When I retired a few years ago, I decided to delve into whether or not the stock market was overvalued. Conventional wisdom at the time suggested that the stock market had been on the rise since 2009, and maybe it was overvalued. It has continued rising since then, so where does it stand now?
Robert J Shiller is the Noble Prize-winning economics professor and author. In his book “Irrational Exuberance”, Shiller presents the time history of real (inflation adjusted) S&P Composite stock prices dating back to 1871. This time series can be found at http://www.econ.yale.edu/~shiller/data.htm. It contains monthly averages of daily closing prices and monthly dividend, earnings and consumer price index data.
Using this data, we can calculate the growth in real total value of the S&P Composite Index since 1871 to the present (Q4 2019). Real total value is just the number of shares owned times the inflation-adjusted or real price/share. The real total value calculation assumes dividends are reinvested to purchase more shares and the value of the index grows as it would in an IRA or 401k, i.e., not taxed until withdrawn. Today, several low cost mutual and exchange traded funds accurately track the S&P 500, which is the current version of the S&P Composite Index.
One share of the index in January of 1871 cost $4.44, which is $91.54 in today’s dollars. If your ancestor had bought that share and willed it to you, you would have 620 shares today, and they would be worth almost $2M. While a growth factor of 21,500-to-1 is impressive, most people have investment horizons less than 149 years! What is important to many people is how the growth occurred, steady or unsteady.
The Steady Growth Perspective
The trajectory of growth is shown in Figure 1. The vertical axis is logarithmic, which proportionately displays percentage changes regardless of index level, e.g., increases from $100 to $110 and from $1M to $1.1M each represent 10% growth, and each change is the same vertical distance in the figure. Another attribute of using a logarithmic vertical axis is that a constant rate of growth is represented by a straight line. The dashed line shows the mean growth line, which is an annualized real growth rate of 6.9%. The value R2 measures how well the mean line represents the data, and R2 here is 98.6%. The passage of time explains 98.6% of the data. From this perspective, American industry as measured by the S&P Composite has been a remarkably steady growth generator over the last 149 years.
Figure 1 Growth in Real Total Value of S&P Composite Index
This steadiness is not due to a lack of underlying changes. Shiller’s dataset (Figure 2) shows that dividend yields have been declining over the decades, averaging around 5% in the late 1800’s while averaging about 2% today. Fewer dividends means fewer shares in the total real value calculation. The decline was gradual prior to the early 1980’s, a period that coincides with the imposition of taxes on dividends, making dividends less advantageous to shareholders who invest for current income. Dividends were not taxed until 1936 (https://www.dividend.com/taxes/a-brief-history-of-dividend-tax-rates/). By 1954, dividends were taxed as ordinary income, with the maximum rate as high as 90% until 1985. Today, qualified dividends are taxed at a 15% rate (20% for individuals with incomes over $400K and households with incomes over $450K) while unqualified dividends are taxed as ordinary income.
Figure 2 Dividend Yield Has Declined Over the Years Until Holding Steady Since 2010
The decline in dividends accelerated after the 1980’s, a period that coincides with the advent of stock buybacks. Buybacks raise a stock’s price since a company’s earnings are spread over fewer shares. In 1982, the Securities and Exchange Commission established Rule 10b-18, which provided corporations doing buybacks protection from lawsuits. The New York Times (https://www.nytimes.com/2018/08/23/opinion/ban-stock-buybacks.html) studied 232 companies in the S&P 500 that were publicly listed from 1981 through 2016. In the early 1980’s, these companies spent 4.3% of profits on buybacks. From 2014 through 2016, these same companies spent 59% of profits on buybacks.
More recently and from 2015 through the first quarter of 2019, Standard & Poors reports (https://us.spindices.com/indices/equity/sp-500) that on average, S&P 500 company’s spent 99.6% of operating earnings on buybacks (59.4%) and dividends (40.3%). Over the same period, S&P 500 total earnings grew at an annualized rate of 12.8%, but earnings per share grew at a 14.5% rate. Buybacks bolstered the earnings growth realized by shareholders, which then induced higher stock prices.
Table 1 contrasts early and recent years of the S&P Composite Index, the late 1800’s and from 2015 through the 1stquarter of 2019. The dividend payout ratio is the ratio of dividends to earnings. This ratio has significantly declined, from 74% to 46%. The dividend yield (dividend/price) has declined with it, from 5.2% to 2.0%, but the buyback yield has gone from nothing to 2.9%, so when you compare total yields, they remain nearly the same at about 5%.
Table 1 Contrasting Early and Recent Years of the S&P Composite Index
Table 1 also shows that we are paying more for earnings today with the price-to-earnings ratio rising from 14.3 to 22.8. This is fallout from the declining dividend payout ratio and yield, and it is also the result of stocks today being a relatively much higher return on investment than government bonds. In the early years with 10-year government bonds (GS10) at a 4% yield and deflation at 1.7%, 10-year government bonds had a real return of 5.7% compared to the Composite Index’s total real return of 8.3%. Bonds were lower by 2.6%, but they do not incur state income tax, and there was little or no principal risk if the bonds were held to maturity.
This trade is much different in recent years. GS10 yield has averaged 2.3% while inflation has been at 1.9%, leaving a real annualized return of just 0.4%. At 2%, the S&P 500 dividend yield has been almost as high as GS10’s yield. Investors realize almost as much current income investing in stocks as they would in bonds, and the S&P 500 total real annualized return has been 7.7%. This 7.3% difference tilts the trade much more in favor of stocks.
The Unsteady Growth Perspective
Another perspective is that growth has not been steady. Figure 1 shows that the Roaring 20’s, post-World War II period and the Internet Bubble were all times of rapid growth driving total value higher. The Great Depression, 1970’s oil crises, early 2000’s bursting Internet Bubble, and the Great Recession were times of rapid real total value contraction.
Figure 3 shows this in even more glaring detail. Figure 3 is the percentage difference (mean value in the denominator) between the real total value and the mean real total value for every month from 1871 to the present. There are swings of 140% or more around the Roaring 20’s and Great Depression, the post-World War II period and the depths of the 1970’s Oil Crises and high inflation 1980’s, and the Internet Bubble and Great Recession. Taken together, these two perspectives show that over the long term the average monthly pricing of the Composite Index is inefficient.
Figure 3 Real Total Value About the Mean Illustrates Market Inefficiency
The Current Level of Market Prices
The unsteady growth perspective shows that the market level at the end of 2019 was 21.2% above the mean line. The steady growth perspective shows that since the beginning of 2015, S&P 500 real annualized total return has been 9.3% compared to the long-term average of 6.9%. Both measures indicate the market is hot. Note that the market gets really hot and peaks (1929, 1966 and 2000) about every 33 years, or about once every human generation. It would seem people forget and display, per Shiller, “Irrational Exuberance”.
Will the next peak occur in the early 2030’s, or earlier? I will address the statistical prospects going forward from here in a later article.
Feels like we are experiencing some “irrational exuberance” in more ways than in the stock market. Does your model give ample warning to move to bonds before the bottom drops out of the market? Won’t the company buyback of stock need to change if they are spending all of their earnings on buyback? Did the recent $600.00 drop show up in your model before it hit the market? What percentage of drop constitutes movement and how do you know when to move back into the stock market? If unemployment is the only thing that will determine weather Trump gets re-elected then it looks like there is not enough time for the employment level to fall to ensure that he doesn’t?
Taking the questions one at a time:
The data presented here looks at the historical record, and it does not serve as a prediction. It identifies level. I do plan on sharing additional historical statistics in a later post.
In general, Companies would have to modify their behavior if earnings declined, but not always. Some companies borrow to fund buybacks and dividends. The key over the long run is whether or not that have enough earnings to fund capital and R&D expenses, dividends and buybacks. My takeaway from looking at the historical data is that companies can return about 7% in real terms over the long run one way or another: buybacks, dividends and price increases.
The unemployment rate would have to go to 3.8% or higher by next October. There is still time for unemployment to turn up, but it gets less likely every month that it stays about where it is now, 3.5% in November and December. January data is released this Friday. It looks like Republicans are solidly behind Trump, so he is likely to receive a very strong first ballot vote at the convention.