This week's letter is by Myles Zyblock, who is Chief Institutional Strategist & Director of Capital Markets Research at the Royal Bank of Canada. I have been reading Myles for a number of years, when he was first at another firm whose quality of work has dropped since he left.
Myles takes a look at what the next ten years might hold for investors in the equity markets. Using both PEs and an accounting based approach he concludes that investors will be very disappointed if they expect returns over the next ten years to be anywhere close to those of the late 1990's.
This is another example of research that supports my belief that we are in a secular bear market for the next 5-10 years and why I picked it for this week's Outside the Box.
John Mauldin, Editor
Outside the Box
Peering into the Next Ten Years: You Aren't Going to Like What You See
Investors who have held onto stocks since the October 2002 lows have been handsomely rewarded, as the S&P 500 and TSX have risen by 48% and 64% respectively. The important question for investors to ponder is whether this powerful bull market can last and, if so, for how long? History argues that rallies starting from low valuations tend to endure, while those starting from high valuations are usually given back and sometimes in anguish. Valuations have come down significantly over the past several years but still sit at the high end of their 100-year range. If we are on the cusp of the next secular bull market, it would be an unprecedented development given current valuations.
So, what can North American investors expect out of the stock market over the long term? Let's start with the next decade. There are a number of ways to approach this question. In this report, we focus on two of the more prominently featured methods found in finance literature. The first is based on the historical relationship between P/Es and future stock market returns. The second uses an accounting-based methodology, which breaks the return on investment into its different components. The findings based on these two methodologies are very similar, and are briefly highlighted below:
- P/Es are strongly and inversely related to subsequent long-term returns. With the help of a simple regression model, we estimate that the S&P 500 at current valuations is priced to deliver about 0.4% in total return terms per annum over the next decade. Unfortunately, we could not conduct a similar exercise for the Canadian market due to the lack of historical data.
- An accounting-based approach delivers an expected annualized total return of 2.7% for the S&P 500 and 2.9% for the TSX Composite over the next decade. Here we assumed that dividends would continue to grow at a trend pace, and that dividend yields would rise in the next 10 years up to where they were a decade ago.
The bottom line is that we are most likely going to see a substantial return shortfall in the broad North American indices over the next decade relative to what investors have become accustomed to. Buy and hold strategies and indexing will probably deliver sub-par returns. In contrast, astute stock picking and concentrating the number of stocks held in portfolios should provide the best opportunity to enhance returns in future years.
Placing this Rally in an Historic Context
Equities have risen sharply since the bear market lows set in October 2002. Investors who have held onto stocks over this two and a half year period have been handsomely rewarded, as the S&P 500 and TSX have risen by 48% and 64%, respectively. The important question for investors to ponder is whether this powerful bull market can last and, if so, for how long? History argues that rallies starting from low valuations tend to endure, while those starting from high valuations are usually given back and sometimes agonizingly so.
*Normalized earnings are derived by taking a 30-year moving average of real reported earnings.
For example, the last major US equity bear market started in February 1966 when P/Es were at the high end of their historical range. This bear market lasted until August 1982. There were four very powerful rallies over this period when the market was up 32%, 66%, 76% and 38%. But, the index lost close to 50% of its value in real terms over the entire period. In the meantime, the P/E was taken down from the high end to the low end of its historical range, thereby setting the stage for the most impressive secular bull market in modern history.
From 1982 to 2000, earnings increased by 120% and stock prices rose six-fold in inflation-adjusted terms. Valuations were pushed up to a level not seen in more than 100 years. In the chart above, we see that valuations have come down significantly over the past several years but still sit at the high end of their historic range. It is doubtful that the rally from October of 2002 marks the beginning of the next secular bull market. It would be an unprecedented development if it were indeed the case.
So, if we are not on the cusp of a major bull market then what are we facing? What can investors expect out of the stock market over the long term, say over the next decade? There are several ways to approach this question, and we will focus on a couple of the methods featured more prominently in finance literature. The first is based on the historical relationship between P/Es and future stock market returns. The second uses an accounting-based methodology, which breaks the return on investment into its components. Both methods rely heavily on the assumption of mean reversion.
High P/Es Beget Low Future Returns
Numerous academic studies have found that the single most powerful force driving markets over the long-term is valuations. This relationship is highlighted in the following chart for the S&P 500, where real normalized price-to-earnings are plotted against subsequent ten-year real stock price returns (annualized) using yearly data back to 1901. Unfortunately, we were unable to find a long enough time series to properly do the work for the Canadian market. Each variable was deflated by the CPI in order to get real values. And we calculated a 30-year moving average of real reported earnings to normalize earnings.
*Based on a model that regresses price-to-normalized real earnings against subsequent 10-year real stock market returns.
In the previous table, we show in greater detail the relationship between valuations and future long-term real returns based on estimates from our regression work. If, for example, the stock market was trading at a normalized P/E of 15, then a best guess for real returns would be 2.5% per annum over the next decade. In actual fact, the S&P 500 is trading at a normalized P/E ratio of 35. Based on this model, stocks are estimated to decline by 3.6% per annum in real terms. Grossing this result up by adding in some inflation (say 2%) and a dividend yield (say 2%), we arrive at 0.4% total returns per annum for the next decade... uninspiring prospects, to say the least.
Bottom Line: Based on the relationship between P/Es and future returns, our work suggests that the S&P 500 will return something in the order of 0.4% per annum over the next decade.
Any way we Slice it, Long-term Return Prospects Look Unattractive
We arrive at similar conclusions to those presented above using an accounting-based approach to estimating long-term returns. With this methodology, we are able to derive and compare long-term return prospects for the S&P 500 and the TSX Composite, using data back to 1956. Let's begin with the US market.
The return on investment over the long term is the sum of three parts: the income rate, the growth rate of income, and the change in how that income is valued. For example, the dividend yield for the S&P 500 was 3.8% in 1956. From 1956 to the present, companies have been able to raise dividends by 5.2% per annum, which is approximately the same as the trend growth rate for earnings. Thus, the fundamental return for the equity market over this period has been 9.0% (i.e., 3.8% dividend income plus 5.2% growth in income). However, the dividend yield has declined from 3.8% to 1.7% in these 49 years, representing an increase in the 'dividend multiple' of 1.7% per annum or, said another way, a 1.7% annual boost to total returns generated from investor speculation. In other words, the total return on investment derived from this decomposition exercise is 10.7% (i.e., = 3.8% + 5.2% + 1.7%). Not surprisingly, this compares closely to the actual total return over the period of 10.5% per year.
This type of work can generate enlightening insights about long-term equity market opportunities. We know that the current dividend yield for the S&P 500 is 1.7%. Let's assume that the growth of income going forward mimics the experience of the past - that is, corporations will continue to grow dividends at a 5.2% annual pace. Let's also assume, and this represents the big leap in the analysis, that the dividend yield over the next decade returns to where it was a decade ago: that is, it rises from 1.7% currently to 2.6%. So, over the next 10 years, an increase in the dividend yield from 1.7% to 2.6% would represent a valuation de-rating for equities equivalent to 4.2% per annum. Adding up current income (+1.7%), its growth (+5.2%) and the assumed change in valuation (-4.2%) provides a total return of 2.7% per year...yikes!
Under much less hostile assumptions, we still arrive at muted return prospects for the broad US equity market. For example, our analysis points to a total return of 6.9% per year over the next 10 years assuming no change in equity valuations (i.e., the current dividend yield plus the trend rate of dividend growth, only). Maybe the ultimate truth lies somewhere in between - that is between the 2.7%, which assumes a return to decade-ago valuations, and the 6.9%, which assumes unchanged valuations. That said, the analysis still points to a substantial return shortfall relative to what investors have become accustomed to over the past several years.
Long-term return prospects can't possibly be as dim for Canadian equities, right? Well, let's run through the same exercise and see what we find. From 1956 to the present, TSX companies have increased their dividends by 4.6% per annum. The dividend yield in 1956 was 3.4%. Adding the two together gives us a fundamental return for the equity market of 8.0% However, the dividend yield has declined from 3.4% to 1.7% in the past 49 years, representing an increase in the 'dividend multiple' of 1.4% per annum. Thus, the total return on investment derived from this decomposition exercise for the TSX is 9.4% (i.e., = 3.4% + 4.6% + 1.4%) versus an actual total return of 9.6% per year.
Let's assume that the dividend yield for the TSX over the next decade rises up to where it was a decade ago (i.e., to 2.4% from 1.7% currently), and that companies will continue to grow dividends at their historic pace of 4.6% per annum. The current yield of 1.7% plus the growth in income of 4.6% delivers a fundamental return for equities of 6.3% per annum. However an increase in the dividend yield from 1.7% to 2.4% represents a valuation derating equivalent to 3.4% per year. Adding together the fundamental return with the impact of the derating delivers an expected total return for the market of 2.9% per annum for the next 10 years - slightly above the 2.7% long-term return estimate we calculated for the US market.
Bottom Line: An accounting-based approach to estimate long-term returns delivers an expected annualized return of 2.7% for the S&P 500 and 2.9% for the TSX Composite over the next decade. Our calculations assume that dividends will continue to grow at a trend pace, and that dividend yields will rise over the next 10 years to where they were a decade ago.