as some analysts currently suggest? The answer provides direction not just for allocation decisions, but for how much and how long investors need to save in order to reach their goals.
Ibbotson and Chen (Source) estimated the forward-looking equity risk premium by constructing a supply-side earnings model with data from 1926 to 2000. They factored in inflation, real earnings growth and the historical dividend yield. Their conclusion: the forward-looking equity premium is 3.97% annualized.
Fama and French (2002) use fundamental (dividends and earnings) growth models to estimate expected returns. Fama and French estimate the expected equity premium from the fundamentals to help them judge whether the realized average return was high, meaning the market was too kind to investors, or low. The logic of their approach is motivated by the Gordon Equation, which fundamentally says that the expected equity premium can be expressed as a sum of the current dividend yield plus the historical real dividend growth rate (Bernstein, “Rational Expectations”). The estimate of the expected premium for 1872 – 2000 from the dividend growth model was 3.54%. The estimate from the realized average return, 5.57%, was almost 60% higher. The difference between the two is largely due to the last fifty years. The equity premium for 1872-1950 from the dividend growth model, 4.17% per year, is close to the estimate from the average return, 4.40%. In contrast, the equity premium for 1951-2000 produced by the average return, 7.43% per year, is almost three times the estimate, 2.55%. What accounted for the difference? Fama and French reasoned that the unexpected higher average 7.43% return on the index from 1951-2000 was due to a positive valuation change rooted in a decline in the discount rate that produced larger than expected capital gains, and historically low dividend yields and earnings-price ratios at the end of the period. Specifically, the dividend yield fell from 7.18% at the end of 1950 to a historically low 1.22% at the end of 2000, as evidenced by a commensurate six-fold increase in the price-to-dividend valuation ratio leading into the Dot-Com market collapse.
This raises the question of whether or not there is a way to profitably time the market by looking at valuation metrics. Some analysts’ look to the Cyclically Adjusted Price Earnings Ratio (CAPE) that Professor Robert Schiller helped create. This is a measure that enables the comparison of stock market valuations from different eras by averaging the earnings over 10 years, thus reducing some of the short-term fluctuations of each market cycle. As of this writing (December, 2020), the CAPE stands at a level that was higher in only two periods, both of which were followed by stock market crashes: the 1920s, in the lead-up to the Great Depression; and early 2000, just before the bursting of the dot-com bubble. So how do we use the CAPE to estimate returns? It all depends on your timeframe and the pace of adjustment. Over the past 60 years, the CAPE has averaged about 20 times. The current CAPE of about 30 times implies that the S&P 500 is overvalued by about 50%. In order to get valuations back in line with historical long-term average valuations, either earnings would have to grow over the next decade at double the historical rate, or the stock market would have to fall about 33%.
In estimating dividend and earnings growth rates, Fama and French found that they are largely unpredictable, and concluded that historical average growth rates are the best forecasts of future growth. Based on this logic, given that the current dividend yield on the S&P 500 index is 1.8% and assuming a 2% real dividend growth rate, it stands to reason that the prospective equity premium for the S&P 500 index is 3.8% before taking into account the risk of valuation mean reversion. As explained by William Bernstein (“Rational Expectations: Asset Allocation for Investing Adults”), if you are looking out over 30 years, CAPE reversion to the mean annualizes out to a lowering of expected returns by about 1%. If you are looking out just 10 years, this annualizes out to a lowering of expected returns by about 3%. This lowers our 3.8% equity premium forecast on the S&P 500 to 0.8% and 2.8% over 10 and 30 years respectively, which is far lower than the historical actual average equity premium of around 5% to 6%.
Clearly, there is a relationship between the CAPE and forward long-term returns, but can you make money off this? Probably not, says Bernstein. The reason is that valuation metrics are not stationary enough to be of any practical use. To understand this better, Dimson et al. examined the performance of their favored valuation metric, the PD10, which is similar to Schillers CAPE except that it uses dividends. They found modest evidence of returns predictability with their “in-sample” database, but when they tested their results “out-of-sample,” predictability dropped to zero. As discouraging as this may sound, Bernstein points out that strategies which lower overall equity allocations at high valuations will almost certainly lower risk. This would seem to make sense for older investors with little human capital or savings potential left, not so much for return enhancement as for risk reduction. On the other hand, for younger periodic savers, adjusting overall equity exposure according to valuations makes little sense.
What does make sense is that all investors will likely benefit from international diversification, and the opportunity to lift expected portfolio returns back to long-term historical levels by systematically tilting portfolios toward more profitable securities and to more promising investment options like small cap and value stocks, two asset classes with above-market expected returns. For more discussion about tilting portfolios toward additional dimensions of expected returns, see Sources of Expected Returns elsewhere on this site.
Long-Term Assumptions May Not Hold
Stocks are thought to perform so much better than bonds over the long run. Is that really true? In “Stocks for the Long Run,” Jeremy Siegel analyzes US stock returns from 1802 to 2001 and makes the case for US equities. According to Siegel’s analysis, the real return on US equities averaged 6.9% percent per year over the past 200 years. This means that purchasing power in the stock market nearly doubled every 10 years. According to Siegel, US stocks have never fallen behind inflation, whereas bonds and bills have fallen as much as 3% per year behind the rate of inflation when the time interval is 20 years or longer. A second conclusion is that for all investment holding periods of around twenty years or more, US equity premiums have always been positive or within a fraction of a percentage point on the negative side of zero. And third, over 20-year horizons, US stocks beat bonds and bills over 90% of the time.
To what extent should we rely on such patterns persisting in the future? In “Triumph of the Optimists,” the authors (Dimson, Marsh and Staunton) present evidence that spans more countries. They show that the equity premium is not constant across markets. And more often than not, they show stock market volatility has been greater in other countries than the United States. From their global perspective, and looking forward to the future, equities are far from risk-free over the long-term. To focus on the international dimension, Dimson et al. looked at the experiences of all 16 countries. Only when they look at periods of 40 years or more does the equity premium turn consistently positive for the Netherlands. Several countries studied in “Triumph of the Optimists” went through even longer periods before showing consistently positive premiums. In fact, for more than half the 16 countries, including Germany, Sweden and Switzerland, there was a greater than 10% chance that the 20-year equity premium would be negative. Poor long-term performance for these countries can be attributed to a variety of social, economic and political factors. For investors, the point is that a twenty-year investment horizon in these countries was far from risk-free. Diversification isn’t a panacea, but by holding a globally diversified portfolio investors can reduce the portion of variance tied to the volatility of a single country, and lower overall portfolio volatility.
What does the future hold for the equity premium? Dimson et al. estimate the prospective equity premium for their sixteen-country world index to be approximately 3 percent annualized. They examined the range of expected risk premium that can be anticipated over various future time horizons. In “Triumph of the Optimists” Dimson et al., retain Siegel’s interpretation that 20 years may be considered a long time, and concluded that based on their assumption of a forward-looking equity premium of 3% and applying a standard deviation of 16% (typical for the United States and UK), there is an 18% probability that global equities will underperform T-bills over 20 years. What happens when the researchers applied a standard deviation of 24%? That standard deviation would not be unrealistic for a poorly diversified portfolio, or if the world became a riskier place in which to do business. With this higher margin for error, there is a 17% probability that equities would underperform T-bills even over 50 years. As the standard of error increases, the assumption doesn’t hold that single country stock returns are always triumphant over longer periods. There is clearly a substantial probability of achieving a negative premium, even over longer investment horizons.
To sum up, history and theory confirm that the expected return on equities exceeds the expected return on fixed income. Yet downside risk in equities is always present, and at times can be extreme. The first line of defense in risk reduction is diversification. Investors reduce the risk of equities in portfolios by diversifying, which means spreading the risk among domestic and foreign equities, large companies and small companies, growth stocks and value stocks, real estate securities, etc. But given the volatility of stock returns, even a highly diversified equity portfolio may have higher volatility than is desired by the investor. Investors are then advised to adjust total risk in the portfolio by adding fixed income to their asset allocation. Adding fixed income, particularly short-term bonds, to an equity portfolio will reduce the overall portfolio volatility. For more on portfolio design, see Asset Allocation elsewhere on this site.
How Should Assumptions About the Equity Premium Guide Investors?
Following are some important take-aways from the accumulated research:
Base Your Investment Strategy on Realistic Assumptions. Equities are likely to outperform risk-free T-bills in the long-run, but current dividend yields suggest that the forward-looking equity premium will be lower than in the past (how much lower is a matter of debate). With a smaller risk premium, the opportunity cost of being out of stocks might look smaller. The reward from risk for society as a whole might therefore be lower in relative terms than was previously thought.
Tilt Allocations to Profitability, Value and Small Cap Stocks. Stock market researchers have identified a number of persistent patterns in stock returns. We know that profitability, and the small and value effect, provide better returns in the long run than the broad market. There is little doubt that these phenomena contribute to explaining stock returns. Skilled investors can therefore buttress their portfolios against the risk of diminished long-term broad market returns by deviating from the market benchmark by targeting profitability and taking on more size and value risk in their portfolios.
Reduce Risk through International Diversification. Investment risk is lowered in worldwide portfolios. Global investors can therefore afford to allocate relatively more of their portfolio to risky assets, such as equities. Since riskier assets have a positive expected reward for risk, investors in globally diversified portfolios have a higher expected return in relation to their risk.
Control Investment Costs. A declining equity premium argues for careful control of costs, including advisory fees, management fees, and trading fees. Investors who fail to diversify efficiently and who overpay for investment management services have a lower expected return than those investors who control these costs. See discussion about our low fixed fees elsewhere on this site.
Manage Taxes. The risk of a lower forward looking equity premium also argues for a tax aware approach, a hallmark of passive and index investment strategies. Those who do pay tax suffer a dead weight burden on their stock returns. That burden can be large. Dimson et al. reference Siegel and Montgomery (1995), for example, who estimate that the twentieth century annualized real return on US equities is roughly halved if the investor is taxable at the maximum rate on income and capital appreciation.
Carefully Explore Assumptions. There is no guarantee that the past is an indicator of future results. And many asset pricing models predict a much lower equity premium than observed in historical returns. This implies that equities should be held as part of a diversified portfolio, including multiple asset classes worldwide.
Cardiff Park simulates portfolio performance against a range of equity premium assumptions. For many investors, knowing the worst that can happen to their portfolio provides clarity about the direction they should take. If long-term historical average returns are reasonable estimates for expected returns, the implication for long-term investors is that real spending levels will rise over time. If stock returns are low for long periods, investors may need to save more, and spend less.
For more information about Cardiff Park Advisors please review our brochure at https://adviserinfo.sec.gov/firm/summary/126752 or visit www.cardiffpark.com